Monitoring and prediction of land cover change in the Anambra River Basin using neural network classification and Ca-Markov model

Abstract

This study investigates the spatio-temporal variations in land cover within the Anambra River Basin of south-eastern Nigeria at three epochs: 1987, 2000, and 2018. The land cover scenario for 2030 was also predicted. Land cover was extracted using neural network classification, while the prediction was implemented with the Cellular Automata (CA) Markov chain modeling tool in Idrisi TerrSet 18.31 software. Results show that between 1987 and 2018, there was a loss of 1431.73km2 in the wetlands. In the same period, there was a gain of 214.59km2, 1,176km2, 34.72km2 and 4.19km2 in the areal extent of built-up areas, vegetation, bare lands, and water bodies respectively. This outcome could be attributed to the rise in the human population within the basin, increasing demand for agricultural land, infrastructural development, and housing. The land cover projection between 2018 and 2030 shows a loss of 0.79km2, 12.62km2, and 4.96km2 in the water bodies, wetlands, and bare lands respectively. In comparison, there was a gain of 23.6km2 in vegetation. It is recommended that sustainable conservation practices and good land cover management policies be established to safeguard the river basin.