Daniel T. Myers

Stream temperature influences aquatic ecosystem health, species distribution, and water quality. In snow-dominated watersheds, rain-on-snow (ROS) events regulate stream temperature by introducing cold meltwater, yet their role under future climate is uncertain. We investigate spatial variability of stream thermal regimes in the Great Lakes Basin (GLB) under historical (1960–2014) and future (2047–2074) scenarios using a basin-scale Soil and Water Assessment Tool (SWAT) enhanced with ROS and stream temperature modules. The model was calibrated with daily streamflow and snowpack water equivalent and driven with downscaled CMIP6 climate data. ROS events strongly cool streams in winter and spring, especially in colder northwest and northeast subbasins. However, under future warming, as ROS intensity declines, cooling weakens and stream temperatures increase across all seasons. Summer temperatures are projected to rise by 3–5°C in the southeast and northeast, with cumulative degree days (above 6°C) increasing and warming occurring earlier. These changes are linked to ROS-driven cooling and diminished summer flows. Thermal homogenization is projected across the basin, as historically cooler northern streams warm more rapidly, reducing spatial variability. Hot day duration shows strong negative correlations (−0.4 to −0.6) with ROS timing and intensity, particularly in the Lake Michigan and Ontario basins. The projected loss of ROS-induced cooling may elevate thermal stress for cold-water species, alter ecosystem metabolism, and increase invasive species risk. Adaptive strategies such as riparian shading and forest management will be critical to sustain thermal resilience.

Land cover-based models that rely on purpose-specific thematic details are common in environmental fields such as hydrology, water quality, and ecology. Global remotely sensed land cover from the Dynamic World dataset on Google Earth Engine has trees and built area classes, but enables modelers to create more thematically detailed classifications based on pixel class probabilities from their convolutional neural network (CNN) classifier. However, more information is needed about how these probabilities relate to actual heterogeneity within a land cover class. We used Dynamic World CNN class probabilities to subclassify temperate and tropical forest types from the trees class in the Eastern United States and Costa Rica, as well as developed area intensities from the built class. Subclassifications were evaluated against reference data and in a watershed they were not trained for. The results on dominant temperate forest type user’s accuracy (i.e., of all the pixels classified as a specific land cover type, how many are actually that type) ranged from 43% to 76%, while producer’s accuracy (i.e., of all the actual pixels of a specific land cover type, how many were correctly classified) ranged from 50% to 70%. In the untrained watershed, the overall accuracy was 85% for temperate forest types and 52% for developed areas, demonstrating reliability in classifying forest and developed land cover types. This approach creates opportunities to access up-to-date land cover information with greater thematic detail.

The Potomac River Basin is home to more than six million people and the second largest drainage of the nine river basins that form the Chesapeake Bay watershed. Although great progress has been made toward water quality improvement in the basin, streams and rivers in the area have continued to show trends in increasing salinity, higher stream temperature, higher concentrations of nutrients, and other indicators of degraded water quality. In 2005, the National Capital Region Inventory & Monitoring Network (NCRN I&M) of the National Park Service (NPS) initiated a long-term water quality and quantity monitoring program, carried out through discrete sampling visits to long-term monitoring stream sites within NCRN parks. Parks monitored with this protocol include Antietam National Battlefield, Catoctin Mountain Park, Harpers Ferry National Historical Park, George Washington Memorial Parkway, Manassas National Battlefield Park, Monocacy National Battlefield, National Capital Parks–East, Prince William Forest Park, Rock Creek Park, and Wolf Trap National Park for the Performing Arts. The streams monitored are all part of the Potomac River watershed. The monitoring sites include streams whose watersheds are undeveloped and within National Park boundaries, as well as streams whose watersheds are mostly external to park boundaries, encompassing a variety of urban and agricultural landscapes outside the parks. The goal of the NCRN is to monitor whether stream water quality in park streams is suitable to meet management goals for well-functioning, native ecological communities. The vital sign selection process of the NCRN I&M identified water quality and quantity information as a critical need for the parks of the NCRN. Data collected using this protocol provides much needed status and trend information for park streams, to better understand how changes in watershed land use, climate, and other factors are affecting the stream ecosystems. This protocol contains background information about the purposes of the NCRN water vital signs monitoring, stream water sampling design, field and laboratory methods, and data quality procedures. Acid neutralizing capacity (ANC), dissolved oxygen (DO), pH, specific conductance (including conductivity, salinity, and total dissolved solids (TDS)), and water temperature are monitored to better understand stressors to stream ecosystems such as acidification, increased temperatures, and pollutant pulses. Nitrogen and phosphorus (as total nitrogen and total phosphorus) are monitored to provide information about nutrient pollution in park streams. Water depth, wetted width, velocity, and discharge are monitored to better understand water quantity, and better explain water physical and chemical dynamics.

