Understanding Drivers and Pathways of Compound Flooding in Estuaries: The Importance of Estuary Shape

Abstract

Estuaries are among the most exposed parts on planet in terms of flood risk because they are locations where sea water meets inland waters. The type of flooding caused by the concurrence of high sea level, river discharge and sometimes local precipitations or groundwater breakthrough is known as compound flooding and it is in the focus of this thesis. Compound flooding has always been a severe threat to communities around estuaries, and with the climate change it is certainly going to increase both in strength and frequency. Traditionally most studied elements in examining the flood risk are drivers and sources of flooding. However, this thesis is focused on the pathways of flooding (flood defenses, floodplains, mouth types) and in the center of the research is the estuary shape.
The issue of compound flooding in estuaries was examined through the flood inundation modelling by using the LISFLOOD-FP hydrodynamic model developed by the Bristol University. Chapter 3 of this thesis is delivering deep calibration of the applied model. 20 m regular grid DEM produced in the previous studies by using data from special bathymetric survey and LiDAR was post-processed and modified to fit for the purpose of this research, while UK CEH (UK Centre for Ecology and Hydrology) land cover map was used to set varying manning roughness coefficients. Model was tested on the three different compound flood events from the 2007, 2020 and 2021 by using different values for the surface and bed roughness. Modelled results on water surface elevations were compared against the observational data from depth loggers in the Dyfi estuary with the mean standard deviation for spatially varying models being around 20 cm, and for the constant roughness models around 30 cm, except for the January 2021 event when its vice-versa. Much larger errors are observed during ebb times which is attributed to the DEM 20 m which could not effectively resolve narrow tidal channels.
Model accuracy was also examined by using the SAR satellite images of the 2021 compound flood event and results have shown that models have underestimated the inundated area by 15 % on average. Largest mispredictions have occurred at the saltmarsh area on the south of the Dyfi estuary where the 20 m DEM resolution has shown insufficient for resolving narrow tidal channels. More accurate flood extent is observed when depths less than 5 cm are excluded from the comparison with the SAR image. Additionally, slightly better results are shown when using the unsupervised image classification scheme.
In the Chapter 4 the effects of the 8 different estuary shapes or management scenarios (Hold, Advance, Remove, Retreat the line, Breach in the line (2), Deeper estuary and No spit scenario) on compound flooding in estuaries were examined. The results have shown that only nature-based solutions such as floodplain restoration and breaches in
4
defense line can mitigate projected future flooding, while the other models significantly increased water levels along the estuary. Also, the effects of different estuary shapes were compared with the effects of different behavior of the selected flood event parameters (relative timing and flood duration). It is concluded that the estuary shape has a larger role in flood control since the variability of all results is greater than the variability of the results for different flood behavior models. Looking individually to the results from various flood behavior models, the most important findings are that 1) 6 h later arrival of peak discharge to peak high water and 2) short but extremely flashy flood duration event have resulted with highest modelled water levels along the estuary, being higher even from the “worst case” estuary shape models, especially in the upper estuary.
Another roughness sensitivity test at the very end of this chapter has confirmed the sensitivity of flooding in estuaries on this parameter and has shown that flood mitigation is proportional to the magnitude of increase in the roughness value and to the size of area with the increase roughness value. Finally, in the Chapter 5, the impacts of the climate projections (UKCP18) on the compound flooding in estuaries were examined. The modelling was performed for the same set of different estuary shapes developed in the previous chapter for the three different future time slices (Baseline 1 in 20, Near future, and Far future). Two major patterns found in the results are that a) the variability in water elevations from all estuary shapes is decreasing with the climate change and b) the mean water surface elevation is markedly increasing for all shapes. In the very last part the impact of the far future flood scenario on housing in the Dyfi area is analyzed, with results indicating large >1000 % increase in the number of flooded buildings (far future). Raising the flood defenses for additional 1 m along the whole estuary has decreased the inundation by 20 %.

Date of Award	10 Mar 2025
Original language	English
Awarding Institution	Bangor University
Sponsors	The Gaynor Cemlyn Jones Trust
Supervisor	Peter Robins (Supervisor) & Martin Skov (Supervisor)