254 
TEMPORAL AND SPATIAL VARIATIONS IN THE SAHELIAN RANGELANDS: 
REMOTE SENSING TECHNIQUES FOR RESOURCE MONITORING IN NIGER 
John Harrington, Jr. 
Indiana State University Remote Sensing Laboratory 
Dept, of Geography & Geology, Indiana State University 
Terre Haute, IN 47809 
Bruce Wylie and Rex Pieper 
Department of Animal and Range Science and 
Center for International Programs, New Mexico State University 
Las Cruces, NM 
Issa Denda 
Ministry of Agriculture and Animal Resources 
B.P. 12868, Niamey, Republique du Niger 
ABSTRACT 
Almost two decades of below normal precipitation receipts have had a pronounced impact on the environment 
and people of the Sahel zone in Africa. The degree of desertification that has occurred varies from year-to- 
year and from place-to-place due to the highly variable nature of Sahelian thunderstorm rainfall. As a result, 
a significant need exists annually for accurate and timely information on the current status of the environment. 
One cost-effective method for providing resource managers with some of the information they need is through 
digital image processing of NOAA AVHRR satellite data. This paper identifies the techniques currently being 
used by Projet Gestion de Ressources Fourragbres to estimate the standing crop of annual grasses at the end 
of each summer rainy season. 
An operational system exists within the Ministry of Agriculture and Animal Resources of the Government of 
Niger to produce summary statistics and maps of biomass production based on satellite data analysis each 
October. To-date, four years (1986-89) have been monitored and this resource assessment will continue. 
Estimates of biomass production are prepared using both remote sensing and GIS technologies combined with 
extensive ground truth data collection from over 30 sites across the Sahelian zone in Niger. The NOAA 
AVHRR data is converted into the Normalized Difference Vegetation Index (NDVT) and used in simple linear 
regression analyses to relate the satellite data to surface observations. Coefficients of determination for the 
four different years have ranged from over .90 to near .60. Inverse regression analysis is used to predict 
biomass for all areas of the grassland zone. The combination of image processing and GIS hardware and 
software are used to produce maps and summary statistics for individual resource management units within 
Niger. 
Results obtained to-date highlight the strong temporal and spatial variability that exists. Included within the 
four years that have been analyzed are a year of extreme drought (1987) and a year of above normal rainfall 
(1988). Due to the spotty nature of each annual rainfall pattern, the best and worst years of grass production 
at many sites do not correspond with regional averages. A few sites had good production in 1987 whereas 
others had localized drought conditions in 1988. Examination of the four years of data provide several insights 
concerning both the time frame necessary and the quality of information that must be collected for global 
monitoring of environmental changes.