Sina NabiZadeh; Ataollah Ebrahimi; Masoumeh Aghababaei; Iraj Rahimi
Abstract
The land use of the watersheds is one of the most affected and highly vulnerable due to developmental process which effect on the other variables such as the hydrological function. The purpose of this research is to monitor land use changes in the past and to investigate predictability of its future ...
Read More
The land use of the watersheds is one of the most affected and highly vulnerable due to developmental process which effect on the other variables such as the hydrological function. The purpose of this research is to monitor land use changes in the past and to investigate predictability of its future using Land Change Modeler (LCM) in the watershed of Farsan County of Chaharmahal-va-Bakhtiari province. For this purpose, the Landsat-5 TM images of 1986 and 2009 as well as the Landsat-8 OLI images of 2017 were analyzed. Land covers including residential areas, agricultural land, dryland farming, rangelands, rocks, water bodies, bare-land and snow were classified for the three periods. The prediction of land cover of 2017 was done using the LCM model based on Artificial Neural Network and Markov chain analysis after assessing model’s accuracy based on Kappa index. The land cover of 2027 was also predicted using a change probability table extracted from occurred changes from 1986-2017. The results show that the rangeland decreased by 4379-ha in the years 1986 to 2017, but the agricultural land increased by 1922-ha. This study proved that the LCM could accurately forecast future changes (85% overall accuracy). An increase of 149-ha of residential area and 100-ha decrease of rangelands in the study area was predicted for 2027. Therefore, while emphasizing the conservation of rangelands, it is necessary to study and use this technique to predict changes, its causes, as well as the consequences of land use changes at the broader scales.
