Flow Duration Curves (FDCs) provide a statistical relationship between the magnitude and frequency of streamflow in a watershed. The shape and scale of an FDC are dependent upon the dynamic relationship between streamflow regimes and the climatic, geological, topographical, and other environmental or anthropogenic attributes of the watershed. While Machine Learning (ML) models can provide enhanced predictive performance in ungaged watersheds over conventional approaches, validating the models’ comprehension of the underlying hydrologic processes is imperative for any stakeholder involvement. In this study, we employed Random Forest (RF) regression on individual Exceedance Percentiles (EPs) and slope of the FDCs for a large sample of watersheds in the contiguous United States. Then, we explored the interactions between watershed attributes and FDCs using SHapley Additive exPlanations (SHAP) to divulge the local (watershed scale), regional, and global (model scale) influence of attributes and their dependence structures. Results indicate that climate attributes (precipitation and aridity index) were the preeminent drivers in predicting FDCs across all quantiles, primarily affecting the scale of FDCs, followed by the baseflow index and geologic attributes that highly influence the low flow regime and control the shape of FDCs. The dominant controls of other watershed attributes, including precipitation seasonality, % snow, and elevation, vary between EPs and regions that were readily discernible in their SHAP values.

Placing appropriate conservation practices in critical source areas of pollutants can benefit stream health in karst environments susceptible to agricultural pollution. Watershed models, such as the Soil and Water Assessment Tool+ (SWAT+), can optimize both practice selection and placement across the landscape, but to our knowledge no studies have tested this model in karst pasturelands. Our goal was to understand pollutant dynamics in pasturelands with karst topography in southwest Virginia, United States. We built a SWAT+ model to predict streamflow and pollutant loads for four watersheds. SWAT+ predicted that streamflow estimates were most affected by setting the available water capacity to zero and increasing the hydraulic conductivity of the soil, revealing that water, and associated pollutants, move primarily through subsurface pathways. Predicted sediment yield was negatively associated with agricultural land cover and was strongly influenced by channel erodibility-indicating that the predominant sediment source may be streambanks. Therefore, practices that stabilize stream banks (e.g., fencing cattle out of streams) may be most effective at reducing sediment loads. The model unsatisfactorily predicted total nitrogen and total phosphorus loads. The utility of SWAT+ for cattle grazing operations in karst regions could be improved by more accurately representing the effects of cattle grazing on streambank erosion and the dynamic subsurface movement of pollutants.

The flow duration curve (FDC) is a fundamental tool in hydrology representing the cumulative distribution function of streamflow with time. As such, it is a valuable indicator of watershed performance, being controlled by rainfall, runoff, and storage. The shape of the FDC represents the complex interactions of climate, geology, topography, vegetation and human activities. This study is a novel effort to predict the FDC using a continuous simulation daily time step model calibrated to mean annual flow for CONUS (Contiguous United States of America). Model performance was evaluated using long-term USGS streamflow records from relatively undisturbed gages, while sites substantially affected by major reservoirs and water diversions were excluded. Model outputs are compared regionally to modeled FDCs and gage data. Output is compiled by NHDPlus reach for CONUS. Evaluations based on representative FDC quantiles demonstrate the model’s ability to simulate high, medium, and low flows with average Nash-Sutcliffe Efficiency (NSE) values of 0.69, 0.84, and 0.56, respectively. The significant advantage of employing a continuous simulation model lies in its ability to simulate diverse climate and land use scenarios, offering valuable insights beyond historical data limitations.

The Soil and Water Assessment Tool (SWAT) includes plant parameters for only a few fruits and vegetables. Testing SWAT parameters for these crops remains very limited, including the need for evaluation of table food crops in historically low production areas. This study tests SWAT simulations for fruits and vegetables in the U.S. Western Corn Belt, to advance model evaluation in a data-limited region. We focus on 24 crops, integrating parameters from related models, expert opinions, and expected versus simulated yields. Parameters for 15 crops in the SWAT database were revised and nine crops were added. The resulting models achieved yield simulation errors within ±37%, a reasonable level of accuracy for underrepresented crops, and statistical tests further supported the robustness of the simulations. Future climate scenario analyses also revealed potential yield impacts and emphasized the need for adaptation strategies. Overall, this study helps fill important gaps in SWAT applications for fruits and vegetables and supports efforts to expand local table food production and crop diversification, although further testing is needed across additional climatic regions and longer crop yield time series.

