Herbaceous biomass assessment in the central pastoral zone of the Sahelian country of Niger, 1986-1988
[Thesis]
B. K. Wylie
R. D. Pieper
New Mexico State University
1991
112
Ph.D.
New Mexico State University
1991
Advanced Very High Resolution Radiometer (AVHRR) Local Area Coverage (LAC) Normalized Difference Vegetation Index (NDVI) was calibrated to herbaceous biomass (kg/ha) for the central pastoral zone of the Sahelian country of Niger in 1986-1988. Ground truth sites (GTS's) were sampled 2-4 times during the rainy season for herbaceous biomass and once for species composition. Double sampling was employed using weighted least squares regressions pooled by ocular weight appraiser within sampling period to estimate herbaceous biomass. GTS's having different regressions coefficients from the pooled regression were represented by separate regression equations. GTS size varied from 100 km2 for 1986-1987 to 25 km2 for 1988. NDVI and GTS herbaceous biomass variables were pooled as a function of GTS size. NDVI variables utilized included (1) maximum value composite (MVC) NDVI for the September herbaceous biomass sample period, (2) integrated NDVI from July-August, (3) integrated NDVI from July-September, (4) integrated NDVI from minimum MVC NDVI-September, and (5) maximum NDVI for the growing season. Herbaceous biomass variables were maximum herbaceous biomass and September herbaceous biomass. NDVI variables were regressed individually on each herbaceous biomass variable using weighted least squares analysis. Integrated-NDVI data for the period July-September regressed on September herbaceous biomass data was the preferred model. The pooled 1986-1987 regression model was utilized as it was not different between years. Inverse predictions of the respective regressions were used to calibrate integrated NDVI for July-September to herbaceous biomass. Maps with 6 herbaceous biomass classes were produced for each year. Average pastoral herbaceous biomass and dry season carrying capacities were estimated for four arrondissements (counties) in central Niger. Estimated 1987 stocking rate was 34% of 1986-1988 calculated carrying capacity for the four arrondissements. Of areas having grazable forage, the percentage areas having moderate herbaceous biomass were utilized as an indicator of forage quality. Consistently high herbaceous biomass producing areas were identified and management implications indicated. Recommendations included maintaining at least 20 GTS's and testing for differences from the pooled double sampling regression for regression coefficients as well as mean estimated weights.