Monitoring The Distribution Of Water Hyacinth, Using Remotely Sensed Data: Case Study Of Lake Victoria, Kenya

Abstract

The objective of this project was to apply the recent Landsat 8 OLI & TRS Thematic Mapper imagery in mapping the distribution, quantify the area covered by invasive floating vegetation commonly known as water hyacinth as well as assess the changing pattern (change detection) of the weed on the Lake Victoria. Selected period of study was from 2013 to 2019, considering the month of April when the cloud cover is relatively less than 10%. The process involved pre-processing of raw downloaded Landsat 8 OLI/TIRS TM satellite images by radiometric and atmospheric correction, geometric rectification, layer stacking, and sub-setting to extract area of interest. Pre-processed images were then subjected to supervised classification by maximum likelihood, Normalized Difference Vegetation Index (NDVI) and post-classification. In supervised classification, spectral signatures for each image were obtained through data training. Both classification techniques were used to obtain water-vegetation maps followed by quantifying the changes through change detection technique. The results indicated a fluctuating but significant percentage in area occupied by floating vegetation. The highest was in 2014 and lowest in 2015 recording 9.7% and 1.9% corresponding to 33211 and 6435 hectares respectively. The change detection analysis results depicted a massive decline in floating vegetation by about 79.8%. In contrast to the table 17 which points out a strong increment of the floating vegetation by 79.913%. However the general change indicate that the floating vegetation decreased overall from 5.779% in 2013 to 4.693% in 2019. In contrast, water did not change a lot, but it increased by approximately 3.937% between 2013 and 2019.