Warming Alters Hydrologic Heterogeneity: Simulated Climate Sensitivity of Hydrology‐Based Microrefugia in the Snow‐to‐Rain Transition Zone

Abstract

In complex terrain, drifting snow contributes to ecohydrologic landscape heterogeneity and ecological refugia. In this study, we assessed the climate sensitivity of hydrological dynamics in a semiarid mountainous catchment in the snow‐to‐rain transition zone. This catchment includes a distinct snow drift‐subsidized refugium that comprises a small portion (14.5%) of the watershed but accounts for a disproportionate amount (modeled average 56%) of hydrological flux generation. We conducted climate sensitivity experiments using a physically based hydrologic model to assess responses of a suite of hydrologic metrics across the watershed. Experiments with an imposed 3.5 °C warming showed reductions in average maximum snow water equivalent of 58–68% and deep percolation by 72%. While relative decreases were similar across the watershed, much greater absolute decreases in snowpack occurred in the drift‐subsidized site than the surrounding landscape. In drift‐subsidized locations, warming caused a shift from a regime that included both energy‐ and water‐limited evapotranspiration conditions to exclusively water‐limited conditions. Warming also resulted in altered interannual variability of hydrologic metrics. The drift‐subsidized unit was more sensitive to warming than the surrounding landscape, with reduced potential for the effects of warming to be offset by increased precipitation. Despite spatially homogeneous changes in climate, the effects of climate change on the hydrological dynamics were spatially heterogeneous in this watershed due to the presence of lateral water transport in the form of drifting snow. These findings suggest an increase in hydrologic homogeneity across the landscape and relatively large changes in snow drift‐subsidized refugia.

Spatial Distribution and Scaling Properties of Lidar‐Derived Snow Depth in the Extratropical Andes

“…If the primary goal is to provide a detailed snow process estimation, the minimum model scale that our data suggest is 1 m (i.e., the lidar‐derived data resolution), while the recommended maximum model scale will depend on directional and omnidirectional L 1 values: 2.5 m in VH West, and 12 m for Las Bayas, and 18 m for the remaining sites. These model resolutions are coherent with the scales of drift formation, which has been suggested to control the timing of melt and runoff generation in alpine basins, particularly late in the melt season when persistent drifts can serve as local ecological refugia (e.g., Marshall et al, 2019). Thus, appropriate modeling of snow drift magnitude and location may be desirable for model applications where localized snow persistence has the potential to impact basin‐aggregated estimates of summer low‐flow conditions, water limitation, and vegetation productivity.…”

Runoff sensitivity to snow depletion curve representation within a continental scale hydrologic model

“…However, at the point scale, areas that hold snow longer through the snowmelt season, such as wind‐induced snow drifts, are likely to experience faster snowmelt rates (Trujillo & Molotch, ), while also being associated with watersheds that have high sub‐grid snow variability. Furthermore, many studies evaluating runoff in Reynolds Creek Experimental Watershed, Idaho, have highlighted that deep wind‐induced snow drifts have high runoff efficiency and are a key source of runoff for the watershed (Chauvin et al, ; Flerchinger, Hanson, & Wight, ; Luce et al, ; Marshall et al, ; Stephenson & Freeze, ). Therefore, in the context of our results suggesting that runoff ratio decreases with increased snow variability at the watershed scale, an important scaling consideration is how hydrologic models distribute simulated snowmelt input across an HRU.…”

Efficient multi-objective calibration and uncertainty analysis of distributed snow simulations in rugged alpine terrain

“…Spatially distributed wind drift correction factors have been successfully applied in an attempt to account for the influence of wind transport processes on snowpack heterogeneity (e.g. Hanzer et al, 2016;Marshall et al, 2019). However, as mentioned earlier, after extensive testing, the wind redistribution algorithm avaliable in WaSiM (developed by Warscher et al (2013)) was not applied in our final model; doing so was found to substantially reduce model fits with respect to the observations.…”

“…This poses something of a headache because in very steep terrain, for example, accounting for gravitational snow redistribution is indispensable to hydrologically realistic simulations of SWE evolution and meltwater patterns (Bernhardt et al, 2012;Kerr et al, 2013); In extremis, failure to do so can lead to unrealistic simulated "snow towers" (Freudiger et al, 2017). Fortunately, a variety of pragmatic empirical correction methods and algorithms have emerged to account for such processes (Bernhardt et al, 2012;Marshall et al, 2019;Vögeli et al, 2016).…”