130
Earth and Planetary Science Letters, 40 (1978) 130-136
© Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands
I61
THE REMOVAL OF DISSOLVED HUMIC ACIDS AND IRON DURING ESTUARINE MIXING
E.R. SHOLKOVITZ l, E.A. BOYLE : and N.B. PRICE 1
i Grant Institute of Geology, University of Edinburgh, West Mains Road, Edinburgh, EH9 3JW (Scotland)
2 Department of Earth and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139 (U.S.A.)
Received December 24, 1977
Revised version received March 15, 1978
The estuarine chemistry of dissolved humic acids was determined by carrying out both field and laboratory
studies. These approaches were combined in an investigation of the Amazon estuary while laboratory mixing experiments were performed using filtered (0.45-0.001 #m) river water fractions of the Water of Luce (Scotland).
The results demonstrate that a small fraction of river dissolved organic matter is preferentially and rapidly flocculated during estuarine mixing. This fraction is the high molecular weight component of dissolved humic acids (0.450.1/zm filtered). Approximately 60-80% of the dissolved humic acid in these rivers flocculates during estuarine mixing. This represents a removal of only 3-6% of river dissolved organic matter and is responsible for the non-conservative behaviour of dissolved humic acid in the Amazon estuary even though total dissolved organic carbon appears
conservative.
The salinity dependence with which humic acid flocculates in estuaries is similar to that of iron. This implies that
both constituents may be removed from river water by a common mechanism of colloid flocculation.
1. Introduction
During estuarine mixing the removal of dissolved
organic matter (DOM), iron and other inorganic trace
constituents appear to be related to a common process [ 1 - 3 ] . The salinity dependence of the removal
o f DOM and iron are similar [ 1,2]. However, dissolved
iron is almost totally removed while only 3 - 1 1 % of
river DOM is flocculated [1,3]. This small reactive
fraction o f DOM appears to be colloidal humic material which forms part o f DOM in river water [ 1,2]. A
better understanding o f the processes which control
DOM behaviour in estuaries will also help to establish
the relationship between DOM and trace metal reactivities.
Apart from the references cited above, little work
has been undertaken to determine either the concentration o f DOM in estuaries or its estuarine chemistry
[4]. For example, Duursma [5] has presented a salini t y - d i s s o l v e d organic carbon (DOC) diagram for an
estuary (Wadden Sea) indicating that DOC behaves
conservatively;he also concludes from laboratory
experiments that there is no measurable precipitation
of DOC from river water when it is mixed with seawater. Sieburth and Jensen [6] and Swanson and
Palacas [7] have demonstrated that land-derived dissolved humic substances precipitate in seawater. Hair
and Bassett [8] and Brown [9,10] have suggested
that the high molecular weight fractions of dissolved
humic acid (DHA) in river water flocculate during
estuarine mixing. This conclusion is based on monthly
averages of the concentrations of dissolved and particulate humic acids [8] and spectrophotometric measurements of estuarine waters [9,10]. Although the
above studies reach what appears to be an intuitively
correct conclusion, in neither case do their results
actually demonstrate the preferentive precipitation of
high molecular weight humic acids during estuarine
mixing.
In this paper we have attempted to extend and
refine our understanding of removal mechanisms and
reactivity of DOM in estuaries by combining both
field and laboratory studies o f dissolved organic matter and humic acid in the Amazon and Water of Luce
131
(Scotland) estuaries. River water-seawater mixing
experiments have proven valuable in establishing the
estuarine reactivity of certain inorganic and organic
constituents [ 1 - 3 ] .
2. Field and laboratory studies
2.1. Amazon estuary field study
The work in this paper is part of a large geochemical study of the Amazon River system which was conducted during May and June, 1976, on the R.V.
"Alpha Helix". Surface water samples from three
transects of the Amazon River and Para River (Brazil)
estuaries were pressure-filtered through 0.4/am Nuclepore filters within 2 hours of their collection to remove
particulate substances. The dissolved humic acids
were then precipitated from solution by acidification.
