Deriving extreme precipitation quantiles in Italian watersheds using a WebGIS tool

The outcome of this thesis is a quantitative spatial and statistical system for analyzing extreme precipitation data inputs that are essential for modelling and assessing extreme and potentially dangerous conditions. Heavy rainfall is among the major drivers of natural disasters such as flash floods and landslides. Spatial information about extreme rainfall amounts associated with a given probability of exceedance at the spatial and temporal scales relevant for a target watershed is a key element for disaster risk reduction. This thesis introduces an automated workflow, based on a WebGIS tool, that allows for the connection of web-based hydrology layers with statistical calculations to efficiently derive precipitation metrics of interest for risk management of the delineated watersheds. The resulting application, the Watershed Explorer, integrates an ESRI-hosted hydrological delineation service with a custom, client-side computation process that extracts rainfall data and aggregates them into precipitation quantile estimates – derived from the following statistical approaches – Generalized Extreme Value (GEV), Metastatistical Extreme Value (MEV – point and native scale), and Simplified Metastatistical Extreme Value (SMEV) distributions – for any user-chosen watershed, across a range of rainfall durations and return periods. Through the automation of the connection between watershed delineation and extreme precipitation statistics, the application avoids the demand for manual GIS operations or computational analysis, making the targeted precipitation estimates readily accessible. The application efficiently produces precipitation outputs across a wide range of Italian catchment areas and parametric configurations, capturing meaningful variations in rainfall magnitude associated with the spatial scale and statistical method.