H53F-01 INVITED 13:40h
The Relative Effects of Hydrology, Ecology, and Climate on Temporal Trends and Spatial Patterns of Stream Nitrate Concentrations in the Catskill Mountains, New York, USA
The Catskill Mountains of New York receive 10 to 15 kg ha$^{-1}$ yr$^{-1}$ of nitrogen (N) in atmospheric deposition, among the highest rates in eastern North America. Consequently, streams in the Catskills have relatively high nitrate (NO$_{3}$$^{-}$) concentrations (mean of stream surveys = 20 to 30 $\mu$mol L$^{-1}$ at baseflow), which contribute to chronic and episodic stream acidification. Stream chemistry monitoring in the 1980s showed an increasing trend in NO$_{3}$$^{-}$ concentrations that was attributed to N saturation, a condition whereby continued high rates of atmospheric N deposition in combination with a maturing forest result in gradually decreasing ecosystem N retention. This increasing trend reversed itself during the early 1990s, a pattern that was partially attributed to an intense region-wide soil freezing event in December 1989. Stream NO$_{3}$$^{-}$ concentrations remained relatively low at long-term monitoring sites from 1992-03 despite relatively constant atmospheric N deposition rates from the 1980s through 2003, suggesting that a simple interpretation of the N saturation model is not valid on decadal time scales. The discovery of groundwater seeps with relatively high NO$_{3}$$^{-}$ concentrations in the 1990s led to a hypothesis that the presence of seeps in the Catskills may be controlling spatial variability in stream NO$_{3}$$^{-}$ concentrations, which range from near 0 to about 50 $\mu$mol L$^{-1}$ at baseflow. Subsequent research indicated these variations in stream NO$_{3}$$^{-}$ concentrations are likely controlled by variations in tree species dominance. Nitrate concentrations in drainage waters are highest in stands of sugar maple and yellow birch and lowest in red oak and hemlock stands. One study, however, showed that NO$_{3}$$^{-}$ concentrations in shallow groundwater were correlated with a topographic index that is a surrogate for soil moisture suggesting that covariance of tree species and soil moisture may amplify the apparent differences in NO$_{3}$$^{-}$ concentrations previously attributed solely to differences in tree species. A conceptual model of controls on spatial variation of stream NO$_{3}$$^{-}$ concentrations in the Catskills indicates that these values are established in the shallow soil at the time of recharge as a result of differences in relative nitrification rates caused by differences in tree species effects on the relative recalcitrance and C/N ratio of litter. Groundwater seeps serve mainly as conduits that drain a deeper reservoir that can provide seasonally higher NO$_{3}$$^{-}$ concentrations during the summer growing season. In spring 2004, stream NO$_{3}$$^{-}$ concentrations reached 125 $\mu$mol L$^{-1}$ in Biscuit Brook, among the highest values ever recorded at this stream during 21 years of monitoring. A preliminary analysis finds no apparent disturbance or extreme climatic event that may have caused such high concentrations. This result suggests that much remains to be learned about the controls on temporal patterns and trends in stream NO$_{3}$$^{-}$ concentrations in this region. Whereas a viable conceptual model explains most spatial variability in NO$_{3}$$^{-}$ concentrations among streams, a simple interpretation of the N saturation model that would indicate a long-term trend of increasing stream NO$_{3}$$^{-}$ concentrations does not explain the observed temporal patterns. A more complex model that includes the role of climate, hydrology, and disturbances caused by insects and tree disease may be necessary to adequately predict long-term trends.
H53F-02 13:55h
The Effect of Ozone on Ecosystem Processes Using Improved Hydrological Cycling Within a Biogeochemical Model
Exposure of plants to ozone reduces their photosynthetic capacity and results in less productivity and carbon sequestration. We have recently used the Terrestrial Ecosystem Model (TEM) to show the magnitude and extent of this effect historically for the U.S. and the world, with the maximum ozone damage occurring in hot-spot regions like the eastern U.S., Europe, and eastern China. In this study we have improved the hydrological model in order to provide a better representation of the mean stomatal conductance of the canopy, which determines the amount of ozone uptake that occurs. We are using a high-resolution ecosystem physiology model (Soil-Plant-Atmosphere (SPA) model) to aggregate stomatal conductance up to the coarse TEM temporal and spatial resolution. Experimental evidence indicates that ozone exposure results in lower stomatal conductance, thus limiting further ozone exposure and water loss. By replacing the current formulation of transpiration in TEM with a Penmon-Monteith approach using the computed canopy conductance, we are able to incorporate all the major links between ozone, stomatal conductance, and the hydrological cycle. The carbon and nitrogen stocks and fluxes are then recalibrated to ensure the correct values at each calibration site. Preliminary equilibrium results for Harvard Forest, MA show a reduction in the magnitude of the ozone effect from a 5.2% decrease in net primary production (NPP) to a 4.0% decrease. The difference is largely due to lower stomatal conductances that produce lower evapotranspiration (EET) rates closer to the observed rates. From 1996-1999, flux tower data show a mean July EET of 84 mm, while the model simulates an EET of 98 mm with a resulting mean stomatal conductance of 1.7 mm/s for July. Indirect effects of ozone on stomatal conductance are small because the reduced conductance, while reducing ozone uptake, also reduces CO2 uptake. We are now validating the model at several other well-studied temperate deciduous forest sites before parameterizing and validating it at other biomes throughout the world.
