Florida State University Libraries
Electronic Theses, Treatises and Dissertations
The Graduate School
2009
Natural Radionuclide Applications for
Riverine and Coastal Marine Investigations
Richard N. (Richard Neil) Peterson
Follow this and additional works at the FSU Digital Library. For more information, please contact [email protected]
FLORIDA STATE UNIVERSITY
COLLEGE OF ARTS AND SCIENCES
NATURAL RADIONUCLIDE APPLICATIONS FOR RIVERINE AND COASTAL MARINE
INVESTIGATIONS
By
RICHARD N. PETERSON
A Dissertation submitted to the
Department of Oceanography
in partial fulfillment of the
requirements for the degree of
Doctor of Philosophy
Degree Awarded:
Spring Semester, 2009
Copyright © 2009
Richard N. Peterson
All Rights Reserved
The members of the Committee approve the Dissertation of Richard N. Peterson defended on
January 30, 2009.
______________________________
William C. Burnett
Professor Directing Dissertation
______________________________
Joseph F. Donoghue
Outside Committee Member
______________________________
Jeffrey P. Chanton
Committee Member
______________________________
Philip N. Froelich
Committee Member
______________________________
Kevin G. Speer
Committee Member
______________________________
Stephen Opsahl
Committee Member
Approved:
___________________________________________
William Dewar, Chair, Department of Oceanography
The Graduate School has verified and approved the above named committee members.
ii
This dissertation is dedicated to my amazing wife, Amanda, and my family who have given me
constant support throughout my time in graduate school.
iii
ACKNOWLEDGEMENTS
The work presented in this dissertation would not have been possible without the help of
many people. First, I can not say enough about the encouragement and guidance that Bill
Burnett has provided for me during my time here at Florida State University. He has truly made
me the man and researcher that I am today. I thank Bill for the wonderful support that he has
given to me during the last 8 years.
I would also like to thank the rest of my committee members (Jeff Chanton, Flip
Froelich, Kevin Speer, Joe Donoghue, and Steve Opsahl) for guiding and directly my education
and research. Jeff and Flip have both opened up their labs to me and encouraged their students
to work with me on various projects. Kevin and Joe both provided an outside point of view on
my research that has been very valuable. Steve has been a great contributor to the project on the
Apalachicola-Chattahoochee-Flint River system and without his assistance, I could have never
collected a single sample.
Our work on the Yellow River delta (Chapters 3-5) was funded by both an NSF grant and
the Research Institute for Humanity and Nature in Kyoto, Japan. I would like to thank my
coauthors and field assistants for their wonderful help with this project: Makoto Taniguchi,
Jianyao Chen, Sambuddha Misra, Tomotoshi Ishitobi, Yoshihiro Fukushima, Jianzhong Cheng,
and Songqing Zeng.
Our colleagues who shared our work on the Big Island of Hawaii certainly made that
project very enjoyable. I would especially like to thank Craig Glenn and Adam Johnson for their
continued support and hard work. Also, Jacque Kelly, Carrie Plath, David Gremminger, Axel
Schmidt, and Benjamin Sellers helped make up our fieldwork teams. George Wilkins, Sallie
Beavers, William Walsh, and Robert Ravenscroft all provided essential logistical aid in the field.
This work was supported by an NSF grant.
The work on the ACF system also relied on several individuals including Axel Schmidt
for his field assistance. I am deeply indebted to the faculty and staff of the Joseph W. Jones
Ecological Research Center at Ichauway, Georgia. Lee Edmiston and his group also provided
necessary aquatic facilities for sample collection. I thank the Antarctic Research Facility at
Florida State University and its curator Dr. Matt Olney. The Geochemistry Group (especially
Nicole Tibbetts and Yingfeng Xu) at FSU assisted greatly in my mass spectrometry work.
My former and current lab mates have been vital to my work. Henrieta Dulaiova taught
me much of what I learned during my early years in the department. Isaac Santos has served as
an incredible inspiration in both the field and the office for creative thought and efficient work
and participated in all the fieldwork projects shown here. Natasha Dimova has been very
valuable to me in both the field and the lab and I certainly appreciate all that she has done to help
me. Ben Mwashote is one of the nicest people in the department and was always there for me.
I also greatly appreciate all the technical education I received from the instrument shop
guys. Dave Oliff, Arden Stephens, and Alan Michels taught me of things I never knew existed.
Anne Womack, Beth Kostka, Fran Bollman, and Michaela Lupiani helped keep my paperwork in
good order. I thank all the rest of my colleagues in the Oceanography Department for making
my time here very successful.
Finally, I never could have survived without the help from my wonderful family. My
lovely wife, Amanda, helped out many times in the field and in the lab and kept my priorities in
line. My parents (Gary and Judy) and brothers (Rob, Mike, and Mark) have also been great
friends and inspiration.
iv
TABLE OF CONTENTS
List of Tables ………………………………………………………………………
List of Figures ……………………………………………………………………...
Abstract ………………………………………………………………………….…
viii
xi
xvii
1. INTRODUCTION TO THE DISSERTATION .………………………………..
1
2. A COMPARISON OF MEASUREMENT METHODS FOR RADIUM-226
ON MANGANESE-FIBER …………………...…………………………………...
3
Abstract ……..……………………………………………………………...
Introduction ………………………………………………………………...
Materials and Procedures
Cartridge Design …………………………………………………...
Standard Preparation ……………………………………………….
RAD7 Method ……………………………………………………...
Radon Emanation Line (Lucas Cell) Method ……………………...
RADECC Coincidence Counting System Method ………………...
γ-spectrometry Method ………………………………………….…
Assessment …………………………………………………………………
Discussion ………………………………………………………………….
Comments and Recommendations ………….……………………………...
3
3
5
5
6
7
9
9
10
11
13
3. DETERMINATION OF TRANSPORT RATES IN THE YELLOW
RIVER – BOHAI SEA MIXING ZONE VIA NATURAL GEOCHEMICAL
TRACERS ………………………………………………………………………….
18
Abstract ……………..……………………………………………………...
Introduction ……….……………………………...………………………...
Methods
Sampling and Analytical Methods ………………………………....
Transport Rate Calculations ……………………………………..…
Results ………...…………………………………………………………...
Discussion ………………………………………………………………….
Conclusions ………………………………………………………………...
20
22
23
25
32
4. RADON AND RADIUM ISOTOPE ASSESSMENT OF SUBMARINE
GROUNDWATER DISCHARGE IN THE YELLOW RIVER DELTA,
CHINA ……………………………………………………………………………..
34
Abstract …………………..………………………………………………...
Introduction ………………………………………………………………...
Methods
Measurements …….………..…………………………………...….
Transport Rate Model ……………………………………………...
v
18
18
34
34
36
38
Radium Time-Series Model ………………………………………..
Radon Time-Series Model …………………………………………
38
39
Offshore Transport …………………………………………………
Radium Time-Series ……………………………………………….
Radon Time-Series …………………………………………………
Offshore Radium Distribution Assessment of SGD ……………….
Discussion
Radium Offshore Distribution ..…………………………………...
Radon Results ..…………………………………………………….
Model Uncertainties …..…………………………………………....
Comparison to Seepage Meters …………………………………….
Regional-Scale Fluxes ……………………………………………...
Conclusions ………………………………………………………………...
40
40
48
49
50
51
52
53
55
56
5. BOHAI SEA COASTAL TRANSPORT RATES AND THEIR INFLUENCE
ON COASTLINE NUTRIENT INPUTS …………………………………………..
57
Results
Abstract …………………………………………………..………………...
Introduction ………………………………………………………………...
Materials and Results
Transport of Yellow River Water Offshore ……………………..…
Nitrogen Uptake Rates ……………………………………………..
Riverine Nutrient Transport ………………………………………..
Submarine Groundwater Discharge Nutrient Transport ..………….
Discussion and Conclusions ………………………………………………..
6. QUANTIFICATION OF POINT-SOURCE GROUNDWATER
DISCHARGES FROM THE WESTERN HAWAII SHORELINE …………...…..
Abstract …………………..………………………………………………...
Introduction ………………………………………………………………...
Methods
Study Site …………………………………………………………..
Measurements ……………………………………………………...
Mass Balance Model to Quantify SGD …………………………….
Results
Kahualoa Bay ………………………………………………………
Manini Beach ……………………………………………………....
Queen’s Bath …………………………………………………….…
Kailua-Kona Harbor ………………………………………………..
Honokohau Harbor …………………………………………………
Kiholo Bay …………………………………………………………
Discussion ………………………………………………………………….
vi
57
57
58
59
60
61
63
64
64
64
65
67
68
71
76
79
79
80
81
83
7. TRACKING SUSPENDED PARTICLES WITH NATURALLYOCCURRING RADIONUCLIDES AND CHEMICAL TRACERS IN THE
APALACHICOLA-CHATTAHOOCHEE-FLINT RIVER SYSTEM ……………
86
Abstract …………..………………………………………………………...
Introduction ………………………………………………………………...
Study Area ………………………………………………………………….
Methods
Sampling Strategy ………………………………………………….
Sample Collection ………………………………………………….
Measurements ……………………………………………………...
Results
TSS Concentrations ……………………………………………...…
Long-Lived Nuclides (Radium Isotopes and 40K) …………………
Radiotracers Derived From Atmospheric Deposition (7Be and
210
PbXS) ………………………………………………………….….
Other Stable Tracers ………………………………………………..
Discussion
Reservoir Sediment Deposition ……………………………………
TSS Load Constituents …………………………………………….
Particle Ages, Fraction of ‘New’ Sediments, and Associated Transit
Times ……………………………………………………………….
Particle Tracing …………………………………………………….
Metal Fluxes to Apalachicola Bay …………………………………
Conclusions ………………………………………………………………...
100
107
110
111
8. CONCLUSIONS OF THE DISSERTATION ………………………………….
113
APPENDIX A. SIMULTANEOUS RADON BOX MODEL EQUATION
DERIVATIONS ……………………………………………………………………
116
APPENDIX B. DETAILED SAMPLE COLLECTION SITE DETAILS ON THE
ACF RIVER SYSTEM …………………………………………………………….
120
APPENDIX C. RAW DATA FROM THE APALACHICOLACHATTAHOOCHEE-FLINT RIVER SYSTEM ………………………………….
123
REFERENCES …………………………………………………………………….
144
BIOGRAPHICAL SKETCH ……………………………………………………....
156
vii
86
86
88
90
92
92
94
94
95
96
98
100
LIST OF TABLES
Table 2.1. Average preparation times and measurement parameters associated
with the sample set…………………………………………………………………..
12
Table 2.2. Important considerations (positive and negative) associated with each
measurement technique under a variety of scenarios………………………………..
16
Table 3.1. Summary of the initial AR parameters used to model the apparent
radium ages of Yellow River plume waters…………………………………………
29
Table 3.2. Summary of the linear regression results from apparent radium age
versus distance offshore plots found in Figure 3.7. Transport rates are converted
from the reciprocal of the regression slopes to transport rates of cm/s……...............
30
Table 4.1. All groundwater samples collected for radium isotopes. Samples
included in upper portion of table were collected from boreholes throughout the
Yellow River delta. Samples named “PW” are collected pore water samples from
the study site. All radionuclide activities have been decay-corrected to the time of
sampling. Uncertainties shown are at the 1-σ level………………………………...
42
Table 4.2. Radium time-series model parameter summary. The values found in the
central column between TS-1 and TS-2 are common to both time-series…………..
47
Table 4.3. Summary of the results of the SGD analyses based on radium isotopes,
radon, and automatic seepage meters deployed in the same area. The rates reflect
the average SGD rate (cm/day) throughout the time-series, with the corresponding
standard deviation for each set. Values in parentheses are based on an endmember equivalent to the highest pore water radium measured. Seepage meter
results from Taniguchi et al. (2008)…………………………………………………
50
Table 4.4. Summary of 226Ra content on collected sediments, measured via
gamma spectrometry. All samples except TS-1 were collected in the Yellow
River, above the maximum salinity front. Uncertainties shown represent 1-σ
measurement uncertainties…………………………………………………………..
53
Table 5.1. Summary of transport distance calculation parameters for riverine DIN
fluxes. Results using our calculated nitrogen uptake rate, and the literature uptake
rate (Raabe et al., 2004) are reported………………………………………………..
62
Table 5.2. Summary of transport distance calculation parameters for groundwaterderived DIN fluxes. Results using our calculated nitrogen uptake rate (46 mmol
N/m2.day), and the literature uptake rate (5.1 mmol N/m2.day) are reported……….
63
viii
Table 6.1. Summary characteristics of the sampled groundwater sources. These
sites are arranged in order from north to south. Reference numbers appear
throughout the text. Those sites listed under the ‘Used for end-member’ category
show the samples that are averaged to provide an end-member value at a specific
site for the SGD model………………………………………………………………
73
Table 6.2. Measured parameters for each of the sampled groundwater sources.
Reference numbers refer to the specific site characteristics listed in Table 7.1.
Some sites were sampled during multiple field trips. n/a indicates that this
particular parameter was not measured during the specific sampling, whereas ‘BD’
indicates that the measurement was below the detectable activity………………….
75
Table 6.3. Model parameters selected for each groundwater plume and the
associated total and freshwater SGD fluxes determined by averaging all positive
SGD fluxes over the time-series. The results are arranged in order from north to
south…………………………………………………………………………………
77
Table 7.1. Description of each of our sampling sites. Reported distances upstream
are relative to the US 98 bridge at the mouth of the Apalachicola River. Included
are the nearest USGS gauging station identification numbers and their distance
from our sampling sites……………………………………………………………...
91
Table 7.2. Reservoir TSS deposition/erosion parameters for West Point Lake,
Lake Blackshear, and Lake Seminole for both June 2006 (top section) and
February 2007 (bottom section). Net TSS Fluxes are derived by subtracting the
flux into the reservoir from the flux out, and negative values indicate deposition
within the reservoir. Discharge estimates and reservoir areas are taken from U.S.
Geological Survey and U.S. Army Corps of Engineers online databases…………..
101
Table 7.3. TSS input parameters for the river segment through Atlanta for both
June 2006 (top section) and February 2007 (bottom section). Discharge estimates
are taken from U.S. Geological Survey online databases…………………………...
103
Table 7.4. Results of the 7Be/210Pb transit time analyses for those riverine
segments in February 2007 that show decreases in the activity ratio as well as in
the TSS concentration. Estimated residence times are based on Equation (1) in the
text and particle transport velocities are converted based on the river length
between sample sites………………………………………………………………...
109
Table 7.5. Estimation of the As and Sb heavy metal flux to Apalachicola Bay.
The fluxes determined from each of the rivers represent the portion of the total
flux to Apalachicola Bay contributed from each river segment…………………….
111
Table C.1. Water quality characteristics from the June 2006 sampling of the ACF
system. ‘nd’ indicates that no data exists for that particular measurement…………
124
ix
Table C.2. Dissolved radium isotope activity concentrations during the June 2006
ACF sampling. ‘BD’ indicates that a measurement was below detection.
Measurement uncertainties are reported at the 1-σ level……………………………
125
Table C.3. Particulate radionuclide analyses during June 2006 in the ACF
sampling. ‘BD’ indicates that a measurement was below detection. Measurement
uncertainties are reported at the 1-σ level…………………………………………...
126
Table C.4. Particulate stable tracer and metal analyses from the June 2006
sampling in the ACF system. ‘BD’ indicates that a measurement was below
detection. Measurement uncertainties are reported at the 2-σ level. ………………
127
Table C.5. June 2006 particulate org-C and N concentrations (%) as well as their
isotopic values……………………………………………………………………….
131
Table C.6. Water quality characteristics from the February 2007 sampling of the
ACF system. ‘nd’ indicates that no data exists for that particular measurement…...
132
Table C.7. Dissolved radium isotope activity concentrations during the February
2007 ACF sampling. ‘BD’ indicates that a measurement was below detection.
Measurement uncertainties are reported at the 1-σ level……………………………
133
Table C.8. Particulate radionuclide analyses during February 2007 in the ACF
sampling. ‘BD’ indicates that a measurement was below detection. Measurement
uncertainties are reported at the 1-σ level…………………………………………...
134
Table C.9. Particulate stable tracer and metal analyses from the February 2007
sampling in the ACF system. ‘BD’ indicates that a measurement was below
detection. Measurement uncertainties are reported at the 2-σ level………………..
135
Table C.10. February 2007 particulate org-C and N concentrations (%) as well as
their isotopic values…………………………………………………………………
139
Table C.11. Dissolved stable tracer and metal analyses from the February 2007
sampling in the ACF system. ‘BD’ indicates that a measurement was below
detection. Measurement uncertainties are reported at the 2-σ level………………..
140
x
LIST OF FIGURES
Figure 2.1. Picture of the cartridge design fully assembled (A). The specific parts
are the brass valves (1), the PVC plug (2), the Union with o-ring (3), the clear PVC
pipe (4), and the end cap (5). Also included is a picture of the open cartridge,
showing the different parts of the Union (B)………………………………………..
6
Figure 2.2. Sample calibration curves for the Rn emanation line, RaDeCC system,
and RAD7 system. Slopes provide a measure of efficiency (cpm/dpm). These
represent one example of several units of these systems that were employed for
this work……………………………………………………………………………..
8
Figure 2.3. Sample set 226Ra results based on each measurement technique. The
RAD7, Rn line, and RaDeCC results represent the average of 2 separate
measurements. The error bars (+ and -) represent the average 1-σ measurement
uncertainties based on counting statistics…………………………………………...
11
Figure 2.4. Minimum detectable activity (solid line) and relative uncertainty
(dotted line) plots for each measurement system showing their dependence on the
counting time. The horizontal dashed lines indicate an MDA of 3 dpm/100L and a
relative uncertainty of 10%, and thus the corresponding measurement time needed
by each system to attain that level of precision. The uncertainty calculations are
based on a fiber containing a total activity of 10 dpm 226Ra (10 dpm/100L activity
and 100 L volume)…………………………………………………………………..
14
Figure 2.5. Results of the sparging time analysis and the corresponding system
efficiencies based on the associated sparging time……..……………………….......
15
Figure 3.1. Map of the study site at the Yellow River mouth and surrounding
Bohai Sea……………………………………………………………………………
20
Figure 3.2. Yellow River hydrograph based on daily average discharge, with each
sampling period highlighted…………………………………………………………
21
Figure 3.3. Horizontal salinity profiles from all three sampling trips. Negative
distances represent salinities up the river, whereas positive distances indicate
offshore values. The river “mouth” is somewhat arbitrary, but is at a constant
position based on GPS coordinates………………………………………………….
24
Figure 3.4. Dissolved silica (a), barium (b), 224Ra (c), and total suspended
sediment concentration (d) as measured across the salinity gradient in September
2006. The solid line in each plot shows the horizontal salinity profile. Error bars
reflect 1-σ measurement uncertainties………………………………………………
25
xi
Figure 3.5. September 2006 barium (nM) vs. 224Ra (dpm/100L) relationship in
samples along offshore transect. Error bars represent 1-σ measurement
uncertainty…………………………………………………………………………...
26
Figure 3.6. Distribution of 223Ra (a), 224Ra (b), 226Ra (c), and 228Ra (d) activities in
dpm/100L for all three sampling campaigns. Both 223Ra and 224Ra peak in the
salinity front and decrease offshore due to decay and mixing. The long-lived
isotopes remain constant (226Ra) or increase (228Ra) offshore………………………
27
Figure 3.7. Horizontal distribution of apparent radium ages calculated from
224
Ra/223Ra (a), 224Ra/226Ra (b), and 224Ra/228Ra activity ratios (c)………………….
28
Figure 3.8. Apparent radium ages calculated from the 224Ra/228Ra activity ratios
from 2004 (a), 2005 (b), and 2006 (c). Best-fit equations are shown with their
95% confidence levels. The transport rates shown are calculated from the slopes
of the best-fit lines…………………………………………………………………...
31
Figure 3.9. All apparent radium ages calculated from the 224Ra/228Ra isotope ratio
plotted against their corresponding salinities………………………………………..
32
Figure 4.1. Map showing the Yellow River delta and the SGD study site (box on
the west side of Laizhou Bay)……………………………………………………….
37
Figure 4.2. Profile of seafloor bottom (a) and distribution of salinity (b), pH (c),
224
Ra (d), 223Ra (e), 226Ra (f), and 222Rn (g) along offshore transect from the study
site. Error bars reflect 1-σ measurement uncertainties……………………………..
41
Figure 4.3. Distribution of apparent radium ages offshore based on the 224Ra/223Ra
AR (a) and the 224Ra/226Ra AR (b). Dashed lines represent the 95% confidence
level of the linear regression. Transport rates are converted into units of cm/sec
from the reciprocal of the slope of the linear regression……………………………
43
Figure 4.4. Temporal variability of the water level (right axis) as well as salinity
(a), pH (b), 224Ra (c), 223Ra (d), 226Ra (e), and 222Rn (f) during TS-1 in September
2006. Error bars represent 1-σ measurement uncertainties………………………...
44
Figure 4.5. Temporal variability of the water level (right axis) as well as salinity
(a), pH (b), 224Ra (c), 223Ra (d), 226Ra (e), and 222Rn (f) during TS-2 in July 2007.
Error bars represent 1-σ measurement uncertainties………………………………..
46
Figure 4.6. Radium time-series results for SGD fluxes during TS-1 (a) and TS-2
(b) as well as the corresponding water level recorded at the study site. Results
from 224Ra are shown by the black circles, those from 223Ra by gray squares, and
those from 226Ra analysis are shown by white triangles. The results shown
represent a 3-point smoothing……………………………………………………….
49
xii
Figure 4.7. Results from the 222Rn SGD model for TS-1 (a) and TS-2 (b) as well
as the corresponding water level recorded at the study site. The results shown
represent a 5-point smoothing. Error bars shown are propagated errors throughout
the model calculations……………………………………………………………….
51
Figure 4.8. Independent analysis of SGD by means of automated seepage meters,
as summarized by Taniguchi et al. (2008) for the time during TS-1 (a) and TS-2
(b) chemical tracer sampling. Error bars represent standard deviation of all
measurements recorded during each measurement interval (30 minutes)…………..
54
Figure 5.1. Distribution of 224Ra/228Ra ages offshore from the mouth of the Yellow
River found in September 2006. The reported transport rate is derived from the
inverse of the slope of the best-fit line. The thin lines represent the 95% confidence
interval of the regression…………………………………………………………….
60
Figure 5.2. Distribution of silica (μM) along offshore transect measured in
September 2006 (A). The dashed line represents the conservative mixing line due
simply to diffusional mixing. The closed symbols are then used to calculate a first
order removal constant (B) by plotting the natural logarithm of the silica
concentrations against their respective apparent radium ages………………………
61
Figure 5.3. Schematic representing the progressing mixing areas from the pointsource Yellow River mouth (a) and the SGD area (b). Tr represents the reported
transport rate (Peterson et al., 2008a; Peterson et al., 2008b)……………………….
62
Figure 6.1. Map of the (A) Hawaii Island chain, featuring the (B) western coast of
the Big Island of Hawaii. Each plus symbol represents the radon platform position
in the groundwater plumes associated with (C) Kiholo Bay, (D) Honokohau
Harbor, (E) Kailua-Kona Harbor, and (F) Queen’s Bath, Manini Beach, and
Kahauloa Bay (north to south) in Kealakekua Bay………………………………….
66
Figure 6.2. Example of an aerial TIR image of a buoyant SGD plume (taken from
Johnson, 2008). Note the temperature of the plume is cooler than the ambient
ocean water………………………………………………………………………….
68
Figure 6.3. Diagram of model variables and their interactions used to simulate the
changing water fluxes into and out of the control volume, defined here as the
‘groundwater plume’. See text for details…………………………………………..
69
Figure 6.4. Offshore transect of (A) salinity, (B) 222Rn, and (C) 224Ra from the
plume emanating from Kahauloa Bay. This transect was collected in August
2005………………………………………………………………………………….
72
xiii
Figure 6.5. Time-series analysis of (A) 222Rn activity, (B) salinity, and (C)
temperature in surface waters of the Kahauloa Bay plume measured in August
2005. (D) Modeled total SGD results are also included, after a 3-point smoothing
function. Solid lines represent tidal variations……………………………………...
74
Figure 6.6. Time-series analysis of (A) 222Rn activity, (B) salinity, and (C)
temperature in surface waters of Kahauloa Bay measured in February 2006. (D)
Modeled total SGD and (E) tidal flux results are also included and both represent a
3-point smoothing. Solid lines represent tidal variations…………………………..
76
Figure 6.7. Offshore transect of (A) salinity, (B) 222Rn, and (C) 224Ra from the
plume at Manini Beach. This transect was collected in August 2006……………...
78
Figure 6.8. Time-series analysis of (A) 222Rn activity, (B) salinity, and (C)
temperature in surface waters at Manini Beach measured in August 2006. (D)
Modeled total SGD results are also included after a 3-point smoothing function.
Solid lines represent tidal variations………………………………………………...
80
Figure 6.9. 222Rn versus salinity plots of the raw time-series measurement values
collected at (A) Queen’s Bath, (B) Kailua-Kona Harbor, and (C) Kiholo Bay. The
regression lines are extrapolated to estimate a radon end-member in the
groundwater at these sites…………………………………………………………...
82
Figure 6.10. SGD flux model results from the (A) Queen’s Bath, (B) Kailua-Kona
Harbor, (C) Honokohau Harbor, and (D) Kiholo Bay time-series deployments. All
results represent a 3-point smoothing. Solid lines represent tidal variations……….
84
Figure 7.1. Map of the Apalachicola-Chattahoochee-Flint River basin through
Florida, Georgia, and Alabama (insert). Each sample site is shown (black dots and
alpha-numeric designations). Major reservoirs (all man-made since 1948) along
the river system are highlighted. The fall line where the Piedmont geologic unit
meets the coastal plain province is designated by the dashed line…………………..
89
Figure 7.2. Hydrograph of the Apalachicola River measured at Chattahoochee,
Florida (below Lake Seminole) during the study period. Sampling periods are
shown by arrows. Data from online USGS river discharge database (02358000 at
Chattahoochee, Florida). Note that high discharge is typically during winterspring except during tropical storm events (e.g., July 2005 = Hurricane
Dennis).……………………………………………………………………………...
93
xiv
Figure 7.3. Total suspended sediment (TSS) concentrations from June 2006 (solid
circles) and February 2007 (open squares) for the Chattahoochee, Flint, and
Apalachicola River sample sites. Gray bars indicate reservoir locations on each
river: (1) Lake Seminole, (2) Walter F. George Reservoir, (3) Lake Worth, (4)
Lake Blackshear, (5) Lake Harding, (6) West Point Lake, and (7) Lake Lanier.
The fall line is designated by the horizontal gray dashed lines. The center of urban
Atlanta, Georgia is shown by the circled ‘A’, and the location of the CFPPs is
indicated by the asterisk (*).……………………………………………………...…
96
Figure 7.4. Particulate radionuclide activities from June 2006 (solid circles) and
February 2007 (open squares) for the Chattahoochee, Flint, and Apalachicola
River sample sites. Shown are 226Ra (a), 228Ra (b), 40K (c), 7Be (d), and 210PbXS
(e). Other symbols are the same as in Figure 7.3. Error bars represent 1σ
measurement uncertainties, but often do not extend beyond the symbol.…………..
98
Figure 7.5. Particulate stable tracer concentrations from June 2006 (solid circles)
and February 2007 (open squares) for the Chattahoochee, Flint, and Apalachicola
River sample sites. Shown are organic-C (a), Ca (b), As (c), Sb (d), and Zn (e).
Other symbols are the same as in Figure 7.3. Error bars represent 1σ measurement
uncertainties, but often do not extend beyond the symbol…………………………..
99
Figure 7.6. Calculated components of the TSS load from each sample site (total
height of bar; mg/L) contributed by lithogenic particles (gray section), organic
particles (white section), and uncharacterized particles (assumed to be composed
of non K-bearing minerals and anthropogenic particles). Each river segment’s
samples are grouped for June 2006 (A) and February 2007 (B). Carbonate
contributions (<1%) are not shown……..…………………………………………...
104
Figure 7.7. Activity ratios of 7Be/210PbXS from June 2006 (solid circles) and
February 2007 (open squares) for the Chattahoochee, Flint, and Apalachicola
River sample sites. Other symbols are the same as in Figure 7.3.………………….
105
Figure 7.8. Calculated particle ages for the bulk suspended sediment samples
(based on Equation 7.1) from June 2006 (solid circles) and February 2007 (open
squares). The samples with no detectable excess 210Pb have been omitted. Other
symbols are the same as in Figure 7.3.……………………………………………...
106
Figure 7.9. Theoretical relationships between piston age of bulk suspended
sediment sample (Equation 7.1; solid line) and the fraction of ‘new’ sediment in a
sample (Equation 7.2; dashed line) relative to an assumed initial value of the
measured 7Be/210PbXS to that in the atmospheric source (from Matisoff et al.,
2005)………………………….……………………………………………………..
107
xv
Figure 7.10. Change in the average ‘new’ sediment percentage (based on Equation
7.2) plotted against the corresponding change in TSS concentration between each
pair of two contiguous sample sites for June 2006 (a) and February 2007 (b). The
sample corresponding to Atlanta, Georgia is shown by the circled ‘A’ and ignored
for the linear regression. Note that the center point of both lines is near (0,0),
suggesting that the mixing model adequately reflects TSS mixture of recent (high
7
Be/210PbXS) and old (low 7Be/210PbXS)
material.……………………………………………………………………………...
xvi
108
ABSTRACT
Naturally-occurring radioisotopes are ubiquitous in nature, and as such, there are many
opportunities for researchers to use them as environmental tracers. Their associated radioactive
decay rates provide an inherent time clock that often makes these radioisotopes more effective
chemical tracers than other substances, since they can help assess the temporal component of a
particular process. Radioisotopes were employed in this study to examine different coastal
oceanographic processes that have far reaching consequences. A review of various measurement
methods for assessing 226Ra concentrations, a valuable tracer employed throughout most of this
dissertation, is presented. This is followed by applying natural radioisotopes to study various
oceanographic processes in China, Hawaii, and a river system in the southeastern United States.
Acrylic fibers impregnated with MnO2 (‘Mn-fiber’) have become a valuable tool for
concentrating dissolved radium for oceanographic applications. Several techniques have been
outlined in the literature describing the measurement of 226Ra on Mn-fiber via its gaseous
daughter, 222Rn. We describe procedures for three radon-based non-destructive measurement
techniques for 226Ra on Mn-fiber (via RAD7, RaDeCC, and Rn emanation line systems) and
perform an intercomparison among them, using the standard technique of γ-spectrometry as a
reference. We find that all methods agree in terms of the measured activity, with the respective
correlation coefficients (r) between any two different methods ranging from 0.78 to 0.93. The
methods vary in their advantages, with the Rn emanation line and RaDeCC techniques offering
the best measurement uncertainties and lowest minimum detection activities, while the RAD7
method is the least operator-demanding technique.
In the reported field activities, radium isotopes were mapped around the mouth of the
Yellow River in China to examine the transport rates of Yellow River water through the estuary
and into the Bohai Sea during three different field excursions, covering a large range in the
discharge patterns of the river. Using radium isotope ratios, horizontal transport rates ranging
from 1.4 to 4.7 cm/s were found throughout the delta and apply over most of the discharge range
of the river while exhibiting no seasonal variability. Time series analyses of radium isotopes and
222
Rn were also used near the Yellow River delta to assess submarine groundwater discharge
(SGD) rates. Modeled vertical SGD rates in this area varied between 4 and 20 cm/day during the
sampling periods, thus delivering significant volumes of groundwater containing elevated
concentrations of nutrients to the coastal zone. We combine these results to show that the
dissolved freshwater nitrate inputs (riverine and SGD) around the Yellow River delta cannot be
directly transported to the central Bohai Sea (where increasing nitrate levels have been
documented) before being biologically or otherwise removed.
222
Rn was the main tracer employed to examine groundwater discharge along the leeward
coast of the Big Island of Hawaii. Here, geological formations act to direct coastal aquifer
waters to point-source discharges, forming buoyant plumes of freshwater extending out into the
coastal waters. A box model was created using mass balance of radon, salt, and water to assess
the discharge rates of these plumes. The point-source inputs were found to discharge thousands
of cubic meters of brackish water to the coastal zone each day, ranging from 1100 m3/day to
12,000 m3/day of total water fluxes. The purely freshwater components of these discharges
varied from 630 m3/day to 8600 m3/day. Considering possible sources of nutrient and industrial
contamination in this area, these discharges can significantly affect the local reefs and
ecosystems along this coast.
xvii
A particle tracing project was performed in the Apalachicola-Chattahoochee-Flint River
system in the southeastern United States. The river system empties into Apalachicola Bay in
northwest Florida, a biodiverse and economically valuable ecosystem to the state of Florida. We
applied various naturally-occurring radionuclides as tracers to reveal suspended particle transport
behavior under both base flow (June 2006) and high discharge conditions (February 2007).
Potassium-40 activities are used to assess the lithogenic/crustal fractions of each suspended
sediment sample (ranging from 4-60%), whereas the organic fraction ranges from 4 to 32% by
mass. Particulate radium isotopes (namely 226Ra and 228Ra) were employed to trace the origin of
the suspended particle flux to Apalachicola Bay. During low discharge, the Flint River
dominates the particulate flux to Apalachicola Bay (70%), whereas the Chattahoochee River
contributes 30% of the flux and the Apalachicola River is net depositional. During high
discharge, the Chattahoochee River contributes the majority of the suspended particles (56%),
with the Apalachicola River contributing 30% and the Flint River only providing 14% of the
particulate flux. The 7Be/210Pb activity ratios were used to assess the residence time of
suspended particles in upstream reservoirs (5.2 days in Lake Blackshear and 60 days in West
Point Lake), and also to examine open channel transport velocities (~ 13 cm/sec in the lower
Flint River) during high discharge.
xviii
CHAPTER 1
INTRODUCTION TO THE DISSERTATION
Researchers have long used naturally-occurring radioisotopes to examine environmental
processes in the fields of geology, biology, and oceanography. The very wide range in
radioactive half-lives within the uranium and thorium decay chains (micro-seconds to billions of
years) offers great applicability of these isotopes in examining processes that span an equally
large range in temporal domains. The natural decay mechanism within these radionuclides
allows researchers to examine the time component of the processes in question, a trait that stable
tracers simply cannot provide. In addition, the different behavioral characteristics exhibited by
the various elements within the natural decay chains (gas versus solid, soluble versus insoluble,
oxidation-reduction behavior) ensure that disequilibrium in the natural environment is often
observed, thereby facilitating the applicability of these radioisotopes for research studies.
Natural radioisotopes have been applied in oceanographic research studies to examine
processes such as water mass mixing, sedimentation and ridge spreading rates, coral reef growth,
and gas transfer across the air-sea interface. Due to the more dynamic nature of the coastal zone,
terrestrial freshwater sources can have a direct impact on local biogeochemical cycles. Mixing
processes near shore (driven by tides, winds, and waves) occur over very short temporal scales,
and sediment movement can affect local biologic, geologic, and anthropogenic behavior. The
highly dynamic nature of the coastal zone thus presents a challenging opportunity to study shortterm processes using natural radioisotopes in an environment where other methods are not nearly
as effective.
A common theme of the research summarized here is the use of naturally-occurring
radioisotopes to examine various water and sedimentary mixing processes in rivers and the
coastal ocean. One process that can be examined in this way is submarine groundwater
discharge (SGD). SGD is defined as the upward water flux from the pore water of the bottom
sediments to the overlying water column, and does not differentiate between origin or driving
force (Burnett et al., 2003b). Generally, SGD has a terrestrial groundwater component (driven
by the hydraulic gradient of aquifer water flowing towards the ocean) and a recirculated seawater
component (driven by various oceanic forcing factors exchanging with the sub-bottom fluids)
(Zektser, 2000; Burnett et al., 2003a; Charette, 2007). As a result, SGD is often variable over
both spatial and temporal scales. Natural radioisotopic tracers have been shown to be very
powerful for quantifying SGD rates because they tend to integrate relatively large areas,
smoothing out any effects of the small scale variability of SGD (Burnett et al., 2006a;
Swarzenski, 2007; Charette et al., 2008).
SGD can deliver dissolved constituents and nutrients to the coastal ocean in quantities
comparable to those delivered via rivers (Swarzenski et al., 2007a). Unlike rivers, however,
SGD can not be seen or gauged and is a much more disseminated input of dissolved constituents.
SGD acts as a transport mechanism for regenerated nutrients to be flushed back into the water
column. Unlike SGD, however, rivers transport particles and any insoluble constituents that may
be attached to them to the coastal ocean including heavy metals, phosphorus, and pesticides
(Bonniwell et al., 1999). The dichotomy between riverine and SGD chemical inputs and their
associated pathways makes for an interesting comparison between these two terrestrial water
1
sources to the coastal ocean. The suite of natural radioisotopes applied here represents a
valuable tool that can be used to make such assessments.
The future of U/Th chain isotope applications to coastal oceanographic studies is
promising, yet somwhat dependent upon the development of new measurement tools. For
example, researchers with the International Atomic Energy Agency (IAEA) have developed an
underwater gamma-ray spectrometer capable of real-time in-situ spectral analysis. With further
enhancements to these types of systems, many new research possibilities would be available to
oceanographersAs another example, researchers at Florida State University are working to
develop a new system for measuring the short-lived 220Rn isotope (T1/2 = 54 seconds) in realtime. This tool would also greatly assist the field by providing yet another tool available to
researchers.
Chapter 2 of this dissertation presents a comparison of several techniques for measuring
226
Ra via its gaseous daughter 222Rn. The common method by which researchers measure 226Ra
is via gamma-spectrometric counting methods, but this technique is very expensive and timeconsuming. We compare here three alternative techniques based on their previously published
methods to determine which technique is most suitable in terms of measurement time,
uncertainty, minimum detectable activity, and operator demand.
We applied all naturally-occurring radium isotopes to examine the freshwater mixing
environment around the Yellow River delta in China. Chapter 3 uses these tracers to estimate
the Yellow River water mixing scales into the Bohai Sea. We estimate the offshore mixing
velocities of the river at three different discharge stages to be similar and conclude that the
riverine mixing within the Yellow River is modulated by tidal effects rather than river forces. In
Chapter 4, we use radium isotopes to examine the SGD rates along the Yellow River delta during
two different sampling periods. We develop a model to estimate SGD rates from high-resolution
(hourly) radium sampling and compare the results to other, more proven techniques 222Rn and
benthic seepage meter estimates). Chapter 5 integrates the results from Chapters 3 and 4 to
examine whether either of these freshwater sources to the Bohai Sea can be providing nitrate
fluxes sufficiently high that the nutrient can be directly transported to the central Bohai Sea prior
to being biologically removed. Though this study cannot account for the cycle of nitrate uptake,
biological burial, subsequent regeneration, and further offshore transport, we conclude that these
shoreline inputs are likely not sufficient to be causing the observed nitrate increases in the Bohai
Sea.
In Chapter 6, we use 222Rn and radium isotopes as tracers to examine point-source SGD
inputs along the western coast of the Big Island of Hawaii. The geology of this area focuses
most of the terrestrial groundwater flow to distinct discharge portals along the coastline as
opposed to the more spatially uniform, diffuse SGD flows that are common along many other
shorelines. We developed a new model that could be used to estimate the SGD rates associated
with these point-source discharges. We then present the model’s application to several sites
along the Hawaiian coastline.
While the previous chapters examined water mixing and the associated dissolved
chemical behavior, Chapter 7 applies natural radioisotopes to riverine suspended particle
transport. We use radium isotopes, 210Pb, 7Be, and 40K to characterize and track the transport of
suspended particles down the Apalachicola-Chattahoochee-Flint River system in the
southeastern United States. The heavy metal contaminant transport (e.g., arsenic and antimony)
on these suspended particles downstream to the fragile ecosystem in Apalachicola Bay is also
examined to determine the main source areas of these contaminants.
2
CHAPTER 2
A COMPARISON OF MEASUREMENT METHODS FOR RADIUM-226 ON
MANGANESE-FIBER
Article published in Limnology & Oceanography: Methods
Abstract
Acrylic fibers impregnated with MnO2 (‘Mn-fiber’) have become a valuable tool for
concentrating dissolved radium for oceanographic applications. With four naturally-occurring
radium isotopes (223Ra, 224Ra, 226Ra, and 228Ra) with vastly different half-lives (3.6 days to 1600
years), radium can be a powerful tool for tracing terrestrial water discharges into the ocean and
studying coastal mixing processes. Several techniques have been outlined in the literature
describing the measurement of 226Ra on Mn-fiber via its gaseous daughter, 222Rn. We describe
procedures for three radon-based non-destructive measurement techniques for 226Ra on Mn-fiber
(via RAD7, RaDeCC, and Rn emanation line systems) and perform an intercomparison among
them, using the standard technique of γ-spectrometry as a reference. The main goal of this work
is to provide information regarding various measurement figures of merit (which are often not
reported in original method publications) so users can choose the best method to meet their
needs. We find that all methods statistically agree in terms of the measured activity. The Rn
emanation line and the RaDeCC systems (both based on Lucas cell counting) provide the lowest
measurement uncertainties and MDA values for a given counting time. The RAD7 technique, on
the other hand, offers the advantage of being an automated system, thus requiring minimal user
interaction. The standard γ-spectrometry technique, while more time-consuming and destructive,
has the advantage of providing a simultaneous measurement for 228Ra.
Introduction
One of the great advancements in the field of marine environmental radioactivity was the
introduction of MnO2-impregnated acrylic fiber (‘Mn-fiber’). Mn-fibers were first introduced in
the early 1970’s (Moore and Reid, 1973). This material is known to quantitatively adsorb
dissolved radium isotopes, thus concentrating them to allow for more accurate measurements.
While other metals are also adsorbed onto Mn-fiber (e.g., Ba, Ca, Pb, Hg, Cu, Zn, Co, Cd, Th,
Ac, Pu, Am; see Moore; 1976; Mann et al., 1984), radium has most often been the intended
target for environmental studies using Mn-fibers. Four naturally-occurring radium isotopes exist
(223Ra, 224Ra, 226Ra, and 228Ra), and their large range in half-lives (11.4 days, 3.6 days, 1600
years, and 5.7 years, respectively) allows researchers to adopt them as useful tracers for many
different oceanographic processes of varying temporal scales.
W.S. Moore pioneered the use of the long-lived radium isotopes (226Ra and 228Ra) for
open ocean studies after the introduction of Mn-fibers. For example, Moore (1976)
demonstrated the sampling protocol of 228Ra in the deep ocean in order to examine thermohaline
3
circulation and the benthic input of 228Ra into bottom waters. He later mapped the North
Atlantic basin for 226Ra and 228Ra in surface and deep waters (Moore et al., 1985). In a study
along the southeastern United States continental shelf, Moore (1996) extended the use of radium
isotopes (namely 226Ra) to quantify submarine groundwater discharge (SGD) inputs along the
continental shelf. That study inspired many researchers to examine the role of SGD as an input
mechanism of dissolved substances to coastal waters and today stands as one of the most
prominent of early SGD studies.
After the adaptation of a pre-existing delayed coincidence counting system (Griffin et al.,
1963) to measure the short-lived radium isotopes (223Ra and 224Ra), termed the Radium Delayed
Coincidence Counter (RaDeCC) (Moore and Arnold, 1996), the applications of these isotopes to
oceanographic studies expanded substantially. By using all four isotopes in the coastal
environment, radium can act as a tracer to determine river plume mixing rates (Krest et al., 1999;
Moore, 2000b; Peterson et al., 2008b), apparent water mass ages and/or residence times (Moore,
2000a; Dulaiova and Burnett, 2008), and saline SGD fluxes (Hwang et al., 2005; Kim et al.,
2005; Mulligan and Charette, 2006; Peterson et al., 2008a). Radium is a useful tracer of
terrestrial waters (both surface and groundwater) because it is supplied by radioactive ingrowth
from its particle-reactive thorium parents contained in sediments, and subsequently desorbs in
the presence of high ionic strengths (Li and Chan, 1979; Martin and Akber, 1999; Nozaki et al.,
2001).
While Mn-fibers containing short-lived radium isotopes are relatively simple to measure
via the RaDeCC system, the analysis of the long-lived 226Ra and 228Ra is far more time
consuming. The traditional measurement protocol for long-lived isotopes on Mn-fiber involves
gamma (γ-) spectrometric counting techniques (Moore, 1984; Dulaiova and Burnett, 2004), but
other options exist, especially for 226Ra. Most of these other counting methods indirectly analyze
226
Ra via its gaseous daughter, 222Rn (half-life of 3.8 days). We evaluate three of these
techniques using measurements based on: (1) a radon-in-air monitor (RAD7 produced by
Durridge Co.; Kim et al. 2001); (2) the RaDeCC delayed coincidence counting system (Moore
and Arnold, 1996; Waska et al., 2008); and (3) a radon emanation line that sparges and traps
radon followed by counting in a Lucas cell (Key et al., 1979). Other possible techniques exist
for 226Ra measurement from Mn-fiber, but these require further wet chemistry procedures to
extract the radium from the fiber for later quantification by either liquid scintillation counting or
γ/α-spectrometry (Sill, 1987; Moon et al., 2003). We focus here on the non-destructive counting
techniques directly from the Mn-fiber.
The techniques described here rely on the concept of radioactive ingrowth of 222Rn and
its emanation from the fiber. A Mn-fiber containing 226Ra, when sealed in an air-tight container
for a known amount of time, will eventually produce 222Rn in the air space at an activity directly
proportional to the 226Ra on the fiber. One can then measure the 222Rn activity in this air space
and use the ingrowth time to calculate the activity of 226Ra on the fiber. By calibrating each
method with known amounts of 226Ra on Mn-fiber, we can determine the efficiency of each
approach.
The objectives of this paper are thus to assemble the protocols for each of these radon
ingrowth-based methods and use a common set of samples to compare the techniques among
themselves and to a standard gamma spectrometry method. It is important to note that we
employ the methods as they appear in the literature and as the measurement systems are
commonly used. Further enhancements to the methods are possible and are suggested later. We
consider the following measurement figures of merit: operator interaction, measurement time,
4
efficiency, uncertainty, MDA, and associated costs. The ultimate goal here is to provide the
necessary information to allow a user to determine the most appropriate method based on one’s
needs and resources.
Materials and Procedures
Cartridge Design
One of the most important aspects of measuring 226Ra via 222Rn is the design of an airtight cartridge that can mount directly to all the different measurement systems. The following
design (Figure 2.1) has proven to be reliable throughout the course of our experiments. This
cartridge is based on the style of those supplied with the RaDeCC system (Moore and Arnold,
1996), but is adapted to allow air-tight sealing.
The foundation of our cartridge is a 15 cm long section of 1¼” clear PVC pipe (Figure
2.1A, #4; e.g., United States Plastic Corp. item # 34105). The advantage of the clear pipe is that
one can visually observe the Mn-fiber inside to assure that it is well-fluffed (Sun and Torgersen,
1998; Moore, 2007). On one end of the pipe, we glue a 1¼” PVC end cap (Figure 2.1A, #5). On
the other end, we glue the base of a 1¼” PVC Schedule 80 Union (Figure 2.1A, #3; e.g., Spears
Manufacturing Co. item #897-012). This union consists of two separate slip couplings, which
are connected with an o-ring seal and held together by a screw lock (Figure 2.1B). The cartridge
must provide open access to the Mn-fiber, but also must remain air-tight when sealed. A PVC
Union with o-ring meets these needs. Into the other end of the union, we glue a 1¼” PVC plug
(Figure 2.1A, # 2). The ends of the cap and plug are each tapped with threaded holes to allow
barbed brass ball valves to screw in with Teflon tape (Figure 2.1A, #1; e.g., MSC Direct item #
09881608). These ball valves allow the cartridge to be mounted directly to a measurement
device via tubing without the risk of losing any 222Rn during a transfer. In order to remain airtight, we have found that brass ball valves are better suited than those made from plastic. The
supplies required to build a cartridge cost less that $40 USD.
Because the Mn-fiber held inside a cartridge will be producing radon gas over time, we
recommend testing to ensure each cartridge is indeed air-tight. For evaluation, we created
positive pressure inside each cartridge with compressed gas, sealed the valves, and submerged
the cartridge under water. If no bubbles are visible, that is a strong indication that the cartridge
was successfully constructed.
