H51D-1162 0800h
Coupled Simulation of Wetland Hydrology, Nutrient and Vegetation Dynamics
Ecohydrological variations such as altered hydrologic regime, invasion of exotic flora, and nutrient enrichment in the Kissimmee-Okeechobee-Everglades aquatic ecosystem in south Florida have been observed. It is important to study the dynamics of wetland hydrology, nutrient and vegetation communities and their interactions over multiple spatial and temporal scales so that wetland restoration, ecological protection, and best management policy decision-making can be effectively accomplished. Hydrologic models are important tools in these decision-making processes. Hydrological components capable of multi-directional overland flow and lateral groundwater flow simulation within the Java-based, object-oriented framework of the ACRU2000 model were developed to make the existing hydrologic model in ACRU2000 more applicable in the Lake Okeechobee Basin where flat topography, high-water-table and sandy soils define the very unique hydrology. In addition nutrient components capable of multi-directional transport and transformation of nitrogen and phosphorus were modified to make the current nutrient model in ACRU2000 more applicable in the Lake Okeechobee Basin. Observed data in the Lake Okeechobee Basin were used to validate the coupled hydrologic and nutrient cycling model's predictions of the spatial and temporal distribution of flow and nutrient concentrations. The simulated results indicate that the coupled model is capable of simulating nutrient, overland, and lateral groundwater flows over a watershed that incorporates wetlands. Future work will focus on the development of a new wetland vegetation model to be integrated into this coupled hydrological and nutrient model. The new model will then be applied in the Lake Okeechobee Basin to simulate the ecohydrological variations due to the implementation of alternative water and land management practices. (More information regarding ACRU2000 and its modification for use in the Southeastern Coastal Plain can be found at http://www.agen.ufl.edu/~klc/acru2k/.)
http://www.agen.ufl.edu/~klc/acru2k/
H51D-1163 0800h
The Ecohydrology of Degraded and Restored Meadow Systems
Meadow systems play a critical buffering role in the hydrology of the northern Sierra Nevada, Plumas County, CA. They serve as surface and subsurface storage zones that mitigate flood flows resulting from the spring snowmelt. They also provide reservoirs of water that contribute to baseflow and transpiration demands during the dry summer in this semi-arid environment. Unfortunately, land use practices such as grazing, logging, and railroad construction have caused stream incision and subsequent draining of meadow sediments. This watertable drop causes a succession from wet meadow species such as sedges and rushes to dryland grasses and sagebrush. Based on water level records at 22 piezometers, we present the critical depth to water hydrograph that determines whether the xeric or mesic/hydrophytic vegetation communities will dominate a site. We call this the "vegetation-threshold hydrograph" (VTH); sites with water levels above those of the VTH support meadow species whereas sites with water levels below those of the VTH are dominated by sagebrush. The most important feature of the VTH is that water levels are very near, or above, the land surface during April and May. If the water table is within 0.5m of the surface during this period of growth initiation, the very high moisture contents in the root zone will cause water-logging and mortality of xeric vegetation. Additionally, the water table must remain within 1m of the surface through June and July, to maintain sufficient soil moisture for the mesic/hydric species to reach full bloom during the annual drought period. Through August and later, the mesic/hydric species begin to senesce and no longer require a shallow watertable. Knowledge of the VTH has potential to improve the success rate of restoration in this environment by providing a minimum target hydrograph. Examples will be used to discuss the hydrologic function of degraded and restored meadow systems.
H51D-1164 0800h
Estuarine Ecohydrology: The Importance of Wetlands in Estuary Robustness
In addressing the interaction of biota, hydrology and sediments in determining estuary functioning, it is important to recognise that human activity has increased the flux of water and sediment (and associated nutrients and pollutants) to the coast, leading to enhanced lake sedimentation, floodplain alluviation, delta progradation, estuary infilling and, in recent times, eutrophication and pollution. These factors contribute to the ephemeral nature of estuaries when viewed on a geological timescale. Recent directions in coastal and river basin management acknowledge the importance of balancing human use and ecosystem function within a large-scale integrated management framework. In addition, the importance of wetlands in regulating water, sediment, pollutant and nutrient transfer has been highlighted. With considerable practical success being achieved in the development of riparian woodlands and the managed realignment of saltmarshes, we highlight the importance of perimarine wetlands as an important buffer zone at the interface between river basin and coastal cell. The benefits of these wetlands as an ecohydrological management tool are considered, and priorities for developing a new realignment continuum are discussed.
H51D-1165 0800h
Ecohydraulics and Estuarine Wetland Rehabilitation
The hydraulics or water flow in wetlands is known to be a key factor influencing ecosystem development in estuarine wetland environments. The relationship is indirect, with the hydraulics of wetlands influencing a host of factors including soil salinity, waterlogging, sediment transport, sediment chemistry, vegetation dispersal and growth and nutrient availability and cycling. The relationship is also not one way, with the hydraulics of wetlands being influenced by plant and animal activity. Understanding these complex interactions is fundamental for the adequate management of estuarine wetlands. Listed as a Wetland of International Importance under the 1971 Ramsar Convention, the Hunter River estuary is regarded as the most significant site for migratory shorebirds in New South Wales, Australia. Over the past 20 years, the number of migratory shorebirds in the estuary has sharply declined from 8,000 to 4,000 approx. Alteration of bird habitat is believed to be one of the reasons for this alarming trend. In 2004 we started a three-year program to investigate the links between hydraulics, sediment, benthic invertebrates, vegetation and migratory shorebird habitat in the estuary. During the first year we have focused on a highly disturbed part of the Hunter estuary wetlands located on Ash Island. The area is one of the major roosting sites in the estuary and is characterized by a complex hydraulic regime due to a restricted tidal interchange with the Hunter River and the presence of infrastructure for the maintenance of power lines (i.e., roads, bridges, culverts). Salt marshes, mudflat and mangroves are the dominant vegetation types. The monitoring program includes measurements of water levels, salinity, discharge, velocity, turbulence, sediment transport and deposition, plant species and density, soil composition and benthic invertebrates coordinated with observations of bird habitat utilization on a number of locations throughout the wetland and for different flow conditions. We present a preliminary analysis of the data aimed at the hydrodynamic and geomorphologic characterization of the different vegetation zones and the resulting habitat properties.
