Search

Watershed models are essential tools to understand, quantify, and predict hydrologic processes and water quality responses from scales ranging from field to large river basins. However, the reliability of watershed models in a management context depends largely on inherent uncertainties in model predictions. The objective of this study is to quantify prediction uncertainty for flow, sediment, total nitrogen (TN), and total phosphorus (TP) resulting from model structure. We do this using using three process-based models: the Soil and Water Assessment Tool-Variable Source Area model (SWAT-VSA), the standard Soil and Water Assessment Tool (SWAT-ST), and the Chesapeake Bay Program’s Phase 6 Watershed Model (CBP-Model). We initialize each of the models using meteorological, soil, and land use data, and analyze outputs of flow, sediment, TN, and TP fluxes at the U.S. Geological Survey stream gauge at the downstream end of the Susquehanna River Basin in Conowingo, Maryland. Using these three models, we develop and compare two types of Bayesian models, a Bayesian Generalized (Non-) Linear Multilevel Model (BGMM), and a Bayesian Model Averaging (BMA) for flow, sediment, TN, and TP and 95% credible intervals. We compare the Bayesian models results against the individual model results, and straight model averaging (SMA) using a split time period analysis to assess their predictive strengths. Both Bayesian models provided substantially better predictions than the individual process-based models, and estimates of prediction uncertainty, which can enhance decision-making and improve watershed management by providing a risk based assessment of outcomes.

Nutrient export from agricultural landscapes is a water quality concern and the cause of mitigation activities worldwide. Climate change impacts hydrology and nutrient cycling by changing soil moisture, stoichiometric nutrient ratios, and soil temperature, potentially complicating mitigation measures. This research quantifies the impact of climate change and climate anomalies on hydrology, nutrient cycling, and greenhouse gas emissions in an agricultural catchment of the Chesapeake Bay watershed. We force a calibrated model with seven downscaled and bias-corrected regional climate models and derived climate anomalies to assess their impact on hydrology and the export of nitrate (NO3-), phosphorus (P), and sediment, and emissions of nitrous oxide (N2O) and di-nitrogen (N-2). Modelaverage (+/- standard deviation) results indicate that climate change, through an increase in precipitation and temperature, will result in substantial increases in winter/spring flow (10.6 +/- 12.3%), NO3-(17.3 +/- 6.4%), dissolved P (32.3 +/- 18.4%), total P (24.8 +/- 16.9%), and sediment (25.2 +/- 16.6%) export, and a slight increases in N2O (0.3 +/- 4.8%) and N-2 (0.2 +/- 11.8%) emissions. Conversely, decreases in summer flow (-29.1 +/- 24.6%) and the export of dissolved P (-15.5 +/- 26.4%), total P (-16.3 +/- 20.7%), sediment (-20.7 +/- 18.3%), and NO3-(-29.1 +/- 27.8%) are driven by greater evapotranspiration from increasing summer temperatures. Decreases in N2O (-26.9 +/- 15.7%) and N-2 (-36.6 +/- 22.9%) are predicted in the summer and driven by drier soils. While the changes in flow are related directly to changes in precipitation and temperature, the changes in nutrient and sediment export are, to some extent, driven by changes in agricultural management that climate change induces, such as earlier spring tillage and altered nutrient application timing and by alterations to nutrient cycling in the soil. (C) 2018 Elsevier B.V. All rights reserved. National Science FoundationNational Science Foundation (NSF) [1360415, 1343802]; USDAUnited States Department of Agriculture (USDA) [2012-67019-19434] We would like to acknowledge high-performance computing support from Yellowstone (http://n2t.net/ark:/85065/d7wd3xhc) provided by NCAR's Computational and Information Systems Laboratory, support from the National Science Foundation under award numbers 1360415 and 1343802, and funding support from the USDA under project number 2012-67019-19434. All data, methods, and code used in this manuscript are available upon request. Public domain – authored by a U.S. government employee

