Snow Data Assimilation System Research Articles

Enabling Advanced Snow Physics Within Land Surface Models Through an Interoperable Model‐Physics Coupling Framework

AbstractAccurate estimation of snow accumulation and melt is a critical part of decision‐making in snow‐dominated watersheds. In this study, we demonstrate a flexible methodology to couple a detailed snow model, Crocus, separately to two different land surface models (LSMs), Noah‐MP and Noah. The original LSMs and the coupled models (Noah‐MP‐Crocus and Noah‐Crocus) are used to simulate snow depth, snow water equivalent, and other water and energy states and fluxes. The results of simulations are compared against a wide range of independent gridded and point scale reference data sets. Our results show that coupling the detailed snow model, Crocus, with the LSMs improves the snow depth and snow water equivalent relative to independent observations. Overall, larger improvements are obtained with coupling Crocus to the Noah LSM, with the coupled Noah‐Crocus configuration reducing the RMSE and bias of snow depth from 2% to 12% and 57% to 75%, respectively, relative to Snow Data Assimilation System (SNODAS) and snow product from the University of Arizona. On the other hand, smaller improvements are obtained by coupling Crocus with Noah‐MP. The Coupled Noah‐MP‐Crocus reduces the snow depth bias but slightly degrades the RMSE of snow depth and snow water equivalent. The corresponding impacts in other water budget terms such as evapotranspiration, soil moisture, and streamflow, however, are mixed, pointing to the significant need to improve the coupling assumptions of these processes within land models. Overall, the interoperable coupling framework demonstrated here offers the opportunity to include more detailed snow physics and processes, and to advance data assimilation systems through improved exploitation of information from snow remote sensing instruments.

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.

Spatio-temporal information propagation using sparse observations in hyper-resolution ensemble-based snow data assimilation

Abstract. Data assimilation techniques that integrate available observations with snow models have been proposed as a viable option to simultaneously help constrain model uncertainty and add value to observations by improving estimates of the snowpack state. However, the propagation of information from spatially sparse observations in high-resolution simulations remains an under-explored topic. To remedy this, the development of data assimilation techniques that can spread information in space is a crucial step. Herein, we examine the potential of spatio-temporal data assimilation for integrating sparse snow depth observations with hyper-resolution (5 m) snow simulations in the Izas central Pyrenean experimental catchment (Spain). Our experiments were developed using the Multiple Snow Data Assimilation System (MuSA) with new improvements to tackle the spatio-temporal data assimilation. Therein, we used a deterministic ensemble smoother with multiple data assimilation (DES-MDA) with domain localization. Three different experiments were performed to showcase the capabilities of spatio-temporal information transfer in hyper-resolution snow simulations. Experiment I employed the conventional geographical Euclidean distance to map the similarity between cells. Experiment II utilized the Mahalanobis distance in a multi-dimensional topographic space using terrain parameters extracted from a digital elevation model. Experiment III utilized a more direct mapping of snowpack similarity from a single complete snow depth map together with the easting and northing coordinates. Although all experiments showed a noticeable improvement in the snow patterns in the catchment compared with the deterministic open loop in terms of correlation (r=0.13) and root mean square error (RMSE = 1.11 m), the use of topographical dimensions (Experiment II, r=0.63 and RMSE = 0.89 m) and observations (Experiments III, r=0.92 and RMSE = 0.44 m) largely outperform the simulated patterns in Experiment I (r=0.38 and RMSE = 1.16 m). At the same time, Experiments II and III are considerably more challenging to set up. The results of these experiments can help pave the way for the creation of snow reanalysis and forecasting tools that can seamlessly integrate sparse information from national monitoring networks and high-resolution satellite information.

Spatiotemporal snow water storage uncertainty in the midlatitude American Cordillera

