-
由于标准激光雷达尚未研制完成,因此采用统计分析的方法对气溶胶激光雷达进行比对观测标定。参加标定的6台激光雷达都有532nm米散射接收通道,可以通过标定得到较为准确的信号,3台雷达有355nm米散射接收通道,1台雷达有1064nm米散射接收通道,较难实现较为准确的标定。参考欧洲EARLINET激光雷达标定后的精度以及我国目前激光雷达标定处于起步阶段的现状,确定本次标定的主要参数目标,如表 1所示。
Table 1. Main parameter targets of aerosol LiDAR calibration (10min cumulative data)
signal name 1km~2km 2km~5km standard deviation/% systematical error/% standard deviation/% systematical error/% 532nm original signal 5 10 5 10 532nm Mie channel backscattering coefficient 15 25 20 30 对于参加试验的气溶胶激光雷达的各种参数的标定,采用统一的方法来计算其误差,主要为系统差和标准差。
假设第1部雷达在距离r1,r2,rk, …, rn处的测量值为x11,x12,x1k, …, x1n;第2部雷达在距离r1,r2,rk, …, rn处的测量值为x21,x22,x2k, …, x2n;第i部雷达在距离r1,r2,rk, …, rn处的测量值为xi1,xi2,xik, …, xin。假设共有m台雷达,则所有m台雷达在r1,r2,rk, …, rn处测量值的平均值为y1, y2,yk, …, yn,其中:
$ y_{k}=\frac{x_{1 k}+x_{2 k}+\cdots+x_{m k}}{m} $
(1) 式中,k表示距离雷达r1,r2, …, rn处的点。每个雷达的系统差的计算方法如下:
$ \sigma_{\mathrm{sys}, i}=\frac{\sum\limits_{k=1}^{n}\left(x_{i k}-y_{k}\right)}{n} $
(2) 式中,i表示第i部雷达。相对系统差的计算方法如下:
$ {R_{{\rm{sys}}, i}} = \frac{{{\sigma _{{\rm{sys}}, i}}}}{{{{\overline y }_k}}} = \frac{{\sum\limits_{k = 1}^n {\frac{{{x_{ik}} - {y_k}}}{{{y_k}}}} }}{n} $
(3) 式中,yk表示所有雷达在单个位置的测量平均值。
标准差的计算方法如下:
$ \sigma_{\mathrm{std}, i}=\left[\frac{\sum\limits_{k=1}^{n}\left(x_{i k}-y_{k}\right)^{2}}{n}\right]^{1 / 2} $
(4) 相对标准差的计算方法如下:
$ {R_{{\rm{std}}, i}} = \frac{{{\sigma _{{\rm{std}}, i}}}}{{{{\overline y }_k}}} = {\left[ {\frac{{\sum\limits_{k = 1}^n {{{\left( {{x_{ik}} - {y_k}} \right)}^2}} }}{n}} \right]^{1/2}}/\left( {\frac{{\sum\limits_{k = 1}^n {{y_k}} }}{n}} \right) $
(5) -
第一批次气溶胶激光雷达的标定在2017年9月开展,参加标定的6台雷达中,任意两台雷达之间的距离不大于100m,为了方便比对过程说明,对6台雷达分别命名为101、102、103、104、105和106。
-
2017-09-08对参加联合比对观测标定的气溶胶激光雷达的信号进行了首次对比分析,取22:00~22:10累加的532nm米散射通道的信号归一化处理,距离平方校正后,进行对比,如图 1所示。可以看出,104号雷达曲线和106号雷达曲线,与计算得到的大气分子线拟合较差,存在明显的问题。因此,对其余4台雷达的信号进行对比显示,并与分子线拟合,可以看出,由于参加标定的雷达的overlap区在0km~1km范围,因此对1km~5km的信号分段进行统计分析,如表 2所示。可以看到,1km~2km高度范围内最大的相对标准差达到了90.8%,2km~5km最大相对标准差为58.8%。
Figure 1. Comparison and analysis of range square correction signals of 532nm Mie scattering channel for 6 LiDARs from 2017-09-08T22:00 to 2017-09-08T22:10(10min accumulation)
Table 2. Comparison and analysis table of 532nm Mie scattering raw signal and backscattering coefficients for 4 LiDARs from 2017-09-08T22:00 to 2017-09-08T22:10
