CHEMISTRY, TRANSPORT AND FATE OF
ALUMINUM IN DILUTE ACIDIFIED LAKE
SYSTEMS
CHARLES T. DRISCOLL
-------
The Chemistry, Transport and Fate of
Aluminum in Dilute Acidified Lake Systems
by Charles T. Driscoll
Gary C. Schafran
Department of Civil Engineering
Syracuse University
Syracuse, NY 13210
315-423-2311
Final report for the USEPA/NCSU Acidic Deposition Program project
APP-0094-1981
-------
Abstract
Elevated levels of aluminum have been observed in acidic surface
waters. Aluminum is of interest because of its role as^/a' toxicant to
t aquatic organisms, a pH buffer, and an adsorbent of orthophosphate and
organic carbon. In this study we evaluated the spatial and temporal
fluctuations in aluminum chemistry and aluminum transport in an acidic
drainage lake.
Elevated levels of nitrate, hydrogen ion and aluminum were largely intro-
duced during snowmelt by drainage water to Dart Lake. During low flow periods,
microbially mediated depletions of nitrate served to neutralize hydrogen ion
and aluminum base neutralizing capacity. Thus, nitrate transformations were
extremely important in regulating short-term changes in pH and inorganic forms
of aluminum in Dart Lake. These transformations resulted in changes in the
inorganic speciation of aluminum. During low pH conditions associated with
snowmelt, A1 was the major form of inorganic aluminum. However, during higher
pH conditions, observed in the summer, fluoride and hydroxide forms of aluminum
predominated.
Although concentrations of organically complexed aluminum were high in
Dart Lake, we know very little about the character and transformations of
these solutes. Organically complexed aluminum was correlated to the dissolved
organic carbon concentration. Alumino-organic solutes were introduced to the
lake from both drainage inputs and sediments. Moreover these materials do not
appear to be conservative within acidic lake systems.
-------
On an annual basis Dart Lake wa9 a net aluminum sink. However, the
magnitude of various aluminum fluxes exhibited significant temporal variation.
During spring snowmelt Dart Lake was highly undersaturated with respect to
readily forming mineral phases and aluminum had a relatively low affinity
for particulate matter. Consequently aluminum was conservative; aluminum
that entered the lake was essentially transported out the outlet. During
stratification, waters were supersaturated with respect to readily forming
mineral phases, aluminum demonstrated a relatively high affinity for particles,
and was generally non-conservative. Within the lake, aqueous aluminum was
converted to particulate aluminum and these substances were deposited to
lake sediments. The non-conservative nature of aluminum would appear to
have significant implications for aquatic organisms.
-------
Introduction
Atmospheric deposition of acidic substances appears to influence bio-
geochemical cycling in ecosystems with soils derived from difficult to weather
'minerals (Braekke, 1976; Hutchinson and Havas, 1980). For example, Cronan and
Schofield (1979) have hypothesized that acidic deposition alters the process
of podzol development. Rather than being retained in soils, aluminum is solu-
bilized by mineral acid inputs and transported to the aquatic environnent.
Elevated levels of aqueous aluminum are significant for aquatic ecosystems
in several respects:
1. Aluminum may be deleterious to fish (Baker and Schofield, 1982) and
other aquatic organisms (Hall et al., 1984) in low ionic strength waters.
Aluminum forms soluble complexes with hydroxide, fluoride, sulfate and organic
ligands (Roberson and Hem, 1969; Lind and Hem 1975). Hydroxy-aluminum monomers
appear to be particularly important to fish toxicity (Baker and Schofield,
1982). Because pH, natural organic and inorganic (e.g. F ) ligands signifi-
cantly influence aluminum speciation,they also influence hydroxy-aluminum
toxicity (Driscoll et al., 1980; Baker and Schofield, 1982).
2. Aluminum is a hydrolyzing metal and therefore elevated concentrations
influence the pH buffering of acidic waters (Dickson, 1978; Johannesson, 1980;
Henriksen and Seip, 1980; Driscoll and Bisogni, 1984).
3. Hydrous aluminum oxides, formed by the hydrolysis and precipitation of
aluminum, may serve as an adsorbent and scavenge substances from the water
column. By this mechanism aluminum may alter the cycling of sorbable solutes
such as orthophosphate (Dickson, 1978), trace metals (Hohl and Stumm, 1976)
and dissolved orgJgfi|ic carbon (Dickson, 1978; Davis, 1982) in acidic lake eco-
systems ,
-------
Considerable information is available through synoptic surveys on total
aluminum levels in acidic surface waters (Schofield, 1976; Hutchinson et al.,
1978; Dickson, 1978; Wright and Snekvik, 1978; Vangenechten and Vanderborght,
1980; Hultberg and Johansson, 1981). Much less information is available on
the speciation of aluminum in surface waters. Reported here are results from
I
a study on the chemistry and transport of aluminum in an acidic drainage lake
in the Adirondack region of New York State.
Materials and Methods
Study Site and Field Program
The study site, Dart Lake (43°47'N, 74051,W), is located in the drainage
basin that forms the North Branch of the Moose River in the Adirondack State
2
Park, New York State. The watershed, which occupies 107 Km , is forested except
in regions of exposed bedrock. The major vegetation is secondary-growth hard-
woods including American beech (Fagus grandifolia), yellow birch (Betula alleg-
haniens is), sugar maple (Acer saccharum) and red maple (Acer rubrum). Bedrock
geology is characteristically crystalline granitic gneiss (Isachsen and Fisher,
1970) which is generally resistant to chemical weathering.
2
Dart Lake has a surface area of 0.144 Km , a major inlet and a single
outlet (Figure 1). The lake bathymetry is characterized by two significant
depressions both reaching a maximum depth of 15 m. The mean depth of Dart
Lake is 7.1 m.
The quantity of precipitation entering the watershed was measured daily
and incident radiation was monitored continuously (29 April - 21 Nov '82) ad-
jacent to the lake. To monitor discharge, staff gages were installed at the
inlet and outlet, and stream height was measured on sampling dates. Stream
velocity was measured, flows were calculated on 9 dates and a stage-discharge
relationship was developed. To provide an estimate of the water flux through
-------
Dart Lake, correlations between our instantaneous discharge measurements/cal-
culations and that of the continuous discharge recorded at the nearby Indepen-
dence River site (at Donnatsberg) and a site on the Hudson River (Newcomb) were
used. Correlations of flow per watershed area were similar to either one or
both continuous discharge sites throughout the study period and thus the
estimated continuous flow for Dart Lake is believed to be representative of
actual flow conditions.
In our study, samples were collected for water quality analysis approxi-
mately every two weeks at the inlet, outlet and from seven depths at a pelagic
sampling station. Water column samples were collected with a battery-operated
submersible pump. Temperature and light were measured in-situ with a thermister
and light meter, respectively. We measured pH and dissolved inorganic carbon
(DIC), fixed samples for dissolved oxygen (DO), ampulated samples for dissolved
organic carbon (DOC) determination and extracted samples for mononuclear aluminum,
shortly after collection.
While it is reasonably well established that pH, DIC, D.O. and DOC are prone
to change, we feel it is important to emphasize that mononuclear aluminum concen-
trations may also fluctuate during sample storage. Natural solutions are generally
supersaturated with respect to atmospheric CO2. Carbon dioxide exolution will
cause an increase in solution pH and a decrease in mononuclear aluminum levels.
To illustrate the extent to which this might occur we made some hypothetical
thermodynamic calculations (see the computative methods section of this report)
using data of a sample collected from the hypolimnion of Dart Lake (Figure 2).
In situ, this sample was highly supersaturated with respect to atmospheric CO2
(log pCO^ = 2.25). As we simulated the exolution of CO^ by reducing pCO^ to
-------
atmospheric levels, pH values Increased and inorganic aluminum concentration
decreased (Figure 2a). Aluminum transformations are very temperature sensitive.
We also simulated the change in aluminum concentration as the temperature of the
hypolimnitic Dart Lake sample increase from the in-lake value (5.5°C) to typical
I
laboratory conditions (up to 25°C; Figure 2b). It is apparent that changes
associated with_sample storage can dramatically affect aluminum levels. There-
fore we extracted samples for mononuclear aluminum as shortly after collection
as possible.
To monitor the gross flux of particulate material from the water column,
we placed triplicate sediment traps (aspect ratio of 11.7; Bloesch and Burns,
1980) adjacent to our water column sampling site at depths of 6 and 14 m.
Sediment trap samples were collected approximately every two weeks. As part
of another research project, six sediment cores were obtained from Dart Lake
with a 5-cm diameter gravity coring device. Sediments were sectioned on site
and placed in water-tight bags for storage. (Because results of the sediment
study supplement the results of this research they will be discussed in this
report. For further details on the sediment investigation see White (1984^)p
Analytical Methods
Water samples were measured for pH potentiometrically by glass electrode.
Dissolved oxygen was measured on field-fixed samples by Winkler titration (Standard
Methods, 1980). Free fluoride was determined by direct measurement with a fluoride
ion selective electrode, while total fluoride was analyzed by using a total ionic
strength adjustor and buffer (TISAB II: Orion, 1976). Dissolved inorganic carbon
was analyzed by sample acidification and extraction of CO^ into helium (Stainton,
1973). Dissolved organic carbon was measured using persulfate oxidation (Menzel
and Vaccaro, 1964) followed by syringe stripping of (»2 (Stainton, 1973).
-------
A gas partitioner was used to detect for both DIC and DOC determinations.
2+ 2+ + +
Basic cations (Ca , Mg , Na , K ) were measured by atomic absorption spectro-
photometry (AAS). Sulfate, nitrate, chloride and dissolved silica were measured
colorimetrically using the methylthymol blue (Lazrus et al., 1968), hydrazine
reduction (Kamphke et al., 1967), thiocyanate (Zall et al., 1956) and the
heteropoly blue (Standard Methods, 1980) methods, respectively, on an autoanalyzer.
The colorimetric analyses of sulfate, nitrate and dissc^^sd silica are prone to
interference by organic solutes. However, DOC levels in Dart Lake were lowj
and little background color was observed when checked by the colorimeter.
Moreover, we compared the colorimetric procedures for anion analysis with
ion chromatography^ which does not have an organic carbon interference, and
observed no statistically significant difference between measured values on
Dart Lake samples.
In our study^ three separate measurements of aluminum were made. 1) Total
aluminum (TA1) was determined by acidifying samples to a pH value of 1 for 1 hour,
chelating aluminum with 8-hydroxyquinoline and extracting the complex in methyl
isobutyl ketone (MIBK), using the procedure of Barnes (1976). Detection of
aluminum was made by AAS using graphite furnace atomization. 2) Monomeric alumi-
num (MAI) was directly chelated by 8-hydroxyquinoline, followed by rapid extrac-
tion in MIBK (Barnes, 1976) in the field. These extracts were analyzed in the
laboratory by AAS with graphite furnace. 3) Monomeric aluminum was separated into
two fractions using a column of strongly acidic cation exchange resin. We term
the aluminum that passes through the column and is detected using the procedure
for mononuclear aluminum, non-labile monomeric aluminum.
-------
With these three measurements, three separate aluminum fractions were
determined. 1) Non-labile monomeric aluminum was measured directly and is an
estimate of mononuclear aluminum that is organically complexed (0MA1). 2)
Labile monomeric aluminum is the difference between MAI and 0MA1. This fraction
thought to be an estimate of inorganic monomeric aluminum (IMA1) and would
include aquo aluminum as well as hydroxide, fluoride and sulfate complexes of
mononuclear aluminum. 3) Acid soluble aluminum (ASA1) is the difference between
TA1 and MAI. This fraction represents aluminum which requires acid dissolution
for detection and would include relatively easy to dissolve particulate
aluminum and strongly bound alumino-organic substances. Further details on
the fractionation procedure used for the determination of aqueous aluminum
are available elsewhere (Driscoll, 1984; Appendix 1).
Selected samples, collected from drainage waters of the North Branch of
the Moose River, were processed using the cation exchange column to remove any
IMA1 and titrated to characterize the proton dissociation of organic solutes.
These data were fit to a monoprotic proton dissociation model using a modified
Gran plot analysis (Driscoll and Bisogni, 1984). A statistically significant
empirical relationship was observed between DOC concentration and mols of pro-
ton dissociation sites per liter (CT = 0.025 DOC + 8.0; where CT is the proton
dissociation sites on organic solutes in ymol « 1 ^ and DOC is in ymol«l
2
n = 13, r = 0.69 p < 0.01). A relatively consistent fit to a proton dissocia-
tion constant was also obtained (pKa = 4.53 t 0.25). Our detailed observations
of DOC were applied to this model to estimate the base neutralizing capacity
(BNC) of organic solutes and organic anion concentration of water samples.
-------
Sediment trap samples were measured for total, volatile and fixed sus-
pended solids using the gravimetric procedure outlined in Standard Methods
(1980). Also the quantity of particulate aluminum deposited in sediment traps
was measured by a brief acid (pH = 1 for 1 hr) extraction followed by direct
determination of aluminum by AAS with graphite furnace.
Sediment samples were measured for total solids content (Standard Methods,
1980). In addition wet sediments were extracted sequentially for "exchangeable
aluminum", "oxide-bound aluminum", "organic-bound aluminum" and residual aluminum
using the procedure of Tessier et al. (1979). Note that while these fractions
are operationally defined^ they do provide some information on the form of par-
ticulate aluminum.
Computational Methods
Thermodynamic calculations used in this study were made with a modified
version of the chemical equilibrium model MINEQL (Westall et al., 1976).
Thermochemical data of aluminum equilibria used in our calculations are sum-
marized elsewhere (Johnson et al., 1981; Driscoll, 1984). Thermodynamic cal-
culations were corrected for temperature. Ionic strength corrections were
made using the Davles equation (Stumm and Morgan, 1981).
To evaluate the distribution of IMA1 we calculated aquo aluminum (Al^ ),
hydroxide bound aluminum (Al—OH), fluoride bound aluminum (Al—F), sulfate bound
aluminum (Al-SO^) concentrations and aluminum base neutralizing capacity (Al-BNC)
(Table 1). We define hydrogen ion and aluminum base neutralizing capacity
(H-A1-BNC) as the amount of strong base required to increase the pH of a liter
of aluminum solution to 8.3.
-------
We computed mineral saturation indices (SI) for Dart Lake solutions, to
assess the potential for equilibrium with mineral phases
SI «= log IAP /K
where: SI is the saturation index for the mineral phase of interest,
IAP is the ion activity product of the solution, and
K is the thermodynamic solubility constant for the mineral
phase of interest.
Positive, negative, and zero SI values suggest that a solution is oversaturated,
undersaturated or in equilibrium, respectively, with the mineral phase of interest.
While it is possible to use any number of minerals for this analysis, past re-
search by Johnson et al. (1981) and Driscoll et al. (1984) suggest that solution
IAP values closely follow aluminum trihydroxide (A1(0H)^) solubility. Therefore,
we have chosen to use the relatively soluble phase microcrystalline gibbsite
(p*Kso = 9.35; Hera and Roberson, 1967) as a reference mineral in our SI cal-
culations.
