基于日气象资料、区域农业系统数据库和野外实地调查问卷，本文构建了干旱指数、农业旱灾恢复力指标体系、灾情指数、收入多样性指数，评价了三种农业典型区主要农作物干旱风险、县级尺度和乡镇尺度的农业旱灾恢复力、旱灾灾情变化情况、收入结构特点，归纳了三种农业不同的旱灾响应模式和恢复特点，并提出相应对策。结果表明：（1）灌溉农业干旱风险最大，其次为雨养农业和水田农业。兴和县年相对降水变率为18.5%，邢台县年相对降水变率为25.9%，鼎城区年相对降水变率为15.5%。雨养农业主要作物中春小麦在拔节期干旱风险最大（发生旱情的概率为88.64%，其中重旱概率为54.56%），马铃薯在定植期风险最大（发生旱情的概率为93.19%，其中重旱概率为65.91%）；灌溉农业冬小麦拔节抽穗灌浆成熟期均面临严重的干旱风险（发生重旱以上的概率均超过60%），其中拔节抽穗期干旱风险最大（发生旱情的概率为96.15%，其中重旱概率为86.54%），夏玉米在播种时节干旱风险最大（发生旱情的概率为65.38%，其中重旱概率为23.08%）；水田农业早稻整体受旱风险较小，在乳熟期风险最大（发生旱情的概率为21.57%，其中重旱概率为9.8%），晚稻在抽穗开花期风险最大（发生旱情的概率为31.37%，其中重旱概率为23.53%）。（2）农业旱灾恢复的核心是水资源，恢复能力包括内部自组织能力和外部输入能力；恢复形式包括自然恢复、生产恢复和生活恢复，恢复结果取决于作物状态和农户对旱灾恢复的依赖性；恢复过程包括灾中抵抗和灾后恢复，其中灾中抵抗主要制约因子是水，灾后恢复主要制约因子是非农收入。乡镇尺度上，地貌因素、水资源、农民收入结构是导致综合恢复力差异的主要因素。兴和县综合恢复力高值区位于中部平原区，其次是南部山区，低值区分布于北部丘陵区；邢台县东部平原恢复力值最高，其次是西部山区，低值区分布于中部丘陵；鼎城区南部山区和西北岗地属于恢复力高值区，其次是东北平原区，低值区分布于中部平原和南部山区无水库之地。全国尺度上，雨养农业主要依赖于灾后恢复过程中的外部组织能力；灌溉农业主要依赖于灾中抵抗过程中的生活恢复力；水田农业主要依赖于灾中抵抗过程中的生产恢复力。（3）不同农业类型对干旱的恢复响应导致不同的灾情。雨养农业干旱风险较大，但由于恢复力低，成灾率高（66.14%），旱灾年份成灾面积比例平均值达32.09%；灌溉农业干旱风险最大，虽然恢复力较高，但由于作物需水量大，成灾率高（69.73%），旱灾年份成灾面积比例平均值为17.93%；水田农业干旱风险最小，而且恢复力最高，成灾率低（28.83%），成灾面积比例平均值仅为2.87%。（4）不同农业类型农民收入结构存在差异：兴和县非农收入比例平均为68%，收入多样性指数平均为0.77；邢台县非农收入比例平均为44%，收入多样性指数平均为0.57；鼎城区非农收入比例平均为72%，收入多样性指数平均为0.65。因此，收入多样性和非农收入比例既与农民从事非农产业的意愿度有关，也与当地经济发达水平相关。本文在研究方法上实现了整合，采用统一的指标体系、模型、地图系列，使得区域之间和区域内部实现可比性。在此基础上，本文对尺度转换进行了探索，以三个典型区代表三种类型农业，通过对不同农业类型的比较，找出了南北方旱灾恢复的区域差异性和规律，实现了从县级尺度向全国尺度的转换。人与自然的适应是减灾的根本，基于本文分析所提出的农业旱灾对策可为区域抗旱资源配置和战略部署提供科学依据。

Based on daily weather data, regional agricultural system database, and farmer’s investigation questionnaires, the thesis built up drought hazard index, agricultural drought resilience index system, income diversity index, and disaster effect index, assessed the drought hazard risk of all major crops, agricultural drought resilience on scales of county and town, drought disaster effects, and characteristics of income structure, and concluded different response patterns to drought of three types of agriculture. The results show as follows.(1) Drought hazard risk of irrigated agriculture is highest, followed by the rain-fed agriculture and paddy field agriculture. The relative variability of annual precipitation in Xinghe County is 18.5%, while it is 25.9% in Xingtai County, and 15.5% in Dingcheng County. Among the major crops of the rain-fed agriculture, the spring wheat suffers highest drought hazard risk on the jointing stage (the total probability of drought is 88.64%, in which the probability of severe drought is 54.56%), while the potato suffers the highest risk on establishment stage (the total probability of drought is 93.19%, in which the probability of severe drought is 65.91%); For the irrigated agriculture, the winter wheat suffers highest drought hazard risk on the stage of jointing and heading is most severe (the total probability of drought is 96.15%, in which the probability of severe drought is 86.54%), while the summer corn only suffers relatively low risk on the sowing stage(the total probability of drought is 65.38%, in which the probability of severe drought is 23.08%); For the paddy field agriculture, the early rice suffer little risk of drought (the total probability of drought is 21.57%, in which the probability of severe drought is 9.8%), while the late rice mainly suffers the high risk on heading flowering stage (the total probability of drought is 31.37%, in which the probability of severe drought is 23.53%). (2) The