Filling gaps in the historical series of precipitation and air temperature in northeastern brazilian states - DOI: 10.7127/rbai.v1901364

Resumo

The availability of historical climate series is essential for planning and decision-making in agricultural activities. This study aimed to compare the efficiency of four methods of filling gaps in a historical series of rainfall and air temperature data. Data from 12 automatic meteorological stations (AWS) in the state of Pernambuco and six in the state of Alagoas belonging to the National Institute of Meteorology (INMET) were excluded so that the filling techniques could be applied and evaluated. The database was composed of monthly values of accumulated rainfall and average air temperature for the period 01/01/2011 to 12/31/2020. The data absence simulation was standardized to two months in summer and two months in winter for each year and locality. The results indicated that Regional Weighting was the best technique for filling gaps in mean air temperature and rainfall data in the states of Alagoas and Pernambuco. The Multiple Linear Regression method presented the worst result in the state of Alagoas however, it was the second method with the lowest error rate in the state of Pernambuco.