Comparison of Methods to Estimate Snow Water Equivalent at the Mountain Range Scale: A Case Study of the California Sierra Nevada

Abstract

Despite the importance of snow in global water and energy budgets, estimates of global mountain snow water equivalent (SWE) are not well constrained. Two approaches for estimating total range-wide SWE over Sierra Nevada, California, are assessed: 1) global/hemispherical models and remote sensing and models available for continental United States (CONUS) plus southern Canada (CONUS+) available to the scientific community and 2) regional climate model simulations via the Weather Research and Forecasting (WRF) Model run at 3, 9, and 27 km. As no truth dataset provides total mountain range SWE, these two approaches are compared to a “reference” SWE consisting of three published, independent datasets that utilize/validate against in situ SWE measurements. Model outputs are compared with the reference datasets for three water years: 2005 (high snow accumulation), 2009 (average), and 2014 (low). There is a distinctive difference between the reference/WRF datasets and the global/CONUS+ daily estimates of SWE, with the former suggesting up to an order of magnitude more snow. Results are qualitatively similar for peak SWE and 1 April SWE for all three years. Analysis of SWE time series indicates that lower SWE for global and CONUS+ datasets is likely due to precipitation, rain/snow partitioning, and ablation parameterization differences. It is found that WRF produces reasonable (within 50%) estimates of total mountain range SWE in the Sierra Nevada, while the global and CONUS+ datasets underestimate SWE.

A New Estimate of North American Mountain Snow Accumulation From Regional Climate Model Simulations

“…Peak mountain SWS from our estimate is 2.94 times greater than the CanSISE mountain estimate. Comparison of WRF with CanSISE over the mountains shows that current global models often have a large negative bias in mountain snow, consistent with previous studies (e.g., Dutra et al, ; Kapnick et al, ; Snauffer et al, ; Wrzesien et al, ). This difference is substantial; incorrect estimates of snow accumulation lead to incorrect water budgets, which preclude accurate description of the water cycle itself (Rodell et al, ).…”

“…Most striking, the WRF mountain estimate is nearly identical to the total continental storage in CanSISE (1,006 km 3 in WRF versus 1,087 km 3 in CanSISE), despite mountains covering ~25% of the total continent. Previous studies indicated that coarse‐resolution models like ERA‐Interim, a component of CanSISE, fail to capture snow accumulation in mountain regions (Ikeda et al, ; Kapnick & Delworth, ; Pavelsky et al, ; Snauffer et al, ; Wrzesien et al, ), so the low estimate from CanSISE, produced at 1°, is not surprising.…”

“…Here we use “snow water storage” (SWS) to refer to SWE summed over all pixels within a mountain range (presented in cubic kilometers). Wrzesien et al () suggest that a “reasonable” estimate of range‐wide SWS is accurate to the first significant figure, that is, ±50%. However, evaluating accuracy of continental‐scale estimates is problematic since in situ point observations measure (and represent) a small fraction of an RCM pixel (Molotch & Bales, ; Nolin, ).…”

Simulating and Evaluating Atmospheric River‐Induced Precipitation Extremes Along the U.S. Pacific Coast: Case Studies From 1980–2017

“…We dynamically downscale these events, using nested WRF simulations at 27, 9, and 3 km forced by the ERA‐Interim reanalysis, to examine extreme precipitation and other relevant fine‐scale thermodynamic and kinematic quantities event by event. Previous studies suggest resolutions of 10 km are needed to capture realistic orographic precipitation processes in mountainous regions (Minder et al, ; Pavelsky et al, ; Rasmussen et al, ; Wrzesien et al, ). Our results reinforce these earlier findings but also provide substantial additional insights into the importance of fine spatial resolution in the representation of precipitation and water vapor transport extrema—as well as resolution‐dependent drivers in topographically complex regions.…”

Daily gridded datasets of snow depth and snow water equivalent for the Iberian Peninsula from 1980 to 2014

“…To increase the spatial resolution of the WRF outputs, they used the MICROMET model (Liston and Elder, 2006b), a submodel of SnowModel in which WRF outputs are interpolated to a new grid, and then corrected physically according to topography. Wrzesien et al (2017) tested the capability of WRF to estimate SWE over complex terrain concluding that WRF simulations can be used over areas with few observational data.…”