Evaluating the effect of stone bunds erosion control on vegetation trend in South-West Burkina Faso - A fine scale remote sensing perspective in the Ioba Province
Soil erosion by water has become a worldwide issue due to its environmental and socioeconomic impact in the light of rising concerns over climate change. To minimize the impact of soil erosion by water in West Africa, several erosion control measures have been adopted and are being practiced. The type of erosion control measure practiced depends on the climatic zone in which the area falls. In South-West Burkina Faso where this study was undertaken, rainfall is relatively high compared to the other areas within the country. As a result, the use of stone bunds/lines is the most commonly practiced erosion control measure. But after the implementation of these erosion controls, very little has been done on evaluating the impact of these erosion controls on vegetation (crops and natural vegetation) improvement using remote sensing data. This is because until recently, organized erosion control measures more especially using stone bunds over thousands of hectares of both agriculture and non-agriculture lands was rare. This study, therefore, investigated the effect of stone bunds erosion control measure on vegetation trend using remote sensing data. A time series analysis of NDVI data from 2004 to 2017 was conducted to find: (i) the trend of vegetation in the whole study area and (ii) the trend of vegetation in areas with stone bunds erosion control and areas without. Subsequently, a comparison using the ANOVA test was done between the trends of NDVI in these two areas. Also, a seasonal analysis of the crop heights of cotton and millet was conducted using photographs from UAV. Lastly, a pixel-wise trend was conducted for climate variables (rainfall and temperature) and a correlation analysis was also performed between NDVI and climate variable time series. The results showed that, the NDVI trend of the whole study area is significantly increasing at a rate of 3.7 x 10-4 ΔNDVI/month at 95% confidence interval (CI). Similarly, areas with stone bunds erosion control and areas without stone bunds erosion control had significant increasing trends ranging from 3.14 x 10-4 to 3.95 x 10-4 ΔNDVI/month and 3.83 x 10-4 to 3.91 x 10-4 ΔNDVI/month respectively. In comparing the NDVI trends of the two areas, the result from the ANOVA test showed that there is no significant difference between the NDVI trends of areas with stone bunds erosion control and areas without stone bunds erosion control (p-value = 0.319). Although, the mean NDVI trends for the whole area gave a positive trend, the results of the pixel-wise analysis showed that, positive, stable and negative NDVI trends were widespread in the study area with a range of -0.001 to a maximum of 0.002 ΔNDVI/month. Only 10.6% of the NDVI trends was statistically significant at 95% CI. In comparing the crop heights in areas with stone bunds erosion control and areas without, at 95% CI, the t-test revealed that there is no significant difference between the means of the crop heights of cotton (p-value = 0.389) and millet (p-value = 0.884) in these two areas. For trends of climate variables, rainfall and temperature had a positive increase in the monthly trend of 0.12mm/month and 0.01°C/month respectively. In terms of the correlation between NDVI and climate variables, there was a positive correlation between NDVI and rainfall (Kendall τ of 0.513), whiles a negative correlation (τ = -0.322) was observed between NDVI and temperature. The results from this study will help future studies of evaluation of erosion control measures in West Africa. By combining data from other satellites such as the Sentinel, this will go a long way to help to bridge the problem of data availability for vegetation time series analysis.
A thesis submitted to the Department of Civil Engineering, College of Engineering, Kwame Nkrumah University of Science and Technology, Kumasi in partial fulfilment of the requirement for the degree of Doctor of philosophy in Climate Change and Land Use.
NDVI, Soil erosion, Socioeconomic, Environmental