Island ecosystems are vulnerable to the influence of goats which are invasive to most islands and are considered the most damaging for island vegetation. High grazing pressure has also been a challenge for the island of Curaçao which decided to remove goats from the Christoffelpark between the late 1980s and early 1990s. After almost 30 years of grazer absence and with no vegetation mapping campaigns in the park, vegetation maps created prior to the exclusion of goats were not representative anymore. Therefore, the vegetation was re-mapped in 2018/19 to analyse changes in vegetation. Results showed evidence of vegetation regrowth in the Christoffelpark. However, the rate of vegetation regrowth and when it first occurred in time was still unknown. This thesis aimed to gain understanding in the 30 years without vegetation data from the field and to compare effects on vegetation trends in areas where grazers are still present (Shete Boka National Park) and where grazers have been removed (Christoffelpark’s coast). Vegetation succession was found in the Christoffelpark’s coast with a decrease in coverage of grasses and herbs and an increase in coverage of trees and large shrubs. Contrary results were found in the Shete Boka park where grazing pressure is still high. Landsat satellite imagery were used time series analysis between 2000 and 2020 to get insight in vegetation trends. Based on a pixel-wise Mann-Kendall trend test, no perfect monotonous trends were found and only 10 vegetation communities showed a significant trend with 7 located in the Christoffelpark’s coast. Change detection analysis was performed to detect shifts in vegetation trends. The LandTrendr algorithm and the bfast01 algorithm showed contrary results. Considering the TWINSPAN vegetation communities, the LandTrendr algorithm detected an average of 18 breakpoints in the Christoffelpark’s coast and 13 breakpoints in the Shete Boka National Park whereas the bfast01 algorithm detected two breakpoints in the Shete Boka park and none in the Christoffelpark’s coast. The contrary results make it difficult to explain the drivers behind the detected breakpoints, especially as the period shortly after the exclusion of goats has not been part of the analysis. Further research is recommended to assess vegetation dynamics prior to 2000 as shifts in vegetation growth may have occurred more closely to the exclusion of goats in the Christoffelpark.
[MSc thesis report]