Navigating Change: Elephant Movement Response to Vegetation Dynamics in Etosha National Park, Namibia

By Jacotte Monroe

Abstract
African savanna elephants in Etosha National Park, Namibia, make use of the large expanses of savanna, foraging for 75% of the day, to sustain their large body mass. Labeled as ecosystem engineers, elephants have been seen to transform the habitats they frequent by creating new openings in vegetation. Etosha’s elephant population peaked in 2015, reaching 2,911 individuals, and has since slightly decreased to 2,355, according to a 2018 census. Such behavioral and demographic characteristics have placed the species on the national park’s priority list for conservation. To improve the effectiveness of conservation efforts, research remains to be done on how elephants respond to changes in vegetation, specifically in their movement path selection at the onset of the growing period. This study modeled the weekly relation between elephant movement and vegetation in a growing season for 12 elephants from Etosha. Furthermore, 4 elephants were studied from 2009 to 2012 to investigate inter-annual trends in the elephant movement response to vegetation. Thirdly, a focus was done on the effect of vegetation spatial detail on model interpretation, to highlight the importance of spatial scale when modeling the elephant-vegetation relation. The weekly elephant response models made use of elephant movement paths, created from series of GPS fixes, and four metrics related to MODIS Normalized Difference Vegetation Index (NDVI) at 250 m. To evaluate the effect of vegetation detail, some models were also built on MODIS NDVI data downscaled to 30 m after fusing MODIS with Landsat 8 images. Results showed that the average NDVI became increasingly influential in the elephant path selection as vegetation grew. While weekly vegetation growth itself was not a persistent driver of elephant movement. Individuals generally preferred more homogeneous landscape structures throughout the year. Overall, a gradual change in elephant response was seen to match the growing vegetation, thereby illustrating the non-stationary relation between elephants and vegetation. Meanwhile, further research should be conducted on a larger sample size than the 6 models that were built on both 250 m and 30 m data to draw conclusions on the effect of spatial resolution.