Bayesian Modeling of Coastal Marine Environment for Sustainable Coastal Development

Abstract

To support sustainable coastal development, we propose a Bayesian method for modelingcoastal marine environments and apply the method to the Saemangeum Coast locatedalong the mid&#8208 western coast of Korea. For the purpose of economic growth, coastaldevelopment has been conducted by the construction of an approximately 33&#8208 km&#8208 long seadyke, reclaiming about 40 Kha of land since 1991. Originally, the Saemangeum Coast had awell&#8208 developed and large area of tidal flats, comprising two estuaries in addition to a chainof small islands. However, a large change in water movement and a reduction of tidalcurrents occurred following construction of the sea dyke. Severe environmental problemshave occurred over time, such as the intensification of vertical stratification, the occurrenceof red tide and oxygen depletion, coastal erosion/deposition, and the death and reductionof fishes. To monitor the environmental changes, continuous spatio&#8208 temporal oceanobservation has been performed periodically since 2002 at several sites using observationplatforms, such as the deployment of buoys, installation of towers with oceanographic andmeteorological sensors, ship surveys, and water sampling. Using the accumulatedobservational data, we developed Bayesian modeling to understand, assess, and predictcoastal marine environment and its changes, quantitatively. Furthermore, Bayesianinferen growth, coastaldevelopment has been conducted by the construction of an approximately 33&#8208 km&#8208 long seadyke, reclaiming about 40 Kha of land since 1991. Originally, the Saemangeum Coast had awell&#8208 developed and large area of tidal flats, comprising two estuaries in addition to a chainof small islands. However, a large change in water movement and a reduction of tidalcurrents occurred following construction of the sea dyke. Severe environmental problemshave occurred over time, such as the intensification of vertical stratification, the occurrenceof red tide and oxygen depletion, coastal erosion/deposition, and the death and reductionof fishes. To monitor the environmental changes, continuous spatio&#8208 temporal oceanobservation has been performed periodically since 2002 at several sites using observationplatforms, such as the deployment of buoys, installation of towers with oceanographic andmeteorological sensors, ship surveys, and water sampling. Using the accumulatedobservational data, we developed Bayesian modeling to understand, assess, and predictcoastal marine environment and its changes, quantitatively. Furthermore, Bayesianinferen