. Adjusted fast Fourier transform
Missing data in the data sequence of the downloaded
NDVI as a result of cloud effects commonly requires gap
filling. In addition, some errors in NDVI exits revealed as
sudden drops and rises in the data sequence. Thus, further
adjustment of the NDVI time-series was necessary. In this
study, we applied the adjusted fast Fourier transform
suggested by Sellers et al. (1994). The adjusted fast Fourier
transform of NDVI time-series is based on two assumptions