SEARCH

SEARCH BY CITATION

Keywords:

  • hydrological characteristics;
  • dissolved organic carbon;
  • chloride ion

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Materials and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusion
  8. Acknowledgments
  9. References

[1] The semiarid prairie pothole region of the North American Great Plains is characterized by millions of small, shallow, closed basin wetlands. These wetlands are hydrologically dynamic, often losing considerable water volume and depth seasonally in response to high evaporative stress and/or infiltration rates. However, the consequences of such water loss on wetland water chemistry parameters, in particular dissolved organic carbon (DOC), remain relatively unstudied. Seasonal changes in DOC concentrations in 12 freshwater and saline wetlands at the St. Denis National Wildlife Refuge near Saskatoon, Saskatchewan, Canada, were examined over an 8-year period (1993–2000). Specific conductivity in the study ponds ranged from 312 μS cm−1 to 33,493 μS cm−1 (seasonal means). DOC concentrations in all study ponds were high (>10 mg L−1) and increased across a gradient of increasing salinity (mean DOC values from fresh water to saline ranged from 19.7 mg L−1 to 102.7 mg L−1). In the majority of ponds, DOC concentrations increased seasonally from spring through fall. On average this increase was 21 mg L−1, with fall values averaging 60% greater than spring. The greatest DOC increases were observed in saline ponds which lost most of their water by evaporation. Although DOC in these ponds was highly correlated with the conservative tracer, chloride, the slopes of these regression lines were always less than 1 as were the DOC:chloride ratios, indicating nonconservative DOC behavior. Additionally, chloride concentrations increased much faster seasonally than did DOC. Taken together, these data indicated that although DOC was not behaving conservatively, at least some of the observed DOC increases could be explained by simple evapoconcentration. These data also suggested that saline ponds appeared to experience net seasonal removal of DOC. Possible removal mechanisms for DOC include infiltration to the pond margin, bacterial utilization, and photolysis. Freshwater ponds, which lost most of their water by infiltration to the pond margin, on the other hand, displayed less seasonal variation in DOC concentrations. In these ponds, the relationship between DOC and chloride ion was not as strong as in the saline ponds; the slope of this relationship was always >1, as were DOC:chloride ratios. These data indicated that although DOC was being lost to the pond margin as water infiltrated, freshwater ponds accumulated DOC seasonally. Decomposition and excretion of DOC by macrophytes, as well as by pelagic and attached phytoplankton, are the likely within pond sources of DOC here. The rapid response of these small, shallow aquatic systems to water loss make them ideal microcosms in which to study effects of climate on DOC concentrations and other water chemistry parameters.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Materials and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusion
  8. Acknowledgments
  9. References

[2] Wetlands are a ubiquitous feature of the North American prairie landscape. Created during the Pleistocene glaciation retreat some 12,000 years ago, the hummocky terrain of the prairies contains millions of closed-basin wetlands known locally as potholes, ponds or sloughs [Mitsch and Gosselink, 1993]. The combination of unique climatic conditions (which initiates wet/dry cycles), depressional wetlands, and rich glacial till make for some of the most productive aquatic ecosystems in the world [Murkin, 1989]. In fact, prairie wetlands are the single most important breeding area for waterfowl on the North American continent [Poiani and Johnson, 1991] producing 50–80% of the North American waterfowl population in any given year [Batt et al., 1989]. Most are shallow, usually <1 m in depth, and occur in hydrologically closed basins (no permanent stream inflow or outflow) [Winter et al., 2001]. They range from permanent to ephemeral (only filled with water for a short time during the ice-free season) and freshwater to saline. Because watersheds of prairie wetlands are not connected by surface water drainage, water balance within individual wetlands is determined by precipitation, evapotranspiration, and interaction with groundwater [LaBaugh et al., 1996]. Most are highly dependent on spring melt for their water supply [Covich et al., 1997]. Prairie wetlands are also located within a semiarid climatic zone, where potential evaporation exceeds precipitation [LaBaugh et al., 1996]. In these topographically closed aquatic systems, response to changes in the balance between precipitation and evaporation is much faster than in those systems with distinct inflows and outflows [Fritz, 1996]. One consequence is that some wetlands are dry by midsummer, while others lose considerable water volume and depth seasonally [Poiani et al., 1996]. Another consequence is that water chemistry may change seasonally with decreasing water volume [Fritz, 1996].

[3] A key characteristic of prairie wetlands is their high dissolved organic carbon (DOC) concentration: typically >10 mg L−1 and sometimes ranging in excess of 100 mg L−1 [Arts et al., 2000]. Such concentrations are somewhat higher than the range published for wetlands (10–50 mg L−1 [Curtis, 1998]) and much higher than those reported for alpine (0.5–3.0 mg L−1 [Curtis, 1998]) and Antarctic lakes (Lake Hoare DOC 1.3–3.9 mg L−1 [McKnight et al., 1991]). Williamson et al. [1999] pointed out that one of the key characteristics which determines how a water body will respond to disturbance (i.e., anthropogenic stress, climate change) is its colored DOC concentration. DOC participates in many biogeochemical reactions in aquatic ecosystems. It flocculates chemical substances (including pesticides), fuels microbial food chains, and takes part in a number of photochemical reactions that produce not only strong oxidants, such as carbon monoxide and peroxide, but low molecular weight carbon compounds suitable for bacterial uptake [Schiff et al., 1990; Wetzel et al., 1995]. Its ability to attenuate photosynthetically active (PAR) and ultraviolet (UVR) radiation restricts the depth of autotrophic production and also protects aquatic organisms from harmful ultraviolet light [Arts et al., 2000].

[4] Some early work on prairie aquatic systems indicated a relationship between DOC and hydrological characteristics. A study of four shallow lakes in the Nebraska Sandhills showed that high DOC concentrations in groundwater only occurred in those areas which received seepage from lakes and these concentrations were also much higher than in groundwater that was moving into lakes [LaBaugh, 1986]. However, most of what is known regarding the relationship between DOC and hydrological characteristics stems from work done in temperate systems. In these zones, precipitation and the way in which aquatic ecosystems are connected to their watersheds affects DOC concentration. In two study streams in the Ontario Precambrian Shield, DOC concentrations increased by as much as 100 and 410% during storm events [Hinton et al., 1997]. Not surprisingly, then, maxima in DOC concentrations in lakes have been linked to the rainy season (spring or fall [Reche and Pace, 2002]), when DOC inputs from streams are expected to be high. In some Precambrian shield wetlands, higher DOC concentrations observed in dry summer and early fall periods have been linked to lower lake flushing rates especially during times of lowered water tables, while lower concentrations at others are associated with high flushing rates throughout the ice-free season [Schiff et al., 1998]. As well, “hydrology controls the geochemistry and export of dissolved organic matter (DOM) by affecting internal redox biochemistry and by controlling water movement along important flowpaths in organic rich soil horizons which are DOM sources” [Schiff et al., 1998]. In those aquatic systems whose catchments are dominated by or connected to wetlands, DOC concentrations are generally higher [Molot and Dillon, 1997]. At a global scale, the number of wetlands in the watershed, in addition to elevation, have been found to be the most important predictors of DOC concentrations in lakes [Xenopoulos et al., 2003].

[5] DOC concentrations may also be influenced by climate. For example, decreases in DOC concentrations may occur in response to drier conditions associated with climate change. With lower precipitation and a decrease in the number of storm events, less DOC is brought into lakes from the surrounding watershed [Schindler et al., 1992]. DOC concentrations in receiving waters also increase after heavy rains in regions which have experienced antecedent drought or dry spells [Schindler et al., 1997].

[6] DOC is important in aquatic ecosystems. Additionally, hydrological characteristics and climate can have a significant effect on its concentration, at least in temperate systems. And there is no doubt that climate change and warming will adversely affect Canadian water quality and quantity [Schindler, 2001]. But in the semiarid prairie ecozone, little is known about how DOC concentration or cycling might be affected by the tremendous inherent interannual and intra-annual variations in wetland water volume, let alone by disturbance, like climate change. In Western Canada, “Climate change means that precipitation is becoming less reliable and more of it is expected to come as rain rather than as snow. What snow there is will melt sooner. There are likely to be more big storms and more severe droughts. Transpiration is expected to increase” [Banks and Cochrane, 2005; Schindler, 2001]. Paleoecological studies on prairie lakes have revealed that significant changes in algae, macrophytes and other aquatic organisms have occurred because of past drought effects on water depth and chemistry [Schindler, 2001]. Given this paleoecological evidence, the current climate change scenarios for the prairies, and the importance of DOC to aquatic ecosystems, it is imperative to gain an understanding now of the relationship between DOC and prairie wetland hydrological characteristics. As Lindeman [1942, p.415] aptly stated “Analyses of food cycle relationships indicate that a biotic community cannot be clearly differentiated from its abiotic environment.”

