Journal of Coastal Research
478-485
Royal Palm Beach, Florida
Spring 1999
Watershed Landuse and Bay Sedimentation
John J. Rooney]', and Stephen V. Smith']
t Department of Geology and
:j:Department of Oceanography
School of Ocean and Earth
Science and Technology
University of Hawaii
2525 Correa Road
Honolulu, HI 96822, U.S.A.
Geophysics
School of Ocean and Earth
Science and Technology
University of Hawaii
2525 Correa Road
Honolulu, HI 96822, U.S.A.
.:
.tllllllll:.
eusu
... b-iUS
~
....
•
~
W
"d
..
ROONEY, J.J. and SMITH, S.V., 1999. Watershed Landuse and Bay Sedimentation. Journal of Coastal Research,
15(2), 478-485. Royal Palm Beach (Florida), ISSN 0749-0208.
This paper discusses a unique synthesis of techniques used to examine sediment dynamics in a temperate watershedestuary system. Sedimentation rates for Tomales Bay, California, were initially calculated using a geographi8c information system (GIS) and digitized bathymetric data from National Ocean Service navigational survey charts published
in 1861, 1931, 1957 and 1994, with a public-domain Unix-based software package, Generic Mapping Tool (GMT). The
surfaces were subtracted from each other to determine the volume of sediment that accumulated in the Bay between
surveys. Average baywide accumulation rates, normalized by watershed area, were found to be about 94 t km 2yr 1
for the 1861-1931 interval (Interval I), 357 t km 2 yr 1 for 1931-1957 (Interval 11), and 101 t km -yr 1 for 1957-1994
(Interval III). GIS-derived rates for Tomales Bay were compared with results hindcast from rating curve equations
between rainfall and contemporaneous sediment yield. GIS results for Intervals I and II agree rather well with the
sediment yield equation, while the rating curve provides a much lower sediment delivery rate than the bathymetric
data for Interval II. Differences between the two sets of rates apparently reflect changes in land use and streambed
sediment storage.
ADDITIONAL INDEX WORDS:
Coastal sedimentation, erosion, landuse, geographic information system.
INTRODUCTON
It has been estimated that worldwide sedimentation rates
in coastal waters have doubled since prehistoric times due to
anthropogenic activities such as crop farming, grazing of livestock, logging, etc. (e.g. MILLIMAN and SYVITSKI, 1992). These
estimates of elevated sediment delivery are in spite of the
sediment trapping effects of reservoirs, which have greatly
reduced sediment delivery from some river systems to the
coastal ocean (MILLIMAN and MEADE, 1983). Increases in the
influx of terrigenous sediment to coastal waters frequently
have a significant and widespread influence on the nearshore
marine environment. These enhanced inputs can have important consequences in terms of the geochemical cycling and
biological interactions within coastal systems (e.g., HaLLIGAN
and REINERS, 1992; HEATH, 1995).
Because alterations in land use have, in general, dramatically increased nearshore sediment loading, profound
changes have occurred in coastal marine environments. To
understand these these changes, one must have knowledge
of the sedimentation rates there over the period of time in
which anthropogenic land use practices may have affected
the system. On a regional scale, changes in sediment accumulation rates strongly influence processes such as nearshore carbon and nutrient loading, which in turn is linked to
97004 received 10 June 1997; accepted in revision 20 February 1998.
This research was funded by NSF grant No. OCE 89-14833. UH
SOEST Contribution No. 4644.
the role of the coastal ocean in the global carbon cycle (SMITH
and HOLLIBAUGH, 1993; HElP et al., 1995; HEATH, 1995).
The Tomales Bay, California watershed has experienced
changes in land use which have affected nearshore sedimentation rates over historic times. The research reported here
used a synthesis of techniques to examine sediment dynamics
in this temperate estuary/watershed system. Sediment accumulation rates in the bay were determined, with the aid of
a geographic information system (GIS), by comparison of
bathymetric surveys at three descreet time intervals. These
were compared with rates derived from a series of models
that relate rainfall to sediment yield, in order to determine
the effects of varying land use on sedimentation rates in the
bay.
MEADE (1982) reports that probably greater than 95% of
the riverine sediment delivered to the U. S. Atlantic coast is
trapped in estuaries and coastal marshlands. This observation suggests that the methods described here may be applicable to a wide selection of coastal environments, at least
those for which other sediment sources and sinks can be
quantified.
Study Area
Tomales Bay (Figure 1) is a straight, narrow, submerged
section of the San Andreas Rift Zone, 50 km north of San
Francisco. The embayment is about 20 km long and an average of 1.4 km across. It trends from southeast at the head
of the bay to northwest at the mouth, has a watershed of 561
479
Tomales Landuse and Sedimentation
California
Bodega
Bay
/
Keys
Creek
San
Francisco
more recent sedimentary rock types (IRWIN, 1990; KASHIWAGI, 1985). Fluvial input from Walker and Lagunitas
Creeks is the major sediment source to Tomales Bay, with
oceanic sediments, cliff erosion, and periodic landslides contributing minor amounts (DAETWYLER, 1966). Agricultural
activity (mostly grazing by cattle) and use of the watershed
as park and recreation areas dominate land use. There is only
light residential and commercial development in the Tomales
watershed, with a population of about 11,000 (SMITH et al.,
1996).