A number of pool and flux calculations were made in this study including
(1) the flux of solutes entering and leaving Dart Lake, 2) water column pools
of solutes, 3) sediment pools of metals, 4) gross deposition of particulate
substances, 5) aqueous-particulate distribution coefficients of metals (Kd),
6) upward transport of solutes, 7) net deposition of substances based on
material balance calculations and 8) net deposition of substances based on
sediment composition and sedimentation rate. Details on how these calculations
are made are beyond the scope on this report but are summarized elsewhere
(Schafran, 1984; White 1984).
-------
Results
Aluminum Chemistry of Dart Lake
Dart Lake solutions may be characterized as acidic, low Ionic strength waters
2+ 2-
in which Ca is the dominant cation while SO^ is the dominant anion (Table 2).
We obtained a relatively close electroneutrality balance, which provides some
independent confirmation of the procedures and assumptions used in the study.
A discussion of the general water chemistry of Dart Lake is beyond the scope
of this report but is available elsewhere (Schafran and Driscoll, 1984).
Although ^CO^ was the major component of BNC, levels of hydrogen and
aluminum BNC (H—Al-BNC) are generally of more interest in acidic lakes because
of their ecological significance (Baker and Schofield, 1982; Hall et al., 1984).
Variations in H-A1-BNC were strongly ^correlated with variations in N0^~ concen-
tration (H-A1-BNC = 0.94 N03" + 2.4; yeq-l"1, r2 = 0^5^ p <0.0001). Note
that this empirical correlation is linear with a slope close to one and an inter-
cept near the origin. H-A1-BNC was positively correlated with organic anion
concentrations (H-A1-BNC = 33.1 (RCOO ) - 481.5; where RCOO represents the
-1 2
organic anion concentration in yeq«l , r = 0.13, p < 0.0001) and no statis-
2_ _
tically significant relationship was observed with SO^ or CI .
It is well known that processes which produce NO^ such as deprotonation
of HNO^, algal respiration, and nitrification result in proton release while
reactions in which nitrate is depleted such as denitrlfication or algal assimila-
tion of nitrate result in proton neutralization (Table 3). Water quality
observations in Dart Lake indicate that biogeochemical processes result in
temporal and spatial variations in NO^ which In turn influence pH and IMA1
concentrations (Figures 3a, 3b, 3c). During the autumn, NO^-, pH and IMA1
-------
exhibited an orthograde distribution. After ice-development, NO^ and
IMA1 levels increased and pH values decreased in the upper waters, presumably
due to freeze-concentration at the ice-water interface or minor snowmelts.
Concentrations of NC>3~ and IMA1 decreased while pH values increased with
increasing lake depth. Although the lake was aerobic throughout the study
period (D.O. > 41 innol-1 minimum value observed, 14 m on 10/1/82),we
attribute these transformations to sediment reduction of NO^ . During snow-
melt, large inputs of acidic water elevated in NO^ were introduced to Dart
Lake. Inlet concentrations of H-A1-BNC and NO^" were significantly greater
during snowmelt (4/1-5/1/1982, 8.9±6.9m*s , H-A1-BNC = 52 ± 11, N0^ =
3 -1
51 ± 11) than during either summer-autumn (6/15-11/1, 1.7 ± 1.7 m »s ,
H-A1-BNC = 18 ± 5; p < 0.05 student t test, NO^- =24+5, p < 0.05) or
winter (11/1-3/1, 1.5 ± 1.0 m^s"1, H-A1-BNC = 23 ± 6, p < 0.05, N03" =
26 - 6, p < 0.1) base flow periods (Figure 4). Johannes et al. (1980)
reported that N03~ storage in snowpack from the Adirondack region was generally
greater than S0^2~. Likewise, Galloway et al. (1980) have attributed the de-
crease in alkalinity that occurs in Adirondack lake outlets during snowmelt
to a reduction in the release of basic cations from soil and the addition of
HNO^ from snowpack. Our observations would appear to be consistent with these
studies. Unlike N03~, there were no statistically significant differences between
the concentrations of SO^ , CI (not shown) and organic anions during snowmelt
and during base flow periods (Figure 4).
During summer stratification pH increased, and N03 and IMA1 were depleted
in the upper and lower waters of Dart Lake, resulting in heterograde distributions.
We attribute the epilimnitic N03 depletion to algal assimilation or denitrifi-
cation in the littoral sediments and the hypolimnetic depletion to sediment deni-
triflcation. Note that in the extreme lower waters (below 12m) during strati-
fication, transformations of Fe, Mn, NH^ , DOC and Ca also contributed to the
-------
changes in H-A1-BNC. However, because this region represents a relatively
small percentage of the total lake volume (-v5%), NO^- depletions appear to
be the most significant mechanism of H-A1-BNC neutralization on a whole-lake
basis. For further details on the role of nitrate in regulating BNC see
Driscoll and Schafran (1984; Appendix 2).
Like other studies of acidic surface waters (Johnson et al., 1981; Driscoll
et al. 1984, we observed that the inorganic aluminum chemistry was rather well
described by A1(0H)^ solubility. The IAP of Dart Lake solutions was generally
similar to the theoretical solubility of microcrystalline gibbsite (SI = 0.17
± 0.28). Johnson et al. (1981) and Driscoll et al. (1984) reported that HBEF
and Adirondack waters were somewhat undersaturated with respect to the solubility
of microcrystalline gibbsite (SI = -0.69 ± 0.36 and -0.85 ± 0.45, respectively).
The higher SI values for Dart Lake may be due to real differences in solution
composition or the fact that we extracted our samples for monomeric aluminum
in the field.
While the removal of aluminum was nearly stoichiometric with depletions
of N0^ and IAP values were well described by microcrystalline gibbsite solu-
bility, there were well defined patterns of disequilibrium in Dart Lake (Figure
5). During snowmelt, the inlet and outlet streams and the entire water column
were highly undersaturated (negative SI values). The short retention time of
water in the soil and relatively slow rates of mineral dissolution undoubtedly
contributed to these low SI values. Moreover, conditions of high supersaturation,
in the hypolimnion during winter and summer stratification and the epilimnion
during summer stratification, coincided with periods of N0^ depletion.
The removal of aluminum during N0^ depletions indicate that heterogeneous
phase transformations of aluminum proceed at relatively slow rates.
-------
In Dart Lake IMA1 was distributed among a number of Inorganic forms.
Mean values of the concentrations of various aluminum forms as well as the
relative distribution as total and monomeric aluminum are summarized in Table 4.
3+
Fluoride and organic complexes were the predominant forms of aluminum. A1 and
A1-0H were present at lower but significant levels, while Al-SO^ was insignifi-
cant. The concentrations and relative distribution of aluminum in Dart Lake
were similar to those values reported by Johnson et al. (1981) for the HBEF
and by Driscoll et al. (1984) for Adirondack surface waters.
As one would expect, the distribution of IMA1 changed markedly with varia-
3+
tions in pH (Figure 6). At low pH values, Al was the predominant form of IMA1,
but concentrations declined at higher pH values. Levels of Al-F were also high-
est at low pH values and the magnitude of the complex under these conditions
was regulated by the level of total fluoride in solution. Concentrations of
Al-F also decreased with increasing pH. A1-0H values were low at acidic pH
values, increased to a maximum value approximately at pH 5.1 and decreased at
higher pH values. The pH dependence of Al-F and A1-0H is less pronounced than
3+
Al . Thus at higher pH values (pH > 5.1) Al-F and to a lesser extent A1-0H pre-
dominated IMA1. Concentrations of A1-S0^ also generally decreased with increasing
pH.
In Dart Lake a substantial portion of mononuclear aluminum appeared to be
complexed with organic ligands. We observed a weak but statistically significant
empirical relationship between 0MA1 and DOC (0MA1 = 0.018 DOC + 1.84; in ymol-1
2
r = 0.14, p < 0.0001). Similar empirical relationships have been reported by
Driscoll et al. (1984) for Adirondack lakes and streams and by Driscoll (1984)
for HBEF streams. The correlation between 0MA1 and DOC was considerably weaker
for Dart Lake than these other systems and may be attributed in part to the
-------
limited range of DOC (220 - 420 ymol-1 *). Also DOC is a relatively coarse
parameter because it encompasses a great variety of organic solutes which
vary greatly In their capacity to complex aluminum.
DOC levels were relatively constant in the inlet and outlet streams while
0MA1 levels were more variable (Figure 4). Concentrations of both parameters
Increased slightly from autumn through the winter season. After snowmelt DOC and
0MA1 concentrations declined through the summer until September and increased
again in autumn. An exception to these relatively smooth trends occurred during
a major rainfall event (6.1 cm on 2 June 1982). pH decreased, and IMA1 and DOC
(and organic anion) concentrations increased in the inlet stream in response to
2-
the storm, while no increase was observed for SO^ , N0^ or CI (Figure 4).
Therefore we presume that organic anions were associated with this increase in
H-A1-BNC. 0MA1 levels in the inlet stream also increased in response to the
rainfall event.
Within Dart Lake there were spatial and temporal trends in DOC and 0MA1
(Figure 7a, b). Elevated concentrations of DOC and 0MA1 were introduced to
Dart Lake during snowmelt. After spring turnover both DOC and 0MA1 were sub-
stantially depleted from the epilimnitic waters. Low DOC water migrated with
the thermocline to increasing lake depths during the summer season. In the
hypolimnion, DOC and particularly 0MA1 concentrations increased dramatically
during summer stratification. By September considerable 0MA1 was introduced to
the epilimnion probably through a combination of entrainment of hypolimntic 0MA1
and Increased concentrations in the inlet stream (Figure 4). 0MA1 was rapidly
depleted form the epilimnion after this event.
-------
The aluminum content of the surficial sediments in Dart Lake was high
and generally declined with increasing depth (Figure 8). Sediment aluminum
was largely in an oxide bound form and to a lesser extent an organic bound
form. The content of residual aluminum was smaller but present in significant
quantities and the exchangeable aluminum content of Dart Lake sediments was
low. The oxide bound and exchangeable aluminum content of sediments increased
with increasing depth while the organic bound aluminum content was highest at
the water-sediment interface, declined to a minimum at a 5 cm depth and gradually
increased with greater sediment depth.
Studies of the aluminum content of acidic lake sediments have produced
inconsistent conclusions. Some investigators report, as observed in Dart Lake,
increased aluminum levels in the upper strata (Dickson, 1980; Heit et al., 1981).
Other researchers have reported aluminum depletions in the upper sediments
(Norton and Hess, 1980) , while still others report no change in aluminum
content with depth (Galloway and Likens, 1980; Heit et al., 1981). The
aluminum content and depth trends that Heit et al. (1981) report for nearby
Woods Lake are nearly identical to our results for Dart Lake.
It is difficult to extrapolate sediment chemistry to in-lake processes.
Moreover, the sediment aluminum fractions are operationally defined and there
may be some overlap associated with the various extractions. Nevertheless, the
sediment distribution between organic and inorganic forms of aluminum was very
similar to our water column observations (Table 3). The rather significant
amounts of both oxide and organic bound aluminum provide evidence that IMA1
and 0MA1 are deposited from the water column. The systematic decline in
organic bound aluminum with recent sediment depth ( < 5 cm) might be attributed
to microbial oxidation. A by-product of the oxidation of organic bound aluminum
-------
would then presumably be oxide bound or residual aluminum. However, this
process is counter to our observed decline in oxide bound aluminum with
increasing sediment depth. It is more likely that deposition of 0MA1 to the
sediments has increased in recent years or that organic bound aluminum is
isolubilized in the sediment, diffuses into the water column, is converted
to a particulate phase and redeposited to the surface sediments. This latter
process would be consistent with our water column observations.
Aluminum Transport and Cycling in Dart Lake
We have computed input/output budgets for various forms of aluminum, NO^ ,
2-
SO^ , DOC and CI for Dart Lake (Table 4). It would seem that the lake was
a net nitrate sink, which resulted in BNC neutralization and net retention of
IMA1. Moreover, on an annual basis Dart Lake retained 0MA1, was conservative
2-
with respect to CI , ASA1 and SO^ and was a net source of DOC. While these
calculations are useful excercises they provide limited information on processes
that are important to aluminum cycling.
We have formulated a conceptual model of reactive aluminum cycling within
an acidic lake (Figure 9). Reactive aluminum is considered to be aluminum
that readily moves between aqueous and particulate phases; this designation
would not include highly crystalline substances. Within the lake there are
various aqueous pools (IMA1, 0MA1, ASA1) and sediment pools (exchangeable Al,
oxide bound Al, organic bound Al and residual Al) that have been previously
discussed. Inputs of aluminum to the lake include direct atmospheric deposition,
seepage, and drainage flux. Within the lake particulate aluminum may be, de-
posited to the sediments (gross deposition), solubilized in sediments and
diffuse upward into the water column (gross upward flux) or may be transported
from the lake with seepage or drainage water.
-------
Using the format of this conceptual model we have estimated the
average annual pools and fluxes of aluminum for Dart Lake (Figure 10).
Unfortunately we do not have estimates of seepage flux to and from the lake.
Neglecting seepage, stream inputs and export would appear to quantitatively be
the most important flux of all forms of aluminum in Dart Lake. The aluminum con-
tent in bulk atmospheric deposition is low (Johannes, 1984) and therefore direct
atmospheric inputs of aluminum to the lake surface were probably insignificant.
Because solute transport to and from the lake were largely dictated by water
flow, there were pronounced variations in the drainage inputs of aluminum over
the annual cycle (Figure 11). Note that 62, 44 and 46% of the annual IMA.1,
0MA1 and ASA1 inputs to the lake occurred during the spring high flow period
(15 March - 15 June 1982).
We also observed pronounced temporal trends in the gross deposition of
particulate aluminum to sediment traps positioned at upper (6 m) and lower
(14 m) water depths (Figure 12). Gross deposition of particulate aluminum
to the lower water sediment traps always exceeded upper water values due
to the greater settling depth of the former. During the winter months the
gross deposition of aluminum was low in the upper waters and considerably
greater in the hypolimnitic waters. This observation is qualitatively con-
sistent with our observation of high SI values in the hypolimnion during
this period (Figure 5). During the high flow period (15 March - 15 June)
upper and lower water values of sedimenting aluminum were quite similar,
probably due to the turbulence associated with the high flux of water during
this period. Also considering high aluminum inputs to the lake the gross
downward flux of particulate aluminum was surprisingly low. The concentrations
-------
of ASAl in the inflowing stream during snowmelt were not statistically
different than during base flow suggesting that the concentration of
particulate aluminum did not increase with increasing water flow.
Moreover, Dart Lake was highly understaturated with respect to the solubility
of microcrystalline gibbsite during this period, so in-lake formation of
particulate aluminum was probably minimal. During summer stratification
gross deposition of particulate aluminum was high for both the upper and
lower waters, which again is qualitatively consistent with our high SI
values. Aluminum deposited within sediment was strongly correlated
with total particulate deposition (Table 5) suggesting that vertical
transport of aluminum and total particulate matter are interrelated in Dart
Lake.
