Analysis of vegetation seasonality in Sahelian environments using MODIS LAI, in association with land cover and rainfall

Abstract

Present-day Sahelian vegetation in a highly anthropized semi-arid region is assessed from local to regional scales, through the joint analysis of MODIS LAI (1km2 and 8-day resolutions), daily rainfall, morphopedological and land cover datasets covering the period 2000–2008. The study area is located in northwest Senegal and consists of the “Niayes” and the northwestern “Peanut Basin” eco-regions, characterized by market gardening and rain-fed cultivated crops, respectively. The objectives are i) to analyse at pixel scale LAI time series and their relation to vegetation and soil types, ii) the estimation of phenological metrics (start of season SOS, end of season EOS, growing season length GSL) and their inter-annual variability, iii) to recognize the vegetation responses to rainfall trends (mean annual precipitation, MAP; frequency of rainy events, K; combination of MAP and K, called F).Pixel-scale analyses show that LAI time series 1) describe the actual phenology (agreeing with ground-truth AGHRYMET data), and thus can be used as a proxy for Sahelian vegetation dynamics, 2) are strongly dependent on soil types. Median maps of SOS and EOS suggest an increase of the GSL from Saint-Louis to Dakar, in agreement with both the North-South rainfall gradient and the intensification of agricultural practices around Dakar. Significant correlations (R: 0.64) between annual variation coefficient of LAI and MAP for both herbaceous crops and natural vegetation are highlighted; this correlation is reinforced (R: 0.7) using the rainfall distribution factors K and F. Rainfall thresholds allowing the SOS can be defined for each type of vegetation. These thresholds are estimated at 0–5mm, 20mm and 40mm for natural herbs, herbaceous crops and shrublands, respectively.If previous works revealed the close link between the MAP and the SOS, our results highlight that LAI dynamics are also controlled by rainfall distribution during the Monsoon season. In this study, climatic indicators are proposed for estimating vegetation dynamics and monitoring SOS. Coupling Earth Observation data, such as MODIS LAI, with rainfall data, vegetation and soil information is found to be a reliable method for vegetation monitoring and for assessing the impact of human pressure on vegetation degradation.