23 October 2024 in Water Resources Research
Hydraulic geometry parameters describing river hydrogeomorphic relationships are critical for determining a channel's capacity to convey water and sediment which is important for flood forecasting. Although well‐established, power‐law hydraulic geometry curves have been widely used to understand riverine systems and mapping flooding inundation worldwide for the past 70 years, we have become increasingly aware of their limitations. In the present study, we have moved beyond these traditional power‐law relationships, testing the ability of machine‐learning models to provide improved predictions of river width and depth. For this work, we have used an unprecedentedly large river measurement data set (HYDRoSWOT) as well as a suite of watershed predictor data to develop novel data‐driven approaches to better estimate river geometries over the contiguous United States (CONUS). Our Random Forest, XGBoost, and neural network models out‐performed the traditional, regionalized power law‐based hydraulic geometry equations for both width and depth, providing R‐squared values of as high as 0.75 for width and as high as 0.67 for depth, compared with R‐squared values of 0.45 for width and 0.18 for depth from the regional hydraulic geometry equations. Our results also show diverse performance outcomes across stream orders and geographical regions for the different machine‐learning models, demonstrating the value of using multi‐model approaches to maximize the predictability of river geometry. The developed models have been used to create the newly publicly available STREAM‐geo data set, which provides river width, depth, width/depth ratio, and river and stream surface area (%RSSA) for nearly 2.7 million NHDPlus stream reaches across the contiguous US.
29 May 2025 in Water Resources Research
The accelerated warming in the Arctic poses serious risks to freshwater ecosystems by altering streamflow and river thermal regimes. However, limited research on Arctic River water temperatures exists due to data scarcity and the absence of robust methodologies, which often focus on large, major river basins. To address this, we leveraged the newly released, extensive AKTEMP data set and advanced machine learning techniques to develop a Long Short‐Term Memory (LSTM) model. By incorporating ERA5‐Land reanalysis data and integrating physical understanding into data‐driven processes, our model advanced river water temperature predictions in ungauged, snow‐ and permafrost‐affected basins in Alaska. Our model outperformed existing approaches in high‐latitude regions, achieving a median Nash‐Sutcliffe Efficiency of 0.95 and root mean squared error of 1.0°C. The LSTM model learned air temperature, soil temperature, solar radiation, and thermal radiation—factors associated with energy balance—were the most important drivers of river temperature dynamics. Soil moisture and snow water equivalent were highlighted as critical factors representing key processes such as thawing, melting, and groundwater contributions. Glaciers and permafrost were also identified as important covariates, particularly in seasonal river water temperature predictions. Our LSTM model successfully captured the complex relationships between hydrometeorological factors and river water temperatures across varying timescales and hydrological conditions. This scalable and transferable approach can be potentially applied across the Arctic, offering valuable insights for future conservation and management efforts.
28 February 2026 in Scientific Data
Excess nitrogen (N) within the landscape can lead to environmental issues such as water pollution and downstream eutrophication. Quantification of landscape N fluxes can help identify areas of excess N and can improve our understanding of the sources, sinks, and transport of N within ecosystems. The gTREND-Nitrogen dataset provides a comprehensive, long-term (1930-2017) N mass balance for the contiguous United States at a spatial resolution of 250 meters. This dataset integrates county-scale estimates of N fluxes with gridded land use and population data to estimate grid-scale surface fluxes of N, including fertilizer, atmospheric deposition, manure inputs, biological N fixation, crop N uptake, and population-based human waste. The downscaled gTREND-Nitrogen data will allow for fine-scale insights into both historical and current N dynamics, addressing the limitations of previous datasets developed at coarser spatial scales. The data is openly available and will help to inform the development of effective policy and management strategies to mitigate the negative impacts of excess N and to promote sustainable nutrient management practices.
20 April 2026 in Global Biogeochemical Cycles
Phosphorus (P) is essential for food production but poses persistent environmental risks due to overapplication and accumulation in agricultural landscapes. Despite decades of management efforts, P losses continue to degrade water quality across the United States. In this study, we use a high‐resolution 250 m gridded data set (gTREND‐P) spanning 1930–2017 to assess long‐term trends in agricultural P use through three complementary indicators: phosphorus use efficiency (PUE), annual P surplus, and cumulative surplus. We identify strong regional patterns: livestock‐dominated regions in the East, and parts of the Midwest exhibit low PUE and high cumulative surplus due to persistent manure inputs. In the central Midwest, fertilizer‐intensive crop systems have also led to substantial P accumulation, contributing to large legacy surpluses despite recent improvements in PUE. In some areas, PUE values exceed 1, suggesting that crop uptake is increasingly supported by soil P reserves rather than current inputs. To better capture both historical and contemporary dynamics, we introduce a quadrant framework that integrates PUE and cumulative surplus to classify agricultural land into four states: Intensification, Recovery from Overapplication, Recovery from Soil Depletion, and Unsustainable Use. This typology reveals spatial shifts over time, with 19% of US croplands transitioning from Intensification to Recovery from Overapplication between 1980 and 2017, while 60% remain in the “Intensification” state, often driven by high manure inputs. Our findings highlight the need for regionally tailored parcel‐level P management strategies that recognize the legacy effects of past practices while enabling targeted interventions for sustainable P use.
8 April 2026 in Scientific Data
Water quantity predictions have advanced rapidly, driven by ready-to-use benchmarks such as CAMELS (Catchment Attributes and Meteorology for Large-Sample Studies). In contrast, large-scale water quality predictions, especially for nutrients, lag behind due to a lack of comparable datasets. Existing water quality datasets face four major limitations: (1) underrepresentation of human-impacted systems, (2) absence of nutrient inputs, (3) incomplete watershed metadata, and (4) sparse monitoring coverage. To address these gaps, we developed IWAND-Nitrogen (Integrated Watershed Attributes and Nutrient Data for Nitrogen) for the contiguous United States. IWAND-Nitrogen integrates 574,767 nitrate records from 1,877 catchments (median 272 samples per gauge; IQR = 231–346) with at least 200 measurements each from 1980–2023, linked with 93 watershed attributes, eight nitrogen input forcings (both basin-averaged and gridded), and eleven climate forcings. Compared to existing benchmarks such as CAMELS-Chem, IWAND-Nitrogen complements prior efforts by extending spatial/temporal coverage and enhancing representation across anthropogenic gradients. IWAND-Nitrogen aims to serve as a nutrient community benchmark, advancing from model development to new insights from catchment to national scales.