DINEOF重构遥感叶绿素a数据质量分析

Data Quality Analysis of Reconstruction of Remote Sensing Chlorophyll a by DINEOF

文章利用2009—2020年南海海域多传感器融合叶绿素a月平均遥感数据,采用DINEOF方法对数据进行整体重构和缺失点重构,通过对整体重构数据质量的对比分析,探讨了缺失点重构数据的可靠性。结果表明,在数据缺失点上,整体重构和缺失点重构得到的重构数据完全相同,说明可仅针对缺失点进行数据重构,建立无缺失叶绿素a遥感数据集。整体重构数据与原始数据的均方根误差和相关系数分别为0.125 7 mg·m^(-3)和0.93。从重构数据分布图可以看出,数据缺失率越高,平滑越明显,但整体重构数据会更明显。重构相对误差在20%范围以内的数据点比例与数据缺失率数据存在一定负相关关系,数据缺失率越高的月份,该比例越低,说明缺失点重构数据的可靠性越差。

Based on the monthly averaged multi-sensor chlorophyll-a data in the South China Sea from 2009 to 2020,the DINEOF method is used to reconstruct the chlorophyll-a data by two different ways, i.e.,the whole reconstruction and the missing-point reconstruction. Through the comparative analysis of the quality of the whole reconstructed data, the reliability of the missing-point reconstructed data is discussed. The results show that at the locations of missing data, the reconstructed data from the whole reconstruction and missing-point reconstruction are exactly the same, which means we can only use the missing-point way to reconstruct the data, and establish the chlorophyll-a dataset without measurement gaps. The root mean square error and correlation coefficient between the reconstructed and the original data are 0.125 7 mg·m^(-3)and 0.93,respectively. It can be seen from the distribution of reconstructed data that the higher the missing data rate, the more significant the spatial smoothing, but the whole reconstructed data is more significant. There is a certain negative correlation between the proportion of data points with relative error less than 20% and the missing data rate. The higher the data missing rate, the lower the proportion, indicating that the reliability of reconstruction data is worse.