phenotemp
phenological trends using NOAA AVHRR time series
Starting from GIMMS values as ascii data (each line will be treated as individual time series), the application can smooth the ts - values by applying different algorithm that are based on Fouriertransformation. Separation of seasonal figure as well as the detection of linear trends is possible. Possible breaks in longterm mean can be detected with a change-point analysis using CuSum algorithms. Phenological events such as start-of-season, day-of-max/day-of-min, end-of-season can be determined via threshold-based algorithms or via analysis of max. increase of NDVI during green-up.
These analyses work also for multimodal phenologies where individual seasons (partly) interfere each other.
User can navigate through all steps of analysis via GUI or can define complex workflows that will be processed in batch-mode.