[7] This study had a number of goals. The first was to ascertain if and how DOC concentrations in these wetlands varied in response to seasonal water loss. To this end, DOC concentrations were monitored at 12 prairie wetlands (ranging from freshwater to saline) over an 8-year period. The second goal of this study was to establish the relationship, if any, between DOC concentration and pond hydrological characteristics. To do this, the relationship between DOC and the biologically conservative chloride ion, in all of the study ponds, was studied. This relationship was examined in more detail in a freshwater and a saline study pond in order to determine what effect hydrological characteristics might have on DOC concentrations at differing times of the year. The third goal was to establish if DOC in saline and freshwater ponds was accumulating or being removed seasonally. The information gained from this study could be a critical step in understanding how climate change will ultimately affect not only water chemistry, but productivity in these ecologically important wetlands.

1.1. Study Sites

[8] All 12 study ponds were located within the St. Denis National Wildlife Area, 40 km east of Saskatoon, Saskatchewan, Canada (106°06′W, 52°02′N). This refuge is located within the mixed-grass prairie-parkland ecotone and covers 385 ha [Driver and Peden, 1977]. Moderately rolling knob and kettle moraines (slopes vary from 10 to 15%) underlain by glacial tills of the Battleford and Floral formations characterize the region [Hayashi, 1996]. Within the St. Denis National Wildlife Area there are more than 100 wetlands, ranging from ephemeral to permanent and freshwater to saline. The 90-year mean annual precipitation for Saskatoon (40 km west of the St. Denis site) is 360 mm while evaporation averages 690–710 mm [Hayashi et al., 1998a]. Summer rainfall is of short duration but high intensity, typical of convective storms in this semiarid climatic zone [Su et al., 2000].

[9] Two ponds, 50 (saline) and 109 (freshwater), were chosen for a more intensive study as hydrological characteristics for these ponds were well known [Woo and Rowsell, 1993; Hayashi et al., 1998a, 1998b]. Pond 50, 18,000 to 31,000 m2 in area, mean depth 73.6 ± 40.2 cm (SD) is a saline (mean specific conductivity 6172 μS cm−1 ± 3108 SD), semipermanent lowland wetland encircled by marsh. Water levels fluctuate widely on a yearly basis, and although filled with water most years, the pond may completely disappear depending on spring water inputs from snowmelt and runoff [Woo and Rowsell, 1993]. Evaporation accounts for approximately 70% of the water loss here [Woo and Rowsell, 1993] and plays a large role in determining seasonal water depths [Su et al., 2000]. Pond 109 is smaller than Pond 50, varying in area from 1000 to 5000 m2. This freshwater pond (mean specific conductivity 370 μS cm−1 ± 45 SD) is located in undulating uplands and is surrounded by willow trees [Hayashi et al., 1998a]. Pond 109 is ephemeral, usually disappearing by late August or earlier depending on the extent of snowmelt and summer precipitation (mean depth 52.7 ± 31.1 cm SD). The period of inundation (defined as the number of months for which water is present at the center) varied between 0 and 8 months for the interval 1980–1996 with an average of 3.6 months [Hayashi, 1996]. Evaporative stress is low mainly owing to the surrounding willows and extensive in-pond growth of macrophytes which effectively reduce its exposure to wind and sun. Evaporation only accounts for 20–30% of its seasonal water loss with the remaining water lost by infiltration to the pond margin [Hayashi et al., 1998a].

1.2. Prairie Wetland Hydrology

[10] One of the unique aspects of prairie wetlands is their lack of stream inflows and outflows [Winter et al., 2001]. Essentially, there has not been enough geologic time in this post-glacial landscape for the creation of streams which would connect wetland basins [Tiner, 2003]. Consequently, inflows usually only occur as a result of overland flow from runoff events. During the summer, for example, rainwater runoff from the catchment occurs only during occasional intense convective storms. But much of this precipitation tends to infiltrate the soil [Hayashi et al., 1998a] or evaporate [Woo and Rowsell, 1993] and is therefore a minor component of the long-term water balance [Su et al., 2000]. Prairie wetlands also have limited groundwater interaction because they are underlain by low-permeability glacial till. Pond 50 at the St. Denis site, for example, although located in the lowlands, exhibits little deep groundwater flow into the catchment due to low hydraulic conductivity of underlying clay-rich glacial deposits [Su et al., 2000]. Similarly, vertical flow to deep groundwater at Pond 109 is impeded by a layer of clay 8 to 9 m below the wetland [Hayashi et al., 1998a]. Of 481 mm of water leaving this wetland, only 2 mm becomes groundwater recharge while the rest infiltrates the pond margin by shallow horizontal subsurface flow (confirmed by chloride tracer movement [Hayashi et al., 1998a]).

[11] The lack of streams, and the relatively small amount of water which moves from groundwater to the wetlands and vice versa, means that water balance in prairie wetlands, such as those at St. Denis, is determined by four major hydrological processes: winter precipitation and the subsequent spring melt, evaporation, evapotranspiration, and infiltration [LaBaugh et al., 1996]. Generally speaking, prairie wetland systems are highly dependent on winter precipitation for their water supply and most are recharged in spring from snowmelt runoff on frozen or saturated soils [Covich et al., 1997]. At Ponds 50 and 109, for example, some spring overland flow is lost via infiltration into abundant cracks in frozen upland and slough soils. However, substantial increases in water storage still occur as a result of rapid release of meltwater and delivery of overland flow from upland areas [Woo and Rowsell, 1993; Hayashi et al., 1998a]. Water balance calculations from Ponds 50 and 109 indicated that anywhere from 12 to 30% of the annual precipitation became effective snowmelt runoff; snowmelt water that was not absorbed into upland soils but contributed to spring wetland water volume [Woo and Rowsell, 1993; Hayashi et al., 1998a]. At Pond 109 this effective runoff represented 30–60% of the winter precipitation in the catchment [Hayashi et al., 1998a]. Ponds also vary in the degree to which spring melt recharges water levels. Some ephemeral ponds, like Pond 109, are highly reliant on spring runoff. Here, large snowmelt runoff (Table 1) over frozen soils to the wetland proper forms a central pond thereby refilling the basin which had dried out the previous fall. Although spring melt is also important in recharging water levels in semipermanent Pond 50, snowmelt runoff is not as large a part of the water balance as at Pond 109 (Table 1).

Table 1. Water Balance for Ponds 109 and 50
PondPrecipitation, mmSnowmelt, mmVertical Groundwater Flow, mmHorizontal Groundwater Flow: Infiltration or Seepage, mmEvapotranspiration, mm
109a+376+400−2−480−300
50b+206+530−92−202

[12] Wetlands also vary in terms of the types of water loss processes which dominate during the ice-free season. Lowland, saline ponds tend to lose most of their water by evaporation (as noted above for Pond 50; see Table 1) while upland, freshwater ponds, like Pond 109, lose most of their water by infiltration to the pond margin and upland slope [Hayashi et al., 1998b] (Table 1).

2. Materials and Methods

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Materials and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusion
  8. Acknowledgments
  9. References

[13] Study ponds were sampled for total phosphorus (TP), total nitrogen (TN), chlorophyll a (Chl a), DOC, chloride, and specific conductivity from May through October during 1993–1996 (4 years). Ponds were sampled within a two-day period every 2 weeks from 1993–1995 and then monthly in 1996. In 1997 and 1998 water samples were collected on an intermittent basis while in 1999 (for Ponds 50 and 109 only) and 2000 (Pond 50 only), water samples were collected on a fortnightly basis during the ice-free season. Pond 109 remained dry in 1993 and 2000 owing to unusually low snowmelt runoff. The number of water samples taken varied between ponds and on a year to year basis depending on how long ponds remained water filled. Water samples were collected from a small boat paddled to the center of each pond. A clean bucket was lowered from the side of the boat to a depth of approximately 30 cm. Collected water was dispensed into clean, 2-L amber Nalgene bottles and then placed in coolers for subsequent transport back to the laboratory.

[14] For TP, water samples were analyzed using persulfate digestion, followed by addition of ammonium molybdate and then reduction using stannous chloride. The molybdenum blue complex was then measured spectrophotometrically at 660 nm [Environment Canada, 1992]. TN samples were oxidized to nitrate using alkaline persulfate, followed by reduction to nitrite using cadmium. Nitrite was then reacted with sulfanilamide and N-1(naphthyl)-ethylenediamine dihydrochloride to form a reddish purple azo dye. Absorbance was then measured spectrophotometrically at 550 nm [Environment Canada, 1992].

[15] Aliquots of water were drawn through 47-mm GF/C filters (Whatman, nominal pore size 1.2 μm) for subsequent Chl a analyses. Filters were wrapped in foil and transported on ice back to the lab. Chl a was extracted using 90% boiling ethanol and analyzed fluorometrically using a Turner Designs (Turner Designs, Sunnyvale, California) Model 10-AU digital fluorometer [Waiser and Robarts, 1997].