Standard Preparation
We planned to calibrate each method to a series of Mn-fiber standards spanning a range
of Ra activities that would more than cover the range anticipated from 100L samples of
natural ocean waters. We started with a NIST-traceable primary standard with an activity of
2285 dpm/mL in 1M HNO3. From this, we conducted a series dilution (with Ra-free water) to
produce 5 1L standards containing a range of activities from 5 dpm to the full strength of the
standard. Each solution was adjusted to pH ~ 8, and then passed (gravity-fed) through pre-rinsed
Mn-fibers three times for quantitative extraction (Dimova et al., 2008).
226
5
Prior to each standard (and sample) run for the methods examined here, the Mn-fiber was
moistened with Ra-free water, inserted into a cartridge, sparged with compressed air (~ 2
minutes), and sealed inside the cartridge by closing the valves. The sparging serves to adjust the
fiber moisture content and also remove any background 222Rn that may be sealed in the cartridge
from the air. Proper humidity of the fiber is important to ensure optimum emanation of radon
from the fiber (Sun and Torgersen, 1998). The time of sealing is recorded for later calculation of
the ingrowth time.
Figure 2.1. Picture of the cartridge design fully assembled (A). The specific parts are the brass
valves (1), the PVC plug (2), the Union with o-ring (3), the clear PVC pipe (4), and the end cap
(5). Also included is a picture of the open cartridge, showing the different parts of the Union
(B).
RAD7 Method
The RAD7 is a radon-in-air monitor containing an internal air pump and an alpha
semiconductor detector that employs energy discrimination to count the daughters of 222Rn and
220
Rn. This tool has been widely used recently in groundwater discharge studies, as 222Rn is a
useful tracer of groundwater flow (Burnett et al., 2001). The RAD7 is a fully automated and
6
portable radon detector, capable of running continuously for days. An important requirement of
this system is that the air stream supplied to the unit remains dry (humidity < 10%).
We connect the Mn-fiber cartridge in a closed air loop with the RAD7 and a tube of
desiccant. Using two 2-way valves, we also plumb in a bypass loop that can isolate the cartridge.
Because the premise is to distribute the 222Rn that has ingrown inside the cartridge throughout
the air loop, the volume of the air loop is a vital part of the calibration and must be held constant
for each successive measurement (Lee and Kim, 2006).
The method of using the RAD7 to measure 226Ra on Mn-fiber was first introduced by
Kim et al. (2001), but the measurement protocol used here is based on that presented by Dimova
et al. (2007). Prior to each measurement, it is important to know the effective background of the
222
Rn in the air loop, which acts as the carrier gas. To measure this, we connect the tubing
system with the bypass valves set to isolate the cartridge. We run the RAD7 to measure the
background 222Rn count rate for roughly 90 minutes and later subtract this value from the sample
count rate. We then switch the bypass valves to introduce the cartridge to the air loop, open the
brass valves on the cartridge, and turn the RAD7 pump to ‘ON’ for 25 minutes to fully distribute
the 222Rn throughout the system. The RAD7 does not count at this time. Once this pumping
period is complete, we isolate the cartridge with the bypass loop, and start a RAD7 run for 15hours. We find that a 15-hour counting time is a convenient overnight counting period and is
usually sufficient to provide reasonable counting statistics, but this can be altered according to a
user’s needs.
We take the background-corrected RAD7 window A count rate (cpm) as the effective
instrument response for the calibration. This window represents the counts that are due to the
decays of 218Po, the direct daughter of 222Rn. The RAD7 has a different window (C) that counts
decays from radon’s great-granddaughter 214Po (after 3 hours for equilibration), but for
simplicity in this application, we only use the window A count rate. The calibration curve for
one RAD7 (Figure 2.2) indicates a linear response of the system to increasing 226Ra activity on
the Mn-fiber, with an efficiency of about 0.1 cpm/dpm (or 10%). For this and all of these
examined methods, we calibrate to only the 4 lower standards, excluding the highest standard to
achieve a calibration within the typical range of environmental samples. The system efficiency
is not only related to the counting efficiency of the RAD7 itself, but is also dependent upon the
specific geometry/volume of the system. One could use window C of the RAD7 to increase the
system efficiency and measurement precision, as there would be 2 measured counts (218Po and
214
Po) for each decay of 222Rn. However, this would require a 3-hour equilibration time before
beginning the counting.
Radon Emanation Line (Lucas cell) Method
The detection device used with a radon extraction line was first proposed by Henry Lucas
(1957) when he invented a radon scintillation cell, later termed the ‘Lucas cell’. The cell is an
air-tight chamber (ours have a volume of ~125 cm3) with inner walls coated with silver-activated
zinc sulfide, which emits photons when struck by alpha decay particles. Key et al. (1979) later
described a system by which radon can be extracted from water samples, concentrated on a trap
held at liquid nitrogen temperatures (or dry ice when used with a charcoal trap), and later
transferred to a Lucas cell for measurement by a photo multiplier tube (PMT) and a timer/scaler.
Later enhancements were made by several authors to enable precise measurements of 222Rn, and
thus its parent 226Ra (Butts et al., 1988; Mathieu et al., 1988).
7
Figure 2.2. Sample calibration curves for the Rn emanation line, RaDeCC system, and RAD7
system. Slopes provide a measure of efficiency (cpm/dpm). These represent one example of
several units of these systems that were employed for this work.
Butts et al. (1988) were the first to suggest measuring 222Rn directly from Mn-fibers with
this system, and provided the method by which we measure 226Ra bound on Mn-fiber via the Rn
emanation line. Prior to measurement, the radon emanation line is purged with compressed
helium while the background of the Lucas cell is counted (60 minutes). Then, the cartridge is
placed inline, the He flow diverted through the cartridge, and the ingrown 222Rn is flushed to the
liquid nitrogen trap for 90 minutes. Afterward, the cartridge is again bypassed, the trap is heated,
and the concentrated 222Rn is transferred to an evacuated Lucas cell via a stream of helium. The
cell is held for 3 hours to allow the 222Rn daughters to grow into equilibrium, then counted for 60
minutes. By allowing full ingrowth of the short-lived alpha-producing daughters of 222Rn, one
can maximize the counting efficiency because there will be three photons produced by each
decay of 222Rn (the 222Rn decay itself and one each from subsequent 218Po and 214Po decays).
Thus, the ideal counting efficiency of this system would be 3 cpm/dpm.
Figure 2.2 illustrates a calibration curve for one port on our emanation line. The
calibration shows a linear system response to increasing 226Ra activity on the fiber, with an
overall efficiency of about 2 cpm/dpm (or 200%). This is the highest efficiency of the three
methods examined here.
8
RaDeCC Coincidence Counting System Method
Waska et al. (2008) have recently introduced the following method of using the RaDeCC
system to measure 226Ra. This system consists of a closed air loop between a cartridge, an air
pump, and a large Lucas cell. The cell is coupled with a PMT and an electronic system that can
discriminate decays of different radon isotopes based on the timing of subsequent decays. The
RaDeCC system was adapted principally to measure 224Ra (via 220Rn decay to 216Po) as well as
223
Ra (via 219Rn decay to 215Po) on Mn-fiber (Moore and Arnold, 1996).
The routinely used RaDeCC short-lived radium protocol calls for purging the system with
ambient air prior to a measurement. For 226Ra measurements, however, this should be avoided,
as it unnecessarily introduces 222Rn from air into the system. Instead, we simply fill the air loop
with He and plug the cartridge inline. The brass valves on the cartridge are opened, and the air
pump is run for 25 minutes, circulating the 222Rn-rich air inside the cartridge throughout the
system. At the completion of the pumping, the brass valves are closed to prevent addition of
ingrown 222Rn from the Mn-fiber, and the system is left to equilibrate for 3 hours. During this
time, the short-lived polonium daughters of 222Rn are growing into equilibrium inside the Lucas
cell, but also the other isotopes of radon (220Rn and 219Rn) are decaying away. After 3 hours, we
start a measurement with the RaDeCC software, and take the total count rate to be the system
response to the 222Rn in the air loop after 3-4 hours of counting time.
As with the radon emanation line, the Lucas cell will emit 3 photons for each 222Rn
decay, so a perfect system would have an efficiency of 3 cpm/dpm. The calibration of a
representative system (Figure 2.2) shows lower values, as a result of counting efficiency and the
dilution effect of the 222Rn through the air loop, but the calibration curve is linear throughout the
intended range of activities, with an efficiency around 1.6 cpm/dpm (or 160 %).
γ-Spectrometry Method
Ra and 228Ra on Mn-fibers are conventionally measured by γ-spectrometry. We follow
the procedure established by Dulaiova and Burnett (2004) where Mn-fibers are ashed (at 550°C
for 8 hours) in special stainless steel crucibles which are later pressed into a specific geometry
for our planar γ-spectrometer. This technique avoids the potential loss of MnO2 ash during a
transfer from one container to another. Once a fiber is ashed and the crucible pressed, the disc
must be sealed with epoxy to prevent 222Rn escape, and held for ~ 21 days to allow all daughters
to grow into equilibrium before a sample can be measured. Of all the techniques summarized
here, this is the only destructive method, as the Mn-fiber is not usable after the ashing procedure.
Other techniques of measuring the long-lived radium isotopes by γ-spectrometry include ashing
followed by a transfer to another counting vessel or by acid leaching the Mn-fiber and barium
sulfate precipitation of the radium, followed by transfer to a counting vessel (Moore, 1984).
The γ-spectrometer used in this study is an Ortec IG detector with a relative efficiency of
20%. We have calibrated the system for 226Ra by preparing Mn-fibers from several aqueous
IAEA radium standards using the combination of three 226Ra daughter peaks: 214Pb (at 295.1 and
351.9 keV) and 214Bi (at 609.3 keV). The average efficiency of all three peaks is very low at
0.025 cpm/dpm. Based on these efficiencies, our detector typically requires ~ 48 hours of
counting time for each sample to achieve a reasonable uncertainty (< 5-10 %). This technique
will serve as the standard method of measurement for comparison among the other techniques
226
9
reviewed here, but note that larger γ-spectrometers and those with lower backgrounds can
produce results with lower uncertainties in shorter counting times.
Assessment
We used a series of seawater samples for a comparison of the previously presented
methods that are reviewed above. In doing so, we have not enhanced any of the techniques as
they are originally presented so as to not bias our results. The sample set consisting of eight
samples collected from 80L of seawater each as a water column profile from the Sea of Japan.
We measured each sample twice according to each of the 222Rn ingrowth methods, and once by
γ-spectrometry (Figure 2.3). The samples indicate a rise in 226Ra activity from the surface to
1000 m, then the activity remains nearly constant to 3500 m. Considering that the error bars in
Figure 2.3 represent 1-σ measurement uncertainties, there are no statistical differences among
measurements for any particular sample. The overall narrow range in activities among the
sample set tends to magnify the differences between measured values, i.e., these correlation
coefficients would likely be higher had there been more variation in the activity range of the
sample set.
Based on the measurement parameters associated with the sample set, we have
summarized the important topics to consider when comparing methods (Table 2.1). In terms of
user interaction, we use the same amount of ingrowth times for all radon-based methods (4
days). The preparation / holding time (for polonium daughter ingrowth) parameter indicates that
the RAD7 method, being much more automated, requires less user time than the RaDeCC and
the Rn line methods. The RAD7 only requires a 15 minute holding time (versus 3 hours for the
Rn emanation line and RaDeCC techniques) because it only counts radon’s daughter 218Po
instead of both 218Po and 214Po. The γ-spectrometric method here incorporates the recommended
21-day ingrowth time. The Rn line method (and to a lesser extent, the RaDeCC method) requires
much less counting time than do the RAD7 and γ-spectrometry systems because of their higher
efficiencies. Even considering their longer counting times, the RAD7 and γ-spectrometric
methods still produce larger measurement uncertainties than do the Rn line and RaDeCC
methods. It should be noted that we are using low-level environmental samples (< 0.15 dpm/L
226
Ra) for these comparisons.
We use the equation derived by Currie (1968) to determine the minimum detectable
activity (MDA) values, based on a blank measurement and the respective efficiencies (Table
2.1). Figure 2.4 plots the theoretical MDA of each method with increasing counting time. In
order to attain an MDA of 3 dpm/100L, the Rn line and RaDeCC methods require much less
counting time than do the other methods. Figure 2.4 also plots the calculated relative uncertainty
for each measurement method with increasing counting time. To achieve a relative measurement
uncertainty of < 10%, the Rn line and RaDeCC methods again require the least amount of
counting time. An important consideration here is that Ge gamma detectors with much higher
efficiencies and lower backgrounds than ours are available (for a much higher price). Both of
these factors would help to substantially lower the counting time required to attain these
measurement results.
10
Figure 2.3. Sample set 226Ra results based on each measurement technique. The RAD7, Rn line,
and RaDeCC results represent the average of 2 separate measurements. The error bars (+ and -)
represent the average 1-σ measurement uncertainties based on counting statistics.
Discussion
One flexible parameter of the methods considered here is the sparging time necessary to
transfer the 222Rn that has ingrown inside the cartridge to the measurement system. This
parameter is incorporated into the preparation / holding time category listed in Table 2.1. The
initial choices for the methods (25 minutes for RAD7 and RaDeCC; 90 minutes for Rn
emanation line) were based on previously established methods for these measurement systems.
We used our highest activity standard to evaluate each of these systems for shorter sparging
times (Figure 2.5), anticipating lower efficiencies at lower sparging times. For the RAD7 and
RaDeCC systems, the sparge time must be sufficiently long to achieve even distribution of the
222
Rn throughout the air loop, and based on Figure 2.5, that time could be shortened somewhat.
For the Rn emanation line, the sparging is required to quantitatively transfer the 222Rn from the
cartridge to the trap. Figure 2.5 shows that this could be shortened to about 15 minutes,
11
significantly shortening the preparation / holding time for the Rn emanation line in Table 2.1 to
just over 3 hours.
We highlight some important advantages and disadvantages of each method under
different practical scenarios in Table 2.2. During routine, day-to-day 226Ra measurements in the
laboratory, we believe that the RAD7 method offers the user the most benefit. As that system is
automated and can run unattended, a user can start the measurement at one’s convenience and
leave it counting as long is necessary to achieve the desired uncertainty level. The Rn emanation
line and RaDeCC require more user interaction and cannot be left unattended for long periods.
While the gamma spectrometric method is also automated and requires little user interaction, we
feel that its extended counting time is not desirable for high volume throughputs.
Table 2.1. Average preparation times and measurement parameters associated with the sample
set.
RAD7
Rn Line
RaDeCC
γ-Spec.
#
4
4
4
21
Rn Ingrowth time (days)
2
4
3.5
8
Preparation / Holding Time## (hr)
15
1
4
48
Counting Time (hr)
17
5
7.5
56
Total Time per Sample (hr)*
**
0.100
2.000
1.500
0.025
Efficiency (cpm/dpm)
***
10.3
3.9
2.5
12.4
Uncertainty (%)
#
- the time from fiber sealing until a measurement begins
- time necessary to prepare the sample and to allow the polonium daughters to equilibrate
* - sum of preparation / holding time and counting time
##
**
- average of our units used to make the reported measurements
***
- average from all sample set measurements
If a user wants to take a measurement system into the field or to sea, other considerations
become important. With all its required shielding, a gamma spectrometry system is not portable,
so taking it to the field is not an option. While the Rn emanation line is portable, taking it to the
field is not ideal for a number of reasons. First, the maze of air lines is vulnerable during
shipment and can be easily compromised and thus cause a leak. Also, many extra parts are
needed for this system such as a vacuum pump, compressed helium, Lucas cells, photomultiplier
tubes, and liquid nitrogen (or dry ice for the cold trap). The RAD7, being completely selfcontained is the most portable system, only requiring a set of tubing to connect it to a drying
column and the Mn-fiber cartridge. However, the RaDeCC system can be set up in a field
laboratory or on board a ship, thus offering a system readily available to measure other radium
12
isotopes. When possible, we feel that this system offers the greatest advantage for field
excursions.
One can also consider the scenario of operating under a limited equipment budget and
setting up a laboratory capable of measuring 226Ra. Using rough estimates and recognizing that
gamma spectrometer prices can range over a wide scale, we estimate that with about the same
funds needed to buy a middle-range gamma spectrometer, one could purchase about 10 RAD7s,
10 ports of a Rn emanation line with counters and Lucas cells, or 6 RaDeCC counters. The Rn
emanation line requires a substantial amount of assembly, but nonetheless, many more counting
facilities of each system could be purchased for the price of one modern gamma spectrometry
system (Table 2.2).
To further develop this comparison, we examine Table 2.1 for the total time required to
measure a sample with each method. In the time required to count a sample by our gamma
spectrometer to below 10% uncertainty (24 hours; Figure 2.4), a single operator could also
measure one sample by RAD7, one sample by RaDeCC, or two samples with a single port on the
Rn emanation line. The total time required for a Rn emanation line measurement is 5 hours, but
most of this time (3 hours) consists of holding the Lucas cell to allow for radioactive
equilibration, so an operator can start another run on the Rn emanation line during this period.
Considering the number of units of each system that one could purchase for the price of a single
gamma spectrometer, it would be feasible for an operator to measure 6 samples by RaDeCC, 10
samples by RAD7, or 20 samples by Rn emanation line in the time it takes to count a single
sample by gamma spectrometry and without any loss in precision. Therefore, we believe that the
Rn emanation line is the most beneficial technique when dealing with a large number of samples
or a limited time to count them (Table 2.2).
Based on the results from Figure 2.4, we believe that the Rn emanation line also offers
the most benefit when counting low activity samples (Table 2.2). Both the RAD7 and γspectrometer techniques require counting times of hours to achieve an MDA below 3 dpm/100L.
The Rn emanation line, due to its higher efficiency and lower Lucas cell background compared
to the RaDeCC, requires the shortest counting time to achieve this MDA. On the other hand, if a
user also wants to measure 228Ra, the gamma spectrometric method is the only option that can
concomitantly measure this other long-lived radium isotope (Table 2.2).
Comments and Recommendations
The sample set presented in Figure 2.3 was measured according to previously published
methods (summarized above) and the resulting measurement figures of merit were determined
based on these established protocols. However, some advancements to the protocols can be
made to improve the counting statistics for each system. For example, while collecting the
radium samples in the field, simply passing more water through the Mn-fiber will yield more
activity on the fiber, thus enhancing the associated counting statistics. Counting for a longer
period of time will help decrease the measurement uncertainty and MDA level (Figure 2.4) of
each system compared here. One could also preconcentrate the 222Rn before counting it, as is
done on the Rn emanation line with the cold trap. Such an improvement could work with the
RAD7 and RaDeCC systems to improve their efficiencies. Also, if one were to allow 214Po
ingrowth, which is done in the other methods, the RAD7 system efficiency would be nearly
13
doubled, providing lower measurement uncertainties than those reported here for similar
counting times.
Figure 2.4. Minimum detectable activity (solid line) and relative uncertainty (dotted line) plots
for each measurement system showing their dependence on the counting time. The horizontal
dashed lines indicate an MDA of 3 dpm/100L and a relative uncertainty of 10%, and thus the
corresponding measurement time needed by each system to attain that level of precision. The
uncertainty calculations are based on a fiber containing a total activity of 10 dpm 226Ra (10
dpm/100L activity and 100 L volume).
14
We believe that the field of oceanography will benefit from this comparison regarding
various options of measuring 226Ra on Mn-fiber. Some researchers do not have access to γ- / αspectrometry systems to analyze for 226Ra, but could use any of these 222Rn ingrowth methods
instead. Some researchers only use 224Ra and 223Ra in their studies, but the later is often subject
to large measurement uncertainties because of its relatively low natural abundance. When
available, 226Ra is a much better long-lived analog to 224Ra than 223Ra.
Figure 2.5. Results of the sparging time analysis and the corresponding system efficiencies
based on the associated sparging time.
Perhaps the main benefit of this comparison is that researchers will be aware of the
benefits that are afforded from each of these systems. Several previous studies could have
benefitted from the information presented here. For example, during a research cruise along the
South Atlantic Bight, Moore (1996) measured over 70 Mn-fiber samples for 226Ra via gamma
spectrometry to examine submarine groundwater inputs to the region. Using one or two
detectors, these analyses likely required a time frame of several months to complete. However,
an investigation of similar magnitude could be undertaken with several RAD7s on board the ship
(or RaDeCC systems to also allow for 224Ra and 223Ra measurements) and the samples could be
15
16
Relatively inexpensive,
but assembly required
Several per unit per day
Very short counting
time required
Not possible
1 per day per unit
Long counting time
required
Only possible after long
delay
Many samples / Short time available
Low activities
Ra
228
Very little user
interaction needed
Limited user interaction
needed, must start in
morning
Only possible after long
delay
Short counting time
required
1 per day per unit
Moderately expensive
Same measurement
Very long counting time
required
1 per unit every
1-2 days
Very expensive
Not portable
γ-spec
RaDeCC
Must have compressed Need compressed gas,
He, LN2 or dry ice;
but allows other Ra
portable with difficulty isotope measurement
Relatively inexpensive
Very portable, few extra
supplies needed
Limited equipment budget
Taking the system to field/sea
Routine laboratory measurements
Rn Line
Automated - limited Requires user attention,
user interaction required must start in morning
RAD7
Table 2.2 Characteristics (positive and negative) associated with each measurement technique under a variety of scenarios.
measured within days of their collection. With this short delay between sample collection and
measurement, researchers could alter the sampling scheme based on the more rapidly-obtained
results. One can imagine that there would be several benefits of making 226Ra measurements
during the cruise instead of much later. For example, results obtained from these rapid
measurements could lead investigators to revisit a particular site of interest or study a new site
revealed by sample trends.
We recommend the cartridge design presented here for measuring 226Ra since it has
proven to be effective at remaining air-tight throughout various holding times. In addition, the
thick-walled PVC prevents 222Rn diffusional losses out of the cartridge. If any such losses
should occur, we feel that they are minimal and inherently considered in our calibration
procedures. With the cartridge described here, other researchers can adopt this design without
fear of leakage.
17
CHAPTER 3
DETERMINATION OF TRANSPORT RATES IN THE YELLOW RIVER – BOHAI SEA
MIXING ZONE VIA NATURAL GEOCHEMICAL TRACERS
Article published in Continental Shelf Research
Abstract
In light of the current problems facing the Yellow River and surrounding areas (e.g., periods of
zero river discharge, increasing nitrate concentrations of the Bohai Sea), we examined the coastal
mixing dynamics around the mouth of the Yellow River. Naturally-occurring radium isotopes
(223Ra, 224Ra, 226Ra, and 228Ra) and other geochemical tracers (Ba, Si, and salinity) were
employed to determine river plume transport scales and rates. Barium and radium exhibit
elevated concentrations within the salinity gradient where they are desorbed from particles via
ion-exchange. Once they are added to the system, they decrease offshore from dilution with
lower concentration Bohai Sea water, and in the case of 224Ra and 223Ra, by radioactive decay.
Using radium “ages” to assess the dissolved material transport scales and rates proved to be a
useful tool in this environment. The ages based on the 224Ra/228Ra activity ratio increased
gradually until salinities reached ~25 when they rapidly increased due to decreasing mixing at
higher salinities. Integrated net transport rates through the salinity front ranged from 1.4 to 1.6
cm/s and did not vary significantly with river discharge. Thus, tidal mixing appears to dominate
in this system, at least over the range of discharges investigated (80 - 600 m3/s). Determining the
temporal scale of flow across the coastal zone in this region is a valuable first step toward
examining whether the Yellow River is contributing to the increasing inorganic nitrogen
concentrations in the central Bohai Sea.
Introduction
The Yellow River (Huanghe) flows for a total of 5464 km, draining a land area of
752,000 km2 into the Bohai Sea (Figure 3.1) (Wang et al., 2006). It is the sixth largest river in
the world in terms of length, and delivers a sediment load of 10.4 x 108 tons/year to the delta and
the Bohai Sea. As a result of continued sedimentation from its very high sediment load, the
Yellow River channel bed lies up to 11 m above the surrounding lands, and dikes must
continually be augmented to prevent catastrophic flooding. Hyperpycnal, or fluid mud flows, are
even known to occur in the Yellow River (Yu, 2002). The annual average discharge at Lijin
(~110 km upstream of the mouth) has decreased from 4.2 x 1010 m3 (1951-1980) to 2.1 x 1010 m3
(1981-2000) (Wu et al., 1998; Chen et al., 2005). Recent increases in upstream water use have
put serious pressure on available water resources, even producing periods of zero flow in the
delta area. In 1997, for example, there was a span of 226 days when the river did not reach the
ocean. Since then, the Chinese government has implemented management measures to regulate
water discharge from upstream reservoirs to prevent any future periods of zero flow.
18
This fluctuating discharge combined with a growing population in the river basin has
undoubtedly altered the riverine nutrient flux to the coastal Bohai Sea. Since 1960, nitrate
concentrations in the central Bohai Sea have increased by a factor of 10, while phosphate
concentrations have decreased by a factor of 2 (Zhang et al., 2004). These authors observed a
shift to a phosphate-limited ecosystem in Laizhou Bay (a large embayment just south of the
Yellow River mouth) and proposed that the entire Bohai Sea could be approaching this
condition. The Yellow River, being the largest river to empty into the Bohai Sea, is a possible
source of these excess nitrate concentrations. Submarine groundwater discharge and
atmospheric sources (wet and dry deposition) are some other possible sources (Raabe et al.,
2004; Wei et al., 2004; Liu and Yin, 2007).
While recent increases in nitrate concentration in the Yellow River have been observed,
the annual river discharge has been steadily decreasing, thus an increased flux of nitrate to the
Bohai Sea from the river is not necessarily the case (Shen and Le, 1993; Huang et al., 2005).
Clearly, there were periods of zero nitrate flux when the river did not meet the ocean. The
question about whether the Yellow River can be a source of the increasing nitrate concentrations
in the central Bohai Sea thus remains unanswered. Before this can be directly addressed, we
must first examine the transport rates of river water-borne components as they are transported
offshore.
Radium isotopes have become increasingly recognized for their utility in examining
mixing processes in the coastal ocean (Carroll et al., 1993; Moore, 1997; 2000b; Nozaki et al.,
2001; Charette et al., 2007) and are the main geochemical tracers employed in this work. There
are four naturally-occurring isotopes of radium (223Ra, 224Ra, 226Ra, and 228Ra) and their range in
half-lives (11.4 days, 3.6 days, 1600 years, and 5.7 years, respectively) makes them especially
useful for studying processes that cover a wide span of time scales.
In freshwater environments, both radium and its nonradioactive analog, barium, are
particle reactive, so are mostly found adsorbed onto suspended solids. Once encountering saline
waters, ion exchange processes result in radium and barium being released into solution. Many
investigators have documented estuarine radium and barium desorption processes from
suspended particles (Hanor and Chan, 1977; Li and Chan, 1979; Coffey et al., 1997; Martin and
Akber, 1999; Nozaki et al., 2001), with the general conclusion that barium and radium tend to
behave similarly with respect to estuarine distribution. This study provided the opportunity to
examine the estuarine relationship between radium and barium in one of the world’s most
sediment-laden rivers.
The main purpose of this study was to examine the scale and timing of the Yellow River
mixing processes with the Bohai Sea and how those transport processes may vary with river
discharge. We used a suite of geochemical tracers (e.g., radium isotopes, barium, silica, salinity)
to examine the timing and distance scales of these transport processes. Knowledge of the
transport rates, in particular, is fundamental to determining whether the Yellow River is
contributing to the increasing nitrate concentrations in the central Bohai Sea. The Yellow River
can only contribute nutrients to the central Bohai Sea if transport rates are fast enough to prevent
significant nitrogen uptake and sedimentation in the coastal zone before reaching the shelf
waters.
19
Figure 3.1. Map of the study site at the Yellow River mouth and surrounding Bohai Sea.
Methods
Sampling and analytical methods
Samples were collected during expeditions in September 2004, May 2005, and
September 2006. The average discharge (± standard deviation) of each sampling interval was
392 ± 147 m3/s, 81 ± 14 m3/s, and 568 ± 69 m3/s, respectively (Figure 3.2). The wet season in
this part of China is typically during the summer and early fall (June – September). The two
September sampling periods were thus timed to coincide with the end of the rainy season while
the May expedition was intended for a dry season sampling. As the Yellow River is heavily
dammed, much of the discharge is artificially controlled.
During each field campaign, samples were collected in transects starting from a floating
bridge 14 km up river, and then focusing on sampling offshore across the salinity gradient. All
samples and measurements were made from local fishing boats hired during each expedition.
Due to Chinese boating regulations, we could only extend the transects 15 km offshore. During
the May 2005 field trip, the winds were too strong to permit sampling more than 2 km offshore.
Nevertheless, since that sampling trip occurred during low discharge conditions, the entire
salinity gradient was sampled within this distance. Samples were collected for radium isotopes
during each campaign, barium (May 2005 and September 2006), and silica (May 2005 and
September 2006).
Radium samples were collected using acrylic fiber impregnated with manganese-dioxide
that quantitatively adsorbs dissolved radium from water (Moore and Reid, 1973; Moore, 1976).
Large volumes of surface water (~100 L) were pumped through cartridges containing this Mn-
20
fiber at around 1 liter per minute to fully extract the radium from the water. In addition, Mnfibers were towed behind the boat in a mesh bag for ~30 minute intervals. These samples cannot
be used to quantify the radium activity in the waters, as the sample volume is unknown, but are
still useful for determining the activity ratio (AR) of the radium isotopes in the water (Dulaiova
et al., 2006; Burnett et al., 2007a). All fibers were then washed thoroughly with Ra-free
deionized water and first counted for the short-lived radium isotopes (223Ra and 224Ra) using a
delayed coincidence counting system (Moore and Arnold, 1996) and later for the long-lived
radium isotopes (226Ra and 228Ra) via gamma spectrometry (Dulaiova and Burnett, 2004).
Figure 3.2. Yellow River hydrograph based on daily average discharge, with each sampling
period highlighted.
Samples for barium and silica were filtered on site through 0.45 μm nylon filters and
stored dark and cold until they could be analyzed. Barium was measured using an Agilent
Model 7500 CS ICP-MS with an Octopole Reaction Cell. Silica was determined
spectrophotometrically using an Agilent Model 8453 UV-visible spectrophotometer. The
method was a modification (Mortlock et al., 1993) of the Si-method reported by Fanning and
Pilson (1973).
Conductivity and temperature were logged continuously during each transect using a YSI
600R Sonde, and these parameters for each bulk radium sample were measured independently
using a YSI 85 handheld conductivity meter. Total suspended solid (TSS) concentrations were
determined by filtering 50 mL of sample water though pre-weighed 0.45 μm membrane filters.
21
Transport rate calculations
Coastal mixing rates are commonly determined according to the method described by
Moore (2000b). Radium activities are highest where freshwater meets salt water via desorption
from particles. The distribution of the short-lived radium isotope activities along a transect from
the estuary decrease exponentially offshore, so by plotting the natural logarithm of the shortlived radium isotopes versus distance offshore produces straight line segments. Assuming that
all the radium desorbs from the suspended particles at nearly the same location in the salinity
gradient, and that there are no other inputs of radium beyond the near shore environment, the
exponential decrease is due to a combination of radioactive decay and diffusive mixing with low
radium ocean water offshore. Since decay rates are known, mixing coefficients can be derived
based on the slope (m) of the best-fit line to the natural log plot by:
m=
λ
(3.1)
Kh
where λ is the decay constant of the isotope (0.189 day-1 for 224Ra, 0.060 day-1 for 223Ra), and Kh
is the mixing rate, with units of distance squared per time.
This model was initially derived to examine regional-scale mixing of the South Atlantic
Bight (Moore, 2000b) by assuming that the mixing is dominated by eddy diffusion and
neglecting advective mixing. In addition, the model also assumes that the system is in steadystate, at least over the time frame integrated by these radium isotopes (i.e., days to weeks).
However, our work on the interactions of the Yellow River with coastal Bohai Sea water is much
more local in scale, and so these assumptions may not hold. In addition, as will be shown later,
the trends of our salinity transects indicate that advection cannot be ignored.
Another approach using radium isotopes is to determine the length of time that has passed
since the short-lived radium isotopes were added to the system, i.e., since the radium was
desorbed from the suspended particles in the salinity gradient (Moore, 2000a). This time is
referred to as the “apparent radium age” of the water mass. These ages are considered apparent
as they are based on radioactive decay of the radium since first being introduced to the water
mass, and so are not direct measurements of the age of the water mass itself. Nonetheless, this
analysis still yields useful information concerning the history of dissolved components in the
water masses sampled (Moore and Krest, 2004).
This model relies on the assumptions that the radium input is dominated by one source
with a constant isotopic composition. In an estuary such as the Yellow River delta, the input to
the system would be mainly from particle desorption occurring at a constant salinity, which is
assigned the apparent radium age of 0 days. The radium “source” may thus be defined as the
point where the activity ratio of a short-lived radium isotope (223Ra or 224Ra) to a longer-lived
isotope is highest. As the waters age and mix offshore, the short-lived isotope will decay while
the longer-lived isotope (as 228Ra) will only experience minimal decay. Thus, by examining the
difference between the AR at any point and the initial ratio, one can derive an apparent radium
age of the water (t) at that location:
⎛ AR
t = Ln⎜⎜ obs
⎝ ARi
⎞
1
⎟⎟ *
⎠ λ228 − λ224
(3.2)
22
where ARobs refers to the activity ratio of the isotopes at a specific sampling location, while ARi
represents the initial activity ratio when the radium first desorbs from the particles. This initial
ratio is taken from the sample with the highest activity ratio collected in the salinity gradient,
since it would show the least amount of decay. As shown, equation (3.2) utilizes the 224Ra/228Ra
AR, but other isotope ratios may be applied as well (e.g., 224Ra/226Ra, 224Ra/223Ra). Both types of
radium samples, collected water samples with measured volumes and towed fibers, are
applicable here, while only the collected and measured volumes would be applicable to the
mixing rate analysis shown by equation (3.1).
Results
Horizontal salinity profiles through the estuary differed significantly among the three
sampling trips (Figure 3.3). In September 2004, during a relative high flow period, salinity did
not increase until offshore of the river mouth. The profile shows a rather smooth increase to full
saline waters roughly 3.5 km offshore. September 2006, while still experiencing relative high
discharge, shows a much more variable horizontal profile. The salinity rises quickly within the
first kilometer offshore from the mouth, but then varies significantly over the next 12 km until
finally stabilizing at full saline conditions. These fluctuations are likely due to eddies created
due to wind and tidal influences on the river discharge plume. During the low discharge period
in May 2005, the entire salinity gradient was contained within the river channel requiring nearly
4 km to shift from fresh to saltwater. Note that in this delta, our definition of the “river mouth”
is a somewhat arbitrary, yet constant (based on GPS coordinates) definition, since the low-lying
coastal delta is constantly shifting its position. In fact, maps of this area are always out-of-date
due to the dynamic nature of the coastline.
As an example of our tracers’ behavior in the estuary, Figure 3.4 shows the September
2006 distribution of silica (a), barium (b), 224Ra (c), and total suspended solid concentration (d)
as well as salinity along a seaward transect from the Yellow River. The river water is enriched
in silica relative to the Bohai Sea, so the silica distribution through
the salinity gradient is likely due to the combined effects of biological uptake and dilution with
low silica offshore waters. Barium and radium, however, exhibit the expected peak within the
salinity gradient where they are presumably desorbed from particles via ion-exchange processes.
Once they are added to the system by desorption, they decrease offshore from dilution with
lower concentration Bohai Sea water and, in the case of 224Ra, by radioactive decay. Around 8 to
10 kilometers offshore of the river mouth, both tracers reach the background concentration of the
shelf waters of the Bohai Sea, and level off at that concentration. The variability in the radium
and barium data is likely the result of short-term tidal and river discharge fluctuations creating
pulses of both fresher and more saline water pockets offshore (revealed by the salinity
fluctuations). Over the time scale of interest (i.e., days), these short-term fluctuations would be
smoothed out in the general trends of the tracers. The estuarine radium and barium peaks occur
in the same region where both the salinity increases and the TSS decreases. Just offshore of the
river mouth, the suspended particle concentration decreases drastically, so sediments in
suspension are not as available to provide any desorbable radium and barium to the surrounding
water column.
23
Figure 3.3. Horizontal salinity profiles from all three sampling trips. Negative distances
represent salinities up the river, whereas positive distances indicate offshore values. The river
“mouth” is somewhat arbitrary, but is at a constant position based on GPS coordinates.
The direct relationship between barium and radium is shown in Figure 3.5. The
significant linear correlation illustrates their similar estuarine behavior, implying that there likely
is not a time lag between radium and barium desorption in this system. More intense sampling
specifically designed to examine this relationship would be required throughout the estuary to
identify any indication that either barium or radium desorb earlier than the other.
Figure 3.6 illustrates the horizontal distribution of all four radium isotopes from the three
sampling trips. The upper two panels (a and b) show the short-lived isotopes have their highest
activities in the salinity gradient due to desorption from suspended particles and then decrease
offshore due to radioactive decay and mixing with a lower activity Bohai Sea water. It is evident
that in May 2005 (low discharge), the absolute activity concentrations reached the highest
measured levels observed in the three sampling campaigns. The bottom two panels of this figure
(c and d) show the distribution of the long-lived isotopes through the estuary. The data from
September 2004 shows much more variability, but the general trends show that the elevated
activities of 226Ra within the salinity gradient remain nearly constant while 228Ra increases
offshore. This indicates that the river and diffusion from bottom sediments are constant sources
of long-lived radium isotopes to the shelf waters of the Bohai Sea. Other authors have observed
an offshore enrichment in 228Ra in similar estuaries (Nozaki et al., 2001; Dulaiova et al., 2006).
Several other important concepts are revealed in Figure 3.6. First, the samples collected at
the extreme locations (farthest upriver and offshore) indicate that regardless of the season of
collection, the activities at these locations remain nearly constant. This has important
implications for the radium age models described above, since we assume that the radium
activity ratios in the primary source waters remain constant. Secondly, the measurement
uncertainties associated with the 223Ra tend to be around 40% of the measured activity due to its
relatively low activity concentrations in these samples. Since the other isotopes show
24
uncertainties on the order of 10%, the short-lived to long-lived ARs would have higher analytical
precision using 224Ra rather than 223Ra.
Figure 3.4. Dissolved silica (a), barium (b), 224Ra (c), and total suspended sediment
concentration (d) as measured across the salinity gradient in September 2006. The solid line in
each plot shows the horizontal salinity profile. Error bars reflect 1-σ measurement uncertainties.
Discussion
In order to use the model approach described by equation (3.1) to determine coastal
mixing rates from the distribution of short-lived radium isotopes, the system must be in steady25
state and advective mixing should be negligible (Moore, 2000b). However, since the salinity
gradients show significant structure in the offshore direction (Figure 3.3), we conclude that
neither of these assumptions is correct for this system. This structure in the salinity gradient is
likely a result of the tidal influence on the discharge of the river, so the assumptions required by
the model represented by equation (3.1) likely do not hold.
Figure 3.5. September 2006 barium (nM) vs. 224Ra (dpm/100L) relationship in samples along
offshore transect. Error bars represent 1-σ measurement uncertainty.
Instead, we can apply the model involving equation (3.2), since there are no limiting
assumptions about steady-state conditions, with the exception of the constant initial AR (Moore,
2000a). The source water to the estuary does not show any significant temporal variation in the
dissolved radium activities, and there is no reason to believe that the suspended sediment-bound
radium activities would change temporally. Provided that the dominant radium source to the
coastal zone is found within the salinity gradient, this model can be applied.
Applying the model based on equation (3.2), we determined apparent radium ages of all
samples from each field excursion. Figure 3.7 shows these results using the 224Ra/223Ra isotope
ratio (a), 224Ra/226Ra ratio (b), and the 224Ra/228Ra ratio (c). The sampling points farther upriver
than the salinity front are not included, as the radium has not fully desorbed from the suspended
particles. While the data from May 2005 (dry season) are much more horizontally compact, it
does not appear that the ages increase faster than those collected during the higher discharge
periods.
26
Figure 3.6. Distribution of 223Ra (a), 224Ra (b), 226Ra (c), and 228Ra (d) activities in dpm/100L for
all three sampling campaigns. Both 223Ra and 224Ra peak in the salinity front and decrease
offshore due to decay and mixing. The long-lived isotopes remain constant (226Ra) or increase
(228Ra) offshore.
27
Figure 3.7. Horizontal distribution of apparent radium ages calculated from 224Ra/223Ra (a),
224
Ra/226Ra (b), and 224Ra/228Ra activity ratios (c).
Assigning an initial AR is a source of uncertainty in using this model. The most common
criteria utilized in assigning the initial ratio is to use the highest measured AR in low to midsalinity waters (Moore and Krest, 2004; Burnett et al., 2007a). The source AR must be at least as
great as the highest sample, so this method provides a conservative estimate. We attempted to
quantitatively desorb radium off Bohai Sea coastal sediments to determine the desorbable radium
activity from suspended particles, but our determined ARs from this experiment were
significantly lower than most ARs we measured in the estuary. This is likely due to using
sediments that have previously encountered saltwater and so their long-lived radium activities
are lower than those on suspended riverine sediments. Instead, for the purposes of assigning a
28
reasonable initial AR to these data, and maintaining the assumption concerning a constant input
AR, we elected to use the highest measured AR found in the low to mid- salinity range for each
of our radium isotope ratios. Table 3.1 summarizes the selected initial ARs, as well as the
corresponding salinities and sampling expeditions during which they were measured.
Table 3.1. Summary of the initial AR parameters used to model the apparent radium ages of
Yellow River plume waters.
We have included the slopes and correlation coefficients (r values) of the best-fit linear
lines for the results presented in Figure 3.7 in Table 3.2. The corresponding transport rate
reported for each group is derived from the reciprocal of these slopes, converted into units of
cm/s. The effects of using 223Ra values in determining mixing rates are apparent, as the r values
for the regressions using the 224Ra/223Ra AR are nearly half of those found for 224Ra/226Ra and
224
Ra/228Ra AR regressions. While we recognize that linear relationships between apparent
radium age and distance offshore from a river are not necessarily the expected result, we find
that in the case of the Yellow River estuary, linear trends fit the data reasonably well.
Because of the improved analytical precision of the 224Ra and 228Ra results (~ 10%)
mentioned earlier, we will use the 224Ra/228Ra AR to further examine transport rates. Note that
using this AR to determine apparent radium ages yields slightly lower transport rates than does
the 224Ra/226Ra AR. Figure 3.8 shows the data from each sampling campaign separately, and
displays the 95% confidence interval of the linear trend of the data, and shows the regression line
together with the standard error of the slope. These slopes (day/km) and uncertainties are then
converted to transport rates in terms of cm/s by a simple unit conversion. The corresponding
range in possible transport rates indicates that there is significant overlap between all three
sampling trips, and therefore we conclude that there is no significant difference between the
transport rates over the range of river discharges investigated (80 – 600 m3/s). In addition, based
on the uncertainty in mixing rates, there is generally no statistical difference between using the
224
Ra/228Ra and the 224Ra/226Ra AR.
A similar study of the mixing of Mississippi and Atchafalaya Rivers into the Gulf of
Mexico (Moore and Krest, 2004) provides a comparison to our transport velocities. These
29
authors determined the transport rate range that would describe all their radium isotopic data
points. We find that our age gradient results are lower than the calculated transport rates for the
Mississippi River, which ranged from 2.3 to 17 cm/s. This is not surprising as the Mississippi
River has a significantly greater discharge than the Yellow River. During the two sampling
periods of the Moore and Krest study, the Mississippi had average discharges of 11,200 m3/s and
14,500 m3/s, roughly 20 times higher than the discharges of the Yellow River.
Table 3.2. Summary of the linear regression results from apparent radium age versus distance
offshore plots found in Figure 3.7. Transport rates are converted from the reciprocal of the
regression slopes to transport rates of cm/s.
In another investigation of the interaction of the Mississippi River plume with the Gulf of
Mexico, Shiller (1993) described how the rate of salinity change slows offshore as the salinity
increases. This effect is due to higher plume salinities requiring more mixing with high salinity
offshore waters to further increase the plume salinity. In the Yellow River estuary, the
224
Ra/228Ra ages increase gradually until the salinities reach ~25 when they rapidly increase due
to decreasing mixing at higher salinities (Figure 3.9) as suggested by Shiller (1993) and also seen
in the data of Moore and Krest (2004). It is also clear that the September 2006 ages (during the
highest sampled discharge) increase the greatest, while the May 2005 (lowest discharge) increase
the least.
30
Figure 3.8. Apparent radium ages calculated from the 224Ra/228Ra activity ratios from 2004 (a),
2005 (b), and 2006 (c). Best-fit equations are shown with their 95% confidence levels. The
transport rates shown are calculated from the slopes of the best-fit lines.
While knowledge of these one-dimensional transport rates are useful, information
regarding the area of the coastal ocean over which the river plume spreads in a certain amount of
time would be more useful. With an estimate of the river plume dispersion angle, we can
extrapolate the modeled linear transport lengths to theoretical effective mixing areas. We
estimated the angle at which the Yellow River plume spreads into the Bohai Sea to be roughly
80° based on measuring the angle of the lighter water off the Yellow River mouth from Google
Earth satellite images. Since the delta area is currently increasing by 20-25 km2/yr (0.054 –
0.068 km2/day) (Chen et al., 2007), we assume that these are areas where most of the sediment
deposition is occurring. We use the calculated transport rates to represent the linear mixing
distance per unit time and calculate the daily mixing area of the river plume based on the angular
relationship. This assumes a semi-circular geometry (hemisphere) (i.e., A = (πr2)/2), with the
transport distance calculated from the 224Ra/228Ra AR being the radius of the circular mixing
31
zone. From the average of our transport rates determined from the 224Ra/228Ra AR, we determine
this mixing area to be about 1.3 km2/day. These estimates are based on the further assumptions
that: (1) the river plume mixes at a uniform rate in all directions within the 80° plume area; and
(2) the plume angle does not vary with river discharge.
Figure 3.9. All apparent radium ages calculated from the 224Ra/228Ra isotope ratio plotted against
their corresponding salinities.
Conclusions
During the wet season in the Yellow River basin (September), we find higher river
discharge, the salinity gradient is located offshore of the mouth of the river, and offshore
transport rates of around 1.6 cm/s based on the 224Ra/228Ra radium age distribution. Conversely,
during the dry season (May), we find much lower river discharge, the salinity gradient contained
within the river channel, but transport rates still around 1.4 cm/s.
Even though the Yellow River estuary is not steady-state on a time scale of up to about
two weeks (based on the salinity gradient), we were able to determine coastal transport rates
using the activity ratio distribution of radium isotopes. For a dynamic system where advective
mixing is likely, one may examine the range of transport rates required to distribute the radium
isotopes through an estuary via the apparent radium ages of the samples. Since the “transport
rates” varied in a narrow range (1.4 – 1.6 cm/s) in spite of relatively large fluctuations in river
discharge (~80 – 600 m3/s), we conclude that tidal mixing must dominate in this system, at least
over the range of discharges investigated.
32
The mixing processes examined here have important implications for the nutrient input to
the open Bohai Sea. One can apply these results to dissolved substances carried by the river
plume (e.g., nutrients, contamination) to examine the rates at which they are transported
offshore. Incorporating nutrient uptake rates from primary productivity data would allow one to
determine how far the riverine nutrients would be transported before being consumed. This type
of calculation would assist area managers determine whether the Yellow River is a significant
contributor to the increasing central Bohai Sea nitrate levels.
