Analysis of Extracting Prior BRDF from MODIS BRDF Data

Remote Sensing

Tiejun Cui¹  Hu Zhang¹  Yi Lian¹  Peng Du¹  Yadong Dong²  Yang Li²  Ziti Jiao² 

[1] College of Urban and Environmental Sciences, Tianjin Normal University, Tianjin 300387, China;State Key Laboratory of Remote Sensing Science, and School of Geography, Beijing Normal University, Beijing 100875, China;

关键词: reflectance anisotropy;    archetypal BRDFs;    NDVI;    land cover;    Anisotropy Flat Index (AFX);    MODIS;   

DOI  :  10.3390/rs8121004

来源: DOAJ

【 摘 要 】

Many previous studies have attempted to extract prior reflectance anisotropy knowledge from the historical MODIS Bidirectional Reflectance Distribution Function (BRDF) product based on land cover or Normalized Difference Vegetation Index (NDVI) data. In this study, the feasibility of the method is discussed based on MODIS data and archetypal BRDFs. The BRDF is simplified into six archetypal BRDFs that represent different reflectance anisotropies. Five-year time series of MODIS BRDF data over three tiles are classified into six BRDF archetype classes according to the Anisotropy Flat indeX (AFX). The percentage of each BRDF archetype class in different land cover classes or every 0.1-NDVI interval is determined. Nadir BRDF-Adjusted Reflectances (NBARs) and NDVIs simulated from different archetypal BRDFs and the same multi-angular observations are compared to MODIS results to study the effectiveness of the method. The results show that one land cover type, or every 0.1-NDVI interval, contains all the potential BRDF shapes and that one BRDF archetypal class makes up no more than 40% of all data. Moreover, the differences between the NBARs and NDVIs simulated from different archetypal BRDFs are insignificant. In terms of the archetypal BRDF method and MODIS BRDF product, this study indicates that the land cover or NDVI is not necessarily related to surface reflectance anisotropy.

【 授权许可】

Unknown