LiDAR number original signal statistics backscattering coefficient statistics 1km~2km 2km~5km 1km~2km 2km~5km systematical error/% standard deviation/% systematical error/% standard deviation/% systematical error/% standard deviation/% systematical error/% standard deviation/% 101 -26.3 39.4 -8.0 16.0 -21.1 27.8 -38.1 69.2 102 -49.7 69.1 -10.3 27.3 -54.7 62.9 -44.0 118.4 103 66.3 90.8 26.8 58.8 65.5 80.3 137.3 244.3 105 9.7 26.1 -8.5 17.9 10.3 35.3 -55.3 86.9 4台雷达的探测数据反演得到后向散射系数廓线,其对比图如图 2所示。对结果进行分析如表 2所示。可以看到,1km~2km高度范围内最大的相对标准差达到了80.3%,2km~5km最大相对标准差为244.3%。分别对雷达在上述2个高度范围内的后向散射系数进行积分,在1km~2km高度范围内,其积分值的变化范围为5.9×10-5~2.2×10-4,变化范围约为最小积分值的272.9%,在2km~5km高度范围内,其积分值的变化范围为6.6×10-6~3.5×10-5,变化范围约为4个雷达在此高度范围内最小积分值的430.3%。
Figure 2. Comparison of range square correction signals of 532nm Mie scattering channel for 4 LiDARs from 2017-09-08T22:00 to 2017-09-08T22:10(10min accumulation)
从表 2可知,参加标定的雷达的532nm通道的原始信号和反演得到的后向散射系数的一致性较差,如不进行标定,组网后的数据无法使用。
-
本次标定试验包括雷达AD卡底噪标定、信号饱和度标定、发射接收光轴同轴度标定、瑞利散射信号合理性标定和联合对比观测标定等5个分项开展的,对每个分项的标定方法和标定结果进行了分析。
-
对于雷达系统的模拟采集通道,其底噪是否水平及噪声的大小,会影响信号的质量,对参加标定的模拟采集通道的底噪进行分析。在所有参加标定的雷达中,101号雷达的532nm米散射信号采用模拟和光子计数融合的方式探测,105号雷达的532nm米散射通道仅使用模拟方式采集。盖上望远镜的盖子,得到采集卡的底噪曲线。图 3是101号雷达P模拟通道的3000s累加底噪图。整个通道包括8000个距离库,距离分辨率为3.75m,把每100个库分成1段,计算该段的噪声的相对标准偏差,以及该段的噪声平均值与整个8000个库噪声的平均值的相对标准差。101号雷达和105号雷达模拟通道的最大相对偏差如表 3所示。
Figure 3. 3000s cumulative background noise map for 532nm P analog receiving channel of No.101 LiDAR
Table 3. Background noise analysis result table of analog channel for No.101 and No.105 LiDARs
LiDAR number calculation results maximum relative standard deviation of each segment/% maximum relative deviation between the average value of each segment of noise and the average value of all noises/% 101 channel P 0.08 0.25 101 channel S 0.05 0.07 105 channel P 21.6 7.6 105 channel S 14.9 5.36 可以看到,两部雷达的模拟通道在进行信号处理时需要定期测量背景噪声,并减除背景噪声,图 4是101号雷达532nm模拟接收通道在校正之前和之后的后向散射系数结果对比图。两者的后向散射系数在2.13km的相对偏差达到了42.7%,在1km~5km高度范围内的平均相对平均偏差为19.4%。
-
对于信号是否饱和的检查方法,是在假设大气稳定的条件下,采用更换衰减片或者降低激光器发射功率的方法。为了缩短测试时间,降低大气气溶胶变化带来的误差,本文中主要采用降低激光器功率的方法来测试,从激光器发射最大的功率开始,逐步降低,测试过程中最小的功率至少要小于最大功率的20%,对不同发射功率对应的不同接收信号,进行距离平方校正,并取对数。作者选取信噪比高的一段数据,且连续分析一段数据的平行度,通过多组数据平行度分析以降低由于激光能量降低带来的信噪比的影响,同时如果大气稳定性不好,可以反复多次测量。
对102号雷达饱和度进行检查标定,标定前后的数据进行处理如图 5a所示,调整发射功率之后,标定的数据如图 5b所示。power 1是接收通道正常工作条件下的回波信号,power 2一直到power 3是逐步降低发射功率后的回波信号。对3条回波信号的距离平方校正信号取对数,分别拟合斜率,分析结果如表 4所示。可以看到, 标定前回波曲线的倾斜角度最大差值为5.1°,标定后的最大差值为1.5°。
Table 4. Comparison of the results before and after saturation calibration for No.102 LiDAR