33
CHAPTER 4
RADON AND RADIUM ISOTOPE ASSESSMENT OF SUBMARINE GROUNDWATER
DISCHARGE IN THE YELLOW RIVER DELTA, CHINA
Article published in Journal of Geophysical Research
Abstract
Naturally-occurring chemical tracers were used to assess the magnitude of submarine
groundwater discharge (SGD) during two different sampling periods at a coastal site south of the
Yellow River delta, China. We used salinity and pH as indicators of the terrestrial and
recirculated seawater components of discharging groundwater and radium isotopes to quantify
offshore transport rates. We then used an hourly time series of multiple radium isotopes (224Ra,
223
Ra, and 226Ra) to quantify SGD rates, and also used 222Rn and seepage meters to
independently quantify SGD rates as a comparison to the radium results. Offshore transport
rates were found to range from 3.3 to 4.7 cm/s. Modeled time-series radium activities indicated
average SGD rates ranging from 4.5 to 13.9 cm/day in September 2006 and from 5.2 to 11.8
cm/day in July 2007. Temporal trends associated with the radium approach agree with SGD
patterns revealed by automated seepage meters deployed nearby, but the absolute fluxes are
about 70% lower than those determined by the seepage meters. Modeled SGD rates based on
222
Rn (mean = 13.8 cm/day in 2006 and 8.4 cm/day in 2007) agree with those determined by the
radium analysis. Differences in derived SGD rates between the different radium isotopes (226Ra
highest; 224Ra lowest) are likely results of uncertainties in the background activities and our
limited selection of appropriate groundwater/pore water end-member values. Scaling our results
to the entire Yellow River delta, we find SGD fluxes (and corresponding nitrate fluxes) 2-3 times
that of the Yellow River.
Introduction
Submarine groundwater discharge (SGD) is now regarded as an important pathway that
transports dissolved substances from sub-seabed fluids to the coastal ocean. Being both spatially
and temporally variable, SGD is very difficult to measure and therefore its relative importance in
coastal ocean chemical budgets is often unknown (Burnett et al., 2006a). Nonetheless, nutrient
fluxes via SGD have been shown to rival those from rivers in some locations (Slomp and Van
Cappellen, 2004; Kim et al., 2005; Swarzenski et al., 2007a). Providing an additional nutrient
source to the coastal ocean can often be beneficial to the coastal ecology (e.g., Santos et al.,
2008), but effects from SGD can also be harmful in areas where terrestrial groundwaters are
contaminated from anthropogenic sources (Hu et al., 2006).
By definition, SGD includes all water moving across the sediment-water interface and
into the overlying water column, regardless of composition or driving force (Burnett et al.,
2003b). This designation includes both fresh, terrestrial groundwater and saline, recirculated
34
seawater. The hydraulic gradient is the main driving force that results in fresh, terrestrial aquifer
waters discharging at the coastline. Driving forces controlling recirculated seawater include tidal
pumping, wave setup, and convective circulation caused by subterranean aquifer density
differences (Michael et al., 2005; Charette et al., 2007).
The Yellow River delta is an area where these driving forces can interact uniquely
because of some unusual geological characteristics. The high rate of sediment supply from the
Yellow River results in an annual progradation of the delta into the Bohai Sea by 20-25 km2
(Chen et al., 2007). This growth creates large expanses of low-lying land extending into the
Bohai Sea each year. The result is a dampening of the hydraulic gradient between the land and
ocean, thereby increasing the relative influence of marine effects as a driving force. This high
sediment supply, however, has also created a perched riverbed in the lower reaches of the river,
so the Yellow River lies up to 11 m above its surroundings. This effect creates the potential for
the river to recharge the groundwater, enhancing the hydraulic gradient between the river and the
coast (Yu, 2002).
The Yellow River delta is a location where SGD can potentially introduce contaminated
groundwater to the coastal ocean. As a result of China’s growing population, more people are
inhabiting the delta than ever before and are contributing to groundwater contamination, mainly
through agricultural practices and sewage disposal. The groundwater environment around the
delta contains elevated nitrate levels (concentrations up to 3.8 mM), mainly in shallow aquifers
(Chen et al., 2007). These nitrate-rich zones often coincide with agricultural land use patterns
and concentrated population centers. Zhang and others (Zhang et al., 2004) determined that
within the last 40 years, DIN concentrations have increased ten-fold, coincident with reductions
in phosphorus (50-60%) and silica (75%) in the central Bohai Sea. This trend has led to portions
of the Bohai Sea shifting to a phosphate-limited ecosystem. The source of the excess nitrogen,
however, remains unknown.
One potential source of this nitrogen could be SGD. Because Chen et al. (2007) found
excessively high concentrations of nitrate in contaminated groundwater, even a relatively small
volumetric flux of SGD could provide a significant input of nitrate to the coastal ocean. Once
these nutrients are introduced to the coastal zone via SGD, they must be transported offshore to
the central Bohai Sea to contribute to the increasing nitrogen concentrations. In addition to
quantifying SGD rates, knowledge of dissolved component transport rates in this region is thus
an important factor for determining whether SGD can be a significant source of these nutrients
on a regional basis.
Taniguchi and others (2008) used automated heat-type benthic seepage meters (Taniguchi
and Iwakawa, 2001) to measure SGD in a down-gradient area of the delta approximately 40 km
south of the Yellow River estuary (Figure 4.1). Combining the results from the seepage meters
with conductivity (salinity) measured inside the chambers allows for the separation of fresh SGD
from total SGD. They found that fresh groundwater fluxes are between 1 and 5% of total flow,
and integrating their results over the estimated zone of discharge in the delta yielded freshwater
SGD flow estimates that ranged from 4.5 to 7.0% of the Yellow River discharge.
The goals of the work presented here are to use naturally-occurring geochemical tracers
(e.g., radium isotopes, 222Rn, salinity, pH) to examine patterns and fluxes of SGD in the same
area studied by Taniguchi et al. (2008). This paper represents one of the first to use a 24-hour
high-resolution time series of multiple radium isotopes in order to quantify SGD rates. Beck et
al. (2007; 2008) have made similar attempts, but only over ~6 hour intervals. We also use
35
parallel 222Rn and benthic seepage meter measurements to provide an independent comparison of
these fluxes.
Chemical tracers measured in the water column integrate the SGD signal over much
larger spatial scales than individual benthic chambers, and are thus an appropriate way to handle
the large spatial variability inherent in groundwater discharge patterns. In addition, horizontal
offshore transects of radium isotopes allow us to quantify the transport rates of the discharged
groundwater offshore. Radon and radium isotopes have been shown to be effective tracers of
SGD because they are concentrated in groundwater relative to surface water, and the decay rates
of the 222Rn and the short-lived radium isotopes are on the same temporal scale as the processes
in question (Burnett and Dulaiova, 2003; Kim et al., 2005; Dulaiova et al., 2006; Burnett et al.,
2007a; Swarzenski et al., 2007a). We use three radium isotopes in this study, with half-lives that
cover a large temporal range: 224Ra (T1/2: 3.6 days); 223Ra (T1/2: 11.3 days); and 226Ra (T1/2: 1600
a). In freshwater, radium will be found mostly adsorbed onto the surface of particles, but once
introduced to saltwater, ion exchange processes displace the radium off particles and into
solution (Li and Chan, 1979; Nozaki et al., 2001).
Methods
Measurements
Samples were collected in September 2006 and July 2007 from a coastal site located 40
km south of the Yellow River mouth (small square on the west side of Laizhou Bay, Figure 4.1)
in an area toward which groundwater contours indicate subsurface flow should occur (Chen et
al., 2007; Taniguchi et al., 2008). In September 2006, we collected a continuous 25-hour timeseries (TS-1; 1 km offshore; av. depth: 1.4 m) of radium isotopes, 222Rn, pH, and conductivity.
During the July 2007 sampling, these tracers were measured in both a 23-hour time-series (TS-2;
2 km offshore; av. depth: 1.7 m) and a 12 km shore-normal transect.
Radium isotopes were collected according to the methods established by Moore and Reid
(1973). Large volume samples (~ 100L) were pumped slowly (1 L/min) through columns
containing acrylic fibers impregnated with manganese-dioxide. This Mn-fiber quantitatively
sorbs the dissolved radium from the water. These fibers were then washed thoroughly to remove
all particles and counted for the short-lived radium isotopes via a delayed coincidence counting
system (Moore and Arnold, 1996). 226Ra was counted by either gamma spectrometry (Dulaiova
and Burnett, 2004) or by measuring 222Rn and its daughters after sealing the fiber in air-tight
columns to allow for ingrowth. These columns were later mounted to a radon emanation line or
a commercially available radon-in-air monitor (RAD7, Durridge Co.) to measure 222Rn as a
proxy for the 226Ra (Kim et al., 2001). All samples from TS-1 were counted by gamma
spectrometry, while the samples from July 2007 were analyzed for 226Ra using the ingrowth
method. Cross-calibration shows that the methods agree within 20%.
36
Figure 4.1. Map showing the Yellow River delta and the SGD study site (box on the west side of
Laizhou Bay).
222
Rn in water was measured continuously in the field using a modified RAD7 radon-inair monitor (Burnett et al., 2001). Surface water (~ 0.5 m below the surface) was pumped to an
air/water equilibrium exchanger system. In this system, the headspace air is circulated to the
RAD7 for analysis of 222Rn activity in the air, and recycled back to the exchanger, creating a
closed air loop. Applying a temperature-dependent solubility coefficient for 222Rn, we can
convert from measured radon in air to the corresponding value in the water. Each data point
attained represents an integrated value over 30 to 60 minutes, depending upon the desired
measurement uncertainty.
Groundwater samples for 222Rn and radium were collected on land using a peristaltic
pump from boreholes around the delta, and pore water samples from offshore (~50 cm below the
sediment-water interface) were collected from a push-point piezometer (Charette and Allen,
2006). Groundwater and pore water samples for radon were measured using a RAD-H2O system
that uses the internal pump of the RAD7 to sparge radon from a 250 mL volume and circulate it
to the counter for measurement. The pH and electrical conductivity were measured using a
handheld YSI Model 85 probe.
37
Transport rate model
Moore (2000a) described a method of using radium isotope activity ratios (AR) to
determine the apparent radium age of water masses. The ages represent the relative time that has
elapsed since the radium first entered the system. They are based on the exponential decay of the
short-lived radioisotopes from their original input signature. If radium is input to the coastal
ocean with a constant short-lived to long-lived AR over time, and no additional sources of
radium exist, we can examine the difference between the initial activity ratio (ARi) and a
measured ratio (ARobs) to calculate the apparent radium age (t):
⎛ ARi
t = Ln⎜⎜
⎝ ARobs
⎞
1
⎟⎟ *
⎠ λ 224 − λ 226
(4.1)
where λ224 and λ226 represent the decay constants of the short-lived radium isotope (e.g., 224Ra)
and that of the longer-lived isotope (e.g., 226Ra), respectively. Defining the ARs as the activity
of a short-lived isotope to that of a longer-lived isotope dictates that they always decrease with
time.
The relative differences between the radium ages across a transect should indicate the
actual mixing times required to distribute the observed isotope ratios. Therefore, plotting the
apparent radium ages against their distance from shore (assuming the source is near the
shoreline) can yield an estimate of the offshore linear transport rate of radium (Moore and Krest,
2004; Peterson et al., 2008b) and all conservative constituents dissolved in the water.
Radium time-series model
In addition to determining coastal transport rates, we can use radium as a tracer to
quantify SGD rates. At our study sites, the only inputs of radium should be from the Bohai Sea
shelf waters (assumed to be very low), desorption from suspended sediments, desorption and
subsequent diffusion from bottom sediments, and advection of groundwater into the overlying
water column. The only important sink for radium on an hourly basis is mixing.
We have assumed that we can neglect riverine inputs at this location because the seasonal
currents in the Bohai Sea carry the Yellow River freshwater plume away from our site and
because there are no other significant rivers discharging in this area. Several authors have
demonstrated that during the summer months, the monsoonal winds blow from the south in this
region, thus creating a cyclonic gyre within the Bohai Sea (Hainbucher et al., 2004; Wang et al.,
2007). Since the Yellow River is located 40 km north of our study site and its plume is directed
further to the north in the summer, it is very likely that it has no influence over the tracers at our
study site.
Our model to quantify SGD from radium adjusts the measured radium activities (per
square meter of seabed) for offshore mixing contributions and sedimentary inputs to define an
excess radium inventory that must be supplied by SGD. It is difficult to accurately assess the
effect of mixing on our measured inventories, because we are sampling at one point within a
large, shallow continental shelf. Currents and tidal forces move water in all directions, so
determining an offshore “end-member” value for radium activity is inappropriate. Instead, we
choose to use the lowest activity measured during the course of each time-series as an indicator
of the shelf-wide “background” radium activity. As some of this radium may be supplied by
38
SGD, this should represent a maximum estimate of non-SGD sources of radium and therefore
will lead to conservative SGD estimates.
Desorption from sediments is likely a significant source of radium to these shelf waters,
especially for the short-lived isotopes. These waters are saline (S ~ 29), so we can assume that
any radium initially present upon original input of the sediments has already been desorbed.
Therefore, the only source of desorbable radium on both the suspended and bottom sediments is
from decay of the radium parents (insoluble thorium). By accounting for the “background”
radium activity as described above we should have already accounted for diffusive and
desorptive inputs from sediments.
The following steps were used to derive SGD rates from a time-series of radium samples:
(1) Adjust each measured radium activity (Ratotal) to represent the excess radium activity
by subtracting out the shelf background radium activity (Rabkgd; taken as the lowest
measured concentration for each time-series). This correction also accounts for the
contribution from bottom and suspended sediments as described above.
(2) Multiply by the measured water depth (d) to convert from excess activities to excess
radium inventories per unit area of seabed. We assume that the shallow (< 3 m) water
column is well mixed.
(3) Divide the excess radium inventory for each time step by the estimated residence time
(τ) to convert to radium fluxes.
(4) Finally, divide this radium flux by the groundwater end-member radium activity
(Ragw) to convert to a water flux.
This approach may be expressed by the following equation:
{[Ra (dpm / m ) − Ra (dpm / m )]⋅ d (m)}
SGD(m / day ) =
τ (days ) ∗ Ra (dpm / m )
3
total
3
bkgd
3
(4.2)
gw
Radon time-series model
The model used to quantify SGD using 222Rn is described in detail by Burnett and
Dulaiova (2003). Briefly, each excess radon inventory (activity minus the radon supported by
dissolved 226Ra) is adjusted for atmospheric losses (MacIntyre et al., 1995). The changes in
inventories between time steps are then taken to be the net radon flux over each sample interval.
The maximum negative fluxes are assumed to be a conservative estimate of offshore mixing
losses. These mixing losses are then added back into the radon flux, which is divided by the
groundwater radon activity to achieve an SGD flux for each time step. This model has proven
effective in many different environments, and therefore serves as a useful comparison (e.g.,
Burnett et al., 2006; Burnett et al., 2007).
39
Results
Offshore transport
Horizontal transects of our tracers reveal some interesting patterns. Figure 4.2 shows the
bathymetry (a) as well as the distribution of salinity (b), pH (c), radium isotope activities (d, e, f),
and 222Rn activity (g) along our offshore transect in July 2007. These plots suggest that near
shore, a source of water contributes relatively fresh water, with lower pH and generally higher
radium content. According to the sampled groundwater parameters shown in Table 4.1, most
groundwaters have lower salinity and pH than the coastal waters. Also, water table contours on
land (Chen et al., 2007; Taniguchi et al., 2008) indicate flow in this direction. Therefore,
groundwater seepage is likely in this coastal region. The short-lived radium isotopes also reveal
possible groundwater inputs around 7.5 km and between 10 and 12 km offshore.
Applying the model represented by equation (4.1) to the radium transect results in a
general increase in apparent radium ages in the offshore direction (Figure 4.3). For this
calculation, the initial AR was taken as the average AR of the saline groundwater wells sampled
in 2007 (N-6 through N-10; Table 4.1). Even if a different initial AR was chosen, the absolute
ages would change but the trend offshore would remain the same. We calculated radium ages
based on both the 224Ra/223Ra AR (a; initial AR = 44.3) and the 224Ra/226Ra ratio (b; initial AR =
6.23) in Figure 4.3. Fitting a linear regression through these data yields a time-integrated
average horizontal transport rate of dissolved substances in the surface water (Moore and Krest,
2004). The resulting transport rates of 4.7 and 3.3 cm/s based on the inverse of the slope of the
respective regression lines are not significantly different. Peterson et al. (2008b) determined that
the coastal mixing rates off the mouth of the Yellow River (1.4 – 1.6 cm/s) did not vary
significantly with river discharge within the range of discharges studied (81 – 568 m3/s), and
concluded that mixing was tidally driven. Because there are no other significant advective
mixing forces in this study area, we find that tidal forces must have an even greater effect at
transporting dissolved substances offshore in the shallow tidal flats where this study was
conducted.
Radium time-series
Radium samples were collected hourly for 25 hours in September 2006 (TS-1) at a fixed
location 1 km from shore, and for 23 hours in July 2007 (TS-2) at another station along the same
line but located 2 km from shore. These sample sets are useful in that they are not simply a
snap-shot image of tracer concentrations at this site, but allow us to examine how the tracers
behave over a complete tidal cycle. One might expect, for example, to find the highest SGD
rates during low tide, as has been found in several other settings (e.g., Burnett et al., 2007;
Swarzenski et al., 2007b). The low tides here during TS-1 correspond with increases in salinity
and pH (Figures 4.4a and 4.4b), so the composition of the enhanced SGD during low tide, as
indicated by the peaks in radium (Figures 4.4c, d, and e), appears to be influenced by recirculated
seawater and saline groundwater. The peak salinity and pH values during the low tides are very
similar to those sampled offshore (>5 km, Figure 4.2).
40
Figure 4.2. Profile of seafloor bottom (a) and distribution of salinity (b), pH (c), 224Ra (d), 223Ra
(e), 226Ra (f), and 222Rn (g) along offshore transect from the study site. Error bars reflect 1-σ
measurement uncertainties.
41
42
#
Location
N37° 33.268' E118° 43.637'
N37° 35.357' E118° 32.535'
N37° 43.737' E118° 45.862'
N37° 43.737' E118° 45.862'
N37° 42.344' E118° 53.224'
N37° 35.698' E118° 43.687'
N37° 39.531' E118° 55.753'
N37° 40.591' E118° 53.964'
N37° 42.344' E118° 53.224'
N37° 43.737' E118° 45.862'
denotes pore water sample used to determine uncertainty range in SGD
7/20/2007
7/20/2007
7/20/2007
7/22/2007
Date Sampled
9/18/2004
9/18/2004
9/18/2004
5/8/2005
5/8/2005
9/22/2006
7/19/2007
7/19/2007
7/19/2007
7/19/2007
Screened
Depth (m)
3.0
12.0
4-10; 14-19
4-10; 14-19
1-6; 8-19
20
2-6; 12-20
1.5-7; 8-12; 16-20
1-6; 8-19
4-10; 14-19
N37° 34.054' E118° 59.264'
0.5
N37° 34.054' E118° 59.264'
0.5
N37° 35.605' E119° 02.370'
0.5
N37° 36.432' E119° 00.866'
0.5
N37° 36.432' E119° 00.866'
0.5
PW-4b
7/22/2007
N37° 36.234' E119° 01.197'
0.5
PW-5
7/22/2007
* denotes groundwater samples used to determine ARi and end-member values
PW-2
PW-3
PW-4a
#
PW-1
Groundwater
Sample
DO-34
DO-33
N-10
N-10
N-8
N-12
N-6*
N-7*
N-8*
N-10*
7.1
7.15
-
-
pH
6.34
6.15
7.44
7.65
6.96
7.18
7.19
6.96
7.18
27.1
28.3
Salinity
3.0
0.7
3.2
12.4
16.3
55.0
21.0
9.1
15.8
17.9
-
15,000 ± 660
4750 ± 270
Ra
3
(dpm/m )
1600 ± 120
441 ± 45
1620 ± 98
3170 ± 190
4100 ± 300
2950 ± 210
1900 ± 120
2100 ± 120
7180 ± 410
3060 ± 150
224
-
311 ± 52
217 ± 46
Ra
3
(dpm/m )
69 ± 25
16 ± 10
49 ± 18
39 ± 13
88 ± 20
102 ± 16
72 ± 16
31 ± 11
158 ± 21
60 ± 12
223
-
780 ± 110
560 ± 120
Ra
3
(dpm/m )
710 ± 110
32 ± 82
440 ± 120
726 ± 54
685 ± 30
590 ± 21
189 ± 18
805 ± 80
344 ± 24
950 ± 57
226
127,000 ± 36,000
137,000 ± 48,000
226,000 ± 58,000
193,000 ± 31,000
192,000 ± 29,000
269,000 ± 30,000
Rn
3
(dpm/m )
244,000 ± 23,000
390,000 ± 60,000
351,000 ± 21,000
361,000 ± 22,000
222
Table 4.1. All groundwater samples collected for radium isotopes. Samples included in upper portion of table were collected from
boreholes throughout the Yellow River delta. Samples named “PW” are collected pore water samples from the study site. All
radionuclide activities have been decay-corrected to the time of sampling. Uncertainties shown are at the 1-σ level.
Figure 4.3. Distribution of apparent radium ages offshore based on the 224Ra/223Ra AR (a) and
the 224Ra/226Ra AR (b). Dashed lines represent the 95% confidence level of the linear regression.
Transport rates are converted into units of cm/sec from the reciprocal of the slope of the linear
regression.
The radium trends indicate input during the low/ebb tides. Sharp peaks occur in the
short-lived radium isotope record at the low tides, while the 226Ra tends to peak on the falling
tide. The model is based on calculating radium fluxes, so the greatest change in inventories
actually occurs during the down-going tide. Because the tidal range in this area is nearly 2 m,
this shallow environment is greatly influenced by tidal forces.
The tracer record from TS-2 (Figure 4.5) shows little temporal variability in the salinity
record, but indicates decreases in pH corresponding with the low tides. These behaviors suggest
a more significant fraction of terrestrial water in the discharging groundwater than that observed
during the 2006 time-series when the salinity and pH increased at low tide. This terrestrial
component is not necessarily fresher than the overlying water. Several groundwaters sampled
(Table 4.1) show saline to hypersaline conditions while still maintaining relatively low pH.
The short-lived radium trends seen in these data are similar to those from TS-1. All three
radium isotopes show sharp increases at or shortly before low tide, i.e., generally on the falling
tides. Although 226Ra tended to precede the short-lived radium peaks in TS-1, it behaved much
more like 223Ra and 224Ra during TS-2 by maintaining a baseline activity throughout the high
43
Figure 4.4. Temporal variability of the water level (right axis) as well as salinity (a), pH (b),
224
Ra (c), 223Ra (d), 226Ra (e), and 222Rn (f) during TS-1 in September 2006. Error bars represent
1-σ measurement uncertainties.
44
tides. One interesting observation from these data, however, is that the radium peaks occur
during the falling tide, whereas the peaks in pH appear a few hours earlier in TS-2.
We used the radium isotopic data with the model described above (equation 4.2) to
quantify the groundwater discharge rates. The calculations are based on a unit area of sea
bottom, and give results in terms of specific fluxes (m3/m2 day) or vertical velocities of the
advecting water (m/day). Table 4.2 summarizes important parameters used in the calculations.
The following section details each of the calculations steps, as well as the choice of the
parameters reported in Table 4.2.
Adjust measured radium activities for background radium. The time-series radium
trends indicate that each isotope achieves a baseline or “background” activity throughout the
tidal cycle, but always remains present above detection. We elect to use the lowest activity
measured for each radium isotope during the course of each time-series sampling to represent
this “background.” In other words, we interpret these activities as those that would be present
with no SGD signal, i.e., 100% of the radium is derived from other sources such as desorption
from particles and shelf waters mixing into the sampling area. Using the lowest measured
activity thus represents a conservative estimate of the background activity, because these
samples could still contain some groundwater-derived radium. The selected shelf background
activities for TS-2 are uniformly lower than TS-1 (Table 4.2). This difference is likely a
reflection of its location being further offshore. These background values are subtracted from all
other measurements within the time-series to correct each sample for the shelf contribution to
yield an excess radium activity.
Diffusive inputs from bottom sediments and desorption of radium from suspended
sediments are already considered when subtracting out the background activity as described
above. These input rates are assumed to be uniform over time and space. There is no direct
source of suspended sediments to our study site, so we assume that the suspended sediments in
this area are derived from resuspended bottom sediments and no further sedimentary radium
input corrections are necessary.
Convert from excess radium activity to excess radium inventory. In order to remove
the effect of different water depths associated with these samples, we multiply each excess
radium activity concentration by the water depth at the time of sampling to convert to an excess
radium inventory. This correction allows for a direct comparison of the absolute radium
activities per unit area of seabed.
Divide the excess radium inventories by the water residence time. In order to convert
excess inventories to a flux rate, we must divide by the effective residence time of the coastal
waters. We elect to use an average apparent radium age of the waters sampled along the offshore
transect based on the 224Ra/226Ra AR (Figure 4.3). The measurement uncertainty associated with
226
Ra (~ 10%) is much better than that of 223Ra (~ 25%), so this provides the best residence time
estimate available (4.7 days).
While we think this is a reasonable estimate based on the data available, there are some
uncertainties. The model used to calculate the apparent radium ages from water samples (Eq.
4.1) assumes only one source of radium with a fixed isotopic composition. However, the
offshore radium trends reveal possible groundwater inputs around 7.5 km and between 10-12 km
offshore (Figure 4.2). If these are indeed related to groundwater inputs, our assumption would
45
Figure 4.5. Temporal variability of the water level (right axis) as well as salinity (a), pH (b),
Ra (c), 223Ra (d), 226Ra (e), and 222Rn (f) during TS-2 in July 2007. Error bars represent 1-σ
measurement uncertainties.
224
46
still hold as long as the radium signature is the same. Unfortunately, we do not have any means
to assess such variations at this time. In view of these difficulties, we later address the model
effects of different effective residence times in a section on uncertainties.
Table 4.2. Radium time-series model parameter summary. The values found in the central
column between TS-1 and TS-2 are common to both time-series.
TS-1
TS-2
Shelf Background Activity (dpm/m3)
224
Ra
Ra
226
Ra
625
24.7
143
223
77.8
1.5
28.8
Groundwater End-Member Radium Activity (dpm/m3)
224
Ra
Ra
226
Ra
3560
80
570
223
Pore water End-Member Radium Activity (dpm/m3)
224
Ra
Ra
226
Ra
15,000
311
780
223
Divide by the end-member value to convert into a water flux. Table 4.1 presents a
wide range in possible groundwater end-members found throughout the delta that could
constitute the advecting groundwater fluids. Only small volumes (~ 10 L) of sediment pore
water could be collected for radium analysis, so we feel that the large range found in pore water
radium results are at least partially due to analytical difficulties associated with insufficient
sample volumes. We have elected to use the average of the saline wells measured in July 2007
(N-6 through N-10). These wells are all directly up-gradient of the study site and were collected
immediately after one of our time-series (TS-2), so likely represent the terrestrial source water
component of the SGD. No detectable freshening of the surface waters during the apparent
discharge intervals of the time-series occurred, so the use of fresh groundwater end-members
(e.g., wells sampled in 2004, Table 4.1) is not appropriate. The initial groundwater ARs used to
find the radium ages from the offshore transect are also based on the average of these selected
groundwater samples.
Dividing the radium fluxes by the groundwater end-member activities yields estimates of
the SGD flux required to support the measured excess inventories. Figure 4.6a and 4.6b show
the model results for TS-1 and TS-2, respectively. The three isotopes show similar patterns,
though their range in SGD rates varies significantly. All isotopes indicate high discharge during
the falling tide for both TS-1 and TS-2. Table 4.3 summarizes the results from each time-series
47
model. During TS-1, SGD rates averaged 4.5, 9.4, and 13.9 cm/day based on 224Ra, 223Ra, and
226
Ra, respectively. TS-2 SGD rates were similar, averaging 5.2, 11.8, and 9.6 cm/day based on
the same isotopes.
Radon time-series
Figures 4.4f and 4.5f show the results of the 222Rn time-series analysis from TS-1 and
TS-2, respectively. Measurements were integrated over 30 minutes for TS-1, and over 1 hour for
TS-2 in order to achieve better analytical uncertainties. In both cases, no clear temporal trends in
the radon activities were observed over the tidal cycle, except a slight increase in radon activity
during the low tides of TS-1. The water depth changes throughout this tidal cycle, however, so
significant changes in the radon inventories do exist. The model presented by Burnett and
Dulaiova (2003) requires knowledge of the offshore radon activity to account for radon
introduced during the flood tide. We use the minimum value sampled along the offshore transect
(980 dpm/m3 at 7.5 and 8.7 km offshore) as the offshore radon end-member. Values in the range
of ~ 1000 dpm/m3 are found at several locations along the transect, so we feel this is a reasonable
choice.
We also must apply a groundwater end-member to this model. Because the radon in
groundwater analysis requires much less water than that for radium (250 mL versus 10+ L), we
were able to collect water samples for radon from additional pore water locations than those
sampled for radium (Table 4.1). The average (and standard deviation) radon activity from 5 pore
waters sampled was 191,000 ± 54,000 dpm/m3, so we take this to represent the groundwater endmember. This value is somewhat lower than the average from the monitoring wells on land (av.
= 337,000 ± 64,000 dpm/m3). Sampling the pore waters in this case is a more direct assessment
of the composition of the actual advecting SGD water, so should provide a better constrained
end-member value than the terrestrial boreholes selected for the radium end-members.
Wind speeds (and thus radon loss via atmospheric evasion) observed during TS-1 (av. =
3.4 m/s) were somewhat higher than those during TS-2 (av. = 2.9 m/s), so we applied a shorter
integrated mixing time for TS-1 (1 hour versus 3 hours for TS-2) as required by the model. This
mixing time determines the number of hours that the maximum negative fluxes are extrapolated
to represent mixing losses. A shorter mixing time minimizes the risk of overestimating mixing
losses for other samples on the basis of a large negative flux from one sample.
48
Figure 4.6. Radium time-series results for SGD fluxes during TS-1 (a) and TS-2 (b) as well as
the corresponding water level recorded at the study site. Results from 224Ra are shown by the
black circles, those from 223Ra by gray squares, and those from 226Ra analysis are shown by
white triangles. The results shown represent a 3-point smoothing.
Offshore radium distribution assessment of SGD
From the offshore transect data presented in Figure 4.2, we can use the distribution of
Ra to independently calculate shelf-wide SGD rates as per Moore (1996). He found
enrichments of 226Ra in shelf waters along the southeastern United States and applied a version
of the following equation to determine regional SGD rates:
226
SGD (m / day ) =
Excess 226Ra (dpm / m 3 ) ⋅ d (m)
τ (days) ⋅ 226 Ra gw (dpm / m 3 )
(4.3)
In this equation, the excess 226Ra represents the average activity of 226Ra that is enriched above
the open ocean background level. Using the lowest activity measured along the transect (115
dpm/m3) as the oceanic background level, we find the average excess activity along the transect
49
to be 322 dpm/m3. The excess 226Ra activity used is therefore 207 dpm/m3. The average depth
(d) along the transect was 3.0 m. We take the average 224Ra/226Ra age of the samples (4.7 days;
Figure 4.3) as the residence time of the shelf waters (τ) and the groundwater end-member value
of 570 dpm/m3 (Table 4.2) as the 226Ragw term. After applying equation (4.3) to our results, we
find the regional vertical velocity of SGD to be 24.2 cm/day. This SGD flux (Table 4.3) is
similar but somewhat higher than those determined by the radium model.
Figures 4.7a and 4.7b show the model results of the SGD rates based on 222Rn from TS-1
and TS-2, respectively. The results from both TS-1 and TS-2 show enhanced SGD during the
falling tides as did the radium isotope analyses. The average SGD rates for the two time-series
are 13.8 cm/day (TS-1), and 8.4 cm/day (TS-2), which are within the range of the radium results.
Discussion
Radium offshore distribution
The short-lived radium isotope distribution in Figure 4.2 shows several peaks farther
offshore, indicating either possible groundwater inputs around 7.5 km and 11 km offshore or
pulses of tidal currents that have carried coastal waters offshore. The relative amplitudes of
Table 4.3. Summary of the results of the SGD analyses based on radium isotopes, radon, and
automatic seepage meters deployed in the same area. The rates reflect the average SGD rate
(cm/day) throughout the time-series, with the corresponding standard deviation for each set.
Values in parentheses are based on an end-member equivalent to the highest pore water radium
measured. Seepage meter results from Taniguchi et al. (2008).
TS-1
SGD Rate
St. Dev.
(cm/day)
224
Ra
Ra
226
Ra
226
Ra Offshore Distribution
222
Rn
223
Seepage Meters (2006 results):
1000 / 1500 m from shore
2000 m from shore
3000 m from shore
4000 m from shore
5000 m from shore
6000 m from shore
7000 m from shore
4.5 (1.1)
3.4 (0.8)
9.4 (2.4)
6.2 (1.6)
13.9 (10.2) 11.1 (8.1)
13.8
17.9
42.2
48
130
n/A
16
33
15
27.9
44
117
n/A
14
33
14
50
TS-2
SGD Rate
St. Dev.
(cm/day)
5.2 (1.2)
11.8 (3.1)
9.6 (7.1)
24.2
8.4
3.0 (0.7)
8.5 (2.2)
6.2 (4.5)
11.8
35.6
n/A
n/A
67.2
15.9
n/A
n/A
19.9
n/A
n/A
43.9
12.6
n/A
n/A
these peaks becomes more pronounced for the longer-lived isotopes (i.e., 224Ra < 223Ra < 226Ra),
as is expected for aging water masses, but we lack sufficient data to determine whether these
peaks are due to local groundwater inputs or are the result of coastal waters mixing offshore.
Further investigations at this and other study sites should look into this possibility further, as
done by Hancock et al. (2006).
Figure 4.7. Results from the 222Rn SGD model for TS-1 (a) and TS-2 (b) as well as the
corresponding water level recorded at the study site. The results shown represent a 5-point
smoothing. Error bars shown are propagated errors throughout the model calculations.
Radon results
A possible source of error in the radon model used to find SGD rates is that significant
inputs of radon could be from diffusion into the water column from the sediments. We have
estimated this input by using an empirical equation reported by Burnett et al. (2003b) to relate
the 226Ra content in sediments (226Rased; units = dpm/g) to the diffusive flux of 222Rn:
51
222
Rn Flux (dpm m −2 day −1 ) = 495 ⋅ 226 Ra sed + 18.2
(4.4)
We have measured several sediment samples via gamma spectrometry, and found the
average 226Ra activity to be 1.62 dpm/g (Table 4.4). Applying equation (4.4), the corresponding
222
Rn flux is thus 820 dpm/m2 day, or 34.2 dpm/m2 hr. Assuming this system is in steady-state,
we can find the 222Rn concentration supported by diffusion in the water column by subtracting
each calculated hourly atmospheric flux from the diffusive flux, then dividing by the decay
constant (0.0076 hr-1) and by the average water depth for each measurement interval (Dulaiova et
al., 2006). This calculation results in nearly every measurement interval showing a negative
value for the supported 222Rn activity. Therefore, we conclude that on average, the atmospheric
evasion losses are greater than the diffusive flux of radon from the bottom sediments, and as
such, we can neglect this possible source of 222Rn.
Mulligan and Charette (2006) have pointed out that using 222Rn as a tracer to model SGD
yields total discharge rates, because radon would be present in both the terrestrial component as
well as the recirculated seawater component of the discharging groundwater. Using radium as
the tracer, however, would only reveal the recirculated seawater (saline) flow, as radium remains
particle-bound in freshwater. This theory agrees with our results and because the radium fluxes
are as high or higher than those determined from radon, the vast majority of SGD at our study
sites is composed of recirculated seawater. The exception to this is the 224Ra results during TS-1,
when the radium result was roughly one-third of that determined by the radon.
Model uncertainties
The radium model results show similar patterns among the different isotopes, but the
absolute fluxes determined by 224Ra are consistently lower than those found by 223Ra and 226Ra.
This difference is likely a result of overestimating the shelf background 224Ra activity by using
the lowest sampled activity during the time-series. If these lowest samples still contained a 224Ra
component from SGD, then we overcompensated and caused the ultimate SGD fluxes to be too
low.
This discrepancy may also be a result of our limited selection of groundwater endmember values. While we are most confident that the saline groundwaters sampled in July 2007
represent the most likely end-member values, the sampled pore waters were consistently higher
in measured radium activity. In order to assess the model uncertainty due to end-member
selection, we applied the model using the highest pore water sample (PW-2; see Tables 4.2, 4.3).
The model results based on this end-member are shown in Table 4.3. Using this extremely high
pore water value as the end-member, the average SGD fluxes decrease by 75% for 224Ra and
223
Ra, and by 25% for 226Ra. Another source of uncertainty in the model involves diffusive
inputs of radium to the water column as inter-tidal sediments are flooded during each high tide.
We do not feel that this process represents an important source of radium as the tidal variations
in 226Ra (too long-lived to regenerate on tidal time scales) are about the same as the observed
224
Ra variations.
52
Table 4.4. Summary of 226Ra content on collected sediments, measured via gamma
spectrometry. All samples except TS-1 were collected in the Yellow River, above the maximum
salinity front. Uncertainties shown represent 1-σ measurement uncertainties.
226
Ra Activity
(dpm/g)
Sediment
Sample Name
Collection
Date
Latitude
Longitude
Salinity
Yellow River Suspended
9/15/2004
N37° 45.666'
E119° 09.327'
0
2.23 ± 0.16
Yellow River Bottom I
5/4/2005
N37° 45.666'
E119° 09.327'
0
1.36 ± 0.08
Yellow River Bottom II
5/4/2005
N37° 45.666'
E119° 09.327'
0
1.39 ± 0.19
Yellow River Bottom III
9/19/2006
N37° 45.666'
E119° 09.327'
0
1.49 ± 0.11
TS-1
9/22/2006
N37° 36.525'
E119° 00.518'
28.6
1.97 ± 0.12
Yellow River Bottom IV
7/19/2007
N37° 45.666'
E119° 09.327'
0
1.26 ± 0.15
Average:
1.62 ± 0.14
One other source of uncertainty in the radium model concerns the residence time of these
coastal waters. We used a value of 4.7 days, based on the average 224Ra/226Ra age from Figure
4.3, but as detailed above, several sources of uncertainty exist in this calculation. If a longer
water residence time were used, for example the average 224Ra/223Ra age from Figure 4.3 (5.3
days; an increase of 13%), the modeled SGD rates would decrease by 11%. Using a shorter
residence time (4.1 days; a decrease of 13%), the corresponding modeled SGD rates would
increase by 15%.
Burnett et al. (2007b) summarize the important uncertainties behind the radon model
used here. As with the radium model, the most significant source of uncertainty lies with
assigning an end-member value to the discharging fluids. In addition, assessing the mixing
losses of radon to both the offshore waters and the atmosphere represent other, yet often less
important, sources of uncertainty.
Comparison to seepage meters
Table 4.3 indicates that our different geochemical tracer analysis techniques for
quantifying SGD rates are in reasonable agreement. We also have some estimates based on a
completely independent approach. Figures 4.8a and 4.8b contain data presented by Taniguchi et
al. (2008) for automatic seepage meter fluxes collected nearby and at the same time as our TS-1
and TS-2 time-series experiments, respectively. The seepage meter location during TS-2 was
500 m inland of the tracer sampling location, whereas that corresponding with TS-1 was ~50 m
from our sampling station. Also, a Darcy’s Law hydrological calculation of the terrestrial
groundwater flow to the ocean in this area using FEFLOW shows results ranging from 0.3 to 5.8
cm/day [J.Z. Cheng, pers. comm.].
53
Figure 4.8. Independent analysis of SGD by means of automated seepage meters, as summarized
by Taniguchi et al. (2008) for the time during TS-1 (a) and TS-2 (b) chemical tracer sampling.
Error bars represent standard deviation of all measurements recorded during each measurement
interval (30 minutes).
Our interpretive SGD trends based on the radium results agree very well with the seepage
meter patterns for both periods. The patterns based on the radon records also resemble those of
the seepage meters, but with less smooth trends. A disadvantage of seepage meters is that they
can only integrate over the small area of sea bottom they cover (~ 0.25 m2), but the chemical
tracers integrate over an unknown yet much greater spatial range (Burnett et al., 2006a).
Nonetheless, the general agreement in patterns of SGD between the tracer methods and the
seepage meters provides confidence in both approaches.
During the period of chemical tracer sampling of TS-1, the seepage meter average SGD
rate was high at 42.2 ± 27.9 cm/day (Table 4.3), of which ~ 3% was fresh groundwater discharge
according to salinity measurements made inside the seepage meter chambers. Averaging the
seepage meter fluxes over several days prior to the tracer sampling provides an average SGD
flux of 20.7 ± 20.5 cm/day, closer to our tracer-derived values. Seepage meter averages from
other nearby locations during TS-1 ranged from 15 to 130 cm/day (Taniguchi et al., 2008).
54
The average seepage meter flux during our sampling of TS-2 was also high at 35.6 ± 19.9
cm/day, of which about 21% was fresh groundwater. The several day integrated average SGD
based on this seepage meter was similar at 41.9 ± 20.4 cm/day. The fraction of fresh
groundwater in the SGD measured from the seepage meters indicates a more pronounced
recirculated seawater signal during TS-1 than during TS-2. This was also supported by our timeseries pH and salinity measurements.
The seepage meter flux results are generally higher than the results found from our
independent radium isotope and radon analyses. One possible reason for this difference is that
the SGD is dominated by recirculated seawater, with a short residence time in the subsurface.
From data presented by Taniguchi et al. (2008) based on all their seepage meter results in the
area, the freshwater component of SGD is never greater than 27%, and all but one sample ranged
between 0.5 and 7.5%. If most of this water is simply recirculating through the sediments over a
tidal cycle, it would not have sufficient time to fully equilibrate with the tracer concentrations in
the aquifer. Therefore, the groundwater end-members that were used in the models above could
be too high compared to their actual values, artificially lowering our modeled SGD rates. In
addition, the 222Rn pore water value is better constrained, lower than the borehole samples
collected on land, yet produced SGD rates that were in the same range as the radium-based
values.
Regional-scale fluxes
Taniguchi et al. (2008) defined the offshore seepage face as being equal to a width of
about 7 km, and used the length of the Yellow River delta coastline (350 km) to scale up their
seepage results for comparison to the Yellow River discharge. Using these same values, and
conservatively assuming that our lowest estimates of SGD (4.5 cm/day from the 224Ra model
during TS-1; 5.2 cm/day from the 224Ra model during TS-2) are uniform over this area, we find a
total flow of approximately 1280 m3/s during TS-1 and 1480 m3/s during TS-2 around the
Yellow River delta, most of which is likely recirculated seawater (Taniguchi et al., 2008). For
comparison, the Yellow River discharge during this time of year is around 600 m3/s. As these
SGD fluxes are conservative values, the actual water exchange from SGD compared to the
Yellow River could be much higher.
As part of this study, nitrate was measured in the groundwater wells and the Yellow
River to assess relative fluxes to the Bohai Sea. The average NO3- concentration in the
groundwater wells that we sampled for radium (N-6 through N-10) was 440 μM and that in the
Yellow River was 430 μM (T.Z. Mi, pers. comm.). Multiplying these concentrations by the
SGD fluxes yields nitrate fluxes 2 to 3 times higher than that delivered by the Yellow River.
Care must be taken in interpreting these results, however, because geochemical reactions often
alter the nutrient character of groundwater within the subterranean estuary (Santos et al., 2008;
Spiteri et al., 2008).
Nonetheless, Chen et al. (2007) have found groundwaters throughout the delta that are up
to an order of magnitude enriched in nitrate concentration relative to that of the Yellow River, so
clear potential exists for SGD to supply excess nitrate to the Bohai Sea. Previous studies have
found SGD to be an important nutrient source, even in river-dominated regions, such as the Gulf
of Thailand (Dulaiova et al., 2006). Application of the offshore transport rates to the flux of
nutrients from a coastal source (e.g., SGD) can help assess whether the rates are sufficient to
transport the nutrients offshore before uptake by primary productivity (Peterson et al., 2008b).
55
Increasing levels of dissolved inorganic nitrogen have been documented as occurring in the
central Bohai Sea over the past few decades (Zhang et al., 2004), so these results can help assess
the source of the excess nitrogen.
Conclusions
We have measured several groundwater tracers in an area ~ 40 km south of the Yellow
River estuary to quantify SGD rates from offshore transects and time-series analyses. Salinity
and pH are useful tracers in this environment, showing increasing values in the offshore
direction. The gradient in apparent radium ages of the water masses with distance from shore
yields horizontal transport rates between 3.3 and 4.7 cm/s.
We show that using radium isotopes to assess SGD rates via a stationary time-series
fashion is a valuable approach. The results from the radium time-series were similar to those
using an established 222Rn model, and followed the patterns of SGD from seepage meter
measurements. During September 2006, the average SGD rates ranged from 4.5 to 13.9 cm/day,
and the discharging water was composed primarily of recirculated seawater. The SGD rates
found during July 2007 averaged between 5.2 and 11.8 cm/day, and apparently have a larger
fraction of terrestrial water. These fluxes and patterns are somewhat lower than those from
individual seepage meters deployed nearby but are similar to average rates reported from
seepage meters positioned in the same general area.
There are some uncertainties associated with the radium time series model that we use to
estimate the SGD fluxes. The most prominent of these uncertainties lies in the assignment of the
appropriate groundwater end-member. If we use measured pore water radium activities as the
end-member instead of the saline groundwaters, our SGD fluxes decrease by 75% for 224Ra and
223
Ra, and by 25% for 226Ra. Other, more minor sources of uncertainty for the model involve the
residence time of the coastal waters and diffusive inputs from tidal inundation of inter-tidal
sediments.
Scaling our SGD fluxes determined from the radium isotopes to the whole Yellow River
delta, we find estimated SGD and nitrate fluxes 2-3 times that of the Yellow River. We suspect
that most of the regional SGD is composed of recirculated seawater. In fact, excessively high
nitrate levels in groundwaters have been measured around the delta, so NO3- fluxes to the Bohai
Sea from SGD are likely higher than those from the Yellow River, at least during periods of low
river discharge.
56
CHAPTER 5
BOHAI SEA COASTAL TRANSPORT RATES AND THEIR INFLUENCE ON
COASTLINE NUTRIENT INPUTS
Article published in From Headwaters to the Ocean, Taylor & Francis, London, 659-664.
Abstract
Recent studies have shown that the central Bohai Sea is becoming enriched in nitrate relative to
other nutrients. This trend is sufficiently dramatic to indicate that the system is approaching a
shift to a phosphate-limited ecosystem. In order to examine possible sources of this nitrate, we
use values of coastal transport rates assessed via natural radioisotopes to determine the extent of
transport of dissolved inorganic nitrogen (DIN) supplied by the Yellow River and coastal
groundwater discharge. We find that in both cases, transport rates are not fast enough to deliver
DIN to the open Bohai Sea before biological uptake. Maximum Yellow River DIN transport
distances are found to be 43 km offshore, whereas that for SGD-derived DIN is 53 km. Effects of
subsequent transformations to organic-N, regeneration to DIN, and losses via particulate-N
settling and denitrification are currently unknown.
Introduction
Extreme yearly fluctuations in discharge of the Yellow River combined with a growing
population in the river basin have altered the recent riverine nutrient flux to the coastal Bohai
Sea. Since 1960, central Bohai Sea nitrate concentrations have increased by a factor of 10, while
phosphate concentrations have decreased by a factor of 2 (Zhang et al., 2004). These authors
observed a shift to a phosphate-limited ecosystem in Laizhou Bay (a large embayment just south
of the Yellow River mouth) and proposed that the entire Bohai Sea could be approaching this
condition. The Yellow River, the largest river to empty into the Bohai Sea, is a possible source
of these excess nitrate concentrations. Submarine groundwater discharge (SGD) and atmospheric
sources (wet and dry deposition) are other possible sources (Raabe et al., 2004; Wei et al., 2004;
Liu and Yin, 2007).
Although recent increases in Yellow River nitrate concentration have been observed, the
annual river discharge has been steadily decreasing, and therefore an increased flux of nitrate to
the Bohai Sea from this river is not necessarily the case (Shen and Le, 1993; Huang et al., 2005).
Periods of zero nitrate flux occurred when the river did not meet the ocean, including 226 days in
1997 when the river experienced zero flow conditions in the delta area (Yu, 2006). Thus,
whether the Yellow River can be a source of the increasing nitrate concentrations in the central
Bohai Sea remains unanswered.
SGD is now regarded as an important transport mechanism that moves dissolved
substances from sub-seabed fluids to the coastal ocean. Being both spatially and temporally
variable, SGD is very difficult to measure and therefore assess its relative importance in coastal
57
ocean chemical budgets (Burnett et al., 2006b). Nonetheless, nutrient fluxes via SGD have been
shown to rival those from rivers in some locations (Slomp and Van Cappellen, 2004; Kim et al.,
2005; Swarzenski et al., 2007a).
