Snowfall Products Research Articles

Multi-year evaluation of CloudSat, ERA5, and IMERG global snowfall products

Snowfall products derived from satellite-based radar observations, reanalysis products and phase-classified precipitation products are the three primary sources of information on surface snowfall. This study compares snowfall rate products from CloudSat, ERA5 and IMERG using the Triple Collocation (TC) method on a global scale covering almost the entire period of CloudSat cloud-profiling radar (CPR) 2C products from 2006 to 2019, especially considering the precipitation phase classification field of IMERG (probability of Liquid Precipitation, pLP). The main conclusions include but are not limited to the following: (1) The pLP threshold significantly impacts the ability of IMERG to identify snowfall from precipitation, with an optimal pLP threshold of 25 percent determined by matching IMERG with CloudSat and ERA5 snowfall events globally. (2) CloudSat presents the highest performance and stability on a global scale, with average TC-estimated CC and RMSE of 0.472 and 1.465, respectively. (3) The TC-estimated CC of IMERG exhibit significant spatiotemporal heterogeneities, but IMERG product with the optimal pLP threshold show comparable performance with CloudSat, with an average TC-estimated RMSE of ∼ 1.533. (4) ERA5 performed worse than the other two products, except in some high-latitude and high-altitude areas. This study provides novel insights into the performances of snowfall rate products from different sources at the global scales and provides a foundation for improving satellite- and model-based snowfall rate retrievals.

Exploring Snowfall Variability through the High-Latitude Measurement of Snowfall (HiLaMS) Field Campaign

Abstract The High-Latitude Measurement of Snowfall (HiLaMS) campaign explored variability in snowfall properties and processes at meteorologically distinct field sites located in Haukeliseter, Norway, and Kiruna, Sweden, during the winters of 2016/17 and 2017/18, respectively. Campaign activities were founded upon the sensitivities of a low-cost, core instrumentation suite consisting of Micro Rain Radar, Precipitation Imaging Package, and Multi-Angle Snow Camera. These instruments are highly portable to remote field sites and, considered together, provide a unique and complementary set of snowfall observations including snowflake habit, particle size distributions, fall speeds, surface snowfall accumulations, and vertical profiles of radar moments and snow water content. These snow-specific parameters, used in combination with existing observations from the field sites such as snow gauge accumulations and ambient weather conditions, allow for advanced studies of snowfall processes. HiLaMS observations were used to 1) successfully develop a combined radar and in situ microphysical property retrieval scheme to estimate both surface snowfall accumulation and the vertical profile of snow water content, 2) identify the predominant snowfall regimes at Haukeliseter and Kiruna and characterize associated macrophysical and microphysical properties, snowfall production, and meteorological conditions, and 3) identify biases in the HARMONIE-AROME numerical weather prediction model for forecasts of snowfall accumulations and vertical profiles of snow water content for the distinct snowfall regimes observed at the mountainous Haukeliseter site. HiLaMS activities and results suggest value in the deployment of this enhanced snow observing instrumentation suite to new and diverse high-latitude locations that may be underrepresented in climate and weather process studies.

A Machine Learning Snowfall Retrieval Algorithm for ATMS

This article describes the development of a machine learning (ML)-based algorithm for snowfall retrieval (Snow retrievaL ALgorithm fOr gpM–Cross Track, SLALOM-CT), exploiting ATMS radiometer measurements and using the CloudSat CPR snowfall products as references. During a preliminary analysis, different ML techniques (tree-based algorithms, shallow and convolutional neural networks—NNs) were intercompared. A large dataset (three years) of coincident observations from CPR and ATMS was used for training and testing the different techniques. The SLALOM-CT algorithm is based on four independent modules for the detection of snowfall and supercooled droplets, and for the estimation of snow water path and snowfall rate. Each module was designed by choosing the best-performing ML approach through model selection and optimization. While a convolutional NN was the most accurate for the snowfall detection module, a shallow NN was selected for all other modules. SLALOM-CT showed a high degree of consistency with CPR. Moreover, the results were almost independent of the background surface categorization and the observation angle. The reliability of the SLALOM-CT estimates was also highlighted by the good results obtained from a direct comparison with a reference algorithm (GPROF).