serial number slope of 1km~2km fitting straight line corresponding angle value/(°) before power 1 calibration 0.0251 1.4 after power 1 calibration -0.6650 -41.7 before power 2 calibration 0.0184 1.1 after power 2 calibration -0.6626 -41.5 before power 3 calibration -0.0645 -3.7 after power 3 calibration -0.7215 -40.2 采用相同的方法对101号~106号雷达的饱和情况进行检查标定,标定后的结果如表 5所示。
Table 5. Calibration result analysis table of saturation check for No.101~106 LiDARs
LiDAR number maximum change of the slope of the fixed-height echo signal fitting straight line maximum change value of the angle corresponding to the slope of the fixed-height echo signal fitting straight line/(°) 101 0.0184 1.0 102 0.0590 1.3 103 0.0013 1.3 104 0.0156 0.9 105 0.0026 0.2 106 0.0083 0.4 -
望远镜对中检查标定的目的是把发射激光的光轴与接收光轴重合。把望远镜分为4个象限,与激光发射口之间的位置关系如图 6所示。包括2种模式: 第1种是发射激光与望远镜非同轴的情况,如图 6a所示,定义为F1方式; 第2种是同轴的情况,如图 6b所示,定义为F2方式。在进行标定时,轮流打开ABCD这4个象限的遮挡板中的一个,接收大气回波信号,采集并比对分析。对于F1的标定方式,由于ABCD 4个象限与发射激光的相对位置不同,在0km~1km,A象限的信号最强,B象限信号最弱,CD象限信号相同,在1km~5km,4个象限的信号应该完全相同; 对于F2模式,0km~5km的信号都应该完全相同。
经过标定后,部分雷达的对中情况有了较大的改善,少数雷达后期需要通过改进接收系统的结构设计,才能进一步优化发射和接收光轴的一致性。104号雷达标定前后雷达ABCD这4个象限的回波信号以及信号的平均值图如图 7所示。由于104号雷达为非同轴结构,AB两个象限在低空的回波信号差值应该较大,CD两个象限在低空理想情况下应该是完全重合的,从图 7a可以看出来,标定前在0km~1km高度B象限的值要高于D象限的值,CD两个象限的信号差值也非常大,标定后(见图 7b)可以看出,CD象限曲线在低空的差值明显降低,光路得到优化。
对标定前后5km以下的4个象限的数据与平均值的相对标准差和相对系统差进行统计分析,如表 6所示。在低空0km~1km高度范围内,CD两个象限平均值的相对系统差从18.7%降到7.8%,标准差从26.7%降到11.3%。
Table 6. Statistical analysis results of different heights before and after four quadrant calibration for No.104 LiDAR
quadrant calculation results 0km~1km 1km~2km 2km~5km systematical error/% standard deviation/% systematical error/% standard deviation/% systematical error/% standard deviation/% before calibration A -3.2 72.6 -2.9 8.3 — — after calibration A -35.0 52.9 -1.3 2.6 1.1 5.3 before calibration B 3.2 72.6 0.3 4.5 — — after calibration B 35.0 52.9 0.4 1.2 -1.0 3.4 before calibration C -18.7 26.7 1.37 3.6 — — after calibration C 7.8 11.3 0.7 1.8 0.2 4.7 before calibration D 18.7 26.7 1.36 7.8 — — after calibration D -7.8 11.3 -0.1 1.2 0.3 3.7 101号、103号、104号气溶胶激光雷达采用F1标定方式,102号、105号和106号雷达采用F2标定方式。根据雷达的硬件特点以及标定时的大气情况,对低、中、高3层回波信号进行统计分析,如表 7所示。
Table 7. Four quadrant calibration results for No.101 LiDAR