We suspect that the Yellow River delta is a location where SGD can potentially introduce
nutrients in the same magnitude to the coastal ocean as the local rivers. More people are
inhabiting the delta than ever before and are contributing to groundwater contamination, mainly
through agricultural practices and sewage disposal. Chen et al. (2007) inspected the groundwater
environment around the delta and found elevated nitrate levels (up to 3.8 mM) occurring mainly
in shallow aquifers, often coinciding with land use patterns that utilize Yellow River irrigation
water and contain concentrated population centers. Therefore, SGD is considered a potential
source of the recent excess nitrogen flux to the Bohai Sea. Once nutrients are introduced to the
coastal zone via river flow or SGD, they may be transported offshore to the central Bohai Sea
and contribute to the increasing nitrogen concentrations found by Zhang et al. (2004).
The main purpose of this paper is to determine the maximum horizontal length scales
associated with the transport of dissolved inorganic nitrogen (DIN) supplied by the Yellow River
and surrounding SGD. As Zhang et al. (2004) measured increasing nitrate level in the middle of
the Bohai Sea, so if these nutrients can be transported roughly half the distance across the Bohai
Sea (~125 km) before being converted to organic-N via primary production, then these sources
of DIN can be considered significant contributors to the recent increasing nitrogen found in the
Bohai Sea. We use our radium isotopic results presented separately (Peterson et al., 2008a;
Peterson et al., 2008b) for transport rates and SGD fluxes based on radium isotope distribution to
determine these DIN horizontal transport scales.
Methods and Results
Transport of Yellow River water offshore
Peterson et al. (2008b) determined apparent radium ages of samples collected along
offshore transects from the mouth of the Yellow River in September 2004, May 2005, and
September 2006. The river discharges associated with these sampling periods were 392, 81, and
568 m3/s, respectively. Apparent radium ages were determined by examining the difference
between a measured radium isotope activity ratio of a short-lived to long-lived isotope (e.g.,
224
Ra/228Ra) along the transect (ARobs) and an initial radium activity ratio (ARi) measured in the
river estuary. In freshwater environments, radium is particle-reactive and tends to be found
adsorbed onto the surface of suspended sediments, but once reaching saline water, ion exchange
processes desorb the radium into solution. At this point of maximum desorption, the initial
(highest) AR is found. Apparent radium ages are thus calculated by:
⎛ AR
t = Ln⎜⎜ obs
⎝ ARi
⎞
1
⎟⎟ *
⎠ λ LL − λ SL
(5.1)
where λSL represents the decay constant of a short-lived isotope (224Ra; T1/2 = 3.66 days) and λLL
is that of a longer-lived radium isotope (228Ra; T1/2 = 5.7 years). Fitting a linear regression to the
58
apparent radium age distribution offshore (e.g., Fig. 5.1) leads to an assessment of the timeaveraged dissolved component transport flux. Using the 224Ra/228Ra AR, these offshore transport
rates (Tr) were found to be 1.60, 1.43, and 1.62 cm/s for the three sampling trips, respectively
(Peterson et al., 2008b).
Nitrogen uptake rates
Prior to assessing the transport of DIN delivered by the Yellow River or SGD, we first
determine local nitrogen uptake rates. Coincident with our September 2006 sampling, we
measured dissolved silica concentrations (Fig. 5.2A) along an offshore transect from the Yellow
River mouth. Since these silica concentrations decrease exponentially offshore, we can use this
distribution as a proxy of biological activity in the region. Simple diffusional mixing would
result in a straight line joining the river and Bohai Sea end-members within the mixing zone, so
any difference between a measured concentration and the anticipated silica concentration along
the conservative mixing line likely represents the effects of biological consumption or production
(Kaul and Froelich, 1984). After calculating this difference for each silica sample, we multiply
by the associated water depth at the collection site to convert to silica uptake as an inventory
through the water column. These inventories are then divided by the corresponding radium age
of the sample to determine the silica uptake rate (the average rate of all samples was 16.4 ± 32.2
mmol Si/m2.day). If we assume that the biota in this region are consuming nutrients in the same
ratio as delivered by the river, we can convert this silica uptake rate to a nitrogen uptake rate by
multiplying by the riverine N:Si ratio of 2.79 (using a river end-member DIN concentration of
348 μM (Tiezhu, unpubl.) and 124.8 μM Si from this study). This N:Si ratio is significantly
higher than those reported for open ocean diatoms (e.g. 0.89 derived from Brzezinski 1985), but
the diatoms in the estuary are likely utilizing the nutrients according to the ratio in which they
are delivered, rather than an open ocean ratio. This conversion results in a nitrogen uptake rate
of 45.8 mmol N/m2.day, which is comparable in magnitude to nitrogen uptake rates found in
other estuarine systems (20 – 100 mmol N/m2.day, calculated from Price et al., 1985; Pennock,
1987; Bode and Dortch, 1996).
Since our value represents estuarine uptake rates and not necessarily coastal marine rates,
we also use a more conservative value for comparison. The net primary production rate for the
Bohai Sea is 402 mg C/m2.day (Price et al., 1985). Converting this value to molar units and
dividing by the Redfield C:N ratio of 106:16, we find a nitrogen uptake rate of 5.1 mmol
N/m2.day, a value one order of magnitude lower than our estuarine nitrogen uptake rate.
The coastal area into which the Yellow River discharges is constantly fed by nutrients, so
it must have reached a steady-state concentration where the river flux is essentially balanced by
mixing and biological removal. To find this steady-state concentration, we derive a first order
removal constant (λ), similar to a radioactive decay constant, by plotting the natural logarithm of
silica concentrations versus their respective apparent radium ages. Using the data points that
were collected within the same day and within the linear mixing zone (Fig. 5.2B), we find the
silica removal constant (the slope of the regression line) to be 0.1735 day-1. We convert this
value into a first order DIN removal constant by multiplying by the riverine N:Si ratio (0.484
day-1). The daily riverine DIN flux is divided by the effective mixing area from day I, then
divided by this first order removal constant to define the steady-state DIN inventory throughout
this mixing area for each sampling period.
59
Figure 5.1. Distribution of 224Ra/228Ra ages offshore from the mouth of the Yellow River found
in September 2006. The reported transport rate is derived from the inverse of the slope of the
best-fit line. The thin lines represent the 95% confidence interval of the regression.
Riverine nutrient transport
We view the geometry of the river plume as a semi-circular area where the transport
distances, Tr, represent the circular radii. Considering an 80° mixing angle (Peterson et al.,
2008b), we can find the daily mixing area for dissolved substances from the river mouth. Figure
5.3A shows a schematic of the mixing area for the river plume. To calculate progressive mixing
areas, for example Day II, we find its area as A = π(2*Tr)2*(80/360), then subtract out the area
for Day I, since this area will contain the nitrogen input from the following day’s river discharge.
We calculate a daily DIN flux from the Yellow River by multiplying the riverine DIN
concentration (348 μM in September, 432 μM in May) by the corresponding discharge (Table
5.1). The steady-state nitrogen inventory is then divided by the mixing area for Day II as
discussed above to yield the nutrient inventory over the effective mixing area. The nitrogen
uptake rate is then subtracted from this inventory to give a net daily nitrogen inventory delivered
by the Yellow River. The remaining nitrogen is then divided over the progressive mixing area for
the next day, less the daily nitrogen uptake rate. This process is continued until there is no net
inorganic nitrogen remaining. The number of days required to distribute and biologically
remove the daily DIN flux from the river for each sampling trip is summarized in Table 5.1, and
the corresponding distance based on the transport rate is also shown. We find that the
conservative estimate of nitrogen uptake rate (5.1 mmol N/m2.day) allows the DIN to be
transported farther into the Bohai Sea before biological uptake, but neither shows the nitrogen
being directly transported far enough into the central Bohai Sea to explain the high
concentrations observed by Zhang et al. (2004). This analysis only applies to the DIN directly
supplied by the river and does not consider subsequent transformations between organic-N and
DIN.
60
Figure 5.2. Distribution of silica (μM) along offshore transect measured in September 2006 (A).
The dashed line represents the conservative mixing line due simply to diffusional mixing. The
closed symbols are then used to calculate a first order removal constant (B) by plotting the
natural logarithm of the silica concentrations against their respective apparent radium ages.
Submarine groundwater discharge nutrient transport
Continuous time-series records of radon and radium isotopes were used by Peterson et al.
(2008a) to quantify SGD rates in the Yellow River delta ~60 km south of the estuary in
September 2006 and July 2007. In order to estimate the maximum possible transport distances in
this study, we base our SGD fluxes on the highest reported rates in Peterson et al. (2008a) : 19.1
and 17.8 cm/day (as in cm3/cm2.day) for September 2006 and July 2007, respectively. Taniguchi
et al. (2008) have estimated the seepage face to extend 7 km offshore in this area, so the daily
integrated seepage fluxes are then 1337 and 1246 m3/day per meter width of shoreline,
respectively.
Offshore transport rates were also determined (Peterson et al., 2008a) according to the
distribution of apparent radium ages as described above. In this area, the offshore dissolved
substance transport rate was found to be 4.7 cm/s from the 224Ra/223Ra activity ratio. Since the
SGD inputs in this region are assumed to be diffuse, and thus spatially homogenous (as opposed
to the point-source Yellow River mouth), the mixing areas are based on unit length of shoreline
61
(1 m). Figure 5.3B shows a schematic of the progressive mixing areas for this area. To calculate
these areas, for example Day II, we find its area as A = 2*Tr*1m then subtract out the mixing
area from Day I as described in section 2.3.
Figure 5.3. Schematic representing the progressing mixing areas from the point-source Yellow
River mouth (a) and the SGD area (b). Tr represents the reported transport rate (Peterson et al.,
2008a; Peterson et al., 2008b).
Table 5.1. Summary of transport distance calculation parameters for riverine DIN fluxes.
Results using our calculated nitrogen uptake rate, and the literature uptake rate (Raabe et al.,
2004) are reported.
River Discharge
m3/s
September 2004
May 2005
September 2006
392
81
568
Transport Rate
cm/s
1.60
1.43
1.62
River DIN
End-member (μM)
348
432
348
Uptake Rates
(mmol N/m2 day)
46
5.1
Uptake Rates
(mmol N/m2 day)
46
5.1
Uptake Time* (days)
13
27
9
19
15
31
Uptake Distance* (km)
18
37
11
24
21
43
* Uptake times and distances calculated for N uptake rates of 46 and 5.1 mmol N/m2 day (see text).
The average DIN end-member concentration measured from inland, saline wells
(Taniguchi, unpubl.) is 746 μM. We would prefer to use pore water measurements for the endmember, but the samples that were collected are in doubt due to sampling inconsistencies, so we
must therefore use the inland saline wells for the end-member. Multiplying this value (746 μM)
by the SGD rates yield daily DIN fluxes per meter width of shoreline. These fluxes are then
62
distributed over the progressive mixing area to determine the number of transport days required
for the biota to take up the daily DIN flux from groundwater. Results using both nitrogen uptake
rates are reported in Table 5.2. As with the riverine transport results, these dissolved nutrients
cannot be directly transported far enough offshore to contribute to the increasing nitrate levels
found in the central Bohai Sea. As with the river conclusions, however, we cannot assess the
impacts of subsequent transformations between the organic-N forms and DIN.
Table 5.2. Summary of transport distance calculation parameters for groundwater-derived DIN
fluxes. Results using our calculated nitrogen uptake rate (46 mmol N/m2.day), and the literature
uptake rate (5.1 mmol N/m2.day) are reported.
SGD Rate
m3/day
September 2006
July 2007
1337
1246
Transport Rate
cm/s
4.7
4.7
SGD DIN
End-member (μM)
746
746
Uptake Rates
(mmol N/m2 day)
46
5.1
Uptake Rates
(mmol N/m2 day)
46
5.1
Uptake Time* (days)
4
13
4
13
Uptake Distance* (km)
16
53
16
53
* Uptake times and distances calculated for N uptake rates of 46 and 5.1 mmol N/m2 day (see text).
Discussion and Conclusions
There are many assumptions and limitations behind these calculations. First, our
calculated nitrogen uptake rate based on the silica distribution is associated with a large degree
of uncertainty, but even with the conservative literature estimate of the average nitrogen uptake
rate for the entire Bohai Sea, the results show that the DIN initially delivered by the Yellow
River and SGD should be depleted before reaching the open Bohai Sea. We are also assuming
that the transport rates applied are uniform and consistent as the nutrients mix further offshore.
Most importantly, we have not considered transformations of DIN to organic-N that could be
transported beyond the limits considered here. The likely scenario in this region is that after
primary producers utilize DIN, they are transported towards other sections of the Bohai Sea
where nutrients can be regenerated. Particulate settling and denitrification would represent N
losses along the path. Regenerated nutrients could well be contributing to the observed increase
in nitrate in the central Bohai Sea, but our data set does not allow us to examine the role of this
process. Nonetheless, based on reported transport rates of riverine and SGD-derived dissolved
nitrogen species, even the most liberal calculations do not suggest that these dissolved nutrients
can be directly transported more than 53 km offshore before biological uptake. While these first
order calculations are admittedly crude, they provide a foundation for future studies to better
examine this process.
63
CHAPTER 6
QUANTIFICATION OF POINT-SOURCE GROUNDWATER DISCHARGES FROM
THE WESTERN HAWAII SHORELINE
Article published in Limnology & Oceanography.
Abstract
Aerial thermal infrared imaging (TIR) reveals that submarine groundwater discharge (SGD)
along the western coast of the Big Island of Hawaii is often focused as point-source discharges
that create buoyant groundwater plumes that mix into the coastal ocean. We quantified the SGD
fluxes associated with several plumes using natural geochemical tracers. Offshore transects of
222
Rn and 224Ra show elevated activities and corresponding low salinities in the near shore waters
within the plumes, indicating that these naturally-occurring radionuclides can be useful tracers of
groundwater inputs in this area. Using a series of simultaneous mass balance equations for
water, salt, and radon we determined the groundwater fluxes of six plumes near Kona, Hawaii.
The average SGD fluxes ranged from 1100 m3 d-1 to 12,000 m3 d-1 of total (fresh + saline) SGD.
The fresh (meteoric) groundwater equivalents for the same flows, modeled by adjusting the
groundwater end-member to reflect fresh rather than brackish groundwater composition, ranged
from 630 m3 d-1 to 8600 m3 d-1. These fluxes are in general agreement with earlier results
obtained by hydrological water budgets and physical oceanographic analyses of fresh SGD rates
in this region.
Introduction
Submarine groundwater discharge (SGD) has become a widely recognized transport
mechanism between land and the coastal ocean. In the past, SGD was considered a hydrologic
curiosity (Kohout, 1966), but over the past few decades, investigations have shown that water
and especially chemical fluxes via SGD may rival those of riverine inputs in some places (Cable
et al., 1996; Church, 1996; Burnett et al., 2006b). Groundwater typically contains elevated
nutrient concentrations several times higher than those found in nearby rivers, so even though
SGD may be volumetrically much less significant than river discharge, the nutrient inputs from
SGD can be significant (Kim et al., 2005; Swarzenski et al., 2007a; Santos et al., 2008).
Continued residential and agricultural development in coastal areas is causing ever increasing N
and, to a lesser degree, P concentrations in the groundwater. The results of this increased
nutrient loading can be two-fold: first, supply of the limiting nutrient increases coastal primary
production; and second, in extreme cases, the limiting nutrient could switch from N to P (Slomp
and Van Cappellen, 2004).
SGD is not simply freshwater, but is a mixture of fresh groundwater that flows from the
land and seawater that is driven landward beneath it. Many factors can influence the rates and
patterns of fresh SGD as it occurs at the coast. High precipitation rates and aquifer permeability
64
lead to elevated groundwater recharge rates relative to surface stream recharge. Steep relief
between inland areas and the coast can generate a strong hydraulic gradient, which drives
groundwater seepage at the coast. Oceanic forcing within the subterranean estuary drives the
recirculated seawater component of SGD. For example, high tidal levels and wave activity can
drive seawater into the subsurface, with subsequent discharge (Moore, 1999). In all cases, the
permeability of the aquifer(s) exerts an important control on hydraulic conductivity. It is the
combined presence of several of these forcing factors that led Zektser (2000) to hypothesize that
tropical islands, such as Hawaii, would have disproportionately high SGD fluxes compared to
continental areas.
By nature, SGD is often spatially and temporally variable. Taniguchi et al. (2008), for
example, used automatic seepage meters to show that SGD rates differ drastically along a shorenormal transect in the Yellow River delta, and that the patterns at any one particular location
change greatly over the course of several years. This characteristic makes quantifying SGD
fluxes a very difficult task for both local and regional assessments. Natural geochemical tracers
have proven to be effective tools for local-scale analysis, but an approach is needed for relating
such results to a regional basis (Burnett et al., 2006a). Radon (222Rn) and radium isotopes (223Ra,
224
Ra, 226Ra, and 228Ra) have been the most extensively used naturally-occurring tracers for
assessing SGD (see reviews in Burnett et al., 2006a; Swarzenski, 2007; Charette et al., 2008).
These radioisotopes are effective tracers because they are often enriched in groundwater relative
to surface water, are relatively easy to measure, and are generally inert in marine settings. In
fresh groundwater environments, radium is particle-reactive and tends to be attached to aquifer
solids. But once encountering saline waters, radium is removed from surfaces via ion-exchange
and released into solution. Gaseous radon is almost always found at elevated activities in
groundwater, regardless of composition. These characteristics led Mulligan and Charette (2006)
to suggest that radon acts as a tracer for total SGD, whereas radium is best suited as a tracer of
saline, recirculated seawater fluxes.
We used these natural geochemical tracers to ground-truth several surface water SGD
plumes of point-source origin along the Kona Coast of the Big Island of Hawaii. We present
offshore trends in 222Rn (T1/2 = 3.8 d), 224Ra (T1/2 = 3.6 d) and salinity to reveal near shore inputs
of these natural tracers. We then use a series of coincident mass balance equations for 222Rn,
salt, and water to model the SGD fluxes of six separate groundwater discharge sites. The model
calculations are based on continuous time-series measurements of these tracers’ concentrations
near the source of each of the discharge plumes over several days. Using different representative
values for the overall and meteoric groundwater end-members in these models, we examine both
the total and the freshwater component of the discharge.
Methods
Study Site
The coastline of western Hawaii (Figure 6.1) is very arid, receiving a mean annual
rainfall of only 25 cm. Within 10 km of the coastline, however, the mountain slopes receive an
average of 102 cm annually. This is the likely area of groundwater recharge, and no rivers or
streams exist in the area to act as conduits for runoff to the coast (Kay et al., 1977). Performing
65
these studies on this side of the island thus eliminated the need to account for precipitation and
riverine inputs to the coastal ocean as virtually all freshwater input to the sea occurs as SGD.
Diffuse groundwater flow occurs along the coastline, but we emphasize here the focused,
point-sourced groundwater discharges found at or very close to the shoreline (Johnson, 2008;
Johnson et al., 2008). In the TIR images examined along this shoreline (e.g., Figure 6.2), it
appears that the point-source discharges dominate the flow. SGD exits at these point sources to
form large surface water plumes of relatively cool, brackish groundwater which buoyantly float
out from their exit points and mix seaward into the coastal ocean (Figure 6.2). Permeability in
the vesicular basalt country bedrock is high and enhanced by lava tubes, clinker zones, and
down-slope flow along and between individual lava flow sheets. In general, these plumes are
commonly associated with coastal embayments, which we believe is the likely manifestation of
subsurface focusing of water table flow lines toward such sites.
Intensive field-based campaigns were carried out at six SGD points and surrounding
regions along the Kailua-Kona coast (Figure 6.1). Kiholo Bay and Kailua-Kona Harbor were
sampled in May 2007. Honokohau Harbor was examined in February 2006. Three distinct SGD
plumes were measured further south in Kealakekua Bay: Queen’s Bath and Manini Beach in
August 2006, as well as Kahauloa Bay, which was examined in both August 2005 and February
2006. Specific site characteristics are described in more detail later. Aerial thermal infrared
(TIR) images of these sites’ freshwater plumes are shown in Johnson (2008).
Figure 6.1. Map of the (A) Hawaii Island chain, featuring the (B) western coast of the Big Island
of Hawaii. Each plus symbol represents the radon platform position in the groundwater plumes
associated with (C) Kiholo Bay, (D) Honokohau Harbor, (E) Kailua-Kona Harbor, and (F)
Queen’s Bath, Manini Beach, and Kahauloa Bay (north to south) in Kealakekua Bay.
66
Measurements
Radon was measured using a RAD7 radon-in-air monitor (Durridge Co.) modified to
analyze radon in water. A submersible pump delivered surface water (~1 m depth) to an airwater exchanger that degassed the radon from the water (Burnett et al., 2001). During each
continuous time-series analysis, we deployed the RAD7 system on a stationary floating platform
within the groundwater plume at a known distance (up to ~160 m) from the source. Radon
activities were integrated over either 30 or 60 minutes to achieve sufficient precision with
reasonable resolution. In addition, we deployed a bottom-mounted pressure-based water level
meter (HOBO Water Level Logger, Onset Corp.) and an in-situ conductivity probe (YSI 600
XLM Sonde) affixed to the submersible pump. Meteorological parameters (wind speed, air
temperature, radon in air activity) were monitored to correct for radon evasional losses to the
atmosphere using standard gas exchange equations (MacIntyre et al., 1995).
For the offshore transects, we followed the procedure outlined by Dulaiova et al. (2005)
to continuously survey for surface water radon activities. Three RAD7 monitors were connected
in parallel to allow for better time integration (10 minutes for each data point), thereby
maximizing the associated spatial resolution as the boat moved along the transect. We
continuously monitored conductivity, temperature, and GPS coordinates during the offshore
transects.
Relative to most continental margin sites we have monitored, radon levels in both
groundwater and surface water in Hawaii are found at very low concentrations. Because of the
low levels, we collected groundwater samples in 6 L polyethylene bottles (Stringer and Burnett,
2004) designed to be mounted directly to a RAD7 via a closed air loop (Lee and Kim, 2006).
The internal pump of the RAD7 circulated air through a diffusion stone in the bottle which
released the radon from the water, and then circulated it continuously through the RAD7 during
measurement. Despite the relatively low groundwater activities, radon was still much higher
than in the coastal ocean waters, and thus serves as a quantitatively effective tracer.
An additional method used to assess the radon activity in groundwater utilized a porous
membrane tubing (Accurel tubing made by Membrana) to collect radon from the water via
diffusion across the membrane (Freyer et al., 2003; Schubert et al., 2008). A closed air-loop was
created between the RAD7 and the membrane tubing to allow for constant recirculation of the air
stream until it reached activity equilibrium with the radon in the water. This method was
deployed only one time in an open, brackish well near the Manini Beach site.
Radium isotopes were collected onto acrylic fibers impregnated with MnO2 (Moore and
Reid, 1973). Large volumes (up to 200 L) of seawater and groundwater were pumped at slow
rates (~ 1 L/min) through cartridges containing this Mn-fiber. Due to the relatively low radium
activities in the water, even these sample volumes were often insufficient to collect enough
radium to detect reliably. Short-lived radium isotopes (223Ra and 224Ra) were counted via a
delayed coincidence counter system (RaDeCC; Moore and Arnold, 1996), and the long-lived
isotopes (226Ra and 228Ra) were measured via gamma spectrometry (Dulaiova and Burnett,
2004).
67
Figure 6.2. Example of an aerial TIR image of a buoyant SGD plume (taken from Johnson,
2008). Note the temperature of the plume is cooler than the ambient ocean water.
Mass balance model to quantify SGD
Groundwater fluxes are commonly modeled using a radon box model introduced by
Burnett and Dulaiova (2003). This model, however, assumes that the flow is occurring in a more
or less uniform pattern upward through the seabed and thus expresses SGD rates in terms of
vertical advection velocities (e.g., cm/day). In order to convert these results into an absolute
volumetric flux, one must know the area over which seepage is occurring. The point-source
nature of the discharges in our study sites, however, dictated the need for a different approach
because the flows originate from discrete portals along the shoreline. We elected to use a
modified version of a LOICZ-style box model approach (Gordon et al., 1996) to derive SGD
rates from these point-source discharges. We define a domain that captures the flow across a
measured cross-section and includes a salt balance to further constrain the assessments. The
model used here was originally presented by Peterson et al.(2007) , but our current version
contains some slight modifications from that original form.
The non-steady-state mass balance box model employed in this study uses the relatively
high radon, low salinity nature of discharging groundwater, and the relatively low radon, high
salinity character of open ocean water to determine the flux of groundwater and seawater into
and out of a specified control volume, which in our case, is described as a ‘plume’ (Figure 6.3).
Water inputs to the plume are considered from open ocean intrusion (QIN) and SGD (QSGD).
These fluxes are balanced by plume water loss to the open ocean (QOUT). By convention, we
consider fluxes directed offshore from the plume to be positive (QSGD and QOUT), whereas
landward fluxes are negative (QIN). We ignore any direct rainfall fluxes that are negligible at
68
these sites compared to the other fluxes over these short time scales (hours to days). Constant
values for each site are assigned to the radon and salinity end-members for groundwater (RnSGD
and SSGD, respectively) and for open ocean water (RnO and SO, respectively). Continuously
monitoring the plume radon (RnC) and salt (SC) concentrations allows us to examine how the
groundwater flux changes over time. Reference to the measured radon activity here implies
“excess” radon, meaning the total radon activity in the water minus that activity supported by its
parent, 226Ra. Water densities (ρSGD, ρO, and ρC) are included to convert salinity to absolute salt
mass.
Figure 6.3. Diagram of model variables and their interactions used to simulate the changing
water fluxes into and out of the control volume, defined here as the ‘groundwater plume’. See
text for details.
Other parameters required by the model include the length, width, and depth of the
groundwater plume. We take the length term (l) to be the distance between the SGD source
location at the shoreline and the site where the radon and salinity are continuously monitored.
The width (w) is taken as the shore-parallel distance across the buoyant water plume at the site of
the radon platform. This is often based on measurements from the TIR images in Johnson
(2008). At sites where the freshwater plume extends to the seabed at the radon platform, we take
the measured depth record (d) as the vertical dimension for the model. In places where vertical
temperature and salinity (T/S) profiling revealed a stratified water column (Johnson, 2008; A.G.
Johnson unpubl.), we instead use the thickness of the buoyant layer as a constant depth, which is
a reasonable approximation within the length scales we are considering. At all of our study sites,
the water depth dropped quickly within a few meters of the shoreline and thereafter maintained a
constant depth to the radon platform. Thus, we estimate the plume volume as V=lwd.
The model is developed using three simultaneous mass balance equations for water, salt,
and radon to solve for three unknowns: QSGD, QIN, and QOUT. We first parameterize the model
using average values over an entire tidal cycle (24-hours in this case) using the following
equations for water balance (see also Appendix A):
69
ΔVC
= QOUT − QIN − QSGD
Δt
(6.1)
for salt balance:
(
lw S C ( t +1) ρ C (t +1) d C ( t +1) − S C (t ) ρ C (t ) d C (t )
Δt
)=S
C
ρ C QOUT − S O ρ O QIN − S SGD ρ SGD QSGD
(6.2)
and for radon balance:
(
lw RnC ( t +1) d C ( t +1) − RnC (t ) d C ( t )
Δt
) = Rn Q
C
OUT
− RnO QIN − Rn SGD QSGD
(6.3)
Terms marked with subscripts (t) and (t+1) refer to consecutive measurements made throughout
the time-series deployment. We ultimately solve these three equations for QSGD. For the
parameterization, the left sides of the equations represent the average hourly change in water
volume (ΔVC/Δt), salt mass (ΔSC/Δt), and radon activity (ΔRnC/Δt) throughout a complete tidal
cycle. The model used previously by Peterson et al. (2007) assumed steady-state conditions over
the course of a tidal cycle (therefore, ΔVC/Δt, ΔSC/Δ1, and ΔRnC/Δt were assumed to be 0), but
parameterizing according to this current approach eliminates any error associated with that
assumption. The right sides of the equations represent the cumulative effects of the average
coastal groundwater plume fluxing offshore, the offshore water flowing into the system, and the
groundwater fluxing into the system. Solving these equations for the three unknowns, QOUT, QIN,
and QSGD, establishes average water flux values over the tidal cycle.
We next examine the dynamics of the flows by using smaller time increments (as in Q +
ΔQ). A few processes inherently require an inverse relationship between tracer end-member
concentration and the corresponding water flux to achieve the expected response. For these
processes, we expect the water flux to decrease if the end-member tracer concentration increases,
so we use (Q – ΔQ). Radon input via SGD and salt input from the open ocean follow this
convention. All other fluxes are represented as (Q + ΔQ).
The resulting equations are used for time-series modeling for water balance:
ΔVC
= (QOUT + ΔQOUT ) − (QIN + ΔQIN ) − (QSGD + ΔQSGD )
Δt
(6.4)
for salt balance:
(
lw SC ( t +1) ρC ( t +1) dC ( t +1) − SC ( t ) ρC ( t ) dC ( t )
Δt
)=S
C
ρC (QOUT + ΔQOUT ) − SO ρO (QIN − ΔQIN )
− S SGD ρ SGD (QSGD + ΔQSGD )
and for radon balance:
70
(6.5)
(
lw RnC ( t +1) d C ( t +1) − RnC ( t ) d C ( t )
Δt
) = Rn
C
(QOUT
+ ΔQOUT ) − RnO (Q IN + ΔQ IN ) − Rn SGD (QSGD − ΔQSGD ) (6.6)
Atmospheric evasion losses of radon as well as decay losses, while minimal (often < 1% of
radon inventory for the time steps used here), were accounted for by including these loss terms in
Eq. (7.6). Solving these equations for ΔQOUT, ΔQIN, and ΔQSGD at each time step (30 or 60
minutes) allows for the calculation of the variation of the water fluxes from their average value
found during the parameterization step, and ultimately for the total water flux over the entire
measurement interval.
Results
Kahauloa Bay
Our most extensively studied SGD plume was in Kahauloa Bay (Figure 6.1F). This is a
very small embayment (130 m long, 40 m wide) with steep, fixed walls constraining the size of
the plume. Field observations indicate that the SGD inputs are limited to the back shore of the
bay. We measured radon, radium, and salinity along an offshore transect from this bay on 16
August 2005, and performed two separate time-series analyses, 15-17 August 2005, and 10-14
February 2006.
The results of the offshore transect are shown in Figure 6.4. The groundwater radium
isotope and radon results from Tables 6.1 and 6.2 suggest that all these tracers are found in
relatively low activities in this part of Hawaii, with 223Ra, 226Ra, and 228Ra often below reliable
detection limits. Thus, for radium, we only present the 224Ra trends offshore, which are very
similar to the 222Rn patterns, showing relatively high activities in the near shore waters and
decreasing quickly offshore. The 224Ra data indicate a few offshore sites where elevated radium
concentrations are found. This is likely caused by the pulsing of groundwater discharge creating
pockets of higher tracer concentration offshore. Excess radon has likely been removed from
these areas by atmospheric evasion. Beyond about 2000 m from shore, open ocean signatures
dominate the 222Rn, 224Ra, and salinity concentrations. We take the average 222Rn and salinity
values from this transect beyond 3000 m offshore (radon = 1.07 Bq/m3; salinity = 35.7) to
represent the open ocean end-members for the SGD model.
We sampled groundwater from 6 sites in the vicinity of this bay (reference numbers 23-28 in
Tables 6.1 and 6.2). These samples were collected from brackish, rock-lined large diameter
wells and ancient “Hawaiian baths” along the coastline, which exhibited dampened tidal
fluctuations indicating that they are hydraulically connected to the ocean. Keei Well D
(reference number 27 in Tables 6.1 and 6.2) is an upland freshwater production well and likely
does not represent the discharging groundwater, so we opt to neglect this sample. We use the
results of the other groundwaters sampled in August 2006 (n = 5) to estimate the average
groundwater end-member for this bay to have a radon activity of 83 ± 25 Bq/m3 and a salinity of
6.5 ± 1.2.
71
Figure 6.4. Offshore transect of (A) salinity, (B) 222Rn, and (C) 224Ra from the plume emanating
from Kahauloa Bay. This transect was collected in August 2005.
The raw data from the 2005 time-series analysis for radon, salinity, and temperature are
shown in Figures 6.5A, B, and C, respectively. The radon activities, while very low, show a
pattern of fluctuating inversely with the tide, increasing during the ebb tide and decreasing
during flood tide. Salinity does not show much response to the smaller amplitude tides, but
varies directly with the extreme tide. Temperature indicates significant cooling following the
largest high tide from input of cold groundwater, and significant warming during the flood tide
leading to this high tide, as a result of warmer ocean water mixing into the system. The lag
shown between water level fluctuations and tracer response is apparently due to the mixing
between the bottom and surface layers at this site.
Figure 6.5D shows the model results for the SGD flux based on this time-series analysis.
Constants and end-member values used for this and all successive data sets are provided in Table
6.3. Positive fluxes represent groundwater discharge into the bay and tend to occur during ebb
tide. Negative fluxes are found during flood tide and indicate an additional loss of the
groundwater tracer parameters. Since the buoyant plumes that are visible via TIR result from
positive groundwater fluxes, we average all positive discharges to find an overall SGD rate from
this plume of 1100 m3/day (Table 6.3).
72
73
Reference
number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Well
name
Kiholo Lagoon
Fish Pond
Lava Tube
Energy lab - Well 6
Energy lab - Well 2
Hualalai Well
KNP-1
KNP-2
Honokohau Well
KNP-3
HW-14
HW-15
HW-16
HW-17
HW-6
HW-9
Kahaluu Well A
Kahaluu Well C
Kahaluu Well B
Kahaluu Well D
Keauhou Bath
Halekii Well
Queen's Bath
Kelly's Well
Manini Bath
Vergi's Well
Keei Well D
City of Refuge
Description
Surface Water Sample
Brackish Pond
Freshwater in Lava tube
Natural Energy Laboratory
Natural Energy Laboratory
Upland Production Well 4258-03
Kaloko National Historical Park Well 4161-01
Kaloko National Historical Park Well 4161-02
Upland Production Well 4158-02
Kaloko National Historical Park Well 4061-03
Honokohau Harbor Expansion Well
Honokohau Harbor Expansion Well
Honokohau Harbor Expansion Well
Honokohau Harbor Expansion Well
Honokohau Harbor Expansion Well
Honokohau Harbor Expansion Well
Upland Production Well 3557-01
Upland Production Well 3557-03
Upland Production Well 3557-02
Upland Production Well 3557-04
Ancient Hawaiian Bath
Upland Production Well
Ancient Hawaiian Bath
Ancient Hawaiian Well
Ancient Hawaiian Bath
Ancient Hawaiian Well
Upland Production Well
Brackish Pond
Time series
site proximity
Kiholo
Kiholo
Kiholo
Honokohau
Honokohau
Honokohau
Honokohau
Honokohau
Honokohau
Honokohau
Honokohau
Honokohau
Honokohau
Honokohau
Honokohau
Honokohau
Kailua-Kona
Kailua-Kona
Kailua-Kona
Kailua-Kona
Queens Bath / Kahualoa
Manini / Kahauloa
Manini / Kahauloa
Kahauloa
Kahauloa
Kahauloa
Kahualoa
Queens Bath / Kahualoa
Kahualoa
Manini / Kahualoa
Kahualoa
Honokohau
Honokohau
Honokohau
Honokohau
Honokohau
Honokohau
Used for groundwater end-member?
Latitude
N19° 51.633'
N19° 51.318'
N19° 51.242'
N19° 43.600'
N19° 42.754'
N19° 42.250'
N19°41.233'
N19° 41.155'
N19° 40.940'
N19° 40.705'
N19° 40.120'
N19° 40.113'
N19° 40.083'
N19° 40.076'
N19° 39.976'
N19° 39.878'
N19° 34.978'
N19° 34.962'
N19° 34.904'
N19° 34.884'
N19° 33.666'
N19° 31.087'
N19° 28.907'
N19° 28.258'
N19° 28.249'
N19° 28.180'
N19° 27.725'
N19° 27.486'
Longitude
W155° 55.334'
W155° 55.185'
W155° 55.386'
W156° 03.569'
W156° 02.906'
W155° 58.438'
W156° 01.750'
W156° 01.409'
W155° 57.865'
W156° 01.330'
W156° 01.083'
W156° 01.155'
W156° 01.211'
W156° 01.265'
W156° 01.419'
W156° 01.491'
W155° 56.988'
W155° 56.971'
W155° 56.983'
W155° 56.954'
W155° 57.703'
W155° 54.974'
W155° 55.997'
W155° 55.238'
W155° 55.243'
W155° 55.261'
W155° 52.820'
W155° 55.469'
Depth
(m)
n/a
n/a
3.0
n/a
4.8
485.6
7.2
17.8
557.4
12.0
n/a
9.2
15.7
16.4
13.1
18.0
283.8
288.7
300.2
295.3
1.5
105.0
0.3
0.5
0.3
1.0
353.3
1.0
Table 6.1. Summary characteristics of the sampled groundwater sources. These sites are arranged in order from north to south.
Reference numbers appear throughout the text. Those sites listed under the ‘Used for end-member’ category show the samples that
are averaged to provide an end-member value at a specific site for the SGD model.
The radon, salinity, and temperature patterns measured in Kahauloa Bay during the
February 2006 time-series are shown in Figure 6.6 (A, B, and C, respectively). The radon
patterns are again inversely correlated to the tidal fluctuations, while the salinity shows a direct
correlation. Water temperatures increase during the high tides, and tend to decrease immediately
following the high tides, but are dominated by 24-hour solar cycles rather than 12-hour tidal
cycles. During this deployment, the radon platform was 5 m farther from shore than during
2005, but all other parameters remain the same for the SGD model. Figure 6.6D shows that the
model results are similar in pattern to those seen in 2005, but the average positive flux is
somewhat higher at 1600 m3/day (Table 6.3).
Figure 6.5. Time-series analysis of (A) 222Rn activity, (B) salinity, and (C) temperature in
surface waters of the Kahauloa Bay plume measured in August 2005. (D) Modeled total SGD
results are also included, after a 3-point smoothing function. Solid lines represent tidal
variations.
74
75
sampled
number
1
2
3
4
5
6
7
Salinity
Temperature
Rn
(Bq/m3)
222
(°C)
20 May 07
9.5
23.0
162 ± 15
20 May 07
2.7
25.3
707 ± 37
17 May 07
2.1
22.7
356 ± 29
16 Feb 06
28.8
23.6
604 ± 40
16 Feb 06
11.3
21.8
424 ± 33
15 Feb 06
0.1
21.1
5,730 ± 120
17 Aug 05
6.8
22.4
n/a
14 Feb 06
7.3
20.9
192 ± 18
8
17 Aug 05
5.2
21.7
n/a
14 Feb 06
5.8
23.8
611 ± 56
15 Feb 06
0.1
21.8
1540 ± 78
9
17 Aug 05
10
11.5
22.5
n/a
14 Feb 06
11.8
21.5
242 ± 22
11
17 May 07
7.1
20.7
500 ± 28
12
16 May 07
8.2
20.9
274 ± 20
13
16 May 07
15.3
20.5
456 ± 26
14
17 May 07
17.9
20.2
431 ± 23
16 May 07
19.0
20.7
371 ± 23
15
16 May 07
16
23.8
20.3
520 ± 28
13 Feb 06
17
0.4
21.0
1450 ± 48
18
13 Feb 06
0.2
21.4
1140 ± 50
19
13 Feb 06
0.5
21.0
1415 ± 75
20
13 Feb 06
0.4
21.0
1150 ± 60
09 Aug 06
4.9
22.0
321 ± 18
21
13 Feb 06
22
0.1
21.8
879 ± 43
10 Aug 06
23
6.7
21.5
49 ± 10
24
07 Aug 06
7.1
21.1
120 ± 16
25
06 Aug 06
5.1
22.2
86 ± 14
19 May 07
6.5
20.4
131 ± 14
19 May 07*
6.5
20.4
106 ± 20
07 Aug 06
5.7
20.6
83 ± 14
26
13 Feb 06
27
0.1
19.8
340 ± 26
15 Feb 06
0.1
19.6
375 ± 32
28
11 Aug 06
8.1
23.4
78 ± 9
* denotes a time-series record average of this well over 24-hours
Date
Reference
Ra
(Bq/m3)
n/a
n/a
2.06 ± 0.17
13.40 ± 0.73
7.22 ± 0.35
0.06 ± 0.03
2.76 ± 0.36
1.48 ± 0.11
6.46 ± 0.55
2.65 ± 0.39
0.22 ± 0.08
1.27 ± 0.23
2.43 ± 0.18
1.20 ± 0.16
1.80 ± 0.18
4.48 ± 0.34
5.17 ± 0.27
9.98 ± 0.62
7.67 ± 0.47
0.13 ± 0.02
0.04 ± 0.02
0.26 ± 0.06
BD
0.28 ± 0.03
BD
n/a
0.23 ± 0.04
0.73 ± 0.09
n/a
n/a
BD
0.12 ± 0.04
n/a
n/a
n/a
n/a
0.12 ± 0.05
1.08 ± 0.21
0.42 ± 0.05
BD
BD
0.08 ± 0.03
0.60 ± 0.12
0.14 ± 0.08
BD
0.19 ± 0.11
0.35 ± 0.08
0.17 ± 0.06
0.19 ± 0.05
0.67 ± 0.12
0.76 ± 0.12
0.79 ± 0.13
0.18 ± 0.07
BD
BD
BD
BD
0.02 ± 0.01
BD
n/a
0.03 ± 0.01
0.06 ± 0.01
n/a
n/a
0.03 ± 0.1
BD
n/a
n/a
224
Ra
(Bq/m3)
223
n/a
n/a
n/A
0.37 ± 1.2
0.65 ± 0.43
BD
n/a
0.80 ± 0.45
n/a
1.3 ± 1.7
BD
n/a
1.02 ± 0.04
n/a
n/a
n/a
n/a
n/a
n/a
0.18 ± 0.14
BD
0.19 ± 0.49
0.21 ± 0.18
BD
BD
n/a
BD
BD
n/a
n/a
BD
0.08 ± 0.33
n/a
n/a
Ra
(Bq/m3)
226
n/a
n/a
n/a
4.8 ± 5.2
6.3 ± 1.7
BD
n/a
1.7 ± 1.6
n/a
0.2 ± 6.3
BD
n/a
2.3 ± 1.3
n/a
n/a
n/a
n/a
n/a
n/a
0.36 ± 0.55
1.06 ± 0.93
1.9 ± 1.7
BD
0.01 ± 0.55
BD
n/a
0.39 ± 0.74
BD
n/a
n/a
BD
BD
n/a
n/a
Ra
(Bq/m3)
228
Table 6.2. Measured parameters for each of the sampled groundwater sources. Reference numbers refer to the
specific site characteristics listed in Table 6.1. Some sites were sampled during multiple field trips. n/a indicates
that this particular parameter was not measured during the specific sampling, whereas ‘BD” indicates that the
measurement was below the detectable activity.
Figure 6.6. Time-series analysis of (A) 222Rn activity, (B) salinity, and (C) temperature in
surface waters of Kahauloa Bay measured in February 2006. (D) Modeled total SGD and (E)
tidal flux results are also included and both represent a 3-point smoothing. Solid lines represent
tidal variations.
Manini Beach
Just north of Kahauloa Bay, inside the confines of Kealakekua Bay, another plume is
discharging at Manini Beach (Figure 6.1F). There are several groundwater inputs along the
southeast shoreline of Kealakekua Bay, but we only examined the one coming directly from
Manini Beach. This is a narrow, rocky beach with lava outcrops lining the shoreline. Field
observations indicate that the groundwater is mostly discharging from the beach face and along
the base of the rock outcrops. We performed an offshore transect from this plume on 03 August
2006, and a time-series analysis 08-11 August 2006.
Figures 6.7A, B, and C show the results of the offshore transect for salinity, 222Rn, and
224
Ra, respectively. The transect represents an offshore to onshore sampling strategy, conducted
during the outgoing tide to maximize the measured tracer concentrations. As with the transect
76
from Kahauloa Bay, we see detectable elevations in the tracer concentrations (222Rn and 224Ra)
and significantly lower salinity within the plume to about 2000 m from shore. Beyond that point,
the salinity remains high and uniform, while the 222Rn and 224Ra remain at background levels.
These transects reveal that a much more intense sampling scheme concentrated within the first
several hundred meters from shore would have been necessary in order to use offshore
distributions of radium isotopes to independently quantify SGD, such as that employed by Street
et al. (2007).
Table 6.3. Model parameters selected for each groundwater plume and the associated total and
freshwater SGD fluxes determined by averaging all positive SGD fluxes over the time-series.
The results are arranged in order from north to south.
Groundwater
End-members*
222
Rn/salinity
(Bq/m3) / (g/kg)
Distance
to
Source
(m)
Plume
Width
(m)
Flow
Type#
Estimated
Discharge
(m3/day)
120
636 / 0
546 / 4.8
130
82
FSGD
TSGD
6300
7100
2.4
170
652 / 0
425 / 15.2
10
375
FSGD
TSGD
8600
12,000
Kailua-Kona Harbor
18-20 May 2007
n = 65
1.0**
83
1264 / 0
713 / 15.2
162
86
FSGD
TSGD
5100
8700
Queens' Bath
6-8 Aug. 2006
n = 44
2.0**
9.8
130 / 0
107 / 6.7
100
40
FSGD
TSGD
2900
3500
Manini Beach
8-12 Aug. 2006
n = 66
1.2**
15
515 / 0
105 / 6.5
70
100
FSGD
TSGD
1100
5100
Kahauloa Bay
15-17 Aug. 2005
n = 36
1.8
15
153 / 0
83 / 6.5
20
40
FSGD
TSGD
630
1100
Kahauloa Bay
10-14 Feb. 2006
n = 90
2.4
13
153 / 0
83 / 6.5
25
40
FSGD
TSGD
940
1600
Site
Avg. Modeled
Depth
(m)
Avg. Excess
222
Rn
(Bq/m3)
Kiholo Bay
15-18 May 2007
n = 92
1.3
Honokohau Harbor
14-16 Feb. 2006
n = 87
† 226
†
Ra activity in surface water = 0.90 Bq/m3 (Kahauloa, Kiholo); 1.4 Bq/m3 for Kailua-Kona; 0.44 Bq/m3 for
3
Manini; 0.52 Bq/m for Queen's Bath
3
* Ocean end-members for radon and salinity were held constant at 1.1 Bq/m and 35.7 g/kg, respectively.
** Thickness of buoyant plume held constant rather than use absolute measured depths.
#
TSGD = Total submarine groundwater discharge (recirculated seawater + freshwater); FSGD = fresh submarine
groundwater discharge
77
Several of the large-diameter rock-lined brackish wells and baths sampled for Kahauloa
Bay are in the same vicinity as Manini Beach. We deployed the membrane tubing system in one
of these (reference number 25 in Tables 6.1 and 6.2) for 24 hours to monitor the radon activity in
the groundwater supplying this bath. We use the average results from this analysis (radon = 106
± 20 Bq/m3; salinity = 6.5 ± 0.1) as the appropriate end-member values for this plume. These
results are very similar to discrete samples collected from two other Manini Beach sites
(reference numbers 24-26 in Tables 6.1 and 6.2).
Temperature and salinity profiling of the plume waters at Manini Beach revealed
stratification of the buoyant water layer. We therefore use the thickness of this plume (1.2 m) at
the radon platform (70 m from shore) as a constant plume thickness in the SGD model. The
width of this plume at the platform was measured from aerial TIR imagery to be 100 m (Johnson,
2008).
The radon activities, salinity, and temperature measured during the Manini Beach timeseries are shown in Figures 6.8A, B, and C, respectively. The tracer patterns for Manini Beach
are very similar to those from Kahauloa Bay, with the radon peaking during low and ebb tide.
Salinity and temperature again tend to vary directly with the tide. Figure 6.8D shows the
resulting SGD fluxes for Manini Beach. The average positive SGD rate for this plume is 5100
m3/day (Table 6.3), but rates often reach closer to 10,000 m3/day during peak flows.
Figure 6.7. Offshore transect of (A) salinity, (B) 222Rn, and (C) 224Ra from the plume at Manini
Beach. This transect was collected in August 2006.
