Flood hazard reduction by mangroves

Abstract

Although mangrove forests have proven to contribute to wave dissipation significantly, these ecosystem are threatened mainly by the expanding aquaculture. International or nationwide applied greenbelt policies are often used for the preservation of the mangroves and hence the flood hazard reduction, without taking local (hydrodynamic) conditions into account. Therefore this research developed insights in the mangrove forests' effect on waves and their contribution to flood hazard reduction as function of forest width. This is done by setting up a globally applicable tool for determining the governing storm wave and water level conditions at the seaward edge of the mangrove forest and the wave attenuation in the forest itself. In contrary to greenbelt policies, the developed method enables the option to locally review on required mangrove forest widths. The application of the wave transformation model SWAN-2D with offshore global wave data from ERA-Interim significantly increased the prediction accuracy of nearshore wave conditions compared to the direct use of ERA-Interim. This holds mainly for the sheltered and shallow locations like embayments, estuaries and behind barrier islands, in which mangroves are located. For less sheltered and deep conditions, no improvement of the ERA-Interim data with the additional use of SWAN was observed. The attenuation in the forest depends on the highly varying physical characteristics of dimensions of the mangrove trees. The large variabilities were included in several vegetation sets, prior to their implementation into a 1D stationary XBeach model. From a comparison with wave height measurements, three characteristic mangrove species were selected. By combining these species, typical coastal interspecific patterns were composed, representing a young dense and an old sparse forest. In addition, bathymetric and hydrodynamic conditions, which are common for mangrove forests, were imposed to the model. It appeared that a 100 m dense forest is potentially able to dissipate waves to acceptable levels, while sparser forests might need 450-900 m. The model also indicated that the effect of landward deforestation on a sloped profile results in a higher wave exposure for the area behind a forest than seaward erosion, although it does not take the effect of breaking and uprooting of trees into account. The method to determine the required forest and evaluate on maintained greenbelt policies was tested on three case study locations, which differ in hydrodynamic conditions and forest characteristics. The policies appeared to result in different levels of hazard for the hinterland, underlining the great uncertainty of acceptable mangrove buffer zones for coastal zone management. The application of global (open source) data, SWAN-2D and XBeach-1D provides the set-up of a (quick) hazard assessment tool. The inclusion of local conditions improved the prediction ability compared to the conservative use of greenbelt policies.