Monitoring spatial-temporal changes of urban land transition using remote sensing : a case study of Lira Municipality

Abstract

Lira municipality is one of the 14 pioneer municipalities earmarked by Uganda Government’s Vision 2040 to attain regional city status. The municipal’s population was reported to be at 99,551 in 2014 and projected to reach113,500 by 2019. Urbanisation provides economic opportunities however; it also leads to emergence of unplanned urban settlements and urban sprawl. This study aims at applying remote sensing and geographical information system techniques to map land cover changes from 2000 to 2020 and predict the future land cover up to 2040. Land cover classification maps for the years 2000, 2010 and 2020 were produced in Google Earth Engine (GEE) with a total of five land cover categories including built-up areas, tree cover, grassland, wetland and cropland. Classification accuracy was estimated using the confusion matrix and the overall accuracy obtained for the land cover maps of 2000, 2010 and 2020 was 82.6%, 85.9% and 77.2% respectively. TerrSet’s Land Change Modeller and Markov Chain matrices were used to predict future changes in the urban composition. Results of urbanization analyses indicated that in 2000, 21.6% (7.44km2) of the total area was built-up. This increased to 46.6% (16.07 km2) in 2010 and 52.5% (18.12 km2) in 2020. A predicted land cover map for the year 2040 was also generated and the results showed that Lira Municipality will mostly be covered by built-up area. Built-up areas are predicted to increase by up to 6.61km2 (19.3%). Cropland, grassland, tree cover and wetland are predicted to reduce by 4.9%, 5.8%, 8.4%, and 0.3% respectively. While urbanisation continues to increase at this rate, there is need for the government to implement sustainable measures to protect the ecosystems and agricultural lands like encouraging adoption of urban or peri-urban agriculture.