78
Queen’s Bath
The SGD plume associated with Queen’s Bath is located in the northern corner of
Kealakekua Bay (Figure 6.1F). A thin, shallow shelf extends seaward from this corner, before
the seafloor drops to about 50 m near the mouth of the bay. Our platform was deployed along
this shelf, about 100 m from the SGD source in the corner of the bay. The width of the plume at
the platform was measured to be 40 m. Like Manini Beach, T/S profiling of this plume revealed
a buoyant layer extending down to a depth of about 2.0 m at the platform, so we take this as a
constant plume thickness throughout the time-series.
As shown in Table 6.1, only one groundwater well was accessible in the vicinity of
Queen’s Bath (reference number 23). It was an ancient Hawaiian bath, built into the ground in
the woodlands near the coast. Unfortunately, sampling was difficult because even at high tide,
the water depth in the bath was only about 30 cm. The sample collected from this well displayed
a similar salinity as other groundwaters sampled in the area (6.7), but the very shallow depth may
have led to substantial degassing of the radon (measured value = 48 Bq/m3). Instead, we assign
the radon end-member for this time-series based on a plot of the measured radon versus salinity
during the time-series (Figure 6.9A). The inverse correlation of this plot displays reasonable
linearity, so we extrapolate the trend to a salinity of 6.7 to find the radon end-member of 107
Bq/m3. Short temporal mixing scales and influence from other discharges likely cause the scatter
in these data. Nonetheless, this value is reasonably close to the values measured in other
groundwater sites around Kealakekua Bay.
Figure 6.10A shows the model result for SGD during this time-series. As seen before,
the calculated SGD rates show the highest flux during the largest ebb tide of each day. The
overall average positive SGD rate for this plume is 3500 m3/day (Table 6.3).
Kailua-Kona Harbor
TIR imagery of Kailua-Kona Harbor (Figure 6.1E) reveals numerous point-source
discharges along the shoreline (Johnson, 2008). We deployed the radon platform in an area
where only the most northeastern plume exiting near the corner of the harbor would affect our
results. This area is considerably deeper than the other sites we have studied (average depth =
4.5 m). Here, T/S profiling was essential to constrain the thickness of the buoyant water plume,
which was found to be a constant 1.0 m. A designated swimming area prevented us from
deploying the platform closer than 162 m from the source, and the width of the plume here was
measured to be 86 m.
No coastal wells were available in this area, and any direct measure of the discharging
groundwater was not possible. Therefore, as with Queen’s Bath, we estimate the radon endmember based on a regression of the time-series linear radon versus salinity plot. This site is
very near Honokohau Harbor, where many coastal groundwater wells were available. The
average salinity of the wells associated with Honokohau Harbor is significantly higher (15.2) as
a result of groundwater pumping and subsequent seawater infiltration in this area. We chose to
use this salinity as the end-member for Kailua-Kona Harbor also, because this site is located in
close proximity to Honokohau Harbor. The corresponding radon activity based on the regression
line at a salinity of 15.2 is thus found to be 713 Bq/m3 (Figure 6.9B). We recognize the inherent
uncertainties associated with these assumptions, but no other end-member sources are available
for this region.
79
The model results for SGD in the Kailua-Kona Harbor plume are shown in Figure 6.10B.
Averaging the positive fluxes yields an average discharge rate of 8700 m3/day (Table 6.3), but
rates often reach close to 20,000 m3/day. As with all other sites, the SGD tends to peak during
the low/ebb tides, and the negative fluxes occur during flood tide.
Figure 6.8. Time-series analysis of (A) 222Rn activity, (B) salinity, and (C) temperature in
surface waters at Manini Beach measured in August 2006. (D) Modeled total SGD results are
also included after a 3-point smoothing function. Solid lines represent tidal variations.
Honokohau Harbor
Honokohau Harbor is a small boat basin located farther north along the coastline (Figure
6.1D). This site is located in a protected area, as the tidal wedge can only pass through an 80 m
wide entrance. Honokohau Harbor was first constructed in 1970, and later expanded in 1978, by
80
blasting and excavating the lava to form a vertical-walled basin (Bienfang, 1983). Several
authors have noted the abundance of groundwater discharging into the basin through the walls as
a result of the landward excavation (Bienfang and Johnson, 1980; Gallagher, 1980; Bienfang,
1983).
Examination of TIR images of this harbor (Johnson et al., 2008) reveals many different
point-source SGD inputs along the entire perimeter of the basin, mostly apparent along the
basin’s back wall. In addition, divers working with our group have identified diffuse flow
occurring nearly ubiquitously around the basin’s perimeter. As a result, we deployed our
platform 10 m from the back of the harbor. Instead of using the width of the plume as the width
dimension for the model, however, we use the perimeter length of the wall along the entire back
section of the harbor (375 m), as this represents the likely source width. At our station, no
detectable stratification of the buoyant layer was observed from T/S profiling, so we use the
actual depths measured during the time-series for the model.
In the near future, harbor managers are planning another expansion project for this
harbor. During the exploratory stages of the project, a series of wells were installed around the
harbor location. We used the average radon and salinity from these wells, measured in May
2007 (reference numbers 11-16 in Tables 6.1 and 6.2), as the groundwater end-member values
(radon = 425 ± 92 Bq/m3; salinity = 15.2 ± 6.5). The end-member salinity is considerably more
saline than those from the other locations, indicating that saltwater intrusion into the aquifer in
this area is more pronounced than at the other sites.
A significant time shift was observed in both the radon and salinity records as those
tracers peak several hours after the tidal peaks. This pattern is likely due to the protected nature
of this harbor, in that mixing is more inhibited within the harbor than the other study sites. In
spite of this time lag, Figure 6.10C shows that the SGD patterns peak at the low tides, indicating
that the hydraulic gradient affected by the changing ocean level inside the harbor influences the
groundwater fluxes. The average positive discharge is estimated to be 12,000 m3/day in
Honokohau Harbor (Table 6.3).
Kiholo Bay
Our northernmost study site was at the mouth of a small lagoon along the north side of
Kiholo Bay (Figure 6.1C). At this site, a peninsula separates a deep inner lagoon from the open
ocean. TIR imaging of this area (Johnson, 2008) reveals that the lagoon is the receiving waters
of the SGD discharge. A small island in the mouth of the lagoon further inhibits mixing of this
groundwater offshore, so the plume exits the lagoon through two narrow openings around the
island. We chose to install our platform in the larger opening, between the peninsula and the
island, about 130 m from the lagoon’s back shore. The width is taken as the distance across the
mouth of the lagoon (82 m). There was no detected stratification in this shallow area, so we use
the measured water depths for the model.
Three sources of possible groundwater end-members were sampled around the Kiholo
area (reference numbers 1-3 in Tables 6.1 and 6.2), including brackish water inside a lava tube (~
400 m south of the bay), an expansive brackish fishpond fed by groundwater, and a relatively
low salinity water sample inside the inner bay. The average salinity of all these samples was 4.8,
but the range in radon activities suggests that either different water masses were sampled, or the
samples were influenced by different amounts of gas exchange. We therefore take this salinity,
81
and use it in the linear regression of the time-series radon versus salinity record to assign a radon
end-member of 546 Bq/m3 (Figure 6.9C).
Figure 6.9. 222Rn versus salinity plots of the raw time-series measurement values collected at
(A) Queen’s Bath, (B) Kailua-Kona Harbor, and (C) Kiholo Bay. The regression lines are
extrapolated to estimate a radon end-member in the groundwater at these sites.
Figure 6.10D shows the calculated SGD fluxes based on this deployment. SGD rates
slightly increase throughout the moderate tidal changes, but sharply decrease to negative fluxes
during the flood tide following the lowest low tides. The average positive SGD rate for the
plume waters exiting the lagoon in Kiholo Bay was estimated to be 7100 m3/day (Table 6.3).
82
Discussion
In addition to determining the total groundwater fluxes (meteoric + recirculated seawater)
associated with each of these plumes, we can adjust the end-members to estimate the flux of the
purely fresh, meteoric component of the groundwater discharge. Extrapolating the linear timeseries radon versus salinity plot for each site to zero salinity should yield an estimate of the radon
activity in the pure meteoric water. This assumes that these trends remain linear throughout the
salinity range. Table 6.3 includes the resultant end-members for each site using an assumed
salinity of 0 as well as the model results based on these end-members. The radon end-member
estimates are still considerably lower than most of the freshwater, upland wells (reference
numbers 6, 9, 17-20, 22, and 27 in Tables 6.1 and 6.2), likely due to different water to rock ratios
or longer residence times that the upland waters spend in the perched aquifers found along the
mountain slopes of the island. The modeled freshwater flux estimates range from roughly 20%
of the total flow (Manini Beach) to nearly 90% (Kiholo Bay).
The estimates of both total and fresh SGD are subject to a number of uncertainties.
Model parameters that exert the most influence over each result are the groundwater radon endmember and the coastal plume volume. Varying the selected groundwater radon end-member for
any particular site leads to an inverse linear change in the model result, i.e., an increase in the
radon end-member of 10% leads to a corresponding decrease in SGD flux by the same
percentage. The lack of representative groundwater wells in the region of some of our study
sites thus creates some uncertainty. This is especially true for sites like Kailua-Kona Harbor,
where the radon end-member could only be assigned from the linear radon versus salinity plot.
One benefit of the model employed here is that it uses a combination of tracers for the
groundwater end-member in order to lessen the impact of this uncertainty. If a groundwater
sample were collected closer to the coastline, for example, it would likely have a higher salinity,
but lower radon content. These factors should balance out in the model calculations to provide
the same result.
All other end-member values have considerably less effect on the modeled SGD results.
Changing the salinity in groundwater as well as both radon and salinity end-members in the open
ocean vary the model result linearly, but within the range associated with the measurement
uncertainty of the parameters. The only requirement of these end-members is that their values
remain outside the range in which they are measured in the coastal groundwater plume; the open
ocean salinity value, for example, must only be greater than the highest measured plume salinity
during the time-series measurement.
The model results are also largely dependent upon our interpretation of the dimensions of
the plume between our platform and the source. Varying a plume dimension results in a linear
change in the corresponding model results for any particular site where a 10% increase in the
dimension creates a subsequent 10% increase in the SGD flux. We are confident that our length
(distance to shore) and depth dimensions are well constrained for each site, because these have
been directly measured. In cases where the depth is assigned based on temperature-salinity
profiling, a relatively small scale uncertainty (e.g., 0.5 m) would translate to a large uncertainty
in the overall volume. The width uncertainty is rooted in our analysis of the TIR images in cases
where that is the only means of assessing this dimension. It may be that such dimensions vary
over time. Kahauloa Bay is not subject to this uncertainty, nor is Kiholo Bay, as the plumes at
these sites are horizontally constrained by physical land masses.
83
Figure 6.10. SGD flux model results from the (A) Queen’s Bath, (B) Kailua-Kona Harbor, (C)
Honokohau Harbor, and (D) Kiholo Bay time-series deployments. All results represent a 3-point
smoothing. Solid lines represent tidal variations.
Previous studies have assessed SGD rates along the west Hawaii coastline, although none
of these was based on tracer measurements. Bienfang (1980) used physical oceanographic
measurements to estimate the SGD flux into Honokohau Harbor. His study suggested that an
equivalent of 5,500 to 7,500 m3/d of total groundwater flow into the harbor each day. This study
was conducted prior to the harbor expansion project of 1978, which likely resulted in higher
discharge because of the larger wall surface area after the expansion. Our estimates of 12,000
m3/day (8600 m3/day fresh SGD) thus are in reasonable agreement with these earlier results
suggested by Bienfang (1980).
84
Kay et al. (1977) conducted a series of water balance studies for the region immediately
north of Kiholo Bay, and concluded that their best estimate of coastal groundwater discharge was
15,000 m3/km.day. Kanehiro and Peterson (1977) reached the same conclusion in their own
water balance estimates. These values would correspond to the fresh component of SGD that we
have measured because they are based upon groundwater recharge rates and do not consider
recirculated seawater. Recent increases of groundwater pumping in this region in the three
decades since those studies were conducted would likely result in a decrease of these estimates.
Combined, the total fresh groundwater flux from the three plumes studied within Kealakekua
Bay (Figure 6.1F) contributes 4900 m3/day (2500 m3/day per km of shoreline). This flux
estimate is a minimum as TIR imagery reveals other large SGD plume inputs within the bay
(Johnson, 2008) that were not quantified. While this comparison is an oversimplification, it
clearly suggests that the fresh SGD fluxes are at least on the same order as the previous studies
on the island.
On the Big Island of Hawaii, the western (leeward) side receives much less precipitation
than does the windward side. As the population continues to increase in these dry regions (e.g.,
Kailua-Kona area), groundwater promises to become one of the most important commodities in
the region for water resources. Golf courses, for example, rely heavily on groundwater for
irrigation, and area managers can benefit from knowing where losses of groundwater are
occurring. Since groundwater is the only link between land and the ocean, any contamination
could begin to affect local ecology and coral reefs.
85
CHAPTER 7
TRACKING SUSPENDED PARTICLES WITH NATURALLY-OCCURRING
RADIONUCLIDES AND CHEMICAL TRACERS IN THE APALACHICOLACHATTAHOOCHEE-FLINT RIVER SYSTEM
Article in preparation for submission to Geochimica et Cosmochimica Acta
Abstract
Suspended sediment in rivers can carry substantial amounts of metals, nutrients, and pollutants
downstream to estuaries. These can then desorb and become bioavailable in coastal marine
waters. A quantitative understanding of suspended sediment transport in the ApalachicolaChattahoochee-Flint (ACF) River system is especially important for water quality management
between Florida, Georgia, and Alabama. We applied various naturally-occurring radionuclides
(226Ra, 228Ra, 7Be, 210Pb, 40K) and other tracers (total suspended solids (TSS), Ca, org-C, As, Sb,
Zn) to investigate suspended particle transport behavior in this river system. During base flow,
most particles introduced to the river system are either “old” - resuspended bottom sediments or
derived from deeper soil horizons. During high flow, however, our tracer evidence shows that
most particles introduced to the ACF system are “new”, i.e., derived from surface runoff and
topsoils. Particulate radium isotopes (226Ra and 228Ra) indicate that during low discharge the
Flint River is the dominant source of the particulate flux to Apalachicola Bay (70%), while the
Chattahoochee River contributes 30% and the Apalachicola River does not appear to contribute
locally-derived sediments. During high discharge, the Chattahoochee River contributes a
majority of the suspended particles (56%) while the Apalachicola River supplies 30% and the
Flint River only provides 14% of the particulate flux. Under higher flow conditions, these
particle fluxes contribute arsenic and antimony in quantities well above the EPA recommended
limit to the aquatic ecosystem and shell fisheries in Apalachicola Bay.
Introduction
Rivers are the largest flux of freshwater to the coastal ocean, and as such, considerable
attention has been given to the associated dissolved constituent input to estuaries and the coastal
zone (Kaul and Froelich, 1984; Hodell et al., 1990; Hope et al., 1994). However, much less
work has focused on the role that suspended sediments play in coastal constituent loading
(Milliman and Meade, 1983; Milliman and Syvitski, 1992). Most heavy metals, phosphorus,
pesticides, and many organic compounds are particle-reactive and tend to be transported downriver in particulate form (Horowitz, 1995; Bonniwell et al., 1999). Some of these components
can desorb upon entering an estuary due to ion exchange processes, thus becoming more
bioavailable (Froelich, 1988).
Most solutes and sediments in rivers are derived from the weathering of continental rock
in upper drainage basins (Martin and Meybeck, 1979; Gaillardet et al., 1999). The headwaters of
86
rivers provide more sediment per unit area than their respective lower reaches because of steeper
stream gradients, higher relief, and higher erosion rates (Robert, 2003). Once introduced to a
river, suspended particles are subject to further physical and chemical weathering as they are
transported downstream. Canfield (1997) provides a quantitative assessment of the different
climatological factors (e.g., temperature, precipitation, runoff) that influence the extent of
riverine weathering. He shows that higher discharge and longer river length lead to more
particulate weathering. The extent of weathering is also dependent upon the relative
susceptibility of different particles to chemical and physical breakdown, often a function of
mineral structure.
For small to medium sized rivers, temporal variations in suspended sediment content are
driven mainly by discharge. This is not generally true for larger rivers with extensive tributaries
and multiple drainage basins. Other factors, such as lithology, topographic relief, vegetative
cover, seasonality, and reservoir volume tend to control the variability in larger rivers (Meybeck
et al., 2003). The range in temporal variability of the suspended sediment concentration
diminishes with larger basin area and reservoir volume, whereas the range is higher for those
rivers fed by snowmelt and glaciers (Meybeck et al., 2003).
We present here an array of geochemical tracers to track suspended sediments through
the Apalachicola-Chattahoochee-Flint (ACF) River system in the southeastern United States.
We focus on the naturally-occurring radioisotopes 226Ra (Τ1/2: 1600 years), 228Ra (Τ1/2: 5.7
years), 210Pb (Τ1/2: 22.2 years), 7Be (Τ1/2: 53.4 days), and 40K (Τ1/2: 1.25 x 109 years) to examine
suspended particle transport. These may be either embedded in the mineral structure (e.g., 40K
and radium isotopes) or adsorbed onto the active surface of the particles in freshwater systems
(Ra isotopes, 210Pb, and 7Be) and are effective tracers of fluvial particle transport (Donoghue et
al., 1989; Bonniwell et al., 1999; Sommerfield et al., 1999; Corbett et al., 2004; Pfitzner et al.,
2004; Matisoff et al., 2005; Wilson et al., 2007) and sedimentary processes (Olley et al., 1993;
Sommerfield and Nittrouer, 1999; Salant et al., 2007). We also use selected metals (As, Sb, Zn)
as additional tracers of different lithologies and anthropogenic contamination sources, which can
alter the metal concentrations found on suspended sediments (Dupre et al., 1996; Councell et al.,
2004).
The long-lived radium isotopes (226Ra and 228Ra) are naturally-derived by decay from the
primordial radionuclides 238U and 232Th, respectively. The parent isotopes are always present in
continental rocks and soils, but their activities relative to one another vary depending on rock
type. For example, sandstones typically contain relatively high 232Th, whereas granitic rocks
contain higher 238U. In freshwater systems, radium is particle-reactive and will be mostly
adsorbed onto particles (Martin and Akber, 1999; Nozaki et al., 2001). Lead-210 is also found in
the 238U decay chain, but is a descendent of gaseous 222Rn that can escape to the atmosphere.
Lead-210 is ultimately deposited on the land surface via an atmospheric pathway through wet
and/or dry deposition. Beryllium-7, a cosmogenic radioisotope, is formed in the upper
atmosphere by cosmic ray spallation of nitrogen and oxygen atoms. It is also delivered to the
land surface by atmospheric deposition. Potassium-40 is a primordial radionuclide that is
incorporated into crustal rocks and sediments in K-bearing minerals. We measure 40K as a
surrogate for K-minerals.
The complexity of the ACF river system adds to the challenging nature of this project.
Water mass tracing is a relatively simple process, as discharge gauges are abundant along this
river network. Sediment tracing, however, is not nearly as straightforward due to the dynamic
cycle between suspended transport, deposition, and resuspension. The abundance of large
87
reservoirs along the ACF system further complicates this cycle by introducing areas of slower
water velocity promoting settling and deposition. Additionally, suspended sediments in a river
are known to integrate the different types of sediment sources throughout the river basin
(Walling et al., 1999). We hypothesize that the different lithologic and anthropogenic sources
within the three rivers of the ACF system flow path impact the nature of the suspended particles
in each river so that they acquire distinct chemical signatures. These different chemical
signatures should then permit quantitative determination of each river’s contribution to the total
sedimentary flux to the estuary. Our goal here is to establish a set of quantifiable particle tracers
that remain chemically inert during transit and permit us to provide source data for the three
rivers of the ACF system.
Study Area
Apalachicola Bay is a shallow estuary in northwest Florida separated from the Gulf of
Mexico by several barrier islands (Kofoed and Gorsline, 1963). The largest source of freshwater
to the bay is the Apalachicola River (Figure 7.1). The ACF river basin drains 58,000 km2 of
mostly agricultural lands in the Florida Panhandle, western Georgia, and eastern Alabama
(Bedosky, 1987). The Chattahoochee and Flint Rivers converge in Lake Seminole, along the
Florida-Georgia state line, and then flow together as the Apalachicola River to the Gulf of
Mexico. The 10-year average discharge for the Apalachicola River (1998-2008) is about 500
m3/s (USGS) making it one of the largest rivers in the United States east of the Mississippi River.
Combined, the Chattahoochee and Flint Rivers contribute 80% of the water flow to the
Apalachicola River (Elder et al., 1988), with the remainder derived from direct runoff,
precipitation, and groundwater discharge below Lake Seminole. During high discharge periods,
the Chattahoochee River dominates the water flux to the Apalachicola River, whereas during low
discharge, the Flint River contributes a greater portion due to an abundance of groundwater
springs along its flow path (Elder et al., 1988; Couch et al., 1996).
The average annual sediment load to Apalachicola Bay from the ACF system is roughly
1.5 million metric tons (Isphording, 1985). The discharge of the ACF system is highly
controlled, with 16 dams spread throughout the river basin, 13 of which are found along the
Chattahoochee River. Isphording (1985) showed that the construction of the dams along this
river system has not significantly altered the sediment flux to Apalachicola Bay, although
certainly the particle dynamics have been altered. For example, the reservoirs effectively
remove the coarser sands and silts from the river system, leaving only the finer fraction to make
it to the bay (Surratt et al., 2008). Schmidt and Wilcock (2008) showed that a river tends to be
particle-deficient downstream of a dam and therefore becomes net erosional through the channel
as a result. We believe that enhanced down cutting in the ACF system below dams may act to
make up for any particle deposition within the reservoirs.
Water discharge has been affected by the damming of the river system and has decreased
about 20% below pre-dam discharges (1930-1950; USGS). Historically, suspended sediment
concentration trends in the ACF system do not correlate with discharge patterns (Leitman et al.,
1984; Donoghue, 1988). This led Harrington (2001) to conclude that factors beyond discharge
and precipitation control the associated suspended sediment behavior throughout this river basin.
The abundance of dams along the rivers may promote sediment deposition in the reservoirs
88
throughout the year until the spring flood waters flush the recently deposited sediments
downstream (Rozengurt and Haydock, 1991).
Figure 7.1. Map of the Apalachicola-Chattahoochee-Flint River basin through Florida, Georgia,
and Alabama (insert). Each sample site is shown (black dots and alpha-numeric designations).
Major reservoirs (all man-made since 1948) along the river system are highlighted. The fall line
where the Piedmont geologic unit meets the coastal plain province is designated by the dashed
line.
The Apalachicola River itself, which flows through the Apalachicola National Forest, has
traditionally been considered a pristine environment. But a recent study has identified increasing
toxic metal concentrations in the shallow sediments of Apalachicola Bay (Harrington, 2001).
89
Apalachicola Bay supplies 90% of Florida’s oyster harvest, and 13% of that of the entire United
States, so with sediment concentrations of arsenic (As), copper (Cu), lead (Pb), iron (Fe) and
several other metals above EPA recommended limits, these findings have far-reaching
implications. Harrington (2001) thus called for future work to identify upstream sources of
metal contamination throughout the ACF system.
Two major coal-fired power plants (CFPPs) situated on the Chattahoochee River have
been previously studied for their input of some dissolved metals into nearby river waters. The
fly ash from these CFPPs is disposed of in retention ponds along the river. These ponds
discharge into the river releasing significant amounts of arsenic (As), selenium (Se), and
antimony (Sb), especially during periods of high rainfall (Lesley, 2002; Lesley and Froelich,
2003). These metals are enriched in coal ash as a result of combustion and are released to the
aqueous phase in ash ponds. In addition, some of these metals (Se and As) are volatile and
escape to the atmosphere during coal combustion and are then deposited locally within the
catchment area to be washed into the river.
Methods
Sampling strategy
We selected fifteen sampling sites throughout the ACF river system to optimize our
ability to characterize the chemical signature of each river (Figure 7.1). Our sample site naming
convention uses a two letter river designation (CR = Chattahoochee River; FR = Flint River; AR
= Apalachicola River; AB = Apalachicola Bay) followed by a numeric value representing the
relative downstream to upstream order of the sample location (Table 7.1). AR-2, for example, is
further upstream in the Apalachicola River than AR-1 (Table 7.1). Samples in the lower
Apalachicola River and Apalachicola Bay were collected within the Apalachicola National
Estuarine Research Reserve (ANERR). During each sampling campaign, we collected two
samples in Apalachicola Bay, two samples in the Apalachicola River, six samples in the
Chattahoochee River, and five samples in the Flint River.
These sampling sites were chosen based on a number of criteria. First, we wanted to
quantify the signature of the headwaters of each river (CR-6, FR-5, and AR-2) as well as the
end-member signature of each river before it flows into its respective receiving basin (CR-1, FR1, and AR-1). Second, we chose sampling sites to allow us to examine the change in suspended
particle concentrations and composition in selected reservoirs, so a sample just upstream of a
reservoir is paired with a sample just downstream of the respective dam (CR-4 to CR-3 for West
Point Lake; FR-4 to FR-3 for Lake Blackshear; CR-1 and FR-1 to AR-2 for Lake Seminole).
Note that the upstream sample sites were located far enough upstream from the reservoir to avoid
any potential lake effects from influencing the suspended particles. For example, AR-1 was
collected in the Apalachicola River above the uppermost extent of the salt wedge. Third, we
selected two sample sites to allow us to examine possible inputs from the CFPPs to the
particulate load in the Chattahoochee River. Sample CR-5 is just upstream of these CFPPs
whereas CR-4 is just downstream. Several of our sample sites along the Chattahoochee River
correspond to those used by Bedosky (1987) to facilitate comparison with previous samples
collected in this area. Note also that the large urban center of Atlanta, Georgia is located
90
between CR-5 and CR-6. Finally, we selected two sample sites in the river plume within
Apalachicola Bay to examine any marine effects on the particle composition.
Table 7.1. Description of each of our sampling sites. Reported distances upstream are relative to
the US 98 bridge at the mouth of the Apalachicola River. Included are the nearest USGS
gauging station identification numbers and their distance from our sampling sites. See Appendix
B.
Sample
Distance
Name
AB-1a
0
29.7000
84.9815
n/A
-
a
0.6
29.7280
84.9797
n/A
-
a
10.6
29.7783
85.0437
n/A
-
b
AR-2
170
30.7006
84.8572
02358000
(80 m upstream)
CR-1c
208
30.9737
85.0061
02343801
(37 km upstream)
CR-2c
291
31.6066
85.0580
CR-3c
484
32.9154
85.1921
02339500
(4 km upstream)
c
540
33.2777
85.1017
02338500
(200 m upstream)
c
581
33.4762
84.9004
02338000
(Same location)
c
688
33.9740
84.2644
02335450
(7 km downstream)
c
FR-1
222
30.9588
84.5604
02356000
(7 km downstream)
FR-2c
326
31.5409
84.1398
02352500
(7 km upstream)
c
362
31.7251
84.0192
02350512
(Same location)
c
414
32.1244
84.0129
02349605
(27 km upstream)
c
607
32.9892
84.5293
02344500
AB-2
AR-1
CR-4
CR-5
CR-6
FR-3
FR-4
FR-5
a
b
c
Upstream (km) Latitude (°N) Longitude (°W) USGS Gauge Proximity to Site
0234296910 (37 km upstream)
(40 km upstream)
-sample collected from a boat; Salinities for AR-1 and AR-2 were 18.6 and 12.1, respectively
in June 2006 and 8.8 and 0.1, respectively in February 2007
-sample collected from a floating dock
-sample collected from a boat ramp
We collected samples from each of these sites in June 2006 and February 2007. These
sampling periods were chosen to represent base flow and high flow conditions, respectively
(Figure 7.2). While June 2006 successfully represented base flow (average discharge: 190 m3/s),
91
the high discharge in February 2007 (average: 660 m3/s) was lower than previous years because
of drought conditions affecting the southeastern U.S. during these years. The 10-year average
(1998-2008) discharge of the Apalachicola River is about 500 m3/s, occasionally reaching 50006000 m3/s during the spring. Fresh water flows through the ACF system with eventual delivery
to Apalachicola Bay controlled by the Army Corps of Engineers. This water flow is the basis for
the contentious ‘water wars’ issue between Georgia, Alabama, and Florida.
Sample Collection
For each sample, 240 L of surface river water was pumped into four 60-L opaque plastic
barrels via a suction pump. We used a floating series of PVC pipes to extend the end of the
pump hose away from the shoreline by ~ 15 m. Thus, we were able to avoid collecting those
particles that may be resuspended along the shoreline by turbulent flow in the shallower waters.
By collecting surface waters, we presumably captured the finest, most mobile of the suspended
particles in the water column. We also collected 1 L samples at each location to be filtered
through preweighed 0.45 μm membrane filters to determine the total suspended sediment (TSS)
concentration.
The samples were immediately transported to the Joseph W. Jones Ecological Research
Center near Bainbridge, Georgia to separate out the suspended particles. We used a Sorvall SS-3
super-speed centrifuge with a KSB Continuous Flow System attached to an SS-34 rotor. This
system continuously pumps water into a distributor which evenly delivers water to each of eight
250 mL centrifuge tubes within the rotor. The displaced water from each tube is then discarded
from the system through the distributor. The particle separation efficiency of this system
depends on both the pumping rate and the centrifuge spin rate. Using a pumping rate of about 1
L/min and a spin rate of 18,000 rotations per minute (rpm), this system captured about 90% (by
weight) of the total TSS (as determined from independent calibration analyses). This spin rate
results in a relative centrifugal force (RFC) of 380,000 m/s2, or roughly 40,000 g’s. Using this
pumping rate, the water inside a 250 mL centrifuge tube has a residence time of 2 minutes. The
entire centrifuging process took about 4 hours for each sample.
Measurements
After centrifuging each sample, the sediments collected in all eight centrifuge tubes were
combined and dried in an oven (at 60°C) for 24 hours. In June 2006, the average sediment yield
was about 2.9 g and ranged from 0.8 to 12.2 g. In February 2007, the average yield was 2.6 g
and ranged between 0.9 and 6.8 g. Once dried, the collected particles were homogenized,
weighed, and packed into a plastic vial to be counted in a well-type germanium gamma
spectrometer. The nominal vial packing height for sediments to be measured in our well-type
detector is 33 mm. In cases when our collected sediment mass was insufficient to fill the vials to
this height, we corrected the detector efficiency for the effective geometry following Kim and
Burnett (1983).
92
Figure 7.2. Hydrograph of the Apalachicola River measured at Chattahoochee, Florida (below
Lake Seminole) during the study period. Sampling periods are shown by arrows. Data from
online USGS river discharge database (02358000 at Chattahoochee, Florida). Note that high
discharge is typically during winter-spring except during tropical storm events (e.g., July 2005 =
Hurricane Dennis).
Once packed into a vial, each sediment sample was sealed by a layer of epoxy to prevent
radon loss. Secular radioactive equilibrium between radium and its daughters occurs within
about three weeks, after which we measure each sample for 226Ra (via 214Pb and 214Bi γemissions at 295.0, 351.9, and 609.3 keV), 228Ra (via 228Ac at 338.4 and 911.2 keV), 210Pb (at
46.5 keV), 7Be (at 477.6 keV), and 40K (at 1460.1 keV). We calculate excess 210Pb (210PbXS) as
the difference between the total measured 210Pb activity and the supported 210Pb, assuming
equilibrium with 226Ra.
We also subjected 1 mg of each sample (in triplicate) to a hot nitric acid extraction. We
heated the sediment samples in disposable polyethylene tubes using ultra-clean 10% HNO3 to
80°C for one hour to remove the desorbable and labile metals into solution. Note that this
process extracts only the leachable, mobile metals on the suspended sediments, whereas the
radiometric analyses quantify the total activity of each. These acid leachates were analyzed for
As, Sb, Zn, and Ca (as an indicator of CaCO3 weathering) concentrations using an Agilent
7500cs Quadrupole Inductively Coupled Plasma Mass Spectrometer (ICP-MS). The ICP-MS
was operated under 1500W hot plasma conditions to achieve maximum ionization. The sample
matrix (2% HNO3) was prepared using MQ water and Optima® HNO3 acid. The standards were
gravimetrically prepared from a multi-element Spectrum® High Purity ICP-MS standard (lot #
426004). Concentration calibrations were performed before and after the sample runs covering
three orders of magnitude for the analytes. The measurement protocol followed a standardsample-standard bracketing scheme to account for instrument and background drift throughout
the analyses.
Sequential extractions were carried out on several samples to quantify the leaching
efficiency for each measured element. For those elements which were effectively removed via
93
this acid leach, but were not completely leached during the first attempt, we adjusted all
measurements to represent the total leachable concentration based on the sequential extraction
efficiencies. Zinc exhibited quantitative removal during the first extraction, so no correction was
made. Arsenic, Sb, and Ca yielded extraction efficiencies between 80 and 90%, so minor
corrections were made for these elements.
Organic carbon analyses were carried out after rinsing 5-7 mg of untreated sediment with
10% HCl and then H2O to remove the soluble inorganic-C fractions (e.g., CaCO3) and salts.
These clean, dry particles were then analyzed for organic-C and –N concentrations and 13C
isotopes with an Elemental Analyzer (NC2500) interfaced to a Finnigan Delta XP mass
spectrometer. Note that this method provides total organic-C and –N concentrations whereas the
ICP-MS method estimates only the leachable As, Sb, Zn, and soluble Ca concentrations.
Results
TSS Concentrations
Trends in TSS concentration through the river system (Figure 7.3) reveal patterns of
suspended particle transport. In nearly every case under both high and low discharge, TSS
concentrations drop markedly after passing through a reservoir, presumably due to longer
residence times per unit length and water volume within the reservoir, and lower water velocities
that allow more particle settling than in the open river channel. During both sampling periods,
an increase in TSS concentration is observed after the Chattahoochee River passes through the
urban center of Atlanta. The increase is more dramatic during higher discharge conditions in
February. One limitation is that by definition, suspensions are not uniform, and our measured
TSS concentrations may not be fully representative of the bulk river waters at these sites.
One somewhat surprising observation is that the TSS concentrations are generally higher
in June 2006 (12.6 ± 7.5 mg/L during base flow) than in February 2007 (7.3 ± 5.6 mg/L during
high flow). One would expect that the higher energy associated with faster water velocities in
February would inhibit particle settling compared to base flow conditions. However, since the
majority of land within the ACF basin is devoted to agriculture (Bedosky, 1987), we suspect that
the enhanced summertime TSS concentrations are related to agricultural activities. The growing
season brings not only tilling of fields which loosens the surface soils but also irrigation which
helps to promote erosion to rivers. Another possible explanation is that the upper reaches of the
ACF system are currently undergoing heavy development leading to enhanced erosion.
Nonetheless, despite many higher concentrations of TSS in June 2006, the flux of suspended
particles to Apalachicola Bay is higher in February 2007 (9600 g/s) than in June 2006 (1200 g/s)
because discharge is greater.
Long-lived nuclides (radium isotopes and 40K)
The radionuclide signature of particles being transported through a river system reveals
additional information about the suspended particle characteristics (Figure 7.4). Once the
suspended particles are introduced to saline waters, one would expect ionic exchange processes
to displace radium from the surface of particles and thus the particulate radium activities should
94
drop. However, no discernable radium activity decreases were observed within the saline waters
of Apalachicola Bay for the particles in either sampling (Figures 7.4a and 7.4b), so we conclude
that the majority of radium is contained within the mineral matrix rather than adsorbed onto
active surfaces.
The middle Flint River suspended particles tend to have somewhat higher radium
contents than those of the Chattahoochee River (Figures 7.4a and 7.4b). The drainage basins of
the Chattahoochee and Flint Rivers are delineated by the ‘fall line’, which separates the upper
basin consisting of metamorphic sediments of the Piedmont and Blue Ridge Provinces from the
lower basin of coastal plain sediments (Bedosky, 1987). This division is about 425 km upstream
from Apalachicola Bay on the Chattahoochee River and about 550 km upstream of the bay on
the Flint River (Figure 7.1). Both radium isotopes indicate an increase in activity of suspended
particles downstream of the fall line.
We measured 40K as a surrogate for total potassium and the mineral content of our
collected suspended sediments. Potassium-40 indicates lower lithogenic particulate
concentrations downstream in both the Chattahoochee and Flint Rivers (Figure 7.4c). This
decrease may be due not only to a dilution of lithogenic sediments with other types of non Kbearing suspended particles (organic material, carbonates, anthropogenic material), but also to
the continued chemical weathering of K out of minerals (Canfield, 1997; Gaillardet et al., 1999).
Radiotracers derived from atmospheric deposition (7Be and 210PbXS)
Short-lived 7Be (Figure 7.4d) can indicate the ‘newness’ of the suspended particles since
it decays with a half-life of only 53 days (Bonniwell et al., 1999; Matisoff et al., 2005). The
overall activity of 7Be in February 2007 (averaging 40.7 ± 27.3 dpm/g) is higher than in June
2006 (11.5 ± 11.3 dpm/g), resulting from higher erosional inputs of young sediment throughout
the ACF basin. Based on the 7Be concentration patterns, the headwaters of the Chattahoochee
River supply young sediment in February 2007, whereas in June 2006, no ‘new’ sediment is
added upstream. Instead, the June 2006 trends indicate gradual input of new sediment along the
entire course of each river.
Excess 210Pb is thought to represent the same erosional processes as 7Be. Some
particulate 210Pb can be produced from 226Ra decay through various daughters, but the major
source of 210Pb to soils is usually from atmospheric deposition. We only account for the 210Pb
that is derived from atmospheric 222Rn decay and neglect the supported component from 226Ra
decay on the particles by examining the 210PbXS activity, i.e., total 210Pb minus 226Ra. This
correction for supported 210Pb assumes that the particles within the river system are in secular
equilibrium with respect to 226Ra decay through gaseous 222Rn to 210Pb. This assumption likely
underestimates 210PbXS (overestimates 210Pbsupported) because some unknown portion of the
particulate 222Rn escapes.
The June 2006 and February 2007 particulate 210PbXS activities (Figure 7.4e) show the
same general behavior as 7Be. Several of the samples in June 2006 exhibited no 210PbXS after
accounting for the supported 210Pb. As with 7Be, the overall 210PbXS activities in February 2007
(averaging 10.0 ± 6.5 dpm/g) are significantly higher than in June 2006 (1.3 ± 3.9 dpm/g),
indicating more runoff-generated ‘new’ particle loading. Because the TSS concentrations in
February 2007 were lower than those in June 2006, we conclude that the suspended particles
throughout the river system in June 2006 are not derived from young, surface soils. This
observation supports our hypothesis that the suspended particles in June 2006 are related to
95
agriculture. Tilling of fields exhumes older buried soils to the surface, thus facilitating erosion
of older (non 7Be-bearing) soils to the river system.
Figure 7.3. Total suspended sediment (TSS) concentrations from June 2006 (solid circles) and
February 2007 (open squares) for the Chattahoochee, Flint, and Apalachicola River sample sites.
Gray bars indicate reservoir locations on each river: (1) Lake Seminole, (2) Walter F. George
Reservoir, (3) Lake Worth, (4) Lake Blackshear, (5) Lake Harding, (6) West Point Lake, and (7)
Lake Lanier. The fall line is designated by the horizontal gray dashed lines. The center of urban
Atlanta, Georgia is shown by the circled ‘A’, and the location of the CFPPs is indicated by the
asterisk (*).
Other stable tracers
During both sampling periods, organic-C concentrations (Figure 7.5a) tend to increase
downstream in the Chattahoochee and Flint Rivers, indicating enhanced biological productivity
in the downstream direction, especially below the fall line. The overall organic-C concentrations
in June 2006 (averaging 64.4 ± 35.9 mg/g) reach higher levels than those in February 2007 (47.6
± 12.2 mg/g), especially in the lower portions of the Chattahoochee River, presumably due to
96
more sunlight during the summer months fostering a better growth environment. Carbon-13
values (not reported here) are typical of terrestrial aquatic systems (δ13C averages -28.0 ± 1.87
‰) and show no significant longitudinal or seasonal trends. Organic C:N ratios range from 6.7
to 15.1 (atom:atom), averaging 9.5 ± 1.8.
We use calcium (Figure 7.5b) as a proxy to represent the particulate calcium carbonate
(limestone) in the river system. The geology of the coastal plain areas in this river system is
mainly composed of limestone, and the calcium concentration trends indicate more particulate
carbonate below the fall line in the Chattahoochee and Flint Rivers.
In a study of the dissolved metal concentrations along the Chattahoochee River, Lesley
(2002) identified areas just downstream of several CFPPs (between sample sites CR-4 and CR-5)
that exhibit elevated concentrations of dissolved and particulate As, Se, and Sb. Our particulate
samples exhibit much the same behavior from both June 2006 and February 2007 in As (Figure
7.5c) and Sb concentration (Figure 7.5d).
During June 2006, As concentration spikes in particles at the site immediately
downstream of the CFPPs, and the base flow discharge conditions support quick deposition of
these As-rich sediments before the next downstream sample site (Figure 7.5c). During February
2007, however, the faster water movement associated with the higher discharge maintains these
As-rich particles in suspension throughout most of the remaining Chattahoochee River length,
only dropping back down to levels found upstream of the CFPPs just prior to draining into Lake
Seminole. Interestingly, during June 2006, the suspended particles leaving Lake Seminole
(sample AR-2) show the highest measured As concentration. We believe that this may be related
to a subterranean pathway where Lake Seminole waters intrude the aquifer and bypass the dam,
discharging into a spring within the Apalachicola River. The surficial aquifer here is the
Hawthorn province, which contains elevated concentrations of As. When flushed with oxygenrich lake waters, the As may be released and transported to the Apalachicola River where we
have observed the highest particulate As concentrations throughout the river system (L. Torak,
pers. comm.).
The particulate behavior of Sb is more complex than As. During June 2006, Sb does not
peak in the Chattahoochee River suspended particles until sample CR-3 (Figure 7.5d), farther
downstream than the As concentration peak. During February 2007, Sb is again highest at
sample site CR-3. This concentration, however, is over 5-fold higher than that measured at the
same point in June 2006. We believe that the differences in downstream patterns between As
and Sb can be attributed to different adsorption/desorption characteristics of these two metals.
The reported distribution coefficient (Kd) value for As (3.2 mL/g; Wilhelm, 2004) is lower than
that of Sb (44.7 mL/g; Sheppard and Thibault, 1990), indicating that Sb is more strongly retained
on suspended particles than As. Our data are supported by this, as the downstream effects of As
are not as far reaching as those of Sb, likely because most of the As is desorbed from the
particles along the course of the river.
Councell et al. (2004) showed that automotive tire wear particles can contribute
significant Zn to rivers and lakes via runoff from urban areas. Our samples, however, do not
exhibit noticeable elevations in Zn concentration at sites around urban areas (Figure 7.5e). The
most notable urban center is Atlanta (between samples CR-5 and CR-6), but neither data set
shows a marked increase in particulate Zn concentration as the Chattahoochee River passes
through Atlanta. Instead, our samples show the highest concentrations at the lower Apalachicola
River (AR-1) in June 2006 and below the fall line in the Flint River in February 2007. However,
97
both of the areas surrounding and preceding these particular sample sites are rural, and thus not
likely to be contributing significant tire wear particles.
Figure 7.4. Particulate radionuclide activities from June 2006 (solid circles) and February 2007
(open squares) for the Chattahoochee, Flint, and Apalachicola River sample sites. Shown are
226
Ra (a), 228Ra (b), 40K (c), 7Be (d), and 210PbXS (e). Other symbols are the same as in Figure
7.3. Error bars represent 1σ measurement uncertainties, but often do not extend beyond the
symbol. See Appendix C.
Discussion
Reservoir sediment deposition
The change in TSS concentration as a river flows through a reservoir can be used to
calculate the rate of sediment deposition or erosion within that reservoir. If, for example, the
flux of suspended particles that enters a reservoir from upstream is greater than the flux leaving
the reservoir, the loss of particulate matter is due to net sediment deposition within the reservoir.
Note that these results are net fluxes, since some erosional inputs could be balanced by greater
depositional rates.
98
Figure 7.5. Particulate stable tracer concentrations from June 2006 (solid circles) and February
2007 (open squares) for the Chattahoochee, Flint, and Apalachicola River sample sites. Shown
are organic-C (a), Ca (b), As (c), Sb (d), and Zn (e). Other symbols are the same as in Figure
7.3. Error bars represent 1σ measurement uncertainties, but often do not extend beyond the
symbol. See Appendix C.
Based on our sampling scheme, we can examine the deposition/erosion rates in the
reservoirs of West Point Lake (between samples CR-4 and CR-3), Lake Blackshear (between
samples FR-4 and FR-3), and Lake Seminole (between samples FR-1, CR-1, and AR-2) (Table
7.2). West Point Lake and Lake Blackshear are net depositional reservoirs during both sampling
periods, but show somewhat lower deposition rates during higher discharge (0.59 vs. 0.67
g/m2·day in West Point Lake; 0.66 vs. 1.29 g/m2·day in Lake Blackshear). Lake Seminole,
however, shifts from slightly depositional during base flow (0.02 g/m2·day) to net erosional
during high flow (-0.11 g/m2·day). These reservoirs occasionally show different input water
fluxes than those leaving the reservoir, indicating that the Army Corps of Engineers was
filling/draining the lake at the time of sampling.
We perform a similar comparison for the river stretch that flows through the urban area
of Atlanta, Georgia (between CR-6 and CR-5). We cannot reliably estimate the area of this
stretch of river, however, so a quantitative comparison is not possible. Instead, we determine the
amount of particulate input to the river through Atlanta relative to that present above the city
(Table 7.3). We find that during base flow conditions, TSS concentrations only increase by
115% due to minimal runoff, whereas high flow conditions show a 33-fold increase in TSS
concentration through Atlanta.
99
TSS load constituents
Our sample collection method for suspended particles in river systems indiscriminately
captures all sediment types that make up the bulk load. We identify and separate major
components of this particle load using measured parameters of the suspended particle samples
(Figure 7.6). We use 40K to determine the lithogenic/crustal fraction, organic-C to estimate the
organic fraction, and Ca to determine how much carbonate (limestone) is contained in these
suspended particles. The remainder, then, should be made up of non K-bearing clay material,
un-leachable silicates, oxides and amorphous material, and any anthropogenic particle sources
(e.g., fly ash, tire wear particles, and other contaminants). We use 40K as an indicator of
lithogenic material instead of total K because measuring total-K via our acid-leaching method
would yield only the exchangeable fraction of K, and not an assessment of the lattice-bound K
concentration.
We take the natural crustal abundance of 40K (0.0117%) to calculate the lithogenic
fraction of the suspended particles. We convert our measured 40K values (Figure 7.4c) to total-K
via the natural abundance and assume a constant value for crustal K2O (2.8%; Rudnick and Gao,
2003) to convert our estimated total-K to a percentage of crustal material (fraction lithogenic;
Figure 7.6). This assessment does not account for any changes in K due to weathering of the
suspended particles. The lithogenic fractions of all our samples range between 4 and 60%. The
fractions of lithogenic material in our riverine particles are highest in those samples near the
headwaters of the Chattahoochee and Flint Rivers, above the fall line where the Piedmont meets
the coastal plain sediments.
The measured organic-C values (Figure 7.5a) are converted to mass of organic material
by assuming that all organics can be approximated by the formula CH2O, and therefore convert
from measured mass of C to assumed mass of dry organic material (x 2.5) (Figure 7.6). Our
organic fraction estimates range between 4 and 32% of the TSS mass. In general, during base
flow conditions, higher organics are found due to the more sluggish flow and more sunlight
during the summer (see Figure 7.5a). Also, both sample sets indicate higher organic fractions
downstream from reservoirs where the slower water velocities and stability promote algal
growth. Finally, we see vastly greater organic fractions at CR-4 and CR-5 compared to CR-6,
likely a result of runoff and sewage-related nutrient inputs from Atlanta.
Our measured values for Ca are used to determine an upper limit for the fraction of
calcium carbonate in our samples. We expect that as the rivers progress from the Piedmont
sediments to the lower coastal plain geologies, the carbonate fraction would increase. We
assume that all measured Ca in the extracts from the samples comes from CaCO3. However, all
estimated values (between 0 and 1%) are orders of magnitude lower than the accompanying
fractions of other components and will be ignored. Figure 7.6 shows the remaining fractions of
each TSS load as ‘uncharacterized’ components as described earlier.