H51D-1166 0800h
Linkages between flood, aquatic organic matter, and food web processes in the Okavango Delta, Botswana
The Okavango Delta of Botswana is a pristine but threatened wetland that ranges in size from 15,000 - 28,000 km2. Previous research has shown that an annual flooding event exerts controls on the quantity and chemical quality of aquatic organic matter (OM) at the Delta scale. In permanently-inundated areas, the perennial water supply maintains low dissolved organic carbon (DOC) concentrations and more microbial sources of OM, as evidenced by low specific UV absorbance and high fluorescence index values. In seasonally-inundated areas, the annual flood causes a pulse of DOC (over 5 mg C/L) and a shift to vegetation-derived DOC, as a result of the inundation of vegetation and soils. Because seasonal floodplains, which encompass 10,000 - 12,000 km2 of the Okavango Delta, become productive grazing areas after the flood and because a dominant portion of fish biomass production takes place in seasonal floodplains, the productivity of these areas is significant for higher trophic levels. Planned water developments upstream of the Delta may shorten/flatten the hydrological pulse and impact the transport and mobilization of organic matter within the Delta. The impacts of reduced flows on extent and duration of flooding have been examined. However, the secondary effects on biological productivity have received less attention. We hypothesize that varying sources of OM, controlled by the hydrologic regime of the Delta, and OM transformations from bacterial and UV degradation exert an influence on floodplain productivity. This study presents results of leaching and photodegradation experiments and observations of changes in algal populations and floodplain standing stock to demonstrate important linkages between biological, ecological, and hydrological processes in the Okavango Delta. Our results support that the DOC that is mobilized by the flood supports heterotrophic microbial populations which, in turn, support the biological productivity of seasonal floodplains.
H51D-1167 0800h
Interactions of Terrestrial and Aquatic Processes Among Lakes of the Northern Highland Lake District, USA
In the Northern Highland Lake District of Wisconsin and Michigan, USA, thousands of connected terrestrial and aquatic systems interact in a complex, dynamic landscape. To understand the flow and storage of carbon and water across this region, we developed a new simulation model linking groundwater, terrestrial processes, and lake processes in a spatially explicit framework. This model integrates ecological processing (on land and in lakes) with physical hydrology (of the land, lakes, and groundwater pools) to estimate, for each watershed, seasonal water cycle and biogeochemical dynamics. Results indicate that fluxes between lakes and watersheds are an important component of hydrologic and carbon budgets, and the model permits the quantification of these influences across a spatial network of connected watersheds. Our understanding of these systems has been greatly enhanced by multi-decade studies in a small subset of the lake district; we show how the results of modeling these processes can be informed by both hydrologic and biogeochemical data collected at varied times and locations. This explicit incorporation of terrestrial and aquatic processes in surface and subsurface connection networks will aid our understanding of the relative roles of on-land, in-lake, and between-lake processes in this lake-rich region.
http://brahms.zoology.wisc.edu/Currentprojects/Fluxes/fluxeslandwater.htm
H51D-1168 0800h
Terrestrial-aquatic Interactions in SE Alaska: Seasonal Stream Response to DOM and Nutrient Inputs From Wetlands and Salmon
Forested and peat-dominated wetlands comprise about 30% of the uplands in Southeastern Alaska. Inputs from these carbon-rich soils presumably have a large influence on the biogeochemistry of streams draining them, the most obvious being the presence of brownwater, high DOM streams in catchments dominated by peatlands. We measured N and DOM concentration and quality within wetland soil pore water and surface water in two stream types: a brownwater stream and a clearwater stream. Surface samples were taken in intermittent tributaries draining wetlands and the perennial streams receiving those inputs. Soil water samples were taken in dominant end member wetland types: forested wetlands, bogs and fens. Seasonal patterns of DOM and forms of nitrogen were tracked to correlate wetland source pool and surface channel patterns. Soil porewater concentrations ranged from 15 to 50 mg C L$^{-1}$. Wetland drainage channels ranged from 10 to 35 mg C L$^{-1}$ and were similar in wetlands draining into both stream types. DOC concentrations varied between 3 and 15 mg C L$^{-1}$ within the brownwater creek, and less than 1 to 5 mg C L$^{-1}$ in the clearwater creek. The close similarity of seasonal and storm-related DOM patterns in wetland drains and streams indicates that, during storm runoff periods, wetland inputs influence DOM within both stream types, but during low-flow periods wetland influences are strong only in the brownwater stream. Another important source of DOM to SE Alaskan streams is decaying salmon carcasses. The summer of 2004 was unusually dry with low flows in the study streams. When salmon were present, DOM exceeded 15-20 mg C L$^{-1}$. DOM quality increased during summer in both stream types. Soil porewater nitrogen concentrations were higher than surface waters. In spring $<$5% of total N in surface waters was in organic forms. During summer the organic fraction increased to 10-15% of total N in the clearwater stream, and up to 50% in the brownwater stream.