Abstract. This work quantifies the uncertainty of accumulation-season peak snow water storage in the portions of the midlatitude American Cordillera where snow is a dominant driver of hydrology. This is accomplished through intercomparison of commonly used global and regional products over the Western United States (WUS) and Andes domains, which have similar hydrometeorology but are disparate with respect to the amount of available in situ information. The recently developed WUS Snow Reanalysis (WUS-SR) and Andes Snow Reanalysis (Andes-SR) datasets, which have been extensively verified against in situ measurements, are used as baseline reference datasets in the intercomparison. Relative to WUS-SR climatological peak snow water equivalent (SWE) storage (269 km3), high- and moderate-resolution products (i.e., those with resolutions less than ∼10 km) are in much better agreement (284±14 km3; overestimated by 6 %) compared to low-resolution products (127±54 km3; underestimated by 53 %). In comparison to the Andes-SR peak snow storage (29 km3), all other products show large uncertainty and bias (19±16 km3; underestimated by 34 %). Examination of spatial patterns related to orographic effects showed that only the high- to moderate-resolution Snow Data Assimilation System (SNODAS) and University of Arizona (UA) products show comparable estimates of windward–leeward SWE patterns over a subdomain (Sierra Nevada) of the WUS. Coarser products distribute too much snow on the leeward side in both the Sierra Nevada and Andes, missing orographic and rain shadow patterns that have important hydrological implications. The uncertainty of peak seasonal snow storage is primarily explained by precipitation uncertainty in both the WUS (R2=0.55) and Andes (R2=0.84). Despite using similar forcing inputs, snow storage diverges significantly within the ECMWF Reanalysis v5 (ERA5) (i.e., ERA5 vs. ERA5-Land) products and the Global Land Data Assimilation System (GLDAS) (modeled with Noah, Variable Infiltration Capacity (VIC), and Catchment model) products due to resolution-induced elevation differences and/or differing model process representation related to rain–snow partitioning and accumulation-season snowmelt generation. The availability and use of in situ precipitation and snow measurements (i.e., in WUS) in some products adds value by reducing snow storage uncertainty; however, where such data are limited, i.e., in the Andes, significant biases and uncertainty exist.

Snowpack relative permittivity and density derived from near‐coincident lidar and ground‐penetrating radar

AbstractDepth‐based and radar‐based remote sensing methods (e.g., lidar, synthetic aperture radar) are promising approaches for remotely measuring snow water equivalent (SWE) at high spatial resolution. These approaches require snow density estimates, obtained from in‐situ measurements or density models, to calculate SWE. However, in‐situ measurements are operationally limited, and few density models have seen extensive evaluation. Here, we combine near‐coincident, lidar‐measured snow depths with ground‐penetrating radar (GPR) two‐way travel times (twt) of snowpack thickness to derive >20 km of relative permittivity estimates from nine dry and two wet snow surveys at Grand Mesa, Cameron Pass, and Ranch Creek, Colorado. We tested three equations for converting dry snow relative permittivity to snow density and found the Kovacs et al. (1995) equation to yield the best comparison with in‐situ measurements (RMSE = 54 kg m−3). Variogram analyses revealed a 19 m median correlation length for relative permittivity and snow density in dry snow, which increased to >30 m in wet conditions. We compared derived densities with estimated densities from several empirical models, the Snow Data Assimilation System (SNODAS), and the physically based iSnobal model. Estimated and derived densities were combined with snow depths and twt to evaluate density model performance within SWE remote sensing methods. The Jonas et al. (2009) empirical model yielded the most accurate SWE from lidar snow depths (RMSE = 51 mm), whereas SNODAS yielded the most accurate SWE from GPR twt (RMSE = 41 mm). Densities from both models generated SWE estimates within ±10% of derived SWE when SWE averaged >400 mm, however, model uncertainty increased to >20% when SWE averaged <300 mm. The development and refinement of density models, particularly in lower SWE conditions, is a high priority to fully realize the potential of SWE remote sensing methods.

A genetic particle filter scheme for univariate snow cover assimilation into Noah-MP model across snow climates

Abstract. Accurate snowpack simulations are critical for regional hydrological predictions, snow avalanche prevention, water resource management, and agricultural production, particularly during the snow ablation period. Data assimilation methodologies are increasingly being applied for operational purposes to reduce the uncertainty in snowpack simulations and to enhance their predictive capabilities. This study aims to investigate the feasibility of using a genetic particle filter (GPF) as a snow data assimilation scheme designed to assimilate ground-based snow depth (SD) measurements across different snow climates. We employed the default parameterization scheme combination within the Noah-MP (with multi-parameterization) model as the model operator in the snow data assimilation system to evolve snow variables and evaluated the assimilation performance of the GPF using observational data from sites with different snow climates. We also explored the impact of measurement frequency and particle number on the filter updating of the snowpack state at different sites and the results of generic resampling methods compared to the genetic algorithm used in the resampling process. Our results demonstrate that a GPF can be used as a snow data assimilation scheme to assimilate ground-based measurements and obtain satisfactory assimilation performance across different snow climates. We found that particle number is not crucial for the filter's performance, and 100 particles are sufficient to represent the high dimensionality of the point-scale system. The frequency of measurements can significantly affect the filter-updating performance, and dense ground-based snow observational data always dominate the accuracy of assimilation results. Compared to generic resampling methods, the genetic algorithm used to resample particles can significantly enhance the diversity of particles and prevent particle degeneration and impoverishment. Finally, we concluded that the GPF is a suitable candidate approach for snow data assimilation and is appropriate for different snow climates.

