A comparison of National Water Model retrospective analysis snow outputs at snow telemetry sites across the Western United States

Abstract

This study compares the US National Water Model (NWM) reanalysis snow outputs to observed snow water equivalent (SWE) and snow-covered area fraction (SCAF) at snow telemetry (SNOTEL) sites across the Western United States. SWE was obtained from SNOTEL sites, while SCAF was obtained from moderate resolution imaging spectroradiometer (MODIS) observations at a nominal 500 m grid scale. Retrospective NWM results were at a 1000 m grid scale. We compared results for SNOTEL sites to gridded NWM and MODIS outputs for the grid cells encompassing each SNOTEL site.Differences between modelled and observed SWE were attributed to both model errors, as well as errors in inputs, notably precipitation and temperature. The NWM generally under-predicted SWE, partly due to precipitation input differences. There was also a slight general bias for model input temperature to be cooler than observed, counter to the direction expected to lead to under-modelling of SWE. There was also under-modelling of SWE for a subset of sites where precipitation inputs were good. Furthermore, the NWM generally tends to melt snow early. There was considerable variability between modelled and observed SCAF as well as the binary comparison of snow cover presence that hampered useful interpretation of SCAF comparisons. This is in part due to the shortcomings associated with both model SCAF parameterization and MODIS observations, particularly in vegetated regions. However, when SCAF was aggregated across all sites and years, modelled SCAF tended to be more than observed using MODIS. These differences are regional with generally better SWE and SCAF results in the Central Basin and Range and differences tending to become larger the further away regions are from this region. These findings identify areas where predictions from the NWM involving snow may be better or worse, and suggest opportunities for research directed towards model improvements.

Evaluation of Snow and Streamflows Using Noah-MP and WRF-Hydro Models in Aroostook River Basin, Maine

“…We note that while Noah-MP's simulation of fSCA was too high relative to observed fSCA, as also noted in other studies [20,50], more realistically simulating fSCA might degrade the simulation of SWE since the surface albedo, rather than the snow albedo, is used to calculate the energy balance at the snow's surface. Per Noah-MP's parameterization: Surface Albedo = Soil Albedo (1 − fSCA) + Snow Albedo (fSCA) (10) where Snow Albedo = ƒ f(snow age, zenith angle) Noah-MP's energy budget calculation would be more accurate if Noah-MP calculated an energy balance for snow-covered and non-snow-covered surfaces separately and then averaged the resulting fluxes of energy and mass within the grid cell, rather than its current method of using a grid cell average albedo (Equation ( 9)).…”

“…Although the precipitation bias was not the dominant factor contributing to the differences in SWE and SD in this study, the biases, if present along with the biases of T2m, would have a major impact on snow versus rain and thus in the simulated SWE and SD. A study of NWM reanalysis snow outputs for observed SWE in the western US showed that the errors in precipitation inputs were more important than the cooler temperature bias, as the NWM generally underpredicted the SWE [20]. This shows that the contribution of inputs in the NWM simulation might have regional differences.…”

“…In an analysis of snowpack change from 1982 to 2016 over the conterminous United States (US), the snow season was significantly shortened due to the earlier ending of the snow season over the western US and its later arrival over the eastern US [18]. Numerous studies have investigated snow storage's impact on snowmelt and streamflow forecasts in snow-dominated regions of the mountainous western US [8,17,19,20] but studies that explored streamflow predictability in the wet northeastern regions of the US are limited. One 40-year study of water budgets in the northeastern US, using Noah-MP and WRF-Hydro with the NWM configuration, documented increases in temperatures and changes in precipitation patterns, which resulted in more water availability during the winter and less during the spring as the snow accumulation dwindled during the winter [20].…”

“…Numerous studies have investigated snow storage's impact on snowmelt and streamflow forecasts in snow-dominated regions of the mountainous western US [8,17,19,20] but studies that explored streamflow predictability in the wet northeastern regions of the US are limited. One 40-year study of water budgets in the northeastern US, using Noah-MP and WRF-Hydro with the NWM configuration, documented increases in temperatures and changes in precipitation patterns, which resulted in more water availability during the winter and less during the spring as the snow accumulation dwindled during the winter [20]. It also highlighted that more substantial changes, including high and low flows, are expected to occur in river reaches closer to head waters or smaller basins compared to larger basins as the larger basins are able to dampen the signals, whereas the regional models are often too coarse to capture small-scale watershed processes [21,22].…”

Drivers of Dust-Enhanced Snowpack Melt-Out and Streamflow Timing

“…Recommended next steps are to understand individual dust layer dynamics [68] in relation to sub-seasonal snow and climate variables, to incorporate weather forecasts into snowpack energy balance modeling [94] and develop a real-time dust-enhanced energy absorption forecast product for operational end-users. Operational streamflow forecasting is challenging [95], and snowpack and hydrological models are often too complex to be very useful in a forecast setting [96,97]. A hierarchal decision tree approach [98] could be constructed to prioritize most important variables as the snow season unfolds.…”