Twenty‐One‐Year Simulation of Chesapeake Bay Water Quality Using the CE‐QUAL‐ICM Eutrophication Model

The CE‐QUAL‐ICM (Corps of Engineers Integrated Compartment Water Quality Model) eutrophication model was applied in a 21‐year simulation of Chesapeake Bay water quality, 1985‐2005. The eutrophication model is part of a larger model package and is forced, in part, by models of atmospheric deposition, watershed flows and loads, and hydrodynamics. Results from the model are compared with observations in multiple formats including time series plots, cumulative distribution plots, and statistical summaries. The model indicates only one long‐term trend in computed water quality: light attenuation deteriorates circa 1993 through the end of the simulation. The most significant result is the influence of physical processes, notably stratification and associated effects (e.g., anoxic volume), on computed water quality. Within the application period, physical effects are more important determinants of year‐to‐year variability in computed water quality than external loads.     

Suspended-Sediment Concentration Regimes for Two Biological Reference Streams in Middle Tennessee1

Diehl, Timothy H. and William J. Wolfe, 2010. Suspended-Sediment Concentration Regimes for Two Biological Reference Streams in Middle Tennessee. Journal of the American Water Resources Association (JAWRA) 46(4): 824-837. DOI: 10.1111/j.1752-1688.2010.00460.x Abstract: Temporal patterns of suspended-sediment concentration (SSC) duration and frequency (SSC regimes) were characterized and compared with biological impairment thresholds for two headwater streams in the Western Highland Rim of Tennessee. The SSC regimes were plotted as curves showing concentrations and durations of the annual longest and tenth-longest SSC excursions above 18 concentrations for water years 2005-2008 in Copperas Branch and water years 2006 and 2008 in Kelley Creek. Both streams have fish communities remarkably diverse for their small drainage basin areas (420 and 565 ha, respectively), and represent biological reference conditions with respect to SSC. SSC-regime curves were similar for the two sites across water years. The measured SSC regimes reached or exceeded published experimentally based SSC impairment thresholds and plotted below a proposed long-term SSC reference regime for the Interior Plateau ecoregion (Ecoregion 71), suggesting that neither the experimentally based thresholds nor the proposed SSC reference regime adequately reflect the relation between SSC and biological impairment for Western Highland Rim headwater streams. The SSC regimes of the two study streams were similar to the estimated SSC regime of an unimpaired East Tennessee trout stream. Additional field studies are needed to describe SSC regimes in streams of varying basin scale, level of impairment, and region.     

A Regional Modeling Framework of Phosphorus Sources and Transport in Streams of the Southeastern United States1

García, Ana María, Anne B. Hoos, and Silvia Terziotti, 2011. A Regional Modeling Framework of Phosphorus Sources and Transport in Streams of the Southeastern United States. Journal of the American Water Resources Association (JAWRA) 47(5):991‐1010. DOI: 10.1111/j.1752‐1688.2010.00517.x Abstract: We applied the SPARROW model to estimate phosphorus transport from catchments to stream reaches and subsequent delivery to major receiving water bodies in the Southeastern United States (U.S.). We show that six source variables and five land‐to‐water transport variables are significant (p < 0.05) in explaining 67% of the variability in long‐term log‐transformed mean annual phosphorus yields. Three land‐to‐water variables are a subset of landscape characteristics that have been used as transport factors in phosphorus indices developed by state agencies and are identified through experimental research as influencing land‐to‐water phosphorus transport at field and plot scales. Two land‐to‐water variables – soil organic matter and soil pH – are associated with phosphorus sorption, a significant finding given that most state‐developed phosphorus indices do not explicitly contain variables for sorption processes. Our findings for Southeastern U.S. streams emphasize the importance of accounting for phosphorus present in the soil profile to predict attainable instream water quality. Regional estimates of phosphorus associated with soil‐parent rock were highly significant in explaining instream phosphorus yield variability. Model predictions associate 31% of phosphorus delivered to receiving water bodies to geology and the highest total phosphorus yields in the Southeast were catchments with already high background levels that have been impacted by human activity.     

