Snow Water Equivalent Simulations Research Articles

On optimization of calibrations of a distributed hydrological model with spatially distributed information on snow

Abstract. In northern cold-temperate countries, a large portion of annual streamflow is produced by spring snowmelt, which often triggers floods. It is important to have spatial information about snow variables such as snow water equivalent (SWE), which can be incorporated into hydrological models, making them more efficient tools for improved decision-making. The present research implements a unique spatial pattern metric in a multi-objective framework for calibration of hydrological models and attempts to determine whether raw SNODAS (SNOw Data Assimilation System) data can be utilized for hydrological model calibration. The spatial efficiency (SPAEF) metric is explored for spatially calibrating SWE. Different calibration experiments are performed combining Nash–Sutcliffe efficiency (NSE) for streamflow and root-mean-square error (RMSE) and SPAEF for SWE, using the Dynamically Dimensioned Search (DDS) and Pareto Archived Dynamically Dimensioned Search multi-objective optimization (PADDS) algorithms. Results of the study demonstrate that multi-objective calibration outperforms sequential calibration in terms of model performance (SWE and discharge simulations). Traditional model calibration involving only streamflow produced slightly higher NSE values; however, the spatial distribution of SWE could not be adequately maintained. This study indicates that utilizing SPAEF for spatial calibration of snow parameters improved streamflow prediction compared to the conventional practice of using RMSE for calibration. SPAEF is further implied to be a more effective metric than RMSE for both sequential and multi-objective calibration. During validation, the calibration experiment incorporating multi-objective SPAEF exhibits enhanced performance in terms of NSE and Kling–Gupta efficiency (KGE) compared to calibration experiment solely based on NSE. This observation supports the notion that incorporating SPAEF computed on raw SNODAS data within the calibration framework results in a more robust hydrological model. The novelty of this study is the implementation of SPAEF with respect to spatially distributed SWE for calibrating a distributed hydrological model.

LSTM-Based Data Integration to Improve Snow Water Equivalent Prediction and Diagnose Error Sources

Abstract Accurate prediction of snow water equivalent (SWE) can be valuable for water resource managers. Recently, deep learning methods such as long short-term memory (LSTM) have exhibited high accuracy in simulating hydrologic variables and can integrate lagged observations to improve prediction, but their benefits were not clear for SWE simulations. Here we tested an LSTM network with data integration (DI) for SWE in the western United States to integrate 30-day-lagged or 7-day-lagged observations of either SWE or satellite-observed snow cover fraction (SCF) to improve future predictions. SCF proved beneficial only for shallow-snow sites during snowmelt, while lagged SWE integration significantly improved prediction accuracy for both shallow- and deep-snow sites. The median Nash–Sutcliffe model efficiency coefficient (NSE) in temporal testing improved from 0.92 to 0.97 with 30-day-lagged SWE integration, and root-mean-square error (RMSE) and the difference between estimated and observed peak SWE values dmax were reduced by 41% and 57%, respectively. DI effectively mitigated accumulated model and forcing errors that would otherwise be persistent. Moreover, by applying DI to different observations (30-day-lagged, 7-day-lagged), we revealed the spatial distribution of errors with different persistent lengths. For example, integrating 30-day-lagged SWE was ineffective for ephemeral snow sites in the southwestern United States, but significantly reduced monthly-scale biases for regions with stable seasonal snowpack such as high-elevation sites in California. These biases are likely attributable to large interannual variability in snowfall or site-specific snow redistribution patterns that can accumulate to impactful levels over time for nonephemeral sites. These results set up benchmark levels and provide guidance for future model improvement strategies.

Impacts of snow surface aerodynamic resistance on snow water equivalent simulations in forested regions

AbstractThe aerodynamic resistance at the snow surface (ras) has an influence on the snowpack energy balance. However, there are no widely accepted approaches for parameterizing ras in forested regions. In this study, we derived ras under the influence of forest vegetation for five cases that depend on the canopy height and structure. Three parameterization schemes for ras were then integrated into a distributed hydrologic model: (1) the aerodynamic resistance above the canopy, ignoring the canopy structure effects on ras, (2) the full aerodynamic resistance formulation for the different canopy height cases, and (3) a simplified formulation that only represented the lower trunk layer. Model simulations were conducted at six snow telemetry (SNOTEL) sites in the Verde River Watershed, Arizona, and evaluated with respect to snow water equivalent (SWE) product from the SNOTEL sites and a gridded product. Results revealed the importance of accurately parameterizing vegetation canopy structure when calculating ras in forested regions. Cases where ras included the vertical structure effects showed improved simulations of SWE across all sites. Finally, ras is mostly determined by trunk layer aerodynamic resistance such that the simplified formulation is sufficient for representing forest canopy effects on ras.

