A novel statistically-based approach to regionalize extreme precipitation events using temperature data

Abstract

Extreme precipitation events have been increasing and intensifying over the pastfew decades, posing challenges for modeling and prediction, as well as for policyand decision making. While traditional approaches often focus solely on study-ing the precipitation process, recent studies advocate for considering multipleprocesses and variables to better understand the drivers and anomalies of pre-cipitation. This is especially underexplored in South America. To address this,we propose a novel approach that combines time series modeling and quantileregression to estimate the extreme quantiles of precipitation based on maximumdaily temperatures. This methodology helps in understanding the relationships between these processes and contributes to identifying gauge stations with coher-ent climatic covariability, offering valuable insights into the regionalization ofextreme events.