Most readily available land use/land cover (LULC) data are developed using growing season remote sensing images often at annual time steps, but seasonal changes in remote sensing data can lead to inconsistencies in LULC classification, which could impact geospatial models based on LULC. We used the Dynamic World near-real-time global LULC dataset to compare how geospatial environmental models of water quality and hydrology respond to LULC estimated from growing vs. non-growing season data for temperate watersheds of the eastern United States. Non-growing season data resulted in LULC classifications that had more built area and less tree cover than growing season data due to seasonal impacts on classifications rather than actual LULC changes (e.g., quick construction or succession). In mixed-LULC watersheds, seasonal LULC classification inconsistencies could lead to differences in model outputs depending on the LULC season used, such as differences in watershed nitrogen yields simulated by the Soil and Water Assessment Tool. Within reason, using separate calibration for each season may compensate for these inconsistencies but lead to different model parameter optimizations. Our findings provide guidelines on the use of near-real-time and high-temporal-resolution LULC in geospatial models.

Most readily available landuse/landcover (LULC) data are developed using growing season remote sensing images often at annual time steps. We used the Dynamic World near real-time global LULC dataset to compare how geospatial environmental models of water quality and hydrology respond to growing vs. non-growing season LULC for temperate watersheds of the eastern United States. Non-growing season LULC had more built area and less tree cover than growing season data due to seasonal impacts on classifications rather than actual LULC changes (e.g., quick construction or succession). In mixed-LULC watersheds, seasonal LULC classification inconsistencies could lead to differences in model outputs depending on the LULC season used, such as an increase in watershed nitrogen yields simulated by the Soil and Water Assessment Tool. Within reason, using separate calibration for each season may compensate for these inconsistencies, but lead to different model parameter optimizations. Our findings provide guidelines on the use of near real-time and high temporal resolution LULC in geospatial models.

Most readily available landuse/landcover (LULC) data are developed using growing season remote sensing images often at annual time steps. We used the Dynamic World near real-time global LULC dataset to compare how geospatial environmental models of water quality and hydrology respond to growing vs. non-growing season LULC for temperate watersheds of the eastern United States. Non-growing season LULC had more built area and less tree cover than growing season data due to seasonal impacts on classifications rather than actual LULC changes (e.g., quick construction or succession). In mixed-LULC watersheds, seasonal LULC classification inconsistencies could lead to differences in model outputs depending on the LULC season used, such as an increase in watershed nitrogen yields simulated by the Soil and Water Assessment Tool. Within reason, using separate calibration for each season may compensate for these inconsistencies, but lead to different model parameter optimizations. Our findings provide guidelines on the use of near real-time and high temporal resolution LULC in geospatial models.

Most readily available landuse/landcover (LULC) data are developed using growing season remote sensing images often at annual time steps. We used the Dynamic World near real-time global LULC dataset to compare how geospatial environmental models of water quality and hydrology respond to growing vs. non-growing season LULC for temperate watersheds of the eastern United States. Non-growing season LULC had more built area and less tree cover than growing season data due to seasonal impacts on classifications rather than actual LULC changes (e.g., quick construction or succession). In mixed-LULC watersheds, seasonal LULC classification inconsistencies could lead to differences in model outputs depending on the LULC season used, such as an increase in watershed nitrogen yields simulated by the Soil and Water Assessment Tool. Within reason, using separate calibration for each season may compensate for these inconsistencies, but lead to different model parameter optimizations. Our findings provide guidelines on the use of near real-time and high temporal resolution LULC in geospatial models.

Most readily available landuse/landcover (LULC) data are developed using growing season remote sensing images often at annual time steps. We used the Dynamic World near real-time global LULC dataset to compare how geospatial environmental models of water quality and hydrology respond to growing vs. non-growing season LULC for temperate watersheds of the eastern United States. Non-growing season LULC had more built area and less tree cover than growing season data due to seasonal impacts on classifications rather than actual LULC changes (e.g., quick construction or succession). In mixed-LULC watersheds, seasonal LULC classification inconsistencies could lead to differences in model outputs depending on the LULC season used, such as an increase in watershed nitrogen yields simulated by the Soil and Water Assessment Tool. Within reason, using separate calibration for each season may compensate for these inconsistencies, but lead to different model parameter optimizations. Our findings provide guidelines on the use of near real-time and high temporal resolution LULC in geospatial models.