Exploring the potential of thermal infrared remote sensing to improve a snowpack model through an observing system simulation experiment

Abstract. The assimilation of data from Earth observation satellites into numerical models is considered to be the path forward to estimate snow cover distribution in mountain catchments, providing accurate information on the mountainous snow water equivalent (SWE). The land surface temperature (LST) can be observed from space, but its potential to improve SWE simulations remains underexplored. This is likely due to the insufficient temporal or spatial resolution offered by the current thermal infrared (TIR) missions. However, three planned missions will provide global-scale TIR data at much higher spatiotemporal resolution in the coming years. To investigate the value of TIR data to improve SWE estimation, we developed a synthetic data assimilation (DA) experiment at five snow-dominated sites covering a latitudinal gradient in the Northern Hemisphere. We generated synthetic true LST and SWE series by forcing an energy balance snowpack model with the ERA5-Land reanalysis. We used this synthetic true LST to recover the synthetic true SWE from a degraded version of ERA5-Land. We defined different observation scenarios to emulate the revisiting times of Landsat 8 (16 d) and the Thermal infraRed Imaging Satellite for High-resolution Natural resource Assessment (TRISHNA) (3 d) while accounting for cloud cover. We replicated the experiments 100 times at each experimental site to assess the robustness of the assimilation process with respect to cloud cover under both revisiting scenarios. We performed the assimilation using two different approaches: a sequential scheme (particle filter) and a smoother (particle batch smoother). The results show that LST DA using the smoother reduced the normalized root mean square error (nRMSE) of the SWE simulations from 61 % (open loop) to 17 % and 13 % for 16 d revisit and 3 d revisit respectively in the absence of clouds. We found similar but higher nRMSE values by removing observations due to cloud cover but with a substantial increase in the standard deviation of the nRMSE of the replicates, highlighting the importance of revisiting times in the stability of the assimilation performance. The smoother largely outperformed the particle filter algorithm, suggesting that the capability of a smoother to propagate the information along the season is key to exploit LST information for snow modelling. Finally, we have compared the benefit of assimilating LST with synthetic observations of fractional snow cover area (FSCA). LST DA performed better than FSCA DA in all the study sites, suggesting that the information provided by LST is not limited to the duration of the snow season. These results suggest that the LST data assimilation has an underappreciated potential to improve snowpack simulations and highlight the value of upcoming TIR missions to advance snow hydrology.

Scale Dependence of Errors in Snow Water Equivalent Simulations Using ERA5 Reanalysis over Alpine Basins

This study aims to evaluate the potential of ERA5 precipitation and temperature reanalysis for snow water equivalent (SWE) simulation by considering the role of catchment spatial scale in controlling the errors obtained by comparison with corresponding SWE simulations from ground stations. This is obtained by exploiting a semi-distributed snowpack model (TOPMELT) implemented over the upper Adige River basin in the Eastern Italian Alps, where 16 sub-catchments of varying sizes are considered. The comparison is carried out from 1992 to 2019. The findings show that ERA5 precipitation overestimated low-intensity rainfall (drizzle problem) and underestimated high-intensity rainfall, while ERA5 temperature underestimated observations. The overestimation of low-intensity rainfall created fictitious low-intensity snowfall events, which, when combined with colder ERA5 temperature, resulted in delayed snowmelt and increased fictitious snow-cover days over the study area. The quantile mapping (QM) technique was used to remove errors in ERA5 variables. It was shown that ERA5 could struggle to resolve the orographic enhancement in precipitation, which may be particularly important during high-SWE years. This reduces the positive precipitation bias during those years, thus reducing comparatively the ability of the quantile mapping technique to correct for bias homogeneously during all years. This study highlighted the importance of temperature correction over precipitation correction in SWE simulation, particularly for smaller basins.

