Discussion

Water Quality

Water-quality in the Wood River Wetland was dynamic and related to seasonal water availability and temperature, major controls of biological activity, and continued release of nutrients and DOC. Potential sources of N and P include leaching from previously oxidized peat soils, the decay of recently grown vegetation, waterfowl feces (Manny and others, 1994), and remnant cattle waste. Evaporation and evapotranspiration water losses further increased nutrient and mineral concentrations in surface waters.

Pelican at the Wood River Wetland. Photograph taken by John Duff, U.S. Geological Survey, April 2005.

Concentrations of nutrients, DOC, and SC were substantially higher than in other area wetlands, including the Lower Klamath National Wildlife Refuge (Mayer, 2005) and restored wetlands surrounding Upper Klamath Lake (for example, Caledonia Marsh [MacLaren and Geiger, 2001] and Agency Lake Ranch [Damion Ciotti, Bureau of Reclamation, oral commun., 2007]). SRP and NH₄ concentrations were particularly high in the South Unit. Benthic nutrient-flux experiments indicated that N and P were released from the soils (figs. 17 and 18).

Despite the high NH₄ concentrations, concerns about toxicity to aquatic life are lessened because at most times, a relatively insignificant amount of the more toxic, un-ionized free ammonia (NH₃) was present. The median ammonium concentrations in the North and South Units were 380 and 510 µg-N/L, respectively, which would produce NH₃ concentrations less than 30 µg-N/L. The maximum surface-water concentration of total ammonia was 5,840 µg-N/L (in the South Unit during August 2004), which would produce about 390 µg-N/L un-ionized ammonia (with concurrent temperature of 25°C and pH of 8.1 pH units). This maximum ammonia level might stress fish, including endangered Lost River and shortnose sucker fish, which have respective 96-hour LC₅₀ (median lethal concentration for one-half the test population) values of 780 and 530 µg/L NH₃ (Saiki and others, 1999). Although no fish were observed in the wetland during the field sampling—possibly because the inlet structure includes a fish screen—water with elevated NH₃ discharged to Wood River or Sevenmile Canal might affect fish in the immediate vicinity of the pump outlet station, depending on the total ammonia concentration, water temperature, pH, and streamflow in the receiving stream.

One logical source of NO₃, NH₄, and SRP during spring and early summer is sediment flux. Increasing nutrient to SC ratios coincided with increased water temperature, which stimulated bacterial activity and decomposition of peat soils. Although this hypothesis was supported by the chamber experiments, other potential sources of inorganic nutrients, including waterfowl feces or N-fixation, may be important at times. The decrease in DIN (NO₃ + NH₄) and phosphorus (SRP) concentrations during mid-August through September and the decrease in the DIN to SC and SRP to SC ratios indicate the presence of sinks such as plant uptake, rapid nitrification-denitrification for N, and sorption for P.

Legacy Effect of Wetland Drainage and Land Management

Intact wetland soils typically are inundated or water saturated, limiting air penetration into deep layers where anaerobic conditions limit decomposition and allow peat to accumulate (Cameron and others, 1989). Draining lowers the water table, increases aeration, and enhances soil microbial activity, which accelerates decomposition of organic matter (Maciak, 1972; Lee and Manoch, 1974; Lévesque and Mathur, 1979; Mathur and Farnham, 1985; Efimov and Lunina, 1988; Laine and others, 1992). Because decomposition is most efficient in aerobic environments, decay rates are typically highest in aerated, unsaturated soils, and lowest in deep, water-saturated anoxic zones. Decomposition rates generally decrease exponentially as oxygen becomes depleted through progressively deeper layers.

