Snow Telemetry Research Articles

The Role of Forcing and Parameterization in Improving Snow Simulation in the Upper Colorado River Basin Using the National Water Model

AbstractThis study assesses snow water equivalent (SWE) simulation uncertainty in the National Water Model (NWM) due to forcing and model parameterization, using data from 46 Snow Telemetry (SNOTEL) sites in the Upper Colorado River Basin (UCRB). We evaluated the newly developed Analysis of Record for Calibration (AORC) forcing data for SWE simulation and examined the impact of bias correction applied to AORC precipitation and temperature. Additionally, we investigated the sensitivity of SWE simulations to choices of physical parameterization schemes through 72 ensemble experiments. Results showed that NWM driven by AORC forcings captured the overall temporal variation of SWE but underestimated its amount. Adjusting AORC precipitation with SNOTEL observations reduced SWE root‐mean‐square error (RMSE) by 66%, adjusting temperature trimmed it by 10%, and adjusting both decreased it by 69%. Among the physical processes, the snow/soil temperature time scheme (STC) demonstrated the highest sensitivity, followed by the surface exchange coefficient for heat (SFC), snow surface albedo (ALB), and rainfall and snowfall partitioning (SNF), while the lower boundary of soil temperature (TBOT) proved to be insensitive. Further optimization of the parameterization combination resulted in a 12% SWE RMSE reduction. When combined with the bias‐corrected AORC precipitation and temperature, this optimization led to a remarkable 78% SWE RMSE reduction. Despite these enhancements, a persistent slow and late spring ablation suggests model deficiencies in snow ablation physics. The study emphasizes the critical need to enhance the accuracy of forcing data in mountainous regions and address model parameterization uncertainty through optimization efforts.

Demonstrating a Hybrid Machine Learning Approach for Snow Characteristic Estimation Throughout the Western United States

AbstractSnow is a critical component of global climate and provides water resources to over 1 billion people worldwide. Yet current measurement methods and modeling techniques lack the ability to fully capture snow characteristics such as snow water equivalent (SWE) and density across variable landscapes. In recent years, physics‐informed machine learning (ML) methods have demonstrated promise for combining data‐driven learning and physical information. However, this capability has not been widely explored within snow hydrology. Here, we develop a “hybrid” model that applies ML informed by outputs from a physical model and assess whether it provides more accurate estimations of SWE and snow density. We trained and evaluated models at 49 SNOw TELemetry locations spanning a range of snow climates in the western US using 9 years of daily data. The research addressed two questions. In the first, the performance of the hybrid model was compared against a plain neural network (long short‐term memory, Long‐Short Term Memory), a high‐quality physical model, and a statistical snow density model. The second question focused on how regionally trained hybrid models compared to a westwide model as well as their transferability between multiple snow regions. The results showed that combining physical information and ML reduced SWE Root Mean Square Error by 35% compared to a physical model and 51% compared to a neural network. Additionally, regional training only provided minimal benefits compared with a westwide model. These findings indicate that a hybrid approach can yield more accurate snowpack characterization than either physical snow models or ML alone.

Enhancing Monthly Streamflow Prediction Using Meteorological Factors and Machine Learning Models in the Upper Colorado River Basin

Streamflow prediction is crucial for planning future developments and safety measures along river basins, especially in the face of changing climate patterns. In this study, we utilized monthly streamflow data from the United States Bureau of Reclamation and meteorological data (snow water equivalent, temperature, and precipitation) from the various weather monitoring stations of the Snow Telemetry Network within the Upper Colorado River Basin to forecast monthly streamflow at Lees Ferry, a specific location along the Colorado River in the basin. Four machine learning models—Random Forest Regression, Long short-term memory, Gated Recurrent Unit, and Seasonal AutoRegresive Integrated Moving Average—were trained using 30 years of monthly data (1991–2020), split into 80% for training (1991–2014) and 20% for testing (2015–2020). Initially, only historical streamflow data were used for predictions, followed by including meteorological factors to assess their impact on streamflow. Subsequently, sequence analysis was conducted to explore various input-output sequence window combinations. We then evaluated the influence of each factor on streamflow by testing all possible combinations to identify the optimal feature combination for prediction. Our results indicate that the Random Forest Regression model consistently outperformed others, especially after integrating all meteorological factors with historical streamflow data. The best performance was achieved with a 24-month look-back period to predict 12 months of streamflow, yielding a Root Mean Square Error of 2.25 and R-squared (R2) of 0.80. Finally, to assess model generalizability, we tested the best model at other locations—Greenwood Springs (Colorado River), Maybell (Yampa River), and Archuleta (San Juan) in the basin.