Particle ages, fraction of ‘new’ sediments, and associated transit times
Particle Ages. Studies have shown 7Be to be a useful tracer for the relative age of
suspended sediments, as its relatively short half-life assures significant activity change over
normal transit time scales for riverine particles (Sommerfield et al., 1999). Since 210PbXS follows
similar atmospheric input and sedimentary behaviors as 7Be but with a much longer half-life, one
100
Table 7.2. Reservoir TSS deposition/erosion parameters for West Point Lake, Lake Blackshear,
and Lake Seminole for both June 2006 (top section) and February 2007 (bottom section). Net
TSS Fluxes are derived by subtracting the flux into the reservoir from the flux out, and negative
values indicate deposition within the reservoir. Discharge estimates and reservoir areas are taken
from U.S. Geological Survey and U.S. Army Corps of Engineers online databases.
Water Flux
(m3/s)
[TSS]
(g/m3)
West Point Lake
CR-4 (Input)
CR-3 (Output)
83
43
13.7
7.5
Lake Blackshear
FR-4 (Input)
FR-3 (Output)
26
25
Lake Seminole
FR-1 (Input)
CR-1 (Input)
AR-2 (Output)
76
95
155
TSS Flux Net TSS Flux
(g/s)
Area
(km2)
Deposition Rate
(g/m2 day)
1136
325
-811
105
0.67
29.6
10.5
776
266
-510
34
1.29
4.5
17.5
12.8
340
1662
1973
-28
152
0.02
(g/s)
JUNE 2006
FEBRUARY 2007
West Point Lake
CR-4 (Input)
CR-3 (Output)
64
64
17.9
6.7
1136
426
-710
105
0.59
Lake Blackshear
FR-4 (Input)
FR-3 (Output)
66
77
7.6
3.2
504
243
-261
34
0.66
Lake Seminole
FR-1 (Input)
CR-1 (Input)
AR-2 (Output)
180
559
650
4.4
7.0
7.5
788
3916
4905
201
152
-0.11
can account for various particle histories through the course of a river by normalizing 7Be to
210
PbXS activities and examining the 7Be/210PbXS ratio instead of the 7Be activity alone
(Bonniwell et al., 1999; Matisoff et al., 2005; Wilson et al., 2007). This normalization also
eliminates the temporal and spatial variability associated with either 7Be or 210PbXS atmospheric
deposition rates (Koch et al., 1996). In other words, the 7Be/210PbXS activity ratio (AR) can be
used to calculate the age of riverine particles, as higher ratios (less 7Be decay) would signify
younger particles within the river. By “age” here we refer to the time when the particles were
last exposed to atmospheric deposition.
101
Matisoff et al. (2005) provide an in-depth assessment of the various assumptions behind
using the 7Be/210PbXS ratio as a geochronometer for riverine particles. Of these, one of the most
notable is that atmospheric dry fallout inputs of these tracers to the soil surface must not be
significantly distinguishable from wet fallout events, or at least that any delay between dry
fallout and precipitation events is short relative to the half-life of 7Be. One also must assume that
once deposited to the soil surface, these tracers undergo rapid adsorption onto particle surfaces
and remain strongly retained on the particles throughout their life times. Inherent in this
assumption is that there can be no partitioning differences between 7Be and 210PbXS onto
different particle types.
Considering these assumptions, we can attribute a decrease in the 7Be/210PbXS ratio
(Figure 7.7) on suspended sediments to one of two possible end-member scenarios which we
examine separately (after Matisoff et al., 2005). This decrease in AR is due either to an aging of
the particles as they are transported downstream, in which case the 7Be would be decaying faster
than 210PbXS, or to dilution of the suspended sediments with other particles that are 7Be-deficient,
possibly from deeper soil horizons or resuspended river bottom material (Matisoff et al., 2005).
While the two processes can not be uniquely separated (aging versus mixing), they do provide a
conceptual framework for discussion.
For the initial case of freshly 7Be-tagged soils aging downstream, we must know the
signature of the atmospheric source. Since we did not measure this initial AR in our study area,
we use the AR measured in precipitation in nearby Weeks Bay, Alabama (17.2) as our initial
input ratio (Matisoff et al., 2005). We can thus estimate the average age of each of our bulk
suspended sediment samples from the time at which they were initially tagged with atmospheric
7
Be (termed a “particle age”) according to the following equation from Matisoff et al. (2005):
Particle age: t =
λ Be−7
1
−1
ln ( AR ) +
ln ( ARO )
− λ Pb − 210
λ Be−7 − λ Pb −210
(7.1)
where (AR) represents the measured 7Be/210PbXS activity ratio, (ARO) is the initial atmospheric
7
Be/210PbXS source activity ratio, and λ represents each isotope’s respective decay constant
(0.013 day-1 for 7Be and 8.51x10-5 day-1 for 210PbXS). Most soil particles have likely experienced
some delay between the time when they are initially tagged with the atmospheric source (at
which time they are considered to be ‘new’) and when they are eroded and transported, so their
input signature ratio to the river would be somewhat lower than that of the atmospheric
deposition.
We observe a rather constant particle age throughout the course of each river in February
2007 (Figure 7.8). Most samples from the June 2006 data set do not contain sufficient 7Be
activities to make the estimates. When comparing individual sample sites in the Chattahoochee
River between June 2006 and February 2007 we find the ages to be somewhat older in June
2006. This agrees with our hypothesis that most sediments input to the rivers in June 2006 are
derived from agricultural activities, as deeper soil horizons tilled to the surface would not have as
much 7Be. The relatively constant age downstream each river in the ACF system during
February 2007 suggests that wherever sediment inputs to the river do occur, the proportion of
new sediments to old sediments is relatively constant. Other factors, such as particle residence
time within the river as well as any delay between a precipitation event and a specific particle’s
input to the river, tend to complicate this interpretation.
102
Table 7.3. TSS input parameters for the river segment through Atlanta for both June 2006 (top
section) and February 2007 (bottom section). Discharge estimates are taken from U.S.
Geological Survey online databases.
Water Flux
(m3/s)
Atlanta
CR-6 (Input)
CR-5 (Output)
58
72
[TSS]
TSS Flux TSS Flux Change
3
(g/m )
(g/s)
(%)
JUNE 2006
7.9
13.7
459
985
+ 115
76
2613
+ 3317
FEBRUARY 2007
Atlanta
CR-6 (Input)
CR-5 (Output)
45
119
1.7
22.0
Fraction of ‘New’ Sediments. Alternatively, we can consider the input of 7Be-deficient
sediments to a river that lowers the overall 7Be/210PbXS ratio in the bulk particles and estimate the
percentage of ‘new’ sediments in our bulk samples using the following equation (Matisoff et al.,
2005):
% ‘new’ sediment = 100 x (AR)/(ARO)
(7.2)
While the decay-based age model responds exponentially to a change in the measured
7
Be/210PbXS ratio, the equation for the fraction of ‘new’ sediment shows a linear response to the
activity ratio (Figure 7.9).
In February 2007, during high runoff and significant soil loading to the river from the
surrounding basin, the downstream change in ‘new’ sediment fraction varies directly with the
change in TSS concentration (Figure 7.10a). This suggests that the majority of the particles in
the river are derived from the upper soil horizons since increasing sediment inputs are associated
with younger ages. However, in June 2006 when runoff was limited, the change in ‘new’
sediment fraction from one sample site to another exhibits a negative correlation with the change
in TSS concentration (Figure 7.10b). This suggests that any increase in particulate input to the
rivers corresponds to older sediments (7Be-deficient), presumably derived mostly from
resuspended bottom sediments. These particles may also be derived from deeper soil horizons in
the drainage basin after being exposed to the surface via agricultural tilling.
Transit Times. The results above are dependent upon an estimated atmospheric source
signature which we did not directly measure. As an alternative not dependent upon initial AR,
we can instead use measured ARs to assess the relative transit time between two sample points
along the river system by taking an upstream 7Be/210PbXS AR as the initial signature and
examining how that ratio changes downstream at the next sample site. Note that the sampled AR
quantitatively time-integrates all suspended sediment sources within the source area. Several
possibilities exist for the change in observed AR between sample sites: (1) the ratio increases,
103
which represents an unquantifiable net addition of ‘new’ riverine particles between the two
sample sites; (2) the AR decreases and the TSS concentration increases, representing net addition
of old, 7Be-deficient sediments to the river, perhaps from bottom sediment resuspension; or (3)
the ratio decreases, along with a constant or net decrease in TSS concentration, thus indicating an
aging of the suspended particle mass as it is transported downstream. Of these possibilities, only
the last one is useful for purposes of determining transit times without explicitly measuring the
two end-member source ratios as the other two situations would represent unquantifiable
additions of riverine particles. Additionally, sample sites where the ARs are within the analytical
uncertainty of zero are not used.
Figure 7.6. Calculated components of the TSS load from each sample site (total height of bar;
mg/L) contributed by lithogenic particles (gray section), organic particles (white section), and
uncharacterized particles (assumed to be composed of non K-bearing minerals and
anthropogenic particles). Each river segment’s samples are grouped for June 2006 (A) and
February 2007 (B). Carbonate contributions (<1%) are not shown.
We calculate the transit time for those sample pairs that show decreasing 7Be/210PbXS
activity ratios based on equation (7.1), where AR is the downstream activity ratio and ARO is the
upstream activity ratio. This calculation assumes that no erosion of reservoir or riverine
sediments adds particles to the water column during the transit time covered.
104
During base flow conditions, no sets of contiguous samples meet the qualifications
necessary to utilize this ratio to calculate transit times. However, three pairs of samples meet
these requirements in the February 2007 data set, and we examine each pair here separately.
Figure 7.7. Activity ratios of 7Be/210PbXS from June 2006 (solid circles) and February 2007
(open squares) for the Chattahoochee, Flint, and Apalachicola River sample sites. Other symbols
are the same as in Figure 7.3.
In the Chattahoochee River, the AR decreases and the TSS concentration drops by over a
factor of 2 as the particles pass through West Point Lake (samples CR-4 to CR-3; Figure 7.7).
This represents a net loss of suspended particles as well as an aging of 60 days while the particles
transit the reservoir (Table 7.4). As these sample sites are 56 linear km apart, this residence time
equates to a particle transport velocity of 1.1 cm/s (Table 7.4). For comparison, West Point Lake
holds 0.75x109 m3 of water. Dividing this volume by the measured water discharge during the
sampling time (142 m3/s) yields a water residence time of 152 days, and an average water
velocity of 0.4 cm/s. As this less than half of that velocity determined for the suspended
particles, we conclude that some ‘new’ sediment has been added to West Point Lake during
transit.
105
Figure 7.8. Calculated particle ages for the bulk suspended sediment samples (based on
Equation 7.1) from June 2006 (solid circles) and February 2007 (open squares). The samples
with no detectable excess 210Pb have been omitted. Other symbols are the same as in Figure 7.3.
In the Flint River, the AR decreases with an accompanying drop in TSS as the suspended
particles pass through Lake Blackshear (FR-4 to FR-3; Figure 7.7). This is a much less dramatic
decrease in the AR, equating to a transit time of only 5.2 days (Table 7.4). Lake Blackshear
holds a volume of approximately 7x106 m3 (Harrington, 2001), and dividing this by the measured
water discharge at the sampling time (82 m3/s) yields an estimated water residence time of 2.5
days through Lake Blackshear, roughly half as long as the suspended sediments. Since these
sample sites are about 52 river km apart, we find the water and sediment transport velocities
through this reservoir during high discharge are about 24 and 12 cm/s, respectively.
The last set of samples satisfying the required qualifications to use the AR to calculate
transit times are along a river segment in the Flint River (FR-2 to FR-1; Figure 7.7). The drop in
the ratio corresponds to a transit time of 9.6 days (Table 7.4). These sample sites are about 105
river km apart, yielding an estimated suspended particle velocity of about 13 cm/s. Thus, it
seems that during high flow, the suspended sediments are only transported slightly faster through
riverine segments than through Lake Blackshear, so we conclude that this reservoir was acting as
a ‘run of the river’ reservoir during the sampling period (Harrington, 2001).
106
Figure 7.9. Theoretical relationships between piston age of bulk suspended sediment sample
(Equation 7.1; solid line) and the fraction of ‘new’ sediment in a sample (Equation 7.2; dashed
line) relative to an assumed initial value of the measured 7Be/210PbXS to that in the atmospheric
source (from Matisoff et al., 2005).
Particle tracing
One major goal of this project was to determine the fraction of the total particle flux to
Apalachicola Bay that is contributed from each of the three major rivers in the ACF system. For
this, we must determine a suitable particle tracer that is chemically conservative through the river
system and also one that qualitatively characterizes the sediment nature of the various geologic
provinces through which the river system flows. As most of the stable tracers measured can
potentially be affected by biological activity and/or human contamination sources, it seems that
the best tracer should be selected from the natural radionuclides. In addition, the half-life of the
selected tracer must be sufficiently long as to not experience significant decay while residing
within the different reservoirs.
The most suitable tracers thus seem to be radium isotopes. The long half-lives of 226Ra
228
and Ra ensure that radioactive decay would be negligible as the particles move downstream.
To ensure that chemical behavioral changes are considered, we use the 228Ra/226Ra activity ratio.
By normalizing one isotope to another, we thus account for any small-scale adsorption or
desorption effects that radium may undergo throughout the freshwater portions of the ACF
system.
107
Figure 7.10. Change in the average ‘new’ sediment percentage (based on Equation 7.2) plotted
against the corresponding change in TSS concentration between each pair of two contiguous
sample sites for June 2006 (a) and February 2007 (b). The sample corresponding to Atlanta,
Georgia is shown by the circled ‘A’ and ignored for the linear regression. Note that the center
point of both lines is near (0,0), suggesting that the mixing model adequately reflects TSS
mixture of recent (high 7Be/210PbXS) and old (low 7Be/210PbXS) material.
Prior to directly examining the particles that enter Apalachicola Bay, we first determine
the mixing effects of the Chattahoochee and Flint Rivers in Lake Seminole. The particles that
leave Lake Seminole and enter the Apalachicola River (sample AR-2) carry radium isotope
activities that are a result of the mixing of two other river end-member radium activities (CR-1
and FR-1). We use a two end-member mixing model adapted from that presented by Dulaiova et
al. (2006) to determine the fraction of each river’s sediments to the particle load of the
Apalachicola River. This model allows us to examine the mixing of radium isotope ratios via the
individual components using the following equations:
f CR + f FR = 1
f CR ⋅ 228 RaCR + f FR ⋅ 228 Ra FR
=
f CR ⋅ 226 RaCR + f FR ⋅ 226 Ra FR
(7.3)
228
226
Ra AR
Ra AR
(7.4)
108
where fCR and fFR refer to the fractions of suspended sediment contributed from the
Chattahoochee River and the Flint River to the Apalachicola River, respectively. Radium
isotope activities for each river are the other variables in equation (7.5), and are derived from
samples CR-1, FR-1, and AR-2 (Figures 7.4a and 7.4b).
Table 7.4. Results of the 7Be/210Pb transit time analyses for those riverine segments in February
2007 that show decreases in the activity ratio as well as in the TSS concentration. Estimated
residence times are based on Equation (7.3) in the text and particle transport velocities are
converted based on the river length between sample sites.
7
River Segment
West Point Lake
CR-4 (Input)
CR-3 (Output)
Lake Blackshear
FR-4 (Input)
FR-3 (Output)
Flint River segment
FR-2 (Input)
FR-1 (Output)
Be/210Pb
Activity Ratio
TSS
mg/L
Residence
Time (days)
Estimated Particle
Transport Velocity (cm/s)
3.36
1.55
17.9
6.7
60
1.1
2.12
1.98
7.6
3.2
5.2
12
3.56
3.14
4.7
4.4
9.6
13
Solving these equations for fCR and fFR indicates that during base flow, the Flint River
contributes the majority of the suspended particle flux to the Apalachicola River (70%), whereas
the Chattahoochee River becomes the dominant source during high flow (80%). These results
are supported by the water fluxes, since the Flint River contributes a greater fraction of water to
the Apalachicola River during low flow periods (46%) than during high discharge (35%) (Elder
et al., 1988; Couch et al., 1996).
Once entering Apalachicola Bay, however, the radium is expected to display some
unconservative behavior because of the presence of saline water (Peterson et al., 2008b). We
thus cannot use this tracer to examine the TSS input to Apalachicola Bay. Instead, we examine
the TSS concentrations between the upper and lower samples in the Apalachicola River. In June
2006, there is a net loss of TSS along this river (net depositional; Figure 7.3), so an assessment
of any particle contribution from the Apalachicola River is not possible. Therefore, we assume
that the contribution of the Chattahoochee and Flint Rivers out of Lake Seminole represents the
same contribution to Apalachicola Bay (70% from the Flint River and 30% from the
Chattahoochee River), and thus the contribution from the Apalachicola River segment itself is
negligible.
During the high flow conditions in February 2007, however, a net increase in TSS
concentration of 30% is observed over the length of the Apalachicola River (Figure 7.3).
Therefore, the Apalachicola River contribution to the particle flux to Apalachicola Bay is
considered to be 30%, and the remaining contribution is divided between the Chattahoochee and
109
Flint Rivers by the same proportion in which they contributed to the upper reaches of the river
from Lake Seminole. Hence, during high flow, the Chattahoochee River is the source of 56% of
the suspended particles to the bay, whereas the Flint River only contributes 14%. These
estimates all assume that our calculated net fluxes of sediment within the Apalachicola River are
correct, and that more deposition is not balanced by enhanced erosion in other parts of the river.
Metal fluxes to Apalachicola Bay
Harrington (2001) reported a detectable increase in heavy metal concentration in the
shallow sediments of Apalachicola Bay above EPA recommended limits, and speculated about
the effects of this continuing trend on the oyster harvest within the bay. She concluded that
additional research is necessary to determine the source of the metal-bearing particles. Using our
data from Figure 7.5, as well as our particle tracing results, we can make an assessment of the
toxic metal fluxes with presumed CFPP sources (As and Sb) and their respective origins
upstream.
During June 2006, we take the As concentration on the suspended particles in the lower
portion of Apalachicola River (7.09 μg As/g sediment; AR-1) as the input signature to
Apalachicola Bay. Combining this with the TSS concentration of this sample (5.5 g/m3) and the
measured discharge for the Apalachicola River (217.2 m3/s), we calculate the particulate flux of
As to Apalachicola Bay to be 730 g/day (Table 7.5). As discussed above, 30% of the particle
flux is derived from the Chattahoochee River whereas 70% comes from the Flint River. Thus,
the Chattahoochee River contributes 220 g/day while the Flint River contributes 510 g/day of As
to Apalachicola Bay (Table 7.5). These fluxes are thought to be natural, as most of the CFPPderived particles are deposited in upstream reservoirs.
Making similar assessments for the flux during high flow conditions in February 2007
yields a flux of arsenic of 6800 g/day (Table 7.5). Of this, the Chattahoochee River contributes
3800 g/day, an increase of 17-fold over that contribution in June 2006. The Flint River
contributes only twice as much As during high flow (1000 g/day), and the Apalachicola River
now supplies some As-bearing particles (2000 g/day). It is important to note that these Asbearing particles may not be originally derived from the Apalachicola River, but may instead be
resuspended in Lake Seminole and then transferred through the Apalachicola River.
We determine the Sb flux to Apalachicola Bay in the same manner (Table 7.5). We find
that under base flow conditions, the flux of Sb to Apalachicola Bay is 44 g/day, whereas that
during high flow is 400 g/day. During February 2007, the Chattahoochee River shows a 17-fold
increase in Sb contribution flux over that in June 2006, while the Flint River shows only a 2-fold
increase in contributed flux to Apalachicola Bay (Table 7.5).
We use our As and Sb particulate loading fluxes to estimate the range in expected
sedimentary concentrations of these metals. For this, we assume an average sediment
accumulation rate of 7 mm/yr (Bedosky, 1987), an effective area of Apalachicola Bay of 243
km2 over which the Apalachicola River deposits its entire sediment load evenly (Gorsline, 1963),
and that no biogeochemical reactions act to alter the As and Sb concentrations in the sediments.
From this calculation, we determine that the As concentration in recent Apalachicola Bay
sediments should be between 0.16 μg/cm3 (based on the June 2006 flux) and 1.5 μg/cm3 (based
on the February 2007 flux). For reference, Harrington (2001) measured As sedimentary
concentrations in the bay ranging from 0.8 to 7.6 ppm (equivalent to 1.6 to 15.2 μg/cm3,
assuming a bulk density of 2 g/cm3), which were 240% higher than pre-industrial levels. For Sb,
110
we determine the range in sedimentary concentration to be between 9.4 ng/cm3 (based on the
June 2006 flux) and 86 ng/cm3 (based on the February 2007 flux). Harrington (2001) measured
Sb concentrations between 0.1 and 0.6 ppm (equivalent to 200 to 1200 ng/cm3) in these
sediments, and also found elevated enrichments in Sb concentration above the pre-industrial
levels. One would expect our estimates to be somewhat lower than direct measurements within
the deposition area of the river because our fluxes are assumed to be evenly distributed
throughout the bay, whereas Harrington (2001) sampled in areas directly affected by the river
plume.
Table 7.5. Estimation of the As and Sb heavy metal flux to Apalachicola Bay. The fluxes
determined from each of the rivers represent the portion of the total flux to Apalachicola Bay
contributed from each river segment.
Metal
Concentration
μg/g
[TSS]
g/m3
June 2006
February 2007
7.09
8.16
5.5
9.8
June 2006
February 2007
0.42
0.48
5.5
9.8
Apalachicola Bay
Discharge Particlulate Flux
m3/s
(g/day)
Arsenic (As)
217
730
983
6800
217
983
Antimony (Sb)
44
400
Chattahoochee
River
Flux (g/day)
Flint
River
Apalachicola
River
Flux (g/day) Flux (g/day)
220
3800
510
1000
0
2000
13
220
31
60
0
120
Conclusions
We show here various naturally-occurring radionuclide applications for examining
suspended particle transport behavior through a complex river system in the southeastern United
States. While the applied radionuclides are found in much lower concentration than other stable
chemical tracers, they are more useful here because of their conservative behavior through the
river system and their internal time clock in the form of radioactive decay. We list below our
main conclusions on riverine particle transport through the ACF system.
1. We find that West Point Lake along the upper reaches of the Chattahoochee River and
Lake Blackshear along the upper reaches of the Flint River are net depositional reservoirs
under both discharge regimes sampled, but Lake Seminole changes from net depositional
under base flow conditions to a net erosional reservoir at higher discharge. Under base
flow conditions, we find a 115% increase in TSS flux from Atlanta, whereas high
discharge increases the TSS flux through Atlanta by a factor of 33.
2. The lithogenic/crustal fraction in our suspended sediment samples ranged between 4 and
60%. We also determined the organic fraction (between 4 and 32%) and the portion of
carbonate in the samples (less than 1%). The remainder in each sample is attributed to
111
clays and other common minerals that do not contain significant amounts of K as well as
anthropogenic particles.
3. Most particle sources to the rivers under high runoff periods (i.e., February 2007) are
from basin surface-layer soils, whereas in June 2006, the rivers show much older particle
input signatures (presumably from resuspended bed and bank material as well as deeper
soil horizons). We find residence times of particles within reservoirs between 5 and 60
days, and open river particle transport velocities on the order of 13 cm/sec during high
discharge.
4. Under base flow conditions, the Flint River contributes the majority of the particles to the
bay (70%), and the Apalachicola River is net depositional, so its contribution to the
particle flux to Apalachicola Bay is negligible. Under higher discharge, the
Chattahoochee River contributes the highest portion of particles (56%), with the
Apalachicola River contributing 30% and the Flint River providing 14% of the suspended
particle flux to Apalachicola Bay. During base flow conditions, we find the Flint River
contributes a higher toxic metal flux than the other rivers, though these fluxes are still
relatively small and presumed to be the natural flux. During high discharge, however, the
Chattahoochee River contributes a much higher flux (17-fold) of As and Sb to
Apalachicola Bay.
112
CHAPTER 8
CONCLUSIONS OF THE DISSERTATION
In this dissertation, I have illustrated how useful natural radioisotopic tracers can be in
the coastal zone. The inherent time clock (radioactive decay) associated with these tracers can
reveal the temporal component of various environmental processes (and thus compute delivery
fluxes) whereas stable tracers can only reveal spatial variations. Within the dynamic coastal
zone, radium, radon, and other radiotracers have allowed us to estimate submarine groundwater
discharge fluxes (in China and Hawaii) as well as riverine dissolved (China) and particulate
(southeastern U.S.) mixing processes.
In Chapter 2, I presented a complete, detailed method description of three techniques
available to the field for measuring 226Ra on Mn-fiber via its gaseous daughter, 222Rn. I
quantified measurement parameters such as efficiency, uncertainty, and MDA for each of these
methods. These results will provide potential users with the information necessary to make an
informed decision for the most appropriate method to employ for one’s needs. While the RAD7
method exhibited a lower efficiency than both the RaDeCC and Rn line systems, such a system
can provide other benefits. In addition to being fully automated, the RAD7 system is also the
most portable, allowing a user to measure 226Ra (and 224Ra) while in the field or at sea. If one
were to allow 214Po ingrowth, which is done in the other methods, the RAD7 system efficiency
would be nearly doubled, providing lower measurement uncertainties than those reported here
for similar counting times.
During the wet season in the Yellow River basin (Chapter 3), I find higher river
discharge, a salinity gradient located offshore of the mouth of the river, and offshore transport
rates of around 1.6 cm/s based on the 224Ra/228Ra radium age distribution. Conversely, during
the dry season, I find much lower river discharge, a salinity gradient contained within the river
channel, and transport rates still around 1.4 cm/s. We determined coastal transport rates using
the activity ratio distribution of radium isotopes. For a dynamic system where advective mixing
is likely, one may examine the range of transport rates required to distribute the radium isotopes
through an estuary via the apparent radium ages of the samples. Since the “transport rates”
varied in a narrow range (1.4 – 1.6 cm/s) in spite of relatively large fluctuations in river
discharge (~80 – 600 m3/s), I conclude that tidal mixing must dominate in this system, at least
over the range of discharges investigated.
I also measured several groundwater tracers in an area ~ 40 km south of the Yellow River
estuary to quantify SGD rates from offshore transects and time-series analyses (Chapter 4).
Salinity and pH were also found to be useful tracers in this environment, showing increasing
values in the offshore direction. The gradient in apparent radium ages of the water masses with
distance from shore yields horizontal transport rates between 3.3 and 4.7 cm/s. I show that using
radium isotopes to assess SGD rates via a stationary time-series fashion is a valuable approach.
The results from the radium time-series were similar to those using an established 222Rn model,
and followed the patterns of SGD from seepage meter measurements. During September 2006,
the average SGD rates ranged from 4.5 to 13.9 cm/day, and the discharging water was composed
primarily of recirculated seawater. The SGD rates found during July 2007 averaged between 5.2
and 11.8 cm/day, and apparently had a larger fraction of terrestrial water. These fluxes and
113
patterns are somewhat lower than those from individual seepage meters deployed nearby but are
similar to average rates reported from seepage meters positioned in the same general area.
Applying these results to nitrate input to the Bohai Sea, I find that based on reported
transport rates of riverine and SGD-derived dissolved nitrogen species, even the most liberal
calculations do not suggest that these dissolved nutrients can be directly transported more than
53 km offshore before biological uptake (Chapter 5). While these calculations are admittedly
first-order approximations, they provide a foundation for future studies to better define this
process.
Along the leeward coast of the Big Island of Hawaii, groundwater inputs are often
focused as point-source discharges, creating buoyant plumes of groundwater mixing into the
coastal ocean. Offshore transects of 222Rn, 224Ra, and salinity from several of these plumes
reveal that these tracers are enriched (depleted for salinity) in the plume waters, but drop to open
ocean values at distances > 2000 m from the coastline. Radium isotope activities in the coastal
waters of Hawaii are very low, so in order to use them as effective tracers for the quantification
of SGD, a more rigorous sampling scheme within a few hundred meters of shore would be
necessary.
Time-series measurements of radon and salinity can be used to quantify SGD rates based
on a model designed for point-source discharges (Chapter 6). In every plume I studied along the
coastline, radon activities varied inversely with the tide. Salinity varied directly with the tides,
increasing on the flood tides and decreasing during ebb tides. These fluctuations are a result of
relatively high radon, low salinity SGD inputs fluctuations in response to the hydraulic gradient
variations between coastal aquifer levels and the ocean. I thus observed a pulsing input of cool,
lower salinity water into the coast with a 12-hour period. In general, the modeled SGD rates
indicate that maximum discharges occur during ebb/low tides and minimum discharges occur
during flood/high tides. Results from simultaneous mass balance modeling of water, salt, and
radon yield total SGD rates ranging from 1100 m3/day to 12,000 m3/day for the different plumes
studied. Setting modeled groundwater end-members to zero salinity results in estimates of the
freshwater component of the SGD that range from 630 m3/day to 8600 m3/day for these same
plumes.
In the Apalachicola-Chattahoochee-Flint River basin, I show various naturally-occurring
radionuclide applications for examining suspended particle transport behavior through a complex
river system. Using the change in TSS concentration as a river passes through a reservoir, I find
that West Point Lake along the upper reaches of the Chattahoochee River and Lake Blackshear
along the upper reaches of the Flint River are net depositional reservoirs under both discharge
regimes sampled, but Lake Seminole changes from net depositional under base flow conditions
to a net erosional reservoir at higher discharge. Employing the long-lived 40K isotope, I calculate
the lithogenic/crustal fraction in each of our suspended sediment samples to range between 4 and
60% by mass.
Examining the change in 7Be/210Pb ratio under certain conditions, I calculate a transit
time necessary for suspended particles to move from one sample site to the next downstream site.
I find residence times of particles within reservoirs between 5 and 60 days, and open river
transport velocities on the order of 13 cm/sec during high discharge. I also find that the
downstream change in ‘new’ sediment percentage between contiguous sample sites varies
directly with the change in TSS concentration during high discharge (indicating surface soil
input) and indirectly with the change in TSS concentration during base flow (indicating
resuspension of bottom sediments and deeper soil horizon inputs, presumably associated with
114
agricultural practices). Utilizing the particulate 228Ra and 226Ra activities, I trace the river origin
of the suspended particle flux to Apalachicola Bay. Under base flow conditions, the Flint River
contributes the majority of the particles to the bay (70%), while the contribution from the
Apalachicola River is negligible. Under higher discharge, the Chattahoochee River contributes
the highest portion of particles (56%), with the Apalachicola River contributing 30% and the
Flint River providing 14% of the suspended particle flux to Apalachicola Bay. Fom these
results, I estimate both the total toxic metal flux of As and Sb to Apalachicola Bay via suspended
particles as well as that contributed from each river segment. During base flow conditions, I find
the Flint River contributes a higher toxic metal flux than the other rivers, though these fluxes are
still relatively small. During high discharge, however, the Chattahoochee River contributes a
much higher flux (17-fold) of As and Sb to Apalachicola Bay.
115
APPENDIX A
SIMULTANEOUS RADON BOX MODEL EQUATION DERIVATIONS
116
Basic Premise: Setup Mass balance equations for time scale of an entire tidal cycle to calculate
QIN, QOUT, and QSGD
Then, go to 1-hr intervals to calculate ΔQIN, ΔQOUT, and ΔQSGD
Water Balance Equation:
ΔVP
= QOUT − QIN − QSGD
Δt
so QIN + QSGD +
ΔVP
= QOUT
Δt
Salt Balance Equation:
lw( S P (t +1) ρ P (t +1)d P (t +1) − S P (t ) ρ P (t )d P (t ) )
= S P ρ PQOUT − SO ρOQIN − S SGD ρ SGDQSGD
Δt
lw( S P (t +1) ρ P (t +1)d P (t +1) − S P (t ) ρ P (t )d P (t ) )
ΔVP ⎞
⎛
= S P ρ P ⎜ QIN + QSGD +
⎟ − SO ρOQIN − S SGD ρ SGDQSGD
Δt ⎠
Δt
⎝
lw( S P (t +1) ρ P (t +1)d P (t +1) − S P (t ) ρ P (t )d P (t ) )
ΔVP
= QIN (S P ρ P − SO ρO ) + QSGD (S P ρ P − S SGD ρ SGD ) + S P ρ P
Δt
Δt
lw( S P (t +1) ρ P (t +1)d P (t +1) − S P (t ) ρ P (t )d P (t ) )
ΔVP
− QSGD (S P ρ P − S SGD ρ SGD )
− SP ρP
Δt
Δt
QIN =
(S P ρ P − S O ρ O )
Radon Balance Equation:
lw( RnP (t +1)d P (t +1) − RnP (t )d P (t ) )
= RnPQOUT − RnOQIN − RnSGDQSGD
Δt
lw( RnP (t +1)d P (t +1) − RnP (t )d P (t ) )
ΔVP ⎞
⎛
= RnP ⎜ QIN + QSGD +
⎟ − RnOQIN − RnSGDQSGD
Δt ⎠
Δt
⎝
lw( RnP (t +1)d P (t +1) − RnP (t )d P (t ) )
ΔVP
= QIN (RnP − RnO ) + QSGD (RnP − RnSGD ) + RnP
Δt
Δt
lw( S P (t +1) ρ P (t +1) d P (t +1) − S P (t ) ρ P (t ) d P (t ) )
(Rn P − RnO )
lw( Rn P (t +1) d P (t +1) − Rn P (t ) d P (t ) )
Δ
t
−
=
(S P ρ P − S O ρ O )
Δt
ΔVP
(RnP − RnO ) Q (S ρ − S ρ )(Rn − Rn )
ΔVP
Δt
SGD SGD
P
O
− SGD P P
+ QSGD (Rn P − Rn SGD ) + Rn P
(S P ρ P − S O ρ O )
(S P ρ P − S O ρ O )
Δt
SP ρP
lw( RnP (t +1)d P (t +1) − RnP (t )d c (t ) )
Δt
(RnP − RnO ) ⎡ lw( S P (t +1) ρ P (t +1)d P (t +1) − S P (t ) ρ P (t )d P (t ) ) − S ρ ΔVP ⎤
ΔVP
−
P P
(S P ρ P − SO ρO ) ⎢⎣
Δt
Δt
Δt ⎥⎦
= QSGD
(RnP − RnSGD ) − (RnP − RnO )(S P ρ P − S SGD ρ SGD )
(S P ρ P − S O ρ O )
− RnP
117
On an hourly time scale, no more steady-state – we solve for the above constants, then use the following equations:
Water Balance Equation:
ΔVP
= (QOUT + ΔQOUT ) − (QIN + ΔQIN ) − (QSGD + ΔQSGD )
Δt
ΔVP
So, ΔQOUT =
+ QIN + ΔQIN + QSGD + ΔQSGD − QOUT
Δt
Salt Balance Equation:
lw(S P (t +1)ρ P (t +1)d P (t +1) − S P (t ) ρ P (t )d P (t ) )
= S P ρ P (QOUT + ΔQOUT ) − SO ρO (QIN − ΔQIN ) − S SGD ρ SGD (QSGD + ΔQSGD )
Δt
lw(S P (t +1)ρ P (t +1)d P (t +1) − S P (t ) ρ P (t )d P (t ) )
= S P ρ PQOUT + S P ρ P ΔQOUT − SO ρOQIN + SO ρO ΔQIN − S SGD ρ SGDQSGD − S SGD ρ SGD ΔQSGD
Δt
lw(S P (t +1) ρ P (t +1) d P (t +1) − S P (t ) ρ P (t ) d P (t ) )
= − S O ρ O QIN + S O ρ O ΔQIN − S SGD ρ SGD QSGD − S SGD ρ SGD ΔQSGD − S P ρ P QOUT + S P ρ P QIN + S P ρ P ΔQIN + S P ρ P QSGD
Δt
ΔVP
+ S P ρ P ΔQSGD + S P ρ P QOUT + S P ρ P
Δt
lw(S P (t +1)ρ P (t +1)d P (t +1) − S P (t ) ρ P (t )d P (t ) )
ΔVP
= QIN (S P ρ P − SO ρO ) + ΔQIN (S P ρ P + SO ρO ) + QSGD (S P ρ P − S SGD ρ SGD ) + ΔQSGD (S P ρ P − S SGD ρ SGD ) + S P ρ P
Δt
Δt
lw(S P (t +1) ρ P (t +1)d P (t +1) − S P (t ) ρ P (t )d P (t ) )
ΔVP
− SP ρP
− QIN (S P ρ P − SO ρO ) − QSGD (S P ρ P − S SGD ρ SGD ) − ΔQSGD (S P ρ P − S SGD ρ SGD )
Δt
Δt
= ΔQIN
(S P ρ P + S O ρ O )
Radon Balance Equation:
lw(RnP (t +1)d P (t +1) − RnP (t )d P (t ) )
= RnP (QOUT + ΔQOUT ) − RnO (QIN + ΔQIN ) − RnSGD (QSGD − ΔQSGD )
Δt
lw(RnP (t +1)d P (t +1) − RnP (t )d P (t ) )
= RnPQOUT + RnP ΔQOUT − RnOQIN − RnO ΔQIN − RnSGDQSGD + RnSGD ΔQSGD
Δt
118
lw(RnP (t +1)d P (t +1) − RnP (t )d P (t ) )
= − RnOQIN − RnO ΔQIN − RnSGDQSGD + RnSGD ΔQSGD − RnPQOUT + RnPQIN + RnP ΔQIN + RnPQSGD
Δt
ΔVP
+ RnP ΔQSGD + RnPQOUT + RnP
Δt
lw(RnP (t +1)d P (t +1) − RnP (t )d P (t ) )
ΔVP
= QIN (RnP − RnO ) + ΔQIN (RnP − RnO ) + QSGD (RnP − RnSGD ) + ΔQSGD (RnP + RnSGD ) + RnP
Δt
Δt
lw(S P (t +1) ρ P (t +1)d P (t +1) − S P (t ) ρ P (t )d P (t ) )(RnP − RnO )
lw(RnP (t +1)d P (t +1) − RnP (t )d P (t ) )
Δt
−
= QIN (RnP − RnO ) + QSGD (RnP − RnSGD ) + ΔQSGD (RnP + RnSGD ) + RnP
ΔVP
+
Δt
Δt
(S P ρ P + S O ρ O )
ΔV P
Δt − QIN (S P ρ P − SO ρ O )(RnP − RnO ) − QSGD (S P ρ P − S SGD ρ SGD )(RnP − RnO ) − ΔQSGD (S P ρ P − S SGD ρ SGD )(RnP − RnO )
(S P ρ P + S O ρ O )
(S P ρ P + S O ρ O )
(S P ρ P + SO ρO )
(S P ρ P + SO ρ O )
S P ρ P (RnP − RnO )
lw(Rn P (t +1) d P (t +1) − Rn P (t ) d P (t ) )
Δt
lw(S P (t +1) ρ P (t +1) d P (t +1) − S P (t ) ρ P (t ) d P (t ) ) ⎤
⎡
(Rn P − RnO ) ⎢S P ρ P ΔV P −
ΔV P
+⎥
− Rn P
− Q IN (Rn P − RnO ) − Q SGD (Rn P − Rn SGD ) +
⎥
(S P ρ P + S O ρ O ) ⎢Q (S Δρ t − S ρ ) + Q (S ρΔt − S ρ )
Δt
⎥⎦
⎢⎣ IN P P
O O
SGD
P P
SGD SGD
= ΔQ SG
(Rn P − RnO )(S P ρ P − S SGD ρ SGD )
Rn SGD + Rn P −
(S P ρ P + S O ρ O )
119
APPENDIX B
DETAILED SAMPLE COLLECTION SITE DETAILS ON THE ACF RIVER SYSTEM
120
AB-1: In Apalachicola Bay, this sample was collected 600 m due west of the #11 channel
marker. This site is approximately 2.6 km SSW from the U.S. 98 bridge that crosses the
Apalachicola River.
AB-2: This sample was collected at the #2 channel marker, approximately 500 m north of the
U.S. 98 bridge.
AR-1: We collected this sample about 500 m past the ‘pin-hook’ in the Apalachicola River.
This is just upstream from the Apalachicola Northern Railroad tracks that cross the lower reaches
of the Apalachicola River.
AR-2: This sample was collected in Chattahoochee, Florida on the east side of the river. From
U.S. 90, turn south onto S. River Landing Rd. (approximately 900 m east of the river). This road
proceeds through a neighborhood for about 800 m before opening into a public boat ramp. We
sampled from the floating dock to the south of the boat ramp.
CR-1: We collected this sample where SR 2 crosses over the Chattahoochee River. Just west of
the river, turn south into an access road to a public boat ramp. We sampled at the bottom of the
boat ramp.
CR-2: This sample site is located on the west side of the Chattahoochee River from Fort Gaines,
Georgia, where HWY 10 crosses over the river. About 600 m west of the river on HWY 10, turn
north onto CR-202, then make another immediate right into a public boat ramp. We sampled at
the bottom of the boat ramp.
CR-3: We collected this sample on the western side of the Chattahoochee River in West Point,
Georgia. From GA 18 west on the west side of the river, turn west onto W 9th St., just past the
Interstate Telephone Company. Take this to 3rd Ave, which is also CR 212 / S State Line Road
and turn north onto the road. Follow this road 4.7 km to a public park to the east. There is river
access here from a set of stairs that lead to the river’s edge.
CR-4: In Franklin, Georgia, we accessed this sample site on the west side of the Chattahoochee
River where U.S. 27 crosses the river. Several hundred meters from the river, an access road to
the south of U.S. 27 leads to a public park with softball fields and a boat ramp. We collected this
sample from the bottom of this boat ramp.
CR-5: This sample was collected near Whitesburg, Georgia where GA 16 crosses the river.
Roughly 50 m to the southeast of the bridge, turn northeast into a public boat ramp. We sampled
from the bottom of the boat ramp.
CR-6: On the eastern side of Atlanta, we sampled from a public boat launch. Just north of the
Holcomb Bridge Road river crossing, this public park is located on the western side of the road.
We sampled at the bottom of this boat launch.
FR-1: This sample site is located on the west side of the Flint River, north of Bainbridge,
Georgia. From GA 253, turn east onto Flint River Heights Rd. A public boat ramp is found at
121
the end of this road, after bearing to the south at an intersection. We sampled from the bottom of
this boat ramp.
FR-2: On the southern side of Albany, Georgia, we collected this sample on the east side of the
Flint River. About 900 m east of the bridge where GA 234 crosses the Flint River, turn south
onto Dixie Highway. After about 1.3 km, turn west onto a dirt road marked with a boat ramp
sign. At the end of this road, we sampled from the river bank on the side of the boat ramp.
FR-3: About 14.8 km east of Leesburg, Georgia, we collected this sample on the western side of
the Flint River where GA 32 crosses the river. About 300 m west of the Flint River, there is a
dirt road to the south of GA 32 that leads to a boat ramp. We sampled from the bottom of this
boat ramp.
FR-4: This sample site is located east of Americus, Georgia, on the western side of the Flint
River where GA 27 crosses the river. Just on the western side of the river on GA 32, turn north
onto Pool Gin House Road. Follow this road until it turns sharply to the left (after 8.5 km), then
turn right onto a red clay road (Reeves Road). At the end of this road, we sampled from the
bottom of the boat ramp.
FR-5: About 25.7 km west of Thomaston, Georgia, we collected this sample on the western side
of the Flint River as GA 18 crosses the river. Just west of the river, turn south onto a dirt road
that leads to a boat ramp. The boat ramp, however, is located on a side arm of the Flint, so we
collected the sample from the edge of the Flint River under the big tree at the end of the dirt
road.
122
APPENDIX C
RAW DATA FROM THE APALACHICOLA-CHATTAHOOCHEE-FLINT RIVER
SYSTEM
123
Table C.1. Water quality characteristics from the June 2006 sampling of the ACF system. ‘nd’
indicates that no data exists for that particular measurement.
Distance Upstream From
Specific
In-situ
Sample Site Apalachicola Bay (km) Conductance (μS/cm) Temperature (°C)
AB-1
-0.6
30,060
29.6
AB-2
0.0
4,050
30.3
pH
nd
nd
TSS
(mg/L)
23.96
9.86
AR-1
AR-2
10.6
169.9
0.169
0.135
30.5
27.8
nd
nd
5.51
12.76
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
208.0
290.9
484.0
540.0
580.7
688.2
0.126
nd
nd
nd
nd
nd
28
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
17.54
5.20
7.50
13.70
13.70
7.90
FR-1
FR-2
FR-3
FR-4
FR-5
222.2
326.3
362.2
414.0
607.1
0.183
0.15
0.05
0.032
nd
27.2
25.5
29.1
28.4
26.4
nd
nd
nd
nd
nd
4.50
5.90
10.49
29.62
20.50
124
Table C.2. Dissolved radium isotope activity concentrations during the June 2006 ACF
sampling. ‘BD’ indicates that a measurement was below detection. Measurement uncertainties
are reported at the 1-σ level.
Sample Site
AB-1
AB-2
Ra-223
dpm/100L
3.18 ± 0.27
0.67 ± 0.11
Ra-224
dpm/100L
29.82 ± 2.00
3.81 ± 0.33
Ra-226
dpm/100L
24.86 ± 0.64
9.07 ± 0.53
Ra-228
dpm/100L
51.77 ± 1.60
12.15 ± 1.27
AR-1
AR-2
0.16 ± 0.05
0.05 ± 0.03
1.08 ± 0.11
1.13 ± 0.11
2.74 ± 0.47
0.35 ± 0.55
3.57 ± 1.07
4.52 ± 1.77
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
0.11 ± 0.04
0.05 ± 0.03
BD
0.09 ± 0.04
0.10 ± 0.04
0.03 ± 0.02
1.94 ± 0.22
0.94 ± 0.11
0.98 ± 0.17
2.02 ± 0.17
1.44 ± 0.18
0.96 ± 0.10
3.07 ± 0.56
1.07 ± 0.61
1.03 ± 0.57
1.18 ± 0.60
BD
BD
2.61 ± 1.79
0.84 ± 2.20
1.41 ± 1.93
3.13 ± 1.49
0.80 ± 1.52
1.32 ± 1.81
FR-1
FR-2
FR-3
FR-4
FR-5
0.09 ± 0.05
0.10 ± 0.05
0.05 ± 0.04
0.25 ± 0.07
0.32 ± 0.09
1.64 ± 0.15
2.51 ± 0.29
1.98 ± 0.17
3.34 ± 0.24
4.09 ± 0.33
2.03 ± 0.54
3.19 ± 0.55
2.08 ± 0.67
8.02 ± 0.60
2.14 ± 0.63
5.40 ± 1.58
2.29 ± 1.62
1.41 ± 1.70
4.41 ± 1.39
1.81 ± 2.28
125
126
Ra-228
dpm/g
10.66 ± 1.44
3.62 ± 1.24
3.26 ± 0.93
3.17 ± 1.54
5.46 ± 1.09
7.64 ± 1.59
4.57 ± 1.33
4.26 ± 1.37
3.32 ± 1.26
2.76 ± 0.73
3.29 ± 2.00
6.82 ± 1.55
3.69 ± 0.87
7.26 ± 1.03
5.63 ± 0.98
Ra-226
dpm/g
4.56 ± 0.72
2.94 ± 0.68
3.17 ± 0.47
3.07 ± 0.70
2.67 ± 0.63
4.77 ± 0.79
3.11 ± 0.65
3.49 ± 0.67
3.19 ± 0.57
2.31 ± 0.38
4.31 ± 1.03
6.33 ± 0.80
4.22 ± 0.49
7.39± 0.57
5.36 ± 0.51
Sample Site
AB-1
AB-2
AR-1
AR-2
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
FR-1
FR-2
FR-3
FR-4
FR-5
6.72 ± 3.64
5.19 ± 3.37
5.23 ± 1.84
1.32 ± 2.27
4.12 ± 1.85
1.80 ± 2.48
8.47 ± 3.68
14.93 ± 2.28
7.56 ± 2.94
6.85 ± 2.43
1.19 ± 1.49
3.81 ± 1.86
3.38 ± 2.50
Pb-210
dpm/g
4.24 ± 2.49
5.28 ± 2.32
16.20 ± 7.96
0.34 ± 7.68
19.71 ± 5.35
BD
3.24 ± 6.39
2.95 ± 6.61
27.52 ± 9.56
9.68 ± 8.28
BD
6.98 ± 8.68
5.33 ± 5.66
18.01 ± 5.37
13.68 ± 13.21
Be-7
dpm/g
9.65 ± 8.51
39.51 ± 9.09
0.64 ± 0.44
0.43 ± 0.34
0.67 ± 0.23
0.18 ± 0.27
0.10 ± 0.22
0.62 ± 0.27
1.54 ± 0.33
0.58 ± 0.29
1.20 ± 0.32
0.05 ± 0.23
0.03 ± 0.19
0.15 ± 0.21
0.35 ± 0.30
Cs-137
dpm/g
0.20 ± 0.30
0.38 ± 0.33
1.86 ± 4.14
15.68 ± 3.42
7.22 ± 2.19
17.93 ± 2.41
15.38 ± 2.40
11.97 ± 2.92
9.53 ± 3.46
16.45 ± 3.00
25.50 ± 3.23
25.61 ± 2.67
24.72 ± 1.96
11.56 ± 2.21
9.56 ± 3.21
K-40
dpm/g
8.53 ± 3.28
13.77 ± 2.80
Table C.3. Particulate radionuclide analyses during June 2006 in the ACF sampling. ‘BD’ indicates that a measurement was
below detection. Measurement uncertainties are reported at the 1-σ level.
