南方双季稻低温灾害风险分析与预测预警研究

发布时间：2018-06-20 10:18

  本文选题：双季稻 + 低温灾害　； 参考：《中国气象科学研究院》2015年硕士论文

【摘要】：本文基于南方双季稻种植区708个气象站1961-2010年的逐日气象资料和双季稻低温灾害发生等级的气象行业标准,通过各气象站低温灾害发生次数倾向率的计算,揭示了近50年双季稻低温灾害发生趋势的地理分布特征。采用信息扩散法,综合考虑各等级低温灾害的发生概率和发生强度,构建了各气象站双季稻低温灾害综合风险指数,揭示了近50年双季稻低温灾害发生风险的地理分布特征。基于1960-2010年74项大气环流特征量资料、采用因子膨化、相关性分析法,建立了双季稻低温灾害高风险区、增加趋势区、低风险和减少趋势区历年第一次低温灾害发生等级的逐步回归预测模型。基于8项气象因子资料,采用Fisher判别分析法,建立了双季稻低温灾害高风险区早稻未来10天、晚稻未来5天的低温灾害等级逐日滚动预警模型。主要结论如下:(1)近50年南方双季稻低温灾害以减少趋势为主,增加趋势相对不明显。早稻春季低温灾害呈增加趋势的地区主要位于陕西种植区、四川种植区北部、云南种植区中部、广西北部和广东中北部等地,增加速率约在0-0.05次/10a;粳稻寒露风呈增加趋势的地区主要位于湖北东部、湖南中东部、江西西部、广西大部、广东和福建交界等地,增加速率基本为0-0.04次/10a;籼稻寒露风呈增加趋势的地区主要位于云南种植区西部和东南部,增加速率超过0.1次/10a。(2)近50年早稻春季低温灾害发生风险高值区主要位于湖南南部、江西南昌以南、浙江南部、福建大部以及广西和广东北部等地,风险指数超过0.15。粳稻和籼稻寒露风风险高值区主要位于陕西种植区、四川种植区北部和云南种植区大部,风险指数分别超过0.2和0.5。(3)基于74项大气环流特征量资料,建立了双季稻Ⅰ区(高风险区)、Ⅱ区(增加趋势区)、Ⅲ区(低风险和减少趋势区)历年第一次低温灾害发生等级的逐步回归预测模型。早稻Ⅰ区、Ⅱ区、Ⅲ区逐步回归预测模型平均外延预测基本一致正确率分别是100%、100%、83.3%。粳稻Ⅰ区、Ⅱ区、Ⅲ区逐步回归预测模型平均外延预测基本一致正确率分别是83.3%、83.3%、100%。籼稻Ⅰ区、Ⅱ区、Ⅲ区逐步回归预测模型平均外延预测基本一致正确率均是83.3%。(4)基于Fisher判别分析建立了双季稻低温灾害高风险区早稻未来10天、晚稻未来5天的低温灾害等级逐日滚动预警模型。早稻春季低温灾害高风险区平均回代基本一致正确率是95.5%,平均外延预测基本一致正确率是90.5%。粳稻寒露风高风险区平均回代基本一致正确率是83.1%,平均外延预测基本一致正确率是74.2%。籼稻寒露风高风险区平均回代基本一致正确率是79.5%,平均外延预测基本一致正确率是80.3%。
[Abstract]:Based on the daily meteorological data of 708 weather stations in southern double-cropping rice growing area from 1961 to 2010 and the meteorological industry standard for the occurrence of low-temperature disasters of double-cropping rice, the paper calculates the frequency tendency of low-temperature disasters in each meteorological station. The geographical distribution characteristics of low temperature disasters of double cropping rice in recent 50 years were revealed. By using the information diffusion method and considering the occurrence probability and intensity of the low-temperature disasters of different grades, the comprehensive risk index of low-temperature disasters of double-cropping rice in various weather stations was constructed, and the geographical distribution characteristics of the low-temperature disasters of double-cropping rice in the last 50 years were revealed. Based on the data of 74 general circulation characteristics from 1960 to 2010, the high risk area of low temperature disaster of double cropping rice was established by factor expansion and correlation analysis, and the trend area was increased. A stepwise regression prediction model for the first low temperature disaster occurrence grade in the low risk and reduction trend area. Based on the data of 8 meteorological factors, Fisher discriminant analysis was used to establish a daily rolling warning model of early rice in the next 10 days and late rice in 5 days in the high risk area of low temperature disaster of double cropping rice. The main conclusions are as follows: (1) in the last 50 years, the trend of low temperature disaster in Southern double cropping rice was mainly decreasing, but the increasing trend was not obvious. The areas with an increasing trend of early rice low temperature disasters in spring are mainly located in Shaanxi, northern Sichuan, central Yunnan, northern Guangxi and central Guangdong, etc. The increasing rate of japonica rice Cold Dew is about 0 to 0. 05 times / 10 a. The areas with increasing trend of Cold Dew wind are mainly located in eastern Hubei, central and eastern Hunan, western Jiangxi, most of Guangxi, the border of Guangdong and Fujian, etc. The increasing rate of Cold Dew in indica rice is mainly located in the west and southeast of Yunnan planting area, and the increase rate is more than 0.1 times / 10a.f.) in the last 50 years, the high risk area of spring low temperature disaster of early rice is mainly located in southern Hunan. South of Nanchang, south of Zhejiang, most of Fujian, and northern Guangxi and Guangdong, the risk index exceeds 0. 15. The high wind risk areas of japonica rice and indica rice were mainly located in Shaanxi province, northern Sichuan and Yunnan provinces, and the risk index exceeded 0.2 and 0.5.3 respectively) based on the data of 74 atmospheric circulation characteristics. A stepwise regression model was established for predicting the first occurrence grade of low temperature disasters in region I (high risk area), region 鈪，

本文编号：2043923