Understanding the influence of hydrologic processes on nutrient transport remains a critical challenge in agricultural watersheds, where nitrate (N) and phosphorus (P) often exhibit contrasting export patterns. In this study, we developed a coupled surface–subsurface modeling framework using SWAT + integrated with the gwflow module to simulate hydrologic fluxes and nutrient transport in the Choptank River watershed, USA. An unstructured (quadtree) groundwater grid was implemented in the SWAT + gwflow module to improve representation of groundwater gradients near tile drains and channels. The model was calibrated and tested using the iterative ensemble smoother (iES) and evaluated against streamflow, groundwater head, and annual nutrient loading observations. The Morris global sensitivity analysis method was applied to identify and rank the dominant hydrologic controls on streamflow, groundwater dynamics, and nutrient transport.The coupled SWAT + gwflow framework identified distinct hydrologic controls governing streamflow, groundwater dynamics, and nutrient transport. Streamflow is primarily controlled by surface runoff generation, soil water retention, and channel conveyance, whereas groundwater dynamics are governed by aquifer properties and drainage mechanisms. The sensitivity analysis indicates that simulated nitrate transport is primarily influenced by soil nutrient processes and subsurface flow pathways (e.g., percolation and groundwater connectivity), whereas simulated phosphorus transport appears to be more strongly associated with fertilizer inputs and surface/aquatic processes, reflecting event-based mobilization through runoff and sediment transport. Basin-scale flux analysis further indicates that surface processes control short-term variability, whereas subsurface processes regulate longer-term (“legacy”) transport.These findings demonstrate the capability of the coupled surface–subsurface framework to diagnose hydrologic pathway partitioning governing nutrient transport within the critical zone. They also highlight limitations of single-objective calibration, as parameters controlling subsurface processes (relevant to nitrate) differ from those governing surface processes (relevant to phosphorus). This study provides a mechanistic framework linking hydrologic controls to nutrient-specific transport pathways and offers guidance for improved model calibration and targeted nutrient management in agricultural systems.

Precipitation forcing uncertainty can substantially influence parameter estimation and predictive performance in coupled surface–subsurface hydrologic models, yet practical strategies for explicitly representing this uncertainty during calibration remain limited. Here, precipitation forcing was parameterized to represent precipitation uncertainty and aspects of model misspecification, thereby reducing compensatory adjustment of hydrologic parameters during calibration.We constructed coupled surface–subsurface hydrologic models (SWAT+ gwflow) for two humid-region watersheds, the Winnebago River basin (Iowa, USA) and the Monocacy River basin (Maryland, USA), and employed the Iterative Ensemble Smoother (iES) for calibration using monthly mean streamflow and remotely sensed evapotranspiration (ET). Three calibration configurations were evaluated: streamflow-only, ET-only, and joint streamflow–ET calibration. Two precipitation parameterization schemes were evaluated: Precipitation-Variable (PV), where daily precipitation inputs were adjusted using time-varying precipitation multipliers, and Precipitation-Fixed (PF), where precipitation remained constant during calibration.In the Winnebago River basin, PV resulted in significant improvement in streamflow simulation during simultaneous streamflow–ET calibration, increasing NSE from 0.48 to 0.79 during calibration and from 0.25 to 0.81 during testing relative to PF, while also decreasing groundwater head bias at several monitoring wells. In contrast, the Monocacy River basin exhibited similarly strong streamflow performance under both PV and PF experiments, demonstrating that the benefits of precipitation parameterization were watershed-dependent rather than universal. Simulated ET performance showed only insignificant variation among PV and PF experiments in both watersheds, with R2 values generally near 0.90. Water-balance partitioning differed significantly between experiments, with PV notably decreasing simulated groundwater recharge in the Winnebago basin relative to PF, showing how precipitation forcing uncertainty influences coupled surface–subsurface hydrologic fluxes and affects parameter sensitivity patterns.The proposed framework extends precipitation-uncertainty analysis to coupled surface–subsurface hydrologic modeling, providing insight into how precipitation forcing uncertainty interacts with parameter estimation, sensitivity patterns, and water-balance partitioning across contrasting watershed settings, with implications for uncertainty-aware model calibration and interpretation.

Remote sensing vegetation indices such as the normalized difference vegetation index (NDVI) are widely used to monitor crop productivity and agroecosystem conditions. Yet, extracting full-field values across millions of fields from high resolution high frequency raster data remains computationally intensive at national scales, and is unrealistic for whole fields at the local level for continuous coverage of in-situ measurements. This study evaluates whether digital sampling (extracting NDVI from a limited number of stratified point locations within each field) can reliably approximate whole-field NDVI means while reducing processing demands. Using Sentinel-2 imagery from 2019 to 2021, we analyzed NDVI for ∼4.1 million crop-producing fields across the contiguous United States and compared whole-field NDVI with sampling schemes of 1 to 6 stratified points. Accuracy increased nonlinearly as points were added, with the largest improvement occurring when increasing from one to two points. Across the continental United States (CONUS), three points were generally sufficient to exceed an R2 of 0.90, though performance varied by Long-Term Agroecosystem Research (LTAR) network production system and individual agroecosystems. Integrated LTAR regions reached high accuracy with fewer points, whereas cropland and grazingland regions required more extensive sampling. Several agroecosystems showed lower coefficients of determination between point-based and field mean NDVI, particularly during the growing and senescence seasons. Digital sampling reduced computational requirements, using far fewer Earth Engine Compute Units (EECUs) than full-field extraction. These findings demonstrate that stratified digital sampling offers an efficient, scalable approach for monitoring vegetation dynamics across U.S. agroecosystems, enabling broad-scale indicator development while reducing computational time, energy use, and analytical costs.

Abstract. Methods for soil test phosphorus (STP) differ globally and even within countries. While soil P tests often correlate, the relationship is often specific to a region or study, making any conversion equation difficult to transfer between contexts. Agronomic recommendations, P transport models, syntheses, and other works lack a strong basis for converting between STP. We fulfil this need for converting between two common STP measurements – Olsen P and Mehlich-3 P – through models which account explicitly for soil properties. Hypothesizing that soil properties at the sample level govern how STP values relate, we built models using a combined dataset of ca. 900 soils across the conterminous US and Canada, spanning 10 soil orders (USDA), 3 to 89 % clay, and pH 4 to 9. Model complexity ranged from conventional ‘region-level’ regressions (no soil data) to models adapting to many facets of soil P chemistry, presenting users several viable options suitable for their data context. Depending on data availability, the user could convert between Olsen P and Mehlich-3 P with half the error or less of conventional ‘region-level’ regressions. This reduction in conversion error impacts agronomic recommendations, environmental risk assessments (e.g., P index), and calibration of P transport models (e.g., SWAT+). While it remains best to simply measure the STP of interest, the conversion models here should prove useful in many contexts where that is not feasible. We provide the models in ready-to-use formats, depending on the covariates available to the user and whether the user wants to apply the equations in a spreadsheet or within model code.

Societal risks from flooding are evident at a range of spatial scales and climate change will exacerbate these risks in the future. Assessing flood risks across broad geographical regions is a challenge, and often done using streamflow time-series records or hydrologic models. In this study, we used a national-scale hydrological model to identify, assess, and map 16 different streamflow metrics that could be used to describe flood risks across 34,987 HUC12 subwatersheds within the Mississippi-Atchafalaya River Basin (MARB). A clear spatial difference was observed among two different classes of metrics. Watersheds in the eastern half of the MARB exhibited higher overall flows as characterized by the mean, median, and maximum daily values, whereas western MARB watersheds were associated with flood indicative of high extreme flows such as skewness, standardized streamflow index and top days. Total agricultural and building losses within HUC12 watersheds were related to flood metrics and those focused on higher overall flows were more correlated to expected annual losses (EAL) than extreme value metrics. Results from this study are useful for identifying continental scale patterns of flood risks within the MARB and should be considered a launching point from which to improve the connections between watershed scale risks and the potential use of natural infrastructure practices to reduce these risks.

Coupled surface-subsurface hydrologic models are used worldwide to study historical patterns of water storage and hydrologic behaviour, investigate the impact of management strategies on water resources, and quantify the impact of changing climate, population, and policies. This study presents a new hydrologic model to simulate surface and subsurface in a physically based spatially distributed manner by linking the popular SWAT+ and MODFLOW modelling codes. Within this new code, SWAT+ simulates processes in the landscape, soils, channels, and reservoirs, whereas MODFLOW simulation groundwater processes and interaction with land surface features (soil, channels, canals, reservoirs, tile drains). Geographic connections between SWAT+ objects and MODFLOW grid cells are established a priori using a GIS and then read into the code to be used throughout the simulation to map hydrologic fluxes (recharge, soil water transfer, groundwater-channel exchange, canal seepage, tile drainage outflow, groundwater-reservoir exchange, pumping for irrigation) on a daily time step. The use and general accuracy of the model is demonstrated for two study regions that are subject to irrigation management: the Arkansas River Basin in Colorado and the San Joaquin River Basin in California. An accompanying tutorial and example model data allow for easy use of the model to other study regions. As both SWAT+ and MODFLOW are widely used worldwide for watershed and groundwater modelling, we expect that this new tool can be an important asset in many water resources projects.

In the Mississippi alluvial plain (MAP) area, the demand for groundwater resources from the alluvial aquifer for agricultural irrigation has led to significant reductions in groundwater-level elevation over time. In this study, we use the hydrologic model SWAT + to quantify long-term changes in groundwater storage within the MAP in United States, wherein groundwater is used extensively for irrigation. We apply a linear quantile regression method to perform trend analysis for wet, dry, and average conditions for the 1982–2020 period. The SWAT + model uses the gwflow module to simulate groundwater storage and groundwater-surface water interactions in a physically based spatially distributed manner, with groundwater pumping linked to field-based irrigation demand. Results indicate significant trends in storage and groundwater fluxes. In wet conditions, significant decline trends are noted in groundwater head (–18.0 mm/yr.) and groundwater evapotranspiration (–0.7 mm/yr.). Under dry conditions, trends are in groundwater head (–28.0 mm/yr.), recharge (–5.5 mm/yr.), and groundwater discharge (–5.5 mm/yr.). For average conditions, decreases include groundwater head (–20.6 mm/yr.), recharge (–6 mm/yr.), and groundwater discharge (–9.3 mm/yr.). This underscores the significance of local management solutions.

Selenium (Se) is an essential micro-nutrient for humans and animals but can be toxic at high levels of intake. Quantifying the transport of Se in environmental systems is essential for understanding and mitigating Se contamination in soils, groundwater, and surface waters. In this study, we investigate the transport, storage, and contamination of Se in highly managed, irrigated watershed systems to explore historical conditions, the dominant environmental controls on Se contamination and transport, and the impact of anthropogenic influences such as irrigation and urbanization. We use a portion (23,600 km2) of the Arkansas River Basin, Colorado as an example river basin system, and the SWAT+ watershed model as the numerical tool, amended in this study with a Se reactive transport module. The module accounts for concentrations and mass loads of selenate (SeO4) and selenite (SeO3) for landscapes, soils, aquifers, channels, and lakes and reservoirs. The model is applied to the 2000-2020 period and tested against streamflow, groundwater head, in-stream Se loads, and in-stream Se concentration within an extensive sample network, and then applied to the 1981-2020 period to examine anthropogenic influences on the Se transport cycle. Canal diversions are the major control on streamflow in the Arkansas River, with up to one-third of the diverted volume replaced by groundwater discharge. Canal seepage to the unconfined aquifer drives much of the groundwater discharge. Similarly, Se mass loading in the Arkansas River, approximately 23 kg/day in a downstream irrigated watershed, is controlled by canal diversions (14.4 kg/ day) and groundwater loading (8.2 kg/day), with Se in groundwater controlled by oxidation of Cretaceous shale present as outcrops or bedrock. If anthropogenic influences are removed for a 40-year period, there is more flow and Se load in the Arkansas River. While groundwater discharge and Se loading decrease under this scenario, canal diversions also are eliminated, leading to a net increase in average flow (13 m3/sec -> 21 m3/sec; 60 %), Se load (14.5 kg/day -> 26.4 kg/day; 80 %), and concentration (13 mu g/L -> 15 mu g/L; 15 %). These results can be used to formulate management plans for Se contamination. While methods are applied to the Arkansas River Basin, we expect that general patterns and relationships between shale, canal diversions, canals, and irrigated fields found in this study are applicable to other semi-arid river basins where Cretaceous shale is present as bedrock.

