Empirical and simulated biomass, forage quality, and grazing intensity in Laikipia, Kenya

We compared grazing intensity estimated by the InVEST beta Rangeland Production model to empirical grazer density and percent green vegetation measured empirically in Laikipia, Kenya. Modeled grazing intensity was estimated through calibration to empirically measured biomass. Empirical measures were collected on 25 properties across the region, and more detailed data were collected on one property, OPC. Biomass, percent green vegetation, and animal dung were measured at 100 m transects. Biomass was measured with a pasture disk meter (PDM) which was dropped once every 10 m along the transect to estimate vegetation height. PDM measurements were transformed to dry weight biomass with a calibration coefficient estimated by regression. Green:brown ratios of grass at each transect were estimated with a pinframe containing ten pins positioned every 10 cm along a 1 m tube. The pinframe was placed at the 0, 50 and 100 m points along each transect line, and the species and color (green or brown) of any grass or forb touching each pin was recorded. Animal dung on each transect was counted if it fell within 0.5 m of the transect centerline. When possible, animal dung was identified to species; the dung of cattle and buffalo could not be reliably distinguished in the field and so their density was combined into a single bovid class. Simulated grazing intensity was estimated by the InVEST beta Rangeland Production model through calibration to empirical biomass at a given location and date. Simulated grazing intensity was compared to empirical biomass that was averaged within 2 km of weather stations on OPC and within regional properties. Dung and percent green hits on OPC were averaged across transects within months of sampling for comparison with modeled grazing intensity.