The monitoring of short-term changes in structural characteristics of forests is important to understand mechanisms of vegetation loss that can be associated with deforestation, and illegal logging. These changes, however, should be differentiated from variations in vegetation activity due to interannual variability. Change detection based on thematic information is limited for this purpose because it depends highly on classification accuracies, and it does not allow a quantitative evaluation of biomass loss. The definition of bio-indicators associated with structural characteristics (such as, leaf area index, vegetation fraction) is at present, the only way to monitor such changes. We developed an evaluation system consisting of 4 bio-physical variables, estimated from visible red and near-infrared observations of the Enhanced Thematic Mapper, to monitor changes in forest biomass. The system is based in the application of algorithms to estimate leaf area index, the fractional vegetation cover, leaf vegetation index, and a sparse vegetation cover index, from radiometrically and atmospherically calibrated data. The algorithms were applied to individual scene images acquired during the dry season (April-May) to maximize the forest vegetation signal, and in order to identify areas of change due to changes in forest biomass rather than changes in understory vegetation conditions. The change detection analysis consisted in comparing pixel-by-pixel scenes of such variables, and the results indicated that changes in structural characteristics of forest can be monitored with Landsat-7, being leaf area index, and fractional vegetation cover the most significant in identifying changes along roadsides and population centers that indicate biomass extraction.