|
Fourier Component Classification (FCC)
Data:
SPOT Vegetation, 10-day interval NDVI time series 2004
Objective: Cluster/split (using additional criteria including
climate, elevation) the output classes into functional vegetation groups
that are relevant for ecological assessments and for hydrologic
modeling. This will be followed by the assignment of functional
characteristics, e.g., various hydrometeorological parameters, to each
classes and to their vegetation coverages.
The
classification shows areas with vegetation covers that share identical
phenologies (identical shape of their NDVI-cycle). No post
classification processes were applied except class merging for web
display due to scale issues. Classes occupying only smaller areas
are not included in the legend. By clicking on the legend boxes all
of the reference NDVI-cycles summarized in a specific class are shown.
(Dark brown = Bare Soil, red = unclassified).
The
classification is the combined result of four classification runs of the
same data. The first run has been performed at the highest accuracy
level (high shape fidelity), for the following runs accuracy parameters
have been continuously relaxed (lower shape fidelity). The level at
which a pixel was classified is shown in the accuracy layer.
Applying
the same set of reference cycles to seperate data sets, and using the
same accuracy levels will create perfectly matching classifications.
Total
area classified: Balkash Basin, Syr-darya/Chu Rivers, Amu-Darya, Caspian
watershed, Euphrates/Tigris

|