Using Temporal Deep Learning Models to Estimate Daily Snow Water Equivalent Over the Rocky Mountains

AbstractIn this study we construct and compare three different deep learning (DL) models for estimating daily snow water equivalent (SWE) from high‐resolution gridded meteorological fields over the Rocky Mountain region. To train the DL models, Snow Telemetry (SNOTEL) station‐based SWE observations are used as the prediction target. All DL models produce higher median Nash‐Sutcliffe Efficiency (NSE) values than a conceptual SWE model and interpolated gridded data sets, although mean squared errors also tend to be higher. Sensitivity of the SWE prediction to the model's input variables is analyzed using an explainable artificial intelligence (XAI) method, yielding insight into the physical relationships learned by the models. This method reveals the dominant role precipitation and temperature play in snowpack dynamics. In applying our models to estimate SWE throughout the Rocky Mountains, an extrapolation problem arises since the statistical properties of SWE (e.g., annual maximum) and geographical properties of individual grid points (e.g., elevation) differ from the training data. This problem is solved by normalizing the SWE with its historical maximum value to alleviate extrapolation for all tested DL models. Our work shows that the DL models are promising tools for estimating SWE, and sufficiently capture relevant physical relationships to make them useful for spatial and temporal extrapolation of SWE values.

Attention-Based Models for Snow-Water Equivalent Prediction

Snow Water-Equivalent (SWE)—the amount of water available if snowpack is melted—is a key decision variable used by water management agencies to make irrigation, flood control, power generation, and drought management decisions. SWE values vary spatiotemporally—affected by weather, topography, and other environmental factors. While daily SWE can be measured by Snow Telemetry (SNOTEL) stations with requisite instrumentation, such stations are spatially sparse requiring interpolation techniques to create spatiotemporal complete data. While recent efforts have explored machine learning (ML) for SWE prediction, a number of recent ML advances have yet to be considered. The main contribution of this paper is to explore one such ML advance, attention mechanisms, for SWE prediction. Our hypothesis is that attention has a unique ability to capture and exploit correlations that may exist across locations or the temporal spectrum (or both). We present a generic attention-based modeling framework for SWE prediction and adapt it to capture spatial attention and temporal attention. Our experimental results on 323 SNOTEL stations in the Western U.S. demonstrate that our attention-based models outperform other machine-learning approaches. We also provide key results highlighting the differences between spatial and temporal attention in this context and a roadmap toward deployment for generating spatially-complete SWE maps.

On the Sources of Water Supply Forecast Error in Western Colorado

Abstract Annual spring and summer runoff from western Colorado is relied upon by 40 million people, six states, and two countries. Cool season precipitation and snowpack have historically been robust predictors of seasonal runoff in western Colorado. Forecasts made with this information allow water managers to plan for the season ahead. Antecedent hydrological conditions, such as root zone soil moisture and groundwater storage, and weather conditions following peak snowpack also impact seasonal runoff. The roles of such factors were scrutinized in 2020 and 2021: seasonal runoff was much lower than expectations based on snowpack values alone. We investigate the relative importance of meteorological and hydrological conditions occurring before and after the snowpack season in predicting seasonal runoff in western Colorado. This question is critical because the most effective investment strategy for improving forecasts depends on if errors arise before or after the snowpack season. This study is conducted using observations from the Snow Telemetry Network, root zone soil moisture and groundwater data from the Western Land Data Assimilation Systems, and a random forest–based statistical forecasting framework. We find that on average, antecedent root zone soil moisture and groundwater storage values do not add significant skill to seasonal water supply forecasts in western Colorado. In contrast, using precipitation and temperature data after the time of peak snowpack improves water supply forecasts significantly. The 2020 and 2021 runoffs were hampered by dry conditions both before and after the snowpack season. Both antecedent soil moisture and spring/summer precipitation data improved water supply forecast accuracy in these years. Significance Statement Seasonal water supply forecasts in western Colorado are highly valuable because spring and summer runoff from this region helps support the water supply of 40 million people. Accurate forecasts improve the management of the region’s water. Heavy investments have been made in improving our ability to monitor antecedent hydrological conditions in western Colorado, such as root zone soil moisture and groundwater. However, results from this study indicate that the largest source of uncertainty in western Colorado runoff forecasts is future weather. Therefore, improved subseasonal-to-seasonal weather forecasts for western Colorado are what is most needed to improve regional water supply forecasts, and the ability to properly manage western Colorado water.

