Application of Markov chain on daily rainfall data in Paraíba-Brazil from 1995-2015

Abstract

This study analyzed the behavior of daily rainfall in the State of Paraíba using the data from five meteorological stations distributed across the mesoregions of this state. We used the three-state Markov Chain model, in which states are defined as dry, wet and rainy. We calculated transition probabilities among states, probabilities of equilibrium of states, and expected lengths of the defined states for all stations and seasons to investigate spatial/seasonal variability. Results showed that for the entire region and for all seasons, the probability of dry days is greater than the probability of rainy days; expected values of rainy spells are low, indicating that the rainfall regime in Paraíba is characterized by high rainfall intensity distributed over short rainy periods. The dry-dry transition probability presents the highest values for all seasons and stations, as well as the corresponding expected dry spell length, indicating that this region is subjected to prolonged dry periods. The transition probabilities that lead to dry condition are higher in the interior of the State, while probabilities that lead to rainy condition are higher in the coastal region as well as the probability of rainy days, which is greater in fall, during the rainy season.

Cumulative distribution function of daily rainfall in the Tocantins–Araguaia hydrographic region, Amazon, Brazil

“…Rainfall is variable in space and time; thus, it is important to recognize its occurrence patterns for a good prediction of the climatic behavior of a region (Santos, Blanco, & Oliveira Junior, 2019). Sectors such as agriculture, flood control, and human supply projects need knowledge of precipitation for planning, management and operation (Jale et al, 2019; Soares, Paz, & Piccilli, 2016). Stochastic models using cumulative distribution functions (CDFs) allow the simulation of hydrological data by means of their frequency of occurrence.…”