Investigating the effects of initial soil moisture and the uncertainty of Manning friction coefficient on flood hazard estimation and mapping.

Abstract

Floods are among the most devastating water-related hazards and are primarily responsible for the loss of human life and destruction of the natural and man-made environment. This study addresses the estimation and mapping of flood hazard in small mountain watersheds with urban areas at the lowlands and the related uncertainty. Specifically, this research studies the flood hazard for the Metropolitan city of Volos in Central Greece, which is frequently affected by intense storms that cause flash floods. The above study area is crossed by three (3) streams.The methodology used in the study is divided into three stages. At first the 24-hour design storm hydrographs were constructed for the three sub-basins of the study area with using the mean IDF parameters and the relevant confidence limits. The Alternating Block Method was used for the design hyetographs for return periods, T = 50-year, T=100-year and T=1000-year (worst-case scenario). The second stage concerns the hydrological analysis using a rainfall-runoff model. Firstly, the net rainfall was estimated by using the U.S. Soil Conservation Service (SCS-CN) method for three (3) soil's Antecedent Moisture Conditions (AMC) for dry-average-wet conditions. Then, the net rainfall was transformed by using the Instantaneous Unit Clark hydrograph into discharge and the flood hydrographs for each return period were estimated. At the final stage, the flood hydrograph estimated for each watershed was routed through the hydrographic network using the HEC-RAS 2D hydraulic-hydrodynamic simulation (2D) model.  For the flow routing, Manning’s n was estimated for various cross sections by visual inspection and corresponding values reported in international reports. The “upper” and “lower” boundaries of Manning’s n were estimated as the -50% and +50% of the average Manning’s n values, respectively. In this simulation approach, flood hazard maps for three return periods, T=50, T=100 and T=1000 years considering three different soil moisture conditions and three different values of Manning’s n have been estimated. The values of Manning’s n in the flood plain were estimated by using land cover/land use data.  The flow routing with in the urban areas was simulated by the block rising method. In total twenty-seven (27) flood scenarios have been simulated for each watershed. The results were validated with the flooded areas during a specific historical flood event using the Critical Success Index (CSI) method and reports and photographs of the historical flood event. The results of hydrological analysis and hydraulic simulation were also compared with the results of the Greek Flood Hazard Management Plans.