H51D-1169 0800h
Soil Hydrologic Response and Nutrient Movement in Three Small Tropical Catchments
The movement of water over and through soils by storm-generated flowpaths in tropical forests not only mediates nutrient movement and physical weathering, but also potentially influences vegetation growth and dynamics with seasonally dry or saturated soil conditions. However, few small-scale catchment studies (10-1000ha) have produced a comprehensive, standardized dataset on soil hydrologic properties among tropical forest catchments, due in part to complexities within tropical systems, and to inconsistencies in methods, data collection, and/or analyses. In response, this study has utilized the global, standardized network of forest dynamics plots of the Center for Tropical Forest Science (CTFS) for the rapid assessment of soil saturated hydraulic conductivity (Ks) and the water chemistry from storm-generated flowpaths. Ks measurements at varying depths help in testing Elsenbeer's (2001) functional classification continuum of tropical forest soilscapes and resulting hydrologic flowpaths. In Barro Colorado Island, Panama, Ks decreased rapidly with soil depth where horizontal surface and near-surface flowpaths were most prevalent. Ks measurements in Yasuni National Park, Ecuador indicated limited vertical movement of water at depths $>$15cm due to an impermeable soil layer. Ks measurements from Lambir Hills National Park, Malaysia, represented both ends of the continuum due to variability in soil type and lithology. In relation to soil hydrology and hydrological flowpaths, runoff chemistry at Yasuni reveals a general pattern of increased nutrient export as water moves through the canopy and over the soil surface, with concentrations of K+ increasing significantly in throughfall, and concentrations of both K+, and NO3- remaining high in overland flow. The results from the composite overland flow samples may indicate a more open nutrient cycle in tropical forest environments than has been suggested from earlier studies using radioactively labeled isotopes.
H51D-1170 0800h
A Comparison of Runoff Pathways and Nutrient Export in Small Tropical Forest Catchments
The Center for Tropical Forest Research (CTFS), a program of the Smithsonian Tropical Research Institute (STRI), has coordinated a global network of 17 tropical forest dynamics plots of approximately 50 hectares in order to collect baseline information and to monitor forest changes. Missing from most past studies of these plots is an integrated soil hydrology and water chemistry component. To fill this gap, we have developed and are testing rapid assessment methods to measure soil and hydrological properties for tropical forest catchments. This assessment includes gaging and sampling first to third order headwater streams with high frequency over several storm events within a 2-4 week period. Detailed flow separations enable us to test Elsenbeer's (2001) functional classification continuum for tropical soils and allow us to test the hypothesis that forest sites with greater overland flow experience greater nutrient loss during storm events. Data from a storm event for the steep Lutz Creek Catchment on Barro Colorado Island, Panama in September 1990 demonstrate that Na$^{+}$ and Si, typical of most solutes, decrease in concentration with increasing discharge. In contrast, the nutrients, K$^{+}$ and NO$_{3}$$^{-}$ increase in concentration with increasing discharge. Results from a 20 ha catchment in Yasuni National Park, Ecuador from November 2003 show a similar pattern during several small events on relatively impermeable soils. Data collected from a more permeable 20 ha catchment in Lambir Hills National Park, Malaysia in July 2004 also show nutrient export, but suggest that rainfall amount, intensity and duration may play a large role in the magnitude of nutrient concentrations. Elsenbeer, H., 2001. Hydrological flowpaths in tropical rain forest soilscapes-a review. Hydrological Processes, 15: 1751-1759.
H51D-1171 0800h
Characterization of Dissolved Organic Carbon Flushing in Three Small Watersheds, Oregon USA
The hydrologic and biogeochemical responses of forested watersheds to inputs of rainfall and snowmelt can be an indicator of internal watershed function. We are studying how the flushing of dissolved organic carbon (DOC) in response to storm events may be sensitive to DOC pools contributing to runoff. We assessed how the quantity and quality, both chemical and spectroscopic, of DOC changed in response to a six day storm event during the wet season of 2003 in three small ($<$1 km2) basins in the H.J. Andrews Experimental Forest, Oregon. The watersheds included one old-growth watershed (WS 2) and two previously logged watersheds (WS 1 and WS 10). Pre-storm concentrations of DOC ranged from 1.5 to 2.2 mg C L$^{-1}$ in the three watersheds and increased approximately three-fold on the ascending limb of the storm hydrograph. Initial and stormflow concentrations of DOC were both highest in the unharvested, old-growth watershed. The specific UV absorbance (SUVA, 254 nm) of DOC in the three watersheds increased with rising DOC concentrations during the storm suggesting that the DOC mobilized from catchment soils during storms is more aromatic than the DOC present during baseflow. The increase in UV absorbance ranged from 8 to 36% and was most pronounced in the catchments that had been previously harvested. Chromatographic fractionation of DOC similarly showed that there was a shift in the chemical character of DOC during the storm, with the percentage of DOC composed of non-humic material decreasing by 8-20% from pre-storm conditions and then rebounding after the conclusion of the storm. Fluorescence properties of DOC during the storm event suggest that there is not a pronounced shift in the relative proportion of streamwater DOC derived from terrestrial versus aquatic precursor material. Taken together, these results suggest that DOC characterization can be used as a decisive tool to investigate changing sources of DOC and water within forested watersheds. Further complementary work assessing within watershed DOC pools could improve upon biogeochemical and hydrologic interpretation of watershed functions.