Evaluating 3 decades of precipitation in the Upper Colorado River basin from a high-resolution regional climate model

Abstract. Convection-permitting regional climate models (RCMs) have recently become tractable for applications at multi-decadal timescales. These types of models have tremendous utility for water resource studies, but better characterization of precipitation biases is needed, particularly for water-resource-critical mountain regions, where precipitation is highly variable in space, observations are sparse, and the societal water need is great. This study examines 34 years (1987–2020) of RCM precipitation from the Weather Research and Forecasting model (WRF; v3.8.1), using the Climate Forecast System Reanalysis (CFS; CFSv2) initial and lateral boundary conditions and a 1 km × 1 km innermost grid spacing. The RCM is centered over the Upper Colorado River basin, with a focus on the high-elevation, 750 km2 East River watershed (ERW), where a variety of high-impact scientific activities are currently ongoing. Precipitation is compared against point observations (Natural Resources Conservation Service Snow Telemetry or SNOTEL), gridded climate datasets (Newman, Livneh, and PRISM), and Bayesian reconstructions of watershed mean precipitation conditioned on streamflow and high-resolution snow remote-sensing products. We find that the cool-season precipitation percent error between WRF and 23 SNOTEL gauges has a low overall bias (x^ = 0.25 %, s = 13.63 %) and that WRF has a higher percent error during the warm season (x^ = 10.37 %, s = 12.79 %). Warm-season bias manifests as a high number of low-precipitation days, though the low-resolution or SNOTEL gauges limit some of the conclusions that can be drawn. Regional comparisons between WRF precipitation accumulation and three different gridded datasets show differences on the order of ± 20 %, particularly at the highest elevations and in keeping with findings from other studies. We find that WRF agrees slightly better with the Bayesian reconstruction of precipitation in the ERW compared to the gridded precipitation datasets, particularly when changing SNOTEL densities are taken into account. The conclusions are that the RCM reasonably captures orographic precipitation in this region and demonstrates that leveraging additional hydrologic information (streamflow and snow remote-sensing data) improves the ability to characterize biases in RCM precipitation fields. Error characteristics reported in this study are essential for leveraging the RCM model outputs for studies of past and future climates and water resource applications. The methods developed in this study can be applied to other watersheds and model configurations. Hourly 1 km × 1 km precipitation and other meteorological outputs from this dataset are publicly available and suitable for a wide variety of applications.

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.

Snowmelt estimation using an empirical radiation model

This study developed a novel approach for estimating snowmelt using an empirical radiation model, referred to as the Empirical Radiation Snowmelt Model, or ERSM. ERSM assumes that energy available to melt snow is equivalent to the sum of the shortwave radiation absorbed by the snowpack and the net longwave radiation at the snow surface. ERSM has two components. The first component is an empirically determined relationship between the observed daily snowmelt and the daily potential shortwave radiation. The longwave radiation component is implicitly estimated through its influence on the observed snowmelt. The daily potential shortwave radiation was found using two calculated parameters: the top-of-the-atmosphere shortwave radiation and the albedo of the snow surface. This first component simulates the average snowmelt that will occur for a given level of the potential shortwave radiation. The second component of ERSM estimates the difference between the average snowmelt simulated from the potential shortwave radiation (first component) and the actual snowmelt that occurred on any day. Observations from 20 SNOTEL11SNOTEL-U.S. Department of Agriculture, Natural Resources Conservation Service Snow Telemetry site. sites in the Truckee River Watershed were used to develop separate ERSM parameters for each site. ERSM was able to robustly predict snowmelt for all sites considered in this study. The error statistics vary with the magnitude of the observed daily snowmelt and the best results were found for magnitudes at which the majority of snowmelt occurs. A particularly valuable application of ERSM is its use to estimate the melt factor required by temperature index snowmelt models. ERSM can explain a large degree of the observed variance of the temperature index melt factor.

