Dr. Giulia Mazzotti

The water equivalent of new snow (HNW) plays a crucial role in various fields, including hydrological modeling, avalanche forecasting, and assessing snow loads on structures. However, in contrast to snow depth (HS), obtaining HNW measurements is challenging as well as time-consuming and is hence rarely measured. Therefore, we assess the reliability of two semi-empirical methods, HS2SWE and ΔSNOW, for estimating HNW. These methods are designed to simulate continuous water equivalent of the snowpack (SWE) from daily HS only, with changes in SWE yielding daily HNW estimates. We compare both parametric methods against HNW predictions from a physics-based snow model (FSM2oshd) that integrates daily HS recordings using data assimilation. Our findings reveal that all methods exhibit similar performance, with relative biases of less than ∼3 % in replicating SWE observations commonly used for model evaluations. However, the ΔSNOW model tends to underestimate daily HNW by ∼17 %, whereas HS2SWE and FSM2oshd combined with a particle filter data assimilation scheme provide nearly unbiased estimates, with relative biases below ∼5 %. In contrast to the parsimonious parametric methods, we show that the physics-based approach can yield information about unobserved variables, such as total solid precipitation amounts, that may differ from HNW due to concurrent melt. Overall, our results underscore the potential of utilizing commonly available daily HS data in conjunction with appropriate modeling techniques to provide valuable insights into snow accumulation processes. Our study demonstrates that daily SWE observations or supplementary measurements like HNW are important for validating the day-to-day accuracy of simulations and should ideally already be incorporated during the calibration and development of models.

We present an hourly hydrometeorological and snow dataset with 100 m spatial resolution from the alpine Dischma watershed and its surroundings in eastern Switzerland, including station measurements of variables such as snow depth and catchment runoff. This dataset is particularly suited for different modelling experiments using distributed and process-based models, including physics-based snow and hydrological models. Additionally, the data are highly useful for testing various snow data assimilation schemes and for developing models representing snow–forest interactions. The dataset covers 7 water years from 1 October 2016 to 30 September 2023. The complete domain spans an area of 333 km² with altitudes ranging from 1250 to 3228 m. The Dischma Basin, with its outlet at 1671 m elevation, occupies 42.9 km². Included in the dataset are high-resolution (100 m) hourly meteorological data (air temperature, relative humidity, wind speed and direction, precipitation, and long- and shortwave radiation) from a numerical weather predication model and rain radar, land cover characteristics (primarily forest properties), and a digital elevation model. Notably, the dataset includes snow depth acquisitions obtained from airborne lidar and photogrammetry surveys, constituting the most extensive spatial snow depth dataset derived using such techniques in the European Alps. Along with these gridded datasets, we provide daily quality-controlled snow depth recordings from seven sites, biweekly snow water equivalent measurements from two locations, and hourly runoff and stream temperature observations for the Dischma watershed. The data compiled in this study will be useful to further develop our ability to forecast snow and hydrological conditions in high-alpine headwater catchments that are particularly sensitive to ongoing climate change. All data are available for download at https://doi.org/10.16904/envidat.568 (Magnusson et al., 2024).

Distributed energy and mass balance snowpack models at sub-kilometric scale have emerged as a tool for snow-hydrological forecasting over large areas. However, their development and evaluation often rely on a handful of well-observed sites on flat terrain with limited topographic representativeness. Validation of such models over large scales in rugged terrain is therefore necessary. Remote sensing of wet snow has always been motivated by its potential utility in snow hydrology. However, its concrete potential to enhance physically based operational snowpack models in real time remains unproven. Wet-snow maps could potentially help refine the temporal accuracy of simulated snowmelt onset, while the information content of remotely sensed snow cover fraction (SCF) pertains predominantly to the ablation season. In this work, wet-snow maps derived from Sentinel-1 and SCF retrieval from Sentinel-2 are compared against model results from a fully distributed energy balance snow model (FSM2oshd). The comparative analysis spans the winter seasons from 2017 to 2021, focusing on the geographic region of Switzerland. We use the concept of wet-snow line (WSL) to compare Sentinel-1 wet-snow maps with simulations. We show that while the match of the model with flat-field snow depth observation is excellent, the WSL reveals a delayed snowmelt in the southern aspects. Amending the albedo parametrization within FSM2oshd allowed for the achievement of earlier melt in such aspects preferentially, thereby reducing WSL biases. Biases with respect to Sentinel-2 snow-line (SL) observations were also substantially reduced. These results suggest that wet-snow maps contain valuable real-time information for snowpack models, complementing flat-field snow depth observations well, particularly in complex terrain and at higher elevations. The persisting correlation between wet-snow-line and snow-line biases provides insights into refined development, tuning, and data assimilation methodologies for operational snow-hydrological modelling.

