Multi-scenario LUCC modeling based on a new LMCM model: A case study of Hefei, China

Abstract

Abstract Detection of land use and land cover change (LUCC) and its future projection have become a critical issue for rational management of land resources. For this purpose, land use mapping in 2010, 2015 and 2020 in Hefei were conducted by an integrated classification approach based on spring Landsat images and digital elevation model (DEM) data, and dynamic LUCC of 2010-2015 and 2015-2020 were characterized. To predict land use change, a new comprehensive hybrid model consisting of Celluar Automata (CA) and Markov chain (M), Logistic Regression (LR) and Multi-Critical Evaluation (MCE), namely Logistic-MCE-CA-Markov (LMCM) model, was proposed to avoid the disadvantages of the previous models such as CA-Markov (CM), Logistic-CA-Markov (LCM) and MCE-CA-Markov (MCM). This new hybrid model LMCM used the fully standardized logistic regression coefficients as importance of the driving factors to represent their impact weight on each land use type. The CM, LCM, MCM and LMCM models were applied to estimate the land use pattern of 2020 based on the states of 2010 and 2015 of the study area, and we noted that the LMCM model performed better than other three versus the classified map of 2020 with a higher accuracy, that is, 1.72-5.4%, 2.14-6.63% and 2.78-9.33% higher than CM, LCM and MCM models respectively. We believed hence that the newly proposed LMCM hybrid model was capable of achieving more reliable prediction of LUCC and was employed to predict the land use and land cover (LULC) situation of 2025 within four scenarios, i.e., business as usual (BAU), economic development (ED), ecological protection (EP), and comprehensive development (CD). The results show that the LUCC modeling using the LMCM model with ED or CD scenario would be pertinent for a socioeconomic development in the study area and the approaches may be extended for such study in other regions.