变化环境下农业干旱响应机理及智能预测方法研究

Agricultural drought disaster is the result of multi-factors coupling,and it is impacted by hydrology, meteorlogoly, resources,environment. The series of hydrology and meteorlogoly have variations on the effects of global climate change and human activities, underlaying surface have changed severely, the intensity of agricultural drought disaster has increased, and the formation mechanism and catastrophe law shows the new features in the condition of global climate change and human interference. Studying it in-depth has important scientific significance and value. The research object is main grain production area in Henan province. We will research what are the internal and external causes of leading drought disaster, what variation features and changing laws the factors have, and how about the contribution rate in statistical method and catastrophe theory. Agiricultural drought prediction model in mid long term will be established on chaos theory and support vector achine method, taking into account the time series and physical causes. The recovery concept of crop suffering drought will be proposed.The ecovery ability and drought resistance of winter wheat and summer maize will be reaearched by field experiment. The critical value will be determined. On this basis, we will build agricultural drought real-time dynamic Wisdom prediction model. At the same time, winter wheat and summer maize yield reduction prediction model because of drought in different growth stages and whole growth period will be builded with improved Jensen model.

农业干旱灾害是多因素耦合作用的结果。在气候变化和人类干扰下，水文、气象序列发生了变异，下垫面发生了剧烈变化，农业干旱灾害发生频率和强度增大，致灾机理和灾变规律呈现新特征，对其进行深入研究具有重要的科学意义和应用价值。以河南粮食主产区为对象，采用先进的理论和方法，从自然和人工角度研究农业干旱灾害形成的内因和外因构成、致灾要素时空关联性、变异特征、变化规律和致灾贡献率，回答变化环境如何影响农业干旱灾害发生；提出农作物受旱恢复性概念，通过田间试验研究冬小麦、夏玉米受旱恢复力、抗旱强度，确定其耐旱力和受旱恢复力临界阈值，揭示农作物对变化环境的响应机理。提出农业干旱智能型预测概念，从现代网络信息技术平台的深度应用、农业干旱的物理成因及对干旱贡献率的量化表达、预测过程实时动态的智能化应用和农作物各生长阶段及全生育期因旱粮食减产损失的精细化预测等方面建立农业干旱中长期、实时动态、减产损失智能型预测模型。

干旱是一个反复出现的极端气候现象，对全球各地区的生态环境、农业生产、水资源供给、社会经济发展等均造成了不同程度的危害。农业干旱响应机理及智能预测技术作为减轻农业灾害损失与影响的重要途径，是干旱研究的薄弱环节之一。因此，本项目基于田间试验、模型、统计分析等方法分析了农业干旱时空演变规律及其成因，揭示了作物关键生育期干旱胁迫和复水的响应机理，构建了农业干旱、旱灾预测模型。主要研究成果包括：（1）分析了典型区干旱的年代际演变特点和区域性特点，揭示了干旱的时空演变规律，并从土壤与地形条件、有效降水与农作物需水间的关系等方面阐述了农业干旱的成因；（2）分析了不同生育阶段不同水分胁迫程度及复水对冬小麦和夏玉米生长和生理的影响，明确了不同干旱逆境下生理响应的水分阈值，揭示了其关键生育期干旱胁迫和复水的响应机理；（3）构建了基于Palmer模式的农业干旱预测模型，通过建立水文帐、分析气候系数与气候特征值，模拟了典型区历史农业干旱状况，并选取不同水文年（丰水年、平水年、偏枯年、枯水年）对典型区农业干旱状况进行预测；（4）基于旱情评估、灾情评估、旱地/水浇地土壤水分预测、水田水量预测等构建了农业旱灾预测模型，通过分析作物种植与生长参数、作物耗水量、有效降水量等对典型区农业旱灾进行预测，并验证了预测结果的合理性。研究结果可为指导防旱减灾措施与合理规划利用灌溉水资源提供科学依据，对促进我国农业可持续发展，保障国家水安全、粮食安全具有重要意义。