H53F-03 14:10h
A two-layer model to simulate seasonal variations in surface water chemistry draining a northern forest watershed
For many forested headwaters in the northeastern U.S., stream flow is a mixture of water derived from different soil layers. Therefore as a result of the seasonal fluctuations in hydrologic flow paths, considerable seasonal variation in surface water chemistry is evident in these headwaters. Especially during high flow season, flow tends to be routed through the upper soil layers, which are usually characterized by higher concentrations of NO3-, DOC, naturally occurring organic anions and Al. High streamflow coupled with shallow flow path results in dilution in base cations, increased leaching of organic acids, NO3- and Al to stream water, and depression in pH and ANC. While during base flow season, stream discharge is primarily generated from lower soil layers with higher base cation concentrations and lower concentrations of organic anions and NO3-. Previously, an integrated biogeochemical model (PnET-BGC) based on single soil-layer formulation was found to be inadequate to simulate these seasonal variations. In order to better simulate the seasonal variations in stream water chemistry draining acid-sensitive forest watersheds, a two soil-layer version of the model (PnET-BGC2) was formulated and applied to a northern forest ecosystem, the Hubbard Brook Experimental Forest (HBEF). End member mixing analysis was used to better understand hydrologic flowpaths contributing to temporal patterns in stream chemistry and to parameterize the model. The resulting two-layer model is generally able to reproduce the seasonal variations in surface water runoff, concentrations of base cations, SO42-, NO3-, pH and ANC.
H53F-04 INVITED 14:25h
Ecosystem Restoration and Denitrification in a Coastal Stream
Substantial removal of nitrate can occur in the riparian and hyporheic zones of small streams. Urbanization may lead to impairment of processes leading to this nitrate removal, however, by altering aboveground sources of carbon, and routing water to deeper flow paths away from denitrification "hot spots" in surface soil horizons. We have been quantifying denitrification in riparian and hyporheic zones of restored and unrestored reaches of an urban stream within the Chesapeake Bay watershed using an in situ 15N tracer technique. Recovery of conservative SF6 tracer showed that this "push-pull" technique could be applied to a wide range of locations using an incubation time of 4 hours. Rates of denitrification in hyporheic wells were higher in the restored reach of the stream as compared to the unrestored reach, which had less hydrologic connection to the stream channel. In the restored reach, sandy riparian areas, which were frequently flooded, showed significantly higher rates of denitrification as compared to less flooded areas. Maximum rates of denitrification in the restored reach ranged from 100-112 micrograms N/kg soil/day. These high rates of denitrification were consistent with significantly lower concentrations of nitrate in hyporheic water in the restored reach relative to the unrestored reach throughout different seasons. Mean concentrations of dissolved organic carbon also differed between the two reaches and were higher in hyporheic and stream water from the restored reach. Maintaining hydrologic connectivity between streams and adjacent ground water may be important when decreasing the amount of N transported to Chesapeake Bay. Considerable denitrification may occur in urban riparian and hyporheic zones following stream restoration.
H53F-05 14:40h
A Novel Indicator of Ecosystem N Status: DIN to DON Ratio in Riverine Waters
We propose that the ratio of the annual flux of DIN to DON in streamwaters provides a robust and sensitive method of determining the N-status of ecosystems from a variety of biomes. Effects of N deposition are generally decoupled from N deposition because of the large variety of N species found in air, deposition, watersheds, and surface waters, as well as the myriad of pathways through which N can be cycled in terrestrial and aquatic ecosystems Moreover, the amount of N in biomass and soils and resident time within these reservoirs varies widely among biomes. Consequently, developing robust indicators of the N-status of ecosystems is difficult. Our results from many biomes suggest that N-limited ecosystems have DIN:DON ratios less than 0.5 and that ecosystems where N is no longer limiting have DIN:DON ratios greater than 2.0. The DIN:DON ratio in annual riverine yields is independent of ecosystem N storage, rates of N cycling, magnitude of N yield in surface waters, and hence a robust indicator of the N-status of ecosystems. We hypothesize that DON export from terrestrial ecosystems is controlled primarily by the standing stock of C and N in soils and hence will change only slowly in response to anthropogenic additions of N. In contrast, atmospheric deposition of N stimulates net nitrification and nitrate export in soil solution and stream waters. Therefore export of dissolved inorganic N responds directly and quickly to increases in anthropogenic deposition of N.