The two transects of the Amazon estuary covered the
full salinity range (0-36%o) while the one in the Para
estuary only covered 0-9%°. The exact locations of
these transects are available from the authors.
A separate set of filtered estuarine water samples
were acidified and stored for the measurement of dissolved iron.
DOC measurements were made on filtered water
samples collected during an earlier study (June 1974)
of the Amazon estuary [11].
2.2. Amazon laboratory studies
To simulate the behaviour ofhumic acids during
estuarine mixing, shipboard laboratory mixing experiments (as in Sholkovitz [1]) were conducted in tire
following manner. Varying amounts of filtered Atlantic seawater (S = 36.0%o) were added to fixed amounts
of filtered (0.45/am) Amazon River water to achieve
twelve solutions of salinities of 0-30%°. The river
water was collected from the main river channel at a
position approximately 100 km upstream from the
mouth of the estuary but below any major tributary.
The river water was immediately filtered upon collection and used in the mixing experiments. The mixtures were left for 1 hour after which they were filtered
through 0.4/am Nuclepore filters to collect the
resulting flocculants. The filtrates were then acidified
and re filtered to collect the humic acid precipitates.
These precipitates were returned to Edinburgh where
their humic acid concentrations were measured.
2.3. Water o f Luee
Previous studies [ 1,2] of this river water used laboratory mixing experiments to establish the extent and
salinity dependence of removal of dissolved organic
matter, Fe, Mn, A1, P and Si. In the present study emphasis is placed on measuring the concentration and
estuarine reactivity of organic matter in fiver fractions which have passed through filters of pore size
0.45-0.001/am. This allows an investigation of the
estuarine reactivity of colloidal fractions within DOM
to be made. "Dissolved" matter is defined to be that
which passes through 0.45/am membrane filters and
is to be regarded solely as a working definition [12].
Separate aliquots of filtered (0.45/am) river water
were passed through Sartorius 0.1,0.05 and 0.01/am
filters and Amicon ultrafilters XM100, XM50, UM20,
UM10, UM2 and UM05. The nominal pore sizes of
the latter filters range from 0.005 to 0.001/am (see
Fig. 6). The following laboratory studies were conducted using the above fractions of river water.
(1) River water-seawater mixing experiments
were carried out to determine the salinity functionaliry with which humic acid is flocculated. These experiments are identical to the one described for the Amazon; the only added feature is that both 0.45 and
0.01/am filtered river waters were used as end-members.
(2) The humic acid concentrations were determined for the above river water fractions (e.g. 0 . 4 5 0.001/am filtered).
(3) The filtered (0.45-0.001/am) river water fractions were mixed with filtered (0.45/am) seawater to
yield solutions of 25%0 salinity. The concentrations
of both organic carbon and humic acid in the seawater flocculants and in the river water end-members
were measured. This allowed the calculation of the
percentage removal of organic substances from each
filtered fractions of river water during estuarine mixing.
3. Analytical methods
The majority of the chemical measurements follow
those of Sholkovitz [1 ] and are briefly outlined
below.
132
(1) DOC and flocculated organic carbon for the
Water of Luce studies were measured on a PerkinElmer 240 analyzer. For DOC the solutions were concentrated by evaporation before analysis. The flocculants were collected on precombusted Whatman glass
fibre filters (GFF), dried and placed directly in the
analyzer. A precision of-+ 10% is estimated for both
types of measurements.
(2) DOC for the Amazon estuary was measured by
using a wet combustion method [ 13]. The analytical
precision, inferred from duplicate measurements of
the samples, is -+0.2 mg/1.
(3) DHA - operationally defined as acid-insoluble
humic matter - was determined by a colorimetric
method. DHA was precipitated from solution by adding HC1 to give pH values of 1-1.5. The humic acid
precipitates were collected on 0.4/~m Nuclepore
filters and then redissolved in sodium pyrophosphate
(0.1/14); the absorbances of the resulting solutions
were measured at 365 nm. Standards were prepared
separately from Amazon and Water of Luce river
waters. From each river humic acid was acid-precipitated from varying volumes of filtered river water.
