Up to now, all analysis of the distribution of water vapor over the Taklimakan desert area only depends on limited ground measurements and radio soundings setting mostly on the outer margin area. This paper establishes an approach to retrieve the water vapor over the desert at high temporal and spatial resolutions by the use of FY2C geostationary satellite split-window channels in cooperation with ground-based GPS water vapor measurement. Results show that the water vapor distribution over the Taklimakan desert is affected highly by topography and surface properties. The outer margin area has generally more water vapor than the inner area. Over the outer margin area, the western part has more water vapor than the eastern part, and the northern part has more than the southern part. The driest area lies to the south of Tazhong, east of Hotan River, and extended to the south boundary of the desert. Similar to elsewhere, water vapor over the desert area shows diurnal, monthly, seasonal and annual variations even at the driest inner area of the desert. In summer, the water vapor is transported from west to east over a long distance along the westerlies at a height between 700–400 hPa and with the average speed of 50 km h−1.

Based on the data of vapor pressure over the Tarim Basin from 1961 to 2007,the changes of surface vapor are analyzed by applying the 5 points third-order trend line,wavelet analysis and Mann-Kendall trend and abrupt change test.The results show that the trend of surface vapor over the Tarim Basin increase conformably,the average trend is 0.024,there are 3 high centers and 3 low centers that strip distribute from northeast to southwest in the basin;The variation and trend of yearly and seasonal surface vapor present the obvious inter-annual and inter-decadal differences,their minimum and increase beginning presented in the end of 1960s and middle of 1970s,the maximum in 1990s;the increase trend of surface vapor is the most significant in summer,then in autumn,and the minimum in spring.The wavelet analysis shows that the yearly and seasonal surface vapor presented obvious periods.The yearly obvious surface vapor periods are 8 a and 4 a;the periods of spring are 8-10 a and 4-6 a;summer 4-6 a,8-10 a and 16-20 a,autumn 4-6 a,8 a and 20 a;and winter 8-10 a.The abrupt changes of the yearly and seasonal surface vapor presented in 1986,the test shows that all the abrupt changes are significant(α=0.05).The variation trend and abrupt change of surface vapor in the Tarim Basin were similar to the change trends of westerly circle,Tibetan Plateau circle,Tarim Basin annual mean temperature,and the time when they took place abrupt changes.The correlation between climate change and yearly surface vapor is significant,the correlation between climate change and spring surface vapor is weak,the correlation between Tibetan Plateau circle and surface vapor over Tarim Basin is the strongest.

Monthly mean water vapor contents were computed using meteorological data from 5 radiosonde stations including Hotan,Kuqa,Ruoqiang,Kashi and Minfeng around the Tarim Basin China during 1976-2006,and a formula was established involving monthly mean water vapor content and surface vapor pressure.By this formula,we obtained water vapor contents of all the 28 weather stations around the Tarim Basin and also performed EOF of water vapor contents to study its spatial pattern.Changes of vapor pressure with the height were also analysed.The water vapor contents obtained by GPS in the center and surroundings of the Tarim Basin were compared with radiosonde reports.Results show that:there are two higher water vapor areas in the Tarim Basin,one is on the west edge and the other is on the north edge,and the contents are both between 13-14 mm.These two higher centers are both in the oases near the Tarim River,Yarkant River and Aksu River.The center of the Tarim River has the lower water vapor,with the content being only between 7-8 mm,and the lowest spot is in Tazhong station.The water vapor content presents an increasing trend from the basin center to the edges,then presents a decreasing trend due to higher altitudes outside the Basin.There is good linear relationship between water vapor contents obtained from GPS and radiosonde.

气象灾害损失与风险大小取决于气象致灾因子危险性、承灾体脆弱性、自然与人为防控在孕灾环境中时空配置格局及交互作用.但对于一定区域与时段而言,后两个因素相对稳定,气象致灾因子多变,其不同时空分布格局很大程度上决定了灾害的地域性及时间变化特征.对致灾因子危险性予以准确诊断是客观评估气象灾害损失与风险大小的基本前提.为此,文中提出了气象致灾因子危险度定义及点面相结合的诊断模型:(1)将危险度定义为事件致灾因子量值与风险阈值场中各级风险水平阈值之间的接近程度;(2)采用随机变量概率分布模型估计各地各种特定概率下的气象事件致灾因子量级,构建气象致灾因子风险阈值场;(3)联合空间相似和距离参量构建危险度诊断模型,以刻划事件致灾因子与各级风险阈值分布形态相似性及数值差异大小,据此计算事件致灾因子与风险阈值场中各级风险阈值的接近程度,以接近度最大为原则确定某过程致灾因子总体危险性水平等级.然后以上海地区风致灾因子危险性诊断为例,计算了上海各地不同风险水平下年最大风速阈值以及各地各级年最大风速的风险水平,构建了上海地区年最大风速的风险阈值场,结果表明:上海沿海地区的南汇、崇明、金山等地为年最大风速高值区,也是一定风险水平下的最大风速高值区,同时又是8级以上强风频发区及高危险区;相对地,本市较内陆的区域,则是年最大风速低值区,也是一定风险水平下的最大风速低值区,同时又是8级以上强风稀遇区及低危险区;一定重现期下最大风速阈值地区分布也有类似规律.最后,应用该模型对影响上海地区热带气旋及其他天气过程共30余个例作出风危险度诊断,结果表明,以1977年9月11日的7708号热带气旋风危险度最高,总体上与风险水平为8年一遇的年最大风速阈值最为接近;1986年8月27日8615号热带气旋与1983年6月3日其他天气过程个例风危险度为第2,总体接近于7年一遇年最大风速阈值;8114号与9711号热带气旋风危险度则与4年一遇年最大风速阈值最为接近;7413号、7503号、7909号、8506号热带气旋的风危险度接近3年一遇年最大风速阈值;而0509号热带气旋"麦莎"、0515号热带气旋"卡努"风危险度总体上接近于2年一遇的年最大风速阈值.其他热带气旋影响个例,其风危险度多数与重现期约2年一遇的年最大风速阈值接近.实际应用结果表明,所提出的这一点面结合的危险度的诊断方法,能较客观定量地评定气象致灾因子的危险性程度.