Historical land management at the former Wood River Ranch included 50 years of wetland draining for beef production, which caused peat decomposition resulting in decreased peat thickness and mass, and promoted C, N, and P loss. Cattle waste also accumulated. Snyder and Morace (1997) determined that the extent of peat decomposition in drained wetlands surrounding Upper Klamath and Agency Lakes is related to the time since drainage, among other factors. In many places, soils have oxidized and lost substantial bulk, resulting in compaction and land subsidence as much as 13 ft (Snyder and Morace, 1997, p. 26). Loss of peat thickness at the Wood River Wetland may be as high as 4–4.4 ft (table 11). Wetland soils in the Twitchell Island restoration wetland in the Sacramento-San Joaquin River Delta, California, show land subsidence as much as 6 m (Rojstaczer and others, 1991; Fleck and others, 2004). In the Wood River Wetland, similar mineralization has released large quantities of carbon, nitrogen, and phosphorus, with respective losses averaging 36, 31, and 18 percent (table 11).

Other studies at the newly restored Twitchell Island have demonstrated that permanently flooded conditions (25–55 cm depth) resulted in an average plant-material accumulation rate of 4 cm/yr in planted test plots (Miller and others, 2008). Such high rates of accumulation in the Sacramento-San Joaquin River Delta are producing land subsidence reversal, storing carbon and other nutrients and minerals in partly decomposed young peat that resemble the older peat in terms of composition and bulk weight. The highest rates of 7–9 cm/yr were in plots with water depth of 55 cm in an area receiving a relatively low degree of flushing (Miller and others, 2008), which contributed to low decomposition rates and high overall accumulation of vegetation. Vegetative growth at the Wood River Wetland is occurring in many habitats less than 55 cm deep, whereas the open-water areas (canals and deep water habitats) are largely unvegetated. Because of the similarities in plant community composition between the Twitchell Island and Klamath Basin wetlands, hydrologic restoration of Wood River Wetland may produce results similar to those at the restoration of Twitchell Island. Although potential growth rates of wetland vegetation may be lower for upper Klamath River basin wetlands because of lower temperatures, light availability may be greater during summer months due to the higher elevation (about 4,000 ft) of the basin. As Miller and others (2008) point out, reversal of land subsidence through hydrologic restoration is an important strategy for reducing carbon emissions that contribute to global warming for two reasons: (1) a large amount of carbon can be stored (1 kg of carbon per square meter per year in the Twitchell Island wetlands) and (2) losses are minimized due to development of anaerobic conditions that limit further loss of carbon dioxide to the atmosphere from oxidation of peat soil.

Wood River Wetland peat beds were often buried beneath layers of silt, ash, pumice, or clay (Snyder and Morace, 1997) that further restricted oxygen penetration. In some areas, seasonally fluctuating water tables may infuse oxygenated surface water to peat layers, promoting aerobic decomposition. Permeable pumice layers also may facilitate the horizontal transport of oxygenated water, also enhancing peat decomposition. Oxygen availability in the deep peat horizons probably is limited, however, because of the high concentrations of ammonium and low nitrate in the deep wells, particularly in the North Unit (fig. 8).

Biogeochemical cycling and (or) diffusive processes have enhanced P content in sediments relative to C or N. The Redfield ratio (C:N:P of 106:16:1 [by atom]) describes their relative proportions in healthy plant tissue. When compared with the Redfield ratios, C losses from the Wood River Wetland were 2–7 times greater than expected relative to losses for N and 2–451 times greater than expected for P (table 12). This disproportionate C loss resulted from decomposition and aerobic respiration of the organic-rich peat soils, producing carbon dioxide (CO₂). Losses of N also were disproportionate compared with P because N is lost in gaseous form (NO, N₂O, and N₂) during denitrification, whereas P, which has no gaseous phase, is continuously recycled by biotic uptake/release or physiochemical sorption-desorption. Two cores (WRR-08, in the North Unit, and WRR-02, in the South Unit) showed significantly greater P-loss (relative to N), indicated by the low N:P Redfield ratio factors for these two cores (0.2–0.3) (table 12). These cores were collected from the deeper inundated areas of the wetlands (see plate 1, Snyder and Morace [1997]), on the north side of each dike. The large P loss might have been due to greater anaerobic release in these deeper areas.

