Spatial heterogeneity in rainfall extreme trends across Santa Catarina, Brazil

Abstract Understanding spatial patterns of extreme precipitation trends is essential for climate risk assessment, particularly in regions characterized by strong climatic heterogeneity and limited monitoring networks, where locally significant trends may be overinterpreted if field significance is not explicitly assessed. This study presents a regional analysis of trends in ten extreme precipitation indices calculated for each of 62 rainfall stations across Santa Catarina, southern Brazil, resulting in a total of 620 daily time series analyzed over a 45-year period (1976–2021). Local trends were assessed using the Mann–Kendall test and Sen’s slope estimator, while field significance was evaluated through the False Discovery Rate (FDR) procedure to control for the multiple-testing problem and reduce false detections. Results indicate that most series exhibit no statistically significant trends; however, among the 24 field-significant signals retained after FDR correction, increasing trends were twice as frequent as decreasing ones. Statistically robust changes were more consistently detected in intensity-related indices, such as R99p and SDII, whereas frequency indices (R25 and R50) and total annual precipitation (PRCPTOT) largely lost significance after multiplicity control. The spatial distribution of these field-significant trends is heterogeneous, with increasing signals concentrated in the Western Mesoregion and predominantly decreasing trends observed in the Northern Mesoregion, while other regions exhibit mixed patterns. These findings demonstrate the importance of explicitly addressing multiple testing in regional trend analyses and highlight the value of long-term, spatially representative daily rainfall observations for distinguishing robust regional signals from statistical noise. The results provide a consistent empirical basis for climate risk assessment and adaptation planning in southern Brazil, while emphasizing the need for region-specific approaches in studies of precipitation extremes.