Long-term reconstruction of satellite-based precipitation, soil moisture, and snow water equivalent in China

Abstract. A long-term high-resolution national dataset of precipitation (P), soil moisture (SM), and snow water equivalent (SWE) is necessary for predicting floods and droughts and assessing the impacts of climate change on streamflow in China. Current long-term daily or sub-daily datasets of P, SM, and SWE are limited by a coarse spatial resolution or the lack of local correction. Although SM and SWE data derived from hydrological simulations at a national scale have fine spatial resolutions and take advantage of local forcing data, hydrological models are not directly calibrated with SM and SWE data. In this study, we produced a daily 0.1∘ dataset of P, SM, and SWE in 1981–2017 across China, using global background data and local on-site data as forcing input and satellite-based data as reconstruction benchmarks. Global 0.1∘ and local 0.25∘P data in 1981–2017 are merged to reconstruct the historical P of the 0.1∘ China Merged Precipitation Analysis (CMPA) available in 2008–2017 using a stacking machine learning model. The reconstructed P data are used to drive the HBV hydrological model to simulate SM and SWE data in 1981–2017. The SM simulation is calibrated by Soil Moisture Active Passive Level 4 (SMAP-L4) data. The SWE simulation is calibrated by the national satellite-based snow depth dataset in China (Che and Dai, 2015) and the Moderate Resolution Imaging Spectroradiometer (MODIS) snow cover data. Cross-validated by the spatial and temporal splitting of the CMPA data, the median Kling–Gupta efficiency (KGE) of the reconstructed P is 0.68 for all grids at a daily scale. The median KGE of SM in calibration is 0.61 for all grids at a daily scale. For grids in two snow-rich regions, the median KGEs of SWE in calibration are 0.55 and −2.41 in the Songhua and Liaohe basins and the northwest continental basin respectively at a daily scale. Generally, the reconstruction dataset performs better in southern and eastern China than in northern and western China for P and SM and performs better in northeast China than in other regions for SWE. As the first long-term 0.1∘ daily dataset of P, SM, and SWE that combines information from local observations and satellite-based data benchmarks, this reconstruction product is valuable for future national investigations of hydrological processes.

Snow Level From Post‐Processing of Atmospheric Model Improves Snowfall Estimate and Snowpack Prediction in Mountains

AbstractIn mountains, the precipitation phase greatly varies in space and time and affects the evolution of the snow cover. Snowpack models usually rely on precipitation‐phase partitioning methods (PPMs) that use near‐surface variables. These PPMs ignore conditions above the surface thus limiting their ability to predict the precipitation phase at the surface. In this study, the impact on snowpack simulations of atmospheric‐based PPMs, incorporating upper atmospheric information, is tested using the snowpack scheme Crocus. Crocus is run at 2.5‐km grid spacing over the mountains of southwestern Canada and northwestern United States and is driven by meteorological fields from an atmospheric model at the same resolution. Two atmospheric‐based PPMs were considered from the atmospheric model: the output from a detailed microphysics scheme and a post‐processing algorithm determining the snow level and the associated precipitation phase. Two ground‐based PPMs were also included as lower and upper benchmarks: a single air temperature threshold at 0°C and a PPM using wet‐bulb temperature. Compared to the upper benchmark, the snow‐level based PPM improved the estimation of snowfall occurrence by 5% and the simulation of snow water equivalent (SWE) by 9% during the snow melting season. In contrast, due to missing processes, the microphysics scheme decreased performances in phase estimate and SWE simulations compared to the upper benchmark. These results highlight the need for detailed evaluation of the precipitation phase from atmospheric models and the benefit for mountain snow hydrology of the post‐processed snow level. The limitations to drive snowpack models at slope scale are also discussed.

Implementation and Evaluation of a Unified Turbulence Parameterization Throughout the Canopy and Roughness Sublayer in Noah‐MP Snow Simulations