SNOTEL, the Soil Climate Analysis Network, and water supply forecasting at the Natural Resources Conservation Service: Past, present, and future

AbstractThe Snow Survey and Water Supply Forecasting (SSWSF) Program and the Soil Climate Analysis Network (SCAN) of the United States Department of Agriculture's Natural Resources Conservation Service (NRCS) generate key observational and predictive information for water managers. Examples include mountain climate and snow monitoring through manual snow surveys and the SNOw TELemetry (SNOTEL) and SNOtel LITE networks, in situ soil moisture data acquisition through the SCAN and SNOTEL networks, and water supply forecasting using river runoff prediction models. The SSWSF Program has advanced continuously over the decades and is a major source of valuable water management information across the western United States, and the SCAN network supports agricultural and other water users nationwide. Product users and their management goals are diverse, and use‐cases range from guiding crop selection to seasonal flood risk assessment, drought monitoring and prediction, avalanche and fire prediction, hydropower optimization, tracking climate variability and change, environmental management, satisfying international treaty and domestic legal requirements, and more. Priorities going forward are to continue innovating to enhance the accuracy and completeness of the observational and model‐generated data products these programs deliver, including expanded synergies with the remote sensing community and uptake of artificial intelligence while maintaining long‐term operational reliability and consistency at scale.

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.

Spring heat waves drive record western United States snow melt in 2021

Throughout the western US snow melted at an alarming rate in April 2021 and by May 1, hydrologic conditions were severely degraded with declining summer water supply forecasts compared to earlier in the winter. The objectives of this study are to (a) quantify the magnitude and climatological context of observed melt rates of snow water equivalent (SWE) and (b) underpin the hydrometeorological drivers during April 2021 based on atmospheric reanalysis and gridded meteorological data. Peak SWE indicated snow drought conditions were widespread (41% of stations between 5th and 20th percentile) but not necessarily extreme (only 9% of stations less than 5th percentile). Here, using observations from the Snow Telemetry (SNOTEL) network we found record 7 day snow melt rates (median of −99 mm; ±one standard deviation of 61 mm) occurred at 24% of SNOTEL sites and in all 11 Western states. Strong upper atmospheric ridging that began initially in the north Pacific with eastward propagation by mid-April to the Pacific Northwest Coast led to near-surface conditions across the western US conducive to rapid snow loss. One heat wave occurred inland across the Rockies the first week of April and then later in April, a second heat wave impacted the Cascades and northern California. We find that ripening of the snowpack by both record high surface solar radiation and air temperatures were factors in driving the rapid snow melt. Equatorial Pacific sea surface temperatures and the La Niña pattern that peaked in winter along with an eastward propagating and intensifying Madden–Julian Oscillation were likely responsible for driving the placement, strength, and progression of the north Pacific Ridge. This study documents the role of two extreme spring ‘sunny heat wave’ events on snowpack, and the cascading drought impacts which are anticipated to become more frequent in a warming world.

Estimating snow cover from high-resolution satellite imagery by thresholding blue wavelengths

The extent and duration of snow cover, a critical component of the hydrologic cycle and the global climate system, is expected to shift dramatically under climate change. Therefore, developing high-resolution assessments of snow cover change is crucial for estimating the impact of changing snow cover on watershed and ecosystems processes in cold regions. Remote sensing tools provide a powerful method for mapping snow-covered area (SCA) across a landscape. The most common method for estimating SCA utilizes the normalized difference snow index (NDSI), which relies on spectral measurements in the shortwave-infrared wavelengths (SWIR). NDSI can effectively estimate catchment- to regional-scale SCA, but it cannot be used to assess fine-scale SCA because of current limitations on the spatial resolution of satellite-derived SWIR measurements. Here, we map SCA using a threshold of blue wavelengths and high-resolution satellite imagery. The thresholding method, which we call the Blue Snow Threshold algorithm (BST), has previously been used with digital camera imagery. We refine and automate the algorithm for use with cloud-free high-resolution satellite imagery and find that the BST can be used to assess fine-scale SCA. For validation, we compared BST-derived estimates of SCA to a) airborne lidar surveys, b) Landsat fractional SCA, and c) snow disappearance dates from Snow Telemetry (SNOTEL) stations. When compared to airborne lidar surveys of SCA, the BST predicted SCA had a range of F-scores between 0.81 and 0.94 in four study areas in California and Colorado. We also found general agreement between SCA and snow disappearance at multiple SNOTEL sites across the western United States. Given the relatively recent availability of high-resolution satellite imagery with spectral measurements in the visible wavelengths but lacking in SWIR, the BST offers a reliable and easy-to-apply tool for examining fine-scale snow-related processes.

