Modeling LULC using Multi-Layer Perceptron Markov change (MLP-MC) and identifying local drivers of LULC in hilly district of Manipur, India

Abstract

Identification and prediction of future land use and land cover (LULC) changes and their drivers are required for land resources management, formulation of better policy, practicing sustainable management strategies, and modeling future LULC. The present study has focused on the prediction of future LULC and assessment of local drivers of LULC change in the hill agroecosystem Senapati district of Manipur, Northeast India. The prediction of future LULC for 2029 was achieved based on LULC data of 1999, 2009, and 2019 satellite data using multi-layer perceptron Markov change (MLP-MC) of Land Change Modeler (LCM). The socio-economic survey and field observation were performed to identify local drivers of LULC changes in the study area. The total area covered by open and dense forests in 2029 was 69.06% of the total geographical area of the district, a slight increase from the 2019 assessment of 61.01 km2. The drivers of LULC were categorized by assigning rank. The proximate drivers directly influence land use dynamics, such as increases in the settlement area, firewood collection, and construction. The underlying drivers-population growth and poverty-have also indirectly influenced LULC change in the district. It was observed that the livelihoods of local communities depend on forest products and agriculture. Therefore, the study finding will help create a future LULC information database and identification of local drivers of LULC. It will encourage the communities to participate in proper land resource management and create awareness for the sustainable use of the land.