Hydrological Risk Transfer Model for droughts impacts on Agriculture in Southeastern Brazil

We simulated hydrologic multiyear lumped index insurance for safeguarding farmers from drought events in Southeastern Brazil. The model is divided into three modules: (i) hazard, (ii) vulnerability, and (iii) financial module. In the hazard module, we present 20 series of 50 years synthetic generated low-flows generated by Monte Carlo normal generated extreme events probabilities nested to a quantile Generalized Extreme Value (GEV) function. The 3-parameter GEV is estimated using MLE methods using observed daily streamflows. In the vulnerability module, using a water balance, which considered demands for drinking water supply, livestock and irrigation, we assessed potential economic losses on crops due to droughts. Water demands, economic indicators of crop production were obtained via freely available datasets from Govermental agencies. The data is aggregated by municipality. Finally, these losses were used in a cashflow equation to estimate optimized premium values in using different deductibles from 0 to 30% in a sensitivity analysis fashion.本研究针对巴西东南部的干旱灾害，模拟了用于保障农户生计的多年水文集总指数保险方案。该模型分为三大模块：（一）致灾因子模块、（二）脆弱性模块与（三）金融模块。在致灾因子模块中，我们依托嵌套于分位数广义极值（Generalized Extreme Value, GEV）函数的蒙特卡洛正态极端事件概率模型，生成了20组时长50年的合成枯水序列；并采用极大似然估计（Maximum Likelihood Estimation, MLE）方法，基于实测日径流数据对三参数GEV分布完成参数校准。在脆弱性模块中，我们通过考量饮用水供给、畜牧用水与灌溉用水需求的水量平衡模型，评估了干旱引发的作物潜在经济损失。作物生产需水量与经济指标均取自政府机构公开共享的数据集，所有数据均按行政县域完成聚合处理。最后，我们将上述经济损失代入现金流方程式，以敏感性分析的方式，针对0至30%区间内不同免赔比例设置，估算出优化后的保险费率值。