基于改进三次样条插值的探空资料质控研究  

RESEARCH ON QUALITY CONTROL OF RADIOSONDE DATA BASED ON IMPROVED CUBIC SPLINE INTERPOLATION ALGORITHM

作　　者：杨荣芳 马瑞琪 张光磊 YANG Rongfang;MA Ruiqi;ZHANG Guanglei(Hebei Meteorological Technology and Equipment Center,Shijiazhuang 050000,China;Beijing Huayun Oriental Detection Technology Co.,Ltd.,Beijing 100081,China)

机构地区：[1]河北省气象技术装备中心,河北石家庄050000 [2]北京华云东方探测技术有限公司,北京100081

出　　处：《热带气象学报》2023年第2期183-192,共10页Journal of Tropical Meteorology

基　　金：河北省气象局科研项目(19ky09);河北省气象局地基遥感探测数据融合创新团队共同资助。

摘　　要：由于L波段探空资料存在缺测及稀疏、零散的数据,造成信息量在时空分布上的不均匀性。考虑到参与插补计算的权重因子个数和插补半径对探空观测数据影响较大,在三次样条插值基础上将插补半径和参与计算权重因子个数作为约束条件引入算法中,以被插值点±40gpm为插值半径,取有效因子个数≥3个,提出了改进的三次样条插补算法。通过交叉验证分时段对插补结果进行检验评估,气温AE在±0.1℃之间,RMSE值小于0.08;相对湿度AE在-1.0%~+0.7%之间,RMSE值小于0.3;风速数据AE在±0.6m/s之间,RMSE值小于0.2,表明改进三次样条插补算法对探空资料的质控效果在时间和空间上一致性较好。在晴空条件下相对湿度的插补效果优于有降水天气,气温和风速的插补效果不受晴空条件和降水条件的影响。Due to the lack of detection and sparse and scattered data in the L-band radiosonde data,the spatial and temporal distribution of information is not uniform.Considering that the number of weighting factors and interpolation radius involved in the interpolation calculation has great influence on the radiosonde observation data,they are introduced into the algorithm as constraints on the basis of cubic spline interpolation,and an improved cubic spline interpolation algorithm is proposed.The interpolation results and the original data were tested and evaluated by cross-validation at different time intervals.The absolute error(AE)of air temperature is±0.1℃and the RMSE value is less than 0.08.The relative humidity AE ranges from-1.0%to+0.7%,and the RMSE is less than 0.3.The AE of wind speed is±0.6 m/s,and the RMSE is less than 0.2,indicating that the improved cubic spline interpolation algorithm has good consistency in both time and space.The interpolated relative humidity is better under the condition of clear sky than with precipitation weather,and the interpolated temperature and wind speed is not affected by either the clear sky or rainfall conditions.

关 键 词：权重因子 插补半径 改进的三次样条插补算法 

分 类 号：P413.2[天文地球—大气科学及气象学]