Distribution of natural disturbance due to wave and tidal bed currents around the UK

The UK continental shelf experiences large tidal ranges and winter storm events, which can both generate strong near-bed currents. The reg- ular tidal bottom currents from tides plus wind driven ‘benthic storms’ (dominated by wave-driven oscillatory currents in shallow water) are a major source of disturbance to benthic communities, particularly in shal- low waters. We aim to identify and map the relative impact of the tides and storm events on the shallower parts of the North West European con- tinental shelf. A ten-year simulation of waves, tides and surges on the continental shelf was performed. The shelf model was validated against current meter observations and the Centre for Environmental, Fisheries and Aquaculture Science (CEFAS) network of SmartBuoys. Next, the model performance was assessed against seabed lander data from two sites in the Southern North Sea; one in deep water and another shallow water site at Sea Palling, 15 and a third in Liverpool Bay. Both waves and currents are well simulated 16 at the offshore Southern North Sea site. A large storm event was also well captured, though the model tends to underpredict bottom orbital velocity. 18 Poorer results were achieved at the Sea Palling site, thought to be due to 19 an overly deep model water depth, and missing wave-current interactions. 20 In Liverpool Bay tides were well modelled and good correlations (average 21 R – squared =0.89) observed for signiﬁcant wave height, with acceptable 22 values (average R – squared =0.79) for bottom orbital velocity. 23 Using the full ten-year dataset, return periods can be calculated for ex- 24 treme waves and currents. Mapping these return periods presents a spatial 25 picture of extreme bed disturbance, highlighting the importance of rare 26 wave disturbances (e.g. with a return period of 1 in 10 years). Annual 27 maximum currents change little in their magnitude and distribution from 28 year to year, with mean speeds around 0.04 ms − 1 , and maximums exceed- 29 ing 3 ms − 1 . Wave conditions however are widely variable throughout the 30 year, depending largely on storm events. Typical signiﬁcant wave heights 31 ( Hs ) lie between 0.5 – 2 m,


