Publication Date:
2008
abstract:
The first objective of the current work was to confirm the existence of consistent NDVI
trends in the Sahelian region by applying per-pixel linear regression analyses to an 8-km
dataset ranging from 1986 to 2000. These analyses ascertained that evident NDVI
increases took place at the peak of the rainy season (August) in most of the Sahelian
belt. Three sample areas were then selected in order to investigate the main origin of
these increases, which was identified in corresponding rainfall variations during the
previous periods, and particularly July. On this basis, per-pixel correlation analyses
were conducted to identify zones where July rainfall and August NDVI values were
concurrently increasing during the study period. A rainfall/NDVI correlation map was
thus obtained that highlighted areas where vegetation productivity was particularly
variable from year-to-year depending on the development of the rainy season. Such 'hot
spot' areas were demonstrated to correspond to those where the predictive power of
NDVI data on crop (millet) yield was highest. The paper finally discusses the possible
developments of the analytical methodology applied, as well as the environmental
relevance of the thematic maps produced.
Iris type:
01.01 Articolo in rivista
Keywords:
NDVI trends; Sahelian region; per-pixel linear regression analyses; vegetation productivity
List of contributors:
Capecchi, Valerio; Crisci, Alfonso; Vignaroli, Patrizio; Genesio, Lorenzo; Maselli, Fabio
Published in: