PMP估算中大气可降水量计算方法的探讨

干旱区水汽变化影响区域水资源系统的结构和演变，基于2020年1月—2022年12月中昆仑山北坡地区4个地基GPS遥感大气可降水量资料(GPS-PWV)、2个探空站观测资料和108个地面气象观测站逐时水汽压资料，利用一元线性拟合方法建立了适用于中昆仑山北坡地区的大气水汽含量(W-PWV)和地面水汽压计算模型(W-e)并对计算结果进行评估，分析了中昆仑山北坡地区东段、中段、西段W-PWV的时空分布特征及降水开始时刻与W-PWV峰值的关系。结果表明：(1) W-PWV年平均高值区位于研究区西段，中段次之，东段沙漠南缘W-PWV最低。海拔高度大于1 500 m测站W-PWV随高度升高逐渐减少。夏季地面气象观测站平均W-PWV是春、秋季的2倍左右；(2) 研究区W-PWV月变化具有单峰型特征，其中海拔高度1 300~1 500 m测站的W-PWV在7月和8月达到峰值，其余测站的W-PWV在8月达到峰值，海拔低于2 000 m和高于2 000 m测站W-PWV分别在夜间和白天维持较高值；(3) 水汽含量模型计算的测站W-PWV与降水开始时刻有较好的对应关系，降水前各站W-PWV均存在不同程度跃变过程，降水过程前1~2 h内W-PWV峰值达到测站W-PWV月平均值的1.5倍以上。

Comparison of calculation methods for estimating atmospheric precipitable water in estimation of PMP

The water vapor changes in arid areas could affect the structure and evolution of water resource systems in their surrounding areas. Based on the precipitable atmospheric water vapor (GPS-PWV) of 4 ground-based GPS stations, the observation data of 2 sounding stations and the hourly surface pressure water vapor data of 108 meteorological observation stations on the north slope of the Middle Kunlun Mountains from January 2020 to December 2022, this study established the atmospheric water vapor content and surface water vapor pressure (W-e) model suitable for the north slope of the Middle Kunlun Mountains using the unary linear fitting method. The results of water vapor content calculated by this model were verified. Then we analyzed the distribution characteristics of atmospheric water vapor content in the western section, the middle section, and the eastern section of the study area, as well as the relationship between the beginning time of precipitation and the W-PWV peak value. The results show that: (1) The annual mean W-PWV is largest in the western section of the study area, followed by the middle section, and the smallest in the eastern section which located in the southern edge of the desert. The W-PWV of the stations with altitude greater than 1 500 m gradually decrease with altitude increasing. The average W-PWV of each meteorological observation station in summer is about twice than that in spring and autumn. (2) The monthly variation of W-PWV shows a unimodal distribution characteristic. The W-PWV of the stations with an altitude higher than 1 300 m but lower than 1 500 m reached its peak in July and August, while that of the other stations reached its peak in August. The W-PWV of stations with an altitude below 2 000 m and above 2 000 m maintained a high value at night and during the day, respectively, which may be related to the thermal difference between mountain and basin from daytime to nighttime. (3) There is a good correspondence between the W-PWV calculated by W-e model and the beginning time of precipitation. Before precipitation, the W-PWV of each station is jumped varying degrees, and the peak value of W-PWV within 1-2 h before the precipitation is more than 1.5 times of the monthly average value of W-PWV.