AbstractThe Noah‐MP land surface model (LSM) relies on the Monin‐Obukhov (M‐O) Similarity Theory (MOST) to calculate land‐atmosphere exchanges of water, energy, and momentum fluxes. However, MOST flux‐profile relationships neglect canopy‐induced turbulence in the roughness sublayer (RSL) and parameterize within‐canopy turbulence in an ad hoc manner. We implement a new physics scheme (M‐O‐RSL) into Noah‐MP that explicitly parameterizes turbulence in RSL. We compare Noah‐MP simulations employing the M‐O‐RSL scheme (M‐O‐RSL simulations) and the default M‐O scheme (M‐O simulations) against observations obtained from 647 Snow Telemetry (SNOTEL) stations and two AmeriFlux stations in the western United States. M‐O‐RSL simulations of snow water equivalent (SWE) outperform M‐O simulations over 64% and 69% of SNOTEL sites in terms of root‐mean‐square‐error (RMSE) and correlation, respectively. The largest improvements in skill for M‐O‐RSL occur over closed shrubland sites, and the largest degradations in skill occur over deciduous broadleaf forest sites. Differences between M‐O and M‐O‐RSL simulated snowpack are primarily attributable to differences in aerodynamic conductance for heat underneath the canopy top, which modulates sensible heat flux. Differences between M‐O and M‐O‐RSL within‐canopy and below‐canopy sensible heat fluxes affect the amount of heat transported into snowpack and hence change snowmelt when temperatures are close to or above the melting point. The surface energy budget analysis over two AmeriFlux stations shows that differences between M‐O and M‐O‐RSL simulations can be smaller than other model biases (e.g., surface albedo). We intend for the M‐O‐RSL physics scheme to improve performance and uncertainty estimates in weather and hydrological applications that rely on Noah‐MP.

Data assimilation of satellite-based terrestrial water storage changes into a hydrology land-surface model

Accurate estimation of snow mass or snow water equivalent (SWE) over space and time is required for global and regional predictions of the effects of climate change. This work investigates whether integration of remotely sensed terrestrial water storage (TWS) information, which is derived from the Gravity Recovery and Climate Experiment (GRACE), can improve SWE and streamflow simulations within a semi-distributed hydrology land surface model. A data assimilation (DA) framework was developed to combine TWS observations with the MESH (Modélisation Environnementale Communautaire – Surface Hydrology) model using an ensemble Kalman smoother (EnKS). The snow-dominated Liard Basin was selected as a case study. The proposed assimilation methodology reduced bias of monthly SWE simulations at the basin scale by 17.5% and improved unbiased root-mean-square difference (ubRMSD) by 23%. At the grid scale, the DA method improved ubRMSD values and correlation coefficients for 85% and 97% of the grid cells, respectively. Effects of GRACE DA on streamflow simulations were evaluated against observations from three river gauges, where it effectively improved the simulation of high flows during snowmelt season from April to June. The influence of GRACE DA on the total flow volume and low flows was found to be variable. In general, the use of GRACE observations in the assimilation framework not only improved the simulation of SWE, but also effectively influenced streamflow simulations.

Comparing calibration strategies of a conceptual snow hydrology model and their impact on model performance and parameter identifiability

Having a realistic estimation of snow cover by conceptual hydrological models continues to challenge hydrologists. The calibration of the free model parameters is an unavoidable step and the uncertainties resulting from the use of this optimal set remains a source of concern, especially in forecasting applications and climate changes impact assessments. This study seeks to improve the calibration of the conceptual hydrological model GR4J coupled with the Cemaneige snow model, in order to obtain a more realistic simulation of the snow water equivalent (SWE) and to reduce the uncertainty of the free parameters. The performance of the two models was tested over twelve snow-dominated basins in southern Quebec, Canada. Four calibration strategies were adopted and compared. In the first two strategies, the parameters were calibrated against observed streamflow alone using a local and a global algorithm. In the third and fourth strategies the calibration of snow and hydrological parameters was performed against observed streamflow and snow water equivalent (SWE) measured at snow course transects, first separately, and then with a multiobjective approach. An ensemble of equifinal parameters was used to compare the capacity of the global and multiobjective algorithms to improve the parameters identifiability and to assess the impact of parameter equifinality on the temperature sensitivity of spring peak streamflow. The large number of equifinal parameters found during calibration underscores the importance of structural non-identifiability of the coupled GR4J-Cemaneige model. The inclusion of snow observations within a multiobjective calibration improved the simulation of SWE, the identifiability of the parameters and their correlation with basins characteristics. Parameter equifinality caused a small but non negligible uncertainty in the simulated response of spring peak flow to warming temperatures. Parameter equifinality should be considered in climate impact studies in snow-dominated basins where poorly constrained snow parameters can affect the temperature sensitivity of streamflow.

