Scaling effect of area-averaged NDVI derived from ALOS-AVNIR2 data

Integration of multiple optical sensors enhances spatial, spectral, and temporal resolutions of NDVI products; it enables to monitor dynamics in global vegetation over several decades with high data qualities. However, systematic errors appear in NDVI among sensors owing to discrepancies in, for example, spatial resolution called scaling effects. The mechanisms underlying scaling effects in NDVI had been investigated under two-endmember (vegetation and non-vegetation) linear mixture model in previous studies. Since those studies have been limited to the theoretical analysis, validations with actual satellite data is needed. Objective of this study is to validate the theory of scaling effects in NDVI for practical use in the context of inter-sensor calibration by using ALOS-AVNIR2 data. Results of numerical simulation by AVNIR2 well agreed with theory of scaling effects, indicating that the theory can be applied into inter-sensor calibration of multiple resolutions.