Qianqiu Longyang

and 8 more

In many regions globally, snowmelt-recharged mountainous karst aquifers serve as crucial sources for municipal and agricultural water supplies. In these watersheds, complex interplay of meteorological, topographical, and hydrogeological factors leads to intricate recharge-discharge pathways. This study introduces a spatially distributed deep learning precipitation-runoff model that combines Convolutional Long Short-Term Memory (ConvLSTM) with a spatial attention mechanism. The effectiveness of the deep learning model was evaluated using data from the Logan River watershed and subwatersheds, a characteristically karst-dominated hydrological system in northern Utah. Compared to the ConvLSTM baseline, the inclusion of a spatial attention mechanism improved performance for simulating discharge at the watershed outlet. Analysis of attention weights in the trained model unveiled distinct areas contributing the most to discharge under snowmelt and recession conditions. Furthermore, fine-tuning the model at subwatershed scales provided insights into cross-subwatershed subsurface connectivity. These findings align with results obtained from detailed hydrogeochemical tracer studies. Results highlight the potential of the proposed deep learning approach to unravel the complexities of karst aquifer systems, offering valuable insights for water resource management under future climate conditions. Furthermore, results suggest that the proposed explainable, spatially distributed, deep learning approach to hydrologic modeling holds promise for non-karstic watersheds.

Neelarun Mukherjee

and 4 more

Seasonally warm summers in the Arctic produce supra-permafrost aquifers within the active layer. However, the magnitude of groundwater flow, the amount of dissolved carbon and nutrients, and the solute flow paths are largely unknown, but critical to quantifying downgradient contributions to surface waters (lakes and rivers). To develop approachable methods to quantify groundwater inputs in continuous permafrost watersheds, we selected Imnavait Creek watershed on the North Slope of Alaska as a representative headwater drainage. We conducted 1000 groundwater flow simulations based on topography of the watershed and varying aquifer hydraulic conductivity and saturated thickness values. We fitted a lognormal distribution to the resulting 1000 model outputs, and we derived n=1e6 possible discharge values based on Monte Carlo random sampling on the model outputs. The groundwater discharge values integrated across the watershed generally agree with observed streamflow in Imnavait Creek over 2 months.  When groundwater discharge estimates were combined with in-situ measurements of groundwater-dissolved organic carbon and nitrogen concentrations, we found that Imnavait Creek’s organic matter load is also dominantly sourced from groundwater. Thus, riverine and lacustrine ecological and biogeochemical processes relate strongly to groundwater phenomena in these continuous permafrost settings. As the Arctic warms and the active layer deepens, it will become more important to understand and predict supra-permafrost aquifer dynamics.

Sara Alger

and 2 more

Irrigation activities are a major control on water movement and storage in irrigated river valleys in the Intermountain West, USA. Particularly in dry years, surface water diversions can deplete streams over the summer irrigation season, leading to more variable stream temperatures and increased risk for resident aquatic species. Cooler lateral inflows derived from irrigation activities can mitigate the impacts of depletion by buffering main channel stream temperatures. Given the increasing susceptibility of depleted streams to climate and land use changes, understanding stream temperature patterns and controls in these systems is critical. We used intensive field monitoring over three summers and thermal aerial imagery to characterize stream temperature patterns and irrigation influences in a 2.5 km reach of a small agricultural stream in northern Utah. Considering variable hydrology, weather, channel morphology, diversions, and lateral inflows we found stream temperatures to be relatively insensitive to flow depletion or lateral inflows in a wet year but very sensitive in drier years. Irrigation-related lateral inflows reduced longitudinal warming and diel variability during drier years and at times prevented temperatures from reaching stressful or lethal limits. Reaches with substantial lateral inflow contributions also had a greater areal proportion of low temperatures and spatial temperature diversity. These trends were enhanced by differences in channel morphology, with greater spatial and temporal variability in multi-thread than single-thread reaches. Study results highlight critical flow and weather conditions driving increased temperature variability that will likely become more extreme with additional climate change related reductions in baseflow. Regardless of the cause, this study highlights that decreased instream flows increase the importance of identifying, quantifying, and maintaining lateral inflows to maintain instream temperatures and preservation of these inflows should be considered in future water management decisions.