Byrne, DM
(2019)
Using real time and remotely sensed data to improve operational storm surge forecasting in the tropics and mid-latitudes.
PhD thesis, University of Liverpool.
|
Text
200667610_May2019.pdf - Unspecified Download (13MB) | Preview |
Abstract
Weather conditions can generate increases in coastal sea level called storm surges. These events can lead to extensive coastal inundation resulting in the destruction of homes, infrastructure and life and have done so on many occasions in the past. Storm surge risk for coastal communities is predicted to increase globally due to climate change and sea level rise. As such, it is more important than ever that regional agencies are able to accurately forecast coastal flooding and assess risk for coastal defence and policy-making. This thesis investigates how real-time and remotely-sensed data can be used to improve operational forecasting and risk assessment. Specifically, much of this work looks at the modification of atmospheric forcing and sea surface height within operational models. The thesis begins by providing an essential background of storm surge forecasting and data assimilation. This includes details on storm surge generation, numerical modelling, operational techniques such as parameterisation of wind fields and the theory and application of data assimilation. Some of the major challenges for storm surge forecasting are also set out. The parametric representations of tropical cyclone wind fields used in operational models of the tropics are modified using analysis fields derived from remotely sensed data. Three case studies using two methods around the US are considered: Hurricane Ike, Hurricane Gustav and Hurricane Sandy. The first method simply replaces past wind forcing with available analysis fields and the second uses some simple assumptions to extrapolate them into the future. Improvements in predicted maximum surge height are achieved at most locations, reaching up to 0.27m in some cases. Extrapolating information from analysis wind fields into the future yields the best results. In the midlatitudes, we assimilate tide gauge data into a North Sea model using a new technique for dealing with coastal boundaries. We focus on a single case study: the Cyclone Xaver event of December 2013. Forecast root mean square errors are improved at most locations during the first 24 hours of forecast, in some cases up to 0.05m. However, any improvements do not persist after this period due to assimilation perturbations moving around and leaving the North Sea as a shallow water wave. In the final results chapter a novel metric for quantifying North Sea storm surges is investigated: the difference in total volume due to atmospheric forcing. It was possible to use this to identify and compare North Sea storm surges and use it to estimate storm surge persistence in the North Sea to be around 30 hours. Additionally, evidence is presented that suggests that the majority of a storm surge (in terms of sea level) is generated internally within the North Sea and that the presence of tides slows volume transport in and out of the basin.
| Item Type: | Thesis (PhD) |
|---|---|
| Divisions: | Faculty of Science and Engineering > School of Environmental Sciences |
| Depositing User: | Symplectic Admin |
| Date Deposited: | 15 Jul 2019 09:49 |
| Last Modified: | 19 Jan 2023 00:39 |
| DOI: | 10.17638/03046227 |
| Supervisors: |
|
| URI: | https://livrepository.liverpool.ac.uk/id/eprint/3046227 |
Dimensions
Dimensions