Land surface processes, crucial for exchanging carbon, nitrogen, water, and energy between the atmosphere and terrestrial Earth, significantly impact the climate system. Many of these processes vary considerably at small spatial and temporal scales, in particular in mountainous terrain and complex topography. To examine the impact of spatial resolution and representativeness of input data on modelled land surface processes, we conducted simulations using the Community Land Model 5 (CLM5) at different resolutions and based on a range of input datasets over the spatial extent of Switzerland. Using high-resolution meteorological forcing and land use data, we found that increased resolution substantially improved the representation of snow cover in CLM5 (up to 52 % enhancement), allowing CLM5 to closely match performance of a dedicated snow model. However, a simple lapse-rate-based temperature downscaling provided large positive effects on model performance, even if simulations were based on coarse-resolution forcing datasets only. Results demonstrate the need for resolutions higher than 0.25° for accurate snow simulations in topographically complex terrain. These findings have profound implications for climate impact studies. As improvements were observed across the cascade of dependencies in the land surface model, high spatial resolution and high-quality forcing data become necessary for accurately capturing the effects of a declining snow cover and consequent shifts in the vegetation period, particularly in mountainous regions. This study further highlights the utility of multi-resolution modelling experiments when aiming to improve representation of variables in land surface models. By embracing high-resolution modelling, we can enhance our understanding of the land surface and its response to climate change.

Snow hydrological regimes in mountainous catchments are strongly influenced by snowpack heterogeneity resulting from wind- and gravity-induced redistribution processes, requiring them to be modelled at hectometre and finer resolutions. This study presents a novel modelling approach to address this issue, aiming at an intermediate-complexity solution to best represent these processes while maintaining operationally viable computational times. To this end, the physics-based snowpack model FSM2oshd was complemented by integrating the modules SnowTran-3D and SnowSlide to represent wind- and gravity-driven redistribution, respectively. This new modelling framework was further enhanced by implementing a density-dependent layering to account for erodible snow without the need to resolve microstructural properties. Seasonal simulations were performed over a 1180 km² mountain range in the Swiss Alps at 25, 50 and 100 m resolution, using appropriate downscaling and snow data assimilation techniques to provide accurate meteorological forcing. In particular, wind fields were dynamically downscaled using WindNinja to better reflect topographically induced flow patterns. The model results were assessed using snow depths from airborne lidar measurements. We found a remarkable improvement in the representation of snow accumulation and erosion areas, with major contributions from saltation and suspension as well as avalanches and with modest contributions from snowdrift sublimation. The aggregated snow depth distribution curve, key to snowmelt dynamics, significantly and consistently matched the measured distribution better than reference simulations from the peak of winter to the end of the melt season, with improvements at all spatial resolutions. This outcome is promising for a better representation of snow hydrological processes within an operational framework.

