Open access gridded climate datasets can help to reveal meteorological anomalies’ role in landslides initiation 

Abstract

In mountainous areas, landslides initiation is strongly influenced by temperature and precipitation. Liking temperature and precipitation anomalies to landslides occurrence is among the viable methods to predict the future occurrence frequency of such hazards under climate change. Currently, most of the methods to detect such anomalies rely on in-situ measurements. This demands significant efforts for data retrieval and homogenization, which, together with the constraints on data sharing, pose important challenges to the replicability of the studies and to the comparison among different methods. Open access gridded datasets could overcome these limitations, but their ability to capture meteorological anomalies needs to be assessed. Here we address this issue. By means of a consolidated statistical-based approach, we here exploit: a) half-hourly precipitation estimates from the Integrated Multi-Satellite Retrievals from GPM (IMERG), b) daily temperature observations from ENSEMBLES OBServation (E-OBS) and c) daily temperature and total precipitation  of global reanalysis ERA5 to demonstrate that open access gridded climate datasets can complement or even replace in-situ data in studies linking meteorological anomalies (defined as percentiles above 0.9 or below 0.1) with the occurrence of geomorphic hazards. We focus on a vast catalogue of 483 different geomorphic hazards (mainly landslides, rockfalls and debris flows) occurred along 2000-2020 over the Italian Alps. Findings indicate that the statistical significance of the paired anomalies derived by observations and gridded datasets is often achieved. Mismatches are related to limited sample sizes. In general, E-OBS and IMERG demonstrate to provide information on temperature and precipitation anomalies, respectively, that is comparable or even better than the one provided by in-situ observations and ERA5 reanalyses. Additionally, our findings reveal that IMERG, by providing information directly on the initiation zone, can detect precipitation anomalies at the daily scale that in-situ measurements fail to detect, especially in the case of debris/mud flows events triggered by small-scale convective processes. Overall, gridded datasets can help to improve our knowledge on the statistical relation between landslide initiation and meteorological anomalies, which in the future can be adopted to quantify changes in the occurrence probability of these events under a changing climate.