Characterising the spatial pattern of phenology for the tropical vegetation of India using multi-temporal MERIS chlorophyll data

The annual growth cycles of terrestrial ecosystems are related to long-term regional/global climatic patterns. Understanding vegetation phenology and its spatio-temporal variation is required to reveal and predict ongoing changes in Earth system dynamics. The study attempts to characterize the phenology of the major tropical vegetation types in India, since such information is not yet available for India. Multi-temporal Medium Resolution Imaging Spectrometer (MERIS) Terrestrial Chlorophyll Index (MTCI) data were utilized to derive onset of greenness (OG) and end of senescence (ES) for four major tropical vegetation types. The study found that Fourier-smoothed results using the first four components revealed adequately the annual phenological variation of the natural vegetation types in India. From these smoothed data, inflection points were located iteratively through a spatio-temporal search, spanning over 18 months of 8-day composite data, per pixel such as to derive the OG and ES. The median OG and ES was extracted from the available annual results for the years 2003–04, 2004–05, 2005–06 and 2006–07. The GLC2000 land cover map (1 km spatial resolution) was utilized to determine the locations of the major vegetation types. The percentage of each vegetation type falling beneath a MTCI composite pixel (4.6 km spatial resolution) was calculated. MTCI composite pixels with homogeneity of ≥80% vegetative cover were used for examining pattern of phenology in different regions, different years and at different latitudes. The most common dates for the occurrence of OG for the tropical evergreen, semi-evergreen, moist-deciduous, and dry-deciduous vegetation types were found to be during February–April, January–April, March–May, and February–May, respectively. Similarly, for ES the most common dates were in February–April, January–April, February–April, and December–April, respectively. The phenological pattern was uniquely different for each vegetation type, as expected, and also differed with regions and latitudes. A general trend of early occurrence of OG in the lower latitudes was observed.