Using Wavelet Transform for MODIS Time-Series Analysis
Freitas, Instituto Nacional de Pesquisas Espaciais, firstname.lastname@example.org
Shimabukuro, Instituto Nacional de Pesquisas Espaciais, email@example.com
The Normalized Difference Vegetation Index (NDVI) and the Enhanced Vegetation Index - EVI are widely accepted as good indicators of vegetation properties at regional scale. This work reports preliminary findings of a science training research project carried out as part of the LBA-Ecology program. The experiment was performed to assess the suitability of wavelets transform for Moderate Resolution Imaging Spectroradiometer - MODIS time-series filtering and analysis in the Amazon Region. In that, MOD13 Vegetation Indices product, 1km spatial resolution, 16-Days composite images from 2000 to 2005 were used as input data. The test site was a west - east transect (75º W to 35º W at 6º South latitude) encompassing several landscape types as tropical forest, deforestation area, transitional forest and semi-arid vegetation (“caatinga”). The Discrete Wavelets Transform (DWT) were used for filtering MODIS time-series and Continuous Wavelets Transform (CWT) for time-series analysis. The results showed that wavelets transform provides a gain in multitemporal analysis and visualization on interannual NDVI and EVI variability patterns.
Science Theme: LC (Land Use and Land Cover Change)