Modelling landscape ecological assessments of land use and cover change scenarios. Application to the Bojnourd Metropolitan Area (NE Iran)

Abstract Assessing the spatiotemporal dynamics of Land Use and Cover Change (LUCC) and their socio-ecological consequences on the landscape is essential for understanding the relationships between social and natural phenomena, especially in metropolitan areas. All research into the effects of metropolitan land-cover dynamics on landscape ecology should include two basic parts: LUCC simulation and Landscape Ecological Assessment (LEA). This paper proposes a novel integrated LEA-LUCC model that includes Multi-Layer Perceptron Neural Network (MLPNN), Markov Chains (MC) and Ecological Connectivity Index (ECI) able to cope with the complex and non-linear nature of metropolitan socio-ecological systems. Based on Landsat satellite images (1988, 1998, 2008, 2018) of the Bojnourd Metropolitan Area (NE Iran), and using MLPNN and MC to simulate LUCC (2028, 2038), we compare historical and future scenarios and assess their impact on ECI. Our research showed the potential of the tested models for recognizing spatiotemporal LUCC and the efficiency of remote sensing data in LUCC assessment, especially in regions lacking land-cover databases. As the outcome of the integrated models reveals, LUCC in the metropolis of Bojnourd could cause an important depression of landscape ecological functioning. The proposed LEA–LUCC model is a valuable tool for enhancing landscape sustainability in dynamic metropolitan areas and for predicting the effects of land-use planning. The LEA–LUCC model should help policy-makers and land-managers choose appropriate strategies in the socio-ecological transition of complex metropolitan systems.