Previsão de safra de arroz e introdução de um submodelo de brusone no modelo SimulArroz

Abstract

The objectives of this thesis were (i) to adapt a methodology for rice yield forecast in Rio Grande do Sul – Brazil; (ii). to use hourly hydro-thermal time (HTT) to assess interannual and regional weather variability for rice blast risk in southern Brazil. For the objective (i) it was used Morell’s et al. (2016) yield forecast methodology, adapted for rice using SimulArroz v1.1 rice model, actual and historic weather data. Six different scenarios considering different levels of field information were used, changing number of sowing dates (1 to 4), number of cycle length or cultivars (1 to 3) during four growing seasons (2015 to 2018). Root mean square error (RMSE) comparing actual yield versus simulated yield for Rio Grande do Sul ranged from 618.3 kg ha-1 (8%) to 1024.8 kg ha-1 (13%). The recommended scenario for rice yield forecast was Complex 1 that used 3 sowing dates and the 3 most representative rice cultivars (C1), presenting good forecast predictability (RMSE = 618.3 kg ha-1 or RMSE (%) = 8). For objective (ii) a large data collected from multiple locations and years in southern Brazil were used. For each year x site x cultivar combination, HTT was calculated using hourly data of air temperature, relative humidity and wind speed collected from nearby weather station. The HTT was correlated with blast onset to define a threshold for blast onset. The seasonal HTT between years x sites ranged from 5.1oC h-1 year-1 to 725.3oC h-1 year-1. Blast risk started after HTT of 33.6oC h-1, 66.8oC h-1 and 75.6oC h-1 from Jun 1st until rice emergence (EM) and after HTT of 12.5oC h-1, 55.3oC h-1 and 121.8oC h-1 after EM for susceptible, medium-resistant and resistant cultivar, respectively. Based on the results, it would possible to improve fungicide management using the HTT approach over the calendarization approach, once we would be able to discriminate the cultivar and year influence, the two most important factors on rice blast epidemics.