A quantitative method to decompose SWE differences between regional climate models and reanalysis datasets

The simulation of snow water equivalent (SWE) remains difficult for regional climate models. Accurate SWE simulation depends on complex interacting climate processes such as the intensity and distribution of precipitation, rain-snow partitioning, and radiative fluxes. To identify the driving forces behind SWE difference between model and reanalysis datasets, and guide model improvement, we design a framework to quantitatively decompose the SWE difference contributed from precipitation distribution and magnitude, ablation, temperature and topography biases in regional climate models. We apply this framework within the California Sierra Nevada to four regional climate models from the North American Coordinated Regional Downscaling Experiment (NA-CORDEX) run at three spatial resolutions. Models generally predict less SWE compared to Landsat-Era Sierra Nevada Snow Reanalysis (SNSR) dataset. Unresolved topography associated with model resolution contribute to dry and warm biases in models. Refining resolution from 0.44° to 0.11° improves SWE simulation by 35%. To varying degrees across models, additional difference arises from spatial and elevational distribution of precipitation, cold biases revealed by topographic correction, uncertainties in the rain-snow partitioning threshold, and high ablation biases. This work reveals both positive and negative contributions to snow bias in climate models and provides guidance for future model development to enhance SWE simulation.

A High-Resolution Data Assimilation Framework for Snow Water Equivalent Estimation across the Western United States and Validation with the Airborne Snow Observatory

Abstract Numerical simulations of snow water equivalent (SWE) in mountain systems can be biased, and few SWE observations have existed over large domains. New approaches for measuring SWE, like NASA’s ultra-high-resolution Airborne Snow Observatory (ASO), offer an opportunity to improve model estimates by providing a high-quality validation target. In this study, a computationally efficient snow data assimilation (DA) approach over the western United States at 1.75-km spatial resolution for water years (WYs) 2001–17 is presented. A local ensemble transform Kalman filter implemented as a batch smoother is used with the VIC hydrology model to assimilate the remotely sensed daily MODIS fractional snow-covered area (SCA). Validation of the high-resolution SWE estimates is done against ASO SWE data in the Tuolumne basin (California), Uncompahgre basin (Colorado), and Olympic Peninsula (Washington). Results indicate good performance in dry years and during melt, with DA reducing Tuolumne basin-average SWE percent differences from −68%, −92%, and −84% in open loop to 0.6%, 25%, and 3% after DA for WYs 2013–15, respectively, for ASO dates and spatial extent. DA also improved SWE percent difference over the Uncompahgre basin (−84% open loop, −65% DA) and Olympic Peninsula (26% open loop, −0.2% DA). However, in anomalously wet years DA underestimates SWE, likely due to an inadequate snow depletion curve parameterization. Despite potential shortcomings due to VIC model setup (e.g., water balance mode) or parameterization (snow depletion curve), the DA framework implemented in this study shows promise in overcoming some of these limitations and improving estimated SWE, in particular during drier years or at higher elevations, when most in situ observations cannot capture high-elevation snowpack due to lack of stations there.

Assimilation of passive microwave AMSR-2 satellite observations in a snowpack evolution model over northeastern Canada

Abstract. Over northeastern Canada, the amount of water stored in a snowpack, estimated by its snow water equivalent (SWE) amount, is a key variable for hydrological applications. The limited number of weather stations driving snowpack models over large and remote northern areas generates great uncertainty in SWE evolution. A data assimilation (DA) scheme was developed to improve SWE estimates by updating meteorological forcing data and snowpack states with passive microwave (PMW) satellite observations and without using any surface-based data. In this DA experiment, a particle filter with a Sequential Importance Resampling algorithm (SIR) was applied and an inflation technique of the observation error matrix was developed to avoid ensemble degeneracy. Advanced Microwave Scanning Radiometer 2 (AMSR-2) brightness temperature (TB) observations were assimilated into a chain of models composed of the Crocus multilayer snowpack model and radiative transfer models. The microwave snow emission model (Dense Media Radiative Transfer – Multi-Layer model, DMRT-ML), the vegetation transmissivity model (ω-τopt), and atmospheric and soil radiative transfer models were calibrated to simulate the contributions from the snowpack, the vegetation, and the soil, respectively, at the top of the atmosphere. DA experiments were performed for 12 stations where daily continuous SWE measurements were acquired over 4 winters (2012–2016). Best SWE estimates are obtained with the assimilation of the TBs at 11, 19, and 37 GHz in vertical polarizations. The overall SWE bias is reduced by 68 % compared to the original SWE simulations, from 23.7 kg m−2 without assimilation to 7.5 kg m−2 with the assimilation of the three frequencies. The overall SWE relative percentage of error (RPE) is 14.1 % (19 % without assimilation) for sites with a fraction of forest cover below 75 %, which is in the range of accuracy needed for hydrological applications. This research opens the way for global applications to improve SWE estimates over large and remote areas, even when vegetation contributions are up to 50 % of the PMW signal.