Increased adoption of denitrifying bioreactors to mitigate nutrient export in agricultural drainage can improve water quality but also raises concerns about potential unintended consequences of widespread implementation. Greenhouse gas (GHG) emissions, especially as nitrous oxide (N2O) produced during denitrification, are of particular concern. To gain understanding of the variables controlling bioreactor GHG production, the effects of nutrient loading, hydraulic residence time, and biochar addition on emissions of N2O, methane (CH4), and carbon dioxide (CO2) were tested in 6550 cm3 laboratory columns. Biochar, an organic carbon pyrolysis product, was hypothesized to reduce GHG emissions, as has been reported with its use as an agricultural soil amendment. GHG fluxes were measured during 5-day trials using the closed dynamic chamber technique. The 36 combinations of three media types (woodchips, 10% biochar, and 30% biochar), three residence times (3, 6, and 12 h), and four influent formulations, all combinations of high and low nitrogen as nitrate (16.1 and 4.5 mg NO3-N L−1) and phosphorus as orthophosphate (1.9 and 0.6 mg PO4-P L−1) concentrations, were tested in triplicate. Treatment effects were assessed with linear mixed effects models. Biochar addition, particularly at the 30% rate, was found to increase N2O and CO2 emissions and the proportion of removed NO3-N released as N2O-N. However, quantified GHG flux rates were not environmentally concerning. Methane fluxes were negligible, and CO2 fluxes were not considered to increase net GHG emissions. Nitrous oxide fluxes normalized to surface area averaged 2.92 mg N2O-N m−2 d−1, and 98% of measured fluxes were within the range previously reported for woodchip bioreactors and considered acceptable.

Agricultural conservation practices (CPs) are commonly implemented to reduce diffuse nutrient pollution. Climate change can complicate the development, implementation, and efficiency of agricultural CPs by altering hydrology, nutrient cycling, and erosion. This research quantifies the impact of climate change on hydrology, nutrient cycling, erosion, and the effectiveness of agricultural CP in the Susquehanna River Basin in the Chesapeake Bay Watershed, USA. We develop, calibrate, and test the Soil and Water Assessment Tool-Variable Source Area (SWAT-VSA) model and select four CPs; buffer strips, strip-cropping, no-till, and tile drainage, to test their effectiveness in reducing climate change impacts on water quality. We force the model with six downscaled global climate models (GCMs) for a historic period (1990-2014) and two future scenario periods (2041-2065 and 2075-2099) and quantify the impact of climate change on hydrology, nitrate-N (NO3-N), total N (TN), dissolved phosphorus (DP), total phosphorus (TP), and sediment export with and without CPs. We also test prioritizing CP installation on the 30% of agricultural lands that generate the most runoff (e.g., critical source areas-CSAs). Compared against the historical baseline and with no CPs, the ensemble model predictions indicate that climate change results in annual increases in flow (4.5±7.3%), surface runoff (3.5±6.1%), sediment export (28.5±18.2%) and TN export (9.5±5.1%), but decreases in NO3-N (12±12.8%), DP (14±11.5), and TP (2.5±7.4%) export. When agricultural CPs are simulated most do not appreciably change the water balance, however, tile drainage and strip-cropping decrease surface runoff, sediment export, and DP/TP, while buffer strips reduce N export. Installing CPs on CSAs results in nearly the same level of performance for most practices and most pollutants. These results suggest that climate change will influence the performance of agricultural CPs and that targeting agricultural CPs to CSAs can provide nearly the same level of water quality effects as more widespread adoption.

Denitrifying bioreactors are recently-established agricultural best management practices with growing acceptance in the US Midwest but less studied in other agriculturally significant regions, such as the US Mid-Atlantic. A bioreactor was installed in the Virginia Coastal Plain to evaluate performance in this geographically novel region facing challenges managing nutrient pollution. The 25.3 m3 woodchip bed amended with 10% biochar (v/v) intercepted subsurface drainage from 6.5 ha cultivated in soy. Influent and effluent nitrate-nitrogen (NO3–N) and total phosphorus (TP) concentrations and flowrate were monitored intensively during the second year of operation. Bed surface fluxes of greenhouse gases (GHGs) nitrous oxide (N2O), methane (CH4), and carbon dioxide (CO2) were measured periodically with the closed dynamic chamber technique. The bioreactor did not have a statistically or environmentally significant effect on TP export. Cumulative NO3–N removal efficiency (9.5%) and average removal rate (0.56 ± 0.25 g m−3 d−1) were low relative to Midwest tile bioreactors, but comparable to installations in the Maryland Coastal Plain. Underperformance was attributed mainly to low NO3–N loading (mean 9.4 ± 4.4 kg ha−1 yr−1), although intermittent flow, periods of low HRT, and low pH (mean 5.3) also likely contributed. N removal rates were correlated with influent NO3–N concentration and temperature, but decreased with hydraulic residence time, indicating that removal was often N-limited. GHG emissions were similar to other bioreactors and constructed wetlands and not considered environmentally concerning. This study suggests that expectations of NO3–N removal efficiency developed from bioreactors receiving moderate to high NO3–N loading with influent concentrations exceeding 10–20 mg L−1 are unlikely to be met by systems where N-limitation becomes significant.

