高级搜索
周永水, 原野, 杨林, 王瑾, 严小冬. 2022: 晴空时微波辐射计温度廓线反演误差分析. 暴雨灾害, 41(6): 720-726. DOI: 10.12406/byzh.2020-194
引用本文: 周永水, 原野, 杨林, 王瑾, 严小冬. 2022: 晴空时微波辐射计温度廓线反演误差分析. 暴雨灾害, 41(6): 720-726. DOI: 10.12406/byzh.2020-194
ZHOU Yongshui, YUAN Ye, YANG Lin, Wang Jin, YAN Xiaodong. 2022: Bias analysis of retrieval profile of temperature with microwave radiometer in clear sky. Torrential Rain and Disasters, 41(6): 720-726. DOI: 10.12406/byzh.2020-194
Citation: ZHOU Yongshui, YUAN Ye, YANG Lin, Wang Jin, YAN Xiaodong. 2022: Bias analysis of retrieval profile of temperature with microwave radiometer in clear sky. Torrential Rain and Disasters, 41(6): 720-726. DOI: 10.12406/byzh.2020-194

晴空时微波辐射计温度廓线反演误差分析

Bias analysis of retrieval profile of temperature with microwave radiometer in clear sky

  • 摘要: 利用2017年1月—2019年5月逐日08∶00(北京时,下同)、20∶00贵阳地基微波辐射计和无线电探空仪、地面自动站观测数据,评估了微波辐射计在晴空时温度廓线的反演误差及误差的标准差的季节变化以及逆温与误差的关系,结果显示:(1)从春季到冬季4个季节垂直温度廓线的误差分别为-1.71~1.35℃、-1.69~0.78℃、-2.24~1.29℃、-4.98~1.49℃,误差的标准差分别为1.13~2.93℃、0.43~1.86℃、0.76~1.92℃、0.62~3.48℃,春季到秋季的误差及误差的标准差在垂直方向的变化较为一致,冬季误差和误差的标准差均大于其它季节。(2)逆温与误差有较好的相关性。在逆温的3种表达方式中,逆温层厚度为490~740 m时,逆温层厚度与误差有最优的相关系数,最大相关系数出现在逆温层厚度为650 m时;逆温层厚度小于490 m或大于740 m且逆温差阈值为1.2℃时与误差有最优的相关系数。(3)逆温平均高度和误差平均高度的拟合度较差,不能由逆温发生的高度推测误差所在高度。

     

    Abstract: With the Guiyang ground-based microwave radiometer, radiosonde and ground observation data at 08:00 and 20:00 from January 2017 to May 2019, this paper evaluates the seasonal variation and the standard deviation of bias of temperature profile retrieved from the microwave radiometer in clear sky. The results show that: (1) the average relative biases of the vertical profiles of temperature from spring to winter are as follows: -1.71-1.35℃, -1.69-0.78℃, -2.24-1.29℃ and -4.98-1.49℃, and the standard deviations of biases are as follows: 1.13-2.93℃, 0.43-1.86℃, 0.76-1.92℃, 0.62-3.48℃. The vertical changes and standard deviations of biases from spring to autumn are the same, while the changes and standard deviations of error in winter are larger than those in other seasons. (2) The samples of inversion and biases have the best correlation. Among the three expressions of inversion, the inversion thickness and the bias have an best correlation when the inversion thickness is 490-740 m. The largest correlation coefficient appears when the inversion thickness is 650 m. The inversion magnitude and the error have the best correlation when the inversion thickness is less than 490 m or more than 740 m and inversion threshold is taken as 1.2℃. (3) The correlation between the average height of inversion and bias is analyzed. It is found that the correlation between the two is poor. The height of bias can not be reflected by the height of inversion.

     

/

返回文章
返回