[16] DOC samples were screened through 153-μm-mesh-size Nitex netting, filtered through combusted GF/C filters (nominal pore size 1.2 μm), and then analyzed using UV oxidation followed by infrared detection [Environment Canada, 1992]. DOC values obtained using this method are highly correlated with those using high temperature combustion [Lean, 1998]. The coefficient of variation for DOC was calculated for each pond and then coefficients for waters of differing salinity statistically compared to determine if variation in DOC was higher in saline as compared to freshwater systems (t-test; P < 0.05)

[17] Water for chloride ion was filtered through 0.45-μm cellulose acetate filters and then analyzed using ion exchange chromatography according to methods from Environment Canada [1992]. Specific conductivity was measured using a Hydrolab H20 submersible probe with attached Surveyor 3 water quality data logging system (Hydrolab Corp., Austin, Texas).

3. Results

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Materials and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusion
  8. Acknowledgments
  9. References

3.1. Water Chemistry and Quality

[18] TP was highest in the freshwater ponds and appeared to decrease as salinity increased (Figure 1a). Pond 109 (fresh) exhibited the highest average TP concentration (1.8 mg L−1) while Pond 67, one of the most saline ponds, had the lowest (0.1 mg L−1). In all study ponds average TN concentrations were always >1.5 mg L−1 but no trends were apparent (Figure 1b). Chl a, on the other hand, was generally highest at either end of the salinity spectrum with lowest values in between (Figure 1c). Highest average Chl a concentrations were noted at Gursky's Pond (88.9 μg L−1) and lowest at Pond 65 (6.2 μg L−1). Number of samples taken for each parameter are noted on Figure 1b.

image

Figure 1. Box whisker plots illustrating the median and range in concentration of (a) Total P, (b) Total N, and (c) Chl a for each of the 12 study ponds at the St. Denis site (1993–2000). The boundary of the box closest to zero indicates the 25th percentile, the line within the box marks the median, and the boundary furthest from zero indicates the 75th percentile. Whiskers above and below the box (error bars) represent the 90th and 10th percentiles, respectively. Numbers within the Total N box whisker plot represent the number of samples taken at each pond for all parameters (TP, TN, and Chl a). From left to right, ponds increase in salinity.

Download figure to PowerPoint

3.2. DOC Concentration

[19] On average, DOC concentrations in all ponds were >10 mg L−1 and increased across a gradient of increasing salinity (Figure 2a). Average DOC concentrations ranged from a low of 19.7 mg L−1 ± 5.3 SD in Pond 86 (freshwater) to a high of 102.7 mg L−1 ± 32.2 SD in Pond 66 (most saline) (Figure 2a). A t-test, used to test for differences in DOC concentrations across the salinity gradient, indicated that DOC concentrations were higher in saline as compared to freshwater ponds (P < 0.05). Number of samples taken are noted in Table 2.

image

Figure 2. Box whisker plots indicating the median and variation in (a) DOC, (c) specific conductivity, and (d) and chloride ion for each of the 12 study ponds at the St. Denis National Wildlife Refuge. The boundary of the box closest to zero indicates the 25th percentile, the line within the box marks the median, and the boundary farthest from zero indicates the 75th percentile. Whiskers above and below the box (error bars) represent the 90th and 10th percentiles, respectively. (b) Average seasonal increases in DOC and chloride ion for each study pond (expressed as a percentage of the spring value). From left to right, ponds increase in salinity and decrease in percentage of water loss due to infiltration. Ponds in which the range in concentrations are obscured by the scale of the larger plot have been replotted on a smaller scale in the insets in Figures 2c and 2d.

Download figure to PowerPoint

Table 2. Number of Samples (n) Taken for Specific Conductivity, DOC, and Chloride Ion at the 12 Study Ponds for 1993–2000a
PondFresh or SalineSpecific ConductivityDOCChloride
  • a

    Wetlands appear in order of increasing salinity. Notation:, freshwater; IS, intermediate salinity; S, highly saline.

60F131414
109F254040
86F172727
4857IS313634
Gursky'sIS323531
25IS353633
1IS353631
65S353630
50S547155
67S333630
90S273331
66S323628

[20] In 10 of the 12 study ponds, DOC concentration increased from spring until fall or when the pond dried out (data not shown). Greatest average seasonal increase in DOC concentration (45.2 mg.L−1) was seen in Pond 90 (one of the most saline ponds), while the least increase (4.8 mg L−1) at Pond 109 (fresh water) (Figure 2b). In Pond 86 and Gursky's Pond, DOC concentrations decreased, on average, by 1.2% and 0.7%, respectively, from spring until fall (Figure 2b). The coefficient of variation for DOC was calculated for each pond and then compared. The results indicated higher variation in DOC concentrations in saline as compared to freshwater wetlands (t-test; P < 0.05).

3.3. Specific Conductivity and Chloride Ion Concentration

[21] According to the specific conductivity results, three study ponds were freshwater (specific conductivity <600 μS cm−1, Ponds 60, 109 and 86, designated as “F”) and the remaining nine were saline (specific conductivity >600 μS cm−1 [Curtis and Adams, 1995]). Of the saline ponds, it appeared that four were of intermediate salinity (mean specific conductivity values ranging from 600 to 3000 μS cm−1, Ponds 1, 25, 4857 and Gursky's, designated as “IS”) and the remaining five highly saline (mean specific conductivity values >3000 μS cm−1, Ponds 65, 50, 67, 90 and 66, designated as “S”). Conductivities ranged from 312 ± 77 (SD) μS cm−1 in Pond 60 (most freshwater pond) to 33,493 ± 14,332 (SD) μS cm−1 in Pond 66 (most saline pond) (Figure 2c).

[22] Chloride ion concentration ranged from 4.7 mg L−1 ± 2.3 SD in Pond 86 (freshwater) to 927 mg L−1 ± 570 SD in highly saline Pond 66 (Figure 2d). The greatest seasonal increases in chloride ion were seen in one of the more saline ponds (Pond 50, 200% over spring time values) while the least (9%) was seen in Gursky's pond (large pond of intermediate salinity) (Figure 2b). There was less variation in chloride ion concentration in freshwater as opposed to saline wetlands (t-test; P < 0.05). Number of samples taken for conductivity and chloride ion are noted in Table 2.

3.4. Relationship Between Chloride Ion and Pond Volume

[23] In order to validate the usage of chloride as a surrogate for pond hydrological characteristics, pond volume was regressed against chloride ion concentration in two ponds, 50 and 109, which spanned the freshwater-saline continuum (Figure 3). Pond volumes for each date when water levels were available (1993–2000 for Pond 50; 1994–1999 for Pond 109) were estimated using the simple power equation developed by Hayashi and van der Kamp [2000] for small depressional wetlands,

  • equation image

where V is volume of the pond; s is scaling constant equal to the area of the pond measured when h = h0; h0 = 1 m; h is measured pond depth; and p is constant related to the geometry of the depression (values range between 1 and 5; smaller wetlands tend to have lower values of p and vice versa). Values of p and s were derived by the least squares method from the Volume-Area-depth (V-A-h) relationship established for each pond [Hayashi and van der Kamp, 2000]. Theoretically, the slope of chloride versus pond volume line should range from 0, if all water loss was due to infiltration, to 1 if water loss processes were dominated by evaporation.

image

Figure 3. (a) Differing relationships between pond volume and chloride ion concentration in Ponds 50 (1993–2000) and 109 (1994–1999). (b) Owing to the large difference in pond volumes, Pond 109 data have also been plotted on a separate graph for ease of interpretation. Dashed lines on both graphs represent the linear regressions of chloride ion versus pond volume while solid lines indicate the theoretical slopes of the lines given conservative behavior of chloride ion.

Download figure to PowerPoint

[24] In Pond 50, slope of the chloride versus pond volume regression line was 0.50 and the number of observations (n) was 44 (Figure 3). The intercept of this regression was 149.9 and the r2 was 0.52. Probability of a greater t and F value were both <0.001 indicating that there was a significant relationship between chloride ion concentration and pond volume. For Pond 109 the slope was 0.02, n was 40, the intercept was 5.2, and r2 was 0.13. Probabilities of greater t and F values were both 0.024 indicating that there was also a relationship here (albeit weak; see r2 values above) between pond volume and chloride ion concentration. It is clear from these relationships that chloride ion can now be used as a surrogate for pond hydrological characteristics.

3.5. Relationship Between DOC and Hydrological Characteristics

[25] In order to evaluate the effect of changing pond hydrological characteristics on DOC concentrations, DOC was regressed against chloride ion and DOC:chloride ratios were examined in all of the study ponds. If DOC was behaving conservatively, then the slope of the regression line of DOC versus chloride ion should be 1. In this case, observed seasonal increases in DOC concentrations could be ascribed to evaporation alone. If, however, the slope of the line was greater or less than 1, indicating nonconservative DOC behavior, then other within pond factors might be affecting DOC concentrations. Seasonal trends in the DOC:chloride ratios were also examined to provide possible clues as to whether ponds experienced net loss or accumulation of DOC seasonally. This data would not only provide greater information on what other factors might be affecting DOC concentrations but also how DOC tended to cycle seasonally in these ponds.