In mountain regions, forests that overlap with seasonal snow mostly reside in complex terrain. Due to persisting major observational challenges in these environments, the combined impact of forest structure and topography on seasonal snow cover dynamics is still poorly understood. Recent advances in forest snow process representation and increasing availability of detailed canopy structure datasets, however, now allow for hyper-resolution (<5m) snow model simulations capable of resolving tree-scale processes. These can shed light on the complex process interactions that govern forest snow dynamics. We present multi-year simulations at 2m resolution obtained with FSM2, a mass- and energy-balance-based forest snow model specifically developed and validated for metre-scale applications. We simulate an ∼3km2 model domain encompassing forested slopes of a sub-alpine valley in the eastern Swiss Alps and six snow seasons. Simulations thus span a wide range of canopy structures, terrain characteristics, and meteorological conditions. We analyse spatial and temporal variations in forest snow energy balance partitioning, aiming to quantify and understand the contribution of individual energy exchange processes at different locations and times. Our results suggest that snow cover evolution is equally affected by canopy structure, terrain characteristics, and meteorological conditions. We show that the interaction of these three factors can lead to snow accumulation and ablation patterns that vary between years. We further identify higher snow distribution variability and complexity in slopes that receive solar radiation early in winter. Our process-level insights corroborate and complement existing empirical findings that are largely based on snow distribution datasets only. Hyper-resolution simulations as presented here thus help to better understand how snowpacks and ecohydrological regimes in sub-alpine regions may evolve due to forest disturbances and a warming climate. They could further support the development of process-based sub-grid forest snow cover parameterizations or tiling approaches for coarse-resolution modelling applications.

The seasonal evolution of snow cover has significant impacts on the hydrological cycle and microclimate in mountainous regions. However, snow processes also play a crucial role in triggering alpine mass movements and flooding, posing risks to people and infrastructure. To mitigate these risks, many countries use operational forecast systems for snow distribution and melt. This paper presents the Swiss Operational Snow-hydrological (OSHD) model system, developed to provide daily analysis and forecasts on snow cover dynamics throughout Switzerland. The OSHD system is a sophisticated snow hydrological model designed specifically for the high-alpine terrain of the Swiss Alps. It leverages exceptional station data and high-resolution meteorological forcing data, as well as various reanalysis products to combine snow modeling with advanced data assimilation and meteorological downscaling methods. The system offers models of varying complexity, each tailored to specific modeling strategies and applications. For snowmelt runoff forecasting, monitoring snow water resources, and research-grade purposes, the OSHD system employs physics-based modeling chains. For snow climatological assessments, a conceptual model chain is available. We are pleased to present two comprehensive datasets from the conceptual and physics-based models that cover the entirety of Switzerland. The first dataset comprises a snow water equivalent climatology spanning 1998-2022, with a spatial resolution of 1 km. The second dataset includes snow distribution and snow melt data spanning 2016-2022 at a high spatial resolution of 250 m. To meet the needs of a multi-purpose snow hydrological model framework, the OSHD system employs various strategies for process representation and sub-grid parameterizations at the snow-canopy-atmosphere interface, particularly in complex terrain. Recent and ongoing model developments are aimed at accounting for complex forest snow processes, representing slope and ridge-scale precipitation and snow redistribution processes, as well as improving probabilistic snow forecasts and data assimilation procedures based on remote sensing products.

This study presents a new model for calculating canopy shortwave radiation transmissivity from synthetic hemispheric images using only information contained within a canopy height model (CHM) - CanopyHeightModel2Radiation (C2R). The enhanced version calculates synthetic hemispherical images based on the geometric arrangement of the surrounding canopy while applying a statistical correction for canopy transmissivity using canopy thickness and tree species leaf area. The simple input data and statistical correction make this model suitable for estimating canopy transmissivity across large spatial extents typical of land surface models for which canopy transmissivity or radiation is a primary input variable. Performance of C2R-enhanced is assessed against hemispherical photographs, and compared to a basic version of C2R without transmissivity correction, and two versions of a Lidar2Radiation model (L2R-enhanced, L2R-basic) with either a basic representation of canopy structure or an enhanced representation including trunks and branches within tree crowns. The two enhanced models (L2R-enhanced and C2R-enhanced) perform best compared to hemispherical photographs, while the L2R-basic and C2R-basic models over- and underestimate canopy transmissivity, respectively. At 1-meter and 10-minute resolution, the two enhanced models perform similarly, but exact timing and location of transmissivity controlled by canopy structure is better represented in the physically explicit L2R-enhanced model. Across hourly and 25 × 25 m grid-averaged scales, both enhanced models achieve similar estimates of canopy transmissivity. Based on these results, it is recommended that the purely physically-based representation in the L2R-enhanced model is used when estimates of canopy transmissivity at high spatial and temporal (meter and minute) resolutions are necessary, while the computationally more efficient C2R-enhanced model is used when calculating canopy transmissivity within spatially aggregated grid cells, for example, as input into coarser-resolution land surface models. Incorporating C2R-enhanced into existing forest energy balance models creates exciting opportunities for investigating forest structure changes on forest hydrology and ecosystems across previously impossible spatial extents.

