A Satellite View of an Intense Snowfall in Madrid (Spain): The Storm ‘Filomena’ in January 2021

• Published in: Remote sensing (Basel, Switzerland) Switzerland), 2021-07, Vol.13 (14), p.2702
• Main Authors: Tapiador, Francisco J., Villalba-Pradas, Anahí, Navarro, Andrés, Martín, Raúl, Merino, Andrés, García-Ortega, Eduardo, Sánchez, José Luis, Kim, Kwonil, Lee, Gyuwon
• Format: Article
• Language: English
• Subjects: Algorithms; COVID-19; Earth; Global Precipitation Measurement (GPM) mission; Integrated Multi-satellitE Retrievals for GPM (IMERG); Microwave sensors; Natural disasters; Neural networks; Observatories; Precipitation; Qualitative analysis; Radar; Radiometers; Remote sensing; Remote sensors; Satellites; Sensors; Sentinel; severe events; Severe weather; Snow; Snow cover; Snowfall; Snowstorms; Spacecraft recovery; Storms; Temporal resolution; Time response
• ISSN: 2072-4292; 2072-4292
• DOI: 10.3390/rs13142702

Evaluating satellite ability in capturing sudden natural disasters such as heavy snowstorms is a topic of societal interest. This paper presents a rapid qualitative analysis of an intense snowfall in Madrid using data from the Global Precipitation Measurement (GPM) mission, specifically the GPM IMERG (Integrated Multi-satellitE Retrievals for GPM) Late Precipitation L3 Half Hourly 0.1° × 0.1° V06 estimates of precipitation (IMERG-Late), and Sentinel-2 imagery. The main research question addressed is the consistency of ground observations, model outputs and satellite data, a topic of major interest for an appropriate and timely societal response to severe weather episodes. Indeed, the choice of the ‘Late’ product over the IMERG ‘Final’ or other GPM datasets was motivated by the availability of data for near real-time response to the storm. Additionally, the 30-min temporal resolution of the product would in principle allow for a detailed analysis of the dynamic processes involved in the snowstorm. Using several complementary data sources, it is shown that optical remote sensing sensors (Sentinel) add value to existing ground data and that is invaluable for rapid response to severe meteorological events such as Filomena. Regarding the GPM precipitation radar, the sampling of the GPM-core satellite was insufficient to provide the IMERG algorithm with enough quality data to correctly represent the actual sequence of precipitation. Without corrections, the total precipitation differs from observations by a factor of two. The difficulties of retrieving precipitation with radiometers over snow-covered surfaces was a major factor for the mismatch. Thus, the calibrated precipitation product did not fully capture the historic storm, and neither did the IR-based element of the IMERG-Late product, which is a neural network merging of microwave and infrared data. It follows that increased temporal resolution of spaceborne microwave sensors and improved retrieval of precipitation from radiometers are critical in order to provide a complete account of these sorts of extreme, significant, short-duration cases. Otherwise, the high-quality, radar and radiometer data feeding the high temporal resolution algorithms simply slip through the grasp of the ascending and descending orbits, leaving little quality data to be interpolated into successive overpasses.