Automated analysis of geomorphological changes in South Korea’s estuaries using remotely sensed surface water data

Nicholas Wellbrock, United States, Timothy Michael Dellapenna, Texas A & M University-Galveston Campus, Marine and Coastal Environmental Science Dept., Galveston, United States and David Retchless, Texas A&M University at Galveston, Marine Science, Galveston, TX, United States
Abstract:
As countries deal with accelerated sea level rise, climate change and expanding coastal populations, a range of estuarine engineering solutions have been applied. However, these alterations have associated feedbacks and impacts to the natural system. On the Korean Peninsula, engineered solutions include extensive tidal flat land reclamation, installation of extensive seawalls and sea dikes and the building of estuarine dams and barrages. Over the past century these practices have considerably modified the shoreline and altered both net transport of sediment and freshwater from these systems and modulated the timing and intensity of the discharge. Of the South Korean estuaries with drainage basins >500 km2, 56% contain estuarine dams, and 71% with drainage basins larger than 1000 km2. These closures have resulted in conversion of significant portions of the estuaries into freshwater lakes and the significant loss of tidal flats. In an effort to quantify these changes in a systematic manner, algorithms have been created to analyze a combination of surface water data, stream data, and geographic information systems data to automatically identify, isolate, and quantify surface water changes within South Korean estuaries. By pinpointing these changes, we can determine the extent of spatial anthropogenic impacts within each system. The current state of the algorithm is capable of delineating the majority of estuarine systems within the country, with an approximate identification rate of 90% depending upon the data resolution, water type, and temporal range of the surface water data inputted. Further modification of the algorithm and adjustment of the identification components are expected to yield an identification rate of approximately 95% for estuaries with tributaries greater than or equal to 90 meters wide.