Previous Research
OCEAN
N
I
o
10km
Figure 1. Tomales Bay, located in the western portion of Marin County,
California.
km", and a total surface area of about 28 km 2 (FISCHER et al.,
1996; SMITH et al., 1996). The average depth in the bay is
about 3 m with a deeper basin in the middle third of the bay
and shallower areas at either end, particularly at the head.
Both the head and mouth of the bay have intertidal mudflats
and there is a predominant tidal channel up to 21 m deep
that runs along the southwest side in the outer third of the
bay. The average tidal range for the bay is 1.4 m between
mean lower low and mean high water (Steve Lyles, National
Ocean Service, pers. comm.i. Two major streams flow into the
bay; Walker Creek, 5 km from the mouth of the bay on the
northeastern shore, and Lagunitas Creek at the head of the
bay. Rainfall varies across the watershed from about 600 mm
per year in the northeastern to about 1300 mm per year in
the southern portion, with average runnoff calculated as
about 500 mm per year over the past 20 years (FISCHER et
al., 1996). The highest point in the watershed is found in the
western portion, the 794 m peak of Mt. Tamalpais, with several other peaks over 300 m found nearby. In general the
watershed has fairly steep topography and high drainage
density, as is typical of much of the Coastal Range. Roughly
half the watershed is covered with shallow clay loam soils,
primarily on the upland areas. The other soils are moderately
deep and lie directly above bedrock CKASHIWAGI, 1985). The
San Andreas fault divides the area into two distinctive geologic terranes. The area northeast of the fault is characterized by graywacke sandstone of the Franciscan Formation
and the southwestern side by late Cretaceous granitic rock
of the Salinian block, overlain in many places by a variety of
This study was part of a larger research effort, the Land
Margin Ecosystem ResearchlBiogeochemical Reactions in Estuaries (LMER/BRIE) program. The LMER/BRIE studyevaluated the role of the land-sea interface (LSI) in biogeochemical processes and the importance of various forcing functions
in altering these processes within the LSI (HOLLIBAUGH et
al., 1988). It has been found that terrigenous sediments in
runoff water are a major component of exogenous nutrient
input to the bay (SMITH et al., 1996).
PLANT (1995) developed a particle aging model to calculate
sediment accumulation rates for Tomales Bay. Dating of the
younger portions of his sediment cores and estimates of bioturbation depth required by his model were done using 137CS
profiles. His modeled bay-wide sediment accumulation rates
over the period of this study are equivalent to about 300 t
km2 y-l.
MUDIE and BYRNE (1980) used pollen grains as tracers in
cores to study sedimentation rates at Drakes Estero, 5 km
south of Tomales Bay. Sedimentation rates there were determined to be 0.5 mm per year for the period preceeding the
introduction of European-style agriculture, followed by an increase to 5 mm per year, between 1820-1860. This ten-fold
increase in sediment accumulation apparently reflects accelerated erosion caused by overgrazing.
NIEMI and HALL (1996) compared charts and topographic
maps of the head of Tomales Bay to document historical
changes in tidal marshes there. They found a dramatic progradation of the tidal marshes between 1860 and 1954, particularly prior to 1918. They attributed this increase to enhanced sedimentation caused by the onset of European-style
agricultural practices and logging in the watershed. From
1954 to 1982 they found no significant progradation, apparently the result of decreased sediment input due to sediment
trapping by reservoirs in the watershed and revegetation of
hillsides.
METHODS
Digitized bathymetric data were obtained directly from the
National Oceanic and Atmospheric Administration (NOAA),
National Ocean Survey (NOS) for navigational surveys of Tomales Bay conducted in 1931, 1957, and 1994. We digitized
data for 1860-1 using NOS survey charts from those years.
All soundings were converted to meters, adjusted to a common datum to allow for sea-level rise, and georeferenced to
the North American Datum of 1983. The data files were analyzed with the aid of a software package with GIS capabilities called Generic Mapping Tools (GMT), Version 3.0 (WESSEL and SMITH, 1991; WESSEL and SMITH, 1995).
Journal of Coastal Research, Vol. 15, No.2, 1999
480
Rooney and Smith
A gridded surface of the entire bottom of the bay was made
for each survey, with grid dimensions of 0.02 minutes of latitude and longitude (about 30 x 36 m). The mean depth at
each grid location for a given year was compared with the
mean depth of that grid for other years. This yielded areaspecific bathymetry changes, or sediment accumulation volumes, for the entire bay for the intervals between navigational surveys. The head of the bay, approximately 3% of the
bay area, was not surveyed during the 1931 survey, so these
data were estimated from a time-weighted interpolation between 1861 and 1957.
