Quantitative assessment of grassland foliar moisture parameters as an inference on rangeland condition in the mesic rangelands of southern Africa

Abstract

ABSTRACT Understanding Grass Water Content (GWC) is critical in assessing rangelands health as well as utilization rates and patterns. Spatial explicit information on grassland ecology, particularly in southern Africa remains rudimentary due to the lack of affordable data primary sources and robust user-friendly data processing algorithms. These are required to enable efficient and routine monitoring and assessment of GWC to inform policy and management. The Sentinel 2 Multispectral Instrument (MSI) in conjunction with environmental variables acquired at moderately high resolution, provide an attractive primary data source for determining grassland water content at landscape scale. This study, therefore, evaluated the utility of Sentinel 2 MSI data (e.g. Red-Edge (RE) and the Shortwave Infrared (SWIR) bands and environmental variables) in estimating GWC, herein referred to as Canopy Water Content (CWC), Dry Moisture Content (DMC), Canopy Storage Capacity (CSC), and Live Fuel Moisture Content (LFMC) in the Mesic rangelands of southern Africa. The results illustrated that both the RE and SWIR bands were critical in characterizing GWC variables in rangelands. Further, the best models for estimating CWC, DMC, CSC and LMFC yielded a relative root-mean-square error (rRMSE) of 20.8%, 19.5%, 9.3% and 18.5%, respectively. The most optimal variables for characterizing these plant selected variables included RE and SWIR bands centred at 740 nm (B6), 783 nm (B7), 1610 nm (B11) and 2190 nm (B12) as well as their derivatives in concert with environmental variables such as Topographic Wetness Index, maximum curvature and aspect. These findings are critical towards developing a robust, efficient and spatially explicit monitoring framework for southern African rangelands.