LiDAR number quadrant calculation results 0km~1km 1km~2km 2km~5km systematical error/% standard deviation/% systematical error/% standard deviation/% systematical error/% standard deviation/% 101,F1 A -24.2 40.8 -1.6 5.4 1.1 12.4 B 24.2 40.8 2.4 6.2 -0.3 12.6 C -12.1 24.7 0.3 4.8 -0.6 12.2 D 12.1 24.7 -0.1 4.4 -0.1 12.3 102, F2 A 3.3 6.1 -0.4 2.9 -0.7 8.56 B 1.1 2.0 0.2 1.6 -0.2 8.27 C -2.0 3.2 -0.2 2.7 -0.9 8.39 D -2.3 4.2 0.4 2.0 1.7 9.63 103, F1 A -79.1 147.3 2.7 5.5 -2.2 13.9 B 79.1 147.3 -3.0 7.0 1.4 16.5 C 1.8 4.6 -5.6 15.9 -7.7 20.9 D -1.8 4.6 6.4 13.9 9.8 25.4 104, F1 A -35.0 52.9 -1.3 2.6 1.1 5.3 B 35.0 52.9 0.4 1.2 -1.0 3.4 C 7.8 11.3 0.7 1.8 0.2 4.7 D -7.8 11.3 -0.1 1.2 0.3 3.7 105, F2 A 27.2 39.6 0 2.8 0.2 1.7 B -16.5 24.0 -0.8 3.1 0 2.5 C -29.3 39.9 -0.8 3.1 -0.5 3.3 D -38.5 69.1 0.2 6.7 -0.2 4.6 106, F2 A 18.6 30.7 0 9.4 -5.1 51.4 B -0.6 5.3 0 7.9 2.6 45.0 C 3.9 15.2 0 9.7 -7.0 51.4 D -18.7 37.0 0 8.0 7.9 37.5 -
在大气边界层之上,大气气溶胶的含量非常少,可以假设为0,利用计算出来的大气分子回波信号与雷达测量得到的边界层以上的大气回波信号进行对比,来判断雷达的光路和光电器件是否存在问题。106号雷达532nm米散射通道的回波信号与大气分子信号对比图如图 8所示,图 8a是标定之前的对比结果,图 8b是标定之后的对比结果,标定之前和标定之后雷达回波信号与计算的大气分子信号之间的标准差和系统差,如表 8所示。
Table 8. Comparison of Rayleigh signal of No.106 LiDAR before and after calibration
status calculation results 3km~5km 5km~7km 7km~10km systematical error/% standard deviation/% systematical error/% standard deviation/% systematical error/% standard deviation/% before No.106 calibration -43.4 61.7 54.7 104.7 166.6 259.9 after No.106 calibration 9.3 10.2 5.3 7.9 -5.4 15.2 对101号~106号雷达的大气分子回波信号进行检查标定,标定后的结果如表 9所示。可以看到,在3km~7km高度系统差一般都小于10%,标准差都小于20%,个别雷达标准差较大,主要是由于远场信号的信噪比降低造成的。
Table 9. Calibration results of Rayleigh signal for No.101~106 LiDARs
LiDAR number calculation results 3km~5km 5km~7km 7km~10km systematical error/% standard deviation/% systematical error/% standard deviation/% systematical error/% standard deviation/% 101 -2.4 7.1 -1.6 9.2 -0.4 14.3 102 -4.4 10.6 1.2 19.6 -13.1 44.3 103 -7.8 14.6 -3.4 19.0 -10.9 38.8 104 9.3 10.2 5.3 7.9 -5.4 15.2 105 -1.8 8.4 -0.2 11.7 -5.4 21.1 106 7.3 10.3 4.7 6.8 -7.2 18.6 -
在完成了雷达的AD卡底噪、信号饱和度、望远镜四象限对中、瑞利散射信号拟合等几项标定之后,对6部雷达在同一时刻同一时间段内累加的实测信号,进行对比分析。2017-09-14T03:00~2017-09-14T03:10,6部雷达在南郊观象台开展观测,对532nm米散射通道回波信号,进行距离平方校正和归一化处理,如图 9所示。去掉0km~1km的overlap区,对雷达1km~5km高度范围内的信号进行相对系统差和相对标准差的统计分析,结果如表 10所示。可以看到, 1km~2km范围内,最大相对标准差为10.8%,2km~5km范围内最大相对标准差为9.3%,6部雷达观测原始强度数据的一致性有了明显提高。
Figure 9. Comparison of 532nm normalized signal from 2017-09-14T03:00 to 2017-09-14T03:10(10min accumulation)
Table 10. Calibration results of 532nm Mie scattering signal for 6 LiDARs (10min accumulation)
LiDAR number calculation results 1km~2km 2km~5km systematical error/% standard deviation/% systematical error/% standard deviation/% 101 -4.5 10.8 -1.3 7.2 102 -0.5 4.4 3.3 5.1 103 2.8 9.2 -4.2 9.3 104 -3.1 8.3 4.3 6.1 105 2.8 7.3 -3.0 4.7 106 2.6 8.9 0.8 3.9 -
