Análise do impacto da previsão do tempo corrigida pelo método Model Output Calibration em modelos de doenças da maça

Abstract

Simulation models generally serve as tools to assist the different areas in predicting results and, consequently, in decision making. However, the simulations often present results that do not reflect reality. Among the possible causes are the quality of the input data, which may contain inaccuracies when it comes to forecasts. In the case of models that simulate diseases in apple culture, one of the data sources is those generated by weather forecast models, which are subject to these inaccuracies. An improvement in the quality of these data could lead to a consequent improvement in the final outcome of the disease model simulations as it has been found that these weather forecast data can be improved or corrected by statistical methods. This work aimed at the use of some models of diseases of the apple crop applied to the city of Vacaria (RS) that use among their input data the numerical weather prediction (NWP) generated by the regional model Eta 15km of CPTEC/INPE. The analysis of the impact on the final result of the simulation using the original Eta model prediction data in relation to these data corrected by a statistical model was the main objective of the study. In order to perform such correction or calibration the statistical method of model output calibration (MOC) applied on temperature and relative humidity variables was used. The results application of the MOC presented a satisfactory correction of the forecast generated for the city of Vacaria. And in the evaluation of the results generated by the epidemiological models that predict the occurrence of some diseases in the apple crop, the results of such analysis have proved to be satisfactory, making it possible to assist in the decision making of the producers and researchers involved with this crop.