3.6. DOC Versus Chloride Regressions

[26] When DOC was plotted against chloride ion for each pond, the ponds divided into three groups based on slope (Table 3). The freshwater ponds (60, 86 and 109) all had slopes greater than 1 (3.3, 1.8 and 1.9, respectively). Ponds of intermediate salinity (1, 25, 4857 and Gursky's) had slopes ranging between 0.6 and 1.0 (0.9, 1.0, 1.0 and 0.6, respectively) while the slope of the relationship for the most saline ponds (65, 50, 67, 90 and 66) were all less than 0.5 (0.2, 0.3, 0.45, 0.2 and 0.03, respectively). Examples of DOC versus chloride regression lines for a representative freshwater pond (109) and a saline pond (50) are shown in Figures 4a and 4b, respectively.

image

Figure 4. Relationship between DOC and chloride ion in representative (a) freshwater (Pond 109) and (b) saline (Pond 50) ponds. Conservative behavior of DOC is indicated by the 1:1 slope of the DOC versus chloride regression line (solid line).

Download figure to PowerPoint

Table 3. Intercepts, Slopes, t and F Probabilities, and r2 Values of the DOC Versus Chloride Regressions for All St. Denis Study Pondsa
POND nInterceptSlopeP (t) SlopeP (F)r2
  • a

    Wetlands appear in order of increasing salinity. Notation: n, number of samples; F, freshwater; IS, intermediate salinity; S, highly saline.

  • b

    These regressions are significant.

60bF1412.93.3<0.001<0.0010.84
109F4011.91.80.0070.0070.16
86bF2713.51.90.0040.0040.23
4857bIS338.60.9<0.001<0.0010.75
Gursky'sIS3118.21.00.110.110.05
1bIS348.31.0<0.001<0.0010.60
25bIS3111.80.6<0.001<0.0010.75
65bS3022.20.20.010.010.20
50bS5519.90.3<0.001<0.0010.91
67bS3019.40.5<0.001<0.0010.59
90bS3120.90.2<0.001<0.0010.63
66S2876.10.030.060.060.09

3.7. DOC: Chloride Ratios

[27] The study ponds also divided into three groups based on DOC:chloride ratios (Figure 5). In the highly saline ponds these ratios were always <1.0, while in the freshwater ponds they were always >2. In the ponds of intermediate salinity, however, ratios ranged between 1 and 2 (Figure 5). Using these three groups, (fresh, intermediate and highly saline) the DOC:chloride ratio data was compared using a Kruskal-Wallis ANOVA on ranks and a post hoc Dunn's test (Sigmastat). According to the results, the median values of the groups were greater than would be expected by chance (P < 0.001) and the ratios in each group were significantly different from each other (P < 0.05).

image

Figure 5. Box whisker plots indicating the median and range of DOC:chloride for each of the 12 study ponds. The boundary of the box closest to zero indicates the 25th percentile, the line within the box marks the median, and the boundary furthest from zero indicates the 75th percentile. Whiskers above and below the box (error bars) represent the 90th and 10th percentiles, respectively. Conservative behavior of DOC occurs when DOC:chloride ratios = 1 (as indicated by the line on the graph). From left to right, ponds increase in salinity and decrease in percentage of water loss due to infiltration. Ponds in which the range in concentrations are obscured by the scale of the larger plot have been replotted on a smaller scale in the inset.

Download figure to PowerPoint

3.8. DOC and Chloride Mass

[28] Estimates of DOC mass for Ponds 109 and 50 were made by multiplying pond volume estimates (see above) by DOC concentration. Estimates of DOC sources and losses (from runoff, sediments, photolytic decomposition, primary production) could then be made in order to determine the relative importance of DOC inputs and losses with respect to overall DOC mass in these two ponds.

[29] In Pond 109, DOC and chloride mass generally tracked increases or decreases in pond volume (Figure 6a). Mass of both DOC and chloride usually increased in spring and early summer and then both declined dramatically with seasonal decreases in pond volume.

image

Figure 6. Estimated water volume and total mass of DOC and chloride for (a) Pond 109, 1994–1999, and (b) Pond 50, 1993–2000. Asterisks in Figure 6b indicate large excursions in DOC mass from 1995 to 1996 associated with large increases in pond volume. The arrows (1996 and 1998, Figure 6b) indicate years in which DOC and/or chloride mass increased when pond volume decreased.

Download figure to PowerPoint

[30] An examination of the DOC and chloride mass data revealed some generalized trends in Pond 50. In most years, for example, mass of both constituents appeared to increase or decrease concomitantly with changing pond volume (Figure 6b). Large excursions in mass were particularly apparent when there were significant increases in volume from the fall of one year to the following spring and early summer of the next year, for instance increase in DOC mass from fall 1995 to spring 1996 (see asterisks in Figure 6b). At times, however, DOC and/or chloride mass increased when pond volume decreased, i.e., chloride ion in 1996 (arrow in Figure 6b) and chloride and DOC in 1998 (arrows in Figure 6b). In all years, peaks in DOC mass usually occurred between late June and early July. In most instances, DOC and chloride mass did not fall below the mass initially present in the spring (Figure 6b).

4. Discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Materials and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusion
  8. Acknowledgments
  9. References

4.1. Variation in DOC Concentrations

[31] This study clearly shows that DOC concentrations in prairie wetlands vary across a salinity gradient and also in a dynamic fashion seasonally (Figures 2b and 2b). Saline ponds in this study exhibited significantly higher DOC concentrations than their freshwater counterparts, similar to observations made by Curtis and Prepas [1993] and Curtis and Adams [1995] in their studies of Alberta freshwater and saline lakes. As well, DOC concentrations increased from spring until fall, by an average of 21 mg L−1 in those wetlands where increases were noted. In some cases DOC concentrations increased by as much as a factor of 3 (Figure 2b). As well, the DOC coefficient of variation was much greater in the saline systems as compared to their freshwater counterparts. Unfortunately, there are few studies with which to compare these observations. A study of 30 emergent and aquatic bed wetlands in central and northeastern Minnesota and northwestern Wisconsin, for example, revealed constant DOC concentrations from spring through until fall [Peterson et al., 2002]. Similarly, DOC concentrations in two small lakes on the Pocono Plateau, in NE Pennsylvania were relatively unchanged throughout the ice-free season [Osburn et al., 2001]. When increases are reported, however, they are much smaller than those detailed here. In Lake Skjervatjern, Sweden, for example, DOC concentrations only increased from spring to fall by 6 mg L−1 [Hessen et al., 1997].

[32] What then was the explanation for such large seasonal increases in DOC concentration? In lakes, DOC can be concentrated or diluted when water is exchanged with the atmosphere [Curtis, 1998]. In some aquatic systems, DOC decreases in response to increased precipitation (dilution) while in others it decreases in response to drought [Schindler et al., 1992]. Dissolved organic matter (DOM) concentrations in lakes can be estimated using the following equation:

  • equation image

where YDOM is yield of organic matter as DOC (g m−2 yr−1) from catchment, P-Et is excess precipitation or precipitation minus evapotranspiration, P is precipitation, E is evaporation, Ao is lake area, and Ad is drainage area [Curtis, 1998]. This equation illustrates how moisture balance in a climatic zone could theoretically regulate within system DOC concentrations and explains the increasing DOC concentrations observed as one moves from humid, temperate Ontario to the semiarid climes of the prairies [Curtis, 1998]. However, in prairie wetlands, the equation terms associated with DOC inputs from the basin (i.e., from streamflow), or as a result of precipitation are likely minimal as discussed above (see section 1.2). When these terms are ruled out, the equation simplifies to one in which DOC concentration depends, for the most part, on terms defining the manner in which water is lost from the basin. Therefore that is where this story begins.

[33] Most prairie wetlands in this study lost considerable water volume and therefore depth during the ice free season. Additionally, the manner in which water is lost from shallow wetlands has a great effect on water chemistry. Those which lose most of their water by evaporation, for example, tend to exhibit greater seasonal increases in salinity than those where the majority of water is lost via infiltration to the surrounding slope [Fritz, 1996]. The effect of differing types of water loss on salinity was strikingly similar to observations of greater seasonal increases in DOC concentrations in saline as compared to freshwater ponds (Figures 2a and 2b). Concomitant increases and variation in DOC and chloride ion concentration in most ponds were also observed (Figures 2a and 2d). When this evidence was taken together, it would appear that the simplest explanation for the large seasonal increases in DOC concentrations observed in some of the study ponds was that DOC was simply concentrating as water was lost (evapoconcentration). It seemed reasonable, therefore, to look for possible relationships between pond hydrological characteristics and DOC concentrations.

