【作者】 张丹；

【导师】 周惠成；

【作者基本信息】 大连理工大学 ， 水文学及水资源， 2011， 博士

【摘要】 我国是一个水资源短缺的国家,随着社会经济的迅速发展和气候变迁,极端天气现象频繁发生,干旱成为威胁经济发展与农业生产的主要制约因素。因此,区域干旱预报和风险评价已变为人类关心的问题和研究重点,了解区域干旱发展规律、提前预知未来干旱演进趋势,了解区域干旱风险级别已成为抗旱减灾中最主要的研究内容之一。本文对干旱中长期预报和农业干旱风险等问题进行了研究,主要研究成果如下：(1)区域干旱演变规律分析本文归纳总结了干旱的定义、干旱的分类以及各种干旱类型的代表性指标,并根据SPI标准化降水指标能反映不同时间、空间尺度干旱的特点,将其作为干旱划分指标,以辽宁省朝阳地区为例,通过功率谱法,探索了朝阳地区的干旱周期和演变规律,最终得出朝阳地区的干旱呈现5年～10年的波动周期。(2)基于小波分解的干旱等级组合预测模型研究干旱最主要的影响因素是降水量,降水量本身呈现出趋势性、周期性和波动性的特征,基于小波理论较强的分解能力,首先应用小波理论对降水量的趋势性、周期性和波动性成分进行分离,然后采用不同的模型对各成分分别进行预测,将预测结果进行重构,作为最终的预测结果。对于较难预测的波动性成分,本文对影响因子进行了筛选,将筛选后的因子作为降水量波动分量预测的因子集合。与其它模型的对比结果表明,基于小波分解的降水量组合预测模型,降低了降水量值预测误差,提高了预测合格率。根据干旱分级标准,将降水量转换成干旱等级,实现了干旱等级的间接预测。(3)基于土壤水量平衡模型的土壤墒情模拟从年际、季节和垂向的角度,归纳总结了土壤墒情变化规律,得出土壤墒情不但与降水量的大小存在着密切的关系,还与降水的季节分布有关。针对土壤水量平衡模型中作物蒸散发量ET确定复杂的问题,应用BP神经网络对其进行了模拟,并将模拟结果应用于土壤墒情的模拟。结果表明,其模拟结果达到了与基于Penman-Monteith公式计算得到的ET对土壤墒情的模拟结果相近的水平。故在资料不完备情况下,将BP神经网络拟合得到的ET输入到土壤水量平衡模型中,进行土壤墒情预报的方法可行。(4)基于GFS降水预报信息的土壤墒情预报研究土壤墒情是衡量农业干旱较为重要的指标,若能提前预报土壤墒情,将为区域抗旱减灾提供更可靠的依据。本文基于GFS全球降水中期数值预报系统对无雨和少雨预报精度较高的特点,将其预报的降水信息用于逐旬土壤墒情的定量预报；在抗旱部署时,决策者最关心的问题是干旱发展程度、何时得以缓解,本文根据国家旱情划分标准,将定量预报结果转化为相应的干旱等级。结果表明,定性的干旱预报结果合格率较高,而后将量级预报结果应用于全省2009年8月份的干旱等级预报中,也得到了较满意的结果。(5)基于可变模糊集合的农业干旱风险综合评价基于农业干旱风险存在模糊性的特点,综合考虑区域干旱风险的危险性、暴露性、区域脆弱性及抗旱减灾能力,将可变模糊集合应用于农业干旱风险综合评价中,综合数学权重和主观权重确定指标综合权重,以朝阳地区为例,对区域农业干旱风险进行综合评价,并将该方法应用于辽宁省14个市的农业干旱风险综合评价中,形成全省农业干旱风险区划图。结果表明,可变模糊集合的应用为区域农业干旱风险评价提供了新的思路,全省干旱风险图的生成为区域未来的农业干旱风险管理及抗旱减灾决策提供了有力的理论支持和可靠的技术保障。最后,对全文进行了总结,并对有待于进一步研究的问题进行了展望。更多还原

【Abstract】 China is a water shortage country. With the social and economic development and environmental changing, the extreme weather phenomenon in China is more and more frequent. Drought is the main restraining factors which affects the economic development and agriculture production. Consequently, regional drought forecasting and risk assessment have been the key to the researches. Knowning the regional drought developing rules and drought risk is the main contents of drought-resisting. This article studies on the drought long-term forecasting and risks assessment. The research results are listed as follows:(1) The regional drought development law analysisThis paper summarizes the drought definition, classification and the representative indices. According to the reflection different spaces or times distribution merits of SPI (Standardization Precipitation Index), it is used for drought development law analysis of Chaoyang City. Based on the power methods, it concludes that a five to ten years fluctuation drought cycle has been presented.(2) Drought grade combining precipitation model based on wavelet decomposition methodPrecipitation is showing a trendy, periodicity and volatility characteristics. Based on a strong decompose ability of wavelet theory, precipitation trendy, periodicity and volatility components are separated. Then use different prediction models to forecast the different components of precipitation, respectively, and take the predicted results of reconstruction of the three parts as the final predictions. For the difficult volatility characteristics of precipitation, the paper screened the circumfluence factors, and used these factors sets for precipitation model building. Comparison of results with the other models shows that the combination of wavelet decomposition precipitation precipitation model reduces the precipitation forecasted error and improved the forecast passing rate. Then, the precipitation is converted to drought grade according to national standard. And the study result shows that this method is feasible.(3) Soil moisture simulation based on soil-water balance modelThe annual and seasonal changing rules of soil moisture are aummarized. We found that the soil moisture is not only relates to the precipitation quantity, but also relates to the precipitation seasonal distributing characters. According to the complex calculations of crop water requirement in the soil water balance model, BP network is used for crop water requirement simulation. It shows that the BP simulation gets a similar result to the Penman-Monteith. The usage of ET based BP network for the soil moisture simulations will be feasible when the information is incomplete.(4) Soil moisture prediction based on precipitation prediction system of GFSSoil moisture is an important index for agricultural assessment. It will provide more dependable information if known soil moisture the before drought occurs. Based on the characteristics of higher prediction precision of no rain and light rain of global forecasting system, the precipitation forecasting information of GFS is used for soil moisture quantity forecasting. During making a detailed plan to fight the drought, policy-makers concern more about the drought degree and when the drought degree will be reduced. According to the national drought standard, the quantity forecasting result is translated into drought grade. It shows that the prediction passing percent is high. The study results are used for August drought prediction and the satisfying outcomes.(5) Agricultural comprehensive drought risk assessment based on variable fuzzy setsBased on the fuzzy characteristics of comprehensive drought risk in agriculture, variable fuzzy sets theory was used for agriculture comprehensive drought risk assessment of the fourteen cities in Liaoning province of China. Multi-layers and multi-indices variable were established using fuzzy evaluation model and estimated the agriculture drought risk in terms of dangerousness, vulnerability, exposure and drought-resistibility of the fourteen cities respectively. According to the combination weights of the four drought risks factors, agriculture comprehensive drought risk grade of each city was obtained. According to the assessment results, agriculture drought risk zoning map was drawn by the software of MapInfo. The assessment result shows that agriculture comprehensive drought risk of Liaoning province has strong regional features. The western region of Liaoning is a high-risk area for agricultural drought, the central region of Liaoning is moderately risky area, and the eastern region of Liaoning is a low risk area.Finally, the conclusions are drawn and the problems to be further studies are discussed.更多还原