This paper describes (1) how the normalized difference vegetation index (NDVI) evaluation changes and (2) how the NDVI-based evapotranspiration estimation changes, as the spatial resolution of remote sensing images becomes coarser. This analysis uses a degradation algorithm based on the MTF. With the airborne MSS images of 6.25-m resolution, images simulated have resolutions of 16, 30, 50, 80, and 250m, which are corresponding to the resolution of ADEOS/AVNIR, Landsat/TM, MOS-1/MESSR, Landsat/MSS, and ADEOS-II/GLI, respectively. The study area is the Seto-Nagakute area (3×6km) consisting of urbanized areas, forest and cultivated areas. From the pixel-based analysis the authors have found that the mean of NDVI does not change remarkably as the spatial resolution becomes coarser. Meanwhile, the means of estimated evapotranspirations for the 6.25- and 250-m images are about 48 and about 47mm/month respectively. The standard deviation of estimated evapotranspirations gradually becomes smaller as the spatial resolution becomes coarser. Consequently, the NDVI and the estimated evapotranspiration averaged over the area concerned do not depend much on the spatial resolution of the remotely sensed image. The authors also assess the effect of urbanization on monthly evapotranspiration, using a diagram based on the forest, cultivated and urbanized land classes.
View full abstract