Snow Profiles Research Articles

Post-depositional modification on seasonal-to-interannual timescales alters the deuterium-excess signals in summer snow layers in Greenland

Abstract. We document the isotopic evolution of near-surface snow at the East Greenland Ice Core Project (EastGRIP) ice core site in northeast Greenland using a time-resolved array of 1 m deep isotope (δ18O, δD) profiles. The snow profiles were taken from May–August during the 2017–2019 summer seasons. An age–depth model was developed and applied to each profile, mitigating the impacts of stratigraphic noise on isotope signals. Significant changes in deuterium excess (d) are observed in surface snow and near-surface snow as the snow ages. Decreases in d of up to 5 ‰ occur during summer seasons after deposition during two of the three summer seasons observed. The d always experiences a 3 ‰–5 ‰ increase after aging 1 year in the snow due to a broadening of the autumn d maximum. Models of idealized scenarios coupled with prior work indicate that the summertime post-depositional changes in d (Δd) can be explained by a combination of surface sublimation, forced ventilation of the near-surface snow down to 20–30 cm, and isotope-gradient-driven diffusion throughout the column. The interannual Δd is also partly explained with isotope-gradient-driven diffusion, but other mechanisms are at work that leave a bias in the d record. Thus, d does not just carry information about source-region conditions and transport history as is commonly assumed, but also integrates local conditions into summer snow layers as the snow ages through metamorphic processes. Finally, we observe a dramatic increase in the seasonal isotope-to-temperature sensitivity, which can be explained solely by isotope-gradient-driven diffusion. Our results are dependent on the site characteristics (e.g., wind, temperature, accumulation rate, snow properties) but indicate that more process-based research is necessary to understand water isotopes as climate proxies. Recommendations for monitoring and physical modeling are given, with special attention to the d parameter.

Ground-penetrating radar as a tool for determining the interface between temperate and cold ice, and snow depth: a case study for Hurd-Johnsons glaciers, Livingston Island, Antarctica

Abstract We analyze the internal structure of two polythermal glaciers, Hurd and Johnsons, located on Livingston Island, Antarctica, using 200 and 750 MHz GPR data collected in 2003/04, 2008/09 and 2016/17 field campaigns. Based on the different permittivities of snow and ice, we determined the thickness distribution of the end-of winter snow cover and of the cold ice layer. Their knowledge is fundamental for mass balance and glacier dynamics studies due to the different densities and rheological properties of such media. The average measured thicknesses for the snow and cold ice layers (the latter including the snow layer) were of 1.44 ± 0.09 and 29.1 ± 1.5 m, and their corresponding maxima were of 2.45 ± 0.21 and 80.8 ± 2.5 m. GPR snow profiling allowed for extension of the coverage of the snow thickness survey, but added little information to that supplied by snow pits, stake readings and manual snow probing, because of the multiplicity of reflections within the seasonal snowpack caused by internal ice layers and lenses. The polythermal structure determined for Hurd Glacier fits into the so-called Scandinavian type, seldom reported for the Antarctic region.

Development of an Autonomous Measurement System for Estimation of Snow Profile on Road Surfaces

This paper presents the development of a novel prototype system to measure the profile of snow and detect ice on road surfaces. While researchers have previously developed methods to estimate the tire-road friction coefficient, methods to create a high resolution map of snow cover and ice on roads have not been developed. The prototype and data processing methods presented in this study provide a potential solution that would greatly benefit winter road maintenance teams. Having a device capable of creating a detailed map of snow and ice cover on roads would allow for more efficient plowing and salt deployment, better informing of the public about current road conditions, as well as better analysis of current plowing techniques to improve winter road maintenance operations. The measurement of the snow profile is based on the use of two types of inexpensive ranging sensors mounted on a rotationally controlled stepper motor platform. The technical challenges addressed in the development include analysis of the scan path over which the system scans the road, differentiation between asphalt, concrete, and snow surfaces using reflection amplitudes, accurate transformation of raw distance measurements to inertial coordinates, and identification of the Euler angles of the sensors using parameters associated with the scanned surface profile. Extensive experimental results and test data are presented, which show that the system is capable of measuring the snow profile on a road with an accuracy of ±1 cm when a reasonable portion of bare road is visible, and ±2 cm when only small portions of bare road are visible.