Cropland Riparian Buffers throughout Chesapeake Bay Watershed: Spatial Patterns and Effects on Nitrate Loads Delivered to Streams

We used statistical models to provide the first empirical estimates of riparian buffer effects on the cropland nitrate load to streams throughout the Chesapeake Bay watershed. For each of 1,964 subbasins, we quantified the 1990 prevalence of cropland and riparian buffers. Cropland was considered buffered if the topographic flow path connecting it to a stream traversed a streamside forest or wetland. We applied a model that predicts stream nitrate concentration based on physiographic province and the watershed proportions of unbuffered and buffered cropland. We used another model to predict annual streamflow based on precipitation and temperature, and then multiplied the predicted flows and concentrations to estimate 1990 annual nitrate loads. Across the entire Chesapeake watershed, croplands released 92.3 Gg of nitrate nitrogen, but 19.8 Gg of that was removed by riparian buffers. At most, 29.4 Gg more might have been removed if buffer gaps were restored so that all cropland was buffered. The other 43.1 Gg of cropland load cannot be addressed with riparian buffers. The Coastal Plain physiographic province provided 52% of the existing buffer reduction of Bay‐wide nitrate loads and 36% of potential additional removal from buffer restoration in cropland buffer gaps. Existing and restorable nitrate removal in buffers were lower in the other three major provinces because of less cropland, lower buffer prevalence, and lower average buffer nitrate removal efficiency.     

The Shallow‐Water Component of the Chesapeake Bay Environmental Model Package

The shallow‐water component of the Chesapeake Bay Environmental Model Package emphasizes the regions of the system inside the 2‐m depth contour. The model of these regions is unified with the system‐wide model but places emphasis on locally significant components and processes, notably submerged aquatic vegetation (SAV), sediment resuspension, and their interaction with light attenuation (Ke). The SAV model is found to be most suited for computing the equilibrium distribution of perennial species. Addition of plant structure and propagation are recommended to improve representation of observed trends in SAV area. Two approaches are taken to examining shallow‐water Ke. The first compares observed and computed differences between deep‐ and shallow‐water Ke. No consistent difference in observations is noted. In the preponderance of regions examined, computed shallow‐water Ke exceeds computed deep‐water Ke. The second approach directly compares Ke measured in shallow water with modeled results. Model values are primarily lower than observed, in contrast to results in deep water where model values exceed observed. The shortfall in computed Ke mirrors a similar shortfall in computed suspended solids. Improved model representation of Ke requires process‐based investigations into suspended solids dynamics as well as increased model resolution in shallow‐water regions.     

Factors Affecting Stream Nutrient Loads: A Synthesis of Regional SPARROW Model Results for the Continental United States1

Preston, Stephen D., Richard B. Alexander, Gregory E. Schwarz, and Charles G. Crawford, 2011. Factors Affecting Stream Nutrient Loads: A Synthesis of Regional SPARROW Model Results for the Continental United States. Journal of the American Water Resources Association (JAWRA) 47(5):891‐915. DOI: 10.1111/j.1752‐1688.2011.00577.x Abstract: We compared the results of 12 recently calibrated regional SPARROW (SPAtially Referenced Regressions On Watershed attributes) models covering most of the continental United States to evaluate the consistency and regional differences in factors affecting stream nutrient loads. The models – 6 for total nitrogen and 6 for total phosphorus – all provide similar levels of prediction accuracy, but those for major river basins in the eastern half of the country were somewhat more accurate. The models simulate long‐term mean annual stream nutrient loads as a function of a wide range of known sources and climatic (precipitation, temperature), landscape (e.g., soils, geology), and aquatic factors affecting nutrient fate and transport. The results confirm the dominant effects of urban and agricultural sources on stream nutrient loads nationally and regionally, but reveal considerable spatial variability in the specific types of sources that control water quality. These include regional differences in the relative importance of different types of urban (municipal and industrial point vs. diffuse urban runoff) and agriculture (crop cultivation vs. animal waste) sources, as well as the effects of atmospheric deposition, mining, and background (e.g., soil phosphorus) sources on stream nutrients. Overall, we found that the SPARROW model results provide a consistent set of information for identifying the major sources and environmental factors affecting nutrient fate and transport in United States watersheds at regional and subregional scales.     