Intercomparison of snow water equivalent products in the Sierra Nevada California using airborne snow observatory data and ground observations

Whereas many independent methods are used to estimate snow water equivalent (SWE) and its spatial distribution and seasonal variability, a need exists for a systematic characterization of inter-model differences at annual, seasonal, and regional scales necessary to quantify the associated uncertainty in these datasets. This study conducts a multi-scale validation and comparison, based on Airborne Snow Observatory data, of five state-of-the-art SWE datasets in the Sierra Nevada, California, including three SWE datasets from retrospective models: an INiTial REConstruction model (REC-INT), an improved REConstruction model based on the ParBal energy balance model (REC-ParBal), and a Sierra Nevada SWE REConstruction with Data Assimilation (REC-DA), and two operational SWE datasets from the U.S. National Weather Service, including the Snow Data Assimilation System (SNODAS) and the National Water Model (NWM-SWE). The results show that REC-DA and REC-ParBal provide the two most accurate estimates of SWE in the snowmelt season, both with small positive biases. REC-DA provides the most accurate spatial distribution of SWE (R2 = 0.87, MAE = 66 mm, PBIAS = 8.3%) at the pixel scale, while REC-ParBal has the least basin-wide PBIAS (R2 = 0.79, MAE = 73 mm, PBIAS = 4.1%) in the snowmelt season. Moreover, REC-DA underestimates peak SWE by −5.8%, while REC-ParBal overestimates it by 7.5%, when compared with the measured peak SWE at snow pillow stations across the Sierra Nevada. The two operational SWE products—SNODAS and NWM-SWE—are less accurate. Furthermore, the inter-model comparison reveals a certain amount of disagreement in snow water storage across time and space between SWE datasets. This study advances our understanding of regional SWE uncertainties and provides critical insights to support future applications of these SWE data products and therefore has broad implications for water resources management and hydrological process studies.

Evaluation of E3SM land model snow simulations over the western United States

Abstract. Seasonal snow has crucial impacts on climate, ecosystems, and humans, but it is vulnerable to global warming. The land component (ELM) of the Energy Exascale Earth System Model (E3SM) mechanistically simulates snow processes from accumulation, canopy interception, compaction, and snow aging to melt. Although high-quality field measurements, remote sensing snow products, and data assimilation products with high spatio-temporal resolution are available, there has been no systematic evaluation of the snow properties and phenology in ELM. This study comprehensively evaluates ELM snow simulations over the western United States at 0.125∘ resolution during 2001–2019 using the Snow Telemetry (SNOTEL) in situ networks, MODIS remote sensing products (i.e., MCD43 surface albedo product), the spatially and temporally complete (STC) snow-covered area and grain size (MODSCAG) and MODIS dust and radiative forcing in snow (MODDRFS) products (STC-MODSCAG/STC-MODDRFS), and the snow property inversion from remote sensing (SPIReS) product and two data assimilation products of snow water equivalent and snow depth – i.e., University of Arizona (UA) and SNOw Data Assimilation System (SNODAS). Overall the ELM simulations are consistent with the benchmarking datasets and reproduce the spatio-temporal patterns, interannual variability, and elevation gradients for different snow properties including snow cover fraction (fsno), surface albedo (αsur) over snow cover regions, snow water equivalent (SWE), and snow depth (Dsno). However, there are large biases of fsno with dense forest cover and αsur in the Rocky Mountains and Sierra Nevada in winter, compared to the MODIS products. There are large discrepancies of snow albedo, snow grain size, and light-absorbing particle-induced snow albedo reduction between ELM and the MODIS products, attributed to uncertainties in the aerosol forcing data, snow aging processes in ELM, and remote sensing retrievals. Against UA and SNODAS, ELM has a mean bias of −20.7 mm (−35.9 %) and −20.4 mm (−35.5 %), respectively, for spring, and −13.8 mm (−27.8 %) and −10.2 mm (−22.2 %), respectively, for winter. ELM shows a relatively high correlation with SNOTEL SWE, with mean correlation coefficients of 0.69 but negative mean biases of −122.7 mm. Compared to the snow phenology of STC-MODSCAG and SPIReS, ELM shows delayed snow accumulation onset dates by 17.3 and 12.4 d, earlier snow end dates by 35.5 and 26.8 d, and shorter snow durations by 52.9 and 39.5 d, respectively. This study underscores the need for diagnosing model biases and improving ELM representations of snow properties and snow phenology in mountainous areas for more credible simulation and future projection of mountain snowpack.