A Snowfall Detection Algorithm for ATMS Over Ocean, Sea Ice, and Coast

This study developed a snowfall detection algorithm over ocean, sea ice, and coast for the advanced technology microwave sounder (ATMS) onboard both NPP and NOAA-20 satellites. The detection algorithm was trained from collocated observations between ATMS-NPP and CloudSat cloud profiling radar (CPR) snowfall product. Both brightness temperature (TB) variables and global forecast system (GFS) output variables are evaluated for snowfall detection in this algorithm. Results show that combining TB variables and GFS variables provide the optimal snowfall detection performance. The Heidke skill score (HSS) values are about 0.56 over all three surface types, and the probability of detection (POD) values are 0.76, 0.70, and 0.72 over ocean, sea ice, and coast, respectively. The importance of the GFS variables differs greatly among these three surface types. The detection algorithm primarily depends on TB variables over ocean and HSS only increased by 0.05 by adding GFS variables. In contrast, GFS variables are critically important to snowfall detection over sea ice and coastal regions. Without GFS variables, the HSS values over both sea ice and coastal regions decrease sharply from about 0.56 to about 0.40. Over ocean, we also developed a regional snowfall detection model in each 10° grid box, which greatly outperform the global detection model over certain regions (e.g., sea of Okhotsk and Labrador Sea). Case studies and validation against NOAA-20 observations showed that the snowfall detection algorithm performs well, which will benefit coastal communities by providing information on snowstorms offshore before they transition to land.

Biases in CloudSat Falling Snow Estimates Resulting from Daylight-Only Operations

Falling snow is a key component of the Earth’s water cycle, and space-based observations provide the best current capability to evaluate it globally. The Cloud Profiling Radar (CPR) on board CloudSat is sensitive to snowfall, and other satellite missions and climatological models have used snowfall properties measured by it for evaluating and comparing against their snowfall products. Since a battery anomaly in 2011, the CPR has operated in a Daylight-Only Operations (DO-Op) mode, in which it makes measurements primarily during only the daylit portion of its orbit. This work provides estimates of biases inherent in global snowfall amounts derived from CPR measurements due to this shift to DO-Op mode. We use CloudSat’s snowfall measurements during its Full Operations (Full-Op) period prior to the battery anomaly to evaluate the impact of the DO-Op mode sampling. For multi-year global mean values, the snowfall fraction during DO-Op changes by −10.16% and the mean snowfall rate changes by −8.21% compared with Full-Op. These changes are driven by the changes in sampling in DO-Op and are very little influenced by changes in meteorology between the Full-Op and DO-Op periods. The results highlight the need to sample consistently with the CloudSat observations or to adjust snowfall estimates derived from CloudSat when using DO-Op data to evaluate other precipitation products.

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.

A new approach for bias adjustment of IMERG remotely sensed snowfall product

Due to uncertainty in the nature of indirect remotely sensed precipitation products, bias adjustment is a crucial step. In this research, bias adjustment of Integrated Multi-satellitE Retrievals for GPM (Global Precipitation Measurement) (IMERG) snowfall estimates with 0.1° spatial resolution and half-hour temporal resolution was performed over a basin located in the Western United States based on Snowpack Telemetry (SNOTEL) ground station snow accumulation (SA) observations using the Gridded based Bias Adjustment using Copula (GBAC) approach. After dividing the study area into several boxes, usual and unusual pairs of cumulative distribution functions (CDFs) were separated in station pixels. Then, simulated biases were obtained using multivariate copula functions in station boxes. Next, based on the T copula function, bias adjustment bands were generated during the calibration period in station pixels. This step was followed by the validation stage. Also, bias adjustment bands were generated in pixels with no station. Based on calibration results, in boxes located west of 120° W, the range between the lower bound and the 10th percentile of the band had the most impact on satellite SA adjustment. Conversely, in boxes located east of 120° W, the middle range of the band (between 10th and 90th percentiles) was more effective. Thus, longitude was an effective factor in satellite SA bias adjustment. The validation results showed that in station boxes with a longitude over 120° west, the adjustment band was narrower. This is while in station boxes with longitude less than 120° west, the bias adjustment band better encompassed the ground SA data.

