Non-parametric statistical methods as a tool for rainfall monitoring

Abstract

Rainfall is an important role in the global energy and water cycle. Its monitoring is of special importance for the assessment and management of fresh water related to land use, agriculture and hydrology. As a result, the high interest in analyzes of long-term precipitation variability arises from the need to assess climate change. For this purpose, statistical methods have been used to monitor the behavior of climatic variables and to monitor possible trends. Given the above, this study aims to use non-parametric statistical methods, such as Mann-Kendall and Sen's Curvature and observe trends and intensities of rainfall changes in the Região Geográfica Intermediária de Londrina (RGIL) - Brazil. Daily, monthly and annual rainfall data from 12 meteorological stations located at RGIL were used and Mann-Kendall and Sen Curvature tests were applied using the Auto_MK_Sen software. The results indicated trends of increased rainfall in 5 of RGIL's 11 municipalities, all located to the north, in addition to a reduction in rainfall in the fall under a 90% statistical confidence level.