Greenhouse gas (GHG) emissions from agroecosystems, particularly nitrous oxide (N2O), are an increasing concern. To quantify N2O emissions from agroecosystems, which occur as a result of nitrogen (N) cycling, a new physically-based routine was developed for the Soil and Water Assessment Tool (SWAT) model to predict N2O flux during denitrification and an existing nitrification routine was modified to capture N2O flux during this process. The new routines predict N2O emissions by coupling the carbon (C) and N cycles with soil moisture/temperature and pH in SWAT. The model uses reduction functions to predict total denitrification (N2+ N2O) and partitions N2 from N2O using a ratio method. The modified SWAT nitrification routine likewise predicts N2O emissions using reduction functions. The new denitrification routine and modified nitrification routine were tested using GRACEnet data at University Park, Pennsylvania, and West Lafayette, Indiana. Results showed strong correlations between plot measurements of N2O flux and the model predictions for both test sites and suggest that N2O emissions are particularly sensitive to soil pH and soil N, and moderately sensitive to soil temperature/moisture and total soil C levels. Nitrous oxide emissions are modeled with functional relationships between soil N, C, pH, temperature and moisture levels.The model is tested against high spatial and temporal resolution GRACEnet data at two sites.The model is particularly sensitive to soil pH and soil N levels.

Few current modeling tools are designed to predict short-term, high-risk runoff from hydrologically sensitive areas (HSAs) in watersheds. This study couples the Soil and Water Assessment Tool-Variable Source Area model with the Climate Forecast System Reanalysis model and the Global Forecast System-Model Output Statistics model short term weather forecast, to develop a HSA prediction tool designed to assist producers, landowners, and planners in identifying high-risk areas generating storm runoff and pollution. Short-term predictions for stream flow and soil moisture level were estimated in the South Fork of the Shenandoah river watershed. Daily volumetric flow forecasts were found to be satisfactory four days into the future, and distributed model predictions accurately captured sub-field scale HSAs. The model has the potential to provide valuable forecasts that can be used to improve the effectiveness of agricultural management practices and reduce the risk of non-point source pollution. A watershed model-Global Forecast System model are coupled to develop a critical source area prediction tool.All software is open source and the application applies globally.The tool predicts watershed discharge and spatially distributed soil moisture conditions 14 days into the future.The tool is designed to assist land managers in identifying high-risk areas generating storm runoff and pollution.

Streamflow forecasts are essential for water resources management. Although there are many methods for forecasting streamflow, real-time forecasts remain challenging. This study evaluates streamflow forecasts using a process-based model (Soil and Water Assessment Tool-Variable Source Area model-SWAT-VSA), a stochastic model (Artificial Neural Network -ANN), an Auto-Regressive Moving-Average (ARMA) model, and a Bayesian ensemble model that utilizes the SWAT-VSA, ANN, and ARMA results. Streamflow is forecast from 1 to 8 d, forced with Quantitative Precipitation Forecasts from the US National Weather Service. Of the individual models, SWAT-VSA and the ANN provide better predictions of total streamflow (NSE 0.60-0.70) and peak flow, but underpredicted low flows. During the forecast period the ANN had the highest predictive power (NSE 0.44-0.64), however all three models underpredicted peak flow. The Bayesian ensemble forecast streamflow with the most skill for all forecast lead times (NSE 0.49-0.67) and provided a quantification of prediction uncertainty. National Science FoundationNational Science Foundation (NSF) [1360415, 1343802]; USDAUnited States Department of Agriculture (USDA) [2012-67019-19434] We would like to acknowledge high-performance computing support from Yellowstone (http://n2t.net/ark:/85065/d7wd3xhc) provided by NCAR's Computational and Information Systems Laboratory, support from the National Science Foundation under award numbers 1360415 and 1343802, and funding support from the USDA under project number 2012-67019-19434. All data and methods used in this manuscript are available upon request. Public domain – authored by a U.S. government employee

