复杂地形下C波段雷达定量降水估计算法

Quantitative precipitation estimation algorithm for C-band radar situated in complex topographical regions

中文关键词:  天气雷达  定量降水估计  质量控制  降水类型分类

英文关键词:Weather radar  Quantitative precipitation estimation  Quality control  Precipitation type classification

中文摘要:

      C波段雷达定量降水估计（QPE）精度受到很多因素的影响，主要包括：（1）雷达标定，（2）非气象回波的干扰，（3）降水物垂直空间变化，（4）地形或地物的严重遮挡，（5）Z-R关系的代表性，（6）雷达拼图的质量，（7）雷达观测回波衰减等。文中雷达定量降水估计算法基于陕西省C波段天气雷达展开，从雷达探测数据质量控制、地形遮挡、Z-R关系和雷达拼图等方面提高C波段天气雷达定量降水估计的精度，产生降水类型产品和1 h定量降水估计产品，产品空间分辨率为0.01°×0.01°，时间分辨率为6 min。通过对7次降水过程进行评估，结果表明：基于混合仰角反射率因子处理模块和降水类型分类模块进行雷达定量降水估计，得到的结果与地面雨量站观测降水接近，1 h累计降水量的统计评分指标均方根误差稳定在3 mm以下，相对误差稳定在50%左右，相对偏差保持在-30%以内，雷达定量降水估计产品的离散度和绝对偏差都较低，表明该算法得到的雷达定量降水估计稳定可靠。

英文摘要:

      The accuracy of C-band radar quantitative precipitation estimation (QPE) is affected by many factors, including the following: (1) Calibration of radar observations, (2) interference from non-meteorological echoes, (3) vertical spatial variation of precipitation, (4) severe radar beam blockage by terrain or ground objects, (5) Z-R relationship, (6) multi-radar mosaic method, (7) radar observation echo attenuation, etc. In this paper, the radar quantitative precipitation estimation is conducted based on the C-band weather radar in Shaanxi Province. It improves the accuracy of the C-band weather radar QPE from the aspects of radar quality control, terrain blockage, Z-R relationship and multi-radar mosaic, and produces precipitation type and 1 h QPE products. The spatial and temporal resolutions of the product are 0.01°×0.01° and 6 min, respectively. Through the evaluation of seven precipitation processes, the results show that radar QPE based on hybrid scan reflectivity processing module and precipitation type classification module is close to the precipitation observed by the ground gauge stations. The statistical indicators RMSE of hourly precipitation is stable and below 3 mm, RMAE is about 50% and remains stable, and RMB remains within -30%. The dispersion and absolute deviation of radar QPE products are both low, indicating the radar QPE obtained by this algorithm is stable and reliable.