H51D-1172 0800h
Separation in storm-event trajectories of DOC and nitrate concentrations with seasons
The temporal evolution of DOC and nitrate concentrations during storm events and the flow paths responsible for their release have received considerable attention. Some researchers observed DOC and nitrate peaks on the hydrograph rising limb and attributed it to "flushing". Flushing was defined as the release of solutes from near-surface soil layers associated with a rising water table. In contrast, others found a nitrate peak on the rising limb and a delayed DOC peak at or after discharge peak. In this case, the early nitrate peak was attributed to "displacement" of groundwater into the stream and the DOC peak was attributed to flushing. We studied the temporal patterns of storm-event DOC and nitrate for more than 20 storm events over a year in a watershed in Western New York, USA. Storm event sampling was conducted across four partly nested catchments - 696 (s1), 3.4 (s2), 1.6 (s3) and 1.9 (s5) ha. The four watersheds contained varying proportions of hillslope and valley-bottom saturated areas. Early spring rainfall events showed very similar nitrate and DOC trajectories with a peak in concentration at or after the discharge peak. Riparian groundwater elevations showed peak water tables coinciding with solute peaks suggesting flushing of both DOC and nitrate from the catchment. For summer and fall events however, there was a clear separation between the DOC and nitrate trajectories. Nitrate was at its highest early on in the storm event and then followed a dilution trajectory as the event proceeded. DOC concentrations increased through the event (as nitrate dropped) with a peak in concentration just after discharge peak. This seasonal shift in DOC and nitrate trajectories was observed across catchment scales. These observations show that both flushing and displacement mechanisms could occur in a watershed and their expression is regulated by the availability of the solutes in the soil profile. DOC concentrations were found to be high in upper soil layers across all seasons, but nitrate build-up in the upper soil layers was limited to the dormant period before spring onset. We hypothesize that the temporal expression of solutes (especially DOC) is regulated by the total extent of the saturation in the watershed, the areal distribution of the saturated areas and their connectedness, and the seasonally regulated availability of the solutes. These hypotheses are currently being evaluated by comparing the temporal expressions across catchment scales.
http://www.buffalostate.edu/orgs/glc/gowanda/overview.htm
H51D-1173 0800h
Landscape regulation of nitrogen export during snowmelt in boreal watersheds of northern Sweden
In boreal watersheds of northern Sweden, the four week period of snowmelt provides nearly 50% of the annual runoff and a majority of the annual flux of carbon and nutrients to streams. Since the average snow cover is 170 days per year, changes in the depth of snow cover are likely to influence winter conditions such as soil moisture and soil frost which in turn affect microbial processes. Changes in the amount and timing of snow melt can also affect the thawing of seasonal frost and dominant water flow paths from hillslopes to streams. We examined dissolved organic carbon (DOC), dissolved organic nitrogen (DON), and dissolved inorganic nitrogen (nitrate and ammonium) in three streams which differ in catchment characteristics. We found that the stream with the greatest proportion of wetland coverage had higher concentrations of DOC, DON, and ammonium, but undetectable amounts of nitrate during snow melt. Conversely, two streams with low wetland coverage had three-fold lower DOC and two-fold lower DON concentrations, but nearly equal amounts of ammonium and nitrate. Across all catchments, nitrate was inversely related to the ratio of DOC to DON, suggesting that soil organic matter quality and microbial processes influence nitrate flux during snow melt. Ongoing research will expand upon current knowledge of stream chemistry by incorporating snow pack, lysimeter, and groundwater chemistry to examine the mechanisms which regulate nitrogen flux during the snow melt period in northern Sweden.
H51D-1174 0800h
Hydrochemical Modeling of Dissolved Organic Carbon in a Small, Undisturbed, Forested Watershed in Southern Chile
The objective of the present study is to model Dissolved Organic Carbon (DOC) concentrations in a stream draining a small, undisturbed, old-growth forested watershed in Southern Chile and test model results against measured data. DOC plays an important role in several processes in terrestrial and aquatic ecosystems. For example, through the formation of organic complexes, DOC can influence nutrient availability, affect the solubility, mobility, and toxicity of metals, and control the absorption of pesticides to soils. DOC also influences biological activity by absorbing UV-B radiation and can contribute significantly to freshwater acidity. Additionally, DOC is linked to the formation of trihalomethanes as by-products of the disinfection of drinking water with chlorine, which constitutes a potentially serious threat to human health. Despite plentiful research on biogeochemical processes controlling DOC production and consumption, there is little information from minimally impacted environments, which can provide valuable baseline information from which to evaluate the more impacted ecosystems. Our study focused on a virtually unpolluted old-growth forested watershed in Southern Chile. We developed a conceptual model that assumes DOC production in forest soils is a function of temperature and DOC transport from soil to stream water is a function of discharge and hydrological flow paths. The hydrological response of the catchment under study was simulated using a simple lumped model, based on two years of meteorological data previously collected. Three different equations were used to simulate DOC concentrations in soil water as a function of temperature, and ultimately to derive DOC concentrations in streamflow. Model results were tested against two years of measured DOC concentrations in streamflow, and all three models provided a reasonably good representation of the DOC response of the small studied watershed and a better agreement to the observed DOC than previously published, similarly simple models.
H51D-1175 0800h
Eco-hydrologic Modeling of Nutrients, Oxygen, and Temperature Across a Range of Michigan Streams
Ecological responses to human alterations of the landscape are not well understood. In Michigan, agricultural land use is a common alteration, but there are only limited methods of linking water quality changes caused by such alterations to ecological responses at different scales. Numerical models can be used to explore the dominant processes and process rates creating ecological stress. Such a model must consider the dynamic relationships between oxygen, nutrients, temperature, and resident organisms. Also, multiple spatial and temporal measurements are required to characterize a stream system due to dramatic seasonal and diurnal variability. Multiple streams in Michigan were chosen to model based on land use and data availability, including the regional Muskegon River and Grand Traverse Bay watersheds. Pre-dawn synoptic sampling was used to identify stream segments with low dissolved oxygen. In addition, temperature was recorded and chlorophyll and nutrient samples were collected to link oxygen stress to multiple parameters including, like nutrient conditions. The preliminary eco-hydrologic models used for analysis of this stream data were QUAL2E and AQUATOX. QUAL2E was used to simulate the water quality assuming well-mixed, dendritic streams. AQUATOX was used to simulate changing concentrations of nutrients, other solutes, and sediments in the water, as well as organism populations. Preliminary results for mid-July conditions compare favorably to with observed dissolved oxygen and temperature levels.