After drying and weighting the precipitates, they
were redissolved in the sodium pyrophosphate solution. Weight versus absorbance gave linear calibration
lines for both rivers and resulted in a precision of
-+5% for the DHA concentration.
(4) The concentrations of humic acid in the Water
of Luce seawater flocculants were determined by
directly dissolving the flocculants in the sodium pyrophosphate solution and measuring the absorbances of
the resulting solutions.
(5) Dissolved iron was determined on board the
R.V. "Alpha Helix" using a colorimetric (ferrozine)
method [14].
4. Results and discussions
4.1. Amazon estuary
Figs. 1 and 2 show that DHA has a pronounced
non-conservative behaviour whereas DOC is conservative. The major part o f the DHA removal occurs
between 0 and 5%0. The extent of removal is estimated (by the method of Boyle et al. [16]) to be
80%. But this represents only 5% of the river DOM
40(
30C
AMAZON
e-Leg~
O-teg~I~
400~
o
PARA AMAZON
~:~300~1
~20C
• LegVi~
L)
o
}
IOC
%o
°
;°°f
o
L
o
0
1 2
3 4
5 6
S%o
7 8
9 10
o
2's
2'0
~
3's
Fig. 1. The concentration of dissolved h u m i c acid vs. salinity
from two transects o f the A m a z o n estuary and one o f the
Para estuary (May 1976).
(DOM = 2 × DOC = 6 ppm) and explains why DOC
appears to be conservative. The small extent of DOM
removal is similar to that in Scottish estuaries [1 ].
The salinity dependence of removal o f DHA is
very similar to that of dissolved (0.45/~m rdtered)
iron (Fig. 3). As with humic acid, iron removal is
greatest between 0 and 5%° and is almost complete
by 15%° with a greater than 95% removal of the river
dissolved iron. This F e ~ % o relationship is of the
type generally observed in estuaries [1,3,15-19].
DOC
(mgtl
AMAZON
JUNE 1974
4£
•
2.fl
l
•
•
.,o~
1.0
o
-~
i'o
l's
•#
•
•
•
20
$
•
2'5
3~o
|
3'5
S%o
Fig. 2. Dissolved organic carbon vs. salinity from the Amazon
estuary (June 1974). Two data points at the same salinity
represent duplicate measurements of the same sample. The
extent of adsorption of DOC onto filters has not been determined.
133
~M/kg
L
T est analytiCol
pended loads and (b) long residence time of water
relative to the 1-hour mixing experiment.
The laboratory experiments confirm the observation of the Amazon estuary and establish unequivocally that river DOM (0.45/lm filtered) contains a
small fraction (e.g. colloidal humic acid) which flocculates during estuarine mixing. This process was
studied further using the Water of Luce.
error
4.2. Water of Luce (Scotland)
05
qb
oo
In Fig. 5 the laboratory mixing experiments demonstrate that humic acid flocculates rapidly and
•
I0
S~
20
1200
30
Fig. 3. The concentration of dissolved iron vs. salinity from
the A m a z o n estuary.
A - 0.45.urn
1000
~800
The laboratory mixing experiments, using the
Amazon river water, also demonstrate that DHA is
precipitated rapidly during estuarine mixing (Fig. 4).
The extent of removal is approximately 60%. The
removal gradient (in terms of salinity) are less steep in
the mixing experiment than in situ. This may be
attributed to the enhanced removal of DHA under
estuarine conditions of(a) large (up to 500 mg/1) sus-
600
U
200
1'0
1'5
2'0
v
2'5
3'0
2'5
3'0 35
35
5%o
600
°° t
B-
500'
O.01pm
600[
~400
500~ •
i
~t
elm
~ 300
U
U,~
--u300~
1-
200[
100
100I
0
200
lb
l's
S~.