使用相同的反演算法对6部雷达标定好的原始信号进行反演,得到6部雷达同址同时刻观测的后向散射系数曲线,如图 10所示。对1km~5km高度范围内的反演结果进行相对系统差和相对标准差的统计分析,如表 11所示。可以看到,1km~2km高度范围内的最大的相对标准差为21.0%,2km~5km最大相对标准差为35.9%。
Figure 10. Curves of backscattering coefficient for 532nm signal channel of 6 LiDARs from 2017-09-14T03:00am to 2017-09-14T03:10(10min accumulation)
Table 11. Comparison analysis results of retrieved backscattering coefficient curves for 532nm signal channel of 6 LiDARs from 2017-09-14T03:00 to 2017-09-14T03:10(10min accumulation)
LiDAR number calculation results 1km~2km 2km~5km systematical error/% standard deviation/% systematical error/% standard deviation/% 101 -10.6 20.1 -17.6 28.1 102 -2.7 11.2 3.7 12.4 103 7.6 19.0 -6.1 28.1 104 -0.7 16.6 30.8 35.9 105 5.3 15.9 -5.7 10.8 106 1.2 21.0 -5.1 18.0 分别对雷达在上述2个高度范围内的后向散射系数进行积分,在1km~2km高度范围内,其积分值的变化范围为3.4×10-5~4.1×10-5,变化范围约为6部雷达在该高度范围内积分最小值的20.6%,在2km~5km高度范围内,其积分值的变化范围为3.7×10-5~5.9×10-5,变化范围约为6部雷达在该高度范围内积分最小值的32.4%。
超大城市试验气溶胶激光雷达标定及结果分析
Calibration and result analysis of aerosol LiDAR in megacity experiment
-
摘要: 为了保证雷达观测数据的准确性, 2017年9月中国气象局气象探测中心在国内首次开展了组网气溶胶激光雷达标定实验, 通过采用单部雷达的光电系统标定和基于统计分析策略的多部雷达比对观测标定方法, 对组网试验雷达共有的532nm米散射通道进行了检查标定。结果表明, 标定完成后, 532nm通道的后向散射系数的相对标准差在1km~2km高度范围从90.8%降低到20.4%, 在2km~5km高度范围从244.3%降低到了35.9%, 数据质量得到了较大提高。此次气溶胶激光雷达标定试验, 使雷达后向散射系数差异显著减小, 大大地改善了雷达观测数据的一致性, 这将为气溶胶激光雷达组网应用提供很好的硬件质控保证。Abstract: In order to ensure the accuracy of the radar observation data, the first calibration experiment of networked aerosol light detection and ranging (LiDAR) was carried out by Meteorological Observation Center of China Meteorological Administration in september 2017. By adopting the method of single radar's photoelectric system calibration and multi-radar comparison observation calibration based on statistical analysis strategies, the 532nm scattering channel shared by the networked experiment radars was checked and calibrated. After calibration, the relative standard deviation of backscattering coefficient in 532nm channel was reduced from 90.8% to 20.4% in height range of 1km~2km, and from 244.3% to 35.9% in height range of 2km~5km. And the data quality has been greatly improved. In the aerosol lidar calibration experiment, the radar backscattering coefficient difference was significantly reduced, and the consistency of data observation was greatly improved.This work will provide good hardware quality control assurance for aerosol lidar networking applications.