The Multiple Snow Data Assimilation System (MuSA v1.0)

Abstract. Accurate knowledge of the seasonal snow distribution is vital in several domains including ecology, water resources management, and tourism. Current spaceborne sensors provide a useful but incomplete description of the snowpack. Many studies suggest that the assimilation of remotely sensed products in physically based snowpack models is a promising path forward to estimate the spatial distribution of snow water equivalent (SWE). However, to date there is no standalone, open-source, community-driven project dedicated to snow data assimilation, which makes it difficult to compare existing algorithms and fragments development efforts. Here we introduce a new data assimilation toolbox, the Multiple Snow Data Assimilation System (MuSA), to help fill this gap. MuSA was developed to fuse remotely sensed information that is available at different timescales with the energy and mass balance Flexible Snow Model (FSM2). MuSA was designed to be user-friendly and scalable. It enables assimilation of different state variables such as the snow depth, SWE, snow surface temperature, binary or fractional snow-covered area, and snow albedo and could be easily upgraded to assimilate other variables such as liquid water content or snow density in the future. MuSA allows the joint assimilation of an arbitrary number of these variables, through the generation of an ensemble of FSM2 simulations. The characteristics of the ensemble (i.e., the number of particles and their prior covariance) may be controlled by the user, and it is generated by perturbing the meteorological forcing of FSM2. The observational variables may be assimilated using different algorithms including particle filters and smoothers as well as ensemble Kalman filters and smoothers along with their iterative variants. We demonstrate the wide capabilities of MuSA through two snow data assimilation experiments. First, 5 m resolution snow depth maps derived from drone surveys are assimilated in a distributed fashion in the Izas catchment (central Pyrenees). Furthermore, we conducted a joint-assimilation experiment, fusing MODIS land surface temperature and fractional snow-covered area with FSM2 in a single-cell experiment. In light of these experiments, we discuss the pros and cons of the assimilation algorithms, including their computational cost.

Modeling Spatial Distribution of Snow Water Equivalent by Combining Meteorological and Satellite Data with Lidar Maps

Abstract An accurate characterization of the water content of snowpack, or snow water equivalent (SWE), is necessary to quantify water availability and constrain hydrologic and land surface models. Recently, airborne observations (e.g., lidar) have emerged as a promising method to accurately quantify SWE at high resolutions (scales of ∼100 m and finer). However, the frequency of these observations is very low, typically once or twice per season in the Rocky Mountains of Colorado. Here, we present a machine learning framework that is based on random forests to model temporally sparse lidar-derived SWE, enabling estimation of SWE at unmapped time points. We approximated the physical processes governing snow accumulation and melt as well as snow characteristics by obtaining 15 different variables from gridded estimates of precipitation, temperature, surface reflectance, elevation, and canopy. Results showed that, in the Rocky Mountains of Colorado, our framework is capable of modeling SWE with a higher accuracy when compared with estimates generated by the Snow Data Assimilation System (SNODAS). The mean value of the coefficient of determination R2 using our approach was 0.57, and the root-mean-square error (RMSE) was 13 cm, which was a significant improvement over SNODAS (mean R2 = 0.13; RMSE = 20 cm). We explored the relative importance of the input variables and observed that, at the spatial resolution of 800 m, meteorological variables are more important drivers of predictive accuracy than surface variables that characterize the properties of snow on the ground. This research provides a framework to expand the applicability of lidar-derived SWE to unmapped time points. Significance Statement Snowpack is the main source of freshwater for close to 2 billion people globally and needs to be estimated accurately. Mountainous snowpack is highly variable and is challenging to quantify. Recently, lidar technology has been employed to observe snow in great detail, but it is costly and can only be used sparingly. To counter that, we use machine learning to estimate snowpack when lidar data are not available. We approximate the processes that govern snowpack by incorporating meteorological and satellite data. We found that variables associated with precipitation and temperature have more predictive power than variables that characterize snowpack properties. Our work helps to improve snowpack estimation, which is critical for sustainable management of water resources.

