TY - JOUR
T1 - Analysis of land change in the dryland agricultural landscapes of eastern Botswana
AU - Akinyemi, Felicia O.
AU - Mashame, Gofamodimo
PY - 2018
Y1 - 2018
N2 - This study’s focus is on Palapye, a predominantly dryland agricultural but urbanising region in eastern Botswana. Maps of ten land use-land cover (LULC) categories at three time points (1986, 2000, and 2014) were produced based on the ISO 19144 land cover classification scheme. Land change intensities were examined at the time interval, category and transition levels using Intensity Analysis. A combination of multi-layer perceptron neural network and Markov chain analysis was used to project LULC to 2028 and investigate future changes. The rate of land change was faster during the second time interval (2000–2014) than during the first time interval (1986–2000) as the region transforms from an agrarian to a service economy. In the first time interval cropland was a net losing category, whereas it was a net gaining category during the second time interval. Cropland expanded into shrublands in the southwestern part of the study area. The built-up category was active in gains during the two time intervals as it targeted grasslands and shrublands. Built-up is also projected to gain an additional 272 km2 by 2028. The bareland and paved/rocky materials categories were also active in gains during both time intervals. A loss of 26% of natural land cover over the study period was recorded due mainly to transition into croplands, built-up areas, and grassland. The policy implications of findings are discussed as this region is important for biodiversity, ecosystem services, food production, mining, and tourism.
AB - This study’s focus is on Palapye, a predominantly dryland agricultural but urbanising region in eastern Botswana. Maps of ten land use-land cover (LULC) categories at three time points (1986, 2000, and 2014) were produced based on the ISO 19144 land cover classification scheme. Land change intensities were examined at the time interval, category and transition levels using Intensity Analysis. A combination of multi-layer perceptron neural network and Markov chain analysis was used to project LULC to 2028 and investigate future changes. The rate of land change was faster during the second time interval (2000–2014) than during the first time interval (1986–2000) as the region transforms from an agrarian to a service economy. In the first time interval cropland was a net losing category, whereas it was a net gaining category during the second time interval. Cropland expanded into shrublands in the southwestern part of the study area. The built-up category was active in gains during the two time intervals as it targeted grasslands and shrublands. Built-up is also projected to gain an additional 272 km2 by 2028. The bareland and paved/rocky materials categories were also active in gains during both time intervals. A loss of 26% of natural land cover over the study period was recorded due mainly to transition into croplands, built-up areas, and grassland. The policy implications of findings are discussed as this region is important for biodiversity, ecosystem services, food production, mining, and tourism.
U2 - 10.1016/j.landusepol.2018.03.010
DO - 10.1016/j.landusepol.2018.03.010
M3 - Article
SN - 0264-8377
VL - 76
SP - 798
EP - 811
JO - Land Use Policy
JF - Land Use Policy
ER -