DISCUSSION

Based on our results, we can conclude that the vegetation health on Hawaii Island has changed over time, with a more prominent decline in health being observed in the dry season than in the wet season. 

 

Our October 1977, January 1978, and January 2020 images had minimal cloud cover, meaning that our results for these images are not likely to be skewed by cloud interference. Calculating NDVI and Change Detection in ENVI for the Landsat 8 imagery (2020 images) went smoothly. We also decided to do a comparison between the wet and dry seasons upon finding relatively good imagery for both of these periods. 

 

Our October 2020 image had more cloud cover than our other images, which could have impacted the results of our dry season study. Additionally, we had to use Landsat 2 imagery for our 1977 and 1978 imagery, which is not the most conducive to doing NDVI work, but we were still able to run the tools successfully after choosing appropriate bands for the calculations. We originally tried to use Landsat 7 for our 2020 images, but due to problems with the sensor, we had to find new images from Landsat 8. Overall, the months we could use in this study were very limited due to the lack of imagery without a substantial presence of clouds. 

 

Other work studying forest change in Hawaii has been mixed. The University of Maryland’s Global Forest Change map is shown on the next page.