We can not identify a specific process of particulate aluminum formation
in this study. Mechanisms by which in-lake formation of particulate aluminum
may occur include direct chemical precipitation, coprecipitation and adsorption
on particulate matter. The precipitation or coprecipitation of metals in nature
generally occurs by homogenous nucleation (Corey, 1981). Foreign particulate
matter serve to catalyze the precipitation process by lowering the activation
energy required for precipitation. Thus, regardless of the mechanism of in-lake
removal of aqueous aluminum, particles would appear to play an essential role.
Our SI calculations indicate that direct precipitation of aluminum is thermo-
dynamic ally feasible in Dart Lake (Figure 5). We have also computed aluminum
distribution coefficients (Kd; mol A1 gm particulate matter ^"-(mol aqueous
Al^crn-3)-1) which are a measure of the affinity of aluminum for particulate
matter (Figure 13). In-lake distribution coefficients for aluminum were generally
consistent throughout the study period. An exception to this occurred during
-------
enowmelt when Kd values decreased, particularly in the upper water, probably
due to the low pH conditions. Kd values were highest during summer stratifica-
tion and upper water values were considerably greater than full water column
values. This difference indicates that epllimnitic particles had a higher
affinity for aluminum than hypolimnitic particles which again is qualitatively
consistent with our SI calculations. In the epilimnion, high SI values
persisted for only a few weeks indicating relatively rapid removal of
aluminum. In the hypolinmion, elevated SI values were observed for several
months Indicating that mechanisms of aluminum removal were not as efficient
as the epilimnion.
Using the heat-budget approach of Powell and Jassby (1975) we computed
the diffusivity and, from concentration gradients, estimated the upward flux
of aluminum during summer stratification in Dart Lake. To illustrate the
relative significance of the upward flux of aluminum, both upward and downward
gross flux as well as the net vertical flux (the difference between upward and
downward gross flux) of aluminum is indicated in Figure 14 for both 6 and 14 m
depths at monthly intervals during the summer stratification period. (Values
of upward flux are not available for October and November because the thermocline
migrated below the 6 meter depth and therefore was within the well mixed upper
waters). Note that the upward flux of aluminum generally increased during the
stratification period. However, this mechanism of vertical aluminum transport
was insignificant when compared to the downward flux of aluminum. Although
estimates of the upward flux of aluminum are not available for turnover or
winter stratification periods, it would appear that the vertical transport
of aluminum in Dart Lake is characterized by a very pronounced downward com-
ponent .
-------
Over an annual cycle aluminum as retained in Dart Lake. Using surficial
sediment pools and observed sedimentation rates (0.35 cm yr ^ Heit et al., 1981;
Charles, 1983), and mass balance calculations, ve made two independent estimates
of the annual deposition of aluminum in Dart Lake (Table 6). Moreover, the
sediment trap monitoring program provides an estimate of the gross deposition
of particulate aluminum. In view of the evidence that upward transport of
aluminum is minor it is probably not unreasonable to assume that gross and
net deposition rates of aluminum should be similar. While these estimates
of aluminum deposition were obtained from independent approaches they are
in relatively close agreement (Table 6) and suggest that our estimates of
aluminum transport within Dart Lake may be reasonable.
General Discussion
It is apparent that there were pronounced temporal and spatial changes
in the water chemistry of Dart Lake. Such variations pose a problem to inter-
preting data from synoptic surveys. Depending on time of year and region of
the lake where the sample is collected, the water quality could vary markedly.
For example most surveys are conducted in the summer and samples are collected
from the epilimnion. On the basis of our observations such as sampling would
be a poor representation of "typical" lake conditions. Even if surveys are
accomplished over a few weeks and are only designed to provide information on
relative variations in water quality, short term events like precipitation, wind
entrainment, and biological transformations may substantially change lake water
chemistry. Therefore any interpretation of data collected from surveys of
acidic waters should be made with an understanding of spatial and temporal
fluctuations in water quality that may occur.
-------
When evaluating effects of elevated levels of aqueous aluminum considera-
tion should also be given to temporal variations that occur within acidic
lake ecosystems. Using our conceptual model we have computed aluminum pools
and fluxes for snowmelt (3/28 - 6/29; Figure 15) and summer base-flow conditions
(7/28 - 10/29; Figure 16). There were profound differences in the chemistry
and transport of aluminum for these two periods. During high flow conditions
the lake was accumulating aluminum and aluminum was generally conservative in
the lake system. Most of the drainage flux of aluminum during high flow was
exported from the lake outlet. During summer stratification the emphasis of
aluminum transport shifted from predominantly a horizontal flux to a substantial
vertical flux. Microbially mediated depletions of nitrate increased pH values
and induced in-lake formation of particulate aluminum which was to a large extent
retained in the lake. During this period a very high fraction of the fixed
solids deposited from the water column may be attributed to particulate aluminum
(Figure 17).
Elevated levels of aluminum may effect aquatic organisms by 1) acting as
a toxicant or 2) modifying the cycling of ecologically important substances
like phosphorus and organic carbon. Obviously snowmelt, with its associated
high concentration of aluminum, is a critical period for aquatic organisms in-
habiting acidic lakes. But less obvious and maybe of comparable importance are
the stratification periods. During these periods considerable aluminum is con-
verted from aqueous to particulate phases and solutes like DOC and orthophosphate
are probably removed from solution as a consequence. For example the removal of
DOC from the epilimnion of Dart Lake during the summer might be attributed to
coprecipitation with aluminum. Such a process is potentially significant for
acidic lake ecosystems because organic anions contribute to already low levels
-------
of acid neutralizing capacity (ANC) (Johannessen 1980; Driscoll and Bisogni,
1984). Organic anions can also complex aluminum and mitigate toxicity (Baker
and Schofield, 1982). Therefore, any process that results in DOC or organic
anion removal probably acts to diminish the water quality of acidic lakes.
While it is reasonably well established that aluminum can remove ortho-
phosphate from solution (Dickson, 1978) it may also act as a coagulant and
directly facilitate the removal of algae and other forms of particulate
phosphorus from the water column. Sorption and coagulation processes of
phosphorus removal, would be most pronounced during the summer season when
this element is in greatest demand by acidic lake ecosystems. Like DOC,
any removal of phosphorus would probably diminish water quality. In acidic
lakes, like Dart Lake, algal assimilation of nitrate is a significant mechanism
of H-A1-BNC removal (Driscoll and Schafran 1984, Appendix 2). Thus, any reduction
in the algal growth limiting nutrient phosphorus, would retard this process.
Baker and Schofield (1982) have suggested that hydroxyl complexes of mono-
meric aluminum are particularly toxic to fish. Fish mortality has been correlated
to elevated SI values (Baker and Schofield, 1982) and the sorption of aluminum
to gill surfaces has been observed during mortality (Muniz and Leivestad, 1980).
Within acidic lakes, like Dart Lake, high SI values are common during stratifica-
tion, particularly in the hypolimnion. Our large Kd values suggest that aluminum
has a high affinity for particulate surfaces; an observation that can undoubtedly
be extrapolated to biological surfaces. Grahn (1980) reported that a fish kill
of Cisco (Coregonus albula) occurred in two Swedish lakes when large inputs of
aluminum associated with a rainfall event, were followed by increases in pH
mediated by phytoplankton activity. Grahn (1980) hypothesized that formation
of hydroxy-aluminum and sorption to gill surfaces contributed to this fish kill.
-------
It is apparent that the extent to which aluminum is conservative within
acidic lake ecosystems varies considerably over the annual cycle. During
summer and winter stratification, considerable aqueous aluminum is converted
to particulate aluminum which is deposited to lake sediments. This non-con-
servative nature of aluminum has significant implications for the biogeochemistry
of acidic lake ecosystems.
Conclusions
From this study we conclude that:
1) there were significant temporal and spatial variations in the chemistry
and transport of aluminum in Dart Lake,
2) hydrogen ion and aluminum base neutralizing capacity were strongly correlated
with nitrate concentration,
3) during snowmelt, considerable quantities of nitrate, hydrogen ion and
aluminum entered Dart Lake through drainage inputs and passed through
the lake with limited retention,
4) during both summer and winter stratification microbially mediated depletions
of nitrate resulted in removal of inorganic aluminum
5) fluoride and organic complexes were the predominant form of aluminum in
Dart Lake and
6) there appears to be both allochtonous and autochtonous sources of alumino-
organic solutes in Dart Lake.
-------
Acknowledgement s
We would like to thank L.G. Barnum, D. Dickerson, K. Eager, N. Peters,
F.J. Unangst, L. Cordone, E. Southard and J. White for their assistance in
this study. Moreover we would like to thank the Project Officers R. Linthurst
and J. Baker for their encouragement during the investigation. The research
described in this report has been funded in part by the EPA/NCSU Acid
Deposition Program. It has not been subjected to the EPA's required peer
and policy review and therefore does not necessarily reflect the views of
the Agency and no official endorsement should be inferred.
-------
Table 1 Inorganic Aluminum Forms
3+
Aquo aluminum = [A1 ]
Hydroxide bound aluminum = A1-0H - [A1(0H)^+] + [A1(0H)2+] + [A1(0H)^]
Fluoride bound aluminum = Al-F » [A1F^+] + [A1F ] + [A1F ] + [A1F ] + (AlF.^ ] + [AlF,^ ]
I j 4 5 o
Sulfate bound aluminum = A1-S0^ = [A1S0^+] + [AlCSO^^]
3+ 2+
Aluminum base nuetralizing capacity » Al-BNC = 3[A1 ] + 3[A1»F] + 3[A1S0^] + 2[A1(0H) ] +
[A1(0H)2+] - [A1(0H)A"]
Hydrogen ion-aluminum base » H-A1-BNC = [H+] + Al-BNC - [OH ]
neutralizing capacity
-------
Table 2 Average Chemical Composition (peq'l of
Dart Lake 10/1981 - 11/1982 (n = 172)
Mean Standard Deviation
Base Neutralizing Capacity
(to reference pH = 8.3)
hydrogen ion 7.4 3.7
aluminum 19.6 8.2
carbonic acid 62.0 58.9
organic acids 3.0 1.2
total 92.0 59.0
Basic Cations
calcium 99.5 4.6
magnes ium 28.9 2.2
sodium 28.3 4.9
potassium 13.0 1.1
total 169.7 10.0
Anions
sulfate 137.6 11.0
nitrate 24.4 8.6
chloride 13.5 3.4
organic anions 12.2 0.4
bicarbonate 7.2 11.9
free fluoride 0.9 0.7
195.8 25.6
Sum of Cations^" 196.7 13.9
Sum of Anions 195.8 25.6
^"Cationic equivalence were computed to be hydrogen ion and aluminum BNC
plus basic cations.
-------
Table 3 Mean Concentration and Standard Deviation of Aluminum
(in yraol»l ) and Relative Distribution as Total and Monomeric Aluminum
Relative Distribution Relative Distribution
Aluminum Form
Concentration
as TA1
as MAI
Total Aluminum (TA1)
14.8
+
4.4
Monomeric Aluminum (MAI)
11.5
+
3.7
0.78
Inorganic Monomeric Aluminum
8.0
+
3.2
0.54
0.70
ai3+
1.9
+
1.7
0.13
0.17
Al-F
3.6
+
0.7
0.24
0.31
A1-0H
2.5
+
1.3
0.17
0.22
A1-S0,
4
0.1
+
0.1
0.01
0.01
Organic Monomeric Aluminum
3.5
+
1.8
0.24
0.30
Acid Soluble Aluminum
3.3
3.1
0.22
-------
Table 4 Annual Flux and Retention of Solutes
for Dart Lake 1981-1982
Parameter
TA1
IAM1
0MA1
ASA1
CI
DOC
Input''"
(mol* ha ^*yr *")
114.3
65.2
28.3
20.8
221
489
96.5
1976
Output
(mol-ha yr
110.9
62.0
27.4
21.5
213
509
100.2
2096
Retention Coefficient"''
0.064
0.085
0.066
-0.001
0.070
-0.008
-0.005
-0.026
1. Solute input was calculated by assuming that solute concentration in
non-inlet water (^ 3% of the input discharge) was similar to observed
values in inlet water.
-------
Table 6 Relationships Between Gross Deposition of
11 2
Aluminum (Dg-Al) and Total Solids Deposition (DG - TS)
2
6 Meter Depth r p <
Dg - A1 = 0.056 (Dg - TS) - 5.9; 0.44 0.0001
14 Meter Depth
Dg - A1 = 0.032 (Dg - TS) - 0.7; 0.72 0.0001
Total
Dg-Al - 0.036 (Dg - TS) - 1.4; 0.69 0.0001
1 ai -2 a-1
rag Al*m ¦d
2 -2 ,-1
mg *111 • q
-------
-2 -1
Table 7 Net Deposition of Aluminum (mmol»m • yr ) In Dart Lake 1981—1982
Method of Computation
1 2 3
Aluminum Form Mass Balance Recent Sediment Sediment Trap
Total
95
200
230
Inorganic
32
150
-
Organic
25
50
-
Acid Soluble
40
1. See Table 5
2. Computed assuming a sedimentation rate of 0.35 cm yr ^
3. Computed by multiplying the 0-6 m surface area (63 ha) by the sedimentation rate
observed in the 6 m sediment trap and the 6-15 m surface area (81 ha) by the
sedimentation rate observed in the 14 m sediment trap.
-------
Figure Titles
1. A morphometric map of the study site Dart Lake. Water samples were
collected at the inlet, outlet and a single deep water station.
2. Hypothetical calculations illustrating the potential effects of
changes in (a) the partial pressure of CO^ from ambient to atmospheric
levels and (b) the temperature from ambient to typical laboratory
levels on hypolimnitic Dart Lake solutions. The calculations were
made for a sample with an in-situ temperature of 5.5 °C and pH = 5.41,
-4 -6
dissolved inorganic carbon = 3.6 x 10 M, IMA1 = 6.3 x 10 M, fluoride =
4.0 x 10 ^ M and sulfate = 5.5 x 10 collected on 10/1/82 from a
depth of 14 m in Dart Lake with the chemical equilibrium model MINEQL
(Westall et al., 1976).
3. Isopleths of (a) nitrate, (b) pH and (c) inorganic monomeric aluminum
(XMA1) within Dart Lake for the study period. Samples were collected
at 0, 2, 4, 6, 9, 12, and 14 meters. Sampling dates are indicated by
A.
4. Values of (a) pH,(b) inorganic monomeric aluminum (IMA1), (c) nitrate,
(d) sulfate, (e) dissolved organic carbon (DOC) and (f) organic
monomeric aluminum (0MA1) for the inlet and outlet streams of Dart
Lake over the study period.
5. An isopleth of the saturation index (SI) of Dart Lake water with
respect to the solubility of microcrystalline gibbsite (Roberson
and Hem, 1969).
6. The concentration of (a) inorganic monomeric aluminum (IMA1) and
(b) aquo aluminum (Al^+), (c) fluoride bound aluminum (Al-F),
(d) hydroxide bound aluminum (A1-0H) and (e) sulfate bound aluminum
(Al-SO^,) which comprise IMA1 in Dart Lake.
-------
7. Isopleths of (a) dissolved organic carbon (DOC) and (b) organic
monomeric aluminum (0MA1) for Dart Lake.