4.2. Relationship Between DOC Concentration and Pond Hydrological Characteristics

4.2.1. Influence of Spring Recharge Events

[34] Spring runoff events are important in recharging water levels in prairie wetlands [Hayashi et al., 1998a]. As such, they may figure largely in resetting DOC concentrations, especially in small, ephemeral ponds (like Pond 109) which are totally refilled in spring. In order to assess the importance of spring recharge as an input of DOC at Pond 109, the change in water volume from the fall of one year to the following spring was first calculated. It was then assumed, on the basis of what was known regarding its hydrological characteristics, that the observed increase in volume arose solely from snowmelt runoff to the pond. Average concentration of DOC in spring runoff water (12.8 mg L−1, M. J. Waiser, unpublished data, 1999) was then multiplied by the spring volume increase in order to calculate the possible contribution of meltwater to the within pond DOC pool. According to these calculations, the mass of DOC accruing to Pond 109 from spring runoff accounted for, on average, 68% of the total DOC mass present within the water of the wetland proper at this time. In years when runoff was much higher (e.g., 1995 and 1996), its contribution to DOC mass was also higher (80 and 125%, respectively). It is interesting to note that the percentage of humic acids in pond DOC is at its highest (11%) in spring [Waiser, 2001] providing further support for the contention that runoff DOC, which originates in the uplands, is an important source of pond DOC in the spring.

[35] Some of the remaining DOC mass, not accounted for by runoff inputs, probably originates in the top few cm of pond sediments, diffusing into the water column once the dry basin is flooded. In 1996, for example, within pond spring DOC concentrations are about 16 mg L−1 (M. J. Waiser, unpublished data, 1999). Because pore water DOC concentrations were never measured in Pond 109, we used a value of 59 mg L−1 a value derived from the top 5 cm of sediment pore waters at Pond 50 (T. Mayer, unpublished data, 1999). Using Fick's diffusion equation, assuming a diffusion coefficient of 3.5 × 10−5 cm s−1 [Burdidge and Homstead, 1994], and using the concentration difference between the sediment pore waters and overlying waters, diffusion of DOC mass at this time could be 1.6 mg m−2 d−1. Taking into account a 45-day spring time frame, and the surface area of the sediments, diffusion from sediments could only account for about 1% of the total DOC mass present in the pond. This value could be higher, however, depending on the degree to which sediments were initially disturbed during spring refill. From the data presented here, it appears that an important source of DOC to small, ephemeral ponds like Pond 109 is spring runoff. As well, runoff events are probably quite important in setting spring DOC concentrations in these systems.

[36] In Pond 50, calculations for spring runoff contribution to DOC mass were slightly different because this semipermanent pond never totally dried up. DOC mass contained in the water and ice over winter could therefore be carried over to the following spring. “Carryover” DOC mass (that which was present in the pond the previous fall) was first subtracted from the total spring DOC mass to give estimate of how much ‘new’ DOC mass had been added to the pond. DOC mass in spring runoff was then calculated in the following way. First, it was assumed that the increase in volume from the fall of one year to spring of the following year could be attributed to runoff. Second, this volume was then multiplied by the DOC concentration in runoff water (7.6 mg L−1, M. J. Waiser, unpublished data, 1999) in order to estimate the DOC mass added to the pond as a result of runoff. This amount of DOC was then divided by the “new” DOC mass (see above) to give an estimate of the importance of runoff DOC to “new” spring DOC mass. According to these rough calculations, there were only 3 years (1994, 1996 and 1997) during which spring runoff contributed in any way to new DOC mass in the pond (Figure 6b). DOC mass in these high runoff years represented 28% (1994) and 60% (1996 and 1997) of new spring DOC mass. In all of the other study years (except 1995) during which there were small increases in volume, there were no corresponding increases in DOC mass

[37] Some of the remaining unaccounted for DOC mass in spring in Pond 50 may originate in the sediments, moving from the sediment into the overlying water in response to a negative concentration gradient. Major ions can and do enter sediments during freezeup and then return to overlying waters when sediments thaw in the spring [Flicken, 1967]. The same process might apply to DOC. Using 1997 as an example, within pond spring DOC concentrations were about 24 mg L−1, much lower than sediment pore waters (59 mg L−1, T. Mayer, unpublished data, 1999). Using Fick's first law as above, assuming a diffusion coefficient of 3.5 × 10−5 cm s−1 [Burdidge and Homstead, 1994], and using the concentration difference between the sediment pore waters and overlying waters, diffusion of DOC mass at this time could be as high as 1.4 mg m−2 d−1. Taking into account the estimated sediment area, and a 45-day time frame for spring, the total amount of DOC moving from the sediments into the overlying water could be about 1.6 × 106 mg or only about 2% of the new DOC mass once runoff has been accounted for. This number, however, could rise, depending on the extent to which water column mixing at this time affects diffusion from sediments into the upper water layers.

[38] At Pond 50, DOC in runoff appears to be important in setting spring time DOC concentrations only in high runoff years. In other years, springtime DOC concentrations are largely determined by DOC present in the pond the previous fall.

4.2.2. Influence of Infiltration and Evaporation After Spring Runoff

[39] After spring recharge events, water levels during the ice free period in prairie wetlands are influenced by infiltration and evaporation/evapotranspiration [Woo and Rowsell, 1993; Hayashi et al., 1998a]. On the basis of what is known regarding these hydrological processes, how can their connection to DOC concentrations be made? One way is to first look for relationships between a conservative tracer and pond hydrological characteristics and then, the relationship between this conservative tracer and DOC concentration. Of all the solutes which characterize natural waters, chloride is usually conserved over the widest concentration range [Eugster and Jones, 1979]. As a result, it has been used extensively to monitor water balance components like evaporation and infiltration [Classen and Helm, 1996; Hayashi et al., 1998b]. When chloride ion was regressed against pond volume for Ponds 50 and 109, the resulting slopes of the lines were quite different (Figure 3). For Pond 109, the slope approached zero indicating little change in chloride concentration with changing pond volume (Figure 3), and this was consistent with previous intensive hydrological investigations which revealed that the majority of chloride entered this pond with spring runoff (∼93% of total; the remaining chloride enters through atmospheric deposition) but was subsequently transferred to the margin and upland areas by infiltration (shallow horizontal subsurface flow [Hayashi et al., 1998a]). In contrast, the slope for the Pond 50 regression line was higher than that for Pond 109 (Figure 3) indicating the effect of evaporation on chloride ion concentration: As pond volume decreased, chloride ion concentration increased. On the basis of this evidence, chloride ion could now be used as a surrogate for differing pond hydrological characteristics and the relationship between DOC and these characteristics in all of the study ponds could be explored.

[40] If DOC were behaving conservatively, then observed increases in DOC concentration would be solely due to evaporation (evapoconcentration). Both DOC:chloride ratios and the slopes of the regressions of DOC versus chloride ion were examined in order to make inferences regarding the responsiveness of DOC concentrations to evaporative water loss processes (Figures 4 and 5 and Table 3). Ratios and slopes of 1, for example, would indicate conservative behavior of DOC. On the basis of the slopes of the DOC versus chloride regression lines and average DOC:chloride ratios (Table 4 and Figures 4 and 5), the study ponds divided into three groups: freshwater, intermediate salinity, and saline. The slopes of the lines and the ratios approached 1 in just three of the ponds of intermediate salinity (4857, 1 and Gursky's) indicating that DOC is behaving more or less conservatively in these systems (Table 3). Seasonal increases in DOC concentrations in these systems can therefore be ascribed to evaporation. The story was different in the freshwater and more saline ponds.

Table 4. Estimates of Possible Carbon Source and Loss Terms for Ponds 109 and 50 After Spring Runoff
PondEstimated Carbon Source Term, mgPossible SourcesEstimated Sources, mgPercent of Total Source Term
1097.24 × 106excretion of labile DOC from pelagic primary production6.36 × 106≥42
 excretion of labile DOC from macrophytes and biofilmsunknownunknown
 decomposition (POC to DOC)unknownunknown
 sediments4.7 × 1040.6
estimated carbon loss term (mg)possible lossesestimated loss (mg)percent of total loss term
 
502.94 × 108bacterial decomposition of DOC1.66 × 10857
 infiltration of DOC to pond margin1.11 × 10839
 photolytic breakdown of DOC to CO22.91 × 10710
 sediments2.94 × 106∼1
4.2.3. Freshwater Ponds as Sources of DOC After Spring Runoff

[41] Using Pond 109 as a representative freshwater pond, several observations were made. First, despite large seasonal decreases in pond volume, concomitant large increases in DOC and chloride ion concentrations were not seen (Figures 2a, 2b, and 2d). For example, on average, DOC concentrations just prior to the pond drying up later in the season were only 23% higher than spring values. As well, the relationship between DOC and chloride ion was weak (r2 = 0.16; n = 40; Table 3). Finally, steep declines in both DOC and chloride mass occurred every year as the pond dried up (Figure 6a). Hayashi et al. [1998b] noted that the concentration of chloride in this pond did not increase significantly during the summer owing to the fact that ∼80% of the water infiltrates the pond margin with little evaporitic enrichment. This strong infiltration effect also appears to keep DOC concentrations fairly constant within this pond on a seasonal basis.