Mapping snow in forests is important for understanding the snow cover dynamics in these environments in view of hydrological applications and water resources management. Today, Uncrewed Aerial Vehicles (UAVs) are widely used for snow studies due to their rather cheap and flexible operation. UAV-borne Light detection and ranging (LiDAR) systems are a promising technology for sub-canopy snow mapping at high temporal resolutions, concurrently providing information on both the canopy and the below-canopy snow surface. In this pilot study, we used a UAV-LiDAR system to investigate the snow cover dynamics within two steep forested slopes of opposing aspects in the Swiss Alps at both high spatial resolution and unprecedented temporal resolution. Using a Distance to Canopy Edge (DCE) algorithm to characterize local forest structure, snow depth was analyzed in terms of relative position within variable forest cover. The north-exposed site had higher mean snow depths throughout the season compared to the south-exposed site, especially in canopy gaps. Whereas snow depletion rate was consistent throughout the north-exposed site, snow depletion was much faster in the gaps at the south-exposed slope. Correlation coefficients between snow depths and local canopy closure were weaker at the south-exposed (between -0.5 and - 0.7) than at the north-exposed site (between -0.7 and - 0.9), and rapidly deteriorated right after the peak of winter at the south-exposed slope. This indicates shortwave radiation dominates snowmelt processes at this site, which was thought to be spatially uncorrelated to local canopy cover, unlike accumulation and melt processes on the north-exposed slope that generated snow patterns with a high spatial correlation to local canopy cover throughout the entire season. Calculations of incoming sub-canopy shortwave radiation (SWR) for both sites confirmed this assumption. While our findings encourage the use of UAV-borne LiDAR for further investigations of snow cover dynamics in steep forested slopes, we also outline and discuss technical challenges specific to this application. Our insights allow deriving useful recommendations for future studies using UAV-borne LiDAR over a similar environment.

Snow depth mapping in Alpine forests is of high importance for hydrogeology, ecology, tourism, and natural hazards prevention. Different remote sensing approaches have been employed for the precise mapping of snow depth within forests. However, optical sensors cannot provide below-canopy information. While Airborne Laser Scanning (ALS) systems have been used successfully in this context and allow obtaining data below canopies, the costs of acquisitions are very high, not allowing frequent data acquisitions. UAV-based Lidar technology potentially can provide the critical below-canopy information at lower cost and allows for frequent acquisitions. First attempts to employ a UAV-based Lidar system in forests have proven promising, but they are limited to flat forests and to grid-level snow depth calculations. In this study, we present UAV-based Lidar data of both flat and steep forests. Different Lidar processing workflows are analyzed and compared, and snow depth algorithms are used both at the point and the grid level. Whereas the UAV-Lidar system proved capable of mapping snow in both environments, the steep forests' data processing comes with greater challenges, especially for the 3D registration, ground classification, and point-to-point snow depth calculations.

Land Surface Albedo (LSA) of forested environments continues to be a source of uncertainty in land surface modeling, especially across seasonally snow covered domains. Assessment and improvement of global scale model performance has been hampered by the contrasting spatial scales of model resolution and in‐situ LSA measurements. In this study, point‐scale simulations of the Community Land Model 5.0 (CLM5) were evaluated across a large range of forest structures and solar angles at two climatically different locations. LSA measurements, using an uncrewed aerial vehicle with up and down‐looking shortwave radiation sensors, showed canopy structural shading of the snow surface exerted a primary control on LSA. Diurnal patterns of measured LSA revealed strong effects of both azimuth and zenith angles, neither of which were adequately represented in simulations. In sparse forest environments, LSA were overestimated by up to 66%. Further analysis revealed a lack of correlation between Plant Area Index (PAI), the primary canopy descriptor in CLM5, and measured LSA. Instead, measured LSA showed considerable correlation with the fraction of snow visible in the sensor's field of view, a correlation which increased further when only considering the sunlit fraction of visible snow. The use of effective PAI values as a simple first‐order correction for the discrepancy between measured and simulated LSA in sparse forest environments substantially improved model results (64%-76% RMSE reduction). However, the large biases suggest the need for a more generic solution, for example, by introducing a canopy metric that represents canopy gap fraction rather than assuming a spatially homogeneous canopy.

