LAND COVER CHANGE ACROSS BARBADOS USING REMOTE SENSING AND GIS TECHNOLOGY
Presentation Time: Wed, 08/03/2022 - 14:00
Keywords: NDVI, vegetation change, supervised classification, Barbados, Landsat 8, Caribbean
This paper focuses on the use of GIS (Geographical Information Systems) technology to determine the land cover change in Barbados between 2014 and 2021. The island has experienced drought and urban expansion over the years which has raised concern about the availability of agricultural land on the island. Data acquired from the U.S geological survey earth explorer portal for February 2nd 26th and March 2nd, 2021, were used to compute the Normalized Difference Vegetation Index (NDVI) for both years. Supervised classification using Support Vector Machines was used to determine seven (7) identified classes and their changes over the eight (8) year period. Results from the NDVI showed a general decrease in healthy vegetation from 2014 to 2021. 43.22% of the island experienced vegetation loss with 56.52% having vegetation remaining unchanged. Interestingly, only 0.26% of vegetation experienced regrowth mainly in forested areas. The validation of the supervised classification method used yielded an overall medium level of agreement with between 64% and 67% accuracy. The most significant change in land cover was from bare soil/barren land to urban areas which accounted for 23.2% change. 10.4% of grassy areas in 2014 changed to urban areas in 2021 with less than 10% change from forest to urban and agriculture to urban.
