论文部分内容阅读
由于气溶胶的影响,传统的瑞利散射法测量低空大气温度有一定的局限,为此开展了纯转动拉曼法测量低空大气温度。利用纯转动拉曼激光雷达在北京进行了2个月的大气温度观测,由观测数据反演了温度廓线。在基于N_2和O_2的纯转动拉曼谱线特征进行大气温度反演过程中,分析了平滑窗口、定标范围和定标常数对温度反演精度的影响。结果显示随着平滑窗口的增大,雷达和无线电探空仪测量的温度之间的平均绝对偏差先减小后增加,为有效去除信号中随机误差的影响,同时保留温度廓线的垂直结构,平滑窗口应选择600~1 200m比较好。定标范围不同,雷达和无线电探空仪测量的温度之间的平均绝对偏差就不同,相对变化约为0.07K。当定标常数a,b都增大或都减小时,雷达和无线电探空仪测量的温度之间的平均偏差增大,当一个增大另一个减小时,平均偏差相互抵消;a,b的变化不是等几率的,在符号上总是相反的;平均偏差对a的变化不敏感,对b的变化也不敏感,对a与b的整体变化敏感,约91.7%平均偏差落入-3~3K之间。该研究分析结果对纯转动拉曼激光雷达数据反演中涉及的平滑窗口、定标范围的选择提供了理论依据,对激光雷达定标常数造成实际温度反演结果的误差提供了参考。
Due to the influence of aerosol, the traditional Rayleigh scattering method has some limitations in the measurement of low-altitude atmospheric temperature. For this reason, purely rotational Raman method was used to measure the temperature of low-altitude atmosphere. Using a purely rotating Raman lidar in Beijing for 2 months, the atmospheric temperature was observed and the temperature profile was retrieved from the observed data. In the process of atmospheric temperature inversion based on purely rotational Raman spectra of N_2 and O_2, the influence of smooth window, calibration range and calibration constant on temperature inversion accuracy was analyzed. The results show that with the increase of the smoothing window, the average absolute deviation between the temperature measured by the radar and the radiosonde decreases first and then increases. In order to effectively remove the influence of random errors in the signal while preserving the vertical structure of the temperature profile, Smooth window should choose 600 ~ 1 200m is better. The calibration range is different and the average absolute deviation between the temperatures measured by the radar and the radiosonde is different, with a relative change of about 0.07K. When the scaling constants a, b both increase or decrease, the average deviation between the temperature measured by the radar and the radiosonde increases, and when one increases the other decreases, the average deviations cancel each other out; a, b The change is not equal to the probability, the sign is always the opposite; the average deviation is not sensitive to changes in a, b is also not sensitive to changes in the overall changes in a and b sensitive to about 91.7% average deviation of -3 ~ 3K between. The results of this study provide a theoretical basis for the selection of smooth windows and calibration ranges involved in the inversion of purely rotated Raman lidar data, and provide reference for the error of actual temperature inversion results caused by the calibration constant of Lidar.