Constrained Inversion of a Microwave Snowpack Emission Model Using Dictionary Matching: Applications for GPM Satellite

This article presents a new algorithmic framework for multilayer inversion of the dense media radiative transfer (DMRT) equations of snowpack emission, with particular emphasis on the role of high-frequency microwave channels above 60 GHz. The approach relies on dictionary matching and locally constrained least squares. The results demonstrate that the algorithm can invert the DMRT model and retrieve depth, density, and grain size of a single-layer snowpack when dependencies of density and grain size on depth are properly accounted for. However, as the number of layers increases, the sensitivity of the inversion to observation noise grows markedly. Using observations, over the Great Plains in the United States, from the microwave imager onboard the global precipitation measurement (GPM, 10-166 GHz) core satellite, the initial results demonstrate that under a clear-sky condition and no vegetation canopy, the algorithm is capable to retrieve the snow depth and water equivalent of seasonal snow with a mean absolute error (MAE) of less than 0.15 m--when compared to the high-resolution analysis data from the SNOw Data Assimilation System (SNODAS).

Projections of Snow Water Equivalent Using a Process-Based Energy Balance Snow Model in Southwestern British Columbia

Abstract Information about snow water equivalent in southwestern British Columbia, Canada, is used for flood management, agriculture, fisheries, and water resource planning. This study evaluates whether a process-based, energy balance snow model supplied with high-resolution statistically downscaled temperature and precipitation data can effectively simulate snow water equivalent (SWE) in the mountainous terrain of this region. Daily values of SWE from 1951 to 2018 are simulated at 1-km resolution and evaluated using a reanalysis SWE product [Snow Data Assimilation System (SNODAS)], manual snow-survey measurements at 41 sites, and automated snow pillows at six locations in the study region. Simulated SWE matches observed interannual variability well (R2 > 0.8 for annual maximum SWE), but peak SWE biases of 20%–40% occur at some sites in the study domain, and higher biases occur where observed SWE is very low. Modeled SWE displays lower bias relative to SNODAS reanalysis at most manual survey locations. Future projections for the study area are produced using 12 downscaled climate model simulations and are used to illustrate the impacts of climate change on SWE at 1°, 2°, and 3°C of warming. Model results are used to quantify spring SWE changes at different elevations of the Whistler mountain ski resort and the sensitivity of annual peak SWE in the Metropolitan Vancouver municipal watersheds to moderate temperature increases. The results both illustrate the potential utility of a process-based snow model and identify areas where the input meteorological variables could be improved. Significance Statement Using high-resolution (1 km) climate data, we evaluate and apply a snow model in the mountainous terrain of coastal, southwestern British Columbia, Canada. Modeling snow water equivalent at high-resolution enables better representation of snow conditions that can vary widely over short distances and elevations. At 1°, 2°, and 3°C of warming, future snow water equivalent levels at sites nearer the coast are more vulnerable to temperature increases than sites slightly higher in elevation and farther inland. Future efforts to improve the climate data may yield better agreement between simulated and observed snow levels in certain locations.

Combining ground-based and remotely sensed snow data in a linear regression model for real-time estimation of snow water equivalent

Effective water resources management in California relies substantially on real-time information of snow water equivalent (SWE) at basin scale and mountain ranges given that mountain snowpacks provide the primary water supply for the State. However, SWE estimation based solely on remote sensing, modeling, or ground observations does not meet contemporary operational requirements. In this context, this study develops a data-fusion framework that combines multi-source datasets including satellite-observed daily mean fractional snow-covered area (DMFSCA), snow pillow SWE measurements, physiographic data, and historical SWE patterns into a linear regression model (LRM) to improve SWE estimates in real-time. We test two LRMs: a baseline regression model (LRM-baseline) that uses physiographic data and historical SWE patterns as independent variables, and an FSCA-informed regression model (LRM-FSCA) that includes the DMFSCA from Moderate Resolution Imaging Spectroradiometer (MODIS) satellite imagery as an additional independent variable. By incorporating the satellite-observed DMFSCA, LRM-FSCA outperforms LRM-baseline with increased median R2 from 0.54 to 0.60, and reduced median PBIAS of basin average SWE from 2.6% to 2.2% in the snow pillow SWE cross-validation. LRM-FSCA explains 87% of the variance in the snow course SWE measurements with 0.1% PBIAS, while LRM-baseline explains a lower 81% variance with 1.4% PBIAS, both of which show higher accuracy than SWE estimates from the two operational SWE datasets: the Snow Data Assimilation System (SNODAS, 73% and -2.4%, respectively) and Nationtional Water Model (NWM, 75% and -15.9%, respectively). Additionally, LRM-FSCA explains 85% of the median variance in the Airborne Snow Observatory SWE with -9.2% PBIAS, which is comparable to the LRM-baseline (86% and -11.3%, respectively) and considerably better than SNODAS (64% and 28.2%, respectively) and NWM (33% and -30.1%, respectively). This study shows a substantial model improvement by constraining the geographical and seasonal variation on snow-cover via satellite observation and highlights the values of using multi-source observations in real-time SWE estimation. The developed SWE estimation framework has crucial implications for effective water supply forecasting and management in California, where climate extremes (e.g., droughts and floods) require particularly skillful monitoring practices.

