【作者】 苟诗薇；

【导师】 许吟隆；

【作者基本信息】 中国农业科学院 ， 作物气象学， 2012， 博士

【摘要】 近年来，气候变化对农业产生了以弊为主的影响，而已有的风险研究仅从气象灾害或产量进行分析，更少有对未来风险变化的定量评估。本文尝试结合旱涝灾害与产量变化，以宁夏地区玉米生产为例，进行风险辨识和评估，在此基础之上评估未来的风险变化状况，为生产实践、决策制定提供一定科学依据和参考。本研究基于1988～2009年（历史时期）宁夏各县区的实际玉米单产，提出历年平均减产率、歉年出现频率和变异系数3个指标评价玉米生产的脆弱性，同时以气象台站的降水资料分析玉米生育期内（4～9月）的旱灾和涝害，评价旱涝灾害，在此基础上计算风险指数，并进行风险评价和分区。用同样的方法，以1961～1990年（气候基准时段）和2011～2040年（2020s时段）A2和B2情景下模拟的格点气象数据和玉米单产为基础，评价脆弱性、旱涝灾害和风险指数，比较气候基准时段模拟的风险与实际状况的差异，以及与未来两种情景下的风险差异，定量评估2020s时段A2和B2情景下宁夏地区玉米生产的风险较气候基准时段的变化状况。主要结论如下：1.基于历史时期实际气象数据和玉米单产分析的旱涝灾害和减产状况与文献记载的结果基本一致，即本文提出的评价方法适用于宁夏地区。PRECIS模式系统模拟的降水日值经剔除小于等于0.5mm的数值后分析的旱涝灾害与实际情况相符程度显著提高。2.历史时期、气候基准时段和2020s时段A2、B2两种情景下宁夏玉米生产的脆弱性分析结果表明，在历史时期和气候基准时段，脆弱性均在南部山区最高，北部灌区最低；而2020s时段的两种情景下，平均脆弱性指数均以中部干旱带最高、北部灌区最低，表明未来玉米生产的脆弱性中心由南部山区移至中部干旱带。3.历史时期和气候基准时段，玉米生育期内的干旱均在北部灌区最严重，南部山区最轻，多发于4～5月；涝害中心则分别为南部山区和中部干旱带，且均以灌区为少涝害，多发于8～9月；旱涝灾害指数呈现与涝害相似的分布。2020s时段两种情景下的旱、涝灾害发生格点增多；干旱以北部灌区最重、南部山区最轻，并在全区有加重的趋势；涝害则呈相反的分布状况，无明显加重或减轻的趋势；旱涝灾害指数均呈山区高、灌区低的分布，且在灌区和干旱带呈减轻、山区加重的趋势，同时旱涝灾害的高值中心由气候基准时段的中部干旱带移至南部山区。4.宁夏地区玉米生产的气候风险评价结果显示，历史时期宁南山区和中部干旱带的玉米生产风险高、灌区低；气候基准时段和2020s时段两种情景下的高风险区均位于中部干旱带，北部灌区风险低。以气候基准时段的风险划分标准对未来两种情景下的风险进行分区显示，B2情景下的高风险范围远大于A2情景。5.未来风险指数与气候基准时段比较显示，在A2情景下其程度减轻且在更多的格点上呈下降，3个区域的平均风险程度降低；而在B2情景下其程度增加且在更多的格点上呈上升，干旱带和山区风险增大、灌区降低。表明2020s时段B2情景下的气候状况不利于宁夏地区的玉米生产，旱涝灾害增多，生产脆弱性增加，风险也增大；而A2排放情景则是有利于宁夏的玉米生产。平均状况下，未来增温的变化对中部干旱带的玉米生产影响大，而对北部灌区的影响不明显。更多还原

【Abstract】 The impact of climate change on agriculture has been mainly adverse in recent years. There havebeen lots of climate change risk studies based on meteorological disasters and crop yield, but thequantitative assessment of climate change risk in the future is still insufficient. Considering droughtand waterlogging disasters and crop yield, taking maize in Ningxia province for example, this paperaims to identify and assess risk and its changes in the future so as to provide scientific evidence formaize production and policy decision in Ningxia.Three indicators of average yield reduction rate, lean year occurrence rate and coefficient ofvariance have been used to evaluate vulnerability of crop production based on the per unit area yieldof maize. At the same time, the data from observation stations were used into analyze drought andflooding happened during growth stages of maize. Based on the analysis, the climate change riskindex and risk zoning can be figured out. The same method is applied to assess risks in1961-1990(baseline) and2011-2040(2020s) based on grid data under A2and B2and the per unit area yield ofmaize. Its aim is to find out the difference of risk between baseline and actual situation, between twoscenarios of A2and B2and between baseline and two scenarios. The main conclusions are as follows:1. The result based on observation data and the per unit area yield of maize is consistent with thehistorical document, which proves the methods used in this paper is appropriate for Ningxia. Throughrejecting precipitation data which were less than or equal to0.5mm simulated from PRECIS model,the analysis of drought and waterlogging disasters and the actual situation are almost identical.2. The highest vulnerability of crop production in baseline appears in Ningxia southern mountainregion and the lowest vulnerability appears in Ningxia northern irrigation region. In contrast, thehighest one in2020s shows in Ningxia middle arid area and the lowest one still shows in Ningxianorthern irrigation region. The center of vulnerability of crop production in Ningxia moved fromsouthern mountain region to middle arid area.3. The worst meteorological droughts during baseline and2020s both happened in northernirrigation region during growth of maize. And the situation is better in southern mountain regionhappened in April and May. There were worse water loggings happened in southern mountain regionand middle arid area in August and September. The distribution of drought and waterlogging disastersindex is similar with it of flood. In condition, the number of grid happened drought and flood isincreasing under two scenarios. The worst drought will appear in northern irrigation region and themilder drought will show in southern mountain region. But the situation of flood shows the contrary.Drought and waterlogging disasters index is high in mountain areas and low in irrigation areas. What’smore, the drought will lighten in irrigation region and middle arid area and aggravate in mountainareas. The high value center will move from middle arid area in baseline to southern mountain regionin the future. 4. The conclusion of climate change risk of maize production in Ningxia shows that the risk ishigh in southern mountain region and middle arid area in the past and low in northern irrigationregion. The high risk in both baseline and2020s under two scenarios is in middle arid area and thelow one appears in northern irrigation region. Based on the criteria for risk distinction in baselineunder two scenarios, the risk scale under B2is larger than it under A2.5. The risk index in the future compared with which in baseline showed that, under the A2scenario the extent of risk would be mitigated and the value of risk index would be reduced on moregrids, while average level of risk would be decreased in three regions. Under the B2scenario, theextent of risk would be enhanced and the value of risk index would be increased on more grids, whilelevel of risks would be decreased in irrigation area, but increased in arid zones and mountainous areas.Which means that the climatic conditions in the2020s period under the B2scenario is not conduciveto the maize production in Ningxia, drought and waterlogging disasters, vulnerability and risk wouldbe increased. But the climatic condition under the A2emission scenario is in favor of maizeproduction. At average conditions, changes of the future warming would influence much more tomaize production in the central arid zone, while there would be no significant effects on the northernirrigation area.更多还原