Importance of time-series images for change detection
The figure below shows an example of why a long term trend derived from time series analysis is important to understand the true nature of change in vegetation properties. The trend observed from few discrete images could be affected by seasonal and inter-annual climatic variability even when anniversary images are available for analysis.
Figure showing (a) The average NDVI value of Eucalyptus woodland in Barakula state forest area for four anniversary date Landsat TM images and a downward trend observed from those data. (b) The same four values as the part of NDVI time series and a long term trend derived from the time series using STL algorithm.
References
Santosh Prasad Bhandari, Stuart Phinn and Tony Gill (in preparation): Monitoring temporal trajectories and trend of foliage projective cover (FPC) in forest communities using Landsat image time series in Queensland