[42] However, this was not the whole story. Although DOC concentrations remained relatively constant, there was evidence that within pond cycling of DOC was occurring on a seasonal timescale. For example, the average DOC:chloride ratio here was 5, while the slope of the DOC/chloride regression was 1.8 (Table 3). Both indicated nonconservative DOC behavior. As well, DOC:chloride ratios tended to increase seasonally meaning either that DOC was increasing with respect to chloride ion or chloride was decreasing. Examination of seasonal changes in DOC and chloride ion concentration showed that DOC was actually increasing. Consequently, observed seasonal increases were due to the fact that the pond was experiencing net DOC accumulation. In other words, slightly more DOC was being produced within the pond every year than was lost via infiltration and hence the increase in ratios with time. In fact, the mass of DOC in excess of chloride in Pond 109 averaged about 7.24 × 106 mg per year over the study period.

[43] There are a number of possible sources for this “excess” DOC mass, including excretion of DOC by macrophytes and attached and pelagic phytoplankton (during photosynthesis), decomposition of particulate organic carbon (POC) to DOC and flux from the sediments. At Pond 109, pelagic primary production (PP) was measured every week from May until the pond dried up in both 1998 and 1999 [Waiser, 2001] and the results indicated that PP rates in this pond were high: on average 888 mg C m−3 d−1 [Waiser and Robarts, 2004a]. Under normal growth conditions, phytoplankton excrete between 5 and 25% of carbon incorporated during photosynthesis [Søndergaard et al., 1985]. The percentage contribution of this excreted carbon to the DOC source term in this pond can now be calculated by assuming that 5–25% of 888 mg C is excreted per day, then multiplying this number by 150 (estimated growing season) and then by 480 (average pond volume in m3). According to these calculations, between 42 to 100% of the average excess DOC mass in this pond (noted above) could theoretically come from pelagic primary production (Table 4). On the basis of the 450:500 fluorescence ratio of DOC, however, it is likely that pelagic primary production only accounts for the lower percentage of the excess DOC term here. The 450:500 fluorescence ratio of DOC from Pond 109 averages 1.42 [Waiser, 2001] which is closer to the signal for an allochthonous (i.e., terrestrial) DOC source (fluorescence ratio of 1.4) than an autochthonous one (fluorescence ratio of 1.9) [McKnight et al., 2001].

[44] In some aquatic ecosystems, increased primary production of macrophytes and periphyton (biofilms) may represent an important within pond DOC source [Mann and Wetzel, 1995]. A recent article states that although there are no estimates of the rate of flow of littoral DOC to the pelagic zone, there is substantial experimental evidence suggesting that both macrophytes and benthic algae may be important DOC sources [Vadeboncoeur et al., 2002]. Although DOC excreted by macrophytes or microbial biofilms may be important, these production rates have not been measured and so it is impossible to estimate their importance to the DOC source term at Pond 109.

[45] Another possible within pond source of DOC is flux from sediment pore waters. DOC sediment flux in Pond 109 was calculated using the same equation, diffusion coefficient and pore water DOC concentration (59 mg L−1) as noted above. According to these calculations, maximum DOC flux from the sediments could be about 1.6 mg of DOC per m3 d−1. Taking into account the sediment area and 150 days (which approximates the ice free season), sediment flux amounts to roughly 4.7 × 104 mg of DOC, or only about 0.6% of the excess DOC mass in this pond (Table 4). Although mixing events within the pond could increase these estimates somewhat, it appears that sediment pore waters are not a major source of DOC in this pond.

[46] Given the shallow nature of Pond 109, its high average ice-free temperatures (20.0°C [Waiser, 2001]), high bacterial numbers (mean = 5.5 × 106 mL−1 [Waiser, 2001]) and rates of bacterial production (seasonal average = 197 mg C m−3 d−1 [Waiser and Robarts, 2004a], rates of microbial decomposition of POC to DOC could be high as well. Consequently, decomposition processes may be an additional DOC source in this pond. Although decomposition rates have not been measured here, ratios of alkyl to O-alkyl carbon from 13C NMR spectra of hydrophobic DOC constituents extracted from Pond 109 water [Waiser, 2001] are high (∼1.10) when compared to similar ratios in the surrounding soil (0.55 [Jokic et al., 2003]). The percentage of total carbon atoms in the paraffinic region of the 13C NMR DOC spectra (0–55 ppm [Orem and Hatcher, 1987]) is also high (36%). Such elevated ratios and percentages are thought to indicate a high degree of decomposition [Orem and Hatcher, 1987]. Such evidence may provide some indirect support for the idea that decomposition may be a DOC source here and would certainly be consistent with the 450:500 fluorescence signal which indicates an allochthonous DOC source (see above).

[47] In summary, Pond 109 appears to experience net accumulation of DOC on seasonal timescales. The major sources for this DOC could be excretion by pelagic phytoplankton, macrophytes and attached biofilms, as well as microbial conversion of POC to DOC (decomposition).

4.2.4. DOC Losses From Saline Ponds After Spring Runoff

[48] In Pond 50, maxima in DOC mass seen every year usually in late June or early July (Figure 6b) may be related to the cooccurrence of peaks in primary production and algal biomass as noted by Waiser and Robarts [2004a] in their 3-year study of primary production and algal biomass here. However, large concomitant seasonal increases in chloride and DOC concentrations (on average, fall concentrations were twice that in spring; Figure 2b) and a robust relationship between these two variables (r2 = 0.91; n = 55; Table 4 and Figure 4) were also noted here. These data suggested that dramatic seasonal increases in DOC concentration were a direct result of evaporative water loss from the pond (evapoconcentration). Evaporation dominates water loss processes in this pond mainly because the ratio of inundated shoreline (where infiltration occurs) to pond area is so small [Millar, 1971]. Evapoconcentration of DOC and other water constituents has certainly been reported in other studies. A long-term study of wetlands in North Dakota revealed that conductivity in one of the study ponds increased dramatically seasonally and over longer timescales due to seasonal water losses and drought [LaBaugh et al., 1996]. During a 4-year drought in Wisconsin, increases in conservative cations (Mg and Ca) in five of seven study lakes were linked to evaporative water loss [Webster et al., 1996]. As well, the increase in DOC concentrations across a freshwater to saline gradient in Alberta lakes has been attributed to evapoconcentration of DOC over long timescales [Curtis and Prepas, 1993; Curtis and Adams, 1995].

[49] Although DOC was evapoconcentrating, there were indications, that it was not behaving conservatively. DOC:chloride ratios, for example, were always <1 (mean = 0.54; SD = 0.14; n = 55) as was the slope of the DOC versus chloride ion regression (Table 3: 0.3; n = 55). Furthermore, DOC:chloride ratios decreased over seasonal timescales (data not shown). Given that both chloride and DOC concentrations increase seasonally every year, one constituent must be increasing faster than the other, and in fact, plots of both constituents versus Julian day for all study years reveal that chloride ion concentration increases faster (slope = 0.51) than does DOC (slope = 0.20) in response to water loss (data not shown). These findings tend to support the idea that although seasonal increases in DOC concentrations are likely due to evapoconcentration, some DOC is actually being lost. In other words, saline ponds, like Pond 50, appear to experience net losses of DOC over short timescales. The average DOC loss term in this pond (i.e., the difference between mean chloride and DOC mass) is about 2.9 × 108 mg DOC per year.

[50] Possible within pond loss mechanisms for DOC include: utilization by bacteria (i.e., incorporation of DOC into POC), photolytic decomposition of DOC to DIC, infiltration to the pond margin, and flux to the sediments. A number of “bioavailable” carbon experiments were carried out as part of another study in an effort to estimate the percentage of ambient DOC which could be utilized by bacteria in Pond 50 [Waiser, 2001]. According to the results, bacteria here utilized, on average, 6% of the ambient DOC pool over a 21-day incubation period. Using this number and an estimated growing season figure of 150 days, bacterial DOC utilization could theoretically amount to about 1.66 × 108 mg of DOC which represents 57% of the DOC loss term here (Table 4).

[51] Photolytic breakdown of DOC to dissolved inorganic carbon (DIC) may be another way in which carbon is lost from this pond. A number of recent studies have shown that gaseous products like CO, CO2, as well as other forms of DIC, may be formed when UV light interacts with DOC [Valentine and Zepp, 1993; Granéli et al., 1996]. Using a photolytic production rate for CO2 of 0.039 μmol C L−1 h−1 and assuming that one day receives 7.6 hours of noon time sun (i.e., 7.6 hours of light at an intensity expected at noon), it has been estimated that the rate of photolytic production of CO2 could be as high as 2.3 ± 0.7 μmol L−1d−1 [Mopper and Kieber, 2000]. Using this number and again assuming a 150-day growing season, photolytic breakdown of DOC to CO2 could theoretically account for the loss of about 2.9 × 107 mg of DOC or approximately 10% of the DOC loss term (Table 4).

[52] DOC may also be lost to the sediments in this pond. Such DOC flux likely occurs only in the latter half of the ice free season (mid-July until pond freezeup). During this time frame, average DOC concentrations in pond water are greater than those in pore waters (i.e., 71 mg L−1 in pond water compared to pore water concentrations of 59 mg L−1). Although DOC flux to the sediments is possible at this time, flux estimates (using Ficks' law and the diffusion coefficient noted previously) indicated that this process was relatively unimportant accounting for only about 1% of the total DOC loss term here (Table 4).

