Daily rainfall erosivity as an indicator for natural disasters: assessment in mountainous regions of southeastern Brazil

Abstract

Rainfall erosivity is defined as the rainfall potential to cause erosion. Its concept is based on the kinetic energy of rainfall, rainfall intensity, and maximum rainfall intensity in a 30-min period, and purposes recurrence analyzes involving soil losses. It is a climatic index related to damages caused by erosion, landslides, and flooding. This study sought to: (1) model daily rainfall erosivity in Mantiqueira Range Region (MRR), Southeastern Brazil; and (2) propose the Rmaxday as an indicator of the areas prone to natural disasters. Rmaxday is defined as the maximum daily rainfall erosivity and is determined from the maximum daily rainfall. It should be calculated from a historical series over the least 22 years. Three seasonal models were fitted using observed historical series. The models exhibited consistent statistical performances (CNS = 0.55, on average), thereby indicating that they can be used for further studies regarding natural disasters in the MRR. The Center to Northeastern MRR had the most vulnerable areas, as they experienced Rmaxday values > 1600 MJ ha−1 mm h−1 in the year of 2000 when fatalities were registered. Overall, the first half of January had the greatest Rmaxday in MRR. This period has been the most relevant for provoking natural disasters caused by heavy rainfall in MRR. Rmaxday is a promising indicator to identify areas prone to natural disasters, as it is more robust than the commonly used rainfall depth intervals, i.e., there is a relationship between its magnitude and the damages provoked by natural disasters.