Engineered denitrifying environments, such as constructed wetland, biofilters or bioreactors have been increasingly recognized as an efficient means to improve water quality via facilitated denitrification. These environments function by providing a carbon substrate that serves as an energy source for microbial denitrifiers to utilize during denitrification. However, under sulfate (SO42−) reducing redox potentials, methane (CH4), hydrogen sulfide (H2S) and methyl mercury production can result. Amorphous ferric hydroxide (Fe(OH)3) has been shown to poise soil systems above the redox potential (Eh) of microbial SO42 reduction. This mineral was added in a denitrification column experiment in order to assess the ability of amorphous Fe(OH)3 to prevent or mitigate microbiological SO42− reduction. Given biochar’s ability to enhance microbial activity, this experiment was also designed to determine whether biochar enhances microbiological SO42− and NO3− − N reduction, as well as mitigate N2O emissions. Four denitrification substrate mixtures were evaluated in a laboratory column experiment: woodchips only (WC), woodchips + 15% (v/v) sand (Sand), woodchips + 15%(v/v) hardwood 26 mesh biochar (Biochar) and woodchips + 15%(v/v) hardwood 26 mesh biochar + 15 % (v/v) amorphous Fe(OH)3 (Fe). Columns were sampled for NO3− − N, N2O and SO42− at an initial interval of 3 hrs gradually decreasing to 12 h over a 120 h sampling period. Analysis of the data show that Biochar enhanced SO42− reduction (to sulfide) by 85% compared to the initial concentration while Fe largely prevented SO42− reduction, displaying a 18% reduction, which was not statistically significant. These results indicated that Fe(OH)3 addition significantly mitigated SO42− reduction. Similar to Fe, the Sand column exhibited lower microbially reduced SO42 than Biochar, at 20%, but this schema resulted in significantly higher N2O production. The WC treatment reduced SO42 by 47%, but also produced more N2O than other treatments. As expected, both treatments with biochar (Biochar and Fe) accelerated NO3− − N removal while reducing the production of N2O, and there was no discernable difference between the two, indicating that the addition of Fe(OH)3 to the system does not negatively impact performance. These results indicate that biochar, while enhancing denitrification, significantly increases microbiological reduction of SO42−, a precursor to methyl mercury production in. Thus, methyl mercury production may increase with increased biochar addition unless amorphous Fe(OH)3 is added to the substrate matrix.

Summary Spatially distributed nonpoint source (NPS) pollution indices are used to identify areas in a watershed where potential pollutant loading coincides with runoff generating areas. However, most such indices either ignore the degree of hydrologic connectivity to the stream network or they estimate it based simply on the distance of the pollution generating area from an open channel. We propose an NPS pollution index based on runoff travel times from saturated variable source areas (VSAs) to the natural stream network as a means for including hydrologic connectivity between source areas and streams. Although this method could be generalized to any pollutant transported by storm runoff, here we focus on phosphorus and refer to the index as the travel-time phosphorus index (TTPI). The TTPI was applied to a 38 km2 agricultural watershed in central New York and shown to yield realistic, spatially explicit predictions of critical phosphorus loading areas and routing pathways. One interesting finding is the potential role of man-made drainage networks (e.g., road- or agricultural-ditches) in NPS pollution and the possibilities of targeting water quality protection practices around or within these networks. Because the technique is GIS-based, relatively simple to apply, uses readily available geospatial data, and the theoretical underpinnings are transparent, it can provide a useful screening tool for water resource managers charged with the identification and remediation of critical NPS pollution source areas.

Summary Roadside drainage networks can affect deleterious changes to watershed hydrology and water quality. However, very few studies have explored these effects in agricultural settings where the potential for nonpoint source pollution and the implications of tighter hydrologic linkages between landscape and stream are more critical. In this study, we apply a GIS-based, spatially distributed hydrologic model under a variety of drainage scenarios to quantify the hydrologic effects of roadside ditches at multiple spatial scales. We also qualitatively investigate the relative impacts of ditches on phosphorus transport. Our principle findings indicate that roadside ditches: (i) substantially alter basin morphometry and natural flow pathways, (ii) increase peak and total event discharge and (iii) expedite the transport of agricultural pollutants, thereby short-circuiting natural degradation processes that would otherwise have mitigated their effects. These findings underscore how relatively fine-scale alterations to natural flow pathways can result in substantial impacts to watershed scale hydrology and water quality and may help to inform spatially-targeted water resource management decisions and future modeling efforts.