Evaluation of Simulated Snow and Snowmelt Timing in the Community Land Model Using Satellite-based Products and Streamflow Observations.

The purpose of this study was to evaluate snow and snowmelt simulated by version 4 of the Community Land Model (CLM4). We performed uncoupled CLM4 simulations, forced by Modem-Era Retrospective Analysis for Research and Applications Land-only (MERRA-Land) meteorological fields. GlobSnow snow cover fraction (SCF), snow water equivalent (SWE) and satellite-based passive microwave (PMW) snowmelt-off day of year (MoD) data were used to evaluate SCF, SWE, and snowmelt simulations. Simulated runoff was then fed into a river routing scheme and evaluation was performed at 408 snow-dominated catchments using gauge observations. CLM4 and GlobSnow snow cover extent showed a strong agreement, especially during the peak snow cover months. Overall there was a good correlation between simulated and observed SWE (correlation coefficient, R = 0.6). Simulated and observed SWE were similar over areas with relatively flat terrain and moderate forest density. The simulated MoD agreed (MoD differences (CLM4-PMW) = +/-7 days) with observations over 39.4% of the study domain. Snowmelt-off occurred earlier in the model compared to the observations over 39.5 % of the domain and later over 21.1% of the domain. Large differences of MoD were seen in the areas with complex terrain and dense forest cover. We also found that, although streamflow seasonal phase was accurately modeled (R=0.9), the peaks controlled by snowmelt were underestimated. Routed CLM4 streamflow tended to occur early (by 10 days on average).

Simulation and Assimilation of Passive Microwave Data Using a Snowpack Model Coupled to a Calibrated Radiative Transfer Model Over Northeastern Canada

AbstractOver northern snowmelt‐dominated basins, the snow water equivalent (SWE) is of primary interest for hydrological forecasting. This paper evaluates first the performance of a detailed multilayer snowpack model (Crocus), driven by meteorological predictions generated by the Canadian Global Environmental Multiscale model, for hydrological applications. Simulations were compared to daily snow depth and SWE measurements over Québec, northeastern Canada (56–45°N), for 2012–2016, highlighting an overestimation of the annual maximum snow depth (35%) and of the annual maximum SWE (16%), which is not accurate enough for hydrological applications. To improve SWE simulations, a chain of models is implemented to simulate and to assimilate passive microwave satellite observations. The snowpack model is coupled to a microwave snow emission model (Dense Media Radiative Transfer‐Multilayers model, DMRT‐ML), and the comparison of simulated brightness temperatures (TBs) with surface‐based TB measurements (at 11, 19 and 37 GHz) shows best results when the snow stickiness parameter is set to 0.17 in DMRT‐ML. The overall root‐mean‐square error (RMSE) obtained by the calibrated coupling reaches 27 K, significantly better than the RMSE obtained by considering nonsticky spheres in DMRT‐ML (43.0 K). The relevance of TB assimilation is tested with synthetic observations to evaluate the information content of each frequency for SWE estimates. The assimilation scheme is a Sequential Importance Resampling Particle filter using an ensemble of perturbed meteorological forcing data. The results show a SWE RMSE reduced by 82% with TB assimilation compared to without assimilation.

