Classifying land use/land cover change over time within the watershed boundary of keenjhar lake using supervised, unsupervised, and hybrid classification methods
This study determined the change of land use/land cover over time within the watershed boundary of Keenjhar Lake through the usage of remote sensing thematic classification methods, and, assessed if the changes occurring related to statements made in other studies about wetland and biodiversity decline. The three thematic classification methods utilized and compared in this study were the supervised, unsupervised, and hybrid classifications. Due to its higher overall accuracy, the supervised classification method was chosen to classify the March 07, 2020 Landsat 8 image and for post classifications from 1997 to 2002, 2002 to 2008, and 2002 to 2020. The results of this study did not support the hypotheses that there was a decline in wetland percent cover over time and that the hybrid classification was more accurate in comparison to other algorithms. The results did support the hypothesis that there was an increase in urban/agricultural land over time. The utilization of remote sensing in this study to assess land use/land cover did not indicate there was a decline in wetland size due to the 1998-2002 drought and other more recent droughts. For future studies, finer spatial resolution imagery is needed to further break down the land use/land cover classes and to determine minor changes that were not recognized in this study.