Application of SPARROW Modeling to Understanding Contaminant Fate and Transport from Uplands to Streams

Understanding spatial variability in contaminant fate and transport is critical to efficient regional water‐quality restoration. An approach to capitalize on previously calibrated spatially referenced regression (SPARROW) models to improve the understanding of contaminant fate and transport was developed and applied to the case of nitrogen in the 166,000 km² Chesapeake Bay watershed. A continuous function of four hydrogeologic, soil, and other landscape properties significant (α = 0.10) to nitrogen transport from uplands to streams was evaluated and compared among each of the more than 80,000 individual catchments (mean area, 2.1 km²) in the watershed. Budgets (including inputs, losses or net change in storage in uplands and stream corridors, and delivery to tidal waters) were also estimated for nitrogen applied to these catchments from selected upland sources. Most (81%) of such inputs are removed, retained, or otherwise processed in uplands rather than transported to surface waters. Combining SPARROW results with previous budget estimates suggests 55% of this processing is attributable to denitrification, 23% to crop or timber harvest, and 6% to volatilization. Remaining upland inputs represent a net annual increase in landscape storage in soils or biomass exceeding 10 kg per hectare in some areas. Such insights are important for planning watershed restoration and for improving future watershed models.     

A Multi‐Agency Nutrient Dataset Used to Estimate Loads,Improve Monitoring Design,and Calibrate Regional Nutrient SPARROW Models1

Saad, David A., Gregory E. Schwarz, Dale M. Robertson, and Nathaniel L. Booth, 2011. A Multi‐Agency Nutrient Dataset Used to Estimate Loads, Improve Monitoring Design, and Calibrate Regional Nutrient SPARROW Models. Journal of the American Water Resources Association (JAWRA) 47(5):933‐949. DOI: 10.1111/j.1752‐1688. 2011.00575.x Abstract: Stream‐loading information was compiled from federal, state, and local agencies, and selected universities as part of an effort to develop regional SPAtially Referenced Regressions On Watershed attributes (SPARROW) models to help describe the distribution, sources, and transport of nutrients in streams throughout much of the United States. After screening, 2,739 sites, sampled by 73 agencies, were identified as having suitable data for calculating long‐term mean annual nutrient loads required for SPARROW model calibration. These sites had a wide range in nutrient concentrations, loads, and yields, and environmental characteristics in their basins. An analysis of the accuracy in load estimates relative to site attributes indicated that accuracy in loads improve with increases in the number of observations, the proportion of uncensored data, and the variability in flow on observation days, whereas accuracy declines with increases in the root mean square error of the water‐quality model, the flow‐bias ratio, the number of days between samples, the variability in daily streamflow for the prediction period, and if the load estimate has been detrended. Based on compiled data, all areas of the country had recent declines in the number of sites with sufficient water‐quality data to compute accurate annual loads and support regional modeling analyses. These declines were caused by decreases in the number of sites being sampled and data not being entered in readily accessible databases.     

Development of the Chesapeake Bay Watershed Total Maximum Daily Load Allocation

Nutrient load allocations and subsequent reductions in total nitrogen and phosphorus have been applied in the Chesapeake watershed since 1992 to reduce hypoxia and to restore living resources. In 2010, sediment allocations were established to augment nutrient allocations supporting the submerged aquatic vegetation resource. From the initial introduction of nutrient allocations in 1992 to the present, the allocations have become more completely applied to all areas and loads in the watershed and have also become more rigorously assessed and tracked. The latest 2010 application of nutrient and sediment allocations were made as part of the Chesapeake Bay total maximum daily load and covered all six states of the Chesapeake watershed. A quantitative allocation process was developed that applied principles of equity and efficiency in the watershed, while achieving all tidal water quality standards through an assessment of equitable levels of effort in reducing nutrients and sediments. The level of effort was determined through application of two key watershed scenarios: one where no action was taken in nutrient control and one where maximum nutrient control efforts were applied. Once the level of effort was determined for different jurisdictions, the overall load reduction was set watershed‐wide to achieve dissolved oxygen water quality standards. Further adjustments were made to the allocation to achieve the James River chlorophyll‐a standard.     