Constraining Reanalysis Snowfall Over the Arctic Ocean Using CloudSat Observations

AbstractIn the absence of widespread snowfall observations over the Arctic Ocean, reanalysis products provide a wide range of estimates of time‐evolving snowfall rates over Arctic sea ice, and it can be difficult to determine which product is most representative. In this work, Arctic snowfall rates retrieved from 2006 to 2016 CloudSat observations and snowfall products from three reanalyses are assessed. The products can be brought into encouraging agreement over the region on interannual time scales once differences in spatial representativeness and temporal sampling are accounted for. This motivates the use of CloudSat's snowfall product to calibrate reanalysis snowfall. The calibration is carried out for four Arctic quadrants and combined to produce regionally resolved and consistent estimates of interannually varying snowfall. Calibrated reanalysis snowfall inputs are then used to drive the NASA Eulerian Snow On Sea Ice Model, reducing the interproduct spread in the resulting simulated snow depths across the Arctic.

Spatial and temporal variability of snowfall over Greenland from CloudSat observations

Abstract. We use the CloudSat 2006–2016 data record to estimate snowfall over the Greenland Ice Sheet (GrIS). We first evaluate CloudSat snowfall retrievals with respect to remaining ground-clutter issues. Comparing CloudSat observations to the GrIS topography (obtained from airborne altimetry measurements during IceBridge) we find that at the edges of the GrIS spurious high-snowfall retrievals caused by ground clutter occasionally affect the operational snowfall product. After correcting for this effect, the height of the lowest valid CloudSat observation is about 1200 m above the local topography as defined by IceBridge. We then use ground-based millimeter wavelength cloud radar (MMCR) observations obtained from the Integrated Characterization of Energy, Clouds, Atmospheric state, and Precipitation at Summit, Greenland (ICECAPS) experiment to devise a simple, empirical correction to account for precipitation processes occurring between the height of the observed CloudSat reflectivities and the snowfall near the surface. Using the height-corrected, clutter-cleared CloudSat reflectivities we next evaluate various Z–S relationships in terms of snowfall accumulation at Summit through comparison with weekly stake field observations of snow accumulation available since 2007. Using a set of three Z–S relationships that best agree with the observed accumulation at Summit, we then calculate the annual cycle snowfall over the entire GrIS as well as over different drainage areas and compare the derived mean values and annual cycles of snowfall to ERA-Interim reanalysis. We find the annual mean snowfall over the GrIS inferred from CloudSat to be 34±7.5 cm yr−1 liquid equivalent (where the uncertainty is determined by the range in values between the three different Z–S relationships used). In comparison, the ERA-Interim reanalysis product only yields 30 cm yr−1 liquid equivalent snowfall, where the majority of the underestimation in the reanalysis appears to occur in the summer months over the higher GrIS and appears to be related to shallow precipitation events. Comparing all available estimates of snowfall accumulation at Summit Station, we find the annually averaged liquid equivalent snowfall from the stake field to be between 20 and 24 cm yr−1, depending on the assumed snowpack density and from CloudSat 23±4.5 cm yr−1. The annual cycle at Summit is generally similar between all data sources, with the exception of ERA-Interim reanalysis, which shows the aforementioned underestimation during summer months.