8. The distribution of sediment fractions of aluminum for Dart Lake.
The fractions include (a) exchangeable aluminum, (b) oxide associated
aluminum, (c) aluminum associated with organic matter and (d) residual
aluminum.
9. Conceptual model of aluminum pools and transport for an acidic
drainage lake.
10. Average annual pools and fluxes of aluminum for Dart Lake.
11. Input and output of aluminum associated with drainage flow of
Dart Lake.
12. The deposition of particulate aluminum collected in upper (6m)
and lower (14m) water sediment traps in Dart Lake. Error bars
indicate standard deviation of triplicate sample.
13. In-lake distribution coefficients (Kd) computed for upper (6m) and
lower (14m) water sediment trap samples in Dart Lake. Aluminum
distribution coefficients are a measure of the affinity of aqueous
aluminum for particulate matter (mol A1 gm particulate matter ^(mol
-3 -1
aqueous Al-Cm )
14. A comparison of downward (as monitored with sediment traps) and upward
(as determined with values of vertical diffusivity and concentration
gradients) flux of aluminum for the summer stratification period in
Dart Lake. The net vertical flux is tabulated as the difference
between downward and upward fluxes.
15. Fluxes and pools of aluminum in Dart Lake during high flow conditions
(3/28/1982 - 6/29/1982).
16. Fluxes and pools of aluminum in Dart Lake during low flow summer conditions
(7/28/1982 - 10/29/1982).
-------
DART LAKE
74° 52' W, 43°48' N
Surface Area I44ha
Volume 1.02x10'°*.
Inlet
v
j
k r _
Outlet
X
- Water Column and Sediment Trap
Sampling Station
Contours are in 3 meter intervals
-------
<->4.0
5.8 6.0 6.2 6.4
10 15 20
TEMPERATURE l°C)
Fi o -
Z
-------
0) NOj ISOPLETH (yeq I"')
E
b) H-BNC ISOPLETH (jueq
C) AI-BNC ISOPLETH Coea-I'1)
16 20 24
~Z Z 6. t\ A i 6 A S"
J F M A M J J
TIME ( months)
P»au,r-e_ 3
-------
Fio^v-e 4
V J I
Inlet
o Outlet
80
„ 60
&
W 40
20
500
400
O 300
Q200
100
8
<.
24
O
2
o
E
c
m
O
D
O
>
-1 1 1 1 I I I L.
-I 1 I I I I I I
1 '
OND JFMAMJ J ASON
-------
SI Microcrystalline Gibbsite (Log Kso = 9.35)
ice cover
tn
o
o
E
.3.5
i
b-
CL
LU
Q
1.5
10
M J
P'i 5"
-------
r t car a G
0
IMAL
Al-
18'
16
14-
12-
10
8
6-
4-
2
9-
8-
7-
G
5-
4
3
2
1
Al-F
5
A
3-
2
1
a)
°° 0 °°&o
CD OD
O
1 O
000
0 o®c
° CD" O ol
o 0aT
O H 5 _ 0 » 0 c
• % «%|f|
o o
b)
Cb O { °
AI-OH
" d)
o o
o.
8 oc
J0=«E e
, So Fo o o oJ ooO o
CD CP
O 0
o o 0
O Or
1ft
Jo o
GD 05
O o
O Oq O
0.4
e)
0.3-
Al-S04
0.2-
0.1
00 0° 0
o ° C of® 0 c
as 2, ^ 0
® O
o
00,
fir®
OO
- tfy
0 *c 1
40 47 48 49 50 51 52 53 54 55 5G
pH
-------
Q) Dissolved Organic Carboi^nimo! C I ')
ice cover (
V////////////////////'/////s f77T
\Y/\v| \y
\ \35 Ws \j
.35
.35
.29
.23
.35
.26
.26
.29
.32
.29
?32
r35'
F M
A M
J
J
S
0
N
A
D
ice cover b) Organic Monomeric Aluminum CumoN"1)
y/////////////////;;;/;;;; //m
I 8
A A
1
-------
SEDIMENT ALUMINUM FRACTIONS
SEDIMENT, depth
(cm)
i i i
1
2
s
y/////////////////A
3
\
1—.
4
5
V////////////////A I—^
6
i—i
8
V/////////////A H
10
y///////////A: ii
12
y///////////A "h
14
y////////////A. .H
16
'///////////z, W
20
mm,..h
i 1 1
0
0.5 1.0 1.5
SEDIMENT ALUMINUM(mmols/g)
Legend 1
ISAWWM EX
OX
Figure Z
-------
vO/llvkP>wi4L »¦ O k. _ . L i R . C 11 i * E r-\ tm U iv § § iJ l/w C I /» w I J. Ci»¦ C K
epillmnion
d r a I n a g
exports of
drainage
inputs of Al
(precipitation)
(oxidation?)
(sorption on
hydrous oxides)
soil
sediment
diffusion
seepage
exports?
hypolimnion
(precipitation)
(oxidation?)
(sorption on
hydrous oxides?)
sedlmentatio
(minor alizatlon?)
(desorptlon?)
(dissolution?)
(desorptlon ?)
organic monomerlc
a I u mI n u m- o M A I
Inorganic monomerlc
aluminum - IMAI
IMA I
organic AI
0 MAI
sediments
sedimentation
0 MAI
IMA I
(dissolution)
hydrous Al
o x I d e 8
hydrous Al
o x I d e 8
DOC
hydrous Al
oxides
IMAI?
OMAI?
OMAl?
IMAI?
hydrous
A I oxides
OMA I
IMA I
hydrous
A I oxides.
OMA I
IMA I
F
. - • 1 ;
Q
-------
ANNUAL FLUXES AND POOLS OF ALUMINUM IN DART LAKEUO'I9B2-iu/I9c«i
TAI
.001
DIRECT
INPUTS
Input
X
Upper Woters (< 6m)
lMA-£iOMAf'ASA£
1.3 i 0.6 !o.5
Gross
Deposition
Lower Waters (>6m)
TA1 •
IMA-f-
OMA?-
ASAJ ¦
TOTAL At
INORGANIC
MONOMERICAI
ORGANIC
MONOMERIC At
ACID SOLUBLE M
Gross
Sediment (I cm)
IMAi OMAi'ASA-l
IMAciOMAi ASA-f
TM
0.74
Net Deposition
'IMAliOMAt'ASAt
TM
Ex
Ox
Org
Res
570
II
370
130
60
Outlet
mM: OMAt'ASAt
Flux mmol-rn - a
Pool mmol • m"
F,.-
1 ^ ' A
C" T™*
10
-------
C4* pA/><
<20
on
Q.
-+->
O 12
"O
Q.
_J L
I I I I I
O N D J F M
r>i •+ /^J i i rr*| j r-u I pp
--~--OUT
-------
SEDIMENT TRAP
o 6m
~ 14m
~o
c\j*
0.3
P"iO .AT-C I 2_
-------
Kvc^re. 13
O Upper Waters (0-6m )
x = 4.91 ± 0.54
~ Total Water Column
( 0-14m ) -
x = 4.93 i 0.18
DJ FMAMJ JASON
-------
Figure I1-)
X
3
k'-8
cr
I A
=> 0
z.
.4
b
' tO Q
-8
•O CL
Wl?
CM
'!§
-------
LI iS JD J0_ OF
ur ur j it _ u<_ ju
CONDITIONS
(3/28/1982 - 6/29/1982 )
3H _C
DIRECT
INPUTS
Inlets
Upper Waters (<6m)
TA-t JIM At jOMAfjASAt
5.2 13.0 U.5 10.8
i i i
Gross
Deposition
Lower Waters (>6m)
TAI - TOTAL M
IMAt- INORGANIC
MONOMERIC At
OMA-t- ORGANIC
MONOMERIC AI
ASAl- ACID SOLUBLE At
Gross
Deposition
At OMAfciASAt
IMAplQMAt'ASAl
TA-t
D.66
Net Deposition
TAE |IMA{i OMAtiASA-t
0.13 lo.48l-.43 I 0.09
Sediments
Outlets
TA-t 'iM AlJ OMA-ti ASA&
5.2 | 2.8 | 1.4 | 1.0
i i
Flux
mmol • rrf2 d~
Pool
m mol • m~2
Ficyj-r-e. /S*
-------
F i_uXLo ANu PC wi_S ui Ai_wi D^-..». L — XEL ^JF—3 M.— R
LOW FLOW CONDITIONS
(7/28/I982-IO/29/1982)
TAi
.001
DIRECT
INPUTS
Waters (< 6m)
Input
TAt 'iMAJjOMAljASAt
1.2 |0.5 i 0.3 104
' ' L
TAt ' IM At ] OM At1 ASAt
38 ! 18 i 10 t 9
Output
TAt jlMAE'OMAl'ASAt
0.9 i0.4 ! 0.2 |0.2
Gross
Deposition
Lower Waters (>6m)
TAt —TOTAL At
IMAt— INORGANIC
MONOMERIC At
OMAt— ORGANIC
MONOMERIC At
ASAt —ACID SOLUBLE At^
TAt jIMAliOMAtlASAt
32 ! 20 i 8 ! 4
Gross
spe position
TAt
105
Net Deposition
TAt llMA^OMAl'ASAt
0.38!0.21 '-.03 ! .20
Sediments
Upward
Flux
Flux mmol• m" d"
Pool mmol-rrr*
F \ xr<£_ I(o
-------
References
Baker, J. and C. Schofield. 1982. Aluminum toxicity to fish in acidic
waters. Water, Air and Soil Poll. 18:289-309.
Barnes, R.B. 1975. The determination of specific forms of aluminum
in natural water. Chem: Geol. 15:177-191.
Bloesch, J. and N.W. Burns. 1980. A critical review of sediment trap
technique. Schwiez. A. Hydrol. 42:15-55.
Braekke, F.H. ed. 1976. Impact of Acid Precipitation on Forest and
Freshwater Ecosystems in Norway, SNSF report FR6/79. Oslo, Norway.
Charles, D. 1983. Recent pH history of Big Moose Lake (Adirondack
Mountains, New York, USA) inferred from sediment diatom assemblages.
Verh. Internat. Verein. Limnol. (in press).
Corey, R.B. 1981. Adsorption vs. precipitation, ^n Adsorption of
Inorganics at Solid-Liquid Interfaces. M.A. Anderson and A.J.
Rubin, eds. Ann Arbor Science, Ann Arbor, MI. pp. 161-182.
Cronan, C.S. and C.L. Schofield. 1979. Aluminum leaching response to
acid precipitation: Effects on high-elevation watershed in the
northeast. Science 204:304-306.
Davis, J.A. 1982. Adsorption of natural dissolved organic matter at
the oxide/water interface. Geochim. et Cosmochim. Acta 46:681-692.
Dickson, W. 1978. Some effects of the acidification of Swedish lakes.
Verh. Internat. Verein. Limnol. 20:851-856.
Dickson, W. 1980. Properties of acidified water, ^n Ecological Impact
of Acid Precipitation. Drablos, D. and A. Tollan, eds. SNSF-project,
Oslo, Norway, pp. 75-83.
Driscoll, C.T. 1984. A procedure for the fractionation of aqueous aluminum
in dilute acidic waters. Intern. J. Environ. Anal. Chem. (in press).
Driscoll, C.T. And G.C. Schafran. 1984. Characterization of short term
changes in the base neutralizing capacity of an acidic Adirondack, NY
lake. Nature (submitted for publication).
Driscoll, C.T. and J.J. Bisogni. 1984. Weak acid/base systems in dilute
acidified lakes and streams in the Adirondack region of New York State.
In Modeling of Total Acid Precipitation Impacts. J.L. Schnoor, ed.
Ann Arbor Science, Ann Arbor, MI. pp. 53-72.
Driscoll, C.T., J.P. Baker, J.J. Bisogni, Jr., and C.L. Schofield.
1980. Effect of aluminum speciation on fish in dilute acidified
waters. Nature 284:161-164.
-------
Driscoll, C.T. , J.P. Baker, J.J. Bisogni and C.L. Schofield. 1984.
Aluminum speciation and equilibria in dilute acidic surface waters of
the Adirondack region of New York State. In Geologic Aspects of
Acid Rain. O.P. Bricker, ed. Ann Arbor Science, Ann ARbor, MI.,
pp. 55-75.
Galloway, J.N. and G.E. Likens. 1979. Atmospheric enhancement of metal
deposition in Adirondack Lake sediment. Limnol. Oceanogr. 24:427-433.
Galloway, J.N., C.L. Schofield, G.R. Hendrey, N.E. Peters, and A.H. Johannes.
1980. Sources of acidity in three lakes acidified during snowmelt.
In Ecological Impact of Acid Precipitation. Drablos, D. and A. Tollan,
eds. SNSF-project, Oslo, Norway, pp. 264-265.
Grahn, 0. 1980. Fishkills in two moderately acid lakes due to high
aluminum concentrations. _IH Ecological Impact of Acid Precipitation.
Drablos, D. and A. Tollan, eds. SNSF-project, Oslo, Norway.
Hall, R.J., C.T. Driscoll and G.E. Likens. 1984. Physical chemical and
biological consequences of episodic aluminum additions to a stream
ecosystem. Limno. Oceanogr. (in press).
Heit, M., Y. Tan, C. Klusek, and J.C. Burke. 1981. Anthropogenic trace
element and polycyclic aromatic hydrocarbon levels in sediment cores
from two lakes in the Adirondack acid lake region. Water, Air and
Soil Poll. 15:441-464.
Hem, J.D. and C.E. Roberson. 1967. Form and stability of aluminum
hydroxide complexes in dilute solution. U.S. Geol. Surv. Water
Supply Pap. 1827-A, Washington, D.C.
Henriksen, A., and H.M. Seip. 1980. Strong and weak acids in surface waters
of Southern Scotland. Water Res. 14:809-813.
2+
Hohl, H., and W. Stumm. 1976. Interactions of Pb with hydrous Y-Al 0,.
J. Coll. Int. Sci. 55:281-288. 3
Hultberg, H., and S. Johansson. 1981. Acid groundwater. Nordic Hydrology,
12:51-64.
Hutchinson, T.C. and M. Havas (eds.) 1980. Effects of Acid Precipitation
on Terrestrial Systems. Plenum Press, New York, 656 pp.
Hutchinson, T., W. Gizyn, M. Havas, and V. Zobens. 1978. Effect of long-
term lignite burns on arctic ecosystems of the Smoking Hills, N.W.T.
_In Trace Substances in Environmental Health XII. D.D. Hemphill, ed.,
University of Missouri, Columbia, MO, pp. 317-332.
Isachsen, Y.W. and D.W. Fisher. 1970. Geologic map of New York - Adirondack
sheep. New York State Museum and Science Map and Chart Series No. 15.
-------
Johannessen, M. 1980. Aluminum, a buffer in acidic waters. In Proceedings
of the International Conference on Ecological Impact of Acid Precipitation,
D. Drablos and A. Tollan, eds., SNSF Project, Oslo, Norway pp. 222-223.
Johannes, A.H. 1984. Unpublished data. Department of Chemical and Environ-
mental Engineering, Rensselaer Polytechnic Institute.
