Spatiotemporal Water body Change Detection Using Multi-temporal Landsat Imagery: Case Studies of Lake Enriquillo and Lake Azuei

Abstract:

One of the most valuable sources of data is Landsat imagery when in-situ data is absent. The Landsat satellite observations are also among the most widely used sources of data in remote sensing of water resources.

The purpose of this study is to investigate the water body changes of the two biggest lakes of Hispaniola Island for the past 30 years, using remote sensing techniques when there are no in-situ measurements available.

Lake Azuei in Haiti and Lake Enriquillo in the Dominican Republic both have been changing constantly in their quality and quantity. Unexpected growth of the two lakes has been observed since 2003, leaving the area with many ecological and socio-economic complications affecting thousands of local peoples’ lives during the past 12 years. Such phenomena are expected to be due to the influence of climate change on the lakes. One of the main key components to investigate this hypothesis is first to detect and map the patterns of changes of the lakes over time.

100 Landsat 4-5 TM and 192 Landsat 7-ETM+ scenes acquired from 1984 to 2014 were analyzed to investigate the surface area changes for each lake. Almost 60% of the images are fully or partially cloudy which makes it difficult to picture the full extent of the lakes and consequently calculate their surface area. Moreover, 65% of images have gaps due to the failure of the ETM+ scan line corrector (SLC) since 2003 which adds to the problem.

To solve this problem, we developed an algorithm to identify and classify clouds and cloud shadows using blue and Thermal bands; remove them from the scene and then detect water body using Normalized Difference Water Index (NDWI) using Green and NIR bands.

The next step was to fill the gaps which were created after removing clouds and stripes from the scenes. Toward this end, we decided to complete each image using the previous or next available image.

95% of the images have been processed and surface area has been calculated for both lakes. Using the algorism we conducted enabled us to map the spatiotemporal changes of both lakes for 30 years. Having such results gives us the ability to start the hydrological studies and further investigate the correlation of such phenomena with climate change.