The loss rate of agricultural drought was calculated, based on the data of crops area covered by drought disaster, affected by drought disaster, and no harvest by drought disaster, and crops yield data as well. The best probability model of drought loss rate was established, according to the goodness of fit test result. The agricultural drought disaster risk in Gansu province was assessed referring to value at risk (VaR) from economics and different levels of drought risk. The results showed that the effect of drought on agricultural production in Gansu province from 1950 to 2011 was relative limited, and the loss rate of agricultural drought was no more than 30% , with an average of 10% . The max loss rate occurred in 1995, with the rate of 26. 8% . The best probability model of drought loss rate was generalized extreme value (GEV) model in Gansu province in last 62 years. The agricultural loss rate was 18. 8% if facing drought once of ten years, 25. 7% facing drought once of 50 years, and 28. 3% facing drought once of 100 years.%利用甘肃省农作物旱灾受灾面积、成灾面积、绝收面积、播种面积及单位面积产量数据,计算农业旱灾损失率,根据优度检验结果拟合旱灾损失率最优概率分布模型,并借鉴经济学风险价值(VaR)方法,强调在统计意义下不同等级旱灾的风险水平,实现对甘肃省农业旱灾风险的有效度量。结果表明：1950－2011年旱灾对甘肃省农业生产的影响相对有限,农业旱灾损失率均在30％以下,平均10％左右,损失率的年际变化呈增长趋势,1995年因旱致灾损失率最大,达26.8％；近62a 甘肃省农业旱灾损失率的最优概率分布模型为广义极值(Gen. Extreme Value,GEV)分布模型；全省面临10a 一遇的旱灾时农业损失率为18.8％,遭遇50a 一遇的旱灾时农业损失率为25.7％,遭遇100a 一遇的极端旱灾时农业损失率高达28.3％,即全省农业产量或粮食产量将面临减少近30％的风险,这将给甘肃省粮食生产带来严峻考验。
展开▼