Seasonal hydrological loading in the Great Lakes region detected by GNSS: a comparison with hydrological models

SUMMARY Crustal deformation caused by hydrological processes has long been detected using space geodetic techniques, yet questions remain about the relative contributions of surface water and groundwater to the geodetic signals in different regions. Here, we investigate forward models of elastic loading deformation caused by a variety of water-storage changes within the Great Lakes region, including fluctuations in lake-water volume, soil moisture and snow load. We use lake-level data from the Great Lakes Environmental Research Laboratory, soil-moisture content from the North American Land Data Assimilation System (NLDAS), snow load from the Snow Data Assimilation System (SNODAS) and background hydrological load at the global scale from Gravity Recovery and Climate Experiment (GRACE). We compare the modelled surface deformation with estimates of hydrological loading deformation inferred from Global Navigation Satellite System (GNSS) measurements. We find that seasonal deformation measured by GNSS is dominated by regional-scale hydrological loading based on strong correlations with the modelled loading displacements. The mean correlation coefficient for the study network is 0.56. The correlation coefficients vary spatially within the study region and exceed 0.9 at some stations near to the Great Lakes. We assess the relative contribution of each individual hydrological component to the total integrated hydrological load. We find that soil moisture consistently explains the largest percentage (27–69 per cent) of the total vertical loading deformation for 87 per cent of GNSS stations in the Great Lakes region. Snow loading and soil moisture contribute relatively equally in the northern reaches of the study area (e.g. Canadian shield, northern Superior basin). Lake loading accounts for about 10–25 per cent of the total loading signal in the immediate vicinity of the lakes. We also investigate the sensitivities of the surface loading displacements to three different Earth models, including two with lateral variations in structure. The structural variations considered here have limited impact (<0.2 mm) on the predicted hydrological loading displacements and could be neglected at the current level of observational precision.

Evaluation and blending of ATMS and AMSR2 snow water equivalent retrievals over the conterminous United States

This study first compares two different passive microwave snow water equivalent (SWE) retrievals, namely the retrieval from the Suomi National Polar-orbiting Partnership (S-NPP) Advanced Technology Microwave Sounder (ATMS) and that from the Global Change Observation Mission – Water (GCOM-W1) Advanced Microwave Scanning Radiometer 2 (AMSR2); it further creates an optimal blending mechanism that merges the two retrievals with in situ observations from the Snow Telemetry (SNOTEL) and Cooperative Observer Program (COOP) networks. The assessments of the two products are done over conterminous United States (CONUS) for the snow seasons (November–June) of the water years 2017–2019 using in situ data and the SNOw Data Assimilation System (SNODAS) SWE analysis. Both satellite products tend to underestimate SWE. Between the two, AMSR2 retrieval outperforms in terms of correlation with observations and depth of saturation, but it exhibits a distinctive, seasonally varying bias that is not seen in ATMS retrieval. The negative bias over the early snow season, as further analysis indicates, most likely stems from AMSR2 retrieval's use of a high frequency channel (i.e., 89 GHz) for shallow snow detection, while the impact of differing assumptions of snow density is marginal. The blending scheme, developed on the basis of the validation experiment, features a histogram-based bias correction as a supplement to optimal interpolation. Cross-validation suggests that interpolated station product without the satellite background broadly underperforms the blended in situ-satellite product, confirming the utility of the satellite retrievals. Furthermore, the a priori bias correction mechanism is shown to be effective in mitigating large fluctuations in bias. Finally, the bias-corrected, blended in situ-satellite product performs comparably or even favorably against SNODAS over many parts of the CONUS, with important implications for joint use of satellite and in situ observations for hydrological monitoring and forecasting.

Do sub-daily temperature fluctuations around the freezing temperature alter macro-scale snowmelt simulations?

