What are the Probabilities of Land-Use Transitions? The Answer Depends on the Classification Method

Thursday Sep 29   03:30 PM to 05:30 PM (2 hours)
Cellular AutomataLandsat imagesLand-Use/Land-Cover changesMarkov transition probabilitiesmatricesValidation of RS classification methods
Based on four time intervals within a 36-year period, we construct Land-Use/Land- Cover (LULC) maps and estimate the transition probabilities between six LULC states: built-up, agriculture, green and open spaces, transportation, and water surfaces. The LULC maps and transition probabilities matrices (TPM) were built based on the manual classification of high-resolution aerial photos and multispectral Landsat images for the same years. We considered the maps and TPM constructed from the aerial photos as a control, and compared them to those constructed from the Landsat images classified with several methods: mean-shift segmentation followed by Random Forest classification methods, and three pixel-based methods of classification: K-means, ISODATA, and maximum likelihood. For each classification the TPM were compared to the TPM constructed from the aerial photos. The goodness of fit of all maps obtained with the pixel-based methods was insufficient for estimating the LULC TPM. The LULC map obtained with the object- based classification method fit well to that based on the aerial photos, but the estimates of TMP were qualitatively different from those constructed from the aerial photos. This article raises doubts regarding the adequacy of Landsat data and standard classification methods for establishing LULC CA model rules, and calls for the careful reexamination of the entire land-use CA framework.
Tel Aviv University

My Schedule

Add to Your Schedule