H51D-1176 0800h
Modeling Carbon and Water Vapor Fluxes and Carbon Isotope Discrimination at the Canopy Scale in a Semi-arid Pine Forest.
Water, energy, and carbon exchange between the biosphere and the atmosphere in forest ecosystems are strongly coupled and affected by stomatal conductance and photosynthesis, which in turn respond to environmental factors such as air humidity, temperature, radiation, and soil water content. In this study, we test biochemical models of photosynthesis, stomatal conductance, and carbon isotope discrimination at canopy scales, using eddy covariance and isotopic data from the AmeriFlux and BASIN networks. Carbon and water vapor fluxes were simulated with an ecophysiologically based model (ISOLSM) driven by half hourly meteorology at the Old Ponderosa flux tower in Metolius (OR). The model was parameterized with half hourly eddy covariance data of carbon, latent and sensible heat fluxes, and foliar carbon isotope ratios (\delta$^{13}$C) from the same site. Carbon isotope discrimination was sensitive to stomatal conductance parameters that also affect estimates of carbon and water vapor fluxes, reducing the parameter space obtained from the eddy flux data. The variability of simulated \delta$^{13}$C was similar to that of observed \delta$^{13}$C, and the relation between vapor pressure deficit (VPD) and simulated \delta$^{13}$C of assimilated carbon was similar to that between VPD and observed \delta$^{13}$C of ecosystem respiration. These simulations support hypotheses of tight atmospheric controls of stomatal conductance, photosynthesis, and carbon isotope discrimination at short time scales, which affect biosphere-atmosphere carbon and water exchange at large spatial scales.
H51D-1177 0800h
Rainfall-Vegetation Coupling in Semi-Arid Regions: Influence of the Nitrogen Cycle on Vegetation Temporal Dynamics
The ability to predict how the terrestrial environment responds to climate variability rests with understanding how rainfall, nutrient availability, and vegetation are dynamically coupled. In semi-arid ecosystems, water or nutrients are typically considered to be responsible for imposing limitations on vegetation productivity, and shifts in the specific limitation can occur abruptly over large-scale climatic gradients. For instance, in the Kalahari region of southern Africa there exists a well-defined linear relationship (R$^{2}$ = 0.94) between mean wet season rainfall and tree cover over the portion of the Kalahari that receives between 300 and 700 mm of wet season rainfall per year. In areas that receive greater amounts of rainfall, nutrients become the limiting factor, as reflected by the dominance of nutrient-poor savanna phenotypes and the fact that water use is not optimized by the ecosystem vegetation. Here vertical drainage from the root zone mobilizes nitrogen, which is considered to be the primary limiting nutrient in this region, and this perpetuates the sub-optimal (in terms of water use) canopy coverage. An analytical soil moisture model is used in conjunction with nitrogen submodel to explore the effects of changing rainfall on the vegetation structure, and in particular how hysteresis in the rainfall-vegetation relationship is mediated by the nitrogen cycle. Savanna ecosystem stability is discussed in this context.
H51D-1178 0800h
Diurnal variations in the \delta$^{18}$O of water vapor observed in the Pacific Northwest coniferous forests
Changes in the $^{2}$H and $^{18}$O of atmospheric water vapor provide information for integrating aspects of gas exchange within forest canopies. In this study we showed that overstory transpiration dominated \delta$^{18}$O values of canopy water vapor (\delta$^{18}O$_{v}$) in an old-growth coniferous forest in the Pacific Northwest of United States. Diurnal fluctuations nearly 5 $\permil$ in \delta$^{18}O$_{v}$ were observed at 3 heights both above and within the canopy. Values of \delta$^{18}O$_{v}$ were $\sim$ -18 $\permil$ (VSMOW scale) before dawn, and gradually decreased to $\sim$ -23 $\permil$ at noon before slowly becoming more enriched again in the late afternoon. This diurnal pattern reflected a balance of \delta$^{18}$O signatures associated with early-transpired water and the water transpired at later hours when isotopic steady state was established. Small transpiration flux relative to leaf water contents in the early morning prevents leaf transpiration quickly approaching steady state. Using a non-steady state model, we compared the measured and predicted \delta$^{18}$O values of leaf water over the course of a day and showed a lagged response of leaf water enrichment. Using an isotopic approach, we determined that canopy transpiration accounted for $\>$ 80 $%$ of total evapotranspiration for two summer days in this old-growth forest, which agreed closely with above- and within-canopy eddy covariance measurements. Partitioning evapotranspiration into transpiration and evaporation fluxes using \delta$^{18}$O as a tracer needs to carefully assess whether steady state assumption is satisfied.