Estimates of groundwater abstraction volumes for irrigation are essential for water management and water supply forecasting. An integrated modeling approach is presented to quantify spatiotemporal field-level abstraction volumes. The approach is both demand-driven, using crop growth algorithms and soil moisture to trigger irrigation events, and supply-constrained, using simulated groundwater storage to constrain extraction. The approach uses the SWAT + hydrologic model, with the use of the gwflow subroutine for simulating spatially distributed groundwater storage and flow. The approach is demonstrated for the Mississippi Delta (northwest Mississippi, USA), a region of high groundwater irrigation and groundwater depletion, for the period 2000–2020. The model is corroborated using system responses that constrain both soil moisture and groundwater storage: streamflow, crop yield, crop evapotranspiration, groundwater level changes, and annual abstraction volumes. Results show good agreement for all system responses, with groundwater level changes matching groundwater depletion magnitudes in the central region of the Big Sunflower Watershed. Simulated annual abstraction volumes generally match measured volumes on an average and frequency basis, although an underestimation of 20

Efficient calibration of hydrological models such as the Soil and Water Assessment Tool Plus (SWAT+) is crucial for accurate watershed simulations and informed water resource management. Traditional calibration strategies often struggle to handle the computational complexities inherent in exploring high-dimensional parameter spaces. This research introduces an approach that harnesses the power of parallel computing for SWAT + model calibration, employing the Dynamically Dimensioned Search (DDS) algorithm. Parallel computing was utilized to distribute computational tasks across multiple processors or computing nodes, significantly reducing calibration time while maintaining precision. By leveraging parallel computing resources, the DDS algorithm explored more extensive parameter spaces for SWAT + calibration in a reasonable timeframe. In a case study conducted in the Upper Mississippi River Basin (UMRB), the effectiveness of parallel computing-enabled DDS was demonstrated for calibrating the SWAT + model. The algorithm efficiently navigated complex parameter spaces, leading to enhanced model performance in simulating critical hydrological processes such as streamflow dynamics, sediment yield, and nutrient transport. Comparisons of simulated and observed data demonstrated satisfactory performance of the calibrated model. The calibration method offered substantial benefits for large-scale hydrologic modeling, enabling researchers and water resource managers to calibrate more complex or fine-grained watershed models more quickly. The integration of parallel computing with DDS in SWAT + calibration facilitated accelerated model calibration and scenario analysis capabilities, enabling more timely decision-making for sustainable water management projects.

Context: The manureshed concept minimizes nutrient imbalance in livestock-intensive agricultural systems by transporting surplus manure to agricultural fields with nutrient demands. The impacts of manureshed-based manure management across the contiguous United States (CONUS) and its potential to improve soil nutrient dynamics and water quality are not well known. Objective: This study developed a framework to evaluate the impacts of manureshed-based manure nutrient management at the CONUS scale. Methods: Across CONUS, county-scale manure imports and exports were balanced by delineating manuresheds according to historic agronomic nitrogen (N) and phosphorus (P) demands and the transportation potentials of the nearest manure types (wet vs. dry). The water quality impacts of manureshed-based nutrient management

The accuracy of soil databases is essential in hydrological modeling, yet limited studies have evaluated the implications of using emerging soil datasets like POLARIS compared to traditional ones such as SSURGO. This study evaluates the performance of POLARIS soil data for simulating the streamflow and sediment yield at both the sub-basin and field scales within the Big Muddy Watershed (BMW), Illinois, U.S.A., using a soft-calibrated SWAT+ model. The field-scale analysis focused on cropland-dominated HRUs from two sub-basins with contrasting POLARIS-SSURGO similarities at the sub-basin scale, optimizing computational efficiency. POLARIS results were compared to those derived from the widely used SSURGO soil database using a soft-calibrated SWAT+ model. At the sub-basin scale, the two datasets showed strong overall agreement for the streamflow and sediment yield over the 81 BMW sub-basins, with minor discrepancies, especially in sediment yield predictions, which exhibited more variability. At the field scale, the agreement between POLARIS and SSURGO was good for both variables, streamflow and sediment yield, though the sediment yield showed greater variability as shown at the sub-basin level. At both scales, the POLARIS and SSURGO outcomes for the streamflow and sediment yield did not always follow the same trend, with discrepancies observed in some sub-basins and HRUs. This suggested that while POLARIS can replicate SSURGO’s streamflow outcomes, this similarity does not always extend to sediment yield predictions and vice versa. At the sub-basin scale, the POLARIS and SSURGO outcomes showed strong alignment (88.9% in “very good” agreement). However, at the field scale, this alignment decreased to 42.9% and 33.3% in specific sub-basins. This indicates that sub-basin aggregation reduces local variability, while finer scales reveal greater sensitivity to soil and hydrological differences. This study highlights POLARIS as a robust alternative to SSURGO for hydrological modeling. Future research should explore its broader application across diverse conditions.

