Estatística espacial aplicada à agricultura de precisão

Abstract

The methods provided by the spatial statistics are of great importance for studies involving data related to agriculture, for they allow one to know the space variability of the study and identify regions that have similar characteristics, which allows completely localized treatment, maximizing productivity and minimizing the impacts of excessive input application. One of the branches of spatial statistics is geostatistics, which uses a set of regionalized variables to model the structure of spatial dependence, allowing the preparation of thematic maps. Currently, geostatistical studies do not end with the preparation of maps, but also estimates monitored the attribute in non-sampled locations. It is necessary to investigate the quality of these maps, investigating influential points and using measurements to compare maps and area estimations. Another form of research is known as spatial statistics of areas where the objects of analysis are polygons representing blocks, neighborhoods, cities, states and others. This type of analysis seeks to identify spatial autocorrelation in global and local levels, and the usual form of reporting is through thematic maps. In this work we used geostatistics to investigate the productivity of wheat in an agricultural area of 13.7 hectares in the municipality of Salto do Lontra PR. Out of the 50 samples, two were identified as influential, and thus, we chose to build two thematic maps and to compare them using metrics derived from the matrix of errors. The results showed that the maps are different and the removal of influential points was essential to improve the quality of thematic map, since the difference between the estimated yield and actual yield was only 40 Kilos. In order to display the resources provided by the spatial statistics of areas we compared to the vegetation rates NDVI and GVI's of soybean yield from 36 cities in Western Paraná in the agricultural year of 2004/2005. The results showed regions with similar characteristics and that soybeans grow at different times in the region.