SIMULATING LAND COVER CHANGES AND THEIR IMPACTS ON LAND SURFACE TEMPERATURE IN AKURE, NIGERIA

Abstract

This study aims at assessing the impact of land cover dynamics on land surface temperature (LST) at past and future years in Akure, Nigeria. The objectives of the research work includes the establishment of patterns of landcover changes and land surface temperature for past years (1986, 2000, and 2014), simulation of land cover changes for future years 2028 and 2042, estimation of their relative impacts on LST in the respective periods, and examination of the changes in temporal behaviour of land surface temperature in relation to land use conversions. The study adopts landsat TM, ETM+, and TIRS/OLI imageries of 1986, 2000, and 2014 from the archives of United State Geological Survey (USGS). The data acquired were classified using supervised classification as built-up, bare soil, vegetation, and water body. Thereafter, the raw landsat data were subjected to image pre-processing before further computations. The required land surface components were computed using Surface Energy Balance Algorithm for Land (SEBAL) models. Quantitative analysis on the modification and extension of various land cover classes was carried out using Land Change Modeler (LCM). The spectral indices NDBI, NDBaI, NDVI, and NDWI were also computed to represent the various land cover classes used in this study in a more robust way. The estimated LST values were classified into thermal zones based on the estimated range of values for each year. Correlation and regression analyses were carried out between LST and spectral indices over the time periods.For future land surface component simulations, the MLP Markov chain projection model was implemented to project for the future years (2028 and 2042). Derived LST over Akure ranged between 20.86oC and 37.07oC in 1986, 23.48oC and 43.54oC in 2000, and 20.86oC and 39.83oC in 2014. The results from the trend of average LST values by landcover type in the three time periods revealed that the built-up area exhibited the highest LST, followed by bare soil, and vegetation in all the three periods. In 1986, the LST shows moderate positive correlations with NDBI and NDBaI with values 0.686 and 0.507 respectively. On the other hand, NDVI and NDWI showed high (0.857) and moderate (0.695) negative correlations with LST. Similar result was observed for 2000 and 2014. From the future simulation of land use/land cover, 5.04% of areas covered by both vegetation and bare soil in 2014 will be modified to built-up areas by 2028. Similarly, about 4.75% of the areas covered by the same landcover categories expected in 2028 will be converted to built-up by 2042. This will be manifested in the corresponding decrease in the areas covered by vegetation and bare soil. Projected LST showed that, the LST ranged between 19.79oC and 40.97oC in 2028, and 20.68oCand 44.15oC in 2042. The study found that, even though there was no increase in the Multi-Layer perceptron (MLP) Markov projected LST in the urban centres when compared with that of the earlier years, the rural and sub-urban (covered by vegetation and bare soil) areas projected as experiencing rapid urbanization processes have been observed as transitioning from a low and medium to high thermal zones in 2028 and 2042.