Publications

Shi, HY; Xiao, ZQ; Wang, Q; Wu, DX (2021). Multiparameter Estimation From Landsat Observations With Topographic Consideration. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 59(9), 7353-7369.

Abstract
The applications of high-spatial-resolution satellite data have been increasing in recent years owing to improvements in sensor techniques, and the errors in estimated parameters induced by ignoring topographic effects are increasingly stressed because their effects are important for parameter retrieval from high-spatial-resolution satellite observations. A coupled surface-atmosphere model is employed to develop a two-step multiparameter estimation scheme to simultaneously estimate multiple parameters (leaf area index, LAI; aerosol optical depth, AOD; photosynthetically active radiation, PAR; incident shortwave radiation, ISR; surface albedo, and fraction of absorbed photosynthetically active radiation, FAPAR) from long-term Landsat 4-8 top-of-atmosphere (TOA) observations. First, the influential parameters of the coupled model are retrieved through optimization retrieval strategies. Then, these estimated parameters are entered into the coupled model to compute the PAR, ISR, surface reflectance, surface albedo, and FAPAR. Validation of this scheme with in situ measurements from 57 sites demonstrates that it can successfully estimate multiple parameters from Landsat TOA data, with root mean square errors (RMSEs) of LAI, AOD, FAPAR, visible albedo, shortwave albedo, PAR, and ISR of 0.69, 0.16, 0.13, 0.034, 0.047, 26.80, and 64.28 W/m(2), respectively. In the two-step multiparameter estimation scheme, atmospheric and topographic corrections of satellite observations are avoided because the atmospheric and topographic effects are incorporated, and the surface anisotropy is also effectively considered. In addition, by using the two-step multiparameter estimation scheme, physical connections among the multiple parameters are ensured since they are estimated from the same physical model.

DOI:
10.1109/TGRS.2021.3057377

ISSN:
0196-2892