Retrospective analysis of landscape dynamics using normalized spectral entropy

Forest dynamics, such as deforestation, reforestation and afforestation, could be caused by the natural and human-induced disturbances. Traditional methods rarely quantify the history of such disturbances along a continuous temporal trajectory; yet, such disturbance history often determines system course and evolution. We adopted a retrospective perspective to study system history by implementing a newly proposed entropy index. The forest dynamics in the Amazon region, expressed by a time-series vegetation index, were taken as an example to demonstrate the application of such an entropy index. The results indicated that high entropy values were mainly distributed around the major cities with intensive land conversion rates, as well as the riverine areas with unsteady physiological and/or phenological fluctuations. Utilization of this index can measure forest dynamics in a retrospective way with accumulated information and therefore can provide a solid basis of information for current research and planning, such as for forest conservation and land-use plans.