摘要：In this study, the change detection techniques were applied by using maximum likelihood supervised classification and post-classification on the Landsat images acquired 1997 and 2000 for the Lar Dam water level changes in Northern part of Iran. Three different classification of minimum distance, parallel piped and maximum likelihood were carried out on two different date of images individually with the aid of ground truth data. Ground truth as control points were collected during four field trips conducted between 2003 and 2007 (the Lar Dam office has helped to provide some ground truth data) which have supported with the recent recorded discharges. Using ancillary data, visual interpretation and expert knowledge of the Lar Dam basin through field further refined the classification results. Calculation of the accuracy was used to produce change image through cross-tabulation. The results revealed that the reservoir level changed about 6 km2 due to the seepage development in karsts formation during the study period.