Response of Nitrogen Loading to the Chesapeake Bay to Source Reduction and Land Use Change Scenarios: A SPARROW‐Informed Analysis

In response to concerns regarding the health of streams and receiving waters, the United States Environmental Protection Agency established a total maximum daily load for nitrogen in the Chesapeake Bay watershed for which practices must be in place by 2025 resulting in an expected 25% reduction in load from 2009 levels. The response of total nitrogen (TN) loads delivered to the Bay to nine source reduction and land use change scenarios was estimated using a Spatially Referenced Regression on Watershed Attributes model. The largest predicted reduction in TN load delivered to the Bay was associated with a scenario in which the mass of TN as fertilizer applied to agricultural lands was decreased. A 25% decrease in the mass of TN applied as fertilizer resulted in a predicted reduction in TN loading to the Bay of 11.3%, which was 2.5–5 times greater than the reductions predicted by other scenarios. Eliminating fertilizer application to all agricultural land in the watershed resulted in a predicted reduction in TN load to the Bay of 45%. It was estimated that an approximate 25% reduction in TN loading to the Bay could be achieved by eliminating fertilizer applied to the 7% of subwatersheds contributing the greatest fertilizer‐sourced TN loads to the Bay. These results indicate that management strategies aimed at decreasing loading from a small number of subwatersheds may be effective for reducing TN loads to the Bay, and similar analyses are possible in other watersheds.     

The Role of Headwater Streams in Downstream Water Quality1

Abstract: Knowledge of headwater influences on the water‐quality and flow conditions of downstream waters is essential to water‐resource management at all governmental levels; this includes recent court decisions on the jurisdiction of the Federal Clean Water Act (CWA) over upland areas that contribute to larger downstream water bodies. We review current watershed research and use a water‐quality model to investigate headwater influences on downstream receiving waters. Our evaluations demonstrate the intrinsic connections of headwaters to landscape processes and downstream waters through their influence on the supply, transport, and fate of water and solutes in watersheds. Hydrological processes in headwater catchments control the recharge of subsurface water stores, flow paths, and residence times of water throughout landscapes. The dynamic coupling of hydrological and biogeochemical processes in upland streams further controls the chemical form, timing, and longitudinal distances of solute transport to downstream waters. We apply the spatially explicit, mass‐balance watershed model SPARROW to consider transport and transformations of water and nutrients throughout stream networks in the northeastern United States. We simulate fluxes of nitrogen, a primary nutrient that is a water‐quality concern for acidification of streams and lakes and eutrophication of coastal waters, and refine the model structure to include literature observations of nitrogen removal in streams and lakes. We quantify nitrogen transport from headwaters to downstream navigable waters, where headwaters are defined within the model as first‐order, perennial streams that include flow and nitrogen contributions from smaller, intermittent and ephemeral streams. We find that first‐order headwaters contribute approximately 70% of the mean‐annual water volume and 65% of the nitrogen flux in second‐order streams. Their contributions to mean water volume and nitrogen flux decline only marginally to about 55% and 40% in fourth‐ and higher‐order rivers that include navigable waters and their tributaries. These results underscore the profound influence that headwater areas have on shaping downstream water quantity and water quality. The results have relevance to water‐resource management and regulatory decisions and potentially broaden understanding of the spatial extent of Federal CWA jurisdiction in U.S. waters.     