-
Table 1. Main parameter targets of aerosol LiDAR calibration (10min cumulative data)
signal name 1km~2km 2km~5km standard deviation/% systematical error/% standard deviation/% systematical error/% 532nm original signal 5 10 5 10 532nm Mie channel backscattering coefficient 15 25 20 30 Table 2. Comparison and analysis table of 532nm Mie scattering raw signal and backscattering coefficients for 4 LiDARs from 2017-09-08T22:00 to 2017-09-08T22:10
LiDAR number original signal statistics backscattering coefficient statistics 1km~2km 2km~5km 1km~2km 2km~5km systematical error/% standard deviation/% systematical error/% standard deviation/% systematical error/% standard deviation/% systematical error/% standard deviation/% 101 -26.3 39.4 -8.0 16.0 -21.1 27.8 -38.1 69.2 102 -49.7 69.1 -10.3 27.3 -54.7 62.9 -44.0 118.4 103 66.3 90.8 26.8 58.8 65.5 80.3 137.3 244.3 105 9.7 26.1 -8.5 17.9 10.3 35.3 -55.3 86.9 Table 3. Background noise analysis result table of analog channel for No.101 and No.105 LiDARs
LiDAR number calculation results maximum relative standard deviation of each segment/% maximum relative deviation between the average value of each segment of noise and the average value of all noises/% 101 channel P 0.08 0.25 101 channel S 0.05 0.07 105 channel P 21.6 7.6 105 channel S 14.9 5.36 Table 4. Comparison of the results before and after saturation calibration for No.102 LiDAR
serial number slope of 1km~2km fitting straight line corresponding angle value/(°) before power 1 calibration 0.0251 1.4 after power 1 calibration -0.6650 -41.7 before power 2 calibration 0.0184 1.1 after power 2 calibration -0.6626 -41.5 before power 3 calibration -0.0645 -3.7 after power 3 calibration -0.7215 -40.2 Table 5. Calibration result analysis table of saturation check for No.101~106 LiDARs
LiDAR number maximum change of the slope of the fixed-height echo signal fitting straight line maximum change value of the angle corresponding to the slope of the fixed-height echo signal fitting straight line/(°) 101 0.0184 1.0 102 0.0590 1.3 103 0.0013 1.3 104 0.0156 0.9 105 0.0026 0.2 106 0.0083 0.4 Table 6. Statistical analysis results of different heights before and after four quadrant calibration for No.104 LiDAR
quadrant calculation results 0km~1km 1km~2km 2km~5km systematical error/% standard deviation/% systematical error/% standard deviation/% systematical error/% standard deviation/% before calibration A -3.2 72.6 -2.9 8.3 — — after calibration A -35.0 52.9 -1.3 2.6 1.1 5.3 before calibration B 3.2 72.6 0.3 4.5 — — after calibration B 35.0 52.9 0.4 1.2 -1.0 3.4 before calibration C -18.7 26.7 1.37 3.6 — — after calibration C 7.8 11.3 0.7 1.8 0.2 4.7 before calibration D 18.7 26.7 1.36 7.8 — — after calibration D -7.8 11.3 -0.1 1.2 0.3 3.7 Table 7. Four quadrant calibration results for No.101 LiDAR
LiDAR number quadrant calculation results 0km~1km 1km~2km 2km~5km systematical error/% standard deviation/% systematical error/% standard deviation/% systematical error/% standard deviation/% 101,F1 A -24.2 40.8 -1.6 5.4 1.1 12.4 B 24.2 40.8 2.4 6.2 -0.3 12.6 C -12.1 24.7 0.3 4.8 -0.6 12.2 D 12.1 24.7 -0.1 4.4 -0.1 12.3 102, F2 A 3.3 6.1 -0.4 2.9 -0.7 8.56 B 1.1 2.0 0.2 1.6 -0.2 8.27 C -2.0 3.2 -0.2 2.7 -0.9 8.39 D -2.3 4.2 0.4 2.0 1.7 9.63 103, F1 A -79.1 147.3 2.7 5.5 -2.2 13.9 B 79.1 147.3 -3.0 7.0 1.4 16.5 C 1.8 4.6 -5.6 15.9 -7.7 20.9 D -1.8 4.6 6.4 13.9 9.8 25.4 104, F1 A -35.0 52.9 -1.3 2.6 1.1 5.3 B 35.0 52.9 0.4 1.2 -1.0 3.4 C 7.8 11.3 0.7 1.8 0.2 4.7 D -7.8 11.3 -0.1 1.2 0.3 3.7 105, F2 A 27.2 39.6 0 2.8 0.2 1.7 B -16.5 24.0 -0.8 3.1 0 2.5 C -29.3 39.9 -0.8 3.1 -0.5 3.3 D -38.5 69.1 0.2 6.7 -0.2 4.6 106, F2 A 18.6 30.7 0 9.4 -5.1 51.4 B -0.6 5.3 0 7.9 2.6 45.0 C 3.9 15.2 0 9.7 -7.0 51.4 D -18.7 37.0 0 8.0 7.9 37.5 Table 8. Comparison of Rayleigh signal of No.106 LiDAR before and after calibration