Climate-driven temperature variations manipulate the snowmelt process and affect various aspects of aquatic and terrestrial life. The objective of this study is to evaluate the impacts of sub-daily temperature fluctuations around the freezing temperature on the quantity and spatial distribution of macro-scale snowmelt simulations (i.e., over macro-scale basins and monthly or annual temporal scales). A macro-scale snow accumulation and ablation model was developed to account for sub-daily temperature fluctuations. Unlike other existing macro-scale models that rely upon a single daily average temperature, the new model incorporates minimum and maximum temperatures and their occurrence timing within the day in a hybrid temperature index method (HTIM). The model was applied to the Missouri River Basin (MRB) to simulate grid-based snowmelt for water years 2011 and 2012 and compared with the monthly snowmelt data from the SNOw Data Assimilation System (SNODAS), which is a physically-based energy and mass balance snow model. Also, the HTIM was compared with a standard TIM, in which only a daily average temperature was used for snowmelt simulations. Although the HTIM provided simulations comparable to the SNODAS snowmelt data (R2 = 0.9), overall results indicated that the TIM slightly outperformed the HTIM. When minimum temperatures fluctuated around the freezing temperature, HTIM underestimated the snowmelt over the MRB. However, HTIM showed a better performance particularly in the months when maximum temperatures fluctuated around the freezing temperature. Results suggested that the role of refreezing in the generation of snowmelt was overestimated by HTIM, which led to a more significant discrepancy between the SNODAS and HTIM snowmelt values in specific months. The findings from this study highlight the influence of sub-daily temperature fluctuations on macro-scale snowmelt simulations.

Application of machine learning techniques for regional bias correction of snow water equivalent estimates in Ontario, Canada

Abstract. Snow is a critical contributor to Ontario's water-energy budget, with impacts on water resource management and flood forecasting. Snow water equivalent (SWE) describes the amount of water stored in a snowpack and is important in deriving estimates of snowmelt. However, only a limited number of sparsely distributed snow survey sites (n=383) exist throughout Ontario. The SNOw Data Assimilation System (SNODAS) is a daily, 1 km gridded SWE product that provides uniform spatial coverage across this region; however, we show here that SWE estimates from SNODAS display a strong positive mean bias of 50 % (16 mm SWE) when compared to in situ observations from 2011 to 2018. This study evaluates multiple statistical techniques of varying complexity, including simple subtraction, linear regression and machine learning methods to bias-correct SNODAS SWE estimates using absolute mean bias and RMSE as evaluation criteria. Results show that the random forest (RF) algorithm is most effective at reducing bias in SNODAS SWE, with an absolute mean bias of 0.2 mm and RMSE of 3.64 mm when compared with in situ observations. Other methods, such as mean bias subtraction and linear regression, are somewhat effective at bias reduction; however, only the RF method captures the nonlinearity in the bias and its interannual variability. Applying the RF model to the full spatio-temporal domain shows that the SWE bias is largest before 2015, during the spring melt period, north of 44.5∘ N and east (downwind) of the Great Lakes. As an independent validation, we also compare estimated snowmelt volumes with observed hydrographs and demonstrate that uncorrected SNODAS SWE is associated with unrealistically large volumes at the time of the spring freshet, while bias-corrected SWE values are highly consistent with observed discharge volumes.

Extreme Value Snow Water Equivalent and Snowmelt for Infrastructure Design Over the Contiguous United States

AbstractSnowmelt‐driven floods result in large societal and economic impacts on local communities including infrastructure failures in the United States. However, the current U.S. government standard design precipitation maps are based on liquid precipitation data (e.g., National Oceanic and Atmospheric Administration's Precipitation‐Frequency Atlas 14; NOAA Atlas 14) with very limited guidance on snowmelt‐driven floods. In this study, we developed 25‐ and 100‐year return level design maps of snow water equivalent (SWE) and 1‐ and 7‐day snowmelt including precipitation events (e.g., rain‐on‐snow) using long‐term observation‐based gridded SWE developed by University of Arizona (UA) incorporating the national snow model product (SNOw Data Assimilation System; SNODAS) over the contiguous United States (CONUS). For the 44 U.S. states where the NOAA Atlas 14 maps are available, the design snowmelt values from this study exceed the standard design values in 23% of the total extent. The snowmelt values exceed the NOAA Atlas 14 design precipitation by up to 171 and 254 mm in the northeastern United States; 127 and 225 mm in the north central United States; and 191 and 425 mm in the western mountain United States for the 25‐ and 100‐year return levels, respectively. A comparison of 7‐day design snowmelt between with and without precipitation shows that including precipitation results in an average increase of 42 and 68 mm for 25‐ and 100‐year return levels, respectively, over snowmelt that do not include precipitation. The design snowmelt maps from this study complement the NOAA Atlas 14 design precipitation and provide additional guidance on infrastructure design for snowmelt‐driven floods in the CONUS.