Vertical accumulation data were corrected for compaction
and converted to mass accumulation rates in order to characterize sediment yield from the watershed. This was done
using two linear regression equations ( <40 em; > 40 em
depth) relating dry bulk density to depth of the sediment below the sediment-water interface. The sedimentation rate is
referred to in this paper as the "mass accumulation" rate and
may be expressed as the total mass of accumulated sediment,
or be normalized by the surface area of the bay. The term
"sediment load," as used in this paper refers to the total
quantity of sediment being exported from the watershed. To
facilitate comparisons between watersheds, the sediment
load may be normalized by watershed area, and is then referred to as the "sediment yield" (e.g. GESAMP, 1993; MILLIMAN and SYVITSKI, 1992).
We also examined the data for evidence of sediment redistribution (on the time scales of our resolution-decades) after
initial accumulation. We noted that there appeared to be areas of the bay that could be characterized by different types
of sedimentological activity. The mouth of the bay and areas
adjacent to Walker and Lagunitas Creeks appeared to have
accumulated significant volumes of sediment. The middle
portion of the bay actually appeared to deepen slightly over
the period of time covered by this study, while in between the
middle and the creek areas were transitions areas that appeared to be characterized by moderate sediment accumulation. Accordingly, the bay was divided into 6 sections along
its length (Figure 2), and sediment accumulations for each
time interval were calculated for each section.
Contemporaneous Estimates of Stream Sediment Yield
The sediment accumulation rates obtained by bathymetric
comparison using a GIS, are compared with sediment yield
estimates from hydrologic models. FISCHER et ale (1996) developed a multi-cell water balance model for the Tomales watershed. They used rainfall data from 1879-1992, to estimate
runoff and calculate the following rainfall-runoff rating
curve:
RUNOFF(mmlyr) = -284
+ 0.92
X
RAINFALL(mmlyr)(I)
Total volumes of annual runoff data were corrected to flow
volumes using the equation:
FLOW = RUNOFF - (WATERUSE
+ GROUNDWATER) (2)
from SMITH et ale (1996). A sediment yield rating curve
(SMITH et al., 1996 }, relates 7 years of daily total suspended
solids (TSS) data to daily flow for a representative portion of
the watershed, from 1987 through 1993:
TSS (g/m") = 17.8 X FLOWo.797(mmlday)
(3)
122 °52
1W
Depth change (m)
12
4
2
38°
1
12'
o
N
-1
-2
-4
-12
38°
08
N
1
'2' 400
~
1861- :
!
I 19571931 I
1994
I
I
I
I
Q)
I
I
I
I
~ 200 ~
o
I
I
~
I
I
I
I
I
~ 100
E
I 1831- :
~
I 1957 I
Q)
E
.s:
~300
~
co
- !
en
38°
04
1
N
Figure 2. Sed im ent accumulation in Tomales Bay, 1861 -1994, and GIS
derived sedimentation rates for the different time intervals. Approximate
locations of cores taken by Plant (1995) are shown as Stations 8, 12, and
16. Pixels in this figure show the size of the grids used for ba thymetric
comparison and illustrate the heterogneous nature of sediment accumulation.
Equation 3 was used in conjunction with daily streamflow
data from 1975-1993 to generate daily TSS values for that
19 year period, on the assumption that the rating curve was
stable over that time interval. The daily data were accumulated into annual yields and fit by log-log regression. A rating
curve from a log-log plot systematically underestimates material yield when compared with continuous measurements,
but FERGUSON (1986, 1987 ) developed a theory to explain
that underestimate and offered a correction factor. These considerations result in the following equation relating annual
sediment yield to runoff:
YIELD (t km ? yr "") = 10 - 3.308 X FLOW1.907 (mm yr : ") (4)
The rainfall record for the Tomales Bay watershed was used,
in conjunction with equation 4, to hindcast sediment yield to
Tomales Bay from 1860 through 1994. The standard deviation is derived from fitting the annualized runoff to sediment
yield data from 1975-1993 to Equation 4. This value was divided by the square root of the number of years in each time
interval to produce the standard errors reported in Table 2.
It must be stressed that Equation 4 is not expected to represent the actual sediment yield for this entire time period,
because it is based on a rating curve established during the
Journal of Coastal Research, Vol. 15, No.2, 1999
481
Tomales Landuse and Sedimentation
Table 1. GIS derived average sediment accumulation rates and uncertainties, with corrections applied. Note that the vertical accumulation rates
have not been corrected for compaction and thus show apparent vertical
accumulation only. The errors are standard errors as discussed in the
Methods.
Average
Vertical
Sed
Accum.
[rnm yr 11
Avg. Sed
Accum. per
km- of
Bay Area
It km ~ yr 11
Avg. Sed
Accum. per
krn" of
Watershed
It km ~ yr II
Period
Average
Sediment
Accum.
rlO:lt yr 11
1861-1931
1931-1957
1957-1994
52 ± 2
200 ± 7
57 ± 6
3.3
13.2
1814 ± 409
6918 ± 254
1961 ± 133
94 ± 21
357 ± 13
101 ± 11
1861-1994
83 ± 1
4.9
2853 ± 286
147 ± 10
relatively dry period between 1987 and 1993. Rather, the
equation provides an estimate of what the sediment yield
would have been, if land use and hydrologic factors were constant over the entire period since 1860.