Biogeochemical Processes and Nutrient Transformation

Wetlands host diverse microbial communities whose metabolism drives carbon and nutrient cycles. The biogeochemical reactions are complex, and are driven by the presence or absence of DO. Reduction-oxidation (redox) conditions dictate the position, composition, and activity of microbes (Cameron, 1970; Cameron and others, 1989). Alternating bands of dark-colored decomposed soil and unmineralized light soil in Wood River Wetland cores may be due to redox gradients. Abiotic processes, including sorption and sequestering of inorganic phosphorus to iron and manganese compounds, also may be occurring in the sediment. In the presence of DO, phosphorus can be bound to iron and manganese, whereas anaerobic conditions may cause phosphorus to be released (fig. 14).

The dried and oxygenated surface soils at the Wood River Wetland have been decomposed by bacteria that metabolize organic matter into CO₂, producing carbonic, humic, and fulvic acids that lower pH. Median pH values in the shallow wells in the North and South Units (6.2 to 6.5 units) were less than in the deep wells (7.2 to 7.6 units) (table 6), which is consistent with this theory. Lower NH₄ and SRP concentrations in the shallow wells may indicate preferential decomposition and loss of nutrients from surficial soil layers by flushing. The high SC and ammonium concentrations in the North Unit deep well (PZ-NU3) (fig. 8) might have been due to a lack of flushing of the deeper soil layers by precipitation or surface-water irrigation. In contrast, the deep water from the South Unit (PZ-SU3) had lower SC than the intermediate or shallow depth wells. Greater flushing of the deeper layer compared with the shallow wells (PZ-SU1 and PZ-SU2) might explain this pattern, especially given the artesian nature of the deeper piezometer well (tables 1 and 2). The influence of deep artesian water on PZ-SU3 appeared for several water quality parameters, including SC, pH, and SRP (fig. 8). Less flushing in PZ-SU3 is consistent with the lower concentrations of DO and higher water temperatures compared with the deeper artesian wells (fig. 8). The differences in DO between the wells, and closer proximity to surficial peat soils for PZ-SU3, may allow ammonia to accumulate (from the breakdown of peat) if DO concentrations limit the conversion of ammonium to nitrate by nitrifying bacteria.

Decomposition of peat releases large quantities of DOC, amino acids, and other forms of dissolved organic nitrogen (DKN), which is mineralized to NH₄ (ammonification) by bacteria. This is one potential source of NH₄ in the wetland. In the presence of oxygen, NH₄ can be oxidized to NO₂ and NO₃ by nitrifying bacteria Nitrosomonas and Nitrobacter, although the relatively low DO levels in ground water (0.8–1.9 mg/L) may suppress this process below surficial sediments, resulting in relatively low NO₃ in ground-water discharge. Low NO₃ concentrations also may result from denitrification. Although denitrification rates at ambient NO₃ levels were not especially high in the chamber experiments (fig. 18), they were stimulated by adding NO₃. This indicates that coupled nitrification-denitrification could maintain the low NO₃ in the wetland throughout the year. Denitrification is the primary removal mechanism for nitrogen in many wetlands (Kadlec and Knight, 1996), with rates determined by factors such as temperature, organic carbon availability, nitrate loading, and anoxic conditions (Bachard and Horne, 2000).

Wetlands often contain water with little or no oxygen, particularly during summer, when biological oxygen demand from bacteria is high. DO gradients form in the water column, producing redox boundaries near or in the sediments where anaerobic microbes generate energy by oxidizing C, N, S, and other elements. During anaerobic metabolism, bacteria produce NO, N₂O, N₂, hydrogen sulfide (H₂S), and methane (CH₄), among other gases. Hydrogen sulfide gas odors were occasionally detected in the mucky soils in the wetland and in water samples collected from the deep North Unit well (PZ-NU3). Anaerobic conditions can also stimulate microbial reduction of nitrate to ammonium in a process called dissimilatory nitrate reduction (DNRA), which is carried out by Clostridia and certain sulfate reducing bacteria (Atlas and Bartha, 1993). The increase in NH₄ during the chamber experiments with the simultaneous addition of nitrate and glucose (fig. 18) indicates that DNRA could be occurring in the wetland sediments.