Automatic snow type classification of snow micropenetrometer profiles with machine learning algorithms

Abstract. Snow-layer segmentation and classification are essential diagnostic tasks for various cryospheric applications. The SnowMicroPen (SMP) measures the snowpack's penetration force at submillimeter intervals in snow depth. The resulting depth–force profile can be parameterized for density and specific surface area. However, no information on traditional snow types is currently extracted automatically. The labeling of snow types is a time-intensive task that requires practice and becomes infeasible for large datasets. Previous work showed that automated segmentation and classification is, in theory, possible but cannot be applied to data straight from the field or needs additional time-costly information, such as from classified snow pits. We evaluate how well machine learning models can automatically segment and classify SMP profiles to address this gap. We trained 14 models, among them semi-supervised models and artificial neural networks (ANNs), on the MOSAiC SMP dataset, an extensive collection of snow profiles on Arctic sea ice. SMP profiles can be successfully segmented and classified into snow classes based solely on the SMP's signal. The model comparison provided in this study enables SMP users to choose a suitable model for their task and dataset. The findings presented will facilitate and accelerate snow type identification through SMP profiles. Anyone can access the tools and models needed to automate snow type identification via the software repository “snowdragon”. Overall, snowdragon creates a link between traditional snow classification and high-resolution force–depth profiles. Traditional snow profile observations can be compared to SMP profiles with such a tool.

Combining modelled snowpack stability with machine learning to predict avalanche activity

Abstract. Predicting avalanche activity from meteorological and snow cover simulations is critical in mountainous areas to support operational forecasting. Several numerical and statistical methods have tried to address this issue. However, it remains unclear how combining snow physics, mechanical analysis of snow profiles and observed avalanche data improves avalanche activity prediction. This study combines extensive snow cover and snow stability simulations with observed avalanche occurrences within a random forest approach to predict avalanche situations at a spatial resolution corresponding to elevations and aspects of avalanche paths in a given mountain range. We develop a rigorous leave-one-out evaluation procedure including an independent evaluation set, confusion matrices and receiver operating characteristic curves. In a region of the French Alps (Haute-Maurienne) and over the period 1960–2018, we show the added value within the machine learning model of considering advanced snow cover modelling and mechanical stability indices instead of using only simple meteorological and bulk information. Specifically, using mechanically based stability indices and their time derivatives in addition to simple snow and meteorological variables increases the probability of avalanche situation detection from around 65 % to 76 %. However, due to the scarcity of avalanche events and the possible misclassification of non-avalanche situations in the training dataset, the predicted avalanche situations that are really observed remains low, around 3.3 %. These scores illustrate the difficulty of predicting avalanche occurrence with a high spatio-temporal resolution, even with the current data and modelling tools. Yet, our study opens perspectives to improve modelling tools supporting operational avalanche forecasting.

Snow profile reconstruction from tomographic UAV SAR

Snow profiles can provide reliable and detailed snow parameters for global change research, but it is still challenging to reconstruct large-scale spatiotemporally continuous snow profiles. Synthetic aperture radar (SAR) tomography (TomoSAR) has been proven to be a promising method for reconstructing the 3D structure of targets and has been used to reconstruct snow profiles from ground-based SAR. However, the experimental condition of ground-based SAR is ideal and the algorithms developed for ground-based SAR have some problems when applied to airborne and spaceborne sensors, such as large computational load and big angular diversity. In this paper, we first collected the Ku-band tomographic SAR dataset in the snow season of Altay, China using the Unmanned Aerial Vehicle (UAV) platform. Then, we reconstructed the snow profile from the dataset using three spectral estimation-based focusing algorithms. According to the non-stationary characteristic of the UAV platform and the dense medium property of snow, phase calibration and medium correction were performed, respectively. Finally, we assessed the reconstruction accuracy of the snow profile and snow depth using the ground synchronous measurement data. The reconstruction accuracy of the three focusing algorithms from high to low is the MUSIC algorithm, the Capon algorithm, and the Beamforming algorithm. Among them, the average error of the MUSIC algorithm to reconstruct the snow profile is 1.13 cm, and the RMSE and MAE of the snow depth retrieval are 3.71 cm and 3.26 cm, respectively. Although TomoSAR is still widely unexplored for snow using air-based sensors, the results show that it has the ability to reconstruct snow profiles accurately, and the proposed method can provide reliable support for reconstructing the large-scale spatiotemporal continuous snow profiles from subsequent spaceborne SAR.