Error Analysis
The method described above has multiple sources of potential error associated with it. Soundings from navigational
charts have a systematic error introduced because values between reported depth intervals are rounded to the next shallowest depth. This bias was compensated for by adding half
the value of the depth increment to all soundings. For example, if soundings on a particular survey were reported in
feet, 0.5 feet was added to all soundings from that survey.
Horizontal and vertical uncertainties introduced during surveying and digitizing were used to calculate the mean depth
uncertainty associated with bathymetric difference data (see
Rooney, 1995 for details). The depth uncertainty multiplied
by the bay area gives volumetric sediment influx uncertainty,
which can then be converted to mass as outlined above. This
introduces another source of potential error, which must be
added to the total uncertainty.
Table 2. Sediment yield rating curve derived average sediment accumulation rates. The errors are standard errors as discussed in the Results.
Rating Curve
Sed. Yield,
per krn"
Watershed
rt km ~ yr 1 ]
Period
1861-1931
1931-1957
1957-1994
1861-1994
94
67
74
87
200
400
o
-200
-400
RESULTS
1
±
±
±
±
o ± 22
290 ± 15
27 ± 13
60 ± 11
94
357
101
147
5
8
7
4
21
13
11
10
~
L-
~
N
IE
::
-
2
~
C
a
~
C
Q)
E
:a
Q)
en
-200
-400
0
400
200
0
-200
-400
400
200
0
-200
-400
0--
Outer Basin (DB)
0
0
-200
-400
o
0
400
200
400
200
0
-200
-400
Figure 3.
o
0
. Mouth
o
The accumulation rates derived from bathymetric comparisons are presented in Table 1 and Figure 2. The basic pattern is that the mass accumulation rate, normalized by watershed area, for 1861-1931 (Interval I, about 90 t km 2 yr")
was quadrupled for the period from 1931-1957 (Interval II).
It then dropped back to about the rate for Interval I during
the 1957-1994 period (Interval III). The data are obviously
insufficient to resolve the timing or maximum magnitude of
the high sedimentation interval in greater detail.
Figure 2 also shows the area-specific sediment accumulation from 1861 to 1994. It indicates that the bulk of sediment
accumulated in the deltaic regions fronting Walker and Lagunitas Creeks, and a few kilometers down the bay from Lagunitas Creek. The center portion of the bay appears to have
deepened very slightly over the same period, probably reflecting resolution limitations of the technique.
Sediment accumulations for each time interval were calculated for each section of the bay in an effort to determine
patterns of sediment redistribution following initial deposi-
GIS minus
Rating Curve
Sed Yield
[t km 2 yr- 1 ]
tion. A plot of the different rates for each section (Figure 3)
does indeed suggest evidence of a several decades-long delay
in sediment redistribution within the Bay.
The Inner Basin (IB), the section most removed from the
two creek deltas, shows the greatest increase in sedimentation rates between Interval I and Interval II, going from the
lowest to the highest rate of any section. The most likely explanation seems to be that the buildup of sediment in the
Lagunitas Creek delta during Interval I took decades to move
downstream to the IB section. The Lagunitas Creek (LAGS)
section shows an anomalous sediment accumulation rate
trend relative to the other sections of the Bay. Its highest
accumulation rate is during Interval I, followed by continual
decrease. The other sections show an increase from Interval
I to Interval II, following the overall bay-wide trend.
Differences between sediment accumulation rates determined by the GIS analysis and sediment yield determined by
the rating curve equations are shown in figure 4. The GIS
method and sediment yield method agree rather well Interval
400
200
Sediment Accwnulation History
±
±
±
±
GIS Sed.
Yield,
per km 2
Watershed
[t km 2 yr 1 ]
v
v
Outer Transition Zone (OTZ)
u
0
0
Inner Basin (IB)
0
0
0
Inner Transition Zone (ITZ)
0
0
0
Lagunitas (LAGS)
1861-1931 :1931·1957: 1957-1994
Sectional sedimentation rates for each time interval.
Journal of Coastal Research, Vol. 15, No.2, 1999
Rooney and Smith
482
400
........-----------------~
1861-1931 •
Ben
• 1957-1994
B
300 ...,
B~
~
E
c
~
~ 200 J
Q)
2
:::J
~
(!)
~c 100Q)
E
:0
Q)
C/)
:0
~
en
-
c
o
~
I
~
~
::J
o
~
B~
Q)
Q)
o
CJ)
c
0>
.5
Q)
o
"§
c
a
~
~
~
-B
:::J
~ ~
o
.5
E3 16
Q)
g
Q)
B~
:0
B
0-1--
1931-1957 •
Figure 4. Comparison of GIS and sediment yield curve derived sediment
accumulation rates, for periods between navigational surveys. The top
and bottom of each box indicate the uncertainty associated with each rate.
I and Interval III. The sediment yield method greatly underestimates infilling for Interval II.
DISCUSSION
Variations in sediment delivery from the watershed to the
bay are responsive to two major components: variation in
runoff or other aspects of system hydrology and change in
land use. If we accept that little sediment is transported into
the bay from the ocean or exported from it, the GIS-derived
data are an integration of all variation in sediment yield. Because the sediment yield calculations are based on present
land use conditions, the difference between predictions based
on the sediment yield rating curve analysis and the GIS analysis provide an estimate of how changing land use has affected the watershed-bay system.