Johannes, A.H., J.N. Galloway and .DE. Troutman. 1980. Snowpack storage
and ion release. In Ecological Impact of Acid Precipitation, Drablos, D.
and A. Tollan, edsTT SNSF Project, Oslo, Norway pp. 260-261.
Johnson, N.M., C.T. Driscoll, J.S. Eaton, G.E. Likens and W.H. McDowell, 1981.
Acid rain, dissolved aluminum and chemical weathering at the Hubbard
Brook Experimental Forest, New Hampshire. Geochim. Cosmochim. Acta
45:1421-1437.
Kamphake, L.J., S.A. Hannah, and J.M. Cohen. 1967. Automated Analyses for
Nitrate by Hydrazine Reduction. Water Res. 1:206-216.
Lazrus, A., E. Lorange and J.P. Lodge. 1968. New automated microanalyses
for total inorganic fixed nitrogen and for sulfate ion in water.
In Trace Inorganics in Water Advances in Chemistry Series 73,
rTa. Baker, ed. American Chemical Society, Washington, D.C.
pp. 164-171.
Lind, C.J. and J.D. Hem. 1975. Effects of organic solutes on chemical
reactions of aluminum. U.S. Geol. Surv. Wat. Sup. Pap. 1827-G.
Menzel, D.W. and R.F. Vaccaro. 1964. The measurement of dissolved organic
and particulate carbon in seawater. Limnol. Oceanogr. 9:138-142.
Muniz, I.P. and H. Leivestad. 1980. Toxic effects of aluminum on the
brown trout, Salmo trutta L. In Ecological Impact of Acid Precipitation.
Drablos, D. and A. Tollan, eds. SNSF-project, Oslo, Norway, pp. 320-321.
Norton, S.A. and C.T. Hess. 1980. Atmospheric deposition in Norway during
the last 300 years as recorded in SNSF lake sediments. Sediment dating
and chemical stratigraphy. _Iri Ecological Impact of Acid Precipitation.
Drablos, D. and A. Tollan, eds. SNSF Project, Oslo, Norway, pp. 268-269.
Orion, Instruction Manual. 1976. Fluoride Electrodes Orion Research Inc.,
Cambridge, MA.
Powell, T., and A. Jassby. 1974. The estimation of vertical eddy diffusivities
below the thermocline in lakes. Water Resour. Res. 10:191-198.
Roberson, C.E. and J.D. Hem. 1969. Solubility of aluminum in the presence of
hydroxide, fluoride, and sulfate. U.S. Geol. Surv. Water Supply Pap.
1827-C, Washington, D.C.
-------
Schafran, G.C. 1984. Aluminum chemistry in a dilute acidic lake in the
Adirondack region of New York State, M.S. Thesis, Department of Civil
Engineering, Syracuse University.
Schafran, G.C. and C.T. Driscoll. 1984. Temporal and spatial variations in
the water chemistry of a dilute acidic drainage lake (in preparation).
Schofield, C.L. 1976. Lake acidification in the Adirondack Mountains of
New York: Causes and consequences, p. 477. _In Proceedings of the
First International Symposium on Acid Precipitation and the Forest
Ecosystem. L.S. Dochinger and T.A. Seliga, eds. U.S.D.A. Forest
Service Tech. Rept. NE-2.
Stainton, M.P. 1973. A syringe gas-stripping procedure for gas-chromato-
graphy determination of dissolved inorganic and organic carbon in
freshwater and carbonate in sediments. J. Fish. Res. Board Can.
30:1441-1445.
Standard Methods for the Examination of Water and Wastewater 1980. 15th Ed.
American Public Health Association.
Stumm, W. and J.J. Morgan. 1981. Aquatic Chemistry. John Wiley and Sons,
Inc., Mew York, NY.
Tessier, A., P.G.C. Campbell and M. Bisson. 1979. Sequential extraction
for the speciation of particulate trace metals. Analytical Chem.
51:844-851.
Vangenechten, J.H.D. and O.L.J. Vanderborght. 1980. Acidification of
Belgian moorland pools by acid sulphur-rich rainwater. _In Ecological
Impact of Acid Precipitation. Drablos, D. and A. Tollan, eds. SNSF-
project, pp. 246-247.
Westall, J.C., J.L. Zachary and F.M.M. Morel. 1976. MINEQL a computer
program for the calculation of chemical equilibrium of aqueous systems.
Ralph M. Parsons Laboratory for Water Resources and Environmental
Engineering, Civil Engineering Department, Massachusetts Institute
of Technology, Technical Note No. 18.
White, J.R. 1984. Trace metal chemistry and cycling in a dilute acidic
lake in the Adirondack region of New York State, Ph.D. Dissertation,
Department of Civil Engineering, Syracuse University.
Wright, R.F. and E. Snekvik. 1978. Acid precipitation chemistry and fish
populations in 700 lakes in southernmost Norway. Verh. Internat. Verein.
Limnol. 20:765-775.
Zall, D.M., D. Fishers, M.D. Gardner. 1956. Photometric determination of
chlorides in water Anal. Chem. 28:1665-1668.
-------
APPENDIX 1
Driscoll, C.T. 1984. A procedure for the fractionation of aqueous aluminum
in dilute acidic waters. Intern. J. Environ. Anal. Chem. (in press).
-------
A Procedure for the Fractionation of Aqueous Aluminum
in Dilute Acidic Waters
Charles T. Driscoll
Department of Civil Engineering
Syracuse University
Syracuse, New York 13210
ABSTRACT
A procedure was developed for the fractionation of aqueous aluminum.
This procedure results in the determination of acid soluble aluminum, non-
labile monomeric aluminum and labile monomeric aluminum. Acid soluble
aluminum is thought to include colloidal aluminum and extremely non-labile
organic complexes. Non-labile monomeric aluminum is thought to include mono-
meric alumino-organic complexes. Labile monomeric is comprised of aquo
aluminum as well as inorganic complexes of aluminum. The inorganic specia-
tion of aluminum may be calculated by using labile monomeric aluminum, pH,
fluoride and sulfate data with a chemical equilibrium model.
This procedure was evaluated using synthetic and natural water solu-
tions. In natural waters, levels of labile monomeric aluminum increased
exponentially with decreases in pH below 6, while non-labile monomeric
aluminum was strongly correlated with organic carbon concentration. Non-
labile monomeric aluminum was observed to be relatively insensitive to
changes in solution pH. Results of the aluminum fractionation procedure
were in relative agreement with an independent evaluation using the fluoride
ion selective electrode.
KEYWORDS: Acidic Deposition, Aluminum, Alumino-organic complexation,
Aluminum Speciation
-------
INTRODUCTION
Aluminum is the third most abundant element within the earth's crust^. It
occurs primarily In aluminosilicate minerals, most commonly as feldspars In
^etamorphic and igneous rocks and as clay minerals in well weathered soils. In
high elevation, northern temperate regions, the soils encountered are generally
2
podzolic . The process of podzolization involves the transport of aluminum from
upper to lower soil horizons by organic acids leached from foliage as well as from
decomposition in the forest floor^*^'^'^. Ugolini et al.^ have observed that
during podzolization there is little mobilization of aluminum from the adjacent
watershed to surface waters. Concentrations of dissolved aluminum are low in
most circumneutral waters due to the relatively low solubility of natural aluminum
8 -1
minerals. Sturam and Morgan report a median aluminum value of 10 yg Al«l for
9
terrestrial waters, while Bowen gives a higher average concentration of
240 yg Al«1 ^ for freshwaters, however this includes bogs.
It has been hypothesized that mineral acids from acidic depostion have remo-
bilized aluminum previously precipitated within the soil during podzolization or
held on soil exchange sites^. Elevated levels of aluminum have been reported
for acidic waters within regions that are receiving elevated inputs of acidic
10,11,12,13,14,15,16
deposition »»»»•'.
Elevated levels of aluminum in dilute (low ionic strength) acidic waters are
13 17 18 19
of interest because: (1) aluminum is an important pH buffer » » » f (2)
13 20
aluminum may influence the cycling of important elements like phosphorus '
13 21 22
and organic carbon ' , and (3) aluminum is potentially toxic to aquatic and
23
terrestrial organisms . An understanding of the speciation of aluminum is
essential for the evaluation of these processes.
-------
Dissolved monoraeric aluminum occurs as aquo aluminum, as well as hydroxide,
Ik 25.
fluoride, sulfate and organic complexes ' * Past investigations of aluminum
10 26 27
have often ignored non-hydroxide complexes of aluminum * ' . More recent
28 29
studies of Driscoll et al. and Johnson et al. have demonstrated the signifi-
cance of organic and fluoride complexes of aluminum in dilute acidic surface waters.
22
Baker and Schofield observed that aluminum toxicity varies with pH and
life history stage of white suckers (Catostomus commersoni) and brook trout
(Salvelinus fontinalis). Aluminum solutions (>100 to 200 yg Al-1 resulted
in reduction of survival of fish larvae. Aqueous hydroxy-aluminum forms were
considered to be the most toxic to fish. pH, natural organic and inorganic
— 22 28
(e.g. F ) ligands significantly influence aluminum toxicity ' . Elevated
levels of soluble ligands restrict aluminum hydrolysis and mitigate aluminum
toxicity to fish.
23
In terrestrial systems, Ulrich et.al. have hypothesized that acidic deposi-
tion has reduced forest productivity. Vegetation mortality is thought to occur
from dissolution of aluminum mediated by mineral acids and subsequent uptake by
fine roots. While considerably less well understood than aquatic toxicity, the
30
work of Moore suggests that aluminum toxicity to vegetation is also linked to
aqueous speciation.
In view of the current interest in the effects of acidic deposition and the
biogeochemistry of aluminum, it is desirable for researchers to be able to
analytically differentiate between forms of aqueous aluminum. In this paper I
will present an analytical procedure that can be used to fractionate aqueous
aluminum, and an evaluation of this procedure using synthetic and natural water
samples. Potential errors associated with this technique will also be discussed.
-------
ANALYTICAL METHODS
In this methodology three measurements of aluminum were made:
Acid Reactive Aluminum (Air)
Solutions were acidified to a pH value of 1 for one hour and analyzed
31
using a modification of the method of Barnes . The original procedure was
developed to detect levels of aluminum down to 2 yg A1 • 1 Because this level
of sensivity is generally not necessary for acidic natural waters I increased
the volume ratio of solvent, methyl isobutyl ketone (MIBK), to sample from
1:27 to 1:4.
Monomeric Aluminum (Ala)
Solutions were rapidly extracted with 8-hydroxyquinoline in MIBK as
31
described by Barnes
Non-labile Monomeric Aluminum (Alo)
Non-labile monomeric and labile monomeric aluminum were separated by passing
an aliquot of sample through a column of strongly acidic cation exchange resin
(Amberlite 120). The cation exchange column used in this study was 1 cm in
diameter and contained 9.5 ml of prepared resin. Batches of resin were pre-
pared by displacing some of the exchangeable hydrogen ion with sodium, resulting
in a resin that contained both hydrogen and sodium ions on exchange sites. The
amount of sodium on the exchanger was adjusted such that when an eluant of
comparable ionic strength to the solutions being analyzed was passed through
the exchanger, the effluent pH was similar to the pH of the solutions being analy-
zed. In this way the pH change that samples experienced when processed through
the exchange column was minimized.
-------
This procedure was originally developed to evaluate aluminum chemistry in
surface waters and soil solutions from the Adirondack region of New York and
the Hubbard Brook Experimental Forest (HBEF), New Hampshire. An eluant ionic
strength of 3 x 10~\ which was comparable to the mean ionic strength of these
29 32
waters was used ' . In addition, two columns were used with prepared exchanger
that resulted in effluent target pH values of 5 and 7. Target pH values of 5
19
and 7 were chosen because poorly buffered waters in New York and in New
29
Hampshire often have pH values in this range. However, an exchange resin should
be prepared so that the pH of processed samples will be comparable to the pH of
the solutions being analyzed. If the ionic strength of a sample was less than
the ionic strength of the eluant, then the pH of a sample after passage through
the exchanger was greater than the target pH. Conversely, if the ionic strength
of a sample was greater than the eluant, then the pH of the processed sample was
generally less than the target pH of the exchanger.
An aliquot of water sample was placed in a sample reservoir and was passed
through the cation exchange column with a peristaltic pump. After an initial
volume of sample (50 ml to displace the eluant) was discarded, a volume of
exchanger processed sample, sufficient to perform the aluminum determination,
-4 -1
was collected. After the sample was processed, eluant (3 x 10 mol NaCl'l )
was passed through the column to rinse the exchanger bed prior to introduction
of the next sample. Samples were extracted for analysis of aluminum immediately
31
after processing through the cation exchange column using the method of Barnes
With these three measurements of aluminum, three aluminum fractions were
determined (Figure 1). Non-labile monomeric aluminum (Alo) was measured directly
and is an estimate of monomeric aluminum that is organically complexed. Labile
monomeric aluminum was determined by the difference between monomeric aluminum
-------
and non-labile raonomeric aluminum (Ala-Alo). This fraction of aqueous aluminum
would include aquo aluminum as well as hydroxide, sulfate and fluoride complexes
of monomeric aluminum. Acid soluble aluminum, which is acid reactive aluminum
less total monomeric aluminum (Air - Ala), represents an estimate of the aluminum
that requires acid dissolution for the determination. This fraction would
include colloidal aluminum, polymeric aluminum and very strongly bound alumino-
organic forms.
The inorganic speciation of aluminum can be calculated by using measured
values of labile monomeric aluminum, pH, fluoride and sulfate with one of the
33
chemical equilibrium models that are currently available . Chemical equilibrium
relationships for aqueous aluminum species used in this study are listed in
16 29
Table 1 and these calculations are summarized by Driscoll and Johnson et al.
To independently evaluate the aluminum fractionation procedure I used the
fluoride ion selective electrode. Free fluoride was determined by direct poten-
tiometric determination using the fluoride ion selective electrode. Total
fluoride was determined by electrode after addition of a total ionic strength
1 I
adjustor and buffer (TISAB II) solution . TISAB decomplexes fluoride, provides
a constant solution ionic strength, and adjusts solution pH so that hydrogen ion
and hydroxyl interference are minimized. With values of free and total fluoride,
aquo aluminum (Al ) activity was calculated with thermodynamic relationships
(Table II). In turn, inorganic species of aluminum and organic aluminum were
calculated (Table II). While it is possible to use the fluoride ion selec-
tive electrode as an independent method to fractionate aluminum, this procedure
has not been fully evaluated. Therefore, I have chosen to use the fluoride elec-
trode method just as a quality assurance check on the cation exchange separation
procedure.
-------
For additional analytical details or description of the study sites from
which the Adirondack New York or HBEF surface water samples were collected see
16 29
Driscoll or Johnson et al. respectively.
