Kustöversvämningar och GIS : en studie om skånska kustnära kommuners arbete samt interpolationsmetodens betydelse av höjddata vid översvämningssimulering

Sammanfattning: Global warming is causing rising sea levels due to melting of land ice and thermal expansion of water. This could lead to serious consequences around the world, such as environmental refugees. In case of a radical sea level rise, Sweden faces big challenges because many people live along the coasts. This thesis includes a study on how the Scanian coastal municipalities attack the problem of preventing coastal flooding. Their work has been explored and the results show that the municipalities are at various stages of progress. A first approach is usually to map the areas at risk in the event of sea level rise. This work is usually distributed to consultants and it is done preferably with coastal flood simulations in GIS. The possibility to simulate coastal flooding was looked at in four GIS programs, and it proved to be limited. It is possible to perform the analysis in most GIS programs but it requires great knowledge and the analysis must be performed in many steps, leading to increased risk of errors. The condition for a costal flooding analysis to be carried out is access to continuous elevation data, usually interpolated from elevation-points. It is important that the elevation data is of good quality, which requires high accuracy of the raw data but also a good interpolation. There are many interpolation methods whereof three (inverse distance weighting, kriging and topo to raster (ArcGIS specific)) were compared in this study. In total 16 interpolations were conducted, whose RMSE (Root Mean Square Error) values were calculated as a measure of accuracy. The best interpolation method was kriging, based on experimental semivariograms. All interpolations had greater RMSE values than the raw data decimeter accuracy. In addition, coastal flooding simulations were performed for all interpolation methods, for 15 levels of sea rise. The result was that the flooded areas differ substantially depending on interpolation method.