Sediment yield calculations for the three different time intervals give very similar results (Figure 4), reflecting relatively similar estimates of runoff for each period. The Interval
I and Interval III periods show insignificant differences between the GIS-based measurements and the calculations
based on sediment yield. Essentially all of the discrepancy
between these two sets of estimates occurred during Interval
II.
Results from this study do not correspond well with baywide sedimentation rates calculated by Plant (1995). He analyzed several cores taken in the bay and found significantly
higher sedimentation rates than those reported here. However, sedimentation rates were determined by the GIS methodology for rectangles several hundred meters on a side,
within which Plant obtained his cores (Figure 2). The rates
within these areas were found to be higher than those from
the bay as a whole, and in fairly good agreement with Plant's
results. This finding illustrates both the variable nature of
sediment accumulation within the bay and the potential
problems associated with using measurements taken from a
few points to characterize an entire area.
Processes Which are Assumed Not to Affect Results
There are two fundamental assumptions required in comparing the bathymetry from different years to determine sediment accumulation rates. One is that sediment entering the
bay from the watershed remains there and is not contaminated by sediment from other sources. The other is that tectonic activity has not altered bay bathymetry significantly
over this timeframe.
The first assumption, that the bay only receives sediment
from the watershed and does not exchange it with the ocean,
was examined as part of the LMERIBRIE program. Using the
methodology reported in McMuRTY et al. (1995), SNIDVONGS
completed an inventory of fallout-derived l:ncs erosion from
the watershed and accumulation in the Bay. He concluded
that the bay acts as a sink for fluvial sediment, retaining 95%
or more of that sediment, while about 5% of the sediment in
Tomales Bay originates from oceanic sources (A. SNIDVONGS,
unpublished). This is consistent with Daetwyler's (1966) conclusion, based on grain size analysis. He attributed the occurrance of sandy sediments in the mouth of Tomales Bay to
southerly longshore currents which carried them in from Bodega Bay, outside the western end of Tomales Bay. He futher
concluded that the finer sediments found within the bay were
derived from terrigenous sources. His diagram of the distribution of grain size modes for northern Tomales Bay suggests
that the vast majority of this marine sand is trapped on the
seaward side of Tom's Point. For this reason we decided to
eliminate the mouth section of the bay prior to determining
bay-wide sediment accumulations and analyzing sectional
sedimentation rates. Based on the above considerations and
with the mouth section of the Bay eliminated, the assumption
of a closed system with respect to sediment appears valid.
Tomales Bay was formed by strike-slip movement along
the San Andreas fault zone. Small relative vertical movements of different sides of the bay could drastically alter apparent sedimentation rates, but orders of magnitude greater
horizontal movement would be required to cause the same
effect. Several lines of evidence indicate that there has not
been vertical tectonic movement in Tomales Bay over the timeframe covered by this study. Horizontal coseismic movement during the 1906 earthquake in the Tomales Bay area
is well documented (about 2.7 m) but despite efforts of at
least two separate parties of geologists shortly after that
earthquake, no convincing evidence of vertical displacement
was found (DICKINSON, 1993; GALLOWAY, 1977). Two different seismic studies of Tomales Bay sediments were also unable to find any evidence of vertical faulting younger than
several thousand years old (DAETWYLER, 1966; ANIMA, pers.
comm.).
NEIMI and HALL (1992) report that "most or all of the late
Holocene slip on the San Andreas fault has occurred on the
1906 trace." They calculate that 1906-type earthquakes occur
every 221 ± 40 years on the North Coast segment of the San
Andreas fault, which includes Tomales Bay. This is consistent with lack of tectonic creep along the fault trace and the
tendancy of this area to experience large, infrequent earthquakes rather than smaller and more frequent ones (GALLOWAY, 1977; KOENIG, 1963). Based on the above, we assume
Journal of Coastal Research, Vol. 15, No.2, 1999
Tomales Landuse and Sedimentation
483
c
c.2
E~103
~ ~
.eg-
a.
10 2
.u.-..---r----.,..---.,.----,----.----,---~
105
~----------------------..,
e
o
CD;::
-cu
=3
cue.
CJ
o
a.10~
.2c 20
~.2
o~
a. cu 10
>
.---
N
.::
E :::s
=..CJ
0
80
2
logging effort = area of logging [km ] x
intensity coefficient of 0.8 (light logging)
or 1.5 (heavy)
40
O-L..~-----r---_+---f---..,-----..,-----,---~
'i:'
('4':>.
E
400
~
2.
Q)
CiS
- - GIS sed. rate
rating curve sed. rate
o
200
~
-gen
0
1860
1880
1900
1920
1940
1960
1980
2000
year
Figure 5. Summary of landuse and population in the Tomales Bay watershed adapted from Smith et al. (1996) and based primarily on U.S. Census
data. GIS-based sedimentation data from this study and the 10 year means derived from the sediment yield rating curve are also shown.
that no tectonic displacement occurred over the timeframe of
this study that had a significant impact on apparent sedimentation rates in the bay. We now move to processes which
probably have affected the system.
