Verification of precipitation type measurements in an inner valley

Abstract

This study examines various methodologies for classifying precipitation types at both surface and upper levels in La Cerdanya valley through the implementation of a fuzzy verification framework. By comparing on-site observations from Micro Rain Radars (MRR2, MRR-Pro) with the operational product Rain or Snow (ROS), distinct performances are revealed. At the surface level, ROS displays a high Probability of Detection (POD) for precipitation but also a high False Alarm Ratio (FAR) when compared to MRR2. ROS detects some events that MRR2 misses, which contributes to the high FAR, and is likely due to its ability to identify light and scattered precipitation across the valley while MRRs are only able to detect it above their position. When compared to MRR-Pro, ROS exhibits a lower FAR but also a relatively low POD, suggesting that MRR-Pro captures finer transitional details that ROS is not able to. In terms of surface-type classification, both MRR2 and ROS identify rain and sporadic mixed and snow events evenly, while MRR-Pro observes a wider distribution of all types. Overall, ROS and MRR2 exhibit comparable performance in discriminating precipitation types at surface level within the valley, unlike MRR-Pro. In addition, ROS’s estimates of height profiles using a digital elevation model show consistency with the observations from the MRRs, accurately indicating descending snow levels in several case studies.