H51D-1179 0800h
Anthropogenic changes to the hydrologic regime and potential effects on anadromous salmonids in California south of the timber forests
Though the natural range of anadromous salmonids on the Pacific coast of North America extends from Alaska to Baja California, most studies to identify factors that affect their viability in coastal watersheds are conducted in timber-dominated catchments from northern California into Canada. These studies have focused especially on anthropogenic additions of fine sediment as limiting salmonid survival. South of the timber forests, the availability of water may be more important than sediment for affecting salmonid viability. The potential influence of human water demands on the natural hydrologic regime may be greater in central California than in the Pacific Northwest because of less total annual rainfall and the persistent drought that occurs from May through September (when human water needs are greatest). Identifying human-caused impacts requires a thorough understanding of regional climate and hydrology, which vary widely across temporal scales (from one year to the next, across seasonal gradients, and even over a daily scale), and across spatial scales as well. By examining historical stream flow and precipitation records, we characterize the effects of the climate on stream flow, and how these effects vary through the drainage network: more than 50% of annual precipitation is transferred to runoff in an average year, but only 5% of this runoff occurs between May and September. Manipulations in the hydrologic regime may affect winter flows, when salmonids need different water levels for passage, spawning, and channel maintenance; but they may have a more profound and widespread effect on spring and summer flows, when water is needed for juvenile rearing. By examining historical stream flow records and documented demands of stream water for human use, we identify the short-term and long-term impacts (through measures of intensity, persistence, and frequency) of potentially adverse conditions to various stages of the salmonid life cycle that manipulations to the natural hydrologic regime may cause as a result of water use. Analyses considering the spatial and temporal variations of the hydrologic regime, and the potential impacts that human water demands may place on natural hydrologic processes, are essential for planning restorative practices and for developing management plans for coastal watersheds throughout this region.
H51D-1180 0800h
A Dose-Structured Population Dynamics Model for Outmigrant Salmon
The response of fish populations to differing levels of exposure to stressor chemical is commonly characterized using canonical dose-response models that are calibrated in laboratory (e.g., toxicological) studies. Use of such information in the study of migrating populations is difficult because dose received in the environment can vary greatly within a population due to the heterogeneity of the mixing between population members and stressor chemicals. Thus direct use of dose-response models is often predicated on assumptions of complete mixing in the environment. To relax this assumption it is required to devise an approach that keeps track of dose as a distributed quantity over a population. Here such a method is described that uses an added dimension of exposure-time to keep track of the mixing time, or dose, between outmigrant Fall Chinook Salmon and stressor organic contaminants. A mathematical model for the fish population dynamics is developed in the form of a first-order partial differential equation in multiple structural dimensions of dose, size, and age, in addition to space and time. A method-of-characteristics solution to the model under some simplifying conditions is presented and described. The results shed light on the nature of the distribution of outmigrant salmon over structural variables on arrival at the ocean.
H51D-1181 0800h
Increases in Stream Flow at Stream Crossings on Forest Roads in Western Oregon, USA
The hydrologic impacts of forest roads on a watershed are uncertain despite decades of research. There is consensus, however, that roads can alter hydrologic pathways by intercepting subsurface flow and by producing overland runoff. Where a forest road crosses a stream the road can channel runoff, via a roadside ditch, directly to the stream. At these stream crossings the peak flow and runoff volume of the stream may be increased. This study monitored ditch flow and stream flow at fifteen stream crossings within an 824 ha watershed in western Oregon. Ditch flow and stream flow at each crossing were matched in time. Stream flow increases were calculated for five storms that occurred during the winter of 2002-2003. Nine of the fifteen streams consistently had greater than a five percent increase in peak flow and stream runoff at the stream crossing. For these nine stream crossings, the average increase in peak flow was 42 percent and the average increase in stream runoff was 77 percent. Ditch flow, on average, contributed the most volume on the rising limb of the stream hydrograph at the stream crossing. Stream flow increases have management implications for flooding and road drainage features.
H51D-1182 0800h
Total Runoff From Roads as an Index for Potential Changes in Watershed Hydrology
Known impacts of forest roads on watersheds include altered magnitude and timing of peak flows. The effect of the contribution of flow from roadside ditches on a watershed is less well understood. Roadside ditches channel Hortonian overland flow that falls on road surfaces. Also, roadside ditches intercept and channel subsurface flow. The objective of this study was to determine the magnitude of the surface runoff from roads that might alter the hydrology of a roaded watershed. The study was carried out in an 824 ha forested watershed in the foothills of the Oregon Coast Range in the MacDonald-Dunn Research Forest approximately 5 km west of Corvallis, Oregon. In the Oak Creek Watershed, there are 4.57 km of road and 4.87 km of stream resulting in a drainage density of 0.64 km/km$^{2}$ and a road density of 0.6 km/km$^{2}$. There are 98 drainage structures on the roads in the Oak Creek Watershed. Twenty-three of the drainage structures are stream-crossing culverts and the balance are cross-drain culverts. Discharge was measured at all of the drainage structures and at the outlet of the watershed from October 2002 through May 2003 for five storms. The hydrology of the roads in the Oak Creek Watershed is highly variable in space and time. Road runoff was measured and expressed as a proportion of total watershed runoff. This metric allows an insight into the potential for the road system to alter watershed hydrology.
