A spatio-temporal assessment of vegetation variability and degradation in Palapye, Botswana using normalized difference vegetation index (NDVI) time series data from 1998 to 2016

Vegetation plays a vital and important role in regulating climate, water balance and energy. This research examines vegetation variability and degradation in semi-arid Palapye, eastern Botswana. These were assessed over an 18 year period (1998 to 2016) using Remote Sensing based Normalized Difference Vegetation Index (NDVI) time series data. NDVI has been widely used as an indicator of vegetation growth because it measures the amount of plant greenness from the spectral responses measured in the red and infrared bands of the electromagnetic spectrum.The annual variation in vegetation growth cycle, derived as average NDVI from July to June the following year over the study period was assessed. Vegetation anomaly maps identifying how and where vegetation development deviates from the normal in comparison to the long-term average conditions were produced annually and for the year 2015/2016. The latter compared the amount of plant greenness for a particular month relative to the historical average of all years for that particular month. These were indicated as large increase, small increase, no change, large decrease and small decrease.

Vegetation conditions were below the average in 1998/1999 and 2002/2003 which coincided with drought stricken years. For the year 2000/2001, the increase in plant greenness was above the average conditions because of ample rainfall as this year had the highest rainfall ever recorded in Palapye.Having established the spatial and temporal trends in vegetation variability in the long and short term, levels of vegetation degradation were calculated from the time series average NDVI using the regression based Vegetation Degradation Index (VDI). A VDI map was produced and categorised into four vegetation degradation classes, namely: low degradation, medium degradation, high degradation, and relatively undegraded. Nine sites representing these varying levels of degradation were selected for validation. These sites are Mogapinyana, Moeng, Lesenepole, Mahibitswana, Malaka, Maunatlala, BIUST, Lecheng and Morupule. For example, the highly degraded nature of Lecheng was confirmed as quarries and burrow pits for mining limestone, sand, clay and stone for road construction were found. The relatively undegraded vegetation category covers 44.3% of the study area, 51.3% for the low degradation, 4.35% for medium degradation and 0.05% for high degradation.Validation of results was conducted using meteorological data, ground based vegetation data and socioeconomic data. The relationship between vegetation and rainfall in the 9 sites was investigated. That rainfall (water) is the key limiting resource constraining plant development in semi-arid regions is confirmed as the association between NDVI and rainfall at most sites was positive. Furthermore, vegetation sampling was conducted in the 9 sites to determine species composition, frequency and density in order to validate the VDI map. The aim was to relate the occurrence of plant species in the sampled sites to their level of degradation. The most abundant species are Colophospermum mopane with a percentage occurrence of 42.8% and Dicrostachys cinerea with 18.3%. For example, Colophospermum mopane which is known to be tolerant of shallow soils in hot, dry low-lying areas occurred in all 9 sites and was densest in Lecheng which represented a highly degraded area. As a rangeland degradation indicator species, Dicrostachys cinerea was densest in Maunatlala and Malaka which are medium degraded areas respectively. As dryland communities depend more on land resources for survival, the land gets degraded through interactions between thesocioeconomic and environmental systems. Focusing on farmers as the predominant land users in Palapye, face to face interviews were conducted with 136 farmers. Their responses were used in better understanding how vegetation has changed over time in Palapye. Ninety two percent (92%) of the respondents indicated the climate variables to have changed from good to worse especially the variability in rainfall.Results further show that parts of Palapye have NDVI values as high as 0.7 to 0.85 which are unusual values for its semi-arid nature. With the Botswana economy dependent on beef export and wildlife tourism, it is imperative to understand vegetation variability and degradation over time. This study provides empirical evidence of vegetation degradation and contributes to the existing limited literature on degradation at the local level. Because vegetation cover varies with time, and vegetation is prone to degradation as a consequence of biophysical and anthropogenic factors, this study demonstrates that Remote Sensing affords its monitoring on a regular basis and in a cost-effective manner.