127
1.07E+00 ± 7.49E-02
1.28E+00 ± 4.96E-02
1.41E+00 ± 7.46E-02
5.87E-01 ± 1.41E-01
4.03E+00 ± 3.04E-01
1.21E-03 ± 6.11E-05
6.79E-04 ± 3.82E-05
8.49E-04 ± 7.88E-05
8.86E-05 ± 6.54E-06
1.30E-03 ± 4.78E-05
FR-1
FR-2
FR-3
FR-4
FR-5
5.54E+00 ± 1.02E-01
4.35E+00 ± 1.28E-01
9.14E+00 ± 2.58E-01
2.96E+00 ± 8.94E-01
BD
2.97E+00 ± 9.90E-02
2.41E+00 ± 6.79E-02
1.70E+00 ± 6.21E-02
8.15E-04 ± 3.29E-04
1.91E+00 ± 2.32E-01
3.25E-02 ± 3.42E-04
2.51E-02 ± 5.09E-04
3.95E-02 ± 4.61E-04
1.09E-02 ± 2.69E-04
5.99E-02 ± 8.42E-04
2.93E-02 ± 4.22E-04
4.61E-02 ± 8.81E-04
6.38E-02 ± 9.82E-04
5.39E-02 ± 5.97E-04
4.98E-02 ± 5.95E-04
4.21E-02 ± 7.49E-04
1.52E+00 ± 1.21E-01
2.52E+00 ± 1.36E-01
2.95E+00 ± 1.33E-01
4.12E+00 ± 3.31E-01
4.74E+00 ± 3.20E-01
4.58E+00 ± 1.60E-01
6.22E-04 ± 5.05E-05
7.24E-04 ± 4.53E-05
7.03E-04 ± 4.96E-05
8.58E-04 ± 3.87E-05
8.86E-04 ± 6.07E-05
6.74E-04 ± 5.44E-05
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
BD
2.05E+00 ± 5.63E-02
1.52E+00 ± 5.61E-02
8.95E-01 ± 1.73E-01
9.51E-01 ± 1.85E-01
4.26E-01 ± 4.36E-02
2.99E+00 ± 8.67E-02
2.09E+00 ± 5.04E-02 4.85E-02 ± 4.69E-04
BD
3.06E+00 ± 8.37E-02 5.17E+00 ± 1.33E-01 2.41E+00 ± 5.06E-02 4.55E-02 ± 3.92E-04
1.40E-03 ± 1.15E-04
1.13E-03 ± 8.02E-05
AR-1
AR-2
9.14E+00 ± 2.65E-01
9.03E+00 ± 2.85E-01
1.35E+01 ± 2.77E-01
2.17E+01 ± 3.35E+00
2.50E+01 ± 3.25E+00
1.36E+01 ± 2.85E-01
Al
Ca
V
mg metal / g of sediment
5.80E+00 ± 9.39E-02 9.75E+00 ± 1.54E-01 1.13E+00 ± 2.95E-02 4.76E-02 ± 7.72E-04
6.22E+00 ± 2.40E-01 2.21E+01 ± 2.19E+00 1.01E+00 ± 6.65E-02 5.69E-02 ± 8.14E-04
Mg
Be
Sample Site
1.00E-03 ± 6.28E-05
AB-1
1.27E-03 ± 4.66E-05
AB-2
2.50E-02 ± 3.98E-03
1.38E-02 ± 1.81E-03
2.07E-02 ± 4.03E-03
6.28E-04 ± 2.38E-04
2.93E-02 ± 4.05E-03
2.95E-02 ± 3.45E-03
1.76E-02 ± 2.45E-03
3.16E-02 ± 2.54E-03
2.14E-02 ± 3.57E-03
2.28E-02 ± 3.49E-03
2.97E-02 ± 3.77E-03
2.43E-02 ± 2.90E-03
7.36E-02 ± 3.72E-03
3.54E-02 ± 3.90E-03
4.42E-02 ± 3.56E-03
Cr
Table C.4. Particulate stable tracer and metal analyses from the June 2006 sampling in the ACF system. ‘BD’ indicates that a
measurement was below detection. Measurement uncertainties are reported at the 2-σ level.
128
1.16E-02 ± 4.93E-04
1.60E-02 ± 8.30E-04
1.43E-02 ± 3.99E-04
2.23E-02 ± 6.05E-04
1.18E-02 ±2.57E-04
1.90E-02 ± 4.40E-04
4.63E-02 ± 7.30E-04
1.92E-02 ± 7.22E-04
1.87E-02 ± 6.66E-04
1.77E-02 ± 3.24E-04
1.35E-02 ± 4.47E-04
1.15E-02 ± 6.00E-04
1.37E-02 ± 2.85E-04
6.29E-03 ± 3.46E-04
2.40E-02 ± 7.68E-04
4.94E+00 ± 4.52E-02
8.85E+00 ± 1.79E-02
9.85E+00 ± 1.37E-01
3.51E+01 ± 6.45E-01
2.40E+01 ± 3.47E-01
3.01E+00 ± 2.75E-02
3.91E+00 ± 1.80E-02
2.59E+00 ± 3.92E-02
6.97E+00 ± 6.83E-02
8.05E+00 ± 7.46E-02
8.61E+00 ± 4.35E-02
1.70E+00 ± 2.32E-02
1.04E+01 ± 6.62E-02
AR-1
AR-2
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
FR-1
FR-2
FR-3
FR-4
FR-5
Co
6.51E+00 ± 4.76E-02
7.58E+00 ± 8.57E-02
Mn
Sample Site
AB-1
AB-2
Table C.4. continued…
2.72E-02 ± 2.79E-03
1.58E-02 ± 1.57E-03
1.54E-02 ± 2.69E-03
9.55E-04 ± 2.97E-04
1.40E-02 ± 2.08E-03
1.66E-02 ± 1.06E-03
2.44E-02 ± 1.12E-03
2.88E-02 ± 1.11E-03
1.75E-02 ± 1.96E-03
1.78E-02 ± 2.02E-03
3.32E-02 ± 1.40E-03
2.79E-02 ± 3.25E-03
1.65E-02 ± 2.05E-03
6.69E-02 ± 8.43E-03
4.77E-02 ± 6.55E-03
2.68E-02 ± 8.06E-03
5.45E-03 ± 2.11E-03
3.44E-02 ± 7.64E-03
1.24E-01 ± 6.81E-03
4.62E-02 ± 6.21E-03
7.38E-02 ± 6.33E-03
3.85E-02 ± 8.46E-03
4.22E-02 ± 8.21E-03
1.75E-01 ± 9.20E-03
7.96E-02 ± 8.98E-03
4.27E-02 ± 7.20E-03
Ni
Cu
mg metal / g of sediment
1.50E-02 ± 2.02E-03 3.90E-02 ± 5.73E-03
2.86E-02 ± 2.06E-03 2.10E-01 ± 8.78E-03
2.58E-01 ± 1.95E-01
7.94E-02 ± 1.85E-01
4.94E-02 ± 1.78E-01
4.41E-03 ± 7.93E-03
2.10E-02 ± 1.28E-01
3.85E-01 ± 4.63E-02
1.51E-01 ± 3.92E-02
3.50E-01 ± 4.63E-02
5.62E-02 ± 1.48E-01
1.88E-01 ± 1.43E-01
9.79E-02 ± 4.15E-02
9.15E-01 ± 1.80E-01
1.89E-01 ± 1.61E-01
BD
BD
Zn
7.57E-03 ± 2.22E-04
5.46E-03 ± 4.20E-04
8.87E-03 ± 4.42E-04
1.19E-04 ± 1.21E-05
6.39E-03 ± 4.25E-04
7.49E-03 ± 2.19E-04
1.03E-02 ± 1.85E-04
6.88E-03 ± 2.19E-04
2.57E-02 ± 4.72E-04
6.85E-03 ± 2.14E-04
3.22E-03 ± 1.33E-04
7.09E-03 ± 7.47E-04
3.43E-02 ± 5.85E-04
1.26E-02 ± 2.06E-04
1.10E-02 ± 3.92E-04
As
129
2.37E-03 ± 3.40E-04
1.58E-03 ± 5.30E-04
8.17E-03 ± 4.07E-04
2.11E-03 ± 1.10E-04
4.02E-03 ± 1.69E-04
8.33E-04 ± 4.15E-04
4.82E-04 ± 3.45E-04
4.57E-04 ± 2.60E-04
4.68E-04 ± 3.61E-04
1.08E-04 ± 1.14E-04
1.23E-03 ± 1.18E-04
2.25E-02 ± 3.38E-04
6.14E-02 ± 1.42E-03
4.89E-02 ± 5.10E-04
1.98E-02 ± 1.37E-03
1.87E-02 ± 1.33E-03
1.21E-02 ± 2.54E-04
1.50E-02 ± 3.95E-04
1.68E-02 ± 2.85E-04
2.14E-02 ± 4.78E-04
7.24E-03 ± 6.65E-04
4.58E-02 ± 1.48E-03
2.34E-03 ± 2.81E-03
1.06E-03 ± 2.68E-03
1.60E-03 ± 2.52E-03
4.64E-03 ± 2.41E-03
8.60E-03 ± 2.29E-03
BD
BD
BD
BD
BD
BD
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
FR-1
FR-2
FR-3
FR-4
FR-5
3.45E-04 ± 3.64E-05
2.26E-04 ± 2.57E-05
1.21E-04 ± 3.16E-05
5.11E-06 ± 4.32E-05
2.23E-04 ± 1.75E-05
BD
BD
BD
BD
BD
BD
8.83E-05 ± 3.23E-05
2.24E-03 ± 2.22E-04
3.39E-04 ± 3.62E-04
1.74E-03 ± 4.10E-04
3.99E-02 ± 5.75E-04
3.10E-02 ± 6.15E-04
BD
BD
1.09E-01 ± 9.37E-04
1.11E-01 ± 2.18E-03
AR-1
AR-2
Mo
Ag
mg metal / g of sediment
6.82E-04 ± 2.93E-04
BD
1.51E-03 ± 2.16E-04 1.78E-03 ± 2.31E-04
Sr
BD
BD
Se
Sample Site
AB-1
AB-2
Table C.4. continued…
2.40E-03 ± 1.56E-04
1.23E-03 ± 1.59E-04
2.07E-03 ± 1.42E-04
BD
BD
1.73E-03 ± 1.44E-03
7.68E-03 ± 1.44E-03
2.93E-03 ± 1.42E-03
3.87E-04 ± 2.38E-04
5.91E-04 ± 2.30E-04
BD
1.81E-03 ± 1.91E-04
2.86E-03 ± 1.93E-04
9.67E-03 ± 1.31E-03
1.36E-03 ± 2.47E-04
Cd
1.68E-03 ± 9.99E-04
1.21E-03 ± 7.09E-04
1.07E-03 ± 1.00E-03
2.78E-05 ± 1.16E-04
1.78E-03 ± 8.63E-04
BD
BD
3.15E-04 ± 2.42E-03
2.30E-03 ± 9.36E-04
1.89E-03 ± 9.08E-04
BD
5.28E-04 ± 1.03E-03
1.21E-03 ± 9.59E-04
BD
2.40E-03 ± 8.11E-04
Sn
130
2.18E-01 ± 2.43E-03
7.98E-01 ± 1.57E-02
4.39E-01 ± 5.43E-03
2.17E-01 ± 1.66E-02
2.16E-01 ± 1.61E-02
2.39E-01 ± 3.97E-03
2.31E-01 ± 4.10E-03
2.70E-01 ± 1.62E-03
3.18E-01 ± 4.97E-03
1.06E-02 ± 8.94E-04
4.94E-01 ± 1.61E-02
4.24E-04 ± 8.39E-05
7.91E-04 ± 1.40E-04
7.21E-04 ± 8.39E-05
1.49E-03 ± 8.01E-05
2.05E-03 ± 1.07E-04
4.41E-04 ± 1.18E-04
4.51E-04 ± 1.08E-04
2.37E-04 ± 4.21E-05
9.31E-04 ± 1.39E-04
2.83E-04 ± 6.78E-05
3.18E-04 ± 1.10E-04
6.68E-06 ± 1.36E-05
1.78E-03 ± 8.63E-04
AR-1
AR-2
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
FR-1
FR-2
FR-3
FR-4
FR-5
2.64E-04 ± 2.77E-05
2.40E-04 ± 2.15E-05
3.11E-04 ± 2.63E-05
1.42E-04 ± 1.32E-05
8.87E-04 ± 5.82E-05
2.78E-04 ± 2.39E-05
4.73E-04 ± 3.34E-05
5.39E-04 ± 2.07E-05
5.66E-04 ± 2.24E-05
4.94E-04 ± 2.09E-05
3.59E-04 ± 1.95E-05
4.91E-04 ± 4.90E-05
4.37E-04 ± 3.73E-05
2.53E-01 ± 1.93E-03
2.77E-01 ± 3.09E-03
1.05E-03 ± 5.90E-05
1.54E-03 ± 1.29E-04
Sample Site
AB-1
AB-2
Sb
Ba
Tl
mg metal / g of sediment
2.66E-01 ± 2.98E-03 3.34E-04 ± 4.01E-05
1.38E-01 ± 1.43E-02 4.13E-04 ± 2.73E-05
Table C.4. continued…
5.68E-02 ± 1.26E-03
3.12E-02 ± 8.53E-04
3.46E-02 ± 1.21E-03
3.10E-03 ± 4.83E-04
5.00E-02 ± 1.75E-03
2.50E-02 ± 2.16E-02
4.83E-02 ± 2.12E-02
8.46E-02 ± 2.11E-02
4.84E-02 ± 1.81E-03
5.40E-02 ± 1.70E-03
2.91E-02 ± 2.07E-02
4.14E-02 ± 1.17E-03
6.07E-02 ± 1.27E-03
1.61E-01 ± 1.78E-02
3.04E-02 ± 1.58E-03
Pb
Table C.5. June 2006 particulate org-C and N concentrations (%) as well as their isotopic
values.
13
Sample Site
AB-1
AB-2
org-C (%)
9.2
8.9
δ CPDB
AR-1
AR-2
15
δ Nair
-23.9
-30.1
N (%)
1.6
1.4
6.5
7.0
-31.7
-29.0
1.0
0.9
9.6
7.1
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
13.1
11.2
8.6
3.4
3.7
1.6
-30.7
-27.1
-24.5
-26.6
-26.8
-27.2
1.8
1.7
1.1
0.4
0.4
0.2
10.2
9.2
8.7
7.7
7.9
11.7
FR-1
FR-2
FR-3
FR-4
FR-5
9.6
2.1
4.6
1.3
2.8
-29.4
-30.3
-28.3
-28.0
-27.4
1.4
0.4
0.7
0.1
0.3
4.0
8.3
7.3
4.2
7.4
131
9.7
6.9
Table C.6. Water quality characteristics from the February 2007 sampling of the ACF system.
‘nd’ indicates that no data exists for that particular measurement.
Distance Upstream From
Specific
In-situ
Sample Site Apalachicola Bay (km) Conductance (μS/cm) Temperature (°C)
AB-1
-0.6
1,395
16.7
AB-2
0.0
175
16.8
pH
8.0
8.0
TSS
(mg/L)
4.72
4.90
AR-1
AR-2
10.6
169.9
129
133
17.5
16.7
7.7
6.0
9.80
7.55
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
208.0
290.9
484.0
540.0
580.7
688.2
124
111
108
110
141
55
14.0
14.2
11.9
14.1
14.4
10.0
5.8
5.8
5.8
5.6
5.8
5.4
7.01
1.81
6.70
17.87
21.96
1.70
FR-1
FR-2
FR-3
FR-4
FR-5
222.2
326.3
362.2
414.0
607.1
134
110
98
78
69
16.4
16.8
15.0
15.7
12.9
5.7
5.8
5.8
5.7
5.7
4.39
4.73
3.16
7.61
5.10
132
Table C.7. Dissolved radium isotope activity concentrations during the February 2007 ACF
sampling. ‘BD’ indicates that a measurement was below detection. Measurement uncertainties
are reported at the 1-σ level.
Sample Site
AB-1
AB-2
Ra-223
dpm/100L
0.49 ± 0.067
0.09 ± 0.03
Ra-224
dpm/100L
2.48 ± 0.16
1.18 ± 0.10
Ra-226
dpm/100L
11.08 ± 0.57
3.73 ± 0.47
Ra-228
dpm/100L
13.28 ± 1.42
0.34 ± 1.28
AR-1
AR-2
0.12 ± 0.04
0.03 ± 0.04
1.86 ± 0.18
nd
4.27 ± 0.57
5.29 ± 2.58
9.04 ± 1.98
8.29 ± 9.52
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
0.11 ± 0.04
0.03 ± 0.02
0.03 ± 0.03
0.05 ± 0.03
0.12 ± 0.05
0.06 ± 0.03
3.58 ± 0.29
1.69 ± 0.11
1.40 ± 0.11
2.15 ± 0.11
3.09 ± 0.19
1.11 ± 0.08
3.99 ± 1.68
1.44 ± 0.52
1.15 ± 0.52
1.73 ± 0.52
2.13 ± 1.11
0.42 ± 0.66
0.26 ± 5.22
1.28 ± 1.81
1.82 ± 1.76
2.18 ± 1.38
6.03 ± 3.92
5.29 ± 2.83
FR-1
FR-2
FR-3
FR-4
FR-5
0.03 ± 0.03
0.06 ± 0.02
0.08 ± 0.03
0.06 ± 0.02
0.17 ± 0.04
0.94 ± 0.17
0.44 ± 0.11
2.78 ± 0.16
2.58 ± 0.13
3.34 ± 0.26
4.89 ± 0.53
4.42 ± 0.54
6.10 ± 0.54
8.27 ± 0.62
3.11 ± 0.57
3.18 ± 1.62
3.81 ± 1.79
3.51 ± 1.78
7.24 ± 1.58
4.70 ± 1.95
133
134
Ra-228
dpm/g
5.54 ± 1.36
4.20 ± 1.06
5.10 ± 1.00
4.59 ± 1.03
4.88 ± 1.09
4.17 ± 1.46
4.12 ± 1.05
3.48 ± 1.15
5.74 ± 1.11
4.40 ± 1.30
4.40 ± 1.01
5.33 ± 1.38
8.71 ± 1.11
9.14 ± 1.12
5.61 ± 1.32
Ra-226
dpm/g
4.70 ± 0.70
5.46 ± 0.53
5.01 ± 0.58
5.33 ± 0.53
5.28 ± 0.50
3.98 ± 0.70
2.73 ± 0.55
4.14 ± 0.59
5.00 ± 0.57
4.19 ± 0.67
6.61 ± 0.58
8.52 ± 0.68
10.59 ± 0.61
13.37± 0.59
6.61 ± 0.64
Sample Site
AB-1
AB-2
AR-1
AR-2
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
FR-1
FR-2
FR-3
FR-4
FR-5
24.34 ± 2.40
18.60 ± 2.61
19.30 ± 2.42
14.38 ± 2.65
16.24 ± 3.08
15.16 ± 2.22
31.03 ± 3.21
9.19 ± 2.79
9.90 ± 2.56
16.65 ± 2.42
19.97 ± 2.48
7.46 ± 2.21
11.99 ± 2.20
Pb-210
dpm/g
10.21 ± 2.53
17.05 ± 2.38
76.47 ± 6.01
66.18 ± 10.78
38.27 ± 6.33
30.48 ± 9.58
24.38 ± 12.32
63.72 ± 5.06
53.48 ± 6.75
14.28 ± 5.41
33.23 ± 7.46
53.82 ± 6.48
97.97 ± 8.77
19.67 ± 8.95
29.75 ± 7.76
Be-7
dpm/g
1.86 ± 11.91
6.81 ± 10.07
0.84 ± 0.27
0.83 ± 0.27
0.93 ± 0.26
0.19 ± 0.24
BD
0.53 ± 0.21
0.03 ± 0.21
0.22 ± 0.27
0.30 ± 0.27
0.06 ± 0.29
0.47 ± 0.26
0.28 ± 0.19
0.11 ± 0.19
Cs-137
dpm/g
0.84 ± 0.26
0.07 ± 0.20
5.46 ± 2.38
8.99 ± 3.03
2.56 ± 2.66
17.65 ± 2.60
12.79 ± 3.05
14.38 ± 2.57
5.87 ± 3.33
11.76 ± 2.38
21.88 ± 3.04
24.19 ± 2.93
22.52 ± 3.21
19.17 ± 2.68
18.16 ± 2.48
K-40
dpm/g
15.91 ± 3.21
15.14 ± 2.57
Table C.8. Particulate radionuclide analyses during February 2007 in the ACF sampling. ‘BD’ indicates that a measurement
was below detection. Measurement uncertainties are reported at the 1-σ level.
135
BD
1.88E+00 ± 2.46E-02
1.82E+00 ± 2.85E-02
1.57E+00 ± 1.73E-02
2.26E+00 ± 1.76E-01
1.62E+00 ± 1.56E-01
1.43E+00 ± 1.38E-01
3.24E+00 ± 1.33E-01
3.74E+00 ± 1.38E-01
3.91E+00 ± 9.43E-02
1.02E+00 ± 4.33E-01
1.10E+00 ± 4.03E-01
1.22E+00 ± 4.58E-01
1.70E+00 ± 3.79E-01
1.72E+00 ± 4.31E-01
1.26E-03 ± 8.18E-05
1.15E-03 ± 1.44E-04
8.88E-04 ± 1.94E-04
9.36E-04 ± 1.50E-04
6.39E-04 ± 3.59E-04
7.59E-04 ± 2.37E-05
9.62E-04 ± 1.00E-05
9.27E-04 ± 9.17E-05
1.56E-03 ± 8.62E-05
1.47E-03 ± 4.04E-05
1.47E-03 ± 7.80E-05
1.20E-03 ± 5.02E-05
1.10E-03 ± 2.35E-04
AR-1
AR-2
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
FR-1
FR-2
FR-3
FR-4
FR-5
Mg
Be
Sample Site
1.35E-03 ± 1.36E-04
AB-1
1.29E-03 ± 1.23E-04
AB-2
7.67E+00 ± 2.19E-01
9.77E+00 ± 1.32E-01
1.10E+01 ± 1.62E-01
9.77E+00 ± 1.09E-01
1.14E+01 ± 1.59E-01
8.83E+00 ± 3.17E-01
8.12E+00 ± 2.21E-01
7.22E+00 ± 1.79E-01
9.38E+00 ± 2.16E-01
1.23E+01 ± 1.79E-01
1.19E+01 ± 3.37E-01
4.09E-02 ± 8.00E-04
6.89E-02 ± 1.36E-03
6.89E-02 ± 9.32E-04
4.97E-02 ± 5.57E-04
6.12E-02 ± 4.82E-04
5.82E-02 ± 1.11E-03
3.14E-02 ± 1.06E-03
5.01E-02 ± 5.72E-04
3.02E-02 ± 4.24E-04
2.85E-02 ± 5.39E-04
3.16E-02 ± 9.36E-04
4.32E-02 ± 1.26E-03
2.81E-02 ± 7.86E-04
2.85E-02 ± 7.55E-04
2.99E-02 ± 7.27E-04
3.62E-02 ± 8.71E-04
3.32E-02 ± 7.41E-04
2.89E-02 ± 1.02E-03
2.63E-02 ± 5.67E-04
BD
4.43E-02 ± 1.06E-03
5.93E-02 ± 7.00E-04
5.94E-02 ± 1.00E-03
5.20E-02 ± 1.11E-03
4.42E-02 ± 9.17E-04
Cr
V
4.25E+00 ± 1.12E-01 5.84E-02 ± 1.23E-03
3.02E+00 ± 5.19E-02 6.62E-02 ± 4.72E-04
2.32E+00 ± 4.47E-02 7.14E-02 ± 5..45E-04
1.04E+00 ± 2.83E-02 5.58E-02 ± 5.65E-04
7.96E+00 ± 2.76E-02 5.64E-02 ± 5.30E-04
2.55E+00 ± 8.68E-02
2.23E+00 ± 7.61E-02
1.24E+00 ± 6.27E-02
1.07E+00 ± 5.41E-02
1.11E+00 ± 5.81E-02
1.18E+00 ± 2.92E-02
8.53E+00 ± 1.19E-01 2.32E+00 ± 3.06E-02
7.15E+00 ± 6.49E-02 2.39E+00 ± 3.93E-02
Al
Ca
mg metal / g of sediment
1.05E+01 ± 7.82E-02 4.89E-01 ± 5.25E-03
9.77E+00 ± 1.37E-01 2.11E+00 ± 2.54E-02
Table C.9. Particulate stable tracer and metal analyses from the February 2007 sampling in the ACF system. ‘BD’ indicates that a
measurement was below detection. Measurement uncertainties are reported at the 2-σ level.
136
1.02E-02 ± 1.62E-04
1.60E-02 ± 2.58E-04
1.57E-02 ± 1.97E-04
1.13E-02 ± 1.82E-04
2.04E-02 ± 4.62E-04
1.16E-02 ± 1.37E-04
1.57E-02 ± 2.34E-04
1.50E-02 ± 2.52E-04
1.86E-02 ± 1.85E-04
2.20E-02 ± 4.60E-04
1.92E-02 ± 4.12E-04
2.10E-02 ± 1.36E-04
2.27E-02 ± 2.00E-04
2.79E-02 ± 2.95E-04
2.14E-02 ± 2.10E-04
2.94E+00 ± 3.01E-02
3.90E+00 ± 3.03E-02
4.26E+00 ± 9.50E-02
3.81E+00 ± 6.30E-02
7.05E+00 ± 7.59E-02
2.18E+00 ± 2.60E-02
3.44E+00 ± 2.15E-02
3.54E+00 ± 7.17E-02
5.30E+00 ± 1.04E-01
8.75E+00 ± 5.09E-02
1.10E+01 ± 9.93E-02
6.30E+00 ± 3.97E-02
7.08E+00 ± 5.94E-02
AR-1
AR-2
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
FR-1
FR-2
FR-3
FR-4
FR-5
Co
Mn
Sample Site
3.47E+00 ± 3.01E-02
AB-1
3.10E+00 ± 4.07E-02
AB-2
Table C.9. continued…
1.53E-02 ± 6.64E-04
1.89E-02 ± 5.19E-04
1.52E-02 ± 4.89E-04
1.34E-02 ± 5.65E-04
3.09E-02 ± 8.34E-04
2.42E-02 ± 1.74E-03
1.15E-02 ± 1.53E-03
3.44E-02 ± 1.55E-03
1.79E-02 ± 1.21E-03
2.44E-02 ± 1.32E-03
2.98E-02 ± 1.20E-03
4.54E-03 ± 3.88E-04
1.66E-03 ± 4.84E-04
2.27E-01 ± 5.98E-03
2.03E-01 ± 3.10E-03
1.10E-01 ± 3.01E-03
5.63E-02 ± 2.35E-03
1.22E-01 ± 3.03E-03
2.89E-01 ± 7.23E-03
1.45E-01 ± 5.23E-03
9.16E-02 ± 4.26E-03
1.40E-01 ± 4.50E-03
5.91E-02 ± 3.66E-03
4.21E-02 ± 1.50E-03
BD
9.64E-04 ± 6.15E-04
Ni
Cu
mg metal / g of sediment
6.25E-03 ± 5.13E-04 5.06E-03 ± 4.50E-04
2.22E-02 ± 1.02E-03 1.78E-02 ± 9.78E-04
4.28E-01 ± 4.82E-02
5.77E-01 ± 4.40E-02
6.36E-01 ± 4.87E-02
1.47E+00 ± 4.19E-02
1.72E-01 ± 4.54E-02
4.79E-01 ± 2.49E-02
9.29E-01 ± 2.81E-02
2.21E-01 ± 2.02E-02
3.18E-01 ± 1.71E-02
3.08E-01 ± 1.85E-02
1.93E-01 ± 5.14E-03
8.74E-02 ± 1.95E-03
8.21E-02 ± 1.46E-03
8.45E-02 ± 7.88E-04
7.29E-02 ± 2.05E-03
Zn
1.03E-02 ± 2.98E-04
1.35E-02 ± 2.75E-04
1.09E-02 ± 2.18E-04
5.79E-03 ± 1.54E-04
5.68E-03 ± 1.51E-04
8.30E-03 ± 3.08E-04
1.32E-02 ± 4.33E-04
1.32E-02 ± 2.73E-04
1.04E-02 ± 2.08E-04
7.23E-03 ± 1.83E-04
5.72E-03 ± 2.22E-04
8.16E-03 ± 1.84E-04
7.53E-03 ± 2.20E-04
1.53E-02 ± 4.12E-04
9.65E-03 ± 2.30E-04
As
137
1.09E-01 ± 1.21E-03
2.16E-02 ± 4.27E-04
1.94E-02 ± 1.90E-04
1.62E-02 ± 1.84E-04
3.23E-02 ± 1.06E-03
3.41E-02 ± 7.44E-04
2.26E-02 ± 5.93E-04
1.93E-02 ± 5.32E-04
1.99E-02 ± 4.75E-04
2.26E-02 ± 5.29E-04
3.42E-02 ± 8.52E-04
2.91E-02 ± 4.87E-04
2.98E-02 ± 5.19E-04
2.52E-02 ± 4.57E-04
2.90E-02 ± 4.43E-04
2.69E-03 ± 1.74E-03
2.89E-03 ± 1.77E-03
3.81E-03 ± 1.80E-03
4.07E-03 ± 1.88E-03
4.14E-03 ± 1.16E-03
2.96E-03 ± 1.48E-03
3.45E-03 ± 1.41E-03
2.57E-03 ± 1.69E-03
3.30E-03 ± 2.23E-03
3.11E-03 ± 2.18E-03
2.92E-03 ± 1.83E-03
2.29E-03 ± 1.58E-03
9.57E-04 ± 2.17E-03
AR-1
AR-2
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
FR-1
FR-2
FR-3
FR-4
FR-5
Sr
Sample Site
6.38E-05 ± 5.46E-05
AB-1
2.71E-03 ± 2.04E-03
AB-2
Se
Table C.9. continued…
3.89E-04 ± 5.07E-05
2.32E-04 ± 4.77E-05
2.66E-04 ± 5.14E-05
1.93E-04 ± 4.18E-05
1.50E-04 ± 4.56E-05
1.37E-03 ± 1.03E-04
3.10E-04 ± 4.60E-05
2.30E-04 ± 3.86E-05
4.65E-04 ± 3.52E-05
5.47E-04 ± 3.67E-05
9.27E-05 ± 1.95E-05
1.12E-04 ± 1.87E-05
1.27E-04 ± 1.42E-05
1.32E-03 ± 1.20E-04
1.19E-03 ± 8.56E-05
1.05E-03 ± 7.94E-05
5.76E-04 ± 3.36E-05
1.98E-04 ± 3.90E-05
2.02E-03 ± 1.10E-04
6.16E-04 ± 1.04E-04
4.93E-04 ± 8.60E-05
6.19E-04 ± 7.10E-05
6.75E-04 ± 9.51E-05
5.96E-04 ± 6.26E-05
4.25E-04 ± 4.04E-05
4.70E-04 ± 3.47E-05
Mo
Ag
mg metal / g of sediment
3.61E-04 ± 1.77E-05 3.19E-04 ± 3.93E-05
3.78E-04 ± 3.52E-05 4.00E-04 ± 5.01E-05
3.64E-03 ± 5.42E-04
3.45E-03 ± 4.41E-04
2.45E-03 ± 5.05E-04
1.44E-03 ± 4.06E-04
3.57E-03 ± 5.45E-04
3.43E-03 ± 1.40E-03
4.14E-03 ± 1.33E-03
4.16E-03 ± 1.16E-03
3.68E-03 ± 9.91E-04
4.82E-03 ± 1.14E-03
1.85E-02 ± 4.53E-04
7.51E-04 ± 1.02E-04
1.28E-03 ± 1.28E-04
5.35E-03 ± 1.26E-04
1.18E-03 ± 1.03E-04
Cd
2.11E-03 ± 9.46E-05
2.25E-03 ± 5.82E-05
2.11E-03 ± 7.54E-05
1.14E-03 ± 6.23E-05
6.89E-04 ± 6.43E-05
4.77E-03 ± 1.99E-04
4.02E-03 ± 2.01E-04
1.17E-02 ± 2.64E-05
1.33E-03 ± 1.25E-04
1.65E-03 ± 1.44E-04
1.68E-03 ± 7.37E-05
4.83E-04 ± 3.24E-05
1.27E-03 ± 5.84E-05
5.20E-04 ± 5.31E-05
6.82E-04 ± 9.18E-05
Sn
138
Sb
2.52E-04 ± 1.39E-05
1.96E-04 ± 2.15E-05
1.99E-04 ± 1.75E-05
1.38E-04 ± 1.40E-05
1.61E-04 ± 7.90E-06
2.88E-04 ± 1.07E-05
2.94E-04 ± 8.57E-06
2.72E-04 ± 7.73E-06
1.98E-04 ± 1.77E-05
2.28E-04 ± 1.63E-05
2.70E-04 ± 1.98E-05
3.92E-04 ± 1.41E-05
3.69E-04 ± 1.87E-05
1.75E-01 ± 1.35E-03
2.16E-01 ± 2.03E-03
2.35E-01 ± 6.22E-03
2.19E-01 ± 3.32E-03
2.67E-01 ± 4.41E-03
1.74E-01 ± 2.86E-03
2.26E-01 ± 2.02E-03
2.26E-01 ± 5.21E-03
2.57E-01 ± 6.92E-03
3.25E-01 ± 2.81E-03
4.27E-01 ± 5.99E-03
3.58E-01 ± 3.46E-03
4.21E-01 ± 4.01E-03
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
FR-1
FR-2
FR-3
FR-4
FR-5
2.62E-01 ± 5.10E-03
1.77E-01 ± 1.50E-03
2.17E-01 ± 2.92E-03
1.78E-01 ± 2.02E-03
1.06E-01 ± 1.08E-03
4.02E-01 ± 1.45E-02
6.50E-01 ± 1.12E-02
4.18E-01 ± 7.19E-3
3.81E-02 ± 1.44E-03
7.63E-02 ± 1.63E-03
3.87E-02 ± 1.05E-03
7.55E-02 ± 5.42E-04
4.11E-02 ± 2.65E-04
Ba
Tl
mg metal / g of sediment
2.06E-04 ± 1.42E-05 9.03E-02 ± 7.95E-04
2.54E-04 ± 1.01E-05 7.13E-02 ± 5.33E-04
AR-1
AR-2
Sample Site
1.14E-01 ± 1.76E-03
AB-1
1.69E-01 ± 2.70E-03
AB-2
Table C.9. continued…
1.73E-03 ± 5.45E-05
2.44E-03 ± 4.91E-05
2.61E-03 ± 8.01E-05
1.72E-03 ± 5.67E-05
2.46E-03 ± 4.30E-05
1.90E-03 ± 2.69E-05
2.38E-03 ± 4.22E-05
1.63E-03 ± 4.95E-05
1.90E-03 ± 3.60E-05
2.53E-03 ± 3.73E-05
1.85E-03 ± 6.18E-05
2.25E-03 ± 3.57E-05
1.51E-03 ± 2.29E-05
1.37E-03 ± 2.96E-05
2.38E-03 ± 4.03E-05
Pb
Table C.10. February 2007 particulate org-C and N concentrations (%) as well as their isotopic
values.
13
15
Sample Site
org-C (%)
N (%)
δ CPDB
δ Nair
AB-1
5.0
-24.2
0.7
7.2
AB-2
3.8
-28.8
0.4
7.7
AR-1
AR-2
3.4
3.9
-28.4
-28.5
0.4
0.4
7.8
7.0
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
7.4
7.2
3.7
3.3
3.7
3.5
-27.8
-28.4
-32.4
-26.9
-27.0
-27.0
0.6
0.9
0.6
0.4
0.4
0.4
5.3
9.6
16.2
9.3
9.6
9.0
FR-1
FR-2
FR-3
FR-4
FR-5
4.6
6.0
5.0
4.0
3.8
-27.4
-27.3
-27.1
-27.2
-28.3
0.5
0.7
0.6
0.4
0.5
6.2
5.0
5.4
2.9
7.5
139
140
1.07E+02 ± 1.11E+00
1.15E+02 ± 5.17E-01
6.76E+01 ± 9.28E-01
6.58E+01 ± 4.44E-01
7.91E+01 ± 5.09E-01
9.26E-04 ± 2.66E-04
1.25E-03 ± 3.97E-04
1.11E-03 ± 3.43E-04
1.58E-03 ± 2.94E-04
6.63E-04 ± 2.18E-04
FR-1
FR-2
FR-3
FR-4
FR-5
3.05E+00 ± 1.87E-01
9.79E+00 ± 3.65E-01
1.40E+01 ± 4.95E-01
1.21E+01 ± 3.80E-01
7.76E+00 ± 4.60E-01
5.67E+00 ± 2.72E-01
3.10E+00 ± 4.76E-01
6.61E-01 ± 1.80E+00
4.66E+00 ± 3.41E-01
4.71E+00 ± 2.79E-01
9.11E-01 ± 2.37E-01
1.35E+02 ± 3.43E+00
1.38E+02 ± 1.92E+00
1.53E+02 ± 1.19E+00
1.50E+02 ± 7.39E-01
1.74E+02 ± 2.57E+00
1.25E+02 ± 1.32E+00
BD
BD
4.40E-05 ± 3.03E-04
1.38E-04 ± 1.96E-04
1.92E-04 ± 2.39E-04
BD
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
1.25E+03 ± 1.92E+01
7.50E+02 ± 3.40E+00
5.63E+02 ± 9.97E+00
2.60E+02 ± 2.50E+00
2.97E+02 ± 2.41E+00
4.76E+02 ± 1.13E+01
4.14E+02 ± 6.20E+00
4.51E+02 ± 2.10E+00
4.57E+02 ± 3.10E+00
5.33E+02 ± 6.42E+00
2.17E+02 ± 2.15E+00
9.48E+02 ± 8.25E+00
1.22E+03 ± 2.29E+01
1.43E+01 ± 4.72E-01
7.35E+00 ± 2.35E-01
8.27E+01 ± 5.14E-01
7.46E+01 ± 1.35E+00
9.95E-04 ± 3.13E-04
7.14E-04 ± 3.26E-04
AR-1
AR-2
Ca
ppb
7.77E-01 ± 4.84E-02 6.75E+03 ± 2.15E+02
OC
1.22E+02 ± 1.61E+00 5.46E+01 ± 1.14E+00 9.74E+02 ± 1.45E+01
Al
BD
1.00E-03 ± 3.41E-04
Mg
Sample Site
AB-1
AB-2
Be
BD
BD
BD
BD
BD
BD
BD
BD
BD
BD
BD
BD
BD
BD
BD
V
3.45E-02 ± 1.57E-03
4.35E-02 ± 2.73E-03
BD
BD
BD
3.27E-02 ± 1.01E-02
3.69E-02 ± 1.73E-03
2.34E-02 ± 1.17E-03
3.08E-02 ± 9.05E-04
3.01E-02 ± 1.12E-03
1.79E-02 ± 8.99E-04
BD
BD
BD
BD
Cr
Table C.11. Dissolved stable tracer and metal analyses from the February 2007 sampling in the ACF system. ‘BD’ indicates that a
measurement was below detection. Measurement uncertainties are reported at the 2-σ level.
141
4.60E-02 ± 7.72E-04
1.54E-02 ± 1.35E-04
1.62E-02 ± 3.13E-04
1.48E-02 ± 5.30E-04
1.31E-02 ± 3.05E-04
7.22E-03 ± 1.88E-04
9.79E-03 ± 6.06E-04
1.31E-02 ± 1.70E-04
1.83E-02 ± 5.71E-04
7.90E-03 ± 5.40E-04
1.20E-02 ± 5.39E-04
1.48E-02 ± 3.32E-04
1.26E-02 ± 3.03E-04
2.78E-02 ± 3.63E-04
2.27E-02 ± 2.20E-04
1.59E+00 ± 1.27E-02
1.84E+00 ± 2.66E-02
2.16E+00 ± 3.92E-02
9.89E-02 ± 2.87E-02
9.68E-02 ± 2.39E-02
2.17E+00 ± 1.71E-02
3.07E+00 ± 5.75E-02
3.61E+00 ± 5.69E-02
6.99E-01 ± 8.35E-03
3.80E+00 ± 3.21E-02
4.32E+00 ± 7.06E-02
7.10E+00 ± 3.49E-02
1.22E+01 ± 7.67E-02
AR-1
AR-2
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
FR-1
FR-2
FR-3
FR-4
FR-5
Co
Mn
Sample Site
8.92E-01 ± 1.72E-02
AB-1
1.44E+00 ± 2.08E-02
AB-2
Table C.11. continued…
1.02E-01 ± 2.45E-03
8.00E-02 ± 1.10E-03
7.08E-02 ± 3.76E-03
7.37E-02 ± 9.92E-04
6.88E-02 ± 7.36E-04
9.48E-02 ± 3.11E-03
8.32E-02 ± 2.02E-03
8.81E-02 ± 1.19E-03
9.57E-02 ± 1.99E-03
1.57E-01 ± 2.11E-03
5.31E-02 ± 1.39E-03
1.35E-01 ± 1.90E-03
1.42E-01 ± 2.07E-03
BD
BD
BD
BD
Cu
2.95E-01 ± 3.34E-03
3.43E-01 ± 2.81E-03
BD
BD
BD
3.22E-01 ± 6.02E-03
3.17E-01 ± 4.03E-03
3.85E-01 ± 2.97E-03
4.39E-01 ± 2.30E-03
4.98E-01 ± 7.31E-03
3.35E-01 ± 4.08E-03
ppb
2.97E-01 ± 2.53E-02
3.54E-01 ± 5.44E-03
Ni
As
1.65E+00 ± 1.81E-02
1.83E+00 ± 1.54E-02
2.48E+00 ± 3.52E-02
1.99E+00 ± 1.13E-02
1.95E+00 ± 9.81E-03
1.76E+00 ± 4.26E-02
1.51E+00 ± 2.24E-02
1.90E+00 ± 8.42E-03
2.86E+00 ± 1.94E-02
2.69E+00 ± 8.43E-02
2.06E+00 ± 8.03E-02
2.10E+00 ± 1.99E-03
2.37E+00 ± 3.25E-02
2.49E-02 ± 1.79E-03
2.91E-02 ± 1.48E-03
1.44E-02 ± 1.15E-03
1.35E-02 ± 4.73E-04
1.15E-02 ± 7.90E-04
2.99E-02 ± 1.88E-03
2.78E-02 ± 8.63E-04
3.37E-02 ± 7.69E-04
4.70E-02 ± 1.57E-03
2.90E-02 ± 1.66E-03
1.48E-02 ± 1.57E-03
2.04E-02 ± 1.03E-03
1.78E-02 ± 6.78E-04
1.34E+00 ± 1.64E-02 1.06E+00 ± 3.64E-02
1.87E+00 ± 2.87E-02 2.35E-02 ± 6.80E-04
Zn
142
Sr
4.08E+00 ± 2.00E-02
3.66E+00 ± 6.58E-02
4.18E+00 ± 8.20E-02
4.26E+00 ± 5.64E-02
4.82E+00 ± 2.47E-02
4.88E+00 ± 3.55E-02
4.94E+00 ± 4.02E-02
3.23E+00 ± 3.88E-02
3.50E+00 ± 3.25E-02
3.33E+00 ± 4.17E-02
3.17E+00 ± 3.26E-02
2.74E+00 ± 2.10E-02
4.06E+00 ± 3.62E-02
4.81E-02 ± 3.62E-02
5.35E-02 ± 4.99E-02
3.42E-02 ± 5.01E-02
3.55E-02 ± 3.80E-02
2.69E-02 ± 2.97E-02
6.38E-02 ± 5.55E-02
2.55E-02 ± 2.70E-02
9.04E-03 ± 2.61E-02
3.37E-02 ± 1.18E-01
3.47E-02 ± 1.18E-01
1.15E-02 ± 3.21E-02
3.77E-02 ± 4.15E-02
2.44E-02 ± 3.75E-02
AR-1
AR-2
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
FR-1
FR-2
FR-3
FR-4
FR-5
Sample Site
4.06E+00 ± 6.36E-01 1.97E+02 ± 2.77E+00
AB-1
7.73E-02 ± 8.95E-02 4.43E+00 ± 5.36E-02
AB-2
Se
Table C.11. continued…
Ag
2.81E-02 ± 1.60E-03
2.55E-02 ± 6.48E-04
2.62E-02 ± 8.49E-04
2.69E-02 ± 6.21E-04
5.99E-02 ± 1.54E-03
2.43E-01 ± 3.59E-03
2.70E-01 ± 4.60E-03
5.06E-01 ± 4.17E-03
4.47E-01 ± 4.94E-03
2.15E-01 ± 2.22E-03
2.26E-02 ± 1.08E-03
1.50E-01 ± 2.10E-03
1.05E-01 ± 2.55E-03
2.80E-03 ± 2.06E-04
3.85E-03 ± 1.52E-04
2.63E-03 ± 1.73E-04
3.93E-03 ± 3.15E-04
4.90E-03 ± 1.72E-04
5.68E-04 ± 3.87E-04
9.76E-04 ± 3.23E-04
1.35E-03 ± 3.23E-04
2.55E-03 ± 3.64E-04
2.21E-03 ± 4.16E-04
1.43E-03 ± 4.80E-04
6.76E-03 ± 1.73E-04
7.95E-03 ± 5.29E-04
ppb
5.56E-01 ± 1.99E-02 2.15E-03 ± 2.85E-04
1.63E-01 ± 3.50E-03 1.06E-02 ± 3.41E-04
Mo
1.41E-03 ± 5.17E-04
1.41E-03 ± 4.03E-04
1.38E-03 ± 5.36E-04
2.42E-03 ± 5.10E-04
3.05E-03 ± 6.84E-04
1.49E-03 ± 3.50E-04
1.33E-03 ± 3.60E-04
1.41E-03 ± 3.72E-04
1.86E-03 ± 5.64E-04
2.31E-03 ± 5.87E-04
1.93E-03 ± 5.89E-04
2.23E-03 ± 1.01E-03
1.97E-03 ± 5.44E-04
3.01E-03 ± 8.41E-04
1.75E-03 ± 8.37E-04
Cd
3.12E-02 ± 2.57E-03
2.69E-02 ± 2.24E-03
3.64E-02 ± 2.02E-03
3.83E-02 ± 1.73E-03
3.74E-02 ± 1.76E-03
3.83E-02 ± 7.57E-03
5.54E-02 ± 2.72E-03
2.20E-02 ± 5.08E-03
3.24E-02 ± 3.81E-03
4.43E-02 ± 5.37E-03
2.70E-02 ± 4.72E-03
4.57E-02 ± 1.44E-03
6.81E-02 ± 2.61E-03
6.53E-02 ± 6.37E-02
6.25E-02 ± 1.53E-03
Sn
143
Sb
Tl
2.83E+00 ± 2.80E-02
2.74E+00 ± 4.83E-02
2.32E+00 ± 4.36E-02
2.25E+00 ± 3.18E-02
2.00E+00 ± 2.65E-02
1.82E+00 ± 1.81E-02
1.90E+00 ± 1.32E-02
1.17E+00 ± 5.44E-03
2.26E+00 ± 1.18E-02
2.82E+00 ± 4.49E-02
3.01E+00 ± 2.72E-02
3.20E+00 ± 4.29E-02
3.71E+00 ± 2.68E-02
3.09E-02 ± 9.52E-04
2.29E-02 ± 8.98E-04
4.07E-02 ± 2.12E-03
4.43E-02 ± 1.98E-03
4.99E-02 ± 1.59E-03
5.47E-02 ± 1.90E-03
3.06E-02 ± 1.02E-03
6.85E-03 ± 8.73E-04
6.74E-03 ± 7.26E-04
7.76E-03 ± 6.26E-04
7.17E-03 ± 8.35E-04
8.39E-03 ± 7.87E-04
1.26E-02 ± 7.69E-04
CR-1
CR-2
CR-3
CR-4
CR-5
CR-6
FR-1
FR-2
FR-3
FR-4
FR-5
7.21E-04 ± 2.22E-04
8.60E-04 ± 2.21E-04
7.97E-04 ± 1.01E-04
7.44E-04 ± 1.67E-04
5.03E-04 ± 2.08E-04
7.76E-04 ± 1.29E-04
7.48E-04 ± 1.12E-04
1.85E-03 ± 1.97E-04
2.59E-03 ± 2.40E-04
6.77E-04 ± 1.17E-04
4.12E-04 ± 1.31E-04
1.03E-03 ± 3.06E-04
8.32E-04 ± 1.51E-04
ppb
1.62E+00 ± 6.50E-02 9.79E-04 ± 1.63E-04
3.03E+00 ± 3.15E-02 9.77E-04 ± 1.33E-04
Ba
AR-1
AR-2
Sample Site
7.61E-02 ± 7.40E-03
AB-1
3.24E-02 ± 8.46E-04
AB-2
Table C.11. continued…
8.46E-02 ± 1.29E-03
1.45E-01 ± 2.34E-03
6.58E-02 ± 1.57E-03
1.82E-01 ± 3.86E-03
3.15E-01 ± 1.21E-03
4.89E-02 ± 2.18E-03
5.08E-02 ± 1.17E-03
1.03E-01 ± 1.18E-03
3.93E-02 ± 4.90E-04
3.61E-02 ± 1.26E-03
2.90E-02 ± 7.28E-04
1.02E-01 ± 1.21E-03
8.23E-02 ± 2.31E-03
2.13E-02 ± 1.55E-03
1.33E-01 ± 2.73E-03
Pb
REFERENCES
Beck, A.J., Rapaglia, J.P., Cochran, J.K. and Bokuniewicz, H.J., 2007. Radium mass-balance in
Jamaica Bay, NY: Evidence for a substantial flux of submarine groundwater. Marine
Chemistry, 106: 419-441.
