Analysis of Land Use Change Using Object Oriented and Markov Chains in the Zilberhay Basin in East Azarbaijan and West

Abstract

The purpose of this study was to extract and explore land use changes from satellite images with object-oriented and knowledge-based models, as well as predict these changes using the Markov chain model by the 2030 in the Zailbirchai basin. Considering that the separation of some LULC from each other, such as irrigated wheat, Rain-fed Agriculture, Irrigated agriculture, and orchard in images with a spatial resolution, such as Landsat images and known pixel-based methods, is rigorous. Therefore, in the present study, a method based on the knowledge based was used for the extraction of the LULC with the OLI image of 2015 and its validation was done with kappa coefficient (0.86). Then the result of method was used in the TM image of 1987 (kappa coefficient equal to 0.83). Finally, after the analysis of the changes occurred, temporal changes was predicted for 2030. The classification of images shows the growth of residential, orchard and irrigated agriculture land uses, respectively, 15, 42 and 50 km2 over the course of 28 years. This growth reflects the high consumption of water in the region. Cross Tab, on the other hand, shows the conversion of low-consumption applications such as irrigated wheat to arable crops. Also, in the next 15 years, based on pattern of 28-year-old using Markov chain analysis model, this growth will be up to 8, 27 and 8 km2, respectively.

Keywords

• land use
• object orientation
• forecasting
• zilbirchay