Development and Application of the 2010 Chesapeake Bay Watershed Total Maximum Daily Load Model

The Phase 5.3 Watershed Model simulates the Chesapeake watershed land use, river flows, and the associated transport and fate of nutrient and sediment loads to the Chesapeake Bay. The Phase 5.3 Model is the most recent of a series of increasingly refined versions of a model that have been operational for more than two decades. The Phase 5.3 Model, in conjunction with models of the Chesapeake airshed and estuary, provides estimates of management actions needed to protect water quality, achieve Chesapeake water quality standards, and restore living resources. The Phase 5.3 Watershed Model tracks nutrient and sediment load estimates of the entire 166,000 km² watershed, including loads from all six watershed states. The creation of software systems, input datasets, and calibration methods were important aspects of the model development process. A community model approach was taken with model development and application, and the model was developed by a broad coalition of model practitioners including environmental engineers, scientists, and environmental managers. Among the users of the Phase 5.3 Model are the Chesapeake watershed states and local governments, consultants, river basin commissions, and universities. Development and application of the model are described, as well as key scenarios ranging from high nutrient and sediment load conditions if no management actions were taken in the watershed, to low load estimates of an all‐forested condition.     

Spatial and Temporal Patterns of Low Streamflow and Precipitation Changes in the Chesapeake Bay Watershed

Spatial and temporal patterns in low streamflows were investigated for 183 streamgages located in the Chesapeake Bay Watershed for the period 1939–2013. Metrics that represent different aspects of the frequency and magnitude of low streamflows were examined for trends: (1) the annual time series of seven‐day average minimum streamflow, (2) the scaled average deficit at or below the 2% mean daily streamflow value relative to a base period between 1939 and 1970, and (3) the annual number of days below the 2% threshold. Trends in these statistics showed spatial cohesion, with increasing low streamflow volume at streamgages located in the northern uplands of the Chesapeake Bay Watershed and decreasing low streamflow volume at streamgages in the southern part of the watershed. For a small subset of streamgages (12%), conflicting trend patterns were observed between the seven‐day average minimum streamflow and the below‐threshold time series and these appear to be related to upstream diversions or the influence of reservoir‐influenced streamflows in their contributing watersheds. Using multivariate classification techniques, mean annual precipitation and fraction of precipitation falling as snow appear to be broad controls of increasing and decreasing low‐flow trends. Further investigation of seasonal precipitation patterns shows summer rainfall patterns, driven by the Atlantic Multidecadal Oscillation, as the main driver of low streamflows in the Chesapeake Bay Watershed.     

Development and Use of a Sedimentation Risk Index for Unpaved Road‐Stream Crossings in the Choctawhatchee Watershed1

Abstract: Unpaved road‐stream crossings increase sediment yields in streams and alter channel morphology and stability. Before restoration and sedimentation reduction strategies can be implemented, a priority listing of unpaved road‐stream crossings must be created. The objectives of this study were to develop a sedimentation risk index (SRI) for unpaved road‐stream crossings and to prioritize 125 sites in the Choctawhatchee watershed (southeastern Alabama) using this model. Field surveys involved qualitative and quantitative observations of 73 metrics related to waterway conditions, crossing structures, road approaches, and roadside soil erosion. The road‐stream crossing risk analyses involved elimination of candidate metrics based on redundancy, skewness, lack of data, professional judgment, lack of nonzero values, unbalanced box plots, and limited ranges of values. A final selection of 12 metrics formed the SRI and weighed factors involving soil erodibility, road sedimentation abatement features, and stream morphology alteration. The SRI was organized into narrative categories (excellent, good, fair, poor, and very poor) based on the distribution of scores. No excellent sites (scores ≥55) were found in this study, 17 (20.7%) were good (low sedimentation risk), 37 (45.1%) were fair (moderate sedimentation risk), 26 (31.7%) were poor (high sedimentation risk), and two (2.5%) were very poor (high sedimentation risk). There was no significant difference in SRI scores among crossing structure type (round culverts, box culverts, and bridges) (H = 4.31, df = 2, p = 0.058). A future study of the Choctawhatchee watershed involving the same study sites could assess the success of restoration plans and activities based on site score improvement or decline.     

