Comparative eﬀects of climate change and tidal stream 1 energy extraction in a shelf sea 2

The environmental implications of tidal stream energy extraction need to be evaluated 15 against the potential climate change impacts on the marine environment. Here, we study 16 how hypothetical very large tidal stream arrays and a “business as usual” future climate 17 scenario can change the hydrodynamics of a seasonally stratiﬁed shelf sea. The Scottish 18 Shelf Model, an unstructured grid three-dimensional ocean model, has been used to re- 19 produce the present and the future state of the NW European continental shelf. Four 20 scenarios have been modelled: present conditions and projected future climate in 2050, 21 each with and without very large scale tidal stream arrays in Scottish Waters (UK). It 22 is found that where tidal range is reduced a few cm by tidal stream energy extraction, 23 it can help to counter extreme water levels associated with future sea level rise. Tidal 24 velocities, and consequently tidal mixing, are also reduced overall by the action of the 25 tidal turbine arrays. A key ﬁnding is that climate change and tidal energy extraction both 26 act in the same direction, in terms of increasing stratiﬁcation due to warming and re- 27 duced mixing, however the eﬀect of climate change is an order of magnitude larger. 28

is found that where tidal range is reduced a few cm by tidal stream energy extraction, 23 it can help to counter extreme water levels associated with future sea level rise. Tidal 24 velocities, and consequently tidal mixing, are also reduced overall by the action of the 25 tidal turbine arrays. A key finding is that climate change and tidal energy extraction both 26 act in the same direction, in terms of increasing stratification due to warming and re-27 duced mixing, however the effect of climate change is an order of magnitude larger. It is now widely recognised that there is a pressing need to mitigate the effects of  This widespread concern has led to a growing interest in alternative energy sources.
of the Scottish continental shelf waters, using an implementation known as the Scottish where ρ is the water density, |u(i, t)| is the depth-averaged tidal current speed, t stands 159 for time-averaging over 30 days. APD has been estimated from a 30 days tide-only run , with 20 m diameter blades, which "weathervanes" into the tidal flow. 174 The hub height has been set to be 15 m above the bed, giving a total height of 25 m.

175
The turbine spacing is required to eliminate wake effects [ is defined as the ratio of the APD to the power density at the turbine rated speed, |u R (i)|: (2) In other words, the capacity factor is the ratio between the average instantaneous power 180 and the maximum power (rated capacity) that can be generated by a turbine. The large scale arrays have been implemented in the SSM using the momentum 205 sink approach, in which a momentum sink term represents the loss of momentum due 206 -7-Confidential manuscript submitted to JGR-Oceans where C T is the thrust coefficient, A is the area swept by the turbine and u is the flow     the climate model output shows regional-and parameter-dependent biases, for both at-307 mospheric and ocean components. Such biases will have a significant impact on processes 308 such as stratification and upwelling. Where these are non-linearly dependent on the forc-309 ing variables, the biases will not cancel when the climate change signal is calculated. An 310 alternative climate impact assessment method is the "delta-change" approach. In this 311 method, the present day climate forcing is provided by a present day reference forcing, 312 derived from the atmospheric ERA-Interim reanalysis alongside appropriate oceanic con-313 ditions (AMM7-NEMO run also forced with ERA-Interim reanalysis). This approach re-314 moves the influence of biases from the climate model forcings and preserves the mean 315 climate change signal, that is the most robust part of the signal from climate models.

316
The climate change forcing is then derived by perturbing the reference forcing with a mul-317 tiplicative (Eqs. 4-5) or an additive spatially varying correction (Eqs. 6-7), that is a func-318 tion of the future climate change forcing in relation to its present day control: -12-Confidential manuscript submitted to JGR-Oceans where φ f is any atmospheric or oceanic model variable and F M and F A are the multi-

350
The power that can be generated is dependent on the vertical cross-sectional area 351 occupied by tidal stream turbines and is the work done by the thrust force per unit of 352 time:  Table 1). The difference in the total amount of power provided 382 is mainly due to the number of turbines virtually deployed in the model (Fig. 2), that   Table 1).

388
-15-Confidential manuscript submitted to JGR-Oceans Looking to the west coast of Scotland, South West Islay and the Solway Firth ( Fig.   389 1) show equal average power per turbine (see Table 1), with the South West Islay array 390 providing more power than the Solway Firth ( Fig. 3 -central panel), due to the larger 391 number of turbines deployed (see Table 1). The Mull of Kintyre site is as energetic as 392 the Orkney Waters locations (Eday, Sanday, Westray), in terms of average and maxi-393 mum power per turbine (see Table 1). However, given the wider area considered avail-394 able for exploitation (Fig. 2), a larger number of turbines were included, leading to a to-395 tal average practical resource of 0.67 GW and a maximum of 3.40 GW. This appears to 396 be the second most energetic location in Scottish Waters. It must be noted, as for Eday, 397 that to achieve just ≈40% of the practical resource available from the Pentland Firth 398 it is necessary to increase by ≈55% the number of turbines used in the Pentland Firth.