Study region: Lower Arkansas River Basin (LARB) in Colorado, USA. Study focus: The process of implementing irrigation in large river basins often results in significant changes in hydrologic pathways and fluxes, such as canal seepage, runoff, recharge, pumping, and groundwater-river exchange. The objective of this study is to quantify the hydrologic fluxes in a highly irrigated river basin and investigate the controls on these fluxes, using the Lower Arkansas River Basin (LARB) (64,000 km2) in Colorado, USA as a demonstration case. We use the SWAT+ watershed model, enhanced with the new groundwater module gwflow, canal seepage, and irrigation application driven by daily canal diversions and groundwater pumping. The model is tested against streamflow and groundwater head, showing good performance along the Arkansas River and the alluvial corridor. New hydrological insights for the region: On average, precipitation in the basin is 380 mm/yr., of which 2 % (10 mm/yr.) becomes recharge and 2 % is irrigation (80 % surface water irrigation). Water yield is 18 mm/yr. (5 %), principally surface runoff and net groundwater discharge. Canal seepage is only 0.2 % of precipitation. Irrigation fluxes, canal and plant ET are highest in the downstream regions. Sensitivity analysis reveals the controlling watershed features on streamflow, groundwater head, and hydrologic fluxes for each region. Main parameters include streambed conductivity, plant uptake factors, snowmelt factors, aquifer properties, soil available water capacity, and soil percolation coefficient, with each parameter ranked by influence for each region within the basin. The calibrated models can be used to explore the impact of changes in climate, irrigation practices, and general water management schemes.

Groundwater seepage can be a major surface runoff mechanism in humid regions, characterized by shallow aquifers and soil profiles that become saturated during wet periods or intense storm events. This process often plays an important role in the creation and maintenance of groundwater-dependent ecosystems and the overall water yield of a watershed. In this paper, we examine the process of groundwater seepage and assess its influence on hydrologic features (wetlands) and temporal patterns of watershed water yield. We do this by applying a surface-subsurface hydrologic model (SWAT+ with the physically based spatially distributed gwflow module for groundwater storage and flow) to the Little River Watershed, Georgia, USA, which has a high baseflow fraction and contains numerous wetlands, and for which groundwater seepage has been noted in past studies. The model is calibrated and tested against measured streamflow and groundwater head for the period 2000-2015, with and without groundwater seepage included in the gwflow module. Model results indicate that including groundwater seepage improves hydrologic estimation, and demonstrate connections between precipitation, recharge, groundwater seepage, and streamflow before and during storm events. Finally, we compare locations of consistent groundwater seepage (simulated) with mapped wetlands, demonstrating that the model can be used to explore impacts of system changes (land use, climate, management) on wetland development and maintenance.

Nutrients such as nitrogen can be harmful to aquatic organisms in excessive amounts. Climate change, through possible increases in temperature and variable rainfall, may cause changes in nutrient loading patterns from watersheds. This study assesses the potential impact of climate and land use change on nitrate (NO3) loading in the Nanticoke River Watershed (NRW), Chesapeake Bay region, USA, using an updated version of SWAT + watershed model that simulates groundwater nitrate fate and transport in a process based spatially distributed manner. The model was calibrated for the 2000-2015 timeframe and tested against measured streamflow and in-stream nitrate loadings, as well as groundwater head measurements from monitoring wells. After calibration and testing, the model simulated hydrological and nitrate (NO3) flux changes under two future climate scenarios-Representative Concentration Pathways (RCP) 4.5 and 8.5 alongside projected land use changes by the FOREcasting SCEnarios of Land-use Change (FORE-SCE) model. The simulations suggest that under RCP 4.5, streamflow could decrease by 18-34 % and NO3 in-stream loading by 4-22 %, while under RCP 8.5, the projected decreases are 22-33 % for streamflow and 4-11 % for NO3 in-stream loading. Streamflow decrease is due to higher temperatures resulting in higher evapotranspiration during summer months, offsetting increases in precipitation. In-stream NO3 loading is influenced by a decrease in NO3 runoff loading, but an increase in groundwater loading due to increased leaching as plant uptake decreases due to higher surface temperatures. Compared to the influence of climate, land use change results in a minor decrease in NO3 loading. These insights can be used for nutrient management in similar landscapes. Additionally, we show that the updated SWAT + model can be a useful tool in quantifying and investigating NO3 fate and transport in surface-soil-aquifer-channel systems.