A finite element model (a configuration of TELEMAC) of the Irish and Celtic seas was forced with Met Office 'rnesoscale meteorological data to investigate storm surges in the region. This is the first time that output from the Met Office meteorological model (NAE) has been used to force a finite element model of the region so that a comparison to the current UK operational storm surge forecasting model (CS3X) could be carried out. Finite element models allow hydrodynamic simulations to be carried out on a graded mesh permitting fine resolution in shallow water regions, thus removing the need for nesting whilst remaining computation ally efficient. A hindcast simulation of la typical storm surge season was carried out for October 2007 to March 2008 and the simulated surge elevations compared to observations at tide gauge stations and to those simulated by the operational model to investigate any advantage of enhanced grid resolution in shallow coastal regions for storm surge prediction. It was found that including external surge generated in regions outside the boundary of the model was important in simulating the magnitude of the observed surge at the coast. External surge generated in the wider area operational model was successfully implemented at the open boundary of the TELEMAC model so that a like-for-like comparison of the surge at the coast simulated by both models could be carried out. The results show that TELEMAC performs as well as the operational model in simulating the observed residual and skew surge elevations at various tide gauge locations at the coast. However, there is no evidence that increasing the resolution in shallow water regions increases the accuracy of the storm surge prediction based on the same meteorological forcing. Both models simulate the observed residual elevations reasonably accurately with an average RMS error of 0.14 m with respect to observations at all the tide gauges used in the study. However, both models often severely under-predict the largest observed residuals, particularly at locations in the Bristol Channel and at Liverpool. It is evident that both models fail to capture the true dynamics of surge generation in these regions and that increasing the grid resolution in shallow water regions eventually becomes limited by the resolution of the forcing meteorology and/or the bathymetry. Examining the meteorological situation it was found that model performance in both models was better at Liverpool when the surge is generated by wind stress events that are not local to the region. Increasing the wind stress by 10% showed that some of the error in predicting the largest residuals could be explained by under-prediction of the peak wind stresses. However, the largest residuals at Liverpool were still under-predicted. Therefore, to simulate local wind-generated surge residuals it is important to have increased accuracy and detail of the forcing meteorology. To investigate the factors limiting performance of enhanced resolution in shallow water regions two further seasonal simulations were carried out with a more accurate representation of the Mersey estuary, with increased bathymetric resolution, and forcing the model with the Met Office's UK4 model providing a three-fold increase in the resolution of the forcing meteorology. It was found that the bathymetry of the Mersey estuary region was not the limiting factor in accurately simulating the observed residual and skew surge elevations at Liverpool. It is evident that main channel in the Mersey estuary adjacent to Liverpool (present in both model grids of the region) is the main feature in controlling the surge distribution, and the accuracy of the simulated surge at Liverpool is more dependent on the accuracy of the increased water levels and the effect of the wind on the tide outside of the Mersey estuary in Liverpool Bay. In a further seasonal simulation using meteorological
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:569258 |
| Date | January 2011 |
| Creators | Maskell, John Henry |
| Publisher | University of Liverpool |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
Page generated in 0.0021 seconds