2'0
2's
3'o
35"
Fig. 4. The concentration o f dissolved h u m i c acid vs. salinity
from laboratory mixing e x p e r i m e n t s using A m a z o n River
water and Atlantic seawater as end-members.
0
½
1'0
1'5
2'0
S%o
IA
Fig. 5. The concentration o f h u m i c acid vs. salinity from
laboratory mixing experiments. Part A and B use 0.45 and
0.01 # m filtered Water o f Luce river water, respectively, as
end-members.
134
TABLE I
Flocculation of humic acids when the Water of Luce filtered
fractions (0.45-0.001/~m) are mixed with seawater
IOg¢l
90O
Filtration pore size (~tm)
0.45
0.10
0.05
0.01
0.005 (XM 100) **
0.0011 (UM 0.5) **
% Flocculation * of humic acid
45
14
14
20
18
20
WATER OF LUCE
2
u
* % Flocculation =
_ #g/1
800
--~ 700
2IX
of humic acid flocculated @S = 25%0× 100.
/~g/1 of humic acid in the filtered water
SarteeiuJ
** Amicon ultrafilters.
extensively (65% removal) from 0.45/am filtered river
water; in contrast, the 0.01/am filtered river water
shows little (<10%) removal of humic acid. This is
convincing evidence that the humic acids, which flocculate during estuarine mixing, are the high molecular
weight ones (between 0.45 and 0.01/am filtered fiver
water). This reactive fraction is responsible for the
non-conservative behaviour of humic acid observed in
the Amazon estuary. The 65% removal of DHA is
close to the values reported in this paper for the Amazon estuary (60-80%). This 65% removal of humic
TABLE 2
Flocculation of DOC when the Water of Luce Filtered fractions (0.45-0.001/~m) are mixed with seawater
Filtration pore size
(tzm)
DOC concentration in each fraction (mg/1)
% Flocculation *
of DOC
0.45
0.15
0.10
0.05
0.003 (XM 50) **
0.0015 (UM 10) **
0.001 (UM 2) **
16.5
18.0
18.0
15.0
13.0
7.5
3.5
6.3
4.3
3.2
1.4
* % Flocculation
= mg/1 of organic carbon flocculated @S = 25%0 × 100.
mg/1 of organic carbon in the filtered water
** Amicon ultrafilters.
~ XMIIKI ~
~
UMI0
AJ~tm
UM2
UM05
(era)
Fig. 6. Concentration of humic acid in filtered (0.45-0.01
#m) and ultrafiltered (0.005-0.001/am) Water of Luce river
water.
acid represents a 3% removal of DOM in the Water of
Luce estuary (Table 2).
The humic acid concentrations remaining in the
size fractions of filtered river water show a rapid
decrease in concentration below 0.45/am (Fig. 6).
Approximately 43% of the humic acid is between
0.45 and 0.05/am ; below 0.005/am there is a rapid
decrease until there is only 5 - 1 0 % in the river water
passing through 0.002/am and less.
When the filtered fractions of river water in Fig. 6
are mixed with seawater, 45% of the 0.45/am filtered
humic acid is flocculated (Table 1); for solutions
below 0.10/am only 14-24% of their humic acid content is flocculated. Not only is there a significantly
greater concentration of humic acid in the 0.45/am
fdtered fraction, but the extent of flocculation is the
largest in this fraction. These observations further
support the idea that the flocculation of fiver DOM in
estuaries is dominated by a fraction consisting of high
molecular weight humic acids.
The DOC concentrations in the filtered fractions
show a fairly constant value (15 -+ 3 mg/1) between
0.45 and 0.003/am (Table 2). Below this pore size the
DOC concentration decreases significantly. The DOC
size distribution therefore is very different from that
o f h u m i c acid which decreases to 30% of its 0.45/am
f'dtered concentration at 0.003/am (Fig. 6). The loss
of 0.7 mg/1 of humic acid between 0.45 and 0.003
135
#m is too small to measurably affect the DOC concentration (15 + 3 mg/1). This explains why in the
Amazon estuary DOC appears conservative while
humic acid shows a 80% removal. This conclusion is
further substantiated by the organic carbon measurements of the seawater flocculants of the filtered fractions (Table 2). These results show that the percent
renroval of DOC decreases with decreasing pore size.
