Evaluation of geostatistics methods for interpolation of agrometeorological indices used to define climatic risks

Abstract

The definition of the best sowing dates for corn, adopted in the Zoning of Climatic Risks of the State of São Paulo, Brazil, was based on space-temporary variations of the water stress index (ISNA), using the pondered average as the interpolation method to spatialize this parameter. This method does not consider the spatial propagation of errors, leading to imprecise interpretations of the best sowing dates, mainly at the beginning and at the end of the cycle. The objective of this work was to compare methods of spatializing numeric values of agrometeorological indices and to evaluate their spatial space variation. The methods of the pondered means ordinary kriging and indication kriging were used. The indication kriging was the most suitable method for spatializing ISNA and to define the best sowing date for corn, in the State of São Paulo.