Processes shaping forest snow cover evolution often vary at small spatial scales which are not resolved by most model applications. Representing this variability at larger scales and coarser model resolutions constitutes a major challenge for model developers. In this study, we use a well‐validated hyper‐resolution forest snow model that explicitly resolves the spatial variability of canopy‐snow interactions at the meter scale to explore adequate representation of forest‐snow processes at coarser resolutions (50m). For this purpose, we assess coarser‐resolution runs against spatially averaged results from corresponding hyper‐resolution simulations over a 150 000 m² model domain. For the coarser‐resolution simulations, we tested alternative upscaling strategies. Our results reveal considerable discrepancies between strategies that utilize generalized canopy metrics versus strategies that apply a more detailed set of process‐specific canopy descriptors. Particularly, the inclusion of canopy descriptors that represent the various scales and perspectives relevant to the individual processes leads to accurate simulation of forest snow cover dynamics at coarse resolutions. Our results further demonstrate that a realistic representation of snow‐covered fraction in snowmelt calculations is important even for relatively small (∼50 m) grid cells. Ultimately, this work provides recommendations for modelling forest‐snow processes in large‐scale applications, which allow coarse‐resolution simulations to approximate spatially averaged results of corresponding hyper‐resolution simulations.

Forested environments feature a highly complex radiation regime. The three-dimensional canopy structure masks solar radiation as it penetrates through the forest. This entails sub-canopy radiation to vary in space, over time, season, and with weather conditions, which challenges both measurements and modelling. Here we combine the experimental acquisition of forest structure information with radiation transfer modelling to achieve efficient and accurate estimation of sub-canopy shortwave radiation at very high temporal resolution. For this purpose we have developed a Matlab-based software tool to analyse digital hemispherical imagery for directional canopy transmissivity which yields, in combination with measured or estimated above-canopy shortwave radiation, respective below-canopy radiation. We demonstrate the utility of the tool to accurately model shortwave radiation in comparison to pyranometer measurements at both the point scale and along a transect under heterogeneous canopy cover. Optimized for computational efficiency the software even enables fully-distributed simulations. Without necessitating high performance computing or complex ray-tracing, this tool enables easy access to detailed sub-canopy radiation information. These should be valuable for validating site-scale radiation transfer models, as input to eco-hydrological and snowmelt models, and for researching light-dependant bio-physiological processes.

The complex dynamics of snow accumulation and melt processes under forest canopies entail major observational and modeling challenges, as they vary strongly in space and time. In this study, we present novel data sets acquired with mobile multisensor platforms in subalpine and boreal forest stands. These data sets include spatially and temporally resolved measurements of shortwave and longwave irradiance, air and snow surface temperatures, wind speed, and snow depth, all coregistered to canopy structure information. We then apply the energy balance snow model FSM2 to obtain concurrent, distributed simulations of the forest snowpack at very high ("hyper") resolution (2 m). Our data sets allow us to assess the performance of alternative canopy representation strategies within FSM2 at the level of individual snow energy balance components and in a spatially explicit manner. We demonstrate the benefit of accounting for detailed spatial patterns of shortwave and longwave radiation transfer through the canopy and show the importance of describing wind attenuation by the canopy using stand‐scale metrics. With the proposed canopy representation, snowmelt dynamics in discontinuous forest stands were successfully reproduced. Hyper‐resolution simulations resolving these effects provide an optimal basis for assessing the snow‐hydrological impacts of forest disturbances and for validating and improving the representation of forest snow processes in land surface models intended for coarser‐scale applications.