399
However, the Pentland Firth would require turbines with a rated power on average of 400 1.5 MW (see Table 1), while turbines rated at 1 MW on average would be suitable for per turbine (see Table 1) and a smaller area to be exploited (Fig. 2). Despite the smaller  Table 1). This is due to the generic turbine design that tion is 12.85 GW (Fig. 4), which is only slightly less than summing up the maximum 420 power from each location (14.83 GW, see Table 1). This tells us that the peak power oc-  the other locations does not increase (see Table 2). The total average power available  Table 2). The maximum power does not change for the Pentland Firth, while the 447 Mull of Kintyre location shows a peak 0.76 GW larger than the tide-only estimation (see 448   Table 2), which might be connected to strong wind events during the year. As expected 449 tides are thus confirmed to be the most important available contribution to the energy 450 available from currents in these highly energetic tidal locations, with spring peak power 451 resources that can be further enhanced if in conjunction with strong wind events.

452
-18-Confidential manuscript submitted to JGR-Oceans Sea [Pugh, 1996] and moves anti-clockwise as a Kelvin wave through the entire basin.  Fig. 6g and Fig. 6h). On the other hand, the southernmost part of the Irish Sea and 528 the Dutch coast are exposed to an increase in tidal range by both tidal stream energy 529 extraction and climate change ( Fig. 6g and Fig. 6h).  (Fig. 7d), generated by the decrease in tidal range (Fig. 6d) and a consequent water depth 560 reduction and a friction increase. Of opposite sign is the change in tidal range at the north-561 ern and southern entrance of the Irish Sea (Fig. 6d), with a consequent increase of wa-562 ter depth, and a reduction of friction, that lead to a slight increase in tidal currents (Fig.   563 7d). These changes to tidal currents due to tidal turbines were found to be broadly the There are no studies available about the change to tidal currents in the North Sea 572 due to SLR. We found that changes in SLR together with consequent changes in tidal 573 amplitudes act to change the tidal currents as well. The general effect is that slightly 574 stronger tidal currents occur with SLR: increased water depth, and consequent reduced 575 friction, lead to an increase in tidal currents. Fig. 7f shows an overall increase of the or- south-east English coast, that show an increase in mean spring tidal currents (Fig. 7e), 581 that is where the increase in mean spring tidal range was also observed (Fig. 6e). For (see Fig. 7g and Fig. 7h).

608
During present climate winter conditions (Fig. 8a), the water is well-mixed over is the main driver of the winter stratification (Fig. 9a). The Norwegian Trench PEA in-651 crease is negligible in terms of percentage change (Fig. 9c). On the other hand, on-shelf west coast, a small detected decrease in PEA (Fig. 9b) can be linked to the increase in 663 mean spring currents previously observed (Fig. 7c).

664
-27-Confidential manuscript submitted to JGR-Oceans to an exacerbation of the impacts of the large turbine arrays in some limited areas (Fig.   682 9f), where changes go in the same direction of those due to climate change. Those changes 683 do not exceed 6 J/m 3 (Fig. 9f) or a 20% PEA increase (Fig. 9h), as was also found for 684 present climate conditions. The summer water column stratification generated by tidal     cations. This is relevant from the perspective of the development of marine renewable 803 energy industry, that can be seen, in some occasions, as a mitigation measure for climate 804 change, not only on a global scale, but also on a local one (e.g. coastal defence). ification on the NW European continental shelf is found to slightly increase, due to the 958 tidal velocities decrease and, as a consequence, tidal mixing. A key finding is that cli-959 mate change effects and tidal energy extraction both act in the same way in terms of in-960 creasing stratification due to warming and reduced mixing. However, the future increase 961 -40-Confidential manuscript submitted to JGR-Oceans in summer water column stratification driven by the temperature increase is ten times 962 larger and over a much wider area than the one generated by tidal stream energy extrac-963 tion during present or future climate conditions.

964
The results presented in this work are the basis for other ongoing studies that eval-965 uate the impacts of the above mentioned physical changes on animal behaviours, in par-966 ticular the distributions of mobile predator and prey species, on sediment dynamics with 967 special attention to water turbidity, and on benthic communities.

968
Acknowledgments 969 This work is part of the EcoWatt2050 project, funded by the Engineering and Physical