Monitoring straw and litter masses with MODIS during the dry season in Sahel

Sahel is a huge tropical arid and semi-arid territory where pastoral activities, most of the time mobile (nomadic, transhumant) are critical for population economy. Assessing forage resources by remote sensing relies classically on the assessment of herbage growth during the 2 to 4 months of the wet season (e.g. Audaet al. this session). Then, dry herbage degradation during the dry season is left to crude models (e.g. Hiernauxet al. this session). Monitoring forage availability during the 8 to 10 months long dry season at local and regional scales is one of the challenge that remote sensing can address. Straw and litter monitoring is also important for dust emission, surface energy budget and biomass burning emissions estimates. Existing remote sensing studies dedicated to dry herbaceous vegetation detection are mainly motivated by the assessment of soil tillage intensity, of risk of soil erosion, and risk of wildfire linked to the mass of dead fuel. None of these studies have dealt with monitoring of straw and litter degradation during the dry season in Sahel. In the frame of the PAILLASAT and ECLIS projects, we have investigated several remotely sensed reflectancesand indexes to quantitatively assess the standing straws and litter masses using MODIS data from TERRA and AQUA satellites. The short-wave infrared bands (B6 centered at 1640 nm, B7 centered at 2130 nm) are especially sensitive to plant dry matter. Several MODIS derived index from literature were tested against a set of field measures carried out in 26 rangeland sites (1km-long lines) from 2004 to 2011 in the Malian Gourma. The best results are observed for index using MODIS bands from the shortwave infrared. In particular, the Soil Tillage Index (STI), ratio of B6 on B7 (Van Deventer et al., 1997), the Normalized Difference Tillage Index (NDTI), normalized difference of B6 by B7 (Van Deventer et al., 1997) and the ratio B7 on B6 (Guerschmanet al., 2009) showed good results. For example, STI explains 66% of the variance of dry masses found over a sample of sandy soils (n=126). The strong inter-annual variability is well captured. In addition, fire occurrencesmatch with marked fall of the index values, meaning that the method may be suitable even in regularly burned areas. Maps of dry herbaceous mass are shown for the Gourma, with an analysis of the role of distance to water resources on dry masses dynamics (see Diawaraet al this session) through grazing occurrence and intensity