Publications

Yang, XY; Meng, F; Fu, PJ; Wang, YQ; Liu, YH (2022). Time-frequency optimization of RSEI: A case study of Yangtze River Basin. ECOLOGICAL INDICATORS, 141, 109080.

Abstract
Remote Sensing Ecological Index (RSEI) is one of the most widely used ecological quality assessment indicators. Due to the noise caused by adverse atmoshperic conditions and other factors, the RSEI calculated from the original image usually has the phenomena of lack of information and unstable image quality. Therefore, based on Google Earth Engine (GEE) cloud platform, this study adopts three common data reconstruction algorithms firstly, namely: Savitory-Golay filter (SG), harmonic analysis of time series (HANTS), Whittaker Smoother (WS), which are used to reconstruct the original MODIS time series data in the Yangtze River Basin (YRB) from 2000 to 2020, in order to optimize the calculation process of RSEI. At the same time, three indicators (correlation coefficient (R), standard deviation (STD), root mean square error (RMSE)) are used for the accuracy evaluation. The results show that data reconstruction can fill gaps in RSEI, the reconstruction performance of WS and SG for four parameters is better than HANTS, and the four SG reconstructed sequences have the strongest correlation with the original sequences (R between 0.8 ~ 1), while the WS reconstruction sequence has the lowest error value (both STD and RMSE are less than 1), both of them can correct the pixel value, which is conducive to maintaining the stability of RSEI in the temporal dimension; the RSEI produced by HANTS has the best accuracy, that is, R, STD, RMSE are respectively 0.898, 0.130, 0.104. As shown by the research, it is necessary to de-noise each parameter before synthesizing RSEI. This study can provide a theoretical basis for applying time-frequency algorithms to optimize the ecological monitoring performance of RSEI.

DOI:
10.1016/j.ecolind.2022.109080

ISSN:
1872-7034