Land Use
Landuse history in the Tomales Bay watershed was examined to gain a general appreciation for the degree of disturbance of the soil surface required to produce the observed
changes in sedimentation rates in the bay. Prior to about
1840 the primary human inhabitants of the area were Native
Americans, who subsisted by hunting and gathering. The
land was then settled by Europeans and European Americans
(MASON, 1971, 1976). With the advent of a predominantly
non-Native American human population in the area, the primary form of land use changed to various forms of agriculture, especially dairy farming. The intensity of agricultural
land use accelerated rapidly in the latter half of the 1800s,
and then leveled out or declined (Figure 5).
The timing of the bathymetric surveys was helpful for determining the effects of varying land use on sedimentation
rates. The 1860-61 survey essentially gives a bathymetric
"snapshot" of the bay, prior to Euroean impact. The 1931 survey shows the bathymetry after decades of both intensive
land-use and high rainfall. The survey of 1957 shows the bathymetry of the bay after continued western influence, and a
dry period, followed by wetter decades. The 1993-94 survey
covers a period of lowered agricultural impact. Thus the surveys bracket time increments that can give some insight into
the influence of different land-use practices on sediment accumulation rates in the adjacent bay.
We originally hypothesized that the greatest rate of sediment accumulation in the bay would coincide with the period
of maximum land use alteration of the watershed. However,
the peaks of these two rates do not coincide (Figure 5). We
Journal of Coastal Research, Vol. 15, No.2, 1999
484
Rooney and Smith
believe that the explanation lies with sediment storage in the
watershed. Changing land use almost certainly elevated watershed erosion, but the erosion products did not have an impact on bay sedimentation, measurable by the methodology
presented above, until after 1931. By 1957, land use effects
were no longer being reflected in bay sedimentation, which
had returned to conditions closely resembling pre-European
contact. We must look to additional information in order to
reconcile these differences between bathymetric changes and
sediment yield.
Sediment Storage in the Stream System
Some amount of sediment storage in the lower hillslopes,
stream channels, and floodplains is known to occur in the
Tomales Bay watershed. Olema Creek and Keyes Creek were
at one time navigable up to the towns of Olema and Tomales
respectively (HART, 1972). By the end of the nineteenth century they had became so choked with sediment that the head
of navigable waters had migrated several kilometers downstream. This is still the case today, indicating that sediment
has been stored in both stream systems for a century or more.
Similar delays have been documented in other drainage systems (MEADE, 1982).
We found a sediment accumulation of 68 x 10 5 m" for Interval I or about 10 5 m'' yr- 1 • The annual sediment accumulation for Interval II was about 3.6 x 10 5 m" yr-\ or an excess
of 2.5 x 10 5 m" yr ' ! for 26 years. This requires a watershed
storage capacity of about 7 x 10 6 m", assuming no other mechanisms are at work to explain the peak in sedimentation during Interval II. An examination of topographic maps reveals
that there is more than 10 7 m" of stream floodplain in the
watershed (i.e., about 2% of the watershed area). Much of the
floodplain contains an accumulation of > 1 m of apparently
recent stream sediments, for a total volume in excess of 10 7
m'' of sediments. Thus it seems reasonable to assume that
the stream beds and floodplains could easily have stored the
sediment from excessive erosion. This lends credibility to the
notion of stream valley storage as a mechanism to explain
the time delay between disruptive land-use activities and
high sediment accumulation rates in the bay.
Assuming that the delay in sediment accumulation in the
bay is due to sediment storage, it is useful to know the duration of storage as precisly as possible. An examination of
the various land usages (Figure 5) shows that the peak in
erosion is unlikely to have occured prior to 1850 or after 1890.
The GIS-derived rates (Figures 2 and 4) indicate that the
peak in sediment delivery to the bay occurred during Interval
II. That pulse of sediment is therefore most likely to have
been in storage for between, roughly, four to eleven decades.
The number and timing of the bathymetric surveys preclude
resolving the duration of sediment storage more precisely.
Sediment Yield From Other Systems
Given the multitude of variables that affect sediment yield
from a watershed, comparisons from one system to another
can be problematic. MILLIMAN and SYVITSKI (1992) addressed this problem by comparing sediment yield and load
versus area and runoff for 280 rivers throughout the world,
and for each of 7 different topographic categories of river basins. These plots show that Tomales Bay has slightly lower
than average runoff. When plotted as a function of runoff, the
sediment load and yield is very typical of other systems of
comparable size. Tomales Bay fits into their "upland" topographic category, with a maximum elevation of 500-1000 m.
When compared to other "upland" river basins with similar
runoff, the sediment yield at Tomales is slightly above the
norm.