Evaluation of IMERG and MRMS remotely sensed snowfall products

ABSTRACTIn this study, snow accumulation (SA) estimates of the Integrated Multi-satellitE Retrievals for GPM (Global Precipitation Measurement) (IMERG) and Multi-Radar/Multi-Sensor (MRMS) products were evaluated against the SNOwpack TELemetry (SNOTEL) ground observations over a Basin in the western United States from October 2016 to February 2017. IMERG underestimated SA in three snowfall probability thresholds of 45%, 65% and 85%. With increasing the threshold from 45% to 85%, MRMS Bias index showed overestimation compared to that of the IMERG. Overall, MRMS presented more accurate results than the IMERG. In categorical analysis, IMERG had better probability of detection (POD) values than the MRMS although MRMS was generally more accurate than the IMERG in all thresholds. Moreover, with respect to Bias, mean absolute error (MAE) and root mean square error (RMSE) indices in various elevation classes, IMERG was more efficient in lower elevation classes while in categorical analysis, MRMS performed worse than the IMERG based on the POD values. This weakness increased in higher elevation classes. Also, in another comparison in different SA classes, the results showed that IMERG had better performance than the MRMS under moderate snowfall condition. However, MRMS estimates improved in heavy snowfall. In general, it was concluded that the IMERG performed better in snowfall detection than the MRMS; while in contrast, the opposite was true in estimating the SA.

Evaluation of the CloudSat surface snowfall product over Antarctica using ground-based precipitation radars

Abstract. In situ observations of snowfall over the Antarctic Ice Sheet are scarce. Currently, continent-wide assessments of snowfall are limited to information from the Cloud Profiling Radar on board the CloudSat satellite, which has not been evaluated up to now. In this study, snowfall derived from CloudSat is evaluated using three ground-based vertically profiling 24 GHz precipitation radars (Micro Rain Radars: MRRs). Firstly, using the MRR long-term measurement records, an assessment of the uncertainty caused by the low temporal sampling rate of CloudSat (one revisit per 2.1 to 4.5 days) is performed. The 10–90th-percentile temporal sampling uncertainty in the snowfall climatology varies between 30 % and 40 % depending on the latitudinal location and revisit time of CloudSat. Secondly, an evaluation of the snowfall climatology indicates that the CloudSat product, derived at a resolution of 1∘ latitude by 2∘ longitude, is able to accurately represent the snowfall climatology at the three MRR sites (biases < 15 %), outperforming ERA-Interim. For coarser and finer resolutions, the performance drops as a result of higher omission errors by CloudSat. Moreover, the CloudSat product does not perform well in simulating individual snowfall events. Since the difference between the MRRs and the CloudSat climatology are limited and the temporal uncertainty is lower than current Climate Model Intercomparison Project Phase 5 (CMIP5) snowfall variability, our results imply that the CloudSat product is valuable for climate model evaluation purposes.

Using the Airborne Snow Observatory to Assess Remotely Sensed Snowfall Products in the California Sierra Nevada

AbstractThe Airborne Snow Observatory (ASO) performed two acquisitions over two mountainous basins in California on 29 January and 3 March 2017, encompassing two atmospheric river events that brought heavy snowfall to the area. These surveys produced high‐resolution (50 m) maps of snow depth and snow water equivalent (SWE) that were used to estimate monthly areal snowfall accumulation. Comparison of ASO snow accumulation with point measurements showed that the ASO estimates ranged from −10 to +16% relative bias across three sites, which is likely inflated by the disagreement in areal representation of the quantities from the actual errors in these products. The aggregated SWE accumulations from ASO are then used to evaluate a suite of in situ based and remote sensing precipitation products. During the study period, Parameter‐Elevation Regressions on Independent Slopes Model (PRISM) and Mountain Mapper estimates had relative bias <10% compared with ASO‐based estimates of snow accumulation, but satellite and radar products largely underestimate snowfall accumulation compared to ASO (up to 50%). Despite their underestimation, satellite and radar products show correlation coefficients >0.8 with ASO snow accumulation over the selected grids at the monthly scale. Finally, we leveraged the fine‐scale sampling of the spatially complete ASO products to show that by moving from 100 m to 2 km spatial scales, the perceived bias errors SWE at point locations increased by an order of magnitude, displaying a nonlinear relationship. The study demonstrates that ASO acquisitions in cold months can bring a new and effective approach to spatial evaluation of precipitation products.