Snow Layer Prediction Model using Back Scattering Co-efficient

Abstract: In this study, the objective is to estimate the physical parameters of each snow layer at a given certain altitude, which is according to the problem will always be between other altitudes, whose profile data is entirely provided. The provided data for the target profile will also consist of the number of layers, height of each layer, the back-scattering coefficient and the altitude. The goal is to estimate the interval [࢔࢏࢓࢟ ,࢞ࢇ࢓࢟ [for each parameter, which are sphericity, diameter and density of the snowflakes. We are using here a multilayer snow profile and then applying the back scattering coefficients to calculate the snow layer altitude and height. Further, by correct prediction we can preserve the snow layers by continuously monitoring the snow model at the correct altitude throughout the year. Our aim demonstrates the comparatively better way to predict the best snow profile at right altitudes.

A Snapshot on the Buildup of the Stable Water Isotopic Signal in the Upper Snowpack at EastGRIP on the Greenland Ice Sheet

AbstractThe stable water isotopic composition in firn and ice cores provides valuable information on past climatic conditions. Because of uneven accumulation and post‐depositional modifications on local spatial scales up to hundreds of meters, time series derived from adjacent cores differ significantly and do not directly reflect the temporal evolution of the precipitated snow isotopic signal. Hence, a characterization of how the isotopic profile in the snow develops is needed to reliably interpret the isotopic variability in firn and ice cores. By combining digital elevation models of the snow surface and repeated high‐resolution snow sampling for stable water isotope measurements of a transect at the East Greenland Ice‐core Project campsite on the Greenland Ice Sheet, we are able to visualize the buildup and post‐depositional changes of the upper snowpack across one summer season. To this end, 30 cm deep snow profiles were sampled on six dates at 20 adjacent locations along a 40 m transect. Near‐daily photogrammetry provided snow height information for the same transect. Our data shows that erosion and redeposition of the original snowfall lead to a complex stratification in the δ18O signature. Post‐depositional processes through vapor‐snow exchange affect the near surface snow with d‐excess showing a decrease in surface and near‐surface layers. Our data suggests that the interplay of stratigraphic noise, accumulation intermittency, and local post‐depositional processes form the proxy signal in the upper snowpack.

Geochemical Markers of Stagnant Zones in an Urban Heat Island

The paper presents data acquired by a comparative study of the vertical variability in the chemical composition and ratios of aerosol subdisperse fractions in snow layers chronologically correlated with the stratigraphically significant snowfall periods. These data highlight features of the concentrating of trace elements at reactive (geochemical) barriers in the snow profile. The ratio of three geochemically close groups of elements, such as siderophile, sulfophile, and lithophile elements, were found out to vary relatively little from one snow layer to another in the growing snow cover. Trajectory analysis of the transfer of air masses to which stratigraphically significant snowfalls were related to the observation site provides no evidence that the identified geochemical phenomenon can be explained by that the winter aerosol field that was formed above an urban area with different trajectories of air masses can be somehow inherited in the snow layers of the growing snow cover and thus affect the vertical distributions of trace elements. Evidence indicates that that the ratios of assemblages of trace elements in the discrete snow layers, which remain stable with the growth of the snow mass, can be employed as geochemical markers of stagnant zones in an urban heat island, and the method of geochemical sampling of a snow cover in its discrete layers at a rare network of urban meteorological observation sites is an efficient additional tool for studying microscale atmospheric processes and recovering information on characteristics of the transfer of pollutants in a spatially limited urban area.

Fingerprints of Frontal Passages and Post‐Depositional Effects in the Stable Water Isotope Signal of Seasonal Alpine Snow

AbstractStable water isotopes are used as a paleothermometer in ice cores for climate reconstructions over the past millennia. The underlying physical processes involved in the isotope‐temperature relation, however, unfold at much shorter timescales. Here, we study the temporary archival of frontal passages in the seasonal Alpine snow cover. We combine five snow profiles sampled in winter 2017 at the Weissfluhjoch with a quantitative snow layer age reconstruction and atmospheric reanalysis data to characterize the circulation and clouds associated with the precipitation producing synoptic‐scale cold and warm fronts. We find that the vertical cloud structure and the air parcels' net cooling during transport leave a distinct imprint in the δ18O and δD vertical profile in the snow. The near‐surface humidity gradient at the moisture source is reflected in the second order isotope parameter deuterium excess. In the cold season, these environmental conditions during cloud formation and at the moisture source are preserved in the snow. In the melt season, significant post‐depositional effects due to wet snow metamorphism, however, leads to an enrichment in heavy isotopes in the snow and a strong smoothing of the initial atmospheric imprint. These findings show that the isotope signal archived in the dry snow cover is strongly modulated by individual weather systems prior to deposition. Major shifts in the upper‐level jet stream and cyclone tracks likely leading to changes in moisture source regions and conditions, could therefore be detectable in the isotope composition of Alpine ice.