In numerous studies of other systems, anthropogenic impacts, particularly crop farming, have been found to be responsible for elevated sediment erosion or export rates of up
to an order of magnitude, or more (e.g. MEADE, 1982; LEHRE,
1982; MILLIMAN et al., 1987; SHEFFIELD et al., 1995). In that
respect, the rates found at Tomales Bay are comparable to
others. LEHRE (1982) calculated a sediment budget for the
Lone Tree Creek watershed, a small (2 km'') drainage basin
35 km southeast of Tomales Bay. The climate, rock type, vegetation, and topography in the two areas are relatively similar. Lehre found that landslides were the most important
erosional agent and that mobilized sediment tended to be removed from the basin during extreme rain events, and stored
within the basin under other conditions. He calculated an
average annual sediment yield of about 210 t km -2 yr- 1 , noting that landslides apparently increased ten fold over the 100
years preceeding 1974. These rates are roughly comparable
with those from Tomales Bay, with average rates of 150 t
krn:" yr- 1 and 100 t km 2 yr 1 for 1861-1994 and 1957-1994
respectively. MILLIMAN and MEADE (1983) report that smaller drainage systems have less ability to store eroded sediment than do large ones. The higher yield from Lone Tree
Creek is probably explained by its tendency to store less of
the mobilized sediment than does the much larger Tomales
Bay watershed.
CONCLUSIONS
Perhaps the single most important lesson out of this study
is the reminder that sediment yield in a watershed over any
period of time is not synonymous with erosion or sediment
denudation. There is an additional storage reservoir between
erosion in the watershed and yield from the watershed:
streambed storage. Even in a relatively small watershed like
Tomales (560 km") the streambed storage time can be several
decades and this storage can obfuscate strongly expected relationships between land use and erosion.
In the Tomales watershed, it was found that although the
disruptive landuse practices apparently did increase erosion
and deposition in the Bay, there was a decades long delay
between the maximum level of soil surface disruption and
maximum sediment deposition. A similar delay was found
between initial deposition of sediments at stream deltas and
subsequent redistribution to other areas of the Bay more geographically remote from stream deltas, although areas closer
to the mouth of the Bay experienced markedly faster sediment redistribution.
Methodologically, this study demonstrates that A GIS can
be an effective tool for investigating anthropogenic impacts
that may be of too short a duration to be approached using
classic geological techniques such as radiometric dating or
core analysis and yet are too variable to be adequately characterized by measuring current rates. GIS estimates of sedimentation rates offer the advantage of a potentially high degree of spatial data density as well, allowing integration of
Journal of Coastal Research, Vol. 15, No.2, 1999
Tomales Landuse and Sedimentation
area-wide rates without the problems associated with extrapolating results of spatially limited techniques, such as core
analysis, to much wider areas.
ACKNOWLEDGEMENTS
We thank the other members of the Tomales LMER program for their helpful comments during the course of this
study, particularly J. Plant. C. Fletcher and F. Sansone
served on JJR's thesis committee. K. Sender undertook the
first analysis of the bathymetric data using GMT and helped
JJR overcome the intricacies of this software package. Lieutenant Guy Noll of NOAA, provided a pre-publication copy of
the 1994 survey data and also offered considerable insight on
the present and historical quality such data, both horizontal
and vertical. D. Livingston provided information on historical
landuse in the watershed.
UTERATURE CITED
CARTER, V., and RYBICKI, N.B., 1990. Light attenuation and submerged macrophyte distribution in the tidal Potomac River and
estuary. Estuaries, 13, 441-452.
CHAMBERS, P.A., and PREPAS, E.E., 1988. Underwater spectral attenuation and its effect on the maximum depth of angiosperm colonization. Canadian Journal of Fisheries and Aquatic Science, 45,
1010-1017.
DAETWYLER, C.C., 1966. Marine geology of Tomales Bay, central California. Scripps Institution of Oceanography and Pacific Marine
Station Research Report No.6., 169p.
DICKINSON, A.B., 1993. Tomales Township-A History. Tomales, California: Tomales History Center, 132p.
FERGUSON, R.I., 1986. River loads underestimated by rating curves.
Water Resources Research. 22( 1), 74-76.
FERGUSON, R.I., 1987. Accuracy and precision of methods for estimating river loads. Earth Surface Processes and Landforms. 12,
95-104.
FISCHER, D.T.; CHURCHILL, R.R., and SMITH, S.V., 1996. Simulation
of a century of runoff across the Tomales watershed, Marin County, California. Journal of Hydrology. 186,253-273.
GALLOWAY, A.J., 1977. Geology of the Point Reyes Peninsula, Marin
County, California. Bulletin 202, California Division of Mines and
Geology. Sacremento, California, 72p.
GESAMP, 1993. Anthropogenic Influences on Sediment Discharge to
the Coastal Zone and Environmental Consequences. GESAMP
(Joint Group of Experts on the Scientific Aspects of Marine Environmental Protection) Reports and Studies No. 52, 167p.
HART, J., 1972. A sense of place. In: Tomales Bay Environmental
Study Compendium ofReports. Washington, D.C.: Prepared for the
Conservation Foundation, p.2.
HEATH, M., 1995. An holistic analysis of the coupling between physical and biological processes in the coastal zone. Ophelia, 42, 95125.
HElP, C.H.R.; GOOSEN, N.K.; HERMAN, P.M.J.; KROMKAMP, J.; MIDDELBURG, J.J., and SOETAERT, K., 1995. Production and consumption of biological particles in temperate tidal estuaries.