Independent Evaluation of Frozen Precipitation from WRF and PRISM in the Olympic Mountains

Abstract Estimates of precipitation from the Weather Research and Forecasting (WRF) Model and the Parameter-Elevation Regressions on Independent Slopes Model (PRISM) are widely used in complex terrain to obtain spatially distributed precipitation data. The authors evaluated both WRF (4/3 km) and PRISM’s (800-m annual climatology) ability to estimate frozen precipitation using the hydrologic model Structure for Unifying Multiple Modeling Alternatives (SUMMA) and a unique set of spatiotemporal snow depth and snow water equivalent (SWE) observations collected for the Olympic Mountain Experiment (OLYMPEX) ground validation campaign during water year 2016. When SUMMA was forced with WRF precipitation and used a calibrated, wet-bulb-temperature-based method for partitioning rain versus snow, its estimation of near-peak SWE was biased low by 21% on average. However, when SUMMA was allowed to partition WRF total precipitation into rain and snow based on output from WRF’s microphysical scheme (WRFMPP), simulations of snow depth and SWE were near equal to or better than simulations that used PRISM-derived precipitation with the calibrated partitioning method. Over all sites, WRFMPP and simulations that used PRISM-derived precipitation had relatively unbiased estimates of near-peak SWE, but both simulated absolute errors in near-peak SWE of 30%–60% at a few locations. Since, on average, WRFMPP had similar errors to PRISM, WRFMPP suggested a promising path forward in hydrology, as it was independent of gauge data and did not require SWE observations for calibration. Furthermore, in similar maritime environments, hydrologic modelers should pay close attention to decisions regarding rain-versus-snow partitioning, wind speed, and incoming longwave radiation.

Snow redistribution for the hydrological modeling of alpine catchments

Modeling snow redistribution by wind and avalanches in hydrological studies in alpine catchments is important, as the spatial variability of the snow cover has an impact on timing and magnitude of the snowmelt runoff. Disregarding snow redistribution in models can lead to the formation of ‘snow towers,’ i.e., multi‐year accumulation of snow at high elevations and an incorrect water balance. The reviewed approaches to deal with snow redistribution in hydrological models were first broadly grouped by the represented physical processes: (1) the correction of the precipitation input data to account mainly for preferential deposition, (2) the description of all wind‐driven processes based on wind field data, (3) the description of gravitational transports and/or wind‐driven processes based on topographic information, and (4) the statistical description of the variability of the snow water equivalent (SWE) to account for all types of snow redistribution. The review further assessed the implementation of these approaches in physically based and bucket‐type hydrological models. Generally, snow redistribution consideration has improved the simulation of snow patterns and SWE and consequently the prediction of discharge in mountain catchments worldwide. Snow redistribution approaches still have some limitations and a large gap exists between the knowledge and processes in highly detailed physically based snow models and the widely used bucket‐type hydrological models used for water resources and climate change studies. There is a real need to bridge this gap using the knowledge earned by snow redistribution modeling with established physically based models to develop more conceptual approaches for the application in bucket‐type models. WIREs Water 2017, 4:e1232. doi: 10.1002/wat2.1232This article is categorized under: Science of Water > Hydrological Processes Science of Water > Methods Science of Water > Water and Environmental Change

Evaluation of snow storage in Western Siberia based on the land-surface model SPONSOR simulation using reanalysis data

Obtaining of reliable information about the characteristics of snow cover with high spatial and temporal resolution for large areas of Northern Eurasia, with rare or absent network of ground-based observations stations is an important and urgent task. Currently estimation of the value of the snow water equivalent (SWE) and the snow depth have a large degree of uncertainty, especially if we are moving from data at the point of observation stations to distributed space values. In this article, the simulations of SWE and the snow depth using Land-Surface Model (LSM) SPONSOR with input meteorological data taken from the ECMWF ERAInterim reanalysis was performed for Western Siberia for the period from 1979 to 2013. Fields of SWE and of the snow depth with high spatial and temporal resolution corresponding to the resolution of meteorological data of the ECMWF ERA-Interim reanalysis (time step of 6 hours, the grid resolution of 0.75° × 0.75° in latitude and longitude) were obtained. For the entire period SWE data were compared with observations, as simulated using the model and taken directly from the reanalysis ERA-Interim at points corresponding of observation stations. Also comparison of observations and satellite data of SWE for points of observation stations was performed. Correlation coefficients between observations and model and satellite data for SWE and the snow depth were calculated for the period from 1979 to 2013. These correlation coefficients between observations and results of simulations using LSM SPONSOR for SWE, and especially for the snow depth are the best of all methods. Maps with high spatial resolution for SWE, obtained by different methods, were constructed for February averaged. Comparing of constructed maps shows significant uncertainty of the SWE fields, besides field’s distortions are not evenly distributed across the region. It appears that no one of these methods currently can be used as a reference (unique) to determine SWE in the absence of data of ground-based observations. Overall, model simulations using LSM SPONSOR somewhat overstate SWE, however, this overestimation is not more than 10–20% for most part of the territory, except in the South. Model data are reasonably well reproduce SWE for Central, Eastern and, most probably, for Northern parts of the region, differing from a real at 10–15%. Data from used satellite archive a few underestimate of SWE. SWE data taken directly from the reanalysis ERA-Interim, give large distortions of the SWE field: these values for Northern parts of the region, are likely greatly underestimated, and for Western and Eastern parts of the region – inflated. It is shown that in general, the method of simulation of snow cover characteristics using LSM SPONSOR with input data taken from the ECMWF ERA-Interim reanalysis gives good results for the region of Western Siberia.