status calculation results 3km~5km 5km~7km 7km~10km systematical error/% standard deviation/% systematical error/% standard deviation/% systematical error/% standard deviation/% before No.106 calibration -43.4 61.7 54.7 104.7 166.6 259.9 after No.106 calibration 9.3 10.2 5.3 7.9 -5.4 15.2 Table 9. Calibration results of Rayleigh signal for No.101~106 LiDARs
LiDAR number calculation results 3km~5km 5km~7km 7km~10km systematical error/% standard deviation/% systematical error/% standard deviation/% systematical error/% standard deviation/% 101 -2.4 7.1 -1.6 9.2 -0.4 14.3 102 -4.4 10.6 1.2 19.6 -13.1 44.3 103 -7.8 14.6 -3.4 19.0 -10.9 38.8 104 9.3 10.2 5.3 7.9 -5.4 15.2 105 -1.8 8.4 -0.2 11.7 -5.4 21.1 106 7.3 10.3 4.7 6.8 -7.2 18.6 Table 10. Calibration results of 532nm Mie scattering signal for 6 LiDARs (10min accumulation)
LiDAR number calculation results 1km~2km 2km~5km systematical error/% standard deviation/% systematical error/% standard deviation/% 101 -4.5 10.8 -1.3 7.2 102 -0.5 4.4 3.3 5.1 103 2.8 9.2 -4.2 9.3 104 -3.1 8.3 4.3 6.1 105 2.8 7.3 -3.0 4.7 106 2.6 8.9 0.8 3.9 Table 11. Comparison analysis results of retrieved backscattering coefficient curves for 532nm signal channel of 6 LiDARs from 2017-09-14T03:00 to 2017-09-14T03:10(10min accumulation)
LiDAR number calculation results 1km~2km 2km~5km systematical error/% standard deviation/% systematical error/% standard deviation/% 101 -10.6 20.1 -17.6 28.1 102 -2.7 11.2 3.7 12.4 103 7.6 19.0 -6.1 28.1 104 -0.7 16.6 30.8 35.9 105 5.3 15.9 -5.7 10.8 106 1.2 21.0 -5.1 18.0 -
[1] GRABBE G C, BOSENBERG J, DIER H, et al. Intercomparison of ozone measurements between lidar and ECC-sondes[J]. Contributions to Atmospheric Physics, 1996, 69(1): 189-203. [2] MCDERMID I S, GODIN S M, BARNES R A, et al. Comparison of ozone profiles from ground-based lidar, electrochemical concentration cell balloon sonde, ROCOZ-A rocket ozonesonde, and stratospheric aerosol and gas experiment satellite measurements[J]. Journal of Geophysical Research Atmospheres, 1990, 95(D7): 10037-10042. doi: 10.1029/JD095iD07p10037 [3] FERRARE R A, WHITEMAN D N, MELFI S H, et al. A comparison of water vapor measurements made by Raman lidar and radiosondes[J]. Journal of Atmospheric & Oceanic Technology, 1995, 12(6): 1177-1195. [4] TANG J. Research on Raman LiDAR for atmospheric temperature and humidity profiles[D]. Xi'an: Xi'an University of Technology, 2012: 9-23(in Chinese). [5] WANG Sh L, SU J, ZHAO P T, et al. A pure rotational Raman-lidar based on three-stage Fabry-Perot etalons for monitoring atmospheric temperature[J]. Acta Physica Sinica, 2008, 57(6): 3941-3947(in Chinese). doi: 10.7498/aps.57.3941 [6] SHANG Zh. Pure rotational Raman lidar for the measurement of atmospheric temperature in the bottom of the troposphere[D]. Beijing: University of Science and Technology of China, 2017: 13-30(in Chin-ese). [7] SHERLOCK V, GARNIER A, HAUCHECORNE A, et al. Implementation and validation of a Raman lidar measurement of middle and upper tropospheric water vapor[J]. Applied Optics, 1999, 38(27): 5838-5850. doi: 10.1364/AO.38.005838 [8] XIA J R, WANG P C, MIN M. Observation and validation of wind parameters