UCLA
Remote Sensing
Vegetation Change
Vegetation Change
Vegetation Change
Results
An analysis of how agricultural productivity and vegetation changed in Yemen from 2015 to 2018 by using NDVI value calculated and Google earth explore as the visual effects.
Extent Maps
Blue: Vegetation extent of the 2018 image
Red: Vegetation loss since 2015
DATA ANALYST
Orange: Vegetation extent of the 2018 image
Blue: Vegetation gain since 2015
Band Math Difference Map
Despite the conflict caused, famine afflicting at least 17 of Yemen 26 million people, agricultural health in Al-Hudaydah, one of the nation’s key agricultural regions appears to have increased throughout the civil war based on the calculations of this study. The study area included the Red Sea’s coastal plains, including several highly cultivated bands following river valleys and wadis, and a section of the country’s central highlands which in this area are widely vegetated with shrub and dry forest. ENVI was used to perform NDVI calculation and classify the extent of vegetation for Landsat 8 data from both 2015 and 2018. Comparing the classified NDVI images it was calculated that over the whole region between 2015 and 2018, vegetation health increased by 6.86% and the extent of vegetation increased by 24.52%. Excluding the uncultivated/forested highlands, vegetation in the region expanded by 7.69%.
DATA ANALYST
An NDVI difference map was also produced for the region using raster algebra through the band math function in ENVI. The regions classified as vegetated for 2015 and 2018 were merged to produce a shapefile that was then used to subset the difference map, and to mask the data for the areas that were not vegetated in either year, as the moderate NDVI change in the desert and oceanic regions affected the calculations. These results corroborated the earlier calculations, implying that vegetation health did in fact increase through the region. Not only did overall NDVI increase, but the areas in which NDVI increased far exceeded the areas in which NDVI had decreased, even when adjusted to remove the forested region. This dataset was then further subset to repeat the calculations for smaller areas of interest, specifically the three agricultural bands easily visible from the imagery included. In all three of these regions an overall NDVI increase was found, and the area of increase outstripped the area of decrease, indicating a region wide increase in vegetation health.