Beck, A.J., Rapaglia, J.P., Cochran, J.K., Bokuniewicz, H.J. and Yang, S., 2008. Submarine
groundwater discharge to Great South Bay, NY, estimated using Ra isotopes. Marine
Chemistry, 109: 279-291.
Bedosky, S., 1987. Recent sediment history of Apalachicola Bay, Florida. unpubl. Masters thesis
Thesis, Florida State University, Tallahassee, 235 pp.
Bienfang, P., 1980. Water quality characteristics of Honokohau Harbor: A subtropical
embayment affected by groundwater intrusion. Pacific Science, 34: 279-291.
Bienfang, P. and Johnson, W., 1980. Planktonic properties of Honokohau Harbor: A nutrientenriched subtropical embayment. Pacific Science, 34: 293-300.
Bienfang, P.K., 1983. The status of water quality, water circulation, and plankton populations in
Honokohau small boat harbor, 1982, A Decade of Ecological Studies Following
Construction of the Honokohau Small Boat Harbor, Kona, Hawaii. US Army Corps of
Engineers, Honolulu, pp. 11-31.
Bode, A. and Dortch, Q., 1996. Uptake and regeneration of inorganic nitrogen in coastal waters
influenced by the Mississippi River: spatial and seasonal variations. Journal of Plankton
Research, 18: 2251-2268.
Bonniwell, E.C., Matisoff, G. and Whiting, P.J., 1999. Determining the times and distances of
particle transit in a mountain stream using fallout radionuclides. Geomorphology, 27: 7592.
Brzezinski, M.A., 1985. The Si:C:N ratio of marine diatoms: Interspecific variability and the
effect of some environmental variables. Journal of Phycology, 21: 347-357.
Burnett, W.C. et al., 2006a. Quantifying submarine groundwater discharge in the coastal zone
via multiple methods. Science of the Total Environment, 367: 498-543.
Burnett, W.C., Bokuniewicz, H., Huettel, M., Moore, W.S. and Taniguchi, M., 2003a.
Groundwater and porewater inputs to the coastal zone. Biogeochemistry, 66: 3-33.
Burnett, W.C., Cable, J.E. and Corbett, D.R., 2003b. Radon tracing of submarine groundwater
discharge in coastal environments. In: M. Taniguchi, K. Wang and T. Gamo (Editors),
Land and marine hydrology. Elsevier, New York, pp. 25-42.
Burnett, W.C. and Dulaiova, H., 2003. Estimating the dynamics of groundwater input into the
coastal zone via continuous radon-222 measurements. Journal of Environmental
Radioactivity, 69: 21-35.
144
Burnett, W.C., Dulaiova, H., Stringer, C.E. and Peterson, R.N., 2006b. Submarine groundwater
discharge: its measurement and influence on the coastal zone. Journal of Coastal
Research, 39: 35-38.
Burnett, W.C., Kim, G. and Lane-Smith, D., 2001. A continuous monitor for assessment of 222Rn
in the coastal ocean. Journal of Radioanalytical and Nuclear Chemistry, 249: 167-172.
Burnett, W.C., Peterson, R.N., Moore, W.S. and de Oliveira, J., 2007a. Radon and radium
isotopes as tracers of submarine groundwater discharge - results from the Ubatuba, Brazil
SGD assessment intercomparison. Estuarine, Coastal and Shelf Science, 76: 501-511.
Burnett, W.C., Santos, I.R., Weinstein, Y., Swarzenski, P.W. and Herut, B., 2007b. Remaining
uncertainties in the use of Rn-222 as a quantitative tracer of submarine groundwater
discharge, IUGG. IAHS Publ. 312, Perugia, Italy, pp. 109-118.
Butts, J., Todd, J.F., Lerche, I., Moore, W.S. and Moore, D.G., 1988. A simplified method for
226
Ra determinations in natural waters. Marine Chemistry, 25: 349-357.
Cable, J.E., Burnett, W.C., Chanton, J.P. and Weatherly, G.L., 1996. Estimating groundwater
discharge into the northeastern Gulf of Mexico using radon-222. Earth and Planetary
Science Letters, 144: 591-604.
Canfield, D.E., 1997. The geochemistry of river particulates from the continental USA: Major
elements. Geochimica et Cosmochimica Acta, 61: 3349-3365.
Carroll, J., Falkner, K.K., Brown, E.T. and Moore, W.S., 1993. The role of the GangesBrahmaputra mixing zone in supplying barium and 226Ra to the Bay of Bengal.
Geochimica et Cosmochimica Acta, 57: 2981-2990.
Charette, M.A., 2007. Hydrologic forcing of submarine groundwater discharge: Insight from a
seasonal study of radium isotopes in a groundwater-dominated salt marsh estuary.
Limnology and Oceanography, 52: 230-239.
Charette, M.A. and Allen, M.C., 2006. Precision groundwater sampling in coastal aquifers using
a direct push shielded-screen wellpoint system. Ground Water Monitoring &
Remediation 26: 87-93.
Charette, M.A., Gonneea, M.E., Morris, P.J., Statham, P., Fones, G., Planquette, H., Salter, I. and
Garabato, A.N., 2007. Radium isotopes as tracers of iron sources fueling a Southern
Ocean phytoplankton bloom. Deep-Sea Research II, 54: 1989-1998.
Charette, M.A., Moore, W.S. and Burnett, W.C., 2008. Uranium- and thorium-series nuclides as
tracers of submarine groundwater discharge. In: S. Krishnaswami and J.K. Cochran
(Editors), U-Th Series Nuclides in Aquatic Systems. Elsevier, Amsterdam, pp. 151-191.
Chen, J., Fukushima, Y. and Taniguchi, M., 2005. Study on sustainability of irrigation
agriculture by diversion in the lower reach of the Yellow River, 5th International
Symposium. Aquifer Recharge, Berlin, Germany, pp. 186-191.
145
Chen, J.Y., Taniguchi, M., Liu, G., Miyaoka, K., Onodera, S.-i., Tokunaga, T. and Fukushima,
Y., 2007. Nitrate pollution of groundwater in the Yellow River delta, China.
Hydrogeology, 15: 1605-1614.
Church, T.M., 1996. An underground route for the water cycle. Nature, 380: 579-580.
Coffey, M., Dehairs, F., Collette, O., Luther, G., Church, T. and Jickells, T., 1997. The
behaviour of dissolved barium in estuaries. Estuarine, Coastal and Shelf Science, 45:
113-121.
Corbett, D.R., McKee, B. and Duncan, D., 2004. An evaluation of mobile mud dynamics in the
Mississippi River deltaic region. Marine Geology, 209: 91-112.
Couch, C.A., Hopkins, E.H. and Hardy, P.S., 1996. Influences of environmental setting on
aquatic ecosystems in the Apalachicola-Chattahoochee-Flint River basin. U.S. Geological
Survey Water Resources Investigations Report 95-4278, U.S. Geological Survey.
Councell, T.B., Duckenfield, K.U., Landa, E.R. and Callender, E., 2004. Tire-wear particles as a
source of zinc to the environment. Environmental Science and Technology, 38: 42064214.
Currie, L.A., 1968. Limits for qualitative detection and quantitative determination. Analytical
Chemistry, 40: 586-593.
Dimova, N., Burnett, W.C., Horwitz, E.P. and Lane-Smith, D., 2007. Automated measurement of
224
Ra and 226Ra in water. Applied Radiation and Isotopes, 65: 428-434.
Dimova, N., Dulaiova, H., Kim, G. and Burnett, W.C., 2008. Uncertainties in the preparation of
224
Ra Mn fiber standards. Marine Chemistry, 109: 220-225.
Donoghue, J.F., 1988. Evaluation of sediment loading processes in the Apalachicola Bay
Estuary, NOAA Tech. Mem. NOS-MEMD-17.
Donoghue, J.F., Bricker, O.P. and Olsen, C.R., 1989. Particle-borne radionuclides as tracers for
sediment in the Susquehanna River and Chesapeake Bay. Estuarine, Coastal and Shelf
Science, 29: 341-360.
Dulaiova, H. and Burnett, W.C., 2004. An efficient method for γ-spectrometric determination of
radium-226,228 via manganese fibers. Limnology and Oceanography: Methods, 2: 256261.
Dulaiova, H. and Burnett, W.C., 2008. Evaluation of the flushing rates of Apalachicola Bay,
Florida via natural geochemical tracers. Marine Chemistry, 109: 395-408.
Dulaiova, H., Burnett, W.C., Wattayakorn, G. and Sojisuporn, P., 2006. Are groundwater inputs
into river-dominated areas important? The Chao Phraya River - Gulf of Thailand.
Limnology and Oceanography, 51: 2232-2247.
146
Dulaiova, H., Peterson, R.N., Burnett, W.C. and Lane-Smith, D., 2005. A multi-detector
continuous monitor for assessment of 222Rn in the coastal ocean. Journal of
Radioanalytical and Nuclear Chemistry, V263: 361-365.
Dupre, B., Gaillardet, J., Rousseau, D. and Allegre, C.J., 1996. Major and trace elements of
river-borne material: The Congo Basin. Geochimica et Cosmochimica Acta, 60: 13011321.
Elder, J.F., Flagg, S.D. and Mattraw, H.C., Jr., 1988. Hydrology and ecology of the Apalachicola
River floodplain, Florida. U.S. Geological Survey Water-Supply Paper 2196-D46.
Fanning, K.A. and Pilson, M.E.Q., 1973. On the spectrophotometric determination of dissolved
silica in natural waters. Analytical Chemistry, 45: 136-140.
Freyer, K., Treutler, H.C., Just, G. and Philipsborn, H.v., 2003. Optimization of time resolution
and detection limit for online measurements of 222Rn in water. Journal of Radioanalytical
and Nuclear Chemistry, 257: 129-132.
Froelich, P.N., 1988. Kinetic control of dissolved phosphate in natural rivers and estuaries: A
primer on the phosphate buffer mechanism. Limnology and Oceanography, 33: 649-668.
Gaillardet, J., Dupre, B. and Allegre, C.J., 1999. Geochemistry of large river suspended
sediments: silicate weathering or recycling tracer? Geochimica et Cosmochimica Acta,
63: 4037-4051.
Gallagher, B., 1980. Physical structure and circulation in Honokohau, a small Hawaiian harbor
affected by groundwater. Pacific Science, 34: 301-311.
Gordon, D.C.J., Boudreau, P.R., Mann, K.H., Ong, J.-E., Silvert, W.L., Smith, S.V.,
Wattayakorn, G., Wulff, F. and Yanagi, T., 1996. LOICZ Biogeochemical Modelling
Guidelines, LOICZ/R&S/95-5, VI. LOICZ, Texel, The Netherlands, 96 pp.
Gorsline, D.S., 1963. Oceanography of Apalachicola Bay, Florida. Ph.D. Thesis Thesis, Florida
State Univ., Tallahassee, 96 pp.
Griffin, C., Kaufman, A. and Broecker, W., 1963. Delayed coincidence counter for the assay of
actinon and thoron. Journal of Geophysical Research, 68: 1749-1757.
Hainbucher, D., Hao, W., Pohlmann, T., Sündermann, J. and Feng, S., 2004. Variability of the
Bohai Sea circulation based on model calculations. Journal of Marine Systems, 44: 153174.
Hancock, G.J., Webster, I.T. and Stieglitz, T.C., 2006. Horizontal mixing of Great Barrier Reef
waters: Offshore diffusivity determined from radium isotope distribution. Journal of
Geophysical Research, 111.
147
Hanor, J.S. and Chan, L.-H., 1977. Non-conservative behavior of barium during mixing of
Mississippi River and Gulf of Mexico waters. Earth and Planetary Science Letters, 37:
242-250.
Harrington, D.J., 2001. Effect of river stage variations and flood events on trace metal
mobilization and transport in the Apalachicola River basin, NW Florida. unpubl. master's
thesis Thesis, Florida State University, Tallahassee, 689 pp.
Hodell, D.A., Mead, G.A. and Mueller, P.A., 1990. Variation in the strontium isotopic
composition of seawater (8 MA to present) - Implications for chemical-weathering rates
and dissolved fluxes to the oceans. Chemical Geology, 80: 291-307.
Hope, D., Billett, M.F. and Cresser, M.S., 1994. A review of the export of carbon in river water Fluxes and processes. Environmental Pollution, 84: 301-324.
Horowitz, A.Z., 1995. The use of suspended sediment and associated trace elements in water
quality studies. International Association of Hydrologic Scientific Specific Publications,
4: 58.
Hu, C., Muller-Karger, F.E. and Swarzenski, P., 2006. Hurricanes, submarine groundwater
discharge, and Florida's red tides. Geophysical Research Letters, 33:
doi:10.1029/2005GL025449.
Huang, H.J., Li, F., Pang, J.Z., Le, K.T. and Li, S.G., 2005. Land and ocean interaction of the
Yellow River delta and Bohai Sea and Yellow Sea. Science Press of China187-207.
Hwang, D.-W., Kim, G., Lee, Y.-W. and Yang, H.-S., 2005. Estimating submarine inputs of
groundwater and nutrients to a coastal bay using radium isotopes. Marine Chemistry, 96:
61-71.
Isphording, W.C., 1985. Sedimentological investigations of the Apalachicola Bay, Florida
estuarine system: prepared for the Mobile District, US Army Corps of Engineers, BER
Report no. 343-60.
Johnson, A.J., 2008. Groundwater discharge from the leeward half of the Big Island, Hawaii.
M.S. Thesis Thesis, University of Hawaii, Honolulu.
Johnson, A.J., Glenn, C.R., Burnett, W.C., Peterson, R.N. and Lucey, P.G., 2008. Aerial infrared
imaging reveals large nutrient-rich groundwater inputs to the ocean. Geophysical
Research Letters, 35.
Kanehiro, B.Y. and Peterson, F.L., 1977. Groundwater recharge and coastal discharge for the
northwest coast of the island of Hawai'i: A computerized water budget approach.
Technical Report No. 110, Water Resources Research Center: University of Hawaii.,
Honolulu.
Kaul, L.W. and Froelich, J., Philip N., 1984. Modeling estuarine nutrient geochemistry in a
simple system. Geochimica et Cosmochimica Acta, 48: 1417-1433.
148
Kay, E.A., Lau, L.S., Stroup, E.D., Dollar, S.J., Fellows, D.P. and Young, R.H.F., 1977.
Hydrologic and ecological inventory of coastal waters of west Hawaii. Technical Report
No. 105., Water Resources Center, University of Hawaii.
Key, R.M., Brewer, R.L., Stockwell, J.H., Guinasso, N.L. and Schink, D.R., 1979. Some
improved techniques for measuring radon and radium in marine sediments and in
seawater. Marine Chemistry, 7: 251-264.
Kim, G., Burnett, W.C., Dulaiova, H., Swarzenski, P. and Moore, W.S., 2001. Measurement of
224
Ra and 226Ra activities in natural waters using a radon-in-air monitor. Environmental
Science and Technology, 35: 4680-4683.
Kim, G., Ryu, J.-W., Yang, H.-S. and Yun, S.-T., 2005. Submarine groundwater discharge
(SGD) into the Yellow Sea revealed by 228Ra and 226Ra isotopes: Implications for global
silicate fluxes. Earth and Planetary Science Letters, 237: 156-166.
Kim, K.H. and Burnett, W.C., 1983. gamma-ray Spectrometric determination of uranium-series
nuclides in marine phosphorites. Analytical Chemistry, 55: 1796-1800.
Koch, D.M., Jacob, D.J. and Graustein, W.C., 1996. Vertical transport of tropospheric aerosols
as indicated by 7Be and 210Pb in a chemical tracer model. Journal of Geophysical
Research, 101: 18651-18666.
Kofoed, J.W. and Gorsline, D.S., 1963. Sedimentary environments in Apalachicola Bay and
viscinity, Florida. Journal of Sedimentary Petrology, 33: 205-223.
Kohout, F.A., 1966. Submarine springs: A neglected phenomenon of coastal hydrology.
Hydrology, 26: 391-413.
Krest, J.M., Moore, W.S. and Rama, 1999. 226Ra and 228Ra in the mixing zones of the
Mississippi and Atchafalaya Rivers: indicators of groundwater input. Marine Chemistry,
64: 129-152.
Lee, J.-M. and Kim, G., 2006. A simple and rapid method for analyzing radon in coastal and
ground waters using a radon-in-air monitor. Journal of Environmental Radioactivity, 89:
219-228.
Leitman, H.M., Sohm, J.E. and Franklin, M.A., 1984. Flood patterns and tree distribution on the
Apalachicola River floodplain, Florida. U.S. Geological Survey Water-Supply Paper
2196-A52.
Lesley, M.P., 2002. The fluxes and fates of arsenic, selenium, and antimony from coal fired
power plants to rivers. unpubl. master's thesis Thesis, Georgia Institute of Technology,
Atlanta, 144 pp.
Lesley, M.P. and Froelich, J., Philip N., 2003. Arsenic, selenium and antimony from coal-fired
power plants to the Chattahoochee River. In: H.J. Hatcher (Editor), Proceedings of the
2003 Georgia Water Resources Conference, University of Georgia, Athens, pp. 473-476.
149
Li, Y.-H. and Chan, L.-H., 1979. Desorption of Ba and 226Ra from river-borne sediments in the
Hudson estuary. Earth and Planetary Science Letters, 43: 343-350.
Liu, H. and Yin, B., 2007. Annual cycle of carbon, nitrogen and phosphorus in the Bohai Sea: A
model study. Continental Shelf Research, 27: 1399-1407.
Lucas, H.F., 1957. Improved low-level alpha-scintillation counter for radon. Review of Scientific
Instruments, 28: 680-683.
MacIntyre, S., Wanninkhof, R. and Chanton, J.P., 1995. Trace gas exchange across the air-sea
interface in freshwater and coastal marine environments. In: P.A. Matson and R.C. Harris
(Editors), Biogenic Trace Gases: Measuring Emissions from Soil and Water. Blackwell
Science Ltd, Boston, pp. 52-97.
Mann, D.R., Surprenant, L.D. and Casso, S.A., 1984. In-situ chemisorption of transuranics from
seawater. Nuclear Instruments and Methods in Physics Research, 223: 235-238.
Martin, J.-M. and Meybeck, M., 1979. Elemental mass-balance of material carried by major
world rivers. Marine Chemistry, 7: 173-206.
Martin, P. and Akber, R.A., 1999. Radium isotopes as indicators of adsorption-desorption
interactions and barite formation in groundwater. Journal of Environmental
Radioactivity, 46: 271-286.
Mathieu, G.G., Biscaye, P.E., Lupton, R.A. and Hammond, D.E., 1988. System for measurement
of 222Rn at low levels in natural waters. Health Physics, 55: 989-992.
Matisoff, G., Wilson, C.G. and Whiting, P.J., 2005. The 7Be/210Pbxs ratio as an indicator of
suspended sediment age or fraction new sediment in suspension. Earth Surface Processes
and Landforms, 30: 1191-1201.
Meybeck, M., Laroche, L., Durr, H.H. and Syvitski, J.P.M., 2003. Global variability of daily
total suspended solids and their fluxes in rivers. Global and Planetary Change, 39: 65-93.
Michael, H.A., Mulligan, A.E. and Harvey, C.F., 2005. Seasonal oscillations in the water
exchange between aquifers and the coastal ocean. Letters to Nature, 436: 1145-1148.
Milliman, J.D. and Meade, R.H., 1983. World-wide delivery of river sediment to the oceans.
Journal of Geology, 91: 1-21.
Milliman, J.M. and Syvitski, J.P.M., 1992. Geomorphic/tectonic control of sediment discharge to
the ocean: the importance of small mountainous rivers. Journal of Geology, 100: 525544.
Moon, D.S., Burnett, W.C., Nour, S., Horwitz, E.P. and Bond, A., 2003. Preconcentration of
radium isotopes from natural waters using MnO2 resin. Applied Radiation and Isotopes,
59: 255-262.
150
Moore, W.S., 1976. Sampling 228Ra in the deep ocean. Deep Sea Research and Oceanographic
Abstracts, 23: 647-651.
Moore, W.S., 1984. Radium isotope measurements using germanium detectors. Nuclear
Instruments and Methods in Physics Research, 223: 407-411.
Moore, W.S., 1996. Large groundwater inputs to coastal waters revealed by 226Ra enrichments.
Nature, 380: 612-614.
Moore, W.S., 1997. High fluxes of radium and barium from the mouth of the GangesBrahmaputra River during low river discharge suggest a large groundwater source. Earth
and Planetary Science Letters, 150: 141-150.
Moore, W.S., 1999. The subterranean estuary: a reaction zone of ground water and sea water.
Marine Chemistry, 65: 111-125.
Moore, W.S., 2000a. Ages of continental shelf waters determined from 223Ra and 224Ra. Journal
of Geophysical Research, 105: 117-122.
Moore, W.S., 2000b. Determining coastal mixing rates using radium isotopes. Continental Shelf
Research, 20: 1993-2007.
Moore, W.S., 2007. Fifteen years experience in measuring 224Ra and 223Ra by delayedcoincidence counting. Marine Chemistry, 109: 188-197.
Moore, W.S. and Arnold, R., 1996. Measurement of 223Ra and 224Ra in coastal waters using a
delayed coincidence counter. Journal of Geophysical Research, 101: 1321-1329.
Moore, W.S., Key, R.M. and Sarmiento, J.L., 1985. Techniques for precise mapping of 226Ra and
228
Ra in the ocean. Journal of Geophysical Research, 90: 6983-6994.
Moore, W.S. and Krest, J., 2004. Distribution of 223Ra and 224Ra in the plumes of the
Mississippi and Atchafalaya Rivers and the Gulf of Mexico. Marine Chemistry, 86: 105119.
Moore, W.S. and Reid, D.F., 1973. Extraction of radium from natural waters using manganeseimpregnated acrylic fibers. Journal of Geophysical Research, 78: 8880-8886.
Mortlock, R.A., Froelich, P.N., Feely, R.A., Massoth, G.J., Butterfield, D.A. and Lupton, J.E.,
1993. Silica and germanium in Pacific Ocean hydrothermal vents and plumes. Earth and
Planetary Science Letters, 119: 365-378.
Mulligan, A.E. and Charette, M.A., 2006. Intercomparison of submarine groundwater discharge
estimates from a sandy unconfined aquifer. Journal of Hydrology, 327: 411-425.
Nozaki, Y., Yamamoto, Y., Manaka, T., Amakawa, H. and Snidvongs, A., 2001. Dissolved
barium and radium isotopes in the Chao Phraya River estuarine mixing zone in Thailand.
Continental Shelf Research, 21: 1435-1448.
151
Olley, J.M., Murray, A.S., MacKenzie, D.H. and Edwards, K., 1993. Identifying sediment
sources in a gullied catchment using natural and anthropogenic radioactivity. Water
Resources Research, 29: 1037-1043.
Pennock, J.R., 1987. Temporal and spatial variability in phytoplankton ammonium and nitrate
uptake in the Delaware estuary. Estuarine, Coastal and Shelf Science, 24: 841-857.
Peterson, R.N., Burnett, W.C., Glenn, C.R. and Johnson, A.J., 2007. A box model to quantify
groundwater discharge along the Kona coast of Hawaii using natural tracers, IUGG
XXIV Earth: Our Changing Planet. IAHS Press, Perugia, Italy.
Peterson, R.N., Burnett, W.C., Taniguchi, M., Chen, J., Santos, I.R. and Ishitobi, T., 2008a.
Radon and radium isotope assessment of submarine groundwater discharge in the Yellow
River delta, China. Journal of Geophysical Research, 113: C09021.
Peterson, R.N., Burnett, W.C., Taniguchi, M., Chen, J.Y., Santos, I.R. and Misra, S., 2008b.
Determination of transport rates in the Yellow River - Bohai Sea mixing zone via natural
geochemical tracers. Continental Shelf Research, 28: 2700-2707.
Pfitzner, J., Brunskill, G. and Zagorskis, I., 2004. 137Cs and excess 210Pb deposition patterns in
estuarine and marine sediment in the central region of the Great Barrier Reef Lagoon,
north-eastern Australia. Journal of Environmental Radioactivity, South Pacific
Environmental Radioactivity Association: 2002 Conference, 76: 81-102.
Price, N.M., Cochlan, W.P. and Harrison, P.J., 1985. Time course of uptake of organic and
inorganic nitrogen by phytoplankton in the Strait of Georgia: comparison of frontal and
stratified communities. Marine Ecology Progress Series, 27: 39-53.
Raabe, T., Yu, Z., Zhang, J., Sun, J., Starke, A., Brockmann, U. and Hainbucher, D., 2004.
Phase-transfer of nitrogen species within the water column of the Bohai Sea. Journal of
Marine Systems, 44: 213-232.
Robert, A., 2003. River Processes: An Introduction to Fluvial Dynamics. Oxford University
Press Inc., New York, 214 pp.
Rozengurt, M.A. and Haydock, I., 1991. Effects of fresh water development and water pollution
policies on the world's river - delta - estuary - coastal zone ecosystems. In: H.S. Bolton
(Editor), Coastal Wetlands, Coastal Zone '91, Seventh Symposium on Coastal and Ocean
Management. American Society of Civil Engineers, New York, pp. 85-98.
Rudnick, R.L. and Gao, S., 2003. Composition of the continental crust. In: R.L. Rudnick
(Editor), Treatise on Geochemistry. Elsevier, New York, pp. 56.
Salant, N.L., Renshaw, C.E., Magilligan, F.J., Kaste, J.M., Nislow, K.H. and Heimsath, A.M.,
2007. The use of short-lived radionuclides to quantify transitional bed material transport
in a regulated river. Earth Surface Processes and Landforms, 32: 509-524.
152
Santos, I.R., Burnett, W.C., Chanton, J.P., Mwashote, B., Suryaputra, I.G.N.A. and Dittmar, T.,
2008. Nutrient biogeochemistry in a Gulf of Mexico subterranean estuary and
groundwater-derived fluxes to the coastal ocean. Limnology and Oceanography, 53: 705718.
Schmidt, J.C. and Wilcock, P.R., 2008. Metrics for assessing the downstream effects of dams.
Water Resources Research, 44: 1-19.
Schubert, M., Schmidt, A., Paschke, A., Lopez, A. and Balcazar, M., 2008. In-situ determination
of radon in surface water bodies by means of a hydrophobic membrane tubing. Radiation
Measurements, 43: 111-120.
Shen, Z.L. and Le, K.T., 1993. Effects of the Yellow River estuary location changes on its
hydrochemical environment. Studia Marina Sinica, 34.
Sheppard, M.I. and Thibault, D.H., 1990. Default soil solid/liquid partition coefficients, KdS, for
four major soil types: A compendium. Health Physics, 59: 471-482.
Shiller, A.M., 1993. A mixing rate approach to understanding nutrient distributions in the plume
of the Mississippi River. Marine Chemistry, 43: 211-216.
Sill, C.W., 1987. Determination of radium-226 in ores, nuclear wastes and environmental
samples by high-resolution alpha spectrometry. Nuclear and Chemical Waste
Management, 7: 239-254.
Slomp, C.P. and Van Cappellen, P., 2004. Nutrient inputs to the coastal ocean through submarine
groundwater discharge: controls and potential impact. Journal of Hydrology, 295: 64-86.
Sommerfield, C.K. and Nittrouer, C.A., 1999. Modern accumulation rates and a sediment budget
for the Eel Shelf: a flood-dominated depositional environment. Marine Geology, 154:
227-241.
Sommerfield, C.K., Nittrouer, C.A. and Alexander, C.R., 1999. 7Be as a tracer of flood
sedimentation on the northern California continental margin. Continental Shelf Research,
19: 335-361.
Spiteri, C., Slomp, C.P., Tuncay, K. and Meile, C., 2008. Modeling biogeochemical processes in
subterranean estuaries: Effect of flow dynamics and redox conditions on submarine
groundwater discharge of nutrients. Water Resources Research, 44: W02430.
Street, J.H., Knee, K.L., Grossman, E.E. and Paytan, A., 2007. Submarine groundwater
discharge and nutrient addition to the coastal zone and coral reefs of leeward Hawai'i.
Marine Chemistry, 109: 355-376.
Stringer, C.E. and Burnett, W.C., 2004. Sample bottle design improvements for radon emanation
analysis of natural waters. Health Physics, 87: 642-646.
153
Sun, Y. and Torgersen, T., 1998. The effects of water content and Mn-fiber surface conditions on
224
Ra measurement by 220Rn emanation. Marine Chemistry, 62: 299-306.
Surratt, D., Cherrier, J., Robinson, L. and Cable, J., 2008. Chronology of sediment nutrient
geochemistry in Apalachicola Bay, Florida (U.S.A.). Journal of Coastal Research, 24:
660-671.
Swarzenski, P.W., 2007. U/Th series radionuclides as coastal groundwater tracers. Chemical
Reviews, 107: 663-674.
Swarzenski, P.W., Reich, C., Kroeger, K.D. and Baskaran, M., 2007a. Ra and Rn isotopes as
natural tracers of submarine groundwater discharge in Tampa Bay, FL. Marine
Chemistry, 104: 69-84.
Swarzenski, P.W., Simonds, F.W., Paulson, A.J., Kruse, S. and Reich, C., 2007b. Geochemical
and geophysical examination of submarine groundwater discharge and associated nutrient
loading estimates into Lynch Cove, Hood Canal, WA. Environmental Science and
Technology, accepted.
Taniguchi, M., Ishitobi, T., Chen, J., Onodera, S.-i., Miyaoka, K., Burnett, W.C., Peterson, R.N.,
Liu, G. and Fukushima, Y., 2008. Submarine groundwater discharge from the Yellow
River delta to the Bohai Sea, China. Journal of Geophysical Research, 113.
Taniguchi, M. and Iwakawa, T., 2001. Measurements of submarine groundwater discharge rates
by a continuous heat - type automated seepage meter in Osaka Bay, Japan. Journal of
Groundwater Hydrology, 42,: 271-277.
Walling, D.E., Owens, P.N. and Leeks, G.J.L., 1999. Fingerprinting suspended sediment sources
in the catchment of the River Ouse, Yorkshire, UK. Hydrological Processes, 13: 955-975.
Wang, Q., Guo, X. and Kidetaka, T., 2007. A numerical study on the seasonal variation of
Yellow River plume path in the Bohai Sea, Proceedings of 3rd International Workshop
on Yellow River Studies. Research Institute for Humanity and Nature, Kyoto, Japan, pp.
7-9.
Wang, S., Hassan, M.A. and Xie, X., 2006. Relationship between suspended sediment load,
channel geometry and land area increment in the Yellow River Delta. CATENA, 65: 302314.
Waska, H., Kim, S., Kim, G., Peterson, R. and Burnett, W.C., 2008. An efficient and simple
method for measuring 226Ra, together with 223Ra and 224Ra, using a delayed coincidence
counter (RaDeCC). Journal of Environmental Radioactivity, in press.
Wei, H., Sun, J., Moll, A. and Zhao, L., 2004. Phytoplankton dynamics in the Bohai Sea-observations and modelling. Journal of Marine Systems, 44: 233-251.
Wilhelm, R.G., 2004. Understanding the variation in partition coefficients, Kd, values. EPA 40204-002C, United States Environmental Protection Agency.
154
Wilson, C.G., Matisoff, G. and Whiting, P.J., 2007. The use of 7Be and 210Pbxs to differentiate
fine suspended sediment sources in South Slough, Oregon. Estuaries and Coasts, 30: 348358.
Wu, K., Xie, X.Q. and Tang, D.Y., 1998. The causes of formation, the regularities, the effect
estimation to periphery agricultural production and the ecological environment and the
countermeasures of the absence of flow in the Huanghe River (in Chinese). Progress in
Geography, 17 (Suppl): 78-84.
Yu, L., 2002. The Huanghe (Yellow) River: a review of its development, characteristics, and
future management issues. Continental Shelf Research, 22: 389-403.
Yu, L., 2006. The Huanghe (Yellow) River: Recent changes and its countermeasures.
Continental Shelf Research, 26: 2281-2298.
Zektser, I.S., 2000. Groundwater and the Environment: Applications for the Global Community.
Lewis Publishers, Boca Raton, 175 pp.
Zhang, J., Yu, Z.G., Raabe, T., Liu, S.M., Starke, A., Zou, L., Gao, H.W. and Brockmann, U.,
2004. Dynamics of inorganic nutrient species in the Bohai seawaters. Journal of Marine
Systems, 44: 189-212.
155
BIOGRAPHIC SKETCH
Richard N. Peterson
Personal Information
Date of birth: December 10, 1982 in Fort Pierce, Florida
Languages: English and Spanish
Education
May ‘04 – present
Ph.D. Candidate (Oceanography)
Florida State University, Department of Oceanography
Research Advisor: Dr. William Burnett
Aug. ‘01 – Apr. ’04
Bachelors of Science Degree (Chemistry, Magna cum Laude)
Florida State University, Department of Chemistry
Minor: Mathematics
Aug. ’97 – May ’01
International Baccalaureate Diploma (Salutatorian)
Lincoln Park Academy, Fort Pierce Florida
Professional Positions
Oct. ’01 – May ’04
Research Laboratory Assistant
Environmental Radioactivity Measurement Facility
Department of Oceanography, Florida State University
Aug. ’00 – July ’01
Laboratory Technician
University of Florida Indian River Research and Education Center
Jan. ’00 – July ’00
Laboratory Technician
Department of Chemistry, Harbor Branch Oceanographic Institution
Volunteer Experience
Aug. ’01 – Aug. ’02 Volunteered over 120 hours in the Evidence section
Florida Department of Law Enforcement Crime Laboratory
Tallahassee, Florida
Academic Awards and Honors
2008
Invited to attend Dissertations Symposium on Chemical
Oceanography (DISCO XXI) in Honolulu, Hawaii in October 2008
156
2007
Chosen as a Featured Student from several thousand nominations to be
featured at www.fsu.edu
2006 – 2007
Named Outstanding Graduate Student in the Department of Oceanography
2005
Awarded a 3-year Graduate Research Fellowship through NOAA
2004
Induction into Pi Beta Kappa Honor Society
Research Grants
Leading Investigator, NOAA Graduate Research Fellowship Program
“Origin and fate of suspended particulates in the Apalachicola River: Impact on
Apalachicola Bay”, 2005-2008, $85,716
Oral Scientific Presentations
Oct. 27-31, ’08
“Comparing Measurement Methods for 226Ra on Mn-fiber”
54th Radiobioassay & Radiochemical Measurement Conference
Sandestin, Florida
Oct. 5-10, ’08
Invited Presentation: “Point-Source Groundwater Discharges from
Leeward Hawaii”
DISCO Symposium 2008
Honolulu, Hawaii
Oct. 1-3, ’08
“Bohai Sea Coastal Transport Rates and Their Influence on Coastline
Nutrient Inputs”
HydroChange 2008
Kyoto, Japan
Apr. 18, ’08
Invited Presentation: “Applications of Naturally-Occurring
Radionuclides in Coastal Oceanography”
Florida Chapter – Health Physics Society Spring Meeting
Gainesville, Florida
Apr. 7-11, ’08
“Just What IS the Best Method for Measuring 226Ra on MnFibers?”
2008 Ra-Rn Workshop
Venice, Italy
157
Nov. 14, ’07
“Water Budget Effects on Estuaries and the Coastal Zone from the
Recent Drought in Florida”
Florida Oceans and Coastal Council
Harbor Branch Oceanographic Institution
July 2-3, ’07
“A Box Model to Quantify Groundwater Discharge Along the Kona Coast
of Hawaii Using Natural Tracers”
International Union of Geodesy and Geophysics (IUGG) XXIV General
Assembly – Earth, Our Changing Planet
Perugia, Italy
June 25, ‘07
“A Box Model to Quantify Groundwater Discharge Along the Kona Coast
of Hawaii Using Natural Tracers”
Woods Hole Oceanographic Institution
Woods Hole, Massachusetts
Feb. 13-15, ’07
“Analysis of Yellow River Mixing Processes into the Bohai Sea via
Barium and Radium Isotopes”
The 3rd International Workshop on Yellow River Studies
Kyoto, Japan
Nov. 6-11, ’04
“Exchange in the Yellow River / Estuary / Bohai Sea System via
Radium Isotopes”
The 2nd International Workshop on Yellow River Studies
Kyoto, Japan
Peer-Reviewed Publications
Peterson, R.N., W.C. Burnett, N. Dimova, and I.R. Santos, 2008. Comparing measurement
methods for radium-226 on manganese-fiber. Limnology & Oceanography: Methods, in
press.
Santos, I.R., W.C. Burnett, J. Chanton, N. Dimova, and R.N. Peterson, 2008. Land or ocean?
Assessing the driving forces of submarine groundwater discharge. Journal of
Geophysical Research, in press.
Santos, I.R., N. Dimova, R.N. Peterson, B. Mwashote, J. Chanton, and W.C. Burnett, 2008.
Extended time series measurements of submarine groundwater discharge tracers (222Rn
and CH4) at a coastal site in Florida. Marine Chemistry, in press.
Peterson, R.N., W.C. Burnett, C.R. Glenn, and A.G. Johnson, 2008. Quantification of coastal
water fluxes from point-source groundwater discharges from western Hawaii. Limnology
and Oceanography, in press.
158
Peterson, R.N., W.C. Burnett, I.R. Santos, M. Taniguchi, T. Ishitobi, and J. Chen, 2009. Bohai
Sea coastal transport rates and their influence on coastline nutrient inputs. In, Taniguchi,
M., W.C. Burnett, Y. Fukushima, M. Haigh, and Y. Umezawa (eds.) From Headwaters to
the Ocean, Taylor & Francis, London, 659-664.
W.C. Burnett, R. Peterson, M. Taniguchi, G. Wattayakorn, S. Chanyotha, and F. Siringan, 2009.
Importance of groundwater discharge in developing urban centers of Southeast Asia. In,
Taniguchi, M., W.C. Burnett, Y. Fukushima, M. Haigh, and Y. Umezawa (eds.) From
Headwaters to the Ocean, Taylor & Francis, London, 289-294.
Waska, H., S. Kim, G. Kim, R. Peterson, and W.C. Burnett, 2008. An efficient and simple
method for measuring 226Ra, together with 223Ra and 224Ra, using a delayed coincidence
counter (RaDeCC). Journal of Environmental Radioactivity, 99 (12), 1859-1862.
Peterson, R.N., W.C. Burnett, M. Taniguchi, J. Chen, I.R. Santos, and S. Misra, 2008.
Determination of transport rates in the Yellow River – Bohai Sea mixing zone via natural
geochemical tracers. Continental Shelf Research, 28 (19), 2700-2707.
Povinec, P.P., H. Bokuniewicz, W.C. Burnett, J. Cable, M. Charette, J.-F. Comanducci, E.A.
Kontar, W.S. Moore, J.A. Oberdorfer, J. de Oliveira, R. Peterson, T. Stieglitz, and M.
Taniguchi, 2008. Isotope tracing of submarine groundwater discharge offshore Ubatuba,
Brazil: Results of the IAEA-UNESCO SGD project. Journal of Environmental
Radioactivity, 99, 1596-1610, doi:10.1016/j.jenvrad.2008.06.010.
Peterson, R.N., W.C. Burnett, M. Taniguchi, J. Chen, I.R. Santos, and T. Ishitobi, 2008. Radon
and radium isotope assessment of submarine groundwater discharge in the Yellow River
delta, China. Journal of Geophysical Research, 113, C09021,
doi:10.1029/2008JC004776.
Johnson, A.G., C.R. Glenn, W.C. Burnett, R.N. Peterson, and P.G. Lucey, 2008. Aerial infrared
mapping of nutrient-rich groundwater plumes in Hawaiian coastal waters. Geophysical
Research Letters, 35, L15606, doi: 10.1029/2008GL034574.
Tanighchi, M., T. Ishitobi, J. Chen, S. Onodera, K. Miyaoka, W.C. Burnett, R. Peterson, G. Liu,
and Y. Fukushima, 2008. Submarine groundwater discharge from the Yellow River
Delta to the Bohai Sea, China. Journal of Geophysical Research, 113, C06025,
doi:10.1029/2007JC004498.
Santos, I.R., L.F. Niencheski, W. Burnett, R. Peterson, J. Chanton, C.F.F. Andrade, I.B. Milani,
A. Schmidt, and K. Knoeller, 2008. Tracing anthropogenically-driven groundwater
discharge into a coastal lagoon from southern Brazil. Journal of Hydrology, 353 (3-4),
275-293.
Santos, I.R., M.I. Machado, L.F. Niencheski, W. Burnett, I.B. Milani, C.F.F. Andrade, R.
Peterson, J. Chanton, and P. Baisch, 2008. Major ion chemistry in a freshwater coastal
159
lagoon from Southern Brazil (Mangueira Lagoon): Influence of groundwater inputs.
Aquatic Geochemistry, 14, 133-146, DOI: 110.1007/s10498-10008-19029-10490.
Peterson, R.N., W.C. Burnett, C.R. Glenn, and A.J. Johnson, 2007. A box model to quantify
groundwater discharge along the Kona coast of Hawaii using natural tracers. In: Sanford,
W., C. Langevin, M. Polemio, and P. Povinec (eds.), A New Focus on GroundwaterSeawater Interactions. IAHS Publication 312, Oxfordshire, UK., 142-149.
Burnett, W.C., R. Peterson, W.S. Moore, and J. de Oliveira, 2007. Radon and Radium Isotopes
as Tracers of Submarine Groundwater Discharge - Results from the Ubatuba, Brazil SGD
Assessment Intercomparison. Estuarine, Coastal and Shelf Science, Special Issue 76,
501-511.
Burnett, W.C., H. Dulaiova, C. Stringer, and R. Peterson, 2006. Submarine groundwater
discharge: its measurement and influence on the coastal zone. Journal of Coastal
Research, Special Issue 39, p. 35-38.
Swarzenski, P.W., W.C. Burnett, W.J. Greenwood, B. Herut, R. Peterson, N. Dimova, Y.
Shalem, Y. Yechieli, and Y. Weinstein, 2006. Combined time-series resistivity and
geochemical tracer techniques to examine submarine groundwater discharge at Dor
Beach, Israel. Geophysical Research Letters, 33, L24405, doi:10.1029/2006GL028282.
Dulaiova, H., R. Peterson, W.C. Burnett, and D. Lane-Smith, 2005. A multi-detector
continuous monitor for assessment of 222Rn in the coastal ocean. Journal of
Radioanalytical and Nuclear Chemistry, 263 (2), 361-365.
Submitted
Burnett, W.C., R.N. Peterson, I.R. Santos, and R.W. Hicks. Use of automated radon
measurements for rapid assessment of groundwater flow into Florida streams. Journal of
Hydrology, submitted February 2009.
In Preparation
Peterson, R.N., P.N. Froelich, W.C. Burnett, S. Misra, and I.R. Santos, 2008. Applications in
suspended particle tracing using naturally-occurring radionuclides and other chemical
tracers – Results from the Apalachicola-Chattahoochee-Flint River System. In
preparation for submission to: Geochimica et Cosmochimica Acta.
Other Publications
Santos, I.R., L.F. Niencheski, R. Peterson, W. Burnett, J. Chanton, C. Amdrade, and I. Milani,
2007. Groundwater discharge into a coastal lagoon in southern Brazil: Evidence from
geochemical tracers. XII Congresso Latino-Americano de Ciencias do Mar – XII
COLACMAR.
160
Andrade, C., L.F. Niencheski, I.R. Santos, R. Peterson, W. Burnett, J. Chanton, and I. Milani,
2007. Influência de aportes subterrâneos nas concentraçöes de nutrients dissolvidos na
Lagoa Mangueira (RS- Brasil). XII Congresso Latino-Americano de Ciencias do Mar –
XII COLACMAR.
Burnett, W.C., R. Peterson, I. Santos, M. Taniguchi, and T. Ishitobi, 2007. Determination of
submarine groundwater discharge (SGD) via natural radionuclides in a region near the
mouth of the Yellow River. In: Proceedings of 3rd International Workshop on Yellow
River Studies. Research Institute for Humanity and Nature. Kyoto, Japan, p. 44-47.
Peterson, R.N., W.C. Burnett, I.R. Sanots, S. Misra, and M. Taniguchi, 2007. Analysis of
Yellow River mixing processes into the sea via barium and radium isotopes. In:
Proceedings of 3rd International Workshop on Yellow River Studies. Research Institute
for Humanity and Nature. Kyoto, Japan, p. 40-43.
Peterson, R., and W.C. Burnett, 2004. Exchange in the Yellow River / Estuary / Bo-Hai Sea
System via Radium Isotopes (abs.). Research Institute for Humanity & Nature (RIHN)
Proc. International Workshop on the Yellow River Project, Kyoto, Japan.
Swarzenski, P., B. Burnett, C. Reich, H. Dulaiova, R. Peterson, and J. Meunier, 2004. Novel
geophysical and geochemical techniques used to study submarine groundwater discharge
in Biscayne Bay, Florida. USGS Fact Sheet 2004-3117.
Peterson, R., 2003. Lake Sevan Sediments. Published in Armenia in Russian as an informative
publication about Lake Sevan.
Miscellaneous
Reviewer for the following scientific journals (in alphabetical order):
1. Ground Water
2. Radiation Measurements
3. Taniguchi, M., W.C. Burnett, Y. Fukushima, M. Haigh, and Y. Umezawa (eds.) From
Headwaters to the Ocean, Taylor & Francis, 2009.
161