Seasonal variability of shallow cumuliform snowfall: A CloudSat perspective

Cumuliform snowfall seasonal variability is studied using a multi‐year CloudSat snowfall rate and cloud classification retrieval dataset. Microwave radiometer sea ice concentration datasets are also utilized to illustrate the intimate link between oceanic cumuliform snowfall production and decreased sea ice coverage. Three metrics are calculated to illustrate seasonal cumuliform snowfall signatures: (a) cumuliform snowfall frequency of occurrence, (b) mean cumuliform snowfall rate, and (c) fraction of snowfall attributed to cumuliform snowfall events. Distinct seasonal cumuliform snowfall cycles are observed over the Northern Hemispheric oceans. Cumuliform snowfall frequency of occurrence (mean snowfall rate) peaks in months SON (DJF) at most latitudes. Maximum mean cumuliform snowfall rates exceed 300 mm/year in various North Atlantic Ocean locations, with DJF exhibiting the largest areal extent of higher snowfall rates. Cumuliform snow occurrence fraction frequently exceeds 0.5, but regional seasonal sensitivity is observed where transient sea ice coverage exists. Annual snowfall rate fraction attributed to cumuliform snow does not vary appreciably during SON, DJF and MAM north of ∼70°N, but seasonal zonal variability is evident south of this latitudinal threshold. Land cumuliform snowfall features do not universally display strong seasonal signals. Southern Hemisphere seasonal results indicate a strong mean cumuliform snowfall rate maximum (minimum) in JJA (DJF) accompanied by a seasonal latitudinal shift in the snowfall rate peak. Maximum regional snowfall rates exceed 300 mm/year over a broader area compared to the Northern Hemisphere. Cumuliform snowfall production is again strongly linked to seasonal sea ice coverage. Southern Hemispheric cumuliform snowfall occurrence and snowfall rate fraction seasonality is not as obvious as in the Northern Hemisphere, but some latitudinal zones experience ∼5–20% seasonal variability in these quantities. Typical cumuliform snowfall fractions range from 0.4 to 0.6 in the prominent cumuliform snowfall belt covering most of the Southern Ocean. Some subtle seasonal cumuliform snowfall signatures are observed over Antarctica.

Validation of GMI Snowfall Observations by Using a Combination of Weather Radar and Surface Measurements

AbstractCurrently, there are several spaceborne microwave instruments suitable for the detection and quantitative estimation of snowfall. To test and improve retrieval snowfall algorithms, ground validation datasets that combine detailed characterization of snowfall microphysics and spatial precipitation measurements are required. To this endpoint, measurements of snow microphysics are combined with large-scale weather radar observations to generate such a dataset. The quantitative snowfall estimates are computed by applying event-specific relations between the equivalent reflectivity factor and snowfall rate to weather radar observations. The relations are derived using retrieved ice particle microphysical properties from observations that were carried out at the University of Helsinki research station in Hyytiälä, Finland, which is about 64 km east of the radar. For each event, the uncertainties of the estimate are also determined. The feasibility of using this type of data to validate spaceborne snowfall measurements and algorithms is demonstrated with the NASA GPM Microwave Imager (GMI) snowfall product. The detection skill and retrieved surface snowfall precipitation of the GPROF detection algorithm, versions V04A and V05A, are assessed over southern Finland. On the basis of the 26 studied overpasses, probability of detection (POD) is 0.90 for version V04A and 0.84 for version V05A, and corresponding false-alarm rates are 0.09 and 0.10, respectively. A clear dependence of detection skill on cloud echo top height is shown: POD increased from 0.8 to 0.99 (V04A) and from 0.61 to 0.94 (V05A) as the cloud echo top altitude increased from 2 to 5 km. Both versions underestimate the snowfall rate by factors of 6 (V04A) and 3 (V05A).