Mapping Three‐Dimensional Water‐Quality Data in the Chesapeake Bay Using Geostatistics1

Abstract:  Data interpretation and visualization software tools with geostatistical capabilities were adapted, customized, and tested to assist the Chesapeake Bay Program in improving its water‐quality modeling protocols. Tools were required to interpolate, map, and visualize three‐dimensional (3D) water‐quality data, with the capability to determine estimation errors. Components of the software, originally developed for ground‐water modeling, were customized for application in estuaries. Additional software components were developed for retrieval, and for pre‐ and post‐ processing of data. The Chesapeake Bay Program uses the 3D mapped data for input to the Bay water‐quality model that projects the future health of the Bay and its tidal tributary system. In determining water‐quality attainment criteria, 3D kriging estimation errors are needed as a statistical measure of uncertainty. Furthermore, given the high cost of installing and operating new monitoring stations, geostatistical techniques can assist the Chesapeake Bay Program in the identification of suitable data collection locations. Following the evaluation, selection, and development of the software components phase, 3D ordinary kriging techniques with directional semi‐variograms to account for anisotropy were successfully demonstrated for mapping 3D fixed station water‐quality data, such as dissolved oxygen and salinity. Additionally, an improved delineation tool was implemented to simulate the upper and lower pycnocline boundary surfaces allowing the segregation of the interpolated 3D data into three separate zones for a better characterization of the pycnocline layer.     

A Biological Assessment of Streams in the Eastern United States Using a Predictive Model for Macroinvertebrate Assemblages1

Abstract: A predictive model (RIVPACS‐type) for benthic macroinvertebrates was constructed to assess the biological condition of 1,087 streams sampled throughout the eastern United States from 1993‐2003 as part of the U.S. Geological Survey’s National Water‐Quality Assessment Program. A subset of 338 sites was designated as reference quality, 28 of which were withheld from model calibration and used to independently evaluate model precision and accuracy. The ratio of observed (O) to expected (E) taxa richness was used as a continuous measure of biological condition, and sites with O/E values <0.8 were classified as biologically degraded. Spatiotemporal variability of O/E values was evaluated with repeated annual and within‐site samples at reference sites. Values of O/E were regressed on a measure of urbanization in three regions and compared among streams in different land‐use settings. The model accurately predicted the expected taxa at validation sites with high precision (SD = 0.11). Within‐site spatial variability in O/E values was much larger than annual and among‐site variation at reference sites and was likely caused by environmental differences among sampled reaches. Values of O/E were significantly correlated with basin road density in the Boston, Massachusetts (p < 0.001), Birmingham, Alabama (p = 0.002), and Green Bay, Wisconsin (p = 0.034) metropolitan areas, but the strength of the relations varied among regions. Urban streams were more depleted of taxa than streams in other land‐use settings, but larger networks of riparian forest appeared to mediate biological degradation. Taxa that occurred less frequently than predicted by the model were those known to be generally intolerant of a variety of anthropogenic stressors.     

Oregon Hydrologic Landscapes: A Classification Framework1

Wigington, Parker J., Jr., Scott G. Leibowitz, Randy L. Comeleo, and Joseph L. Ebersole, 2012. Oregon Hydrologic Landscapes: A Classification Framework. Journal of the American Water Resources Association (JAWRA) 1‐20. DOI: 10.1111/jawr.12009 Abstract: There is a growing need for hydrologic classification systems that can provide a basis for broad‐scale assessments of the hydrologic functions of landscapes and watersheds and their responses to stressors such as climate change. We developed a hydrologic landscape (HL) classification approach that describes factors of climate‐watershed systems that control the hydrologic characteristics of watersheds. Our assessment units are incremental watersheds (i.e., headwater watersheds or areas draining directly into stream reaches). Major components of the classification include indices of annual climate, climate seasonality, aquifer permeability, terrain, and soil permeability. To evaluate the usefulness of our approach, we identified 30 rivers with long‐term streamflow‐gauging records and without major diversions and impoundments. We used statistical clustering to group the streams based on the shapes of their annual hydrographs. Comparison of the streamflow clusters and HL distributions within river basin clusters shows that the Oregon HL approach has the ability to provide insights about the expected hydrologic behavior of HLs and larger river basins. The Oregon HL approach has potential to be a useful framework for comparing hydrologic attributes of streams and rivers in the Pacific Northwest.     