50
The UK continental shelf experiences large tidal ranges, generating periodic and 51 locally large near-bed currents, as well as winter storm events, which generate strong 52 near-bed currents and also wind waves. These 'benthic storms' are a major source of 53 disturbance for benthic communities. The impact of these disturbances will depend on 54 (i) the sediment type present (ii) bottom stress and (iii) the ability of benthic organ-55 isms to cope with displacement or a rapid accretion of sediment ( Cooper et al. 2007;  in press) while this paper focuses on the direct effect of nearbed wave and current ve-61 locities. 62 Many studies have focused on recovery of sites after anthropogenic disturbance, ei-63 ther following dredging for aggregate material, or the disposal of maintenance dredging 64 material e.g. Bolam and Rees (2003), Bolam et al. (2004). Natural disturbances also 65 cause resuspension and restructuring of soft sediments at the seabed ( Hall 1994;Levin 66 1995). If the disturbance is weak, then some fauna can 'dig themselves out' of a burial, 67 generating bioturbation but little change to the overall community. ( Cooper et al. 2007). 68 After a major disturbance the benthic community recovers mainly by re-colonisation, 69 then succession (Levin 1995). Cooper et al. (2007) identify faunal types better suited 70 to life in high-energy environments which display characteristics including rapid re-71 production, short life span and high mobility and dispersal. 72 The natural level of bottom disturbance determines which species will inhabit the 73 seabed (Hemer 2006). Herkul (2010) assesses the impacts of physical bed disturbance 74 on sediment properties and benthic communities in the Baltic Sea. Wave exposure 75 significantly affects the biomass and abundance of benthic animals, with recolonisation raising the issue that faunal recovery rates will depend on local hydrodynamics, which 79 will be very strongly affected by changing weather conditions. 80 This work is motivated by the potential impacts of natural disturbances on benthic 81 habitats and communities. We aim to identify the relative impact of tides and storm 82 events at the sea bed of the UK continental shelf by mapping the exposure over a 10- 83 year period, and calculating a representative measure of bed disturbance. The forces 84 generated by waves and tidal currents will be considered separately, before conclusions 85 are drawn about their potential impact at the bed. While the disturbance generated by 86 tides is regular and predictable, wave generated currents can be produced at the bed 87 irregularly in the form of sudden storm events. These short violent episodes can affect 88 areas of the sea-bed which are not commonly disturbed by the regular tidal currents. 89 Wave and tidal near-bed currents depend on water depth in different ways, and wave 90 induced currents (especially those generated by long period waves) regularly penetrate 91 down to the sea bed in coastal areas (Draper 1967). 92 Before moving to the core issue of bed disturbance, it is important to understand the 93 driving processes of wind-waves and tidal and surge currents. Fortunately the UK con-94 tinental shelf has been the subject of many studies of tides, waves and coastal change 95 using models and observations. The tides and hydrodynamics of the UK continental 96 shelf has been extensively studied, e.g. Flather (1976), Griffiths (1996), Jones (2002 Figure 1 shows the extent of the model domain, and the sites 143 used for model validation. The tide was simulated using the 15 tidal constituents (Q 1 , SonTek ADV-ocean-Hydra 10 minutes Waves 600 kHz RDI ADCP 100 pings every 10 minutes In order to get the bottom velocity spectrum, S u (ω) from the surface elevation 180 spectrum, S(ω), the approach of Wiberg and Sherwood (2008) is followed: The root mean square of the bottom orbital velocity is then equal to the representa-182 tive bottom orbital velocity Madsen (1994), U br , given by The surface wave spectrum can be obtained from bottom velocities by inversion of this where σ D represents the standard deviation of the data. P bias provides a measure of 206 whether the model is systematically over-or under-predicting the measured data. For 207 the Irish Sea, they find a cost function < 0.6, with P bias generally < 30% and often 208 < 10% for POLCOMS. For WAM, a CF < 0.7 is found for significant wave height 209 and P bias < 38%. Less than 10% is thought to be excellent, and 20 − 40% is good   and water-levels can also be examined.      Table 4 shows some statistical analysis of the tidal model performance, including  (Table 5).

261
The correlations and rmse are presented in Table 5, showing that the model captures  is not being used, very long swells will be underpredicted (as seen in e.g. Leake et al.

289
(2007)). The wave period T p is also found to be too short in the model, confirming

Climatology and extreme events 316
Having sampled the data set throughout the modelled period and gained some con-317 fidence in the results we now use the full simulation to produce a 10 year climatology.

318
As well as extracting an overall climatology representing mean, maximum and mini- year, as they tend to be tidally generated. Figure 7 shows the typical distribution of the defined as: By fitting a Weibull distribution to, for example, modelled significant wave height we 361 can make a prediction of the maximum Hs that can be expected at a particular point 362 within a given length of time or 'return period'. Figure 8 shows the maximum signifi- year (i.e. 100 records) we find a 0.5% error in the shape parameter and an error of 4.5%

372
in the scale parameter.

373
The extreme value approach can also be applied to the currents, but little difference 374 is seen between the 1 and 50 year return period (Figure 9), as the currents are dominated 375 by tides, and shallow water wave induced currents are not included in this simulation.

376
Tidally dominated areas, such as the English Channel, Anglesey and the East coast see  The instantaneous force (per unit cylinder height) is given by where the local instantaneous velocity at height z is u = u(t, z), the dot represents a where h c is the cylinder height and r = (a M ωD) 2 + (a D M ) 2 ). It remains to ap- The wave boundary layer is generally thin, with typical thickness δ w < 1-2 cm  and large Hs and T p were also well captured during storm events. The model performance was worst in very shallow water, due to the minimum water depth assumption 522 and models being run uncoupled and therefore unable to capture tidal modulation of 523 the wave field or wave-current interaction.

524
A modelled climatology showed certain areas to be regularly exposed to fast tidal 525 currents, which varied little year on year. The wave climatology was more spatially 526 varied, with South-West exposed coasts, and shallow water areas identified as at risk 527 from large waves.