[53] Finally, DOC in this pond may also be lost as water infiltrates the pond margin seasonally. On the basis of an average DOC concentration of 55 mg L−1, an average pond volume of 7.04 × 106 L and a 30% water loss due to infiltration to the pond margin (infiltration estimate from Woo and Rowsell [1993]), the calculated DOC loss term due to infiltration could be as high as 1.16 × 108 mg DOC or roughly 39% of the DOC loss term for this pond (Table 4). Taken together, it appears that bacterial utilization (DOC to POC), photolytic decomposition of DOC to DIC, and DOC loss via infiltration to the pond margin could, theoretically, account for most of the DOC loss term at Pond 50.

5. Conclusion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Materials and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusion
  8. Acknowledgments
  9. References

[54] This study suggests that hydrological characteristics are the most important factor affecting DOC concentrations in closed basin prairie wetlands. The linkage between DOC concentrations and hydrological characteristics was made by first establishing the relationship between pond volume and a conservative tracer, chloride ion and then the relationship between DOC concentration and chloride ion. This study clearly demonstrated that DOC concentrations increase across a salinity gradient, a finding which reflected the influence that differing hydrological characteristics of freshwater and saline systems have on DOC concentrations over long timescales. Freshwater ponds which lost most of their water via infiltration to the pond margin had lower DOC concentration than saline ponds which lost most of their water by evaporation. The influence of hydrological characteristics on DOC concentrations is also seen on shorter seasonal timescales. DOC concentrations in freshwater ponds vary less seasonally than in their saline counterparts, again mainly owing to their differing hydrological characteristics. Large seasonal increases in DOC concentrations in saline ponds are principally due to evapoconcentration.

[55] The hydrologically dynamic nature of these ponds allows researchers to observe, within the span of a year or several years, changes in water chemistry which other larger aquatic ecosystems might experience over much longer timescales. A distinct aspect of this study, therefore, was the ability to utilize DOC:chloride ratios to track within pond sources and losses of DOC on short timescales. On the basis of these ratios, this study suggests that within pond DOC cycling is occurring and that it is another factor affecting DOC concentration in prairie wetlands. Some ponds tend to be net sources of DOC, while others are net sinks (i.e., DOC is lost). Seasonal increases in DOC:chloride ratios in freshwater ponds indicate that these ponds are likely sources of DOC when this shorter timescale is examined. In contrast, declining seasonal DOC:chloride ratios in saline ponds indicate that on shorter timescales, these ponds are actually losing DOC seasonally. Within pond cycling of DOC also provides an explanation for the fact that DOC behavior in these wetlands is not totally conservative.

[56] Prairie wetlands are unique because linkages can be made between hydrological characteristics and chloride ion over relatively short timescales, and therefore these characteristics and DOC concentration. Because hydrological processes are so intimately linked to climate, the relationship between climate and DOC concentrations can then be made. Gaining an understanding of these relationship is vital given existing paleoecological evidence, current prairie climate change scenarios, and the importance of DOC to aquatic ecosystems. Studies, like this one, can provide us with insights into what might occur under current climate change scenarios. For example, how might prairie wetlands be affected given predictions of decreased snowfall (decreased spring runoff) and increasing summer precipitation? According to this study, spring runoff can be a significant factor in setting spring time DOC concentrations in some prairie wetlands. If less DOC comes in via runoff, concentrations could be lower. If more precipitation occurs in summer as a result of large storm events, dilution of DOC may occur and concentrations could be lowered again. DOC in prairie wetlands already undergoes considerable photobleaching seasonally [Waiser and Robarts, 2004b]. The combination of increased photobleaching, lowered DOC concentrations and water levels, may mean increasing penetration of harmful ultraviolet radiation (UVR). Although the effect on overall wetland productivity is unknown, algal and bacterial production in some prairie wetlands is already negatively affected by UVR, despite high ambient DOC concentrations [Waiser, 2001]. Not only does this research provide and understanding and appreciation of what might happen under current climate change scenarios, it may also give direction to future research. Given the variability of the prairie climate, even on a year to year basis, prairie wetlands might prove to be ideal microcosms in which to study and perhaps model the interacting effects of climate change and hydrological characteristics on DOC concentrations in closed basin aquatic ecosystems. Such research can only improve our understanding of how climate change may affect these tremendously productive wetland ecosystems.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Materials and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusion
  8. Acknowledgments
  9. References

