Gofamodimo Mashame*,a, Felicia Akinyemia TOWARDS A REMOTE SENSING BASED ASSESSMENT OF LAND SUSCEPTIBILITY TO LAND DEGRADATION: EXAMINING SEASONAL VARIATION IN LAND USE-LAND COVER FOR MODELLING LAND DEGRADATION IN A SEMI-ARID CONTEXT Gofamodimo Mashame*,a, Felicia Akinyemia aDepartment of Earth and Environmental Science, College of Sciences, Botswana International University of Science & Technology, Private Bag 16, Palapye, Botswana

LAND DEGRADATION IN PALAPYE

Land use-land cover types -Cropland -Mining -Bare land -Water bodies -Built-up Area -Paved/Rocky Material -Forest -Savanna -Shrub land -Grassland

Soil erosion by water types -Splash -Rill -Sheet -Gully -Stream Bank Figure 3: Different types of soil erosion by water (Source: Mervyn, 2015)

Aim & objectives of the study This study aims to map seasonal land use-land cover variation and assess land susceptibility to erosion by water in semi-arid Palapye area using remote sensing technology. Specific objectives are: 1) To identify and map major LULC types and assess variations in 2014 dry and rainy season 2) Assess land susceptibility to soil erosion by water in 2014 dry and rainy season

Study Location and Description

Study Location and Description The study region covers approximately 4,479 km² represented by the following geographical coordinates: A (492403.670, 7534369.134) metres, B (581956.143, 7530015.889) metres, C( 579572.223, 7480160.866) metres, D(489294.209, 7485654.247) metres in WGS 84 UTM 35S coordinate system.

Land use-land cover Mapping Methodology 1. Review of Palapye land use-land cover classes 5. Satellite image classification 9. Land use-land cover change detection 2. Land use-land cover classification system 6. Accuracy Assessment 10. Land use-land cover change prediction 3. Satellite Image acquisition 7. Post Classification 11. Land use-land cover change analysis 4. Satellite image pre-processing 8. Final Land use-land cover maps

Methodology: Land use-land cover classification No Land use-land cover class 1 Water body 2 Bare land 3 Mining Area 4 Cropland 5 Forest 6 Savannah 7 Shrub land 8 Grassland 9 Built-up Area 10 Paved/Rocky Material

Methodology: Satellite image acquisition The 30m spatial resolution LANDSAT 8 satellite data for 2014 dry and rainy season were downloaded from USGS website (http://earthexplorer.usgs.gov) for Palapye area Satellite Satellite Image Name Sensor Bands Date Acquired Season Landsat 8 LC81710762014215LGN00 Operational Land Image 2,3,4,5,6,7 03-08-2014 Dry LC81710762014311LGN00 Operational Land Imager 07-11-2014 Rainy

Methodology: Satellite image acquisition -Landsat images for 2014 dry and rainy season shown as a and b respectively (Source: USGS, 2015)

Methodology: Satellite image pre-processing Geometric correction -Geo referencing -Re projection -Image registration Radiometric correction -Atmospheric correction -Standard deviation

Methodology: Satellite image Classification Supervised classification method was used in conjunction with the maximum likelihood classifier to sort image pixels because the method is the most common in semi-arid regions and it utilizes the mean, variance, covariance of training sites (Murtaza and Romshoo, 2014).

Methodology: Accuracy assessment Ground Truthing -Stratified sampling Stratified sampling points for validating land use-land cover maps of 2014 dry and rainy season were generated as 202 and 204 respectively. The criterion for sampling was based on the accessibility of points in the real world, i.e. proximity to the road and settlements was considered (RCMRD-SERVIR Africa, 2013). The vector shapefile consisting of settlements was intersected with that of roads for the Palapye area and both datasets were buffered using a radius of 1 km. Stratified samples of reference points for ground truthing were then generated . The overall accuracy for 2014 dry and rainy season land use-land cover maps are 88.12% and 85.29% respectively

Accuracy assessment

Land use-Land Cover Change Detection Terrset software based Land Change Modeller (LCM) was used for land use-land cover change detection between 2014 dry -2014 rainy season e.g. Change detection=Later LULC – Earlier LULC= 2014 dry season-2014 rainy season The accuracy of two land use-land cover change maps will be accurate as the accuracies of the input land use-land cover maps

RESULTS

RESULTS

RESULTS

Land Degradation by water susceptibility methodology 1. Change Analysis 2. Contributors to Net Change by Bare land 3. Transition from all LULC Classes to Bare land

RESULTS

This makes croplands highly susceptible to RESULTS The highest seasonal land use-land cover variation was experienced in bare land from rainy to dry season recording 1% increase and most exchanges occurred in cropland i.e. many croplands converted into bare land in dry season This makes croplands highly susceptible to land degradation by rainfall splash and runoff The least seasonal land use-land cover variation was recorded in water body in 2014 and this is attributed to the semi-aridity of the study region.

RESULTS-Implication Poor LULC practises in semi-arid regions result in land susceptibility to degradation, especially soil degradation. With the onset of degradation, the productive quality of the land is lost over time, thus reducing the chances of attaining food security in the future.

On the basis of the rates and trends of observed land NEXT STEP Results from this study have laid the ground work for examining the susceptibility of land to LD based on seasonal variation in LULC in this semi-arid environment. A link between LULC type and LD susceptibility is suggested, which would require further examination. Therefore, next on our research agenda is to assess and quantify historical changes in LULC patterns in Palapye. On the basis of the rates and trends of observed land change, changes in LULC in the future can be projected.

and texture, topography, erosive nature of rainfall, NEXT STEP By examining the dynamics of LULC, we hope to better identify the principal signals of observed change and driving factors for Palapye. Also, we would like to examine the impact of LULC change on other types of land degradation. For example, modelling the physical and chemical properties of soil such as organic matter content and texture, topography, erosive nature of rainfall, temperature, vegetation cover by using Normalized Difference Vegetation Index (NDVI), conservation factor, livestock and human population as factors of soil erosion by water.

Conclusion It was found that 22% of bare area susceptible to degradation by water exists in semi-arid Palapye region in rainy season and bare area increased by 1% in dry season. Land degradation by water susceptibility occurred different from different land use-land cover types as follows: cropland (5.5%), paved/rocky material (9.5%), bare land (6%), built-up area (1.6%), mining area (0.07%), and water body (0.33%). The major variation of top soil vulnerability to soil erosion by water was mostly registered in cropland.

Thank You