Assessing the factors governing the ability to predict late-spring flooding in cold-region mountain basins

Abstract. From 19 to 22 June 2013, intense rainfall and concurrent snowmelt led to devastating floods in the Canadian Rockies, foothills and downstream areas of southern Alberta and southeastern British Columbia, Canada. Such an event is typical of late-spring floods in cold-region mountain headwater, combining intense precipitation with rapid melting of late-lying snowpack, and represents a challenge for hydrological forecasting systems. This study investigated the factors governing the ability to predict such an event. Three sources of uncertainty, other than the hydrological model processes and parameters, were considered: (i) the resolution of the atmospheric forcings, (ii) the snow and soil moisture initial conditions (ICs) and (iii) the representation of the soil texture. The Global Environmental Multiscale hydrological modeling platform (GEM-Hydro), running at a 1 km grid spacing, was used to simulate hydrometeorological conditions in the main headwater basins of southern Alberta during this event. The GEM atmospheric model and the Canadian Precipitation Analysis (CaPA) system were combined to generate atmospheric forcing at 10, 2.5 and 1 km over southern Alberta. Gridded estimates of snow water equivalent (SWE) from the Snow Data Assimilation System (SNODAS) were used to replace the model SWE at peak snow accumulation and generate alternative snow and soil moisture ICs before the event. Two global soil texture datasets were also used. Overall 12 simulations of the flooding event were carried out. Results show that the resolution of the atmospheric forcing affected primarily the flood volume and peak flow in all river basins due to a more accurate estimation of intensity and total amount of precipitation during the flooding event provided by CaPA analysis at convection-permitting scales (2.5 and 1 km). Basin-averaged snowmelt also changed with the resolution due to changes in near-surface wind and resulting turbulent fluxes contributing to snowmelt. Snow ICs were the main sources of uncertainty for half of the headwater basins. Finally, the soil texture had less impact and only affected peak flow magnitude and timing for some stations. These results highlight the need to combine atmospheric forcing at convection-permitting scales with high-quality snow ICs to provide accurate streamflow predictions during late-spring floods in cold-region mountain river basins. The predictive improvement by inclusion of high-elevation weather stations in the precipitation analysis and the need for accurate mountain snow information suggest the necessity of integrated observation and prediction systems for forecasting extreme events in mountain river basins.

Improvement of operational airborne gamma radiation snow water equivalent estimates using SMAP soil moisture

Knowledge of snow water equivalent (SWE) magnitude and spatial distribution are keys to improving snowmelt flood predictions. Since the 1980s, the operational National Oceanic and Atmospheric Administration's (NOAA) airborne gamma radiation soil moisture (SM) and SWE survey has provided over 20,000 SWE observations to regional National Weather Service (NWS) River Forecast Centers (RFCs). Because the gamma SWE algorithm is based on the difference in natural gamma emission measurements from the soil between bare (fall) and snow-covered (winter) conditions, it requires a baseline fall SM for each flight line. The operational approach assumes the fall SM remains constant throughout that winter's SWE survey. However, early-winter snowmelt and rainfall events after the fall SM surveys have the potential to introduce large biases into airborne gamma SWE estimates. In this study, operational airborne gamma radiation SWE measurements were improved by updating the baseline fall SM with Soil Moisture Active Passive (SMAP) enhanced SM measurements immediately prior to winter onset over the north-central and eastern United States and southern Canada from September 2015 to April 2018. The operational airborne gamma SM had strong agreement with the SMAP SM (Pearson's correlation coefficient, R = 0.69, unbiased root mean square difference, ubRMSD = 0.057 m3/m3), compared to the Advanced Microwave Scanning Radiometer 2 (AMSR2) SM (R = 0.45, ubRMSD = 0.072 m3/m3) and the North American Land Data Assimilation System Phase 2 (NLDAS-2) Mosaic SM products (R = 0.53, ubRMSD = 0.069 m3/m3) in non-forested regions. The SMAP-enhanced gamma SWE was evaluated with satellite-based SWE (R = 0.57, ubRMSD = 34 mm) from the Special Sensor Microwave Imager Sounder (SSMIS) and in-situ SWE (R = 0.71–0.96) from the Soil Climate Analysis Network and United States Army Corps of Engineer (USACE) St. Paul District, which had better agreement than the operational gamma SWE (R = 0.48, ubRMSD = 36 mm for SSMIS and R = 0.65–0.75 for in-situ SWE). The results contribute to improving snowmelt flood predictions as well as the accuracy of the NOAA SNOw Data Assimilation System.