[57] I am grateful for the expert technical and field assistance of Jennifer Holm and Vijay Tumber and indebted to G. van der Kamp and Richard Robarts, NWRI, for helpful discussions and comments on an earlier version of the manuscript. I am also appreciative of valuable comments and insights provided by Sherry Schiff which greatly improved the manuscript. I also thank M. Hayashi, University of Calgary, who developed the V-a-h relationships used for the mass calculations for both Ponds 109 and 50. The paper benefited greatly from comments of three anonymous reviewers. This research was partially funded by an IWWR, Ducks Unlimited, grant to Richard D. Robarts.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Materials and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  • Arts, M. A., R. D. Robarts, F. Kasai, M. J. Waiser, V. P. Tumber, A. L. Plante, H. Rai, and H. J. de Lange (2000), The attenuation of ultraviolet radiation in high dissolved organic carbon waters of wetlands and lakes on the northern Great Plains, Limnol. Oceanogr., 45, 292299.
  • Banks, T., and E. Cochrane (2005), Water in the west: Under pressure–Fourth interim report of the Canadian Standing Senate Committee on Energy, the Environment and Natural Resources, report, 21 pp., Can. Senate, Ottawa.
  • Batt, B. D., M. G. Anderson, C. D. Anderson, and F. D. Caswell (1989), The use of prairie potholes by North American ducks, in Northern Prairie Wetlands, edited by A. G. van der Valk, pp. 204227, Iowa State Univ. Press, Ames.
  • Burdidge, D. J., and J. Homstead (1994), Fluxes of dissolved organic carbon from Chesapeake Bay sediments, Geochim. Cosmochim. Acta, 16, 34073424.
  • Classen, H. C., and D. R. Helm (1996), Estimates of evapotranspiration or effective moisture in Rocky Mountain watersheds from chloride ion concentrations in stream base flow, Water Resour. Res., 32, 363372.
  • Covich, A. P., S. C. Fritz, P. J. Lamb, R. D. Marzolf, W. J. Matthews, K. A. Poiani, E. E. Prepas, M. B. Richman, and T. C. Winter (1997), Potential effects of climate change on aquatic ecosystems of the Great Plains of North America, Hydrol. Processes, 11, 9931021.
  • Curtis, P. J. (1998), Climatic and hydrologic control of DOM concentration and quality in lakes, in Aquatic Humic Substances: Ecology and Biogeochemistry, edited by D. O. Hessen, and L. J. Tranvik, pp. 93105, Springer, New York.
  • Curtis, P. J., and H. E. Adams (1995), Dissolved organic matter quantity and quality from freshwater and saltwater lakes in Alberta, Biogeochemistry, 30, 5976.
  • Curtis, P. J., and E. E. Prepas (1993), Trends of dissolved organic carbon (DOC) concentrations from freshwater to salt water, Verh. Int. Verein. Limnol., 25, 298301.
  • Driver, E. A., and D. G. Peden (1977), The chemistry of surface waters in prairie ponds, Hydrobiologia, 53, 3348.
  • Environment Canada (1992), Analytical Methods Manual, Inland Waters Dir., Water Qual. Branch, Burlington, Ont., Canada.
  • Eugster, H. P., and B. F. Jones (1979), Behavior of major solutes during closed basin brine evolution, Am. J. Sci., 279, 609631.
  • Flicken, J. H. (1967), Winter loss and spring recovery of dissolved solids in two prairie pothole ponds in North Dakota, U.S. Geol. Surv. Prof. Pap., 575-C.
  • Fritz, S. (1996), Paleolimnological records of climate change in North America, Limnol. Oceanogr., 41, 882889.
  • Granéli, W. M., M. Lindell, and L. Tranvik (1996), Photooxidative production of dissolved inorganic carbon in lakes of different humic content, Limnol. Oceanogr., 41, 698706.
  • Hayashi, M. (1996), Surface-subsurface transport cycle of chloride induced by wetland-focused groundwater recharge, Ph.D. thesis, 140 pp. Univ. of Waterloo, Waterloo, Ont. Canada.
  • Hayashi, M., and G. van der Kamp (2000), Simple equations to represent the volume-area-depth relations of shallow wetlands in small topographic depressions, J. Hydrol., 237, 7485.
  • Hayashi, M., G. van der Kamp, and D. L. Rudolph (1998a), Water and solute transfer between a prairie wetland and adjacent uplands: 1. Water balance, J. Hydrol., 207, 4255.
  • Hayashi, M., G. van der Kamp, and D. L. Rudolph (1998b), Water and solute transfer between a prairie wetland and adjacent uplands: 2. Chloride cycle, J. Hydrol., 207, 5667.
  • Hessen, D. O., E. J. Gjessing, J. Knulst, and E. Fjeld (1997), TOC fluctuations in a humic lake as related to catchment acidification, season and climate, Biogeochemistry, 36, 139151.
  • Hinton, M. J., S. L. Schiff, and M. C. English (1997), The significance of storms for the concentration and export of dissolved organic carbon from two Precambrian Shield catchments, Biogeochemistry, 36, 6788.
  • Jokic, A., J. N. Cutler, E. Ponomarenko, G. van der Kamp, and D. W. Anderson (2003), Organic carbon and sulfur compounds in wetland soils: Insights on structure and transformation processes using K-edge XANES and NMR spectroscopy, Geochim. Cosmochim. Acta, 67, 25852597.
  • LaBaugh, J. W. (1986), Limnological characteristics of selected lakes in the Nebraska Sandhills U.S.A., and their relation to chemical characteristics of adjacent groundwater, J. Hydrol., 86, 279298.
  • LaBaugh, J. W., T. C. Winter, G. A. Swanson, D. O. Rosenberry, R. D. Nelson, and N. H. Euliss Jr. (1996), Changes in atmospheric circulation patterns affect mid-continent wetlands sensitive to climate, Limnol. Oceanogr., 41, 864870.
  • Lean, D. R. S. (1998), Attenuation of solar radiation in humic waters, in Aquatic Humic Substances: Ecology and Biogeochemistry, edited by D. O. Hessen, and L. J. Tranvik, pp. 109124, Springer, New York.
  • Lindeman, R. (1942), The trophic dynamic aspect of ecology, Ecology, 23, 399418.
  • Mann, C. J., and R. G. Wetzel (1995), Dissolved organic carbon and its utilization in a riverine wetland ecosystem, Biogeochemistry, 31, 95120.
  • McKnight, D. M., G. R. Aiken, and R. L. Smith (1991), Aquatic fulvic acids in microbially based ecosystems: Results from two desert lakes in Antarctica, Limnol. Oceanogr., 36, 9981006.
  • McKnight, D. M., E. W. Boyer, P. K. Westerhoff, P. T. Doran, T. Kulbe, and D. T. Andersen (2001), Spectrofluorometric characterization of dissolved organic matter for indication of precursor organic material and aromaticity, Limnol. Oceanogr., 46, 3848.
  • Millar, J. (1971), Shoreline area ratio as a factor in rate of water loss from small sloughs, J. Hydrol., 14, 259284.
  • Mitsch, W. J., and J. G. Gosselink (Eds.) (1993), Wetlands, 2nd ed., 722 pp., Van Nostrand Reinhold, Hoboken, N. J.,
  • Molot, L. A., and P. J. Dillon (1997), Color-mass balances and color-dissolved organic carbon relationships in lakes and streams in central Ontario, Can. J. Fish. Aquat. Sci., 54, 27892795.
  • Mopper, K., and D. J. Kieber (2000), Marine photochemistry and its impact on carbon cycling, in The Effects of UV Radiation in the Marine Environment, edited by J. de Mora, S. Demers, and M. Vernet, pp. 101129, Cambridge Univ. Press, New York.
  • Murkin, H. (1989), The basis for food chains in prairie wetlands, in Northern Prairie Wetlands, edited by A. G. van der Valk, pp. 316338, Iowa State Univ. Press, Ames.
  • Orem, W. H., and P. G. Hatcher (1987), Solid state 13C NMR studies of dissolved organic matter in pore waters from different depositional environments, Org. Geochem., 11, 7382.
  • Osburn, C. L., D. P. Morris, K. A. Thorn, and R. E. Moeller (2001), Chemical and optical changes in freshwater dissolved organic matter exposed to solar radiation, Biogeochemistry, 54, 252278.
  • Peterson, G. S., L. B. Johnson, R. P. Axler, and S. A. Diamond (2002), Assessment of the risk of solar ultraviolet radiation to amphibians: I. In situ characterization of exposure in amphibian habitats, Environ. Sci. Technol., 36, 28592865.
  • Poiani, K., and W. C. Johnson (1991), Global warming and prairie wetlands: Potential consequences for waterfowl habitat, Bioscience, 41, 611618.
  • Poiani, K. A., W. C. Johnson, G. A. Swanson, and T. C. Winter (1996), Climate change and northern prairie wetlands: Simulations of long term dynamics, Limnol. Oceanogr., 41, 871881.
  • Reche, I., and M. L. Pace (2002), Linking dynamics of dissolved organic carbon in a forested lake with environmental factors, Biogeochemistry, 61, 2126.
  • Schiff, S. L., R. Aravena, S. E. Trumbore, and P. J. Dillon (1990), Dissolved organic carbon cycling in forested watersheds: A carbon isotope approach, Water Resour. Res., 26, 29492957.
  • Schiff, S. L., R. Aravena, E. Mewhinney, R. Elgood, B. Warner, P. Dillon, and S. Trumbore (1998), Precambrian shield wetlands: Hydrologic control of the sources and export of dissolved organic matter, Clim. Change, 40, 167188.
  • Schindler, D. W. (2001), The cumulative effects of climate warming and other human stresses on Canadian freshwaters in the new millennium, Can. J. Fish. Aquat. Sci., 58, 1829.
  • Schindler, D. W., S. E. Bayley, P. J. Curtis, B. Parker, M. P. Stainton, and C. A. Kelly (1992), Natural and man made factors affecting the abundance and cycling of dissolved organic substances in Precambrian shield lakes, Hydrobiologia, 229, 121.
  • Schindler, D. W., P. J. Curtis, S. E. Bayley, B. R. Parker, K. G. Beaty, and M. P. Stainton (1997), Climate induced changes in the dissolved organic carbon budgets of boreal lakes, Biogeochemistry, 36, 928.
  • Søndergaard, M., M. B. Riemann, and N. O. G. Jørgensen (1985), Extracellular organic carbon (EOC) released by phytoplankton and bacterial production, Oikos, 45, 323332.
  • Su, M., W. J. Stolte, and G. van der Kamp (2000), Modeling Canadian prairie wetland hydrology using a semi-distributed streamflow model, Hydrol. Processes, 14, 24052422.
  • Tiner, R. W. (2003), Geographically isolated wetlands of the United States, Wetlands, 23, 494516.
  • Vadeboncoeur, Y., M. J. Vander Zanden, and D. M. Lodge (2002), Putting the lake back together: Reintegrating benthic pathways into lake food web models, BioScience, 52, 4454.
  • Valentine, R. L., and R. G. Zepp (1993), Formation of carbon monoxide from the photodegradation of terrestrial dissolved organic matter, Environ. Sci. Technol., 27, 409412.
  • Waiser, M. J. (2001), The effect of solar radiation on the microbial ecology and biogeochemistry of prairie wetlands, Ph.D. thesis, 263 pp., Napier Univ., Edinburgh, Scotland.
  • Waiser, M. J., and R. D. Robarts (1997), Impacts of a herbicide and fertilizers on the microbial community of a saline prairie lake, Can. J. Fish. Aquat. Sci., 54, 320329.
  • Waiser, M. J., and R. D. Robarts (2004a), Net heterotrophy in productive prairie wetlands with high DOC concentrations, Aquat. Microb. Ecol., 34, 279290.
  • Waiser, M. J., and R. D. Robarts (2004b), Photodegradation of DOC in a shallow prairie wetland, Biogeochemistry, 69, 263284.
  • Webster, K. E., T. R. Kratz, C. J. Bowser, and J. J. Magnuson (1996), The influence of landscape position on lake chemical responses to drought in northern Wisconsin, Limnol. Oceanogr., 41, 977984.
  • Wetzel, R. G., P. G. Hatcher, and T. S. Bianchi (1995), Natural photolysis by ultraviolet radiation of recalcitrant dissolved organic matter to simple substrates for bacterial metabolism, Limnol. Oceanogr., 40, 13691380.
  • Williamson, C. E., D. P. Morris, M. L. Pace, and O. G. Olson (1999), Dissolved organic carbon and nutrients as regulators of lake ecosystems: Resurrection of a more integrated paradigm, Limnol. Oceanogr., 44, 785803.
  • Winter, T. C., D. O. Rosenberry, D. C. Buso, and D. A. Merk (2001), Water source to four U.S. wetlands: Implications for wetland management, Wetlands, 21, 462473.
  • Woo, M.-K., and R. D. Rowsell (1993), Hydrology of a prairie slough, J. Hydrol., 147, 175207.
  • Xenopoulos, M., D. M. Lodge, J. Frentress, T. A. Kreps, S. D. Bridgham, E. Grossman, and C. J. Jackson (2003), Regional comparisons of watershed determinants of dissolved organic carbon in temperate lakes from the Upper Great Lakes region and selected regions globally, Limnol. Oceanogr., 48, 23212334.