Forest changes assessment using satellite remote sensing imagery

The synergistic use of multi-temporal and multi-spectral remote sensing data offers the possibility of long-term forest change monitoring. Due to natural and anthropogenic disturbances in Romania, forested ecosystems are undergoing accelerated change. Change detection is a remote sensing technique used to monitor and map landcover change between two or more time periods and is now an essential tool in forest management activities. We compared the ability of Multitemporal Spectral Mixture Analysis (MSMA), and maximum likelihood (ML) classification, Principal Components Analysis (PCA) techniques to accurately identify changes in vegetation cover in a south-eastern part of Romania study area between 1984 and 2004 for Landsat TM, ETM and SAR images. Fuzzy logic approach provides a mathematical formalism for combining evidence from various sources to estimate the significance of a detected change Supervised classification accuracy results were high (> 68% correct classification for four vegetation change classes and one no-change class). For spatial patterns of changes assessment, has been applied change vector analysis Classification accuracies are variable, depending on the class and the comparison method as well as function of season of the year. To solve urgent needs in application of remote sensing data, forest cover changes must be detected based on monitoring spatial and temporal regimes across landscapes. Specific aim of this paper is to assess, forecast, and mitigate the risks of forest system changes and its biodiversity as well as on adjacent environment areas and to provide early warning strategies on the basis of spectral information derived from satellite data.