Investigating enhanced atmospheric–sea surface coupling and interactions in the Irish Sea

Abstracts

Enhanced atmospheric–sea surface coupling is investigated in the Irish Sea. The implications for so-called Proudman resonance are considered for a hindcast of an event that produced a significant, pressure-induced storm surge at the port of Liverpool. Time-series of non-dimensional gain along the depression track show resonant enhancement of the pressure-driven residual elevations in the central, deeper region of the Irish Sea when a depression moves at the speed of a shallow water wave (gh^0.5). However, in the relatively shallow eastern Irish Sea the wind stress is the dominant surge-generating mechanism. Wind-generated surge magnitude is influenced by the propagation speed of the depression which controls the timing of momentum input with respect to tidal depth variations. Large surges at Liverpool are mostly caused by an almost linear summation of the wind- and pressure-induced surge components when the wind stress acts over low water and/or the rising tide. However, it is possible for the interaction of wind stress and pressure to reduce the total surge when a pressure-induced sea-level increase reduces the effect of the wind stress. The Irish Sea is too small for significant resonant enhancement due to atmospheric–sea surface coupling and surge magnitudes are strongly dependent on the intensity of the depression and the magnitude of the wind stress.

Introduction

The Irish Sea, to the west of the UK, is a relatively shallow semi-enclosed tidal basin with fast tidal currents (up to 3 ms⁻¹) (Jones and Davies, 1998). The region is susceptible to storm surges particularly in the winter months when mid-latitude depressions propagate over the region. The risk associated with these meteorologically-driven coastal hazards could increase due to climate change with an increase in the frequency of extreme weather events. Recent research on climate projections described in the UK Climate Projections 2009 (Lowe et al., 2009) shows that the statistical significance of any trends in enhanced storminess and storm surge generation is small. However, the global (IPCC, 2007) and local (Woodworth et al., 2009) rise in mean sea-level will increase the magnitude of extreme sea levels and surges will act on a higher coastal sea level and therefore increase the risk to coastal property and infrastructure. Therefore, it is increasingly important to understand the dynamics of storm surges in the region and the relative significance of the surge-generating components in terms of their accurate prediction and the recognition of potentially damaging storm types.

Lennon (1963) suggested that major west coast storm surges are caused by Atlantic secondary depressions propagating from the Atlantic from south–west to north–east over the northern part of the British Isles at a critical speed of 20 ms⁻¹. However, not all significant events satisfy the criteria proposed by Lennon (1963) and other factors associated with a moving depression may enhance storm surge magnitudes in the region. Storm surge magnitudes in the Irish Sea can generally be strongly correlated to the forcing meteorological variables (Amin, 1982). A proportion of the surge component of sea-level at a particular location can be attributed to atmospheric pressure forcing, local wind stress and from far-field wind and pressure effects. Numerical model studies (Lowe et al., 2001, Olbert and Hartnett, 2010) have shown that the main surge-generating mechanism in the eastern Irish Sea is the wind stress, with the effective response to wind stress being inversely proportional to the water depth (see for instance, Pugh, 1987). However, the contribution of the horizontal pressure gradient may become more significant if there is a dynamical response of the Irish Sea to pressure forcing.

Proudman resonance is the resonant response of a water body to a transient atmospheric pressure system travelling at a similar speed to freely propagating shallow water waves generated by the sea-level response to pressure forcing (Proudman, 1953). The forward moving free wave cannot escape the forcing and the resonant surge grows linearly with time as long as linearity holds out (Nielsen et al., 2008). As the speed (U) of the depression approaches that of a shallow water wave, c=(gh)^0.5, there is theoretically a resonant response (see Pugh, 1987, for the full equations). The theory allows for negative surge growth when the speed of the depression exceeds that of a shallow water wave. However, in this paper we consider only positive surges that are generated by moving low pressure systems as these have the most severe implications for coastal flooding. In reality the dynamical increase in the forced long wave is decreased due to factors such as friction, diffusivity and the Coriolis force which breaks the linear assumption of the Proudman theory (Vilibić et al., 2004). However, model studies have shown that this resonant coupling of a moving air pressure disturbance and the sea surface is capable of producing large surges around the world, causing flooding and also strong currents (e.g. Vilibić et al., 2004, Hibiya and Kajiura, 1982, Mercer et al., 2002). These events have become known as “meteotsunamis” due to the fact that they have similar periods, spatial characteristics and frequency content to tsunamis caused by seismic events. These events rely on the combination of Proudman, or some other, resonance and induced harbour resonance and are only observed at certain specific locations; they take place inside bays and harbours and not at the open coast (Monserrat et al., 2006).