Evaluation of MRMS Snowfall Products over the Western United States

Abstract Snow is important to water resources and is of critical importance to society. Ground-weather-radar-based snowfall observations have been highly desirable for large-scale weather monitoring and water resources applications. This study conducts an evaluation of the Multi-Radar Multi-Sensor (MRMS) quantitative estimates of snow rate using the Snowpack Telemetry (SNOTEL) daily snow water equivalent (SWE) datasets. A detectability evaluation shows that MRMS is limited in detecting very light snow (daily snow accumulation <5 mm) because of the quality control module in MRMS filtering out weak signals (<5 dBZ). For daily snow accumulation greater than 10 mm, MRMS has good detectability. The quantitative comparisons reveal a bias of −77.37% between MRMS and SNOTEL. A majority of the underestimation bias occurs in relatively warm conditions with surface temperatures ranging from −10° to 0°C. A constant reflectivity–SWE intensity relationship does not capture the snow mass flux increase associated with denser snow particles at these relatively warm temperatures. There is no clear dependence of the bias on radar beam height. The findings in this study indicate that further improvement in radar snowfall products might occur by deriving appropriate reflectivity–SWE relationships considering the degree of riming and snowflake size.

Evaluation and Uncertainty Estimation of the Latest Radar and Satellite Snowfall Products Using SNOTEL Measurements over Mountainous Regions in Western United States

Snow contributes to regional and global water budgets, and is of critical importance to water resources management and our society. Along with advancement in remote sensing tools and techniques to retrieve snowfall, verification and refinement of these estimates need to be performed using ground-validation datasets. A comprehensive evaluation of the Multi-Radar/Multi-Sensor (MRMS) snowfall products and Integrated Multi-satellitE Retrievals for GPM (Global Precipitation Measurement) (IMERG) precipitation products is conducted using the Snow Telemetry (SNOTEL) daily precipitation and Snow Water Equivalent (SWE) datasets. Severe underestimations are found in both radar and satellite products. Comparisons are conducted as functions of air temperature, snowfall intensity, and radar beam height, in hopes of resolving the discrepancies between measurements by remote sensing and gauge, and finally developing better snowfall retrieval algorithms in the future.

Comparison of snowfall estimates from the NASA CloudSat Cloud Profiling Radar and NOAA/NSSL Multi-Radar Multi-Sensor System

The latest global snowfall product derived from the CloudSat Cloud Profiling Radar (2C-SNOW-PROFILE) is compared with NOAA/National Severe Storms Laboratory’s Multi-Radar Multi-Sensor (MRMS/Q3) system precipitation products from 2009 through 2010. The results show that: (1) Compared to Q3, CloudSat tends to observe more extremely light snowfall events (<0.2mm/h) and snowfall rate (SR) between 0.6 to 1mm/h, and detects less snowfall events with SR between 0.2–0.5mm/h. (2) CloudSat identifies 69.40% of snowfall events detected by Q3 as certain snow and 10% as certain mixed. When possible snow, possible mixed, and certain mixed precipitation categories are assumed to be snowfall events, CloudSat has a high snowfall POD (86.10%). (3) CloudSat shows less certain snow precipitation than Q3 by 26.13% with a low correlation coefficient (0.41) with Q3 and a high RMSE (0.6mm/h). (4) With Q3 as reference, CloudSat underestimates (overestimates) certain snowfall when the bin height of detected snowfall events are below (above) 3km, and generally overestimates light snowfall (<1mm/h) by 7.53%, and underestimates moderate snowfall (1–2.5mm/h) by 42.33% and heavy snowfall (⩾2.5mm/h) by 68.73%. (5) The bin heights of most (99.41%) CloudSat surface snowfall events are >1km high above the surface, whereas 76.41% of corresponding Q3 observations are low below 1km to the near ground surface. This analysis will provide helpful reference for CloudSat snowfall estimation algorithm developers and the Global Precipitation Measurement (GPM) snowfall product developers to understand and quantify the strengths and weaknesses of remote sensing techniques and precipitation estimation products.