measured by Doppler wind lidar windcube[J]. Climatic and Environmental Research, 2011, 16(6): 733-741(in Chinese). [9] STEINHAGEN H, BAKAN S, BOSENBERG J, et al. Field campaign LINEX 96/1-possibilities of water vapor observation in the free atmosphere[J]. Meteorologische Zeitschrift, 1998, 7(6): 377-391. doi: 10.1127/metz/7/1998/377 [10] CHEN Y B, GAO Y C, LIU B Y, et al. The Assessment and analysis of the upper wind data measured the doppler lidar based on the chebyshev function[J]. Journal of Tropical Meteorology, 2014, 32(2): 465-486. [11] LI S T, MA S Q, GAO Y C, et al. Comparative analysis of cloud base heights observed by cloud radar and ceilometers[J]. Meteorological Monthly, 2015, 41(2): 212-218. [12] LI Ch C, LIU Q H, M J T, et al. An aersol pollution episode in Hong Kong with remote sensing products of MODIS and LiDAR[J]. Journal of Applied Meteorological Science, 2004, 15(6): 641-650(in Chinese). [13] ZHU J K, LI L J, LIN X Z. Research on measurement field planning of lidar measurement system[J]. Laser Technology, 2021, 45(1): 99-104(in Chinese). [14] MCDERMID I S, GODIN S M, Walsh D T. LiDAR measurements of stratospheric ozone and intercomparisons and validation[J]. Applied Optics, 1990, 29(33): 4914-4923. doi: 10.1364/AO.29.004914 [15] BOSENBERG J, ANSMANN A, BALDASANO J M, et al. EARLINET: A European aerosol research lidar network in laser remote sensing of the atmosphere[C]//Selected Papers of the 20th International Laser Radar Conference. Palaiseau, France: Edition Ecole Polytechnique, 2001: 155-158. [16] MATTHAIS V, FREUDENTHALER V, AMODEO A, et al. Aerosol lidar intercomparison in the framework of the EARLINET project. 1. Instruments[J]. Applied Optics, 2004, 43(4): 961-976. doi: 10.1364/AO.43.000961 [17] BOCKMANN C, WANDINGER U, ANSMANN A, et al. Aerosol lidar intercomparison in the framework of the EARLINET project. 2. Aerosol backscatter algorithms[J]. Applied Optics, 2004, 43(4): 977-989. doi: 10.1364/AO.43.000977 [18] WANDINGER U, FREUDENTHALER V, BARRS H, et al. EARLINET instrument intercomparison campaigns: Overview on strategy and results[J]. Atmospheric Measurement Techniques, 2016, 9(3): 1001-1023. doi: 10.5194/amt-9-1001-2016 [19] DAMICO G, AMEODEO A, BARRS H, et al. EARLINET single calculus chain-overview on methodology and strategy[J]. Atmospheric Measurement Techniques, 2015, 8(11): 4891-4916. doi: 10.5194/amt-8-4891-2015 [20] PAPPALARDO G, AMODEO A, APITULEY A, et al. EARLINET: Towards an advanced sustainable European aerosol lidar network[J]. Atmospheric Measurement Techniques, 2014, 7(8): 2389-2409. doi: 10.5194/amt-7-2389-2014 [21] ZHOU B, ZHANG L, JIANG D M, et al. Analysis of aerosol optical depth over Lanzhou based on lidar measurement[J]. Journal of Arid Meteorology, 2013, 31(4): 666-671. [22] CHEN B L, ZHANG Y C, CHEN S Y, et al. Fitting aerosol optical depth and PM2.5 in atmospheric boundary layer by using rotational Roman-mie lidar[J]. Transaction of Beijing Institute of Technology, 2016, 36(8): 857-861. [23] LI L H, LIU W Q, ZHANG T Sh, et al. A new micro-pulse lidar for atmospheric horizontal visibility measurement[J]. Chinese Journal of Lasers, 2014, 41(9): 0908005(in Chinese). doi: 10.3788/CJL201441.0908005 [24] GUO W, BU L B, JIA X H, et al. Analyses on sand-dust aerosol properties with ceilometer in Beijing[J]. Meteorology Monthly, 2016, 42(12): 1540-1546. [25] BU L B, YUAN J, GAO A Z, et al. Analysis of haze-fog events based on laser celiometer[J]. Acta Photonica Sinica, 2014, 43(9): 64-69(in Chinese).