Detecting and Quantifying Desertification in the Upper East Region of Ghana using Multi-spatial and Multi-Temporal Normalized Difference Vegetation Index

Abstract

The need for process and indicator-based approach for assessing desertification is paramount to the drive to combat desertification in sub-Sahara Africa. It is in line with this that this study sought an approach based on early indicator, which can be measured in a continuum and at multispation and multitemporal scale in order to eliminate short-term phonological variations which are not direct character of desertification in its measurement. The study focused on surface vegetation change trends, as depicted by NDVI, on the African continent, Sahel Africa and the Upper East Region to detect and quantify desertification.  The study analyzed NDVI from AVHRR GIMMS NDVIg data (1982-2007) and validated that with Landsat TM5 data (1984-2007). The multitemporal and multispatial validation technique was used to analysis annual temporal and spatial mean NDVI change trends over 26 years. The study finds that temporal NDVI of the African continent has a linear relationship with the seasonal rainfall changes of the year. The mean monthly NDVI for the African region also shows greenness disparity between the northern and southern halves of the continent. It was further observed that mean annual NDVI decrease occurred between 1982 -1983, 1988, 1994, and the largest stretch of decrease expanding over 8 years occurred from 1997-2005. The rise and fall of the NDVI trend from 1982-1997 suggest regular drought on the African continent while the 8-year decrease from 1997-2005 suggests a period of desiccation. These notwithstanding, the NDVI trend of the African continent show no evidence of desertification over the study period. The mean NDVI of the Sahel Africa shows that between 1982 and 1990 NDVI was rising and falling, portraying inter-annual rainfall irregularities of the region. The period 1995-2001 saw NDVI recovering above average, with the highest mean NDVI in 2001. NDVI fell below average from 2002 to 2006. On the average, NDVI of the Sahel Africa increased by about 2.6% during the period under study.

Temporal NDVI of the UER also shows rise and fall trends similar to the Africa and Sahel Africa. However, UER NDVI increased by a higher percentage point 6.7%, as compared to the Sahel Africa average of 2.6% over the 26-year study period. The spatial analysis focused on pixels that lost and those that gained surface greenness over the reference period. The study compared three periods; 1982-1990, 1990-1999 and 1999 - 2007.An average of 19 pixels (8km²) equals 152 km² lost vegetation (1982-1990), while 280 pixels, equivalent to 2240 km², gained vegetation. The maximum gain for 1990-1999 was almost 48% of the maximum NDVI for 1987 and 2007. Although GIMMS NDVIg did not find much land degradation, the LTM NDVI shows widespread pockets of spatial degradation in the UER which were not visible in the GIMMS NDVIg. The study concluded that spatial resolution of satellite data changes land degradation dynamics observed in the analysis.

Key Words: Detecting and quantifying desertification, multi-spatial, multi-temporal.