Modelling spatiotemporal dynamics of snow in forests is challenging, as involved processes are strongly dependent on small‐scale canopy properties. In this study, we explore how local canopy structure information can be integrated in a medium‐complexity energy‐balance snow model to replicate observed snow patterns at very high spatial resolutions. Snow depth distributions simulated with the Flexible Snow Model (FSM2) were tested against extensive experimental data acquired in discontinuous subalpine forest stands in Eastern Switzerland over three winters. While the default canopy implementation in FSM2 fails to capture the observed snow depth variability, performance is considerably improved when local canopy cover fraction and hemispherical sky view fraction are additionally accounted for (30% reduction in RMSE). However, realistic snow depth distribution patterns throughout the season are only achieved if effective temperatures of near and distant canopy elements are discerned, and if a mechanism to mimic preferential deposition of snow in canopy gaps is included. We demonstrate that by diversifying the canopy structure input in order to reflect respective portions of the canopy relevant to different processes, even a simple model based on widely used process parameterizations and canopy metrics can be applied for high‐resolution simulations of the sub‐canopy snow cover with just a few modifications. The presented approaches could be implemented in commonly used land surface models, allowing upscaling experiments and development of sub‐grid parameterizations without necessitating complex high‐resolution models.

Forest canopies act as intermediaries in radiation energy exchange between the atmosphere and the snow surface. The size, location and distribution of forest discontinuities are important controls on forest shortwave radiation transmission and subsequent snow surface shading and radiation energy exchange between the atmosphere and the canopy, but challenges arise when accounting for these vegetation characteristics at large spatial scales. Airborne LiDAR datasets contain detailed information about canopy structure across large spatial scales which can be exploited within 2D transmission models. However, airborne LiDAR data typically does not resolve lower canopy elements, leading to unrealistic depictions of individual trees. We present a methodology to enhance airborne LiDAR data by calculating additional trunk and branch points based on segmentation of a canopy height model, allowing more accurate estimates of canopy shortwave transmissivity. To demonstrate this, we deployed a computationally efficient 2D radiation transfer modelling framework that calculates direct and diffuse radiation from a set of distributed synthetic hemispherical images. The model can predict incoming direct and diffuse solar radiation at the snow surface at high spatial (meter-scale) and temporal (minute-scale) resolutions. Comparison between synthetic and real hemispherical photographs showed that synthetic images, if based on enhanced LiDAR data, featured canopy and individual tree crowns that were much denser than the original LiDAR portrays, improving the representation of vegetation structure especially within dense environments and along canopy edges. Corresponding modelled total shortwave radiation matched well with spatially gridded measurements from a moving pyranometer at two sites, where model RMSE was reduced to 59 and 29 W m^-2 from 181 and 138 W m^-2, respectively, compared to the same transmission model with the original LiDAR data. Maps of snow surface shading patterns corresponded well to those seen in aerial photographs, showing the enhanced LiDAR data can be used to solve complex spatiotemporal patterns of sub-canopy incoming radiation. This work demonstrates that canopy structure information from the lower canopy is an important aspect for accurate radiation transfer modelling, and methods presented here can successfully mitigate problems inherent in many airborne LiDAR datasets to improve spatially distributed estimates of sub-canopy shortwave radiation.

NASA’s 2017 SnowEx field campaign at Grand Mesa, CO, generated Airborne Laser Scans (ALS), Terrestrial Laser Scans (TLS), and snow-probe transects, which allowed for a comparison between snow depth measurement techniques. At six locations, comparisons between gridded ALS and TLS observations, at 1-m resolution, had a median snow depth difference of 5 cm, root-mean-square-difference of 16 cm, mean-absolute-difference of 10 cm, and 3-cm difference in standard deviation. ALS generally had greater, but similar snow depth values to TLS, and results were not sensitive to the gridded cell size between 0.5 and 5 m. The greatest disagreements were where snow-off TLS scans had shrubs and high incidence angles, leading to deeper snow depths (>10 cm) from ALS than TLS. The low vegetation and oblique angles caused occlusion in the TLS data and thus produced higher snow-off bare earth models relative to the ALS. Furthermore, in sub-canopy areas where both ALS and TLS data existed, snow depth differences were comparable to differences in the open. Meanwhile, median values from 52 snow-probe transects and near-coincident ALS data had a mean difference of 6 cm, root-mean-square-difference of 8 cm, mean-absolute-difference of 7 cm, and a mean difference in the standard deviation of 1 cm. Snow depth probes had greater, but similar snow depth values to ALS. Therefore, based on comparisons with TLS and snow depth probes, ALS captured snow depth magnitude with better than or equal agreement to what has been reported in previous studies, and showed the ability to capture high-resolution spatial variability.