A model for the spatial distribution of snow water equivalent parameterized from the spatial variability of precipitation

Abstract. Snow is an important and complicated element in hydrological modelling. The traditional catchment hydrological model with its many free calibration parameters, also in snow sub-models, is not a well-suited tool for predicting conditions for which it has not been calibrated. Such conditions include prediction in ungauged basins and assessing hydrological effects of climate change. In this study, a new model for the spatial distribution of snow water equivalent (SWE), parameterized solely from observed spatial variability of precipitation, is compared with the current snow distribution model used in the operational flood forecasting models in Norway. The former model uses a dynamic gamma distribution and is called Snow Distribution_Gamma, (SD_G), whereas the latter model has a fixed, calibrated coefficient of variation, which parameterizes a log-normal model for snow distribution and is called Snow Distribution_Log-Normal (SD_LN). The two models are implemented in the parameter parsimonious rainfall–runoff model Distance Distribution Dynamics (DDD), and their capability for predicting runoff, SWE and snow-covered area (SCA) is tested and compared for 71 Norwegian catchments. The calibration period is 1985–2000 and validation period is 2000–2014. Results show that SD_G better simulates SCA when compared with MODIS satellite-derived snow cover. In addition, SWE is simulated more realistically in that seasonal snow is melted out and the building up of "snow towers" and giving spurious positive trends in SWE, typical for SD_LN, is prevented. The precision of runoff simulations using SD_G is slightly inferior, with a reduction in Nash–Sutcliffe and Kling–Gupta efficiency criterion of 0.01, but it is shown that the high precision in runoff prediction using SD_LN is accompanied with erroneous simulations of SWE.

Assimilation of AMSR-E snow water equivalent data in a spatially-lumped snow model

Summary Accurately initializing snow model states in hydrologic prediction models is important for estimating future snowmelt, water supplies, and flooding potential. While ground-based snow observations give the most reliable information about snowpack conditions, they are spatially limited. In the north-central USA, there are no continual observations of hydrologically critical snow variables. Satellites offer the most likely source of spatial snow data, such as the snow water equivalent (SWE), for this region. In this study, we test the impact of assimilating SWE data from the Advanced Microwave Scanning Radiometer – Earth Observing System (AMSR-E) instrument into the US National Weather Service (NWS) SNOW17 model for seven watersheds in the Upper Mississippi River basin. The SNOW17 is coupled with the NWS Sacramento Soil Moisture Accounting (SACSMA) model, and both simulated SWE and discharge are evaluated. The ensemble Kalman filter (EnKF) assimilation framework is applied and updating occurs on a daily cycle for water years 2006–2011. Prior to assimilation, AMSR-E data is bias corrected using data from the National Operational Hydrologic Remote Sensing Center (NOHRSC) airborne snow survey program. An average AMSR-E SWE bias of −17.91 mm was found for the study basins. SNOW17 and SAC-SMA model parameters from the North Central River Forecast Center (NCRFC) are used. Compared to a baseline run without assimilation, the SWE assimilation improved discharge for five of the seven study sites, in particular for high discharge magnitudes associated with snow melt runoff. SWE and discharge simulations suggest that the SNOW17 is underestimating SWE and snowmelt rates in the study basins. Deep snow conditions and periods of snowmelt may have introduced error into the assimilation due to difficulty obtaining accurate brightness temperatures under these conditions. Overall results indicate that the AMSR-E data and EnKF are viable and effective solutions for improving simulations of the operational forecast model.