Super typhoon Haiyan makes landfall
Super typhoon Haiyan
Subscribe here to receive special images


Lake Maggiore - Italy, 3-6 April 2000

Applications of VEGETATION data to resource management in arid and semi-arid rangelands

Bégué (CIRAD),G. Chehbouni (IRD), R. Escadafal (IRD)
P. Heilman (USDA), B. Mougenot (IRD), Y. Nouvellon (CIRAD/USDA), J. Qi (USDA), A. Royer (AGRHYMET), C. Watts (IMADES)

Paper (pdf file, 410 k)

The main objective of this proposal was to investigate a strategy to use the VEGETATION data to provide rangeland managers with critical information on vegetation and soil to help them in making management decision. For assessing the VEGETATION data on summer grazing, a database of vegetation measurements taken over 3 summers (97-98-99) has been built for 3 countries : The grassland plateau of southwest United States (Arizona) and northern Mexico (Sonora), mainly composed of perennial plants, and the Sahelian grassland in Niger mainly composed of annual grass.

The work is still under progress, and despite an incomplete analysis of the data, the research axis and the main results are illustrated through examples acquired over the different sites.

First, temporal ground, AVHRR and VGT time profiles and compositing techniques are presented and compared. From satellite data, assessment of floristic composition could be done for Niger, allowing the production of species maps. We then derived several time indicators (various Vegetation Indices summation, maximum, amplitude …) to assess biomass production through statistical relationships over Mexico and Niger. The atmospheric and directional effects are discussed. To improve the empirical relationships, different stratifications of Niger grassland have been tested : stratification of soil and vegetation types according to existing maps or to image classification.

To provide timely estimates of biomass during dry seasons, the US and Mexican teams have begun development of an algorithm to estimate a vegetation index that is also sensitive to senescent vegetation. By using SWIR to enhance the vegetation signals of senescent vegetation, a better estimate of total biomass appears possible. Some preliminary results showed that the use of SWIR band might provide critical information for estimating biomass in the arid and semiarid regions.

An operational algorithm was also developed in collaboration with Qi’s VEGETATION project to estimate green vegetation fraction cover at VEGETATION scales validated with multiscale TM and ETM+ data. This algorithm can provide a rapid computation of fractional cover, as a practical technique to circumvent the atmospheric effect.

The third and last research axis concerns the use of VGT data to calibrate ecosystem models developed for semi-arid grasslands. The results obtained for Mexico and US sites are mainly presented in Dedieu's investigation. Unfortunately, Niger modeling work is not achieved at this time.