Detecting Spatiotemporal Expansion of Water Hyacinth (Eichhornia crassipes) in Lake Tana, Northern Ethiopia

Abstract

In the Northern Ethiopia, Lake Tana has serious problems related to the alien water hyacinth invasive weed. The weed species grows in and swims over water and expands very quickly. The purpose of this study is to detect the spatiotemporal trend of the weed during 2013, 2015 and 2017. Landsat 8 satellite images captured in December were used to assess the rate of expansion of the water hyacinth. Training samples were collected for supervised classification of the satellite images and accuracy assessment. Image preprocessing and image enhancement was carried out before classification. The expansion and distribution of the water hyacinth was analyzed using decision tree and maximum likelihood classification techniques. The classification accuracy assessment result for 2017 reported 99.5% overall accuracy and 98% of kappa coefficient. The result depicted that in 2013, water hyacinth coverage was very small in its coverage and was estimated to be 112.1 ha (0.17%) of the study area. In 2015, the coverage of the weed showed little significant change as of 2013 and grew up to 168.7 ha (0.25%). However, after two years, its aerial coverage was dynamically expanded to 1512 ha (2.25%) of the study area in 2017. The result indicated that there was rapid expansion of the weed species from 2015 to 2017s. The coverage increased by 50% in the period between 2013 and 2015 and by 82% in the period between 2015 and 2017. Since the expansion of the weed is extremely fast and endangers the existence of the water, appropriate intervention mechanisms should be urgently introduced.