core of agricultural drought resilience is the water resource, and the capability of recovery includes the inside self-organizational capability and outside input capability; the form of recovery includes natural recovery, production recovery, and life recovery, while the result of recovery depends on the state of crop and farmer’s rely on drought recovery; the process of recovery includes the in-disaster resistance and post-disaster recovery, in which the critical constricting factor of the former is water, while the one of the later is non-agricultural income.On the county level, topography, water resource, and farmer’s income structure are main factors of integrated resilience. In Xinghe County, regions of highest value are distributed at towns of the central plain, followed by twons in the southern mountainous area, and northern hilly area. In Xingtai County, regions of highest value are distributed at towns of the eastern plain, followed by the western mountainous area, and central hilly area. In Dingcheng county, regions of highest value are distributed at towns of the southern mountains and northwest hilly areas, followed by northeast plain, while towns at the central plain and mountainous area without reservoir are low value places,On the country level, the rain-fed agriculture mainly relies on the external organizational capability on the process of post-disater recovery; the irrigated agriculture mainly relies on the life resilience on the process of anti-disaster; the paddy field agriculture mainly relies on the production resilience on the process of anti-disaster.(3) Different drought disaster effects are caused due to responses of different types of agriculture to drought hazard risk. For the rain-fed agriculture, the drought hazard risk is relatively high, but the resilience is lowest, with the disaster rate of 66.14%, so the average ratio of drought disaster area reaches 32.09%. For the irrigated agriculture, the drought hazard risk is highest, and the water requirement of crops is large, but the resilience is relatively high, with the disaster rate of 69.73%, so the average ratio of drought disaster area reaches 17.93%. For the paddy field agriculture, the drought hazard risk is lowest, but the resilience is highest, with the disaster rate of 28.83%, so the average ratio of drought disaster area is only 2.87%. (4) The structure of farmer’s income is different in three areas. The average non-agricultural income ratio of Xinghe County is 68%, with average income diversity index of 0.77; the average non-agricultural income ratio of Xinghe County is 44%, with average income diversity index of 0.57; the average non-agricultural income ratio of Xinghe County is 72%, with average income diversity index of 0.65. The income diversity and non-agricultural income ratio are relative to both the farmer’s desire of recovery and the advanced level of local economy.The uniformed index system, model, and map series used in the thesis realized the comparison among regions and inside regions. The thesis chose three typical areas of three types of agriculture, found out the difference and regularity of agricultural recovery in the south and north, which realizes the transform of scale from county to country. Finally, the adaptation countermeasure against drought will provide scientific reference for resource collocation and strategic decision for regional drought mitigation.