Oceanography and Marine Biology: An Annual Review 1995, 33,
1-149.
HOLLIBAUGH, J.T.; COLE, B.E.; DOLLAR, S.J.; VINK, S.M.; KIMMER,
W.J.; OBREBSKI, S.; SMITH, S.V.; VALENTINO, M., and WALSH,
T.W., 1988. Tomales Bay, California: A macrocosm for examining
biogeochemical coupling at the land-sea interface. EOS, 69, 843845.
HOLLIGAN, P.M. and REINERS, W.A., 1992. Predicting the responses
485
of the coastal zone to global change. Advances in Ecological Research, 22, 211-255.
IRWIN, W.P., 1990. Geology and Plate Tectonic Development. In:
R.E. WALLACE (ed.), The San Andreas Fault System, California.
Washington, D.C.: U.S. Geological Survey Professional Paper
1515. pp. 61-80.
KASHIWAGI, J.H., 1985. Soil Survey of Marin County, California.
Washington, D.C.: Soil Conservation Service, U.S. Department of
Agriculture.
KOENIG, J.B., 1963. Geologic Setting of Bodega Head. California Division of Mines, Mineral Information Service, 16, 1-10.
LEHRE, A.K., 1982. Sediment budget of a small Coast Range drainage basin in North-Central California. In: SWANSON, F.J.; JANDA,
R.J.; DUNNE, T., and SWANSTON, D.N. (eds.), Sediment Budgets
and Routing in Forested Drainage Basins. U.S. Department of Agriculture, Forest Service, Pacific Northwest Forest and Range Experiment Station, General Technical Report PNW-141, pp. 67-76.
MASON, J., 1971. Early Marin. Inverness, California: North Shore
Books, 170p.
MASON, J., 1976. Earthquake Bay. Inverness, California: North
Shore Books, 144p.
McMURTRY, G.M.; SNIDVONGS, A., and GLENN, C.R., 1995. Modeling Sediment Accumulation and Soil Erosion with 137CS and 21°Pb
in the Ala Wai Canal and Central Honolulu Watershed, Hawaii.
Pacific Science. 49(4), 412-451.
MEADE, R.H., 1982. Sources, sinks, and storage of river sediment in
the Atlantic drainage of the United States. Journal of Geology, 90,
235-252.
MILLIMAN, J.D., and MEADE, R.H., 1983. World-wide delivery ofriver sediment to the oceans. Journal of Geology, 91, 1-21.
MILLIMAN, J.D., 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, 525-544.
MILLIMAN, J.D.; QIN, Y.S.; REN, M.E., and SAITO, Y., 1987. Man's
influence on the erosion and transport of sediment by Asian rivers:
The Yellow (Huange) example. Journal of Geology, 95, 751-762.
MUDIE, P.J., and BYRNE, R., 1980. Pollen evidence for historical sedimentation rates in California coastal marshes. Estuarine and
Coastal Marine Science,10, 305-316.
NIEMI, T.M., and HALL, N.T., 1996. Historical changes in the tidal
marsh of Tomales Bay and Olema Creek, Marin County, California. Journal of Coastal Research, 12(1), 90-102.
NIEMI, T.M., and HALL, N.T., 1992. Late Holocene slip rate and recurrence of great earthquakes on the San Andreas fault in northern California. Geology, 20, 195-198.
REED, D.J., 1989. Patterns of sediment deposition in subsiding coastal salt marshes, Terrebonne Bay, Louisiana: the role of winter
storms. Estuaries, 12, 222-227.
ROONEY, J.J., 1995. Determination of Sedimentation Rates in Tomales Bay, California Using a Geographic Information System.
Honolulu, Hawaii. Master's Thesis, University of Hawaii, School
of Ocean and Earth Science, 95p.
SHEAL, I.L., 1991. Reducing costs and improving the industry: goals
of the Dredging Research Program of the United States. Journal
of Coastal Research, 7, 535-542.
SHEFFIELD, A.T.; HEALY, T.R., and MCGLONE, M.S., 1995. Infilling
rates of a steepland catchment estuary, Whangamata, New Zealand. Journal of Coastal Research, 11, 4, 1294-1308.
SMITH, S.V.; FISCHER, D.T., and HOLLIBAUGH, J.T., 1996. Dissolved
and particulate nutrient transport through a coastal watershedestuary system. Journal of Hydrology, 176, 181-203.
SMITH, S.V., and HOLLIBAUGH, J.T., 1993. Coastal metabolism and the
oceanic organic carbon balance. Reviews of Geophysics, 31, 1, 75--89.
UNEP, 1982. River inputs to the west and central African marine
environment. UNEP Regional Seas Reports and Studies, Number 5.
\VESSEL, P., and SMITH, W.F.H., 1991. Free software helps map and
display data. EOS, 72, 441-448.
WESSEL, P., and SMITH, W.F.H., 1995. The GMT-SYSTEM version
3.0, Technical Reference and Cookbook.
Journal of Coastal Research, Vol. 15, No.2, 1999