DISCUSSION OF THE PROCEDURE
Short-term transformations of aqueous aluminum are generally dependent on
the concentration of raonomeric aluminum (or more specifically the activity of
aquo aluminum), rather than particulate aluminum. Because of the tendency
for aluminum hydroxy cations to polymerize through double OH bridging when values
of solution pH exceed 4.5 , a considerable fraction of the "dissolved" aluminum
reported in many analyses of natural water having neutral or slightly acidic pH
values may consist of suspended microcrystals of aluminum hydroxide. Filtration
of samples through 0.4 ytn pore size membranes, a common practice in clarifying
39
natural water prior to analysis, may fail to remove such material . Therefore,
it is desirable to use an analytical technique that is selective for monomeric
aluminum, rather than assuming that filtration or centrifugation will remove all
suspended material. While detection of aluminum by other analytical techniques
may be used with this fractionation procedure, the technique of complexation by
8-hydroxyquinoline and rapid extraction has the advantage of selecting for mono-
meric aluminum. This technique has been evaluated and discussed by numerous
31, 38, 40, 41, 42, 43, 44, 45
researchers
The cation exchange resin has a strong affinity for aluminum. When solutions
are passed through the cation exchange column, there is a competition between
aqueous ligands and the exchanger for aluminum. Addition of synthetic aluminum
solution (18 ymol Al l"1 as A1K(S04)2 and A1C13 at pH 5) to the cation exchange
-------
column resulted in complete removal ( < 0.1 pmol Al-1 ) at all application
rates evaluated (up to 6.3 ml«min \ ml of exchanger bed volume ). Application
of synthetic aluminum fluoride solutions (18 ymol Al«l ^ with 1-10 ymol F- 1
at pH 5) to the cation exchange column revealed that the exchanger competes very
effectively with fluoride for aluminum. Aluminum was completely removed from
solution (< 0.1 ymol • l"1) and total fluoride in the effluent equaled influent
levels. Treatment of synthetic and natural solutions by the cation exchange
column resulted in the conversion of fluoride from predominately aluminum bound
fluoride to free fluoride.
It would appear that aquo aluminum was stripped from inorganic ligands by
the polar cation exchange resin and therefore inorganic forms of aluminum
(e.g. Al3+, AlF2+,-A10H2+) were readily removed from solution. Organic ligands,
however, form strong complexes with aluminum and therefore more effectively com-
pete with the cation exchange resin. Synthetic solutions of aluminum in the
presence of a strong ligand, sodium citrate, (18 ymol A1 • 1 \ 1 mmol sodium
citrate • l-1 at pH 5) were applied to the cation exchange column. The level
of sodium citrate used was comparable to levels of dissolved organic carbon
16 29
(DOC) observed in Adirondack New York and HBEF surface waters on an organic
carbon basis. In this experiment, detection of aluminum in influent and efflu-
ent solutions was complicated because citrate effectively competes with 8-hydro-
xyquinoline for aluminum and therefore interfered with the analytical determina-
tion. As a result, in sodium citrate solutions aluminum had to be determined
using the less sensitive and less reproducible method of direct determination by
atomic absorption spectrophotometry (AAS) with graphite furnace. Although
aluminum determinations by AAS should be viewed with caution, the results of this
experiment suggested that aluminum was effectively transported through the cation
exchange column in the presence of relatively high levels of citrate.
-------
Analog organic ligands with a strong affinity for aluminum (such as citrate)
not only interfere with the 8-hydroxyquinoline determination of aluminum but are
not entirely representative of organic ligands occurring in natural waters.
Therefore, this fractionation procedure was evaluated in more detail using
samples collected from acidic and non-acidic surface waters in the Adirondack
region of New York^* and at the HBEF New Hampshire^'
The measured concentration of non-labile monomeric aluminum in natural
water samples was somewhat dependent on the flow rate of solution through
the column (Figure 2). My results indicate at low application rates to the
column, the amount of monomeric aluminum (non-labile) passing through the
column was sensitive to flow. However, with higher application rates (above
2.7 ml • min 1 • ml of exchanger bed volume the concentration of aluminum
passing through the column became constant. These results suggest that
natural aluraino-organic complexes exhibit a range of stability. Similar
47
results have been demonstrated by Means et al. , who observed labile
metal complexes as well as very stable complexes in natural water samples.
The longer the sample retention time in the column, the greater the dis-
ruption of alumino-organic species by the resin. This disruption was minimized
by operating the exchange column at a relatively high application rate (3.7 -
A.2 ml* min-1* ml of resin bed volume *") . As mentioned previously, leakage
of inorganic aluminum through the column was not observed at these application
rates.
In this analysis It was assumed that organic matter and alumino-organic
complexes passed through the exchanger while aluminum associated with inor-
ganic forms was retained on the exchanger, resulting in an effective parti-
tioning between inorganic and organic forms of monomeric aluminum. Several
-------
sources of evidence suggest that this assumption was reasonably valid. (1) In
samples from several Adirondack lakes and streams containing a wide range of
total organic carbon concentration (0.17 - 1.20 mmol C*1 and pH values
(4.0 - 7.2), the amount of TOC leaving the column ranged from 93 - 105% of
that entering the column. Unfortunately, a mass balance on organic carbon is
unsuitable to assess the extent to which dissolved organic carbon and organic
forms of aluminum are retained within the column due to the potential for organic
carbon leaching from the organic resin.
(2) Perhaps a better indicator of the limited extent to which organic carbon
was retained within the resin was that there was little difference in the ultra-
violet spectral (200 - 400 nm) pattern of exchange column influent and effluent
(Figure 3). It is evident that absorbance of column effluent exceeded influent
absorbance at low wavelengths in some samples. The nature of this discrepancy
was not apparent, although organic carbon leaching from the resin cannot be
disregarded. UV absorbance patterns observed in this study were similar to those
presented by Schnitzer and Khan^ as being typical of naturally occuring organic
carbon.
(3) Additional evidence for the validity of this fractionation technique is
available from field observations. Levels of labile raonoraeric (and aquo)
16 3 2
aluminum increased exponentially with decreases in pH in Adirondack '
Ox ? 29
(p {A1 } = -6.55 + 2.55 pH, n = 321, r = 0.93, p <0.0001) and HBEF
(p { Al3+} = -8.42 + 2.95 pH, n = 34, r2 = 0.92, p< 0.0001) surface waters. Non-
labile monomeric aluminum was not correlated with pH, but was strongly corre-
16 32
lated with organic carbon concentration in both Adirondack ' (Alo = -3.26
-fi -1 2
x 10 + 0.0204 TOC, where Alo and TOC are in mols .1 , n = 322, r = 0.76,
p < 0.0001) and HBEF (G. Lawrence unpublished data, Alo = -7.3 x 10 ^ + 0.0155
DOC, n = 69, r2 = 0.85, p < 0.0001) waters.
-------
Driscoll et. al. observed pronounced temporal variations in aluminum
fractions in Adirondack streams. During snowmelt and autumn rainfall events,
pH values were low and levels of labile monoraeric aluminum were high. During
low flow, high pH (pH > 5.5) periods in the summer, raonomeric aluminum levels
were elevated ( >10 pmol Al • 1 *") but this aluminum was entirely attributed to
32 29
non-labile raonomeric aluminum. Driscoll et al. and Johnson et al. have
reported that when aquo aluminum was calculated from the labile monomeric
aluminum determination, values were compatible with aluminum trihydroxide
solubility.
To evaluate the extent to which non-labile monomeric aluminum was sensi-
tive to variations in pH, I spiked aliquots of Adirondack stream water with
1 mmol • 1 ^ Tris (Tris (Hydroxy methyl) aminomethane) buffer and incrementally
adjusted the pH over a range of values with 0.1 N HC1. This was used to buffer
the solution in the pH range 5.5 - 7.0. Because solution pH values are well
below the proton dissociation constant of Tris (pKa = 8.08), complexation
of aluminum was considered to be insignficant. These solutions were incubated
for one week and analyzed for pH and non-labile monomeric aluminum. Variations
in non-labile raonomeric aluminum were relatively insensitive to variations in
pH (Figure 4). The concentration of non-labile monoraeric aluminum decreased
24% at pH 3.34 and 15% at pH 6.88 from a maximum value at pH 4.96. The reduction
in the magnitude of the alumino-organic complex in the low pH range might be
attributed to the disruption of the complex by hydrogen ion. The subtle reduc-
tion at higher pH values ( > 5) may be due to aluminum hydrolysis and competition
of hydroxide ligands with the organic ligands for the aluminum central metal
cation or possibly due to complexation of aluminum by dissociated Tris.
-------
These results, together with the observation that non-labile monomeric
aluminum levels in natural water samples were not appreciably different when
analyzed with pH 5 and pH 7 columns suggests that the pH change associated with
ication exchange treatment effects the determination of non-labile monomeric
aluminum only to a limited extent.
There are a number of potential sources of error associated with this
fractionation procedure. (1) 8-hydroxyquinoline may desorb aluminum associated
with particulate matter and therefore result in an overestimation of levels of
monomeric aluminum. (2) The exchanger may disrupt relatively labile organic
46
complexes as suggested by Means et al. . Although this error is minimized
by decreasing the sample retention time in the column, it undoubtedly still
occurs to some extent. (3) Alumino-organic complexes may exchange/adsorb on
resin sites. If the pH within the exchanger bed is low, organic matter may com-
plex hydrogen ion and thereby release aluminum to the resin and/or precipitate
within the column. (4) If the pH within the exchanger bed is higher than the
sample pH, aluminum associated with organic ligands may hydrolyze and adsorb
and/or precipitate within the column. It is noteworthy that all of these
sources of error result in an underestimation of the organically complexed
fraction of monomeric aluminum.
As previously mentioned, the fluoride ion selective electrode may be
used to independently evaluate the cation exchange column procedure for the
separation of inorganic and organic forms of monomeric aluminum. However,
as with the exchange column procedure there are a number of potential
sources of error associated with using the fluoride ion selective electrode
to fractionate aluminum. These potential errors include: (1) hydroxide and
-------
non-aluminum cation interference in the determination of free fluoride, (2) the
measurement of fluoride associated with particulate matter in the determination
of free fluoride, (3) changes in the pH, aluminum and free fluoride levels after
sampling and during the period of time the solution is being analyzed, and (4)
I
errors in and lack of thermochemical data used to make the calculations of ion
distribution.
A comparison of the two methods was made on selected Adirondack surface
water samples. To minimize problems associated with changes during sample
storage, pH and free fluoride levels were measured and monomeric aluminum was
extracted in the field shortly after sample collection. Organic monomeric
aluminum determined by the fluoride ion selective electrode was compared with
non-labile monomeric aluminum determined by the column fractionation procedure
(Figure 5). At low aluminum levels, agreement was good between the two methods,
but at higher concentrations, organic monomeric aluminum calculated from the
fluoride ion selective electrode measurements was generally greater than
non-labile monomeric aluminum determined by the cation exchange column proce-
dure. Further research is needed to quantify the errors associated with both
fractionation techniques.
-------
CONCLUSIONS
This procedure can be used to fractionate aqueous aluminum into acid
soluble aluminum, non-labile raonomeric aluminum and labile monomeric aluminum.
lAcid soluble aluminum is thought to include colloidal aluminum as well as very
non-labile organic complexes. Non-labile raonomeric aluminum is considered
to approximate levels of organically complexed monomeric aluminum in solution.
Labile monomeric aluminum would include aquo aluminum as well as inorganic
complexes of aluminum. This procedure was applied with reasonable success to
studies of dilute acidic waters from the Adirondack region of New York and the
Hubbard Brook Experimental Forest in New Hampshire. This methodology is
significant because it enables researchers to gain a better understanding of
the biogeochemistry of aluminum.
ACKNOWLEDGEMENTS
This research was supported in part by the Office of Water Research and
Technology (14-34-001-7068), the United States Environmental Protection Agency/
North Carolina State University Acid Precipitation Program (APP0094-1981)
and the National Science Foundation (DEB82-06980). I would like to thank
Joan Baker, Joe Unangst, Noye Johnson, Gary Schafran, Greg Lawrence and the
Department of Soil Science and Geology, Agricultural University, Wageningen,
the Netherlands for their assistance in this study.
-------
REFERENCES
1. R.M. Garrels, F.T. MacKenzie and C. Hunt, Chemical Cycles and the Global
Environment (Wm. Kaufmann, Los Altos, CA, 1975).
2. H.O. Buckman and N.C. Brady, The Nature and Properties of Soils (MacMillian,
New York, 1961).
3. C. Bloorafield, J. Sci. Food Agric. 8, 389 (1957).
4. C.B. Coulson, R.I. Davies and D.A. Lewis, J. Soil Sci. 11, 20 (1960).
5. C.B. Coulson, R.I. Davies and D.A. Lewis, J. Soil Sci. 11, 30 (1960).
6. A.H. Johnson and T.G. Siccama, Soil Sci. Am. J. 43, 1199 (1979).
7. F.C. Ugolini, R. Minden, H. Dawson and J. Zachara, J. Soil Sci. 124,
291 (1977).
8. W. Sturam and J.J. Morgan, Aquatic Chemistry (John Wiley and Son, New York,
1970).
9. H.J.M. Bowen, Trace Elements in Biochemistry (Academic Press, New York,
1966) .
10. C.S. Cronan and C.L. Schofield, Science, 204, 305 (1979).
11. W. Scheider, J. Adamski and M. Paylor, Reclamation of Acidified Lakes near
Sudbury Ontario (Ontario Ministry of the Environment, Rexdale, Ontario,
Canada, 1975).
12. R.F. Wright, T. Dale, A. Henriksen, G.R. Hendrey, E.T. Gjessing, M. Johannessen,
C. Lysholm and E. Storen, Regional Surveys of Small Norwegian Lakes (SNSF
Project IR33/77, Oslo, Norway, 1977).
13. W. Dickson, Verein., Limnol. 20, 851 (1978).
14. R.F. Wright, N. Conroy, W.T. Dickson, R. Harriman, A. Henriksen and C.L.
Schofield, In: Proceedings of an International Conference on Ecological
Impact of Acid Precipitation, D. Drablos and A. Tollan, Eds. (SNSF Project
Oslo, Norway, 1980) pp. 377-379.
15. J.H.D. Vangenechten and O.L.J. Vanderborght, In: Proceedings of an Interna-
tional Conference on Ecological Impact of Acid Precipitation, D. Drablos and
A. Tollan, Eds. (SNSF Project, Oslo, Norway, 1980) pp. 246-247.
-------
16. C.T. Driscoll, Chemical Characterization of Some Dilute Acidified Lakes
and Streams in the Adirondack Region of New York State, (Ph.D. Thesis,
Cornell University, 1980).
17. M. Johannessen, In: Proceedings of an International Conference on Ecological
Impact of Acid Precipitation, D. Drablos and A. Tollan, eds. (SNSF Project,
Oslo, Norway, 1980), pp. 222-223.
18. A. Henriksen and H.M. Seip, Wat. Res., 14, 809 (1980).
19. C.T. Driscoll and J.J. Bisogni, In: Modeling of Total Acid Precipitation
Impacts, J.L. Schnoor, ed. (Ann Arbor Science, Ann Arbor, MI, 1983), pp. 53-72.
20. C.P. Huang, J. Coll. Int. Sci. 55, 281 (1976).
21. J.A. Davis, Geochim., Cosraochim, Acta, 46, 681 (1982).
22. J.P. Baker and C.L. Schofield, Water Air Soil Poll. 18, 289 (1982).
23. B. Ulrich, R. Mayer and P.K. Khanna, J. Soil Sci., 130, 193 (1980).
24. C.E. Roberson and J.D. Hem, Solubility of Aluminum in the Presence of
Hydroxide, Fluoride and Sulfate (U.S. Geol. Surv. Water Supply Pap.
