Geostatistical models in the assessment of temporal variability of reference evapotranspiration

Abstract

Reference Evapotranspiration (ET0) is a key hydrometeorological variable used in various applications, including integrated water resource management and land-use planning for agricultural and forestry projects. In much of Brazil, ET0 data is not readily available, necessitating alternative methods and techniques for its estimation. This study aimed to estimate ET0 using the FAO Penman-Monteith method, enabling visualization of its spatial behavior, statistical analysis, and evaluation of ET0's temporal variation using geostatistical methods like kriging and variograms in three locations within the State of Pará. The modeling indicated a strong temporal dependence for ET0, with most values ranging between 3.5 mm and 4.0 mm. The demand for ET0 showed similarities across locations, displaying seasonal patterns influenced by input variables affecting its estimation. The kriging method highlighted periods of both higher and lower ET0 values. The adjusted variograms showed an R2 > 0.86 and a sum of squares of residuals (SQR) close to zero.