6% of the DOC (0.45/am filtered) is flocculated; this
diminishes to 1.4% for the 0.003/am filtered fraction.
This 6% removal is in close agreement with values calculated for the Amazon estuary (5%) and for the
Scottish estuaries (3-11%) [ 1].
4.3. General observations
It must be emphasized that the mixing experiments, described in this paper, use filtered river water
and therefore produce flocculants in the absence of
any suspended particles. As previously mentioned,
the presence of particles in the Amazon estuary may
be responsible for the greater extent ofhumic acid
removal at low salinities in the estuary as compared
to laboratory mixing experiments.
In addition, this paper only describes the estuarine
reactivity of DHA. The behaviour of particulate
humic acids has not been directly studied, but recent
studies using 13 C measurements of estuarine and
coastal sediments demonstrate that the deposition of
river-borne suspended organic matter is confined to
near the mouths of rivers [20]. Beyond this point the
bulk of organic matter in sediments results from the
accumulation of oceanic organisms.
5. Conclusions
The field and laboratory studies of the Amazon
and Water of Luce estuaries have produced a set of
consistent observations on the estuarine chemistry of
DOM. The most important process during estuarine
mixing is the rapid and preferential fiocculation of
the high molecular weight fractions of DHA of river
water. Approximately 60-80% of the DHA (0.45/am
f'dtered) in river water will flocculate during estuarine
mixing. This flocculation is responsible for measured
removal of only 3-6% of fiver DOM. As observed in
the Amazon estuary this reactivity will lead to a pro-
nounced non-conservative behaviour of humic acid
while DOC will appear conservative.
The removal of humic acid has the same general
salinity dependence as that of dissolved iron ([ 1-3,
15-19], this study). Although the extent of DOM
removal (3-6%) in the Amazon estuary is small relative to that of iron (>95%), the flocculated organic
matter still dominates the flocculated iron. In the
Scottish estuaries their weight ratio is 2 - 5 : 1 [ 1]
while in the Amazon estuary this ratio is approximately 5 : 1 (Figs. 1 and 3).
Not only are the concentrations ofhumic acid
greatest in the higher molecular weight fractions
(between the 0.45 and 0.10/am fraction) but also the
preferential flocculation of this fraction has been
shown to occur during estuarine mixing. This same
behaviour also holds true for iron as the major proportion of "dissolved" iron is found between the 0.45
and 0.1/am filtered fractions of river water [3,2123]. Moreover, dissolved iron is associated with dissolved humic acids in river water and seawater [1-3,
24,25]. These observations imply that a single chemical process is responsible for the removal of iron and
humic acid in estuaries. The evidence points strongly
to the seawater flocculation of river colloids which
have iron and humic acids in close physical-chemical
association.
Acknowledgements
Particular thanks goes to Dr. John Edmond who
invited us along on the Alpha Helix Expedition of the
Amazon, and to Derek Brown who did most of the
analytical measurements in Edinburgh. This research
was supported by two grants from the Natural Environment Research Council of Great Britain to E.R.S.
and N.B.P.; their continuing support is greatly appreciated. The helpful comments of Dr. John Farrington
were most welcome.
References
1 E.R. Sholkovitz, Flocculation of dissolved organic and
inorganic matter during the mixing of river water and seawater, Geochim. Cosmochim.Acta 40 (1976) 831.
2 J.M. Eckert and E.R. Sholkovitz, The flocculation of
iron, aluminium and humates from river water by electrolytes, Geochim. Cosmochim. Acta 40 (1976) 847.
136
3 E. Boyle, J.M. Edmond and E.R. Sholkovitz, On the
mechanism of iron removal in estuaries, Geochim. Cosmochim. Acta 41 (1977) 1313.