1827-C, Washington D.C., 1969) p. 37.
25. C.J. Lind and J.D. Hem, Effects of Organic Solutes on Chemical Reactions
of Aluminum (U.S. Geol. Surv. Water Supply Pap. 1827-G, Washington, D.C.,
1975) p. 82.
26. N.M. Johnson, Science, 204, 497 (1979).
27. E. Eriksson, Nordic Hydrology, 12, 43 (1981).
28. C.T. Driscoll, J.P. Baker, J.J. Bisogni and C.L. Schofield, Nature 284,
161 (1980).
29. N.M. Johnson, C.T. Driscoll, J.S. Eaton, G.E. Likens and W.H. McDowell,
Geochim Cosraochim, Acta, 45, 1421 (1981).
30. D.P. Moore, In: The Plant Root and its Environment, E.W. Carson, ed.,
(Univ. Press of Virginia, Charlottesville, Virginia, 1974), pp. 135-151.
31. R.B. Barnes, Chem. Geol, 15, 177 (1976).
32. C.T. Driscoll, J.P. Baker, J.J. Bisogni and C.L. Schofield, In: Acid
Precipitation: Geological Aspects, O.M. Bricker, ed. (Ann Arbor Science,
Ann Arbor, MI, 1983), pp. 55-75.
-------
33. D.K. Nordstrom, L.N. Pluramer, T.M.L. Wigley, J.W. Ball, E.A. Jenne, R.L.
Bassett, D.A. Crerar, T.M. Florence, B. Fritz, M. Hoffman, G.R. Holdren,
G.M. Lafon, S.V. Mattigod, R.E. McDuff, F. Morel, M.M. Reddy, G. Sposito
and J. Thrailkill, In: Chemical Modeling in Aqueous Systems, E.A. Jenne,
ed. (Amer, Chem. Soc. Symp. Series 93, 1979), pp. 857-892.
34. Orion Instruction Manual, Fluoride Electrodes, Orion Research, Inc.,
Cambridge, MA (1976).
35. H.M. May, P.A. Helmke and M.L. Jackson, Geochim. Cosmochim. Acta, 43,
861 (1979).
36. J.D. Hem, Graphical Methods for Studies of Aqueous Aluminum Hydroxide,
Fluoride and Sulfate Complexes (U.S. Geol. Surv. Water Supply Pap. 1827-B,
Washington, D.C., 1968) 33p.
37. B. Behr and H. Wendt, Zeitschr. Elektrochemie, 66, 223, (1962).
38. R.W. Smith and J.D. Hem, Effect of Aging on Aluminum Hydroxide Complexes
in Dilute Aquesous Solutions (U.S. Geol. Survey Water Supply Pap. 1827-D,
Washington, D.C. 1972) 51 p.
39. V.C. Kennedy, D.F. Jones and G.W. Zellweger, Water Resour. Res. 10,
785 (1974).
40. R.C. Turner, Can. J. Chem. 47, 2521 (1969).
41. R.C. Turner and W. Salaiman, Can. J. Chem. 49, 1683 (1971).
42. R. C. Turner, Can. J. Chem. 49, 1688 (1971).
43. R.W. Smith, In: Nonequilibrium Systems in Natural Water Chemistry, J.D.
Hem, ed., Advances in Chemistry Series 106 (American Chemical Society,
Washington, D.C., 1971) pp. 250-279.
44. P.R. Bloom, R.M. Weaver and M.B. McBride, Soil Sci. Soc. Am. J.
42, 713 (1978).
45. H.M. May, P.A. Helmke and M.L. Jackson, Chem. Geol, 24, 259 (1979).
46. G.E. Likens, F.H. Bormann, R.S. Pierce, J.S. Eaton and N.M. Johnson,
Biogeochemistry of a Forested Ecosystem (Springer-Verlag, New York, 1977).
47. J.L. Means, D.A. Crerar and J.L. Amster, Lirnnol. Oceanogr. 22, 957
(1977).
48. M. Schnitzer and S.U. Khan, Humic Substances in the Environment
(Marcel Dekker, New York, 1972).
-------
TABLE I
EQUILIBRIUM RELATIONSHIPS USED IN THIS STUDY
Equation
(Hydroxide Ligands)
Equilibrium Constant
Reference
ai3+
+
H20 = A1(0H)2+ + H+;
KO^
S3
1.03 x
10"5
(35)
ai3+
+
2H20 = A1(0H)2+ + 2H+;
koh2
=
7.36 x
10"11
(35)
ai3+
+
AH20 - A1(0H)a" + 4H+;
KOH.
4
O
6.93 x
io-23
(35)
(Fluoride Ligands)
ai3+
+
F" = A1F2+;
KFi
na
1.05 x
107
(36)
ai3+
+
2F~ =¦ A1F2+;
kf2
-
5.77 x
io12
(36)
ai3+
+
3F~ = A1F3;
KF3
o
1.07 x
io17
(36)
ai3+
+
4F = Al. ;
4
KF4
S3
5.37 x
io19
(36)
ai3+
+
5F~ = A1F52";
kf5
a
8.33 x
io20
(36)
ai3+
+
6F" - AlFg3";
(Sulfate Ligands)
KF6
S3
7.49 x
io20
(36)
ai3+
+
S0,° = A1S0.+;
4 4
KS1
B
1.63 x
io3
(37)
ai3+
+
2 S04= = A1(S04)2~;
ks2
a
1.29 x
io5
(37)
-------
TABLE II
CALCULATION OF ALUMINUM SPECIATION USING THE
FLUORIDE ION SELECTIVE ELECTRODE
Aluminum Bound Fluoride
[F - Al] = [Fx] - [F~] = [FT] - {F-} /Yl
[F-Al] = [A1F2+] + 2[AlF2+]+ 3[A1F3J + 4[A1F4~] + 5[A1F52"] + 6[A1F63"]
[ F-Al] = {Al3+} {F~} KF1/y2 + 2 {Al3+} {F_} 2 KF^y 1 + 3 {Al3+} {F~} 3 KF3 +
4{A13+}{F"}4KF4/Yl + 5{Al3+}{ F~}5 KF5/y2 + 6 {A13+}{F-}6 KF^
Aquo Aluminum
{Al3+}= [F-Al] ( {F-} KF1/y2 + 2{F'}2KF2/Tl + 3{F'}3KF3 +
4{F~> 4KF4/Yl + 5{F"}5KF5/Y2 + 6 {F^KF^)"1
Hydroxide Complexed Aluminum
[A1-0H] = {Al3+}KOH1/ {H+}y2 + {Al3+} KOH2/ {H+}2Y]L +
{A13+}K0H4/ {h+}4 yx
Fluoride Complexed Aluminum
[Al-F] = {Al^XF'jKF^Yjj + {Al3+} {F"}2KF2/Yi + {Al3+} {F_}3KF3 +
{A13+}{F*} 4kfa/Yi + {A13+}{F"}5KF5/y2 + {A13+}{f"}6KF6/y3
Sulfate Complexed Aluminum
[A1-S04] = {Al3+} [S042"] y2KS1/Yi + {ai3+}[S041y2 ks2/Yi
Inorganic Monomeric Aluminum
IMAL + {Al3+} /y 3 + [A1-0H] + [Al-F] + [A1-S04]
Organic Monomeric Aluminum
OMAL =¦ Al - IMAL
a
where: { } is species activity (mol* 1 )
[ ] is species concentration (mol • 1
-------
TABLE II (continued)
[F-Al], aluminum bound fluoride (mol . 1
total fluoride (rool-1
[A1-0H], hydroxide complexed aluminum (mol . 1
[Al-F], fluoride complexed aluminum (mol . 1 "*")
[A1-S0^], sulfate complexed aluminum (mol . 1
IMA1, inorganic monomeric aluminum (mol . 1
0MA1, organic monomeric aluminum (mol . 1 "*")
Ala, monomeric aluminum (mol*l
Yi» Y2» mean activity coefficients for monovalent, divalent
and trivalent species, respectively
KF , KOH , KS , thermodynamic stability constants for fluoride,
n n n hydroxide and sulfate complexes, respectively; see
Table I for values
-------
FIGURE TITLES
Figure 1 A schematic representation of the aluminum fractionation
procedure
Figure 2 Concentration of monomeric aluminum from cation exchange
column effluent (non-labile monomeric aluminum) as a func-
tion of the sample application rate to the column. This
sample was collected from an unnamed stream in the
Adirondack region of New York on 7/27/77 (pH = 4.32,
TOC = 1.2 ramol • 1 ^).
Figure 3 Ultraviolet wavelength scans of selected Adirondack
water samples (collected on 7/4/78, West Pond, pH = 6.51,
TOC = 1.2 mmol • 1 unnamed stream, pH = 4.37, TOC =
0.79 mmol • 1 North Lake, pH = 4.96, TOC = 0.41
mmol * l-^") for cation exchange column influent and
effluent.
Figure 4 Non-labile monomeric aluminum concentration of aliquots
of an unnamed stream sample (collected on 5/18/78
initially with pH = 4.31, TOC = 0.70 mmol • 1 , Al^ =
15.4 y mol *1 "*") which were buffered (1 mmol Tris • 1
and solution pH adjusted (0.1 N HC1).
Figure 5 A comparison of non-labile monomeric aluminum levels
determined using the cation exchange column separation
procedure with organic monomeric aluminum concentrations
calculated using free and total fluoride determinations
with thermochemical data (Tables I, II) for selected
Adirondack surface water samples (collected on 1/20-1/22/83).
-------
SCHEMATIC REPRESENTATION OF THE ALUMINUM FRACTIONATION PROCEDURE
aluminum
measurements
aluminum
fraction
TOTAL REACTIVE ALUMINUM, ACID DIGESTED
¦Total monomeric aluminum.no acid digestion-*-
Cation exchange treated
- monomeric aluminumh
Non-labile monomeric
aluminum
Labile monomeric
aluminum
Acid soluble
aluminum
fraction
composition
monomeric
alumino-organic
complexes
free aluminum;
monomeric
aluminum sulfate;
fluoride, and
hydroxide
complexes
colloidal
poly merici
aluminum;
strong alumino-
organic complexes
-------
0 12 3 4
SAMPLE APPLICATION RATE TO CATION EXCHANGE COLUMN
(ml • min-1 • ml resin-1)
-------
UNTREATED SAMPLE
CATION EXCHANGE PROCESSED
SAMPLE
UJ r
o .6
CD
WEST POND
UNNAMED
STREAM
NORTH
LAKE
200
240
280
320
360
WAVELENGTH (nm)
-------
NON-LABILE MONOMERIC ALUMINUM
( moJ * I"1 )
* 01
o
cn
o
o
-------
o
o
LU —
O "o
if 6
LlJ
CO
00
d-o
ORGANIC MONOMERIC ALUMINUM (F electrode)
(/ieq-1"1)
-------
APPENDIX 2
Driscoll, C.T. and G.C. Schafran. 1984. Characterization of short term
changes in the base neutralizing capacity of an acidic Adirondack, NY
lake. Nature (submitted for publication).
-------
Characterization of Short Term Changes in the
Base Neutralizing Capacity of an Acidic Adirondack, NY Lake
C.T. Driscoll and G.C. Schafran
Department of Civil Engineering
Syracuse University
Syracuse, NY 13210 USA
There is currently much concern and controversy over the effects of acidic
deposition on low ionic strength surface waters. As a result, there has been
much discussion in the recent literature on the nature and extent of proton
12 3 4
transformations within acid sensitive ecosystems ' ' ' . The source of base
neutralizing capacity (BNC) within acidic surface waters has been attributed
5 6 7
to atmospheric deposition of H^SO^ (or SC^) ' or HNO^ as well as production
O
of soluble organic acids from soils' . Unfortunately many of these studies have
failed to adequately characterize aluminum, which is often a very significant
Q
component of BNC in acidic waters. Because of the current interest in surface
water acidification, we have evaluated the nature of short term changes in the
BNC of an acidic Clearwater lake. Our results suggest that much of the variation
in hydrogen ion and aluminum BNC can be attributed to changes in nitrate concentra-
tion, rather than variations in sulfate, chloride, or organic anion concentrations.
The study site, Dart Lake, is located in the Adirondack Mountain region of
2
New York State, U.S.A. (74°52'W, 43°48'N). The watershed drainage area is 107 Km ,
and the lake volume is 1.02 x 107m3 with a mean depth of 7.1 m. Dart Lake has
a single outlet and our water budget indicates that 95% of the outlet flow (7.25 x
107 m3 yr"1) enters the lake at the major inlet. Samples were collected for
water quality analysis at the inlet, outlet and at seven depths from a pelagic
-------
sampling station approximately every two weeks over an annual cycle (10/25/81 -
11/21/82). Shortly after collection samples were measured for pH and dissolved
inorganic carbon (DIC), ampulated for dissolved organic carbon (DOC) analysis,
and extracted for monomeric aluminum using the procedure of Barnes^. Monomeric
aluminum was fractionated into labile and non-labile forms using the cation exchange
11 12
column method of Driscoll ' . Samples were analyzed in the laboratory for all
, . 13
major solutes .
Selected samples, collected from drainage waters in the Dart Lake watershed,
Q
were processed and titrated to characterize the proton dissociation of naturally
occurring organic solutes. These data were fit to a monoprotic proton dissociation
g
model using a modified Gran plot analysis . A statistically significant empirical
relationship was observed between DOC concentration and mols of proton dissociation
sites per liter (CT = 0.025 DOC + 8.0; where CT is the proton dissociation sites on
-1 -1 2
organic solutes inpmols 1 and DOC is in ymols 1 ; n = 13, r = 0.69, p <0.01).
A relatively consistent fit to a proton dissociation constant was also obtained
(pKa = 4.53 ± 0.25). Our detailed observations of DOC were applied to this model
to estimate the BNC due to organic solutes and the organic anion concentration of
water samples.
Because of poor precision, slow kinetics associated with heterogeneous phase
reactions, and the inability to characterize individual components, BNC was not
determined by titration. Rather the individual components of BNC were calculated,
using a chemical equilibrium model14, as the amount of strong base required to in-
crease the pH of a liter of solution to a value of 8.3. The thermochemical data
1112
used in our analysis are summarized elsewhere ' and thermodynamic calculations
were corrected for the effects of temperature and ionic strength. In these calcula-
tions we assumed that labile monomeric aluminum was a qood estimate of inorganic
forms of aluminum and that non-labile monomeric aluminum did not significantly
-------
accept or donate protons. The close electroneutrality balance supports the
assumptions and procedures used in this study for the characterization of
aluminum and DOC (Table 1). Further information on the site description,
hydrology, sampling and analytical methods, thermal and water quality data,
13
and data analysis are available elsewhere .