On the 21st January 2008 a depression passed over the Irish Sea causing storm surges (non-tidal residuals) of up to 1.2 m to be observed at tide gauge locations at eastern Irish Sea port locations such as Liverpool (see http://www.pol.ac.uk/ntslf/). A hindcast model study (Maskell, 2011) showed that when the wind forcing was removed from the solution, the remaining atmospheric pressure within the model domain generated a residual elevation of 0.4 m representing 44% of the total simulated surge elevation (Fig. 1). The study also showed that on average over a winter season (October 2007 to March 2008) the atmospheric pressure in the forcing meteorology contributed approximately 15–20%, on average, to the total residual surge elevations in the absence of both wind forcing and surge entering the region from outside the model domain. Further investigation showed that as the depression moved eastwards into the eastern Irish Sea its speed of propagation was at times close to that of a shallow water wave (c=gh^0.5) based on the water depths in the region (Maskell, 2011). Therefore, there is a possibility that in this case the increased contribution of the pressure-induced forcing was due to a dynamical response to the depression propagation. Along the depth-profile of the depression track water depths reach approximately 120 m in the western Irish Sea before decreasing to approximately 40 m in the central eastern Irish Sea. Therefore, the depression propagation speed would have to decrease from approximately 34 ms⁻¹ to 20 ms⁻¹ along the track profile to maintain a resonant response. In the original meteorological forcing it was found that as the central low of the depression was positioned over the central Irish Sea (Fig. 2a) it was propagating at a speed approximately twice the speed of a shallow water based on the water depths in the region (∼120 m) with the strongest westerly wind stresses, of up to 1.4 Nm⁻², concentrated in the south-east quadrant of the low pressure system (Fig. 2b). The horizontal pressure gradient (δP_a/δx) between the minimum low pressure and the coast showed that the pressure was increasing at a rate of 0.025 mb/km. Two hours later as the depression propagated into the eastern Irish Sea (Fig. 2c) the horizontal pressure gradient (δP_a/δx) remained unchanged but the translation speed became closer to that of a shallow water wave (∼0.8(gh)) based on the water depths in the region (∼40 m). The strongest wind stresses (acting on the sea-surface), of up to 1.2 Nm⁻²,were concentrated in the north–west and south–west quadrants of the low pressure system (Fig. 2d), veering south to south–west.

This study investigates whether a more efficient resonant response would have occurred if the depression moved at the resonant speed throughout its passage across the region and the implications for increased surge magnitudes that could cause coastal flooding. The sensitivity of the simulated surge to the characteristics of the depression, determined by the minimum low pressure and atmospheric pressure gradient, was also investigated and the effect these factors had on the original hindcast were quantified. Previous experiments in other regions have neglected wind forcing (e.g. Vilibić et al., 2004) due to the fact that it was insignificant in comparison to the pressure forcing. However, due to the fact that wind forcing is important in the eastern Irish Sea it must be included in the simulation of the total generated surge; its interaction with the pressure forcing examined since pressure-induced sea level increases could reduce the depth-integrated wind stress effect. The depression speed can also affect the wind-driven surge generation. For example, a faster moving depression may produce more resonantly-induced pressure-generated surge but a slower moving depression may produce a larger surge in the combined solution as there will be more time to set up a significant wind-generated sea-level elevation gradient in the eastern Irish Sea (Jones and Davies, 1998, Woth et al., 2006).