Climate models show that global warming will disproportionately influence high-latitude regions and indicate drastic changes in, amongst others, seasonal snow cover. However, current continental and global simulations covering these regions are often run at coarse grid resolutions, potentially introducing large errors in computed fluxes and states. To quantify some of these errors, we have assessed the sensitivity of an energy-balance snow model to changes in grid resolution using a multi-parametrization framework for the spatial domain of mainland Norway. The framework has allowed us to systematically test how different parametrizations, describing a set of processes, influence the discrepancy, here termed the scale-error, between the coarser (5 to 50 km) and finest (1 km) resolution. The simulations were setup such that liquid and solid precipitation was identical between the different resolutions, and differences between the simulations arise mainly during the ablation period. The analysis presented in this study focuses on evaluating the scale-error for several variables relevant for hydrological and land surface modelling, such as SWE and turbulent heat exchanges. The analysis reveals that the choice of method for routing liquid water through the snowpack influences the scale-error most for SWE, followed by the type of parametrizations used for computing turbulent heat fluxes and albedo. For turbulent heat exchanges, the scale-error is mainly influenced by model assumptions related to atmospheric stability. Finally, regions with strong meteorological and topographic variability show larger scale-errors than more homogenous regions.

Radiative processes are substantially altered by the presence of forest canopies, further affecting snow energetics during wintertime. In situ measurements of subcanopy radiation can help improve process‐scale understanding of these complex interactions, which are needed to further constrain and improve land surface models. In this study, a custom‐made cable car was used to measure incoming and outgoing, shortwave and longwave radiation below an evergreen forest stand. Hemispherical photographs taken concurrently from the cable car measured view fractions of shaded snow, sunlit snow, and bare ground. With this setup it was possible to quantify diurnal and seasonal radiation patterns together with their potential drivers at high spatiotemporal resolution. Measurements were performed between January and May 2018, along a 48‐m transect in a discontinuous needleleaf forest in the Swiss Alps. Analysis of diurnal radiation patterns revealed a strong linear relationship (R = 0.94) between outgoing shortwave radiation and sunlit snow‐view fraction, highlighting shading as the main control on the subcanopy shortwave radiation budget. Measurements of outgoing longwave radiation were strongly controlled by the snow cover extent, with locations of diminished snow cover showing an increase in outgoing longwave radiation of up to 60 W/m². The subcanopy radiation budget was shown to be dominated by shortwave radiation when surrounding canopy structure and the position of the sun allowed for direct insolation of the forest floor, but longwave radiation was the dominating component in the absence of direct insolation.

The retrieval of detailed, co‐located snow depth and canopy cover information from airborne lidar has advanced our understanding of links between forest snow distribution and canopy structure. In this study, we present two recent high‐resolution (1 m) lidar data sets acquired in (i) a 2017 mission in the Eastern Swiss Alps and (ii) NASA's 2017 SnowEx field campaign at Grand Mesa, Colorado. Validation of derived snow depth maps against extensive manual measurements revealed a RMSE of 6 and 3 cm for plot‐level mean and standard deviation of snow depth, respectively, demonstrating that within‐stand snow distribution patterns were captured reliably. Lidar data were further processed to obtain canopy structure metrics. To this end, we developed a novel approach involving a continuous measure of local distance to canopy edge (DCE), which enabled creating spatially aggregated nondirectional and directional descriptors of the canopy structure. DCE‐based canopy metrics were correlated to mean and standard deviation of snow depth over areas representing grid‐cell sizes typical of watershed and regional model applications (20–200 m). Snow depth increased along the DCE gradient from dense canopy to the center of canopy gaps for all sites and acquisition times, while directional effects particularly evolved during the ablation season. These findings highlight the control of canopy gap distribution on snow distribution in discontinuous forests, with higher snow depths where the open fraction is concentrated in few large gaps rather than many fragmented small gaps. In these environments, dedicated canopy structure metrics such as DCE should advance spatially distributed snow modeling.