Wildfire Impacts on Snowpack Phenology in a Changing Climate Within the Western U.S.

AbstractSnowpack in the western U.S. is critical for water supply and is threatened by wildfires, which are becoming larger and more common. Numerous studies have examined impacts of wildfire on snow water equivalent (SWE), but many of these studies are limited in the number of observation locations, and they have sometimes produced conflicting results. The objective of this study is to distinguish the net effects of wildfires on snowpack from those of climate. We consider 45 burned sites from the Snow Telemetry network. For each burned site, unburned control sites are identified from the same level III ecoregion. Impacts of climate changes on snowpack are analyzed by comparing pre‐fire and post‐fire snow water equivalent at the unburned sites. Combined climate and wildfire effects are considered by comparing pre‐fire and post‐fire SWE at the burned sites. Wildfire impacts are then isolated by taking the difference between the burned and unburned sites. The wildfire‐induced changes in SWE are also compared to several possible controlling variables including burn severity, leaf‐area index change, dominant pre‐fire tree genus, years since the fire, and site elevation. On average, wildfires have advanced melt‐out dates 9 days and maximum SWE dates 6 days and reduced annual maximum SWE by 10% across the sites considered. On average, the combined effects of climate and wildfire have advanced melt‐out and maximum SWE dates approximately 14 and 10 days, respectively, while decreasing annual maximum SWE by approximately 10%.

GNSS-IR Snow Depth Retrieval Based on the Fusion of Multi-Satellite SNR Data by the BP Neural Network

Compared with previous snow depth monitoring methods, global navigation satellite system-interferometric reflectometry (GNSS-IR) technology has the advantage of obtaining continuous daily observation data, and has great application potential. However, since GNSS satellites are in motion, their position in the sky is constantly varying, and the Fresnel reflection areas about the receiver in different periods alter accordingly. As a result, the retrieving results obtained from different GNSS satellites, and data sets collected in different periods, fluctuate considerably, making the traditional single-satellite-based GNSS-IR retrieving method have limitations in accuracy and reliability. Therefore, this paper proposed a novel GNSS-IR signal-to-noise ratio (SNR) retrieving snow depth method for fusing the available GNSS-IR observations to obtain an accurate and reliable result. We established the retrieval model based on the backpropagation algorithm, which makes full use of the back propagation (BP) neural network’s self-learning and self-adaptive capability to exploit the degree of contribution of different satellites to the final results. Then, the SNR observations of the global positioning system (GPS) L1 carrier from the Plate Boundary Observation (PBO) site P351 were collected to experiment for validation purposes. For all available GPS L1 carrier data, the snow depth values retrieved for each satellite were first obtained by the existing single-satellite-based GNSS-IR retrieval method. Then, four groups of comparison results were acquired, based on the multiple linear regression model, random forest model, mean fusion model, and the proposed BP neural network model, respectively. Taking the snow depth in-situ data provided by snow telemetry (SNOTEL) as a reference, the root mean squared error (RMSE) and mean absolute error (MAE) of the proposed solution are 0.0297 m and 0.0219 m, respectively. Furthermore, the retrieving results are highly consistent with the measured data, and the correlation coefficient is 0.9407.

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

AbstractThis 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.

