The effect of fire disturbances on woody plant encroachment at Loskop, Irene and Roodeplaat Farms, South Africa

The study on bush encroachment goes as far back as 1917, which is ranked among the top three rangeland problems in South Africa with expected increase in affected areas. Bush encroachment is considered one of the most substantial forms of land degradation because it occurs at the expense of beneficial herbaceous layer. Even with substantial number of studies on bush encroachment, the studies have not provided a broad comprehension of the problem, which complicates its management. Climate change, fire regimes herbivory and excessive increase of CO2 in the atmosphere are some of the key drivers of the current levels of bush encroachment. It is estimated that 20 million ha of South Africa’s agricultural productivity and biodiversity is under the threat of bush encroachment. As a result, the economic productivity of affected rangelands is negatively affected. This study investigated the effects of fire frequency/history (the rate of fire occurrence over an area in a given time period) on tree density and plant diversity. It further investigates the contribution of fire to the current extent of bush encroachment using remote sensing data over a nineteen year-period from the year 2000 to 2019. The study sites are based on three Agricultural Research Council (ARC) farms namely Loskop, Irene and Roodeplaat. Firstly, the in-situ and remotely sensed moderate resolution imaging spectroradiometer (MODIS) data were used to determine how fire influences the vegetation structure (tree density and plant diversity) using Analysis of Variance (ANOVA and Kruskal-Wallis (KW-H)). Secondly, in-situ, MODIS and Landsat data were used to build models needed for mapping areas of tree density change. The study investigated the indicators of bush encroachment namely, tree density within the study sites. The study found that there is a low to moderate correlation between burned areas and tree density in Loskop, Irene and Roodeplaat farms with the Pearson correlation coefficients of -0.06, 0.38 and 0.38 respectively. The significant tree density models had moderate to relatively high R-squares of 0.59, 0.49 and 0.82 for Loskop, Irene and Roodeplaat farms respectively. The findings of this study showed that fire frequency did not significantly influence the bush encroachment as measured by tree density and diversity in Loskop and Roodeplaat farms. However, there was evidence of fire frequency significantly influencing an increase in tree density in Irene farm. Due to lack of herbivores in some parts of Loskop and Roodeplaat farms because of water scarcity, fire alone may have not been a frequent enough disturbance to significantly influence tree density. The models calculated in this study serve as a foundation for understanding and calculating the tree density in response to fire. The findings of this study serve as a guide for resource managers to better manage fire regimes and their effect on vegetation cover at a local scale. Keywords: Fire, Plant diversity, Remote sensing, Tree density