Summary This study uses an instrumented (trenched) 0.5 ha hillslope in the southern tier of New York State, USA, to provide new data and insights on how variable source areas and associated flow pathways form and combine to connect rainfall with downstream water flows across a hillslope. Measurements of water fluxes in the trench, upslope water table dynamics, surface and bedrock topography, and isotopic and geochemical tracers have been combined for a four-dimensional (space–time) characterization of subsurface storm flow responses. During events with dry antecedent conditions infiltrating rainwater was found to percolate through a prevailing fragipan layer to deeper soil layers, with much (33–71%) of the total discharge of the hillslope originating from deeper water flow below the fragipan. During storm events with wet antecedent conditions and large rainfall amounts, shallow lateral flow of event and pre-event water above the fragipan occurred and was one magnitude greater than the deeper water flow contribution. Spatial surface and subsurface water quality observations indicate that water from a distance of up to 56 m contributed runoff from the hillslope during storm events. In addition, mobilization of total dissolved phosphorus (TDP) with subsurface flow played a greater role than with overland or near-surface flow. During all events TDP loads were highest in the total discharge during peak flows (8–11.5 kg ha −1 d −1 ), except during the largest storm event, when TDP concentrations were highly diluted. These results have implications for strategies to protect streams and other downstream water recipients from waterborne nutrient and pollutant loading.

Planners advocate best management practices (BMPs) to reduce loss of sediment and nutrients in agricultural areas. However, the scientific community lacks tools that use readily available data to investigate the relationships between BMPs and their spatial locations and water quality. In rural, humid regions where runoff is associated with saturation-excess processes from variable source areas (VSAs), BMPs are potentially most effective when they are located in areas that produce the majority of the runoff. Thus, two critical elements necessary to predict the water quality impact of BMPs include correct identification of VSAs and accurate predictions of nutrient reduction due to particular BMPs. The objective of this research was to determine the effectiveness of BMPs using the Variable Source Loading Function (VSLF) model, which captures the spatial and temporal evolutions of VSAs in the landscape. Data from a long-term monitoring campaign on a 164-ha farm in the New York City source watersheds in the Catskills Mountains of New York state were used to evaluate the effectiveness of a range of BMPs. The data spanned an 11-year period over which a suite of BMPs, including a nutrient management plan, riparian buffers, filter strips and fencing, was installed to reduce phosphorus (P) loading. Despite its simplicity, VSLF predicted the spatial distribution of runoff producing areas well. Dissolved P reductions were simulated well by using calibrated reduction factors for various BMPs in the VSLF model. Total P losses decreased only after cattle crossings were installed in the creek. The results demonstrated that BMPs, when sited with respect to VSAs, reduce P loss from agricultural watersheds, providing useful information for targeted water quality management.

Many water quality models use some form of the Natural Resources Conservation Services (formerly Soil Conservation Service) curve number (CN) equation to predict storm runoff in ways that implicitly assume an infiltration-excess response to rainfall. Because of this, these models may fail to predict variable source areas (VSAs) correctly, i.e. where runoff is typically generated in rural, humid regions. In this study, the Soil and Water Assessment Tool (SWAT) model was re-conceptualized to distribute overland flow in ways consistent with VSA hydrology by modifying how the CN and available water content were defined; the new modeling approach is called SWAT-VSA. Both SWAT and SWAT-VSA were applied to a sub-watershed in the Cannonsville basin in upstate New York to compare model predictions of integrated and distributed responses, including surface runoff, shallowly perched water table depth, and stream phosphorus loads against direct measures. Event runoff was predicted similarly well for SWAT-VSA and SWAT. However, the distribution of shallowly perched water table depth was predicted better by SWAT-VSA and it is this shallow groundwater that governs VSAs. Event based dissolved phosphorus export from the watershed was also predicted better by SWAT-VSA, presumably because the distribution of runoff source areas was better predicted particularly from areas where manure was applied. This has important consequences for using models to evaluate and guide watershed management because correctly predicting where runoff is generated is critical to locating best management practices to control non-point source pollution.