H53F-06 INVITED 14:55h
Nitrogen removal in floodplains through denitrification: results from an isotope field experiment
Biogeochemical transformations of nitrogen within lotic systems influence the speciation and nitrogen load to downstream water bodies. Within a stream network, redox gradients at the interface between surface and subsurface water result in regions of potential denitrification. During periods of high-flow, water and nutrients from the main channel of a stream are routed onto the adjacent floodplain resulting in greater solute interactions with particle and microbial surfaces. We have found that over 80% of the nitrogen load going onto a floodplain in Northeastern Louisiana is removed from the water column. A $^{15}$N-enriched nitrate tracer experiment within the floodplain demonstrated that a large portion of the nitrate is denitrified between the surface water and the top 3cm of soil. The overall nitrate removal within a given floodplain is dependent on the hydrologic residence time, and the field results highlight that for low to medium flood events the residence time is great enough for denitrification to remove the majority of the nitrate pool. Our results demonstrate that active floodplains are biogeochemical hotspots within stream networks, and suggest that floodplains within the Southeastern USA reduce the nitrate load to the coastal seas.
H53F-07 15:10h
Ecohydrological Linkage Between Surface-Derived Dissolved Organic Matter and Subsurface Carrying Capacity for Unattached Subsurface Bacterial Populations: Implications for Biodegradation in Contaminated Groundwater Environments
Degradation by aquifer microbial populations of dissolved organic matter (DOM) transported from the surface environment exerts strong controls on the concentration, composition, and reactivity of DOM as it is transported in ground water. In turn, the quantity of readily degradable DOM may be a major control of the carrying capacity for bacteria in groundwater and, consequently, would affect the fate of lower levels of organic contaminants being advected through the same sediments. For a 5-km long plume of groundwater organic contaminants in a sandy aquifer at Cape Cod, Massachusetts, the partitioning of bacteria between solution and grain surfaces increased with decreasing distance downgradient from the source. In general, the unattached bacterial populations were particularly sensitive to the level of DOM present. Approximately 2/3 of the spatial variability for unattached bacteria within the upgradient 3 km of the contaminant plume can be explained statistically simply by the abundance of the more degradable (non alkyl benzene surfactant [ABS]) DOM, even though predation by groundwater nanoflagellates (protists) and sorptive-filtration of unattached bacteria both vary spatially. However, distance downgradient from the contaminant source, which directly correlates with both the age and abundance of the DOM, could explain ~98% of the variability of unattached bacteria along the longitudinal axis of the plume. Down-well incubations (USGS well site F513) were performed using membrane (100K Dalton cutoff) chambers (initially seeded with undifferentiated, laboratory-grown aquifer bacteria at abundances of 1 x 10$^{5}$ or of 1 x 10$^{7}$ bacteria/mL) that come to chemical equilibrium with surrounding groundwater within ~1 day. Following a 7-day down-well incubation, numbers of unattached bacteria within the chambers converged on a final abundance value of 2.0 $\pm$ 0.6 x 10$^{6}$ bacteria per mL, consistent with the in-situ bacterial abundance within that part of the plume (typically between 1.0 x 10$^{6}$ and 2.0 x 10$^{6}$/mL). Incubation for a longer (21 day) period resulted in final bacterial abundances that were not statistically different than those observed for the shorter (7-day) incubations. Results from the Cape Cod study suggests the carrying capacity of the system for unattached groundwater bacteria that are advecting downgradient with specific organic contaminants may be determined by the more degradable fraction of the DOM.
H53F-08 15:25h
Geological Control of Physical and Chemical Hydrology in Vernal Pools, Central Valley, California
Vernal pools are seasonally-inundated depressional wetlands with perched water tables due to low-permeability deposits such as bedrock, mudflows, clay-rich soils, or hardpans. Vernal pools on these deposits are morphologically similar and, therefore, are treated similarly by land managers. However, these deposits result in distinct hydrological conditions. In this study, we compared the hydrology of vernal pools formed on clay-rich soils and hardpan soils, the two most common types of vernal pools in the Central Valley, California. The vernal pools on clay-rich soils were formed in fine-textured alluvial deposits that originated from sandstones, siltstones, and mudstones of marine origin. The upper 0.1 m of soil had a clay loam texture and was underlain by a saline-alkaline claypan (Aquic Natrixeralfs). The vernal pools on hardpan soils were formed in alluvial deposits that originated from crystalline bedrock. The upper 0.6 m of soil had a gravelly loam texture and was underlain by a silica-cemented duripan (Abruptic Durixeralfs). The vernal pools on clay-rich soils responded rapidly to rainfall. Rainfall perched on the surface and flowed rapidly overland to the vernal pools. There was little soil moisture or groundwater recharge, with soil saturation limited to the upper few cm after months of inundation. Most surface water evaporated, evapoconcentrating vernal pool water and resulting in high electrical conductivity. Sodium concentrations and clay contents were relatively high, resulting in dispersed clays which contributed to high turbidity. The vernal pools on hardpans responded more slowly to rainfall. Rainfall infiltrated in the uplands, perched on the hardpans, and flowed slowly through the vernal pools, discharging at the upgradient ends and recharging at the downgradient ends. Groundwater flow through the vernal pools provided a continuous source of fresh water that limited evapoconcentration, resulting in low electrical conductivity. Sodium concentrations and clay contents were relatively low, resulting in low turbidity. Though morphologically similar, these vernal pools differ hydrologically and, therefore, should be treated differently by land managers.