Projection of Future Changes in Landuse/Landcover using Multi-Layer Perceptron Markov Model over Akure City, Nigeria.
urbanization, MLP, landuse/landcover, landsat
Abstract
The tropical areas of West Africa including Akure city are experiencing rapid urbanization due to the increasing socio-economic growth and opportunities in the cities. As a result, the demand for space has brought about man’s alteration of the natural surface features, and this is likely to continue in the subsequent years. Urbanization processes in relation to Landuse/Landcover changes (LULCC) over Akure city were examined using Landsat TM, ETM+, and TIRS/OLI data for the periods 1986, 2000 and 2014. The Landsat images were subjected to pre-processing and classified based on the widely used supervised maximum likelihood classification scheme. Afterwards, the past and future (2028 and 2042) transitions, potential modification and extension of various land use/land cover types were carried out using Land Change Modeler (LCM) and the Multi-Layer Perceptron (MLP) Markov chain projection model. The projected LULCC indicated a substantial increase in the built-up areas from 5.04% of the area covered in 2014 to 21.72% and 26.47% by 2028 and 2042 respectively. This was evident in the corresponding decrease in the areas covered by vegetation and bare soil. The observed rural (like Ipinsa, Ibule, Shasha, Airport, NTA, etc) and sub-urban (like FUTA, Oba-Ile, Ijoka, Aule, Igoba etc) areas with abundant vegetation in the earlier periods have all experienced significant depletion and surface modifications in the latter years. The study concludes that, unabated vegetation loss and degradation could trigger serious environmental problems that are linked to increased surface thermal response, reduced infiltration and increased surface runoff.
Downloads
How to Cite
References
Published
2017-06-12
Issue
Section
License
Copyright (c) 2017 Authors and Global Journals Private Limited

This work is licensed under a Creative Commons Attribution 4.0 International License.