H51D-1183 0800h
Water table and overbank flow frequency changes due to suburbanization-induced channel incision, Virginia Coastal Plain, USA
Channel incision is a widely observed response to increased flow in urbanized watersheds, but the effects of channel lowering on riparian water tables is not well documented. In a rapidly incising suburban stream in the Virginia Coastal Plain, we hypothesize that incision has lowered floodplain water tables and decreased the overbank flow frequency, and suggest these changes impact vegetation distribution in a diverse, protected riparian habitat. The monitored stream is a tributary to the James River draining 1.3 km$^{2}$, of which 15% is impervious cover. Incision has occurred largely through upstream migration of a one m high knickpoint at a rate of 1-2 m/yr, primarily during high flow events. We installed 33 wells in six floodplain transects to assess water table elevations beneath the floodplain adjacent to the incising stream. To document the impacts of incision, two transects are located 30 and 50 m upstream of the knickpoint in unincised floodplain, and the remainder are 5, 30, 70, and 100 m downstream of the knickpoint in incised floodplain. In one transect above and two below, pressure transducers attached to dataloggers provide a high-resolution record of water table response to storm events. Significant differences have been observed in the water table above and below the knickpoint. Above the knickpoint, the water table is relatively flat and is 0.2-0.4 m below the floodplain surface. Water table response to precipitation events is nearly immediate, with the water table rising to the floodplain surface in significant rainfall events. In the transect immediately downstream of the knickpoint, the water table possesses a steep gradient, rising from ~1 m below the floodplain at the stream to 0.3 m below the surface within 20 m. In the most downstream transects, the water table is relatively flat, but is one m below the floodplain surface, equivalent to the depth of incision generated by knickpoint passage. Upstream of the knickpoint, overbank flooding occurs frequently, while below the knickpoint the majority of storm flow is contained within the incised channel and occupation of the floodplain is rare. Plant diversity surveys reveal differences in the total density of herbaceous growth and species distribution between the floodplain above and below the knickpoint. Results from >100 plots show that there is more leaf litter, less exposed ground, and a decrease in floodplain species cover in the incised portion of the floodplain. The changes in flood frequency and water table elevation appear to have allowed one invasive species, Japanese stilt grass (Microstegium vimineum), to become dominant in the floodplain understory, displacing native wetland species.
H51D-1184 0800h
A Data Mining Approach for Understanding Topographic Control on Climate-Induced Inter-Annual Vegetation Variability Over the United States
The complex feedback relationship between climate variability and vegetation dynamics is a subject of intense investigation for its implications in furthering our understanding of the global biogeochemical cycle. We address an important question in this context: ``How does topography influence the vegetation's response to natural climate fluctuations?'' We explore this issue through the analysis of inter-annual vegetation variability over a very large area (continental United States) using long-term (thirteen year period of 1989-2001), monthly-averaged, biweekly maximum value composite normalized difference vegetation index (NDVI) data. These data are obtained from satellite remote sensing at 1-km resolution. Through the novel implementation of data mining techniques, we show that the Northern Pacific climate oscillation and the ENSO phenomena influence the year-to-year vegetation variability over an extensive geographical domain. Further, the vegetation response to these fluctuations depends on a variety of topographic attributes such as elevation, slope, aspect, and proximity to moisture convergence zones, although the first two are the predominant controls. Therefore, the dynamic response of terrestrial vegetation to climate fluctuations, which shows tremendous spatial heterogeneity, is closely linked to the variability induced by the topography. These findings suggest that the representation of vegetation dynamics in existing climate models, which do not incorporate such dependencies, may be inadequate. Therefore, climate models that are regularly employed to guide policy decisions need to better incorporate these dependencies for the assessment of terrestrial carbon sequestration under evolving climate scenarios.
H51D-1185 0800h
Assessing Ecological Function with Hydrologic Models in Restoration Projects: Matching Expectations with Predictive Capability
The attempt to return the Everglades to a more naturally functioning ecosystem is the world's most ambitious environmental restoration project. Restoration success has been defined in terms of ecological performance measures, the criteria used to judge and select among various alternative project designs. Since ecological function in the Everglades is strongly linked to the hydrologic cycle, performance measures are developed by identifying hydrologic conditions (e.g. critical water depths) that support key ecological functions. Hydrologic models are then used to analyze these performance measures. In many cases the performance measures and the hydrologic models are developed independently. Ecologists develop the performance measures, generally without a full understanding of the predictive capability of the simulation models that will be used in the analysis. Engineers build the numerical models, but with only limited appreciation of how the models will be applied toward decision-making based on ecologic performance measures. This large divide between ecologists and engineers can result in models that are unable to produce the information required to assess the ecological criteria. The hydrologic models fail to produce the hydrologic information necessary to predict ecological response, or more likely, cannot produce the required information at the correct scale or within some required accuracy or range of uncertainty. To illustrate the potential pitfalls of the disconnect between ecologists and hydrologic modelers, we present projects and associated performance measures currently being applied in Everglades restoration. From these illustrations suggestions are made for improving the co-development of performance measures and models. General principles regarding restoration project evaluation and analyses are also derived.
H51D-1186 0800h
Using Residence-Time Distributions to Compare Water-Level Hydrographs
The availability of accurate, high-frequency water-level data has greatly improved the potential for site-specific hydrologic assessments. Traditional comparisons between sites included summary statistics (mean/median water level) and qualitative assessments of hydrographs. In many instances traditional summary statistics are not adequate to convey the importance of water-level fluctuation or the amount of time water levels are at or above a critical level. For instance, plant growth in wetlands depends on water levels reaching the root zone and sustaining inundation for adequate periods during the growing season. In this paper we present a method of analyzing continuous high-frequency water-level data to determine the distribution of residence time for a specific water level. Typically the root-zone elevation is used as the specific water level, resulting in the distribution of root-zone residence times. We compare this approach to traditional methods at a range of hydrological settings in Wisconsin. In addition to the information contained in the graphical display of the distribution functions, pairs of sites can be compared by examining the duration for specific probabilities. Using the Smirnov test it is possible to determine whether there are statistically significant differences between the distributions for two sites. This insight helps describe hydrologic conditions that drive biological responses--a relation difficult to explain using simple summary statistics derived from water-level measurements.