Creation and Validation of a Comprehensive 1° by 1° Daily Gridded North American Dataset for 1900–2009: Snowfall

AbstractThis study details the creation of a gridded snowfall dataset for North America, with focus on the quality of the interpolated product. Daily snowfall amounts from National Weather Service Cooperative Observer Program stations and Meteorological Service of Canada surface stations are interpolated to 1° by 1° grids from 1900 to 2009 and examined for data record length and quality. The interpolation is validated spatially and temporally through the use of stratified sampling and k-fold cross-validation analyses. Interpolation errors average around 0.5 cm and range from less than 0.01 to greater than 2.5 cm. For most locations, this is within the measurement sensitivity. Grid cells with large variations in elevation experience higher errors and should be used with caution. A new gridded snowfall climatology is presented based on in situ observations that capture seasonal and interannual variability in monthly snowfall over most of the North American land area from 1949 to 2009. The Community Collaborative Rain, Hail and Snow Network is used as an independent set of point data that is compared to the gridded product. Errors are mainly in the form of the gridded data underestimating snowfall compared to the point data. The spatial extent, temporal length, and resolution of the dataset are unprecedented with regard to observational snowfall products and will present new opportunities for examining snowfall across North America.

SNOWFALL DETECTABILITY OF NASA'S CLOUDSAT: THE FIRST CROSS-INVESTIGATION OF ITS 2C-SNOW-PROFILE PRODUCT AND NATIONAL MULTI-SENSOR MOSAIC QPE (NMQ) SNOWFALL DATA

This study investigates snowfall detectability and snowfall rate estimation with NASA's CloudSat through the flrst evaluation of its newly released 2C-SNOW-PROFILE products using the National Mosaic and Multisensor QPE System (NMQ) snowfall products. The primary focus is on the detection and estimation of surface snowfall. The results show that the CloudSat product has good detectability of light snow (snow water equivalent less than 1mm/h) but degrades in moderate and heavy snow (heavier than 1mm/h). The analysis suggests that the new 2C-SNOW-PROFILE algorithm is insu-cient in correcting signal losses due to attenuation. Its underestimation is well correlated to snowfall intensity. Issues of sensitivity and data sampling with ground radars, which may afiect the interpretation of the results, are also discussed. This evaluation of the new 2C-SNOW-PROFILE algorithm provides guidance for applications of the product and identifles particular error sources that can be addressed in future versions of the CloudSat snowfall algorithm.

Snowfall in the northwest Iberian Peninsula: synoptic circulation patterns and their influence on snow day trends.

In recent decades, a decrease in snowfall attributed to the effects of global warming (among other causes) has become evident. However, it is reasonable to investigate meteorological causes for such decrease, by analyzing changes in synoptic scale patterns. On the Iberian Peninsula, the Castilla y León region in the northwest consists of a central plateau surrounded by mountain ranges. This creates snowfalls that are considered both an important water resource and a transportation risk. In this work, we develop a classification of synoptic situations that produced important snowfalls at observation stations in the major cities of Castilla y León from 1960 to 2011. We used principal component analysis (PCA) and cluster techniques to define four synoptic patterns conducive to snowfall in the region. Once we confirmed homogeneity of the series and serial correlation of the snowfallday records at the stations from 1960 to 2011, we carried out a Mann-Kendall test. The results show a negative trend at most stations, so there are a decreased number of snowfall days. Finally, variations in these meteorological variables were related to changes in the frequencies of snow events belonging to each synoptic pattern favorable for snowfall production at the observatory locations.