A random forest model to assess snow instability from simulated snow stratigraphy

Abstract. Modeled snow stratigraphy and instability data are a promising source of information for avalanche forecasting. While instability indices describing the mechanical processes of dry-snow avalanche release have been implemented into snow cover models, there exists no readily applicable method that combines these metrics to predict snow instability. We therefore trained a random forest (RF) classification model to assess snow instability from snow stratigraphy simulated with SNOWPACK. To do so, we manually compared 742 snow profiles observed in the Swiss Alps with their simulated counterparts and selected the simulated weak layer corresponding to the observed rutschblock failure layer. We then used the observed stability test result and an estimate of the local avalanche danger to construct a binary target variable (stable vs. unstable) and considered 34 features describing the simulated weak layer and the overlying slab as potential explanatory variables. The final RF classifier aggregates six of these features into the output probability Punstable, corresponding to the mean vote of an ensemble of 400 classification trees. Although the subset of training data only consisted of 146 profiles labeled as either unstable or stable, the model classified profiles from an independent validation data set (N=121) with high reliability (accuracy 88 %, precision 96 %, recall 85 %) using manually predefined weak layers. Model performance was even higher (accuracy 93 %, precision 96 %, recall 92 %), when the weakest layers of the profiles were identified with the maximum of Punstable. Finally, we compared model predictions to observed avalanche activity in the region of Davos for five winter seasons. Of the 252 avalanche days (345 non-avalanche days), 69 % (75 %) were classified correctly. Overall, the results of our RF classification are very encouraging, suggesting it could be of great value for operational avalanche forecasting.

Hansbreen Snowpit Dataset – over 30-year of detailed snow research on an Arctic glacier

Snow cover is a key element in the water cycle, global heat balance and in the condition of glaciers. Characterised by high temporal and spatial variability, it is subject to short- and long-term changes in climatic conditions. This paper presents a unique dataset of snow measurements on Hansbreen, an Arctic glacier in Svalbard. The dataset includes 79 archived snow profiles performed from 1989 to 2021. It presents all available observations of physical properties for snow cover, such as grain shape and size, hardness, wetness, temperature and density, supplemented with organised metadata. All data has been revised and unified with current protocols and the present International Classification for Seasonal Snow on the Ground, allowing comparison of data from different periods and locations. The information included is essential for estimations of glacier mass balance or snow depth using indirect methods, such as ground-penetrating radar. A wide range of input data makes this dataset valuable to the greater community involved in the study of snow cover evolution and modelling related to glaciology, ecology and hydrology of glacierised areas.

A data exploration tool for averaging and accessing large data sets of snow stratigraphy profiles useful for avalanche forecasting

Abstract. Snowpack models can provide detailed insight about the evolution of the snow stratigraphy in a way that is not possible with direct observations. However, the lack of suitable data aggregation methods currently prevents the effective use of the available information, which is commonly reduced to bulk properties and summary statistics of the entire snow column or individual grid cells. This is only of limited value for operational avalanche forecasting and has substantially hampered the application of spatially distributed simulations, as well as the development of comprehensive ensemble systems. To address this challenge, we present an averaging algorithm for snow profiles that effectively synthesizes large numbers of snow profiles into a meaningful overall perspective of the existing conditions. Notably, the algorithm enables compiling of informative summary statistics and distributions of snowpack layers, which creates new opportunities for presenting and analyzing distributed and ensemble snowpack simulations.

Brief communication: Improving ERA5-Land soil temperature in permafrost regions using an optimized multi-layer snow scheme

Abstract. We previously reported a notable warm bias in ERA5-Land soil temperature in permafrost regions that was supposedly being caused by an underestimation of snow density. In this study, we implemented and evaluated a new multi-layer snow scheme in the land surface scheme of ERA5-Land, i.e., HTESSEL, with revised snow densification parameterizations. We compared permafrost soil temperatures from the numerical experiments with observations and the original ERA5-Land with a single-layer snow scheme. The revised HTESSEL significantly improved the representation of soil temperature in permafrost regions compared to ERA5-Land. The daily warm bias in winter was reduced by about 0.6–3.0 ∘C across the 522 observing stations in high-latitude permafrost regions, and the resulting modeled near-surface permafrost extent was improved (11.0–12.9×106 km2 during 2001–2018), comparing reasonably with observed estimates for continuous and discontinuous permafrost areas. We therefore suggest that a better-resolved snow scheme with a multi-layer snow profile should be included in next-generation reanalyses as a first step towards improving the representation of permafrost.