Forest Canopy Density Effects on Snowpack Across the Climate Gradients of the Western United States Mountain Ranges

AbstractControlled field experiments to disentangle the effect of canopy density from the effect of climate on snowpack dynamics are limited by the underlying linkage between canopy density and climate. Thus, based on observations alone, it is not well understood how variations in canopy density can affect snow processes under different climate regimes. To address this knowledge gap, this study uses a physics‐based modeling approach to evaluate the sensitivity of snowpack dynamics to variations in canopy density across the climate gradients of the Western U.S. as represented by 228 Snow Telemetry (SNOTEL) sites. Within the model, we uniformly parameterize the canopy across sites to represent an idealized forest with high, medium, and low canopy density, respectively. The results illustrate that the effect of canopy density on the peak snow water equivalent (SWE) and duration of under‐canopy snowpack is sensitive to winter climate (i.e., climatological winter precipitation and temperature). As canopy density decreases, the greatest increase in peak SWE and snowpack duration is found in wet/warm and dry/cold climates, where snowpack under low‐density forest lasts longer than that in the open. In comparison, peak SWE and snowpack duration in wet/cold climates are less sensitive to changing canopy density. Thus, forest management actions (e.g., thinning and clearing) are likely to have disparate impacts on snow depending on local winter climate. Climate sensitivity of under‐canopy snowpack suggests that snowpack duration under dense canopy in presently warm winter climates is expected to experience the greatest reduction under a warming climate.

Snowpack Distribution Using Topographical, Climatological and Winter Season Index Inputs

A majority of the annual precipitation in many mountains falls as snow, and obtaining accurate estimates of the amount of water stored within the snowpack is important for water supply forecasting. Mountain topography can produce complex patterns of snow distribution, accumulation, and ablation, yet the interaction of topography and meteorological patterns tends to generate similar inter-annual snow depth distribution patterns. Here, we question whether snow depth patterns at or near peak accumulation are repeatable for a 10-year time frame and whether years with limited snow depth measurement can still be used to accurately represent snow depth and mean snow depth. We used snow depth measurements from the West Glacier Lake watershed, Wyoming, USA, to investigate the distribution of snow depth. West Glacier Lake is a small (0.61 km2) windswept (mean of 8 m/s) watershed that ranges between 3277 m and 3493 m. Three interpolation methods were compared: (1) a binary regression tree, (2) multiple linear regression, and (3) generalized additive models. Generalized additive models using topographic parameters with measured snow depth presented the best estimates of the snow depth distribution and the basin mean amounts. The snow depth patterns near peak accumulation were found to be consistent inter-annually with an average annual correlation coefficient (r2) of 0.83, and scalable based on a winter season accumulation index (r2 = 0.75) based on the correlation between mean snow depth measurements to Brooklyn Lake snow telemetry (SNOTEL) snow depth data.

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.

Multi-layer high-resolution soil moisture estimation using machine learning over the United States

The lack of proper understanding of multi-layer soil moisture (SM) profile (signals) remains a persistent challenge in sustainable agricultural water management and food security, especially during drought conditions. We develop a machine-learning algorithm using the concept of learning from patterns to estimate the multi-layer SM information in ungauged locations firmly based on local knowledge of the climatic and landscape controls. The Contiguous United States (CONUS) is clustered into homogeneous regions based on the association between SM and climate and landscape controls. Extreme Gradient Boosting (XGBoost) algorithm is applied to homogenous regions to capture the complex relationship between appropriate predictor variables and in-situ SM at multiple layers over the CONUS. Soil Moisture Active Passive (SMAP) Level 4 (L4) surface (0–5 cm) and rootzone (0–100 cm) SM along with climate and landscape datasets are used as predictor variables. In-situ multi-layer SM recorded by Soil Climate Analysis Network (SCAN), Snow Telemetry (SNOTEL), and U.S. Climate Reference Network (USCRN) networks are utilized as predictands. XGBoost models are then trained region-wise and layer-wise to estimate multi-layer SM information at 5, 10, 20, 50, and 100 cm depths (five layers) at 1-km spatial resolution. Results indicate that the predictor variables have varying levels of influence on SM with changing soil depth, and meteorological variables have the least importance. Validation at 79 independent locations indicates the multi-layer SM estimates successfully capture temporal dynamics of SM, with most locations achieving ubRMSE less than 0.04 m3/m3. The high-resolution SM estimates offer spatial sub-grid heterogeneity compared to SMAP L4 SM.