Evaluating homogeneity and trends in extreme daily precipitation indices in a semiarid region of Brazil

Abstract

A regional investigation of precipitation extremes could help the local authorities to develop strategies against potential climatic disasters, especially in regions of large climate vulnerability. In this context, this study focused on the homogeneity and trend analysis of eleven extreme precipitation indices from a daily rainfall dataset of rain gauges of the state of Ceará, located in the north of Northeast Brazil, from 1974 to 2018 at annual and seasonal time scales. The data were first submitted to gap-filling and quality control processes. Homogeneity, trend, and correlation coefficient were performed subsequently. Homogeneity results showed that most of the precipitation series were classified as “useful.” Significant break years in the series agreed well with moderate and very strong El Niño and La Niña events, suggesting a further investigation of this possible connection. Wet and dry day precipitation indices mainly indicated a decrease in the rainfall regime and an increase in dry days, mainly in the central-eastern, northwestern, and southern regions of the Ceará, especially during the annual time scale and the rainy season. The Locally Weighted Scatterplot Smoothing (LOWESS) curve showed changes in almost all series during the 1980s and 1990s, coinciding with the homogeneity breaks and years of severe droughts that strongly hit the region. Correlation coefficients were strong and significant between rainfall total index and the other precipitation indices.