The Influence of Flow Projection Errors on Flood Hazard Estimates in Future Climate Conditions

Joanna Doroszkiewicz,Renata Romanowicz,Adam Kiczko

doi:10.3390/w11010049

Joanna Doroszkiewicz, Renata Romanowicz + Show 1 more

Open Access

https://doi.org/10.3390/w11010049

Copy DOI

Abstract

The continuous simulation approach to assessing the impact of climate change on future flood hazards consists of a chain of consecutive actions, starting from the choice of the global climate model (GCM) driven by an assumed CO2 emission scenario, through the downscaling of climatic forcing to a catchment scale, an estimation of flow using a hydrological model, and subsequent derivation of flood hazard maps with the help of a flow routing model. The procedure has been applied to the Biala Tarnowska catchment, Southern Poland. Future climate projections of rainfall and temperature are used as inputs to the precipitation-runoff model simulating flow in part of the catchment upstream of a modeled river reach. An application of a lumped-parameter emulator instead of a distributed flow routing model, MIKE11, substantially lowers the required computation times. A comparison of maximum inundation maps derived using both the flow routing model, MIKE11, and its lump-parameter emulator shows very small differences, which supports the feasibility of the approach. The relationship derived between maximum annual inundation areas and the upstream flow of the study can be used to assess the floodplain extent response to future climate changes. The analysis shows the large influence of the one-grid-storm error in climate projections on the return period of annual maximum inundation areas and their uncertainty bounds.

Highlights

Following the EU Floods Directive (2007) [1], flood risk maps should take into account climate changes and should be updated in cycle, every six years
Flood inundation areas derived from The the water level simulations of MIKE11 and the emulator are compared in the following subsection
Relationships between the annual maximum flow at are compared in thethe following The between the levels annual(Hmaximum flow at Ciezkowice, forming input tosubsection

Summary

Introduction

Following the EU Floods Directive (2007) [1], flood risk maps should take into account climate changes and should be updated in cycle, every six years. A standard approach to deriving flood risk maps, applied in Poland, consists of running the 1D MIKE11 model for design flood waves of different return periods (e.g., 500, 100, 10 years). Design flood wave and model simulations have been applied to estimate 100-year inundation outlines [2]. When annual flow series are short, event-based simulation approaches can be applied [3]. This type of approach uses a rainfall generator as an input to a rainfall–runoff model. Annual maximum flow events are subsequently selected from the obtained flow series and used as inputs to a 1-D hydraulic model to derive maximum inundation maps. Due to the relatively large computation time requirements of a distributed flow routing model, a continuous simulation approach is rarely applied to flood risk assessments [4]

Objectives

Methods

Results

Conclusion