4 P.C. Head, Organic processes in estuaries, in: Estuarine
Chemistry, J.D. Burton and P.S. Liss, eds. (Academic
Press, New York, N.Y., 1976) Chapter 3, p. 54.
5 E.G. Duursma, Dissolved organic carbon, nitrogen and
phosphorus in the sea, Neth. J. Sea Res. 1 (1961) 1.
6 J. McN. Sieburth and A. Jensen, Studies on algal substances in the sea, I. Gelbstoff (humic material) in terrestrial and marine waters, J. Exp. Mar. Biol. Ecol. 2 (1968)
174.
7 V.E. Swanson and J.G. Palacas, Humates in coastal sands
of northwest Florida, U.S. Geol. Surv. Bull. 1214-B
(1965) 131.
8 M.E. Hair and C.R. Bassett, Dissolved and particulate
humic acids in an east coast estuary, Estuarine Coastal
Mar. Sci., 1 (1973) 107.
9 M. Brown, Transmission spectroscopy examination of
natural waters, part C, Estuarine Coastal Mar. Sci. 5
(1977) 309.
10 M. Brown, High molecular weight material in baltic
waters, Mar. Chem. 3 (1975) 253.
11 J.D. Milliman and E. Boyle, Biological uptake of dissolved
silica in the Amazon Estuary, Science 189 (1975) 995.
12 V.C. Kennedy and G.W. Zellweger, Filter pore-size effects
on the analysis of A1, Fe, Mn and Ti in water, Water
Resour. Res. 10 (1974) 785.
13 D.W. Menzel and R.F. Vaccaro, The measurement of dissolved organic carbon and particulate carbon in seawater,
Limnol. Oceanogr. 3 (1964) 785.
14 L.L. Stookey, Ferrozine - new spectrophotometric reagent for iron, Anal. Chem. 42 (1970) 779.
15 E. Boyle, R. Collier, A.T. Dengler, J.M. Edmond, A.C. Ng
and R.F. Stallard, On the chemical mass-balance in estuaries, Geochim. Cosmochim. Acta 38 (1974) 1719.
16 E.R. Sholkovitz, E. Boyle, N.B. Price and J.M. Edmond,
Removal of "dissolved" material in the Amazon Estuary,
Trans. Am. Geophys. Union 58 (1977) 423 (abstract).
17 V. Subramanian and B. D'Anglejan, Water chemistry of
the St. Lawrence Estuary, J. Hydrol. 29 (1976) 341.
18 P.A. Yeats and J.M. Bewers, Trace metals in the waters of
the Sagvenay Fjord, Can. J. Earth Sci. 13 (1976) 1319.
19 L.M. Holliday and P.S. Liss, The behaviour of dissolved
iron, manganese and zinc in the Beaulieu Estuary, Estuarine Coastal Mar. Sci. 4 (1976) 349.
20 P. Gearing, F.E. Plunker and P.L. Parker, Organic carbon
stable isotope ratios of continental margin sediments,
Mar. Chem. 5 (1977) 251.
21 J.P. Giesy and L.A. Briese, Metals associated with organic
carbon extracted from Okefenokee Swamp water, Chem.
Geol. 20 (1977) 109.
22 J.B. Andelman, Incidence, variability and controlling factors for trace elements in natural fresh waters, in: Trace
Metals and Metal-Organic Interactions in Natural Waters,
P.C. Singer, ed. (Ann Arbor Science Publishers, Ann
Arbor, Mich., 1974) 57.
23 L.S. Coonley, E.B. Baker and H.D. Holland, Iron in Mullica River and in Great Bay, New Jersey, Chem. Geol. 7
(1971) 51.
24 J.H. Reuter and E.M. Perdue, Importance of heavy metalorganic matter interactions in natural waters, Geochim.
Cosmochim. Acta 41 (1977) 325.
25 G.L. Pickard and G.T. Felbeck Jr., The complexation of
iron by marine humic acid, Geochim. Cosmochim. Acta
40 (1976) 1347.