H51D-1187 0800h
Fractals in Nature: A Spectral Analysis of Nutrient Transport and Landscape Hydrology
The transport of phosphorus (P), a major source of pollution in freshwater ecosystems, is directly linked to the flow paths and travel time of water through a catchment. The objective of this research is to understand P movement for improved management of drinking water resources. Here we used spectral analysis to show the differences in long-term P transport trends between an active agricultural watershed (160 ha) and an abandoned agricultural forested watershed (85 ha) in the Catskills Mountains, NY, specifically, in Delaware County. The watersheds were close to each other, 6.4 km apart, so that hydro-meteorological differences were small. The results suggest interesting shifts in P transport behavior when historically fertilized land is abandoned and allowed to revert to forest. Spectral analysis, a long-term frequency domain time series analysis method, has been successfully used to analyze long-term time series data, quantify travel time distributions, and measure the watershed scale retardation factor for reactive solutes (Kirchner et al. 2000, Nature 43:524-527; Kirchner et al., 2001J. Hydrol. 254:82-101). Here we combined Spectral Analysis of nutrient data and water table fluctuations to provide a perpetual model of the watershed. We found that spectral analysis is a useful tool for analyzing long-term time series of water and chemical fluxes record for understanding ecosystem responses to disturbance and provide a in depth view of long-term effects of changing agricultural and natural resource management practices.
H51D-1188 0800h
Nesting High-resolution Multi-layer Photosynthesis Approaches in Current Forest Productivity Models: A Cost-Benefit Analysis in the Time-Frequency Domain
Ecosystem processes relevant to carbon transfer and storage are known to vary over many time and space scales. In the time domain, processes ranging from seconds, such as turbulent transport, to seasons, such as plant phenology, affect assimilation and respiration, which in turn, control carbon allocation over time scales of days to years. These inter-related processes contribute to the forest development (often measured in years to decades) and long-term carbon sequestration. To date, no single model captures the entire spectrum of variability of these processes; rather, a modular approach is adopted in which the forcing and response variables are mechanistically coupled over an inherent or assumed time scale that is then integrated to longer time scales. The effect of such modular parameterization of the "fast" processes and their cross-scale interaction with the slowly varying processes on long-term carbon sequestration remains a subject of investigation. We address this problem in two ways. First, we perform a multi-model inter-comparison in the time and frequency domains to assess how different parameterizations of photosynthesis and water vapor fluxes in forest growth models (e.g. BGC, SECRETS, PnET and 3PG) reproduce the observed spectrum of these two fluxes from hours to years. These models were chosen because they significantly vary in complexity and integration time step, thereby "filtering" the flux spectrum differently. Next, we explore the consequences of this filtering on cross-scale information flow using a newly proposed nested scheme that employs multi-species allocation routines with assimilation calculated with CANVEG. CANVEG is a multi-layer and multi-species model that resolves the entire canopy microclimate and uses a dynamic leaf area density as an input. The analysis is done in a cost-benefit fashion evaluating the gain in predictive skills of long-term carbon sequestration as result of extra model complexity and added parameterizations. As a case study, we use multi-year CO2 and water vapor flux measurement from a Pine Plantation at the Duke Forest Ameriflux site.
H51D-1189 0800h
Correlation Between in-situ Redox Reaction Rates and Microbial Biomass Distribution in Porous Media Influenced by Different Transport Regimes
Microbially mediated redox transformations of organic carbon play an important role for the fate of reactive species in porous media. The terminal electron acceptors (TEAs) involved in such reactions depend on the amount and degradability of the organic carbon species and lead to a succession of redox reactions where the TEAs are used-up in a temporal or, in case transport is considered, spatial sequence of decreasing energy yields. A direct characterization of redox stratified systems is challenged by our ability to measure reaction rates in-situ. One novel approach consists in quantifying and characterizing microorganisms in aquifers and sediments and to use such results to predict in-situ redox reaction rates. However, the existence of a spatial correlation between microbial abundance and associated in-situ redox reaction rates should be questioned. Here, we investigate this correlation for porous media having different transport regimes. In the environment, these regimes vary between systems such as aquifers, where advective transport in the water phase is the dominant transport mechanism, and aquatic sediments, where close to the sediment water interface the mixing activity of benthic macrofauna contributes significantly to transport. Results from estuarine sediments show that for such systems, the spatial distributions of redox reaction rates and the associated microorganisms are not correlated. This observation is supported by reactive transport simulations, which show that the ratio of the time scale of the mixing processes to the time scale of microbial growth is controlling the spatial correlation between redox reaction rates and the distribution of microorganisms. For sediments highly affected by mixing, the correlation is missing or weak, while in advection controlled systems such as aquifers, a good correlation between redox rates and microbial biomass distribution can be expected.
H51D-1190 0800h
Rain-Impact-Entrainment of Chemicals and Soil into Overland Flow in Saturated Areas: Theory and Experiments
Overland flow from riparian and other frequently saturated areas is a potentially important transport pathway between the landscape and aquatic ecosystems. Both raindrop driven processes and diffusion play important roles in the transfer of chemicals from soil to surface runoff, however, current transport models either do not consider the two processes together, or use "effective" parameters with uncertain physical definitions. We developed a physically based, solute transport model that couples both mechanisms and tested it with experimental data. One unique aspect of this study is that all the parameters needed to apply the model to our experiments were either directly measured or previously published, that is, there was no model "calibration" or "fitting." Our model assumes that chemicals near the surface of the soil are ejected into runoff by raindrop impact and chemicals deeper in the soil diffuse into a surface layer, or "exchange layer," via diffusion. The exchange layer depth and transfer processes are derived from the "shield" concept in the Rose soil erosion model (e.g., Rose, 1985, Adv. Soil Sci. 2,1-63.). The model's governing equations were solved numerically and the results agreed well with experimental data (R2 > 0.90). The model was also successfully tested against previously published experimental data by Leman and Ahuja (1983, J. Environ. Qual. 12(1), 34-40); these data were unique because they provided chemical concentrations in the soil profile as well as in the overland flow. This model provides insights into important processes relevant to landscape-river interactions and water quality protection.