Mudanças climáticas, seus impactos na cultura da mandioca no Semiárido Brasileiro e estratégias de manejo para mitigação das perdas

Abstract

Climate change, its impacts on cassava production in Brazilian Semi-arid region and management strategies for mitigation of losses Cassava is one of the most important crops in tropical countries. Due to its tolerance to adverse weather and soil conditions, it has been cultivated as a strategic crop in semi-arid regions. With the predictions of future climate changes, it is imperative to understand in which ways cassava production might be affected in these regions, in order to anticipate crucial actions to prevent and/or attenuate possible impacts and prepare the population to deal with that. In this context, crop simulation models are effective tools to quantify the future climate effect on crop yields. The objectives of this study were: i) to calibrate and evaluate the cassava simulation models DSSAT CSM-CROPSIM-Cassava and CSM-MANIHOT-Cassava to simulate the yields of the BRS Formosa cassava cultivar and analyze the sensitivity of these models; ii) to apply the DSSAT CSM-CROPSIM-Cassava model to determine cassava yields in the Brazilian Semi-arid region (SAB), under actual (1980-2010) and future climate projections, in medium (2040-2070) and long (2070-2100) terms, for intermediate (RCP4.5) and high (RCP8.5) greenhouse gases emissions (GHGs) scenarios, and to elaborate a vulnerability index (IV) for cassava production in the region; iii) to apply the DSSAT CSM-CROPSIM-Cassava model to evaluate potential management strategies, associated to planting dates and irrigation, in order to minimize possible climate impacts, in four locations of SAB. The DSSAT CSM-CROPSIM-Cassava and CSM-MANIHOT-Cassava models presented good performance for simulating Cassava BRS Formosa yields, with mean absolute error (MAE) of 1193 and 1315 kg ha-1 , and c index of 0,87 and 0,81, respectively. The ensemble had a better performance than CSM-MANIHOT-Cassava, but worse than CSM-CROPSIM-Cassava, presenting an intermediary MAE, of 1239 kg ha-1 , and c index = 0,85. The sensitivity analysis showed that both models are sensitive to variations in mean air temperature and [CO2], in a direct and inverse way, respectively. As for the rainfall, CSM-MANIHOT-Cassava presented low sensitivity, especially for the driest region, whereas CSM-CROPSIM was very responsive. In general, the CSM-CROPSIM model presented higher yield deviations in relation to the variations of air temperature, rainfall and [CO2] than CSM-MANIHOT-Cassava model, being more suitable for evaluation of climate change impacts on cassava yield. The cassava yields were reduced in all future scenarios, reaching-29% for potential yield (PP) and-22% for attainable yield (PA), at the most pessimist one. The PA maps showed that the areas of the regions with lowest yields at the actual scenario (states of PI, CE, PE and BA) tend to remain unchanged at future scenarios, while the ones with the highest yields (Southeast of BA and North of MG) tend to be reduced. The vulnerability of cassava production in SAB to climate change was greater at long term (2070-2100), in both GHGs emission scena...