Assessing Long-Term Rangeland Fire Frequency And Grass Biomass Using Modis And Landsat Images

Determining rangeland fire frequency (FF) is potentially useful to rangeland managers. This paper examines the utility of historical imagery for assessing FF and grass aboveground biomass (AGB) trends, at a 4800-ha grassland-savanna nature reserve. Addition GIS overlay analysis of Boolean-coded (fire event pixel = 1, no-burn pixel = 0) respective images from a 2000-2021 MODIS burned area monthly image archive indicated FF (total fire events per 500 m pixel location). A peak-biomass, surface reflectance Landsat image series (2000, 2002, 2006, 2009, 2013, 2016, 2020, 2021) indicated grass AGB trends. Using 2020 and 2021 peak-biomass grass AGB samples from widely scattered sites, and corresponding Landsat images, a model predicting grass AGB from enhanced vegetation index values was developed and validated (R
²
= 0.523, p < 0.001, RMSE 40.9 g). Sampling site historical AGB values were then estimated. High FF sites (fire return period ≤ 6 years) experienced higher grass AGB gains.