Exploring the Biophysical Drivers of Amazon Phenology: Preparing Data Sets to Improve Dynamic Global Vegetation Models

We explore the relative influence of biophysical drivers on phenology to assist validation and parameterization of Dynamic Global Vegetation Models. Using 6.8 years of MODIS data we created a vegetation index time series to map the spatial variability of vegetation phenology in the Amazon. TRMM and CERES data were used as a measure of two biophysical variables, precipitation and net radiation respectively. Using a Fourier transform and cross spectral analysis two aspects were considered from these data, the coincidence of: (A) spatial patterns, presence and strength in the annual cycle, and (B) the coherency and phase differences between the phenology and the biophysical variables. Using the Amazon as a study area we find that the coincidence between phenology and the drivers in annual power strength was not linear and in an area of high coherency we found radiation and phenology was almost in phase, whilst precipitation was not. The correspondence of slightly subdued annual phenology with strong annual radiation indicated that other drivers also influence the strength of the phenology.