Evaporação no semiárido: análise de sensibilidade e comparação de 7 métodos para estimativa

Abstract

Evaporation is a key variable in understanding the water availability of an environment. Above all, in semi-arid regions, knowledge of water losses is essential and can support the development of activities in these environments. Several models have been proposed to estimate evaporation, these depend on the availability of data, time scale and conditions for which they were developed. This study sought to analyze the sensitivity of the class "A" tank and Penman methods to meteorological variables and compared 7 methods for estimating evaporation for a location in the Brazilian semiarid region. The database used was obtained from a meteorological station located in the municipality of Juazeiro-BA, and comprised the period from 1969-1999. The Penman method was considered standard and correlated with the meteorological variables: temperature (T), humidity (UR), radiation (Rg), wind (Vv) and vapor pressure deficit (DPV), through regression analysis for understanding of the relationships between them. The same procedure was performed for the evaporation measurements obtained through the Class “A” Tank (TCA). Additionally, the information was used to calculate the variable through alternative methods: Stephens–Stewart, Priestley–Taylor, Linacre, Makkink, TCA, Jensen-Haise and Camargo. This information was compared using indices for which Pearson's correlation coefficient (r), coefficient of determination (r²) and Willmont's concordance index are considered. Based on the results obtained, it was possible to verify that the variable DPV presented a high correlation with the evaporation estimated with the Penman and TCA methods. Among the alternative methods, the one that came closest to the method considered standard was Linacre's. Stephens–Stewart, Makkink and Camargo did not show good adequacy in the analyzed region.