Sources and Delivery of Nutrients to the Northwestern Gulf of Mexico from Streams in the South‐Central United States1

Rebich, Richard A., Natalie A. Houston, Scott V. Mize, Daniel K. Pearson, Patricia B. Ging, and C. Evan Hornig, 2011. Sources and Delivery of Nutrients to the Northwestern Gulf of Mexico From Streams in the South‐Central United States. Journal of the American Water Resources Association (JAWRA) 47(5):1061‐1086. DOI: 10.1111/j.1752‐1688.2011.00583.x Abstract: SPAtially Referenced Regressions On Watershed attributes (SPARROW) models were developed to estimate nutrient inputs [total nitrogen (TN) and total phosphorus (TP)] to the northwestern part of the Gulf of Mexico from streams in the South‐Central United States (U.S.). This area included drainages of the Lower Mississippi, Arkansas‐White‐Red, and Texas‐Gulf hydrologic regions. The models were standardized to reflect nutrient sources and stream conditions during 2002. Model predictions of nutrient loads (mass per time) and yields (mass per area per time) generally were greatest in streams in the eastern part of the region and along reaches near the Texas and Louisiana shoreline. The Mississippi River and Atchafalaya River watersheds, which drain nearly two‐thirds of the conterminous U.S., delivered the largest nutrient loads to the Gulf of Mexico, as expected. However, the three largest delivered TN yields were from the Trinity River/Galveston Bay, Calcasieu River, and Aransas River watersheds, while the three largest delivered TP yields were from the Calcasieu River, Mermentau River, and Trinity River/Galveston Bay watersheds. Model output indicated that the three largest sources of nitrogen from the region were atmospheric deposition (42%), commercial fertilizer (20%), and livestock manure (unconfined, 17%). The three largest sources of phosphorus were commercial fertilizer (28%), urban runoff (23%), and livestock manure (confined and unconfined, 23%).     

The Regionalization of National‐Scale SPARROW Models for Stream Nutrients1

Schwarz, Gregory E., Richard B. Alexander, Richard A. Smith, and Stephen D. Preston, 2011. The Regionalization of National‐Scale SPARROW Models for Stream Nutrients. Journal of the American Water Resources Association (JAWRA) 47(5):1151‐1172. DOI: 10.1111/j.1752‐1688.2011.00581.x Abstract: This analysis modifies the parsimonious specification of recently published total nitrogen (TN) and total phosphorus (TP) national‐scale SPAtially Referenced Regressions On Watershed attributes models to allow each model coefficient to vary geographically among three major river basins of the conterminous United States. Regionalization of the national models reduces the standard errors in the prediction of TN and TP loads, expressed as a percentage of the predicted load, by about 6 and 7%. We develop and apply a method for combining national‐scale and regional‐scale information to estimate a hybrid model that imposes cross‐region constraints that limit regional variation in model coefficients, effectively reducing the number of free model parameters as compared to a collection of independent regional models. The hybrid TN and TP regional models have improved model fit relative to the respective national models, reducing the standard error in the prediction of loads, expressed as a percentage of load, by about 5 and 4%. Only 19% of the TN hybrid model coefficients and just 2% of the TP hybrid model coefficients show evidence of substantial regional specificity (more than ±100% deviation from the national model estimate). The hybrid models have much greater precision in the estimated coefficients than do the unconstrained regional models, demonstrating the efficacy of pooling information across regions to improve regional models.