Investigation of effects of repose angle and critical sliding angle of snow on curved roofs

Accompanied by extreme snowy weather occurred frequently, the number of collapse accidents of structures which are caused by uneven snowdrift is increasing. The formation of uneven snowdrift on the roofs is substantially related to snow profile. Especially to curved roofs, snow profile has certain peculiarity due to curvature of roof edge. The inter-particle properties such as repose angle and critical sliding angle of snow also affect snow profile on curved roofs. This study is to systematically investigate laws of inter-particle properties with measuring, provide a revision approach to accurately restore snow profile considering inter-particle properties, and validate the feasibility of approach by comparing simulation results with that of wind tunnel test. Firstly, a series measuring of repose angle is conducted and results show that repose angle is about 60°. Then the wind tunnel test on curved roofs is carried out to investigate properties of critical sliding angle and found that critical sliding angle is around 60°. Based on inter-particle properties, a revision approach of snow profile is proposed. Snow distribution of numerical simulation with revised snow profile is more in line with wind tunnel test. It is found that snow profile which is revised by inter-particle properties influences characteristics of flow field around roofs and snowdrift on curved roofs.

Modelling snowpack stability from simulated snow stratigraphy: Summary and implementation examples

Information on snowpack stability, i.e., on the propensity for failure initiation and crack propagation in a weak layer, is essential for forecasting snow avalanches. To complement field observations, snow cover modelling provides information otherwise unavailable on the present and future state of the snow cover, and can be used to evaluate snowpack stability. The main goal of this paper is to summarize the broad spectrum of models to assess snowpack stability from simulated snow profiles. The basic mechanical concepts behind these stability models include: the maximum stress criterion which characterizes the failure initiation propensity and the critical crack length to evaluate the crack propagation propensity. However, many subtle differences between models, mainly due to additional expert rules or the effective implementation of the concepts, can be confusing. We try to disentangle this diversity in this summary. We discuss the differences and also present an overview of the mechanical parameterizations of snow material properties such as strength or stiffness as they are a key ingredient for stability modelling. In addition, we apply the stability models to typical and simplified snow profiles in order to illustrate the influence of the underlying assumptions and the model sensitivity to the mechanical input. As we point out scientific challenges and model limitations, the examples we discussed can provide guidance on the interpretation of similar model results. Moreover, we draw some guiding lines for future research concerning snowpack stability assessment based on snow cover modelling.

Avalanche danger level characteristics from field observations of snow instability

Abstract. Avalanche danger levels are described in qualitative terms that mostly are not amenable to measurements or observations. However, estimating and improving forecast consistency and accuracy require descriptors that can be observed or measured. Therefore, we aim to characterize the avalanche danger levels based on expert field observations of snow instability. We analyzed 589 field observations by experienced researchers and forecasters recorded mostly in the region of Davos (Switzerland) during 18 winter seasons (2001–2002 to 2018–2019). The data include a snow profile with a stability test (rutschblock, RB) and observations on snow surface quality, drifting snow, signs of instability and avalanche activity. In addition, observers provided their estimate of the local avalanche danger level. A snow stability class (very poor, poor, fair, good, very good) was assigned to each profile based on RB score, RB release type and snowpack characteristics. First, we describe some of the key snowpack characteristics of the data set. In most cases, the failure layer included persistent grain types even after a recent snowfall. We then related snow instability data to the local avalanche danger level. For the danger levels 1–Low to 4–High, we derived typical stability distributions. The proportions of profiles rated poor and very poor clearly increased with increasing danger level. For our data set, the proportions were 5 %, 13 %, 49 % and 63 % for the danger levels 1–Low to 4–High, respectively. Furthermore, we related the local avalanche danger level to the occurrence of signs of instability such as whumpfs, shooting cracks and recent avalanches. The absence of signs of instability was most closely related to 1–Low and the presence of them to 3–Considerable. Adding the snow stability class and the 3 d sum of new snow depth improved the discrimination between the lower three danger levels. Still, 2–Moderate was not well described. Nevertheless, we propose some typical situations that approximately characterize each of the danger levels. Obviously, there is no single easily observable set of parameters that would allow us to fully characterize the avalanche danger levels. One reason for this shortcoming is the fact that the snow instability data we analyzed usually lack information on spatial frequency, which is needed to reliably assess the danger level.