Although H^CO^ was the major component of BNC in Dart Lake (Table 1),
levels of hydrogen ion and aluminum BNC in acidic lakes are generally of
more interest because of potential toxicity to organisms15'16. Variations
in hydrogen ion and aluminum BNC (H-A1-BNC) of all samples collected (n = 178)
were strongly correlated with variations in nitrate concentration (H-A1-BNC =
0.94 NO3- + 2.4; weq l"1, r2 = 0.54, p< 0.0001 ). Note that this empirical
correlation is linear with a slope close to one and an intercept near the
origin. Although S0^ ~ was the dominant anion in Dart Lake (Table 1), no
statistically significant relationship with H-A1-BNC was observed. H-A1-BNC
was weakly correlated with organic anion concentration (H-A1-BNC = 33.3 (RC00 )
-1 2
-481; where RC00" represents the organic anion content in peq 1 , r = 0.13,
p< 0.0001) and no statistically significant relationship was observed with
CI".
Water quality observations in Dart Lake suggest that biogeochemical pro-
cesses result in temporal and spatial variations in NO^", which influence
hydrogen ion and aluminum BNC (Figure 1). In the autumn NO^ , H-BNC, and
Al-BNC exhibited an orthograde distribution in Dart Lake. During ice cover,
NOg", H-BNC, and Al-BNC increased in the upper waters. These trends were
presumably due to freeze-concentration at the ice-water interface and/or
higher concentrations associated with stream inputs, which flow along the
lake surface because of winter thermal stratification1^. Concentrations
of all three constituents decreased with increasing lake depth.
-------
Although the lake was aerobic throughout the study period (a minimum D.O. of
41 ymol l"1 was observed at 14 m depth on 1 Oct. 1982), we attribute these lower
water depletions of BNC to sediment reduction of NO^" and the resulting production
of acid neutralizing capacity (ANC).
During the snowmelt period (3/1-5/1), 34% of the annual discharge entered
Dart Lake. Inlet concentrations of H-A1-BNC and N03~ were significantly greater
during snowmelt (3/1 - 5/1, flow = 8.9 ± 6.9 m3 s~\ H-A1-BNC = 52 ± 11, N03~ =
O "I
51 ±11) than during either summer - autumn (6/1 - 11/1, flow = 1.7 ± 1.7 m s ,
H-A1-BNC = 18 ± 5; p < 0.05 student t test NO^" = 24 ± 5, p < 0.05) or winter
(11/1 - 3/1, flow = 1.5 ± 1.0 m3 s"1, H-A1-BNC = 23 ± 6; p < 0.05; N03' =
26 ±6, p <0.1) base flow periods. These inputs resulted in high levels of
N03", H-BNC and Al-BNC throughout Dart Lake prior to summer stratification
(Figure 1).
During summer stratification, N03~, H-BNC and Al-BNC were depleted in the
upper and lower waters of Dart Lake, resulting in a positive heterograde dis-
tribution. We attribute the upper water depletion to algal assimilation of
18
N03~ or littoral sediment denitrification and the lower water depletion
to pelagic sediment denitrification. A number of investigators have reported
that the net assimilation of nitrate results in the production of ANC and
19 20
an increase in pH values ' . Groundwater inflow, enriched in ANC, basic
Oi pi, ^ ^
cations (Ca , Mg , Na , K ) and dissolved silica from soil weathering
reactions, could also deplete H-A1-BNC during summer low-flow conditions.
However, we observed no statistically significant variations in the inlet
concentrations of basic cations and dissolved silica over the study period,
suggesting that base flow input of ANC was not a significant mechanism of
H-A1-BNC neutralization. In the extreme lower waters (below 12m, which is
-------
5% of the lake volume) during stratification, transformations of Fe, Mn, DOC,
Ca2+, and NH^+ contributed to the depletion of H-BNC and Al-BNC in Dart Lake.
However, on a whole-lake basis nitrogen transformations affecting NO^" con-
centration appear to be the predominant mechanism regulating short-term changes
in hydrogen ion and aluminum in solution.
The results of synoptic surveys indicate that NO^" levels are generally
higher in the Adirondack lake district than other lake districts in eastern
21
North America with low ionic strength waters (Table 2). A difficult but
27
critical problem is to identify the source of this nitrate. Johannes et al.
indicated that N0^~ storage in snowpack from the Adirondack region is generally
2- 1
greater than SO^ . Galloway et al. have attributed the decrease in the ANC
that occurs in the Adirondack lake outlets during snowmelt to a reduction in
basic cation release from soils and the addition of nitric acid from snowpack.
While snowpack nitrate undoubtedly contributes to the elevated levels observed
in Adirondack lake waters, decomposition and oxidation of organic nitrogen within
the adjacent soil may also result in the release of nitrate and BNC to solu-
.. 28,29
tions
Because of these apparently high NO^" levels in Adirondack waters we
wanted to evaluate the extent to which changes in ambient anion concentrations
(S042-, N03", RCOO") result in changes in solution pH. To accomplish this,
we simulated the response of Dart Lake water (Table 1) to additions and
reductions in HNO-j, ^SO^, and the protonated form of naturally occurring
14
organic solutes using a chemical equilibrium model . Modest additions
and depletions of N03" resulted in significant changes in pH, while Dart
-------
Lake appears to be well buffered with respect to changes in DOC (Figure 2).
The latter response may be attributed to the relatively low density of proton
dissociation sites per mole of organic carbon and indicates that deprotonation
of organic acids probably has not contributed significantly to the acidification
lof Dart Lake. When NO^" levels were reduced to values comparable to other lake
districts in Eastern North America (^2 peq 1"^) pH values approached neutrality.
Like nitrate, variations in the concentration of H^SO^ result in an equivalent
2_
change in H-A1-BNC. Because SO^ is the dominant anion in Dart Lake solutions
(Table 1), neutralization can also be accomplished by a reduction in the concen-
tration of H^SO^. Note that in extremely acidic (pH < 4.8) Adirondack lakes
22
H-A1-BNC generally exceeds NO^ concentrations . Therefore surface water
acidification in the Adirondacks cannot be entirely attributed to HNO^, but
more probably a combination of H^SO^ and HNO^ inputs.
2- 22
Although SO^ is generally the dominant anion in Adirondack surface waters
29
and in atmospheric inputs to the region , additions of NO^ during snowmelt
and depletions of NO^" associated with assimilation and reduction processes
appear to be very important in regulating short-term changes in hydrogen ion
and aluminum concentrations in an acidic Adirondack lake. Moreover, elevated
levels of NO^" in the Adirondack Lake district, compared to other lake districts
in Eastern North America, suggest that nitric acid inputs may have also contri-
buted to the long term acidification of the region. However a more complete
understanding of the processes controlling the transport of NO^" and H-A1-BNC
during snowmelt is essential to assess the mechanisms responsible for surface
water acidification.
-------
We would like to thank L.G. Barnum, N.J. Peters, F.J. Unangst, and
J.R. White for their assistance during this study. This study was funded
in part by the USEPA/NCSU Acid Deposition Program (APP 0094-1981). This
paper has not been subjected to the EPA1s required peer review policy and
therefore does not necessarily reflect the views of the Agency, no official
endorsement should be inferred. Contribution No. 31 of the Upstate Freshwater
Institute.
-------
References
1. Rosenqvist, I.T., Sur jord - surt vann, Ingenorforl aget, Oslo, 123 p.
2. Driscoll, C.T. and Likens, G.E. Tellus 34, 283-292 (1982).
3. Kelly, C.A., Rudd, Cook, R.B. and Schindler, D.W. Limnol.
Oceanogr. 27, 5 (1982).
4. van Bre'emen, N., Mulder, J. and Driscoll, C.T. Plant and Soil 75, (in
press, 1984).
5. Gjessing, E.T., Henriksen, A., Johannessen, M. and Wright, R.F. in
Impact of Acid Precipitation on Forest and Freshwater Ecosystems in
Norway (ed. Braekke, F.H.) 64-85 (SNSF Project As, Norway, 1976).
6. Henrikesen, A. Nature, 278, 542-545, (1979).
7. Galloway, J.N., Schofield, C.L., Hendrey, G.R., Peters, N.J., and
Johannes, A.H. in Ecological Impact of Acid Precipitation (eds. Drablos, D.,
and Tollan, A.) 264-265 (SNSF Project, As, Norway, 1980).
8. Glover, G.M. and Webb, A.H. Water Res. 13, 781-783 (1979).
9. Driscoll, C.T. and Bisogni, J.J. in Modeling of Total Acid Precipitation
Impacts (ed. Schnoor, J.L.) 53-72 (Ann Arbor Science, Ann Arbor, MI,
1984).
10. Barnes, R.B. Chem. Geol. 15, 177-191 (1975).
11. Driscoll, C.T. Int. J. Environ. Anal. Chem. (in press, 1984).
12. Johnson, N.M. Driscoll, C.T., Eaton, J.S., Likens, G.E. and McDowell, W.M.
Geochim. Cosmochim. Acta 45, 1421-1437 (1981).
13. Driscoll, C.T. and Schafran, G.C. Final report for USEPA/NCSU Acid Precipita-
tion Program project APP 0094-1981 (1984).
14. Westall, J.C., Zachary, J.L. and Morel, F.M.M. MINEQL, A Computer Program
for the Calculation of Chemical Equilibrium in Aqueous Systems, Ralph M.
Parsons Laboratory for Water Resources and Environmental Engineering,
Civil Engineering Department, Massachusetts Institute of Technology,
Technical Note, No. 18 (1976).
15. Leivestad, H., Hendrey, G., Muniz, I.P. and Snekvik, E. in Impact of
Acid Precipitation on Forest and Freshwater Ecosystems in Norway (ed.
Braekke, F.H.) 86-111 (SNSF Project As, Norway, 1976).
-------
16. Cronan, C.S. and Schofield, C.L. Science 204, 304-305, (1979).
17. Hagen, A. and Langeland, A., Environ. Pollut. 5, 45-57, (1973).
18. Brewer, P.G. and Goldman, J.C. Limnol. Oceanogr. 21, 108-117 (1976).
19. Harvey, H., Pierce, R.C., Dillon, P.J., Kramer, J.R., and Welpdale, D.M.
Acidification of the Canadian Aquatic Environment. Publication NRCC No.
18475 of the Environment Secretariat National Research Council of
Canada (1981).
20. Dillon, P.J. and Scheider, W.A. in Modeling of Total Acid Precipitation
Impacts (ed. Schnoor, J.L.) 121-154 (Ann Arbor Science, Ann Arbor, MI,
1984).
21. Wright, R.F. Predicting Acidification of North American Lakes, Norwegian
Institute for Water Research, 0-81036, Oslo, Norway (1983).
22. Schofield, C.L. Acidification of Adirondack Lakes by Acid Precipitation:
Extent and Magnitude of the Problem. USFWS Fed. Aid Fish Restoration
Project F-28-R (1976).
23. Beamish, R.J., Blouw, L.M. and McFarlane, G.A., A Fish and Chemical
Study of 109 Lakes in the Experimental Lakes Area (ELA), Northwestern
Ontario, with Appended Reports on Lake Whitefish Ageing Errors and
the Northwestern Ontario Baitfish Industry, Tech. Rep. 607, Fisheries
and Marine Service Environment Canada (1976).
24. Conroy, N., Hawley, K. and Keller, W., Extensive Monitoring of Lakes
in the Greater Sudbury Area 1974-1976, Ontario Ministry of the Environ-
ment (1978).
25. Duthie, H.C. and Ostrofsky, M.L., J. Fish Res. Bd. Canada 31, 1105-1117
(1974).
26. Kerekes, J.J., in Ecological Impact of Acid Precipitation (eds. Drablos, D.,
and Toll an, A.) 232-233 (SNSF Project, As, Norway, 1980).
27. Johannes, A.H., Halloway, J.N. and Troutman, D.E. in Ecological Impact
of Acid Precipitation (eds. Drablos, D., and Tollan, A.) 260-261 (SNSF
Project, As, Norway, 1980).
28. Hem, J.D. and Roberson, C.E. U.S. Geol. Surv. Water Supply Pap. 1827-A
(1967).
29. Johannes, A.H. and Altwicker, E.R., in Ecological Impact of Acid Precipita-
tion (eds. Drablos, D., and Tollan, A.) 256-257 (SNSF Project, As, Norway,
1980).
-------
Table 1 Mean Chemical Composition (yeq.l"^) of
Dart Lake inlet, outlet and water column
solutions, 10/1981 - 11/1982 (n = 172)
Mean Standard Deviation
Base Neutralizing Capacity
(to reference pH = 8.3)
hydrogen ion 7.4 3.7
aluminum 19.6 8.2
carbonic acid 62.0 58.9
organic acids 3.0 1.2
total 92.0 59.0
Basic Cations
calcium 99.5 4.6
magnesium 28.9 2.2
sodium 28.3 4.9
potassium 13.0 1.1
total 169.7 10.0
Anions
sulfate 137.6 11.0
nitrate 24.4 8.6
chloride 13.5 3.4
organic anions 12.2 0.4
bicarbonate 7.2 11.9
free fluoride 0.9 0.7
total 195.8 25.6
Sum of Cations^ 196.7 13.9
Sum of Anions 195.8 25.6
^Cationic equivalence were computed to be hydrogen ion and aluminum BNC
plus basic cations.
-------
Table 2
Average Chemical
Composi tion
(ueq»l~^
) of Low
Ionic Strength
Waters From
Lake
Districts in
Eastern
20
North America
Reg i on
n
H+
Ca2+
S042"
no3"
hco3-
Reference
Adirondack NY, USA
206
12
108
135
16
25
22
Experimental Lakes Area
Ontario, Canada
102
1
96
78
1
74
23
Sudbury
Ontario, Canada
208
4
273
252
2
160
24
Churchill Falls
Laborador, Canada
13
1
80
60
2
69
25
Kejimkujik
Nova Scotia, Canada
3
18
22
76
3
0
26
-------
Figure Titles
1. Isopleths of nitrate (a), hydrogen ion BNC (b) and aluminum
BNC (c) in ueq l"1 for Dart Lake (10/25/1981 - 11/21/1983).
i Samples were collected at seven depths. Sampling dates are
indicated (a).
2. Simulated changes in the pH of Dart Lake with increases and reductions
from the ambient nitrate (a) and DOC (b) concentrations. Average Dart
Lake water (Table 1) was assumed to be in equilibrium with microcrystal1ine
gibbsite (p*K = 9.35; 28) and an organic solute (CT = 0.025 DOC + 8.0,
i SO I
in ymols-1" ; pKa = 4.53). The relatively soluble mineral phase, micro-
crystalline gibbsite, was selected for our simulations because it is
considered to represent the solubility of freshly-formed Al(0H).j (28).
Moreover, the mean ion activity (Qso) product for A1(OH)^ of Dart Lake
solutions was very close to the solubility of microcrystal1ine gibbsite
(p*QS0 = 9.52 ± 0.28).
-------
h 5 I
o) NO; ISOPLETH (ueq r'l
b) H-BNC ISCPLETHiuea
_C) AI-BNC ISG3!.?"7^ Cue:
Un
ft 20 24
J F M A M J J
TIME {months)
-------
Dissolved Organic CarbonConcentr.
(/imols CT1)
_ rv> 4s
_ JS oj ^
o o o en o
oo
Nitrate Concentration
[fieq T1)
:
Percent of Ambient Concentration
------- |