Software aplicado a modelos de estimativa de produtividade agrícola

Abstract

Although there are studies that have been seeking modeling the influence of meteorological variables in relation to crop yield, the application of such models on a global scale and the consequent achievement concerning spatial results is a real challenge due to the complexity of the involved variables. Thus, this study aimed to analyze and implement software that allows the application of models to estimate agricultural yield based on remote sensing images, in a spatial and automated way. Yield estimation model using FAO water balance, as well as the needed indices to calculate it were implemented and subsequently tested for soybean cropping in Parana state for the 2011/2012 harvesting season. Design standards and analyses by language Unified Modeling Language (UML) were used to create a software structures and standardize computational tools. Such structures were implemented to data processing, water balance estimation and yield estimation, whose result was a version 1.0.1 software Crop-yield Modeling Platform - CyMP. Its structure makes implementation and use of estimation models more practical, automated, reusable and scalable for cropping yield. According to the proposed software, it was possible to reduce noise in vegetation indices from the orbital sensor images of Moderate Resolution Imaging Spectroradiometer - MODIS, interpolate images from the European Centre for Medium-Range Weather Forecasts ECMWF, climate model to MODIS resolution, determine FAO water balance, estimate dates of development cycle of soybean crop, estimate hydric stress factor, estimate real evapotranspiration, and finally estimate both gross potential yield and attainable yield of soybean crop in Paraná State for the 2011/2012 harvesting season.