Similarity conditions and cube model tests of snow drift and precipitation preferential deposition patterns

We explore the process of modeling aeolian snow transport around buildings by proposing a set of similarity requirements on simulating a single snow fall event with drifting. The similarity criteria are used to guide our wind tunnel tests from the particle selection to the simulation set-up, including the determination of test wind speed, snowfall intensity and transported snow profile. In parallel, a specific field experiment on snow precipitation and drift around a 1-m prototype cube was achieved in Harbin to serve as the reference for the subsequent test validation. After reproducing a target wind and snow environment with a scale of 1:4 versus that of the prototype measurement, two verification tests were put forward using snow and substitution particles, which can mimic the effects of changing wind directions on the snow distribution. Herein, the wind tunnel and outside prototype results show good qualitative agreement in our limited range of length scales and particle properties. In future work, the range of scales and range of parameter values should be increased.

Evaluation of the Effective Microstructure Parameter of the Microwave Emission Model of Layered Snowpack for Multiple-Layer Snow

Natural snow, one of the most important components of the cryosphere, is fundamentally a layered medium. In forward simulation and retrieval, a single-layer effective microstructure parameter is widely used to represent the emission of multiple-layer snowpacks. However, in most cases, this parameter is fitted instead of calculated based on a physical theory. The uncertainty under different frequencies, polarizations, and snow conditions is uncertain. In this study, we explored different methods to reduce the layered snow properties to a set of single-layer values that can reproduce the same brightness temperature (TB) signal. A validated microwave emission model of layered snowpack (MEMLS) was used as the modelling tool. Multiple-layer snow TB from the snow’s surface was compared with the bulk TB of single-layer snow. The methods were tested using snow profile samples from the locally validated and global snow process model simulations, which follow the natural snow’s characteristics. The results showed that there are two factors that play critical roles in the stability of the bulk TB error, the single-layer effective microstructure parameter, and the reflectivity at the air–snow and snow–soil boundaries. It is important to use the same boundary reflectivity as the multiple-layer snow case calculated using the snow density at the topmost and bottommost layers instead of the average density. Afterwards, a mass-weighted average snow microstructure parameter can be used to calculate the volume scattering coefficient at 10.65 to 23.8 GHz. At 36.5 and 89 GHz, the effective microstructure parameter needs to be retrieved based on the product of the snow layer transmissivity. For thick snow, a cut-off threshold of 1/e is suggested to be used to include only the surface layers within the microwave penetration depth. The optimal method provides a root mean squared error of bulk TB of less than 5 K at 10.65 to 36.5 GHz and less than 10 K at 89 GHz for snow depths up to 130 cm.

Cross-validation of active and passive microwave snowfall products over the continental United States

AbstractSurface snowfall rate estimates from the Global Precipitation Measurement (GPM) mission’s Core Observatory sensors and the CloudSat radar are compared to those from the Multi-Radar Multi-Sensor (MRMS) radar composite product over the continental United States during the period from November 2014 to September 2020. The analysis includes: the Dual-Frequency Precipitation Radar (DPR) retrieval and its single frequency counterparts, the GPM Combined Radar Radiometer Algorithm (CORRA), the CloudSat Snow Profile product (2C-SNOW-PROFILE) and two passive microwave retrievals, i.e., the Goddard PROFiling algorithm (GPROF) and the Snow retrievaL ALgorithm fOr gMi (SLALOM). The 2C-SNOW retrieval has the highest Heidke Skill Score (HSS) for detecting snowfall among the products analysed. SLALOM ranks second; it outperforms GPROF and the other GPM algorithms, all detecting only 30% of the snow events. Since SLALOM is trained with 2C-SNOW, it suggests that the optimal use of the information content in the GMI observations critically depends on the precipitation training dataset. All the retrievals underestimate snowfall rates by a factor of two compared to MRMS. Large discrepancies (RMSE of 0.7 to 1.5 mm h-1) between space-borne and ground-based snowfall rate estimates are attributed to the complexity of the ice scattering properties and to the limitations of the remote sensing systems: the DPR instrument has low sensitivity, while the radiometric measurements are affected by the confounding effects of the background surface emissivity and of the emission of supercooled liquid droplet layers.