Flood Mapping: Assessing the uncertainty associated with flood inundation modelling. A case study of the Mora River, Sweden

Abstract: Expansion of cities and major infrastructure projects lead to changes in land use and river flows. The probability of flooding is expected to increase in the future as a result of these changes in combination with climate change. Hydraulic models can be used to obtain simulated water levels to investigate the risk of flooding and identify areas that might potentially be flooded due to climate change. Since a model is a simplification of the reality it is important to be aware of a model’s uncertainties. A part of this study is therefore aimed to perform a sensitivity analysis to determine which parameter has the largest impact on the model result and has to be treated more careful and accurately. In this study the 1-dimensional flow model Hydrologic Engineering Center-River Analysis System (HEC-RAS) were assed to simulate predicted water levels within the studied river. Topographic data was used to draw cross sections in Geographic Information Systems (GIS) with additional tools of HEC-GeoRAS, in order to get information about the streams geometry. The purpose of doing a sensitivity analysis was attained by investigating changes of the model results when changing different input parameters. This work is based on a reach along Mora river, in Södertälje, Sweden, as a case study. The sensitivity analysis indicate that the number of cross sections has a significant effect when simulating water levels of low flows and that the absolute error of simulated water levels increases as the average spacing between cross sections increases. The second part of the study aims to examine the effects of climate change and how it will affect water levels for the studied river. The results of the study showed that simulated water levels with flows of 100, 200 and 500 years return periods stay within the river channel and do not indicate flooded areas. The results also showed that a backwater effect due to sea level rise would affect the water levels in the stream up to a specific critical point along the studied reach. The lower reach indicated results to contain more uncertain region, where floodplain delineation changed easily as the number of cross section was changed. It is therefore important to identify the areas where uncertainties can be more critical for the results. Because of the uncertainties associated to the model used, it is important to notice that the results of this work correspond particularly to the case study in Mora River.