A significant portion of the electrical energy required globally could potentially be supplied by harnessing wave energy and waters of the northeast Atlantic are ideal for ocean energy extraction. Wave Energy Converters (WECs) transform wave energy to mechanical energy and finally to electricity. Wave energy conversion depends highly on average energy available for extraction, direction of incoming waves (for unidirectional WECs) and wave frequency/period because WECs extract the maximum energy if they are tuned to match this frequency. These wave characteristics can be described by wave parameters like wave energy flux, mean or peak direction, and various types of wave periods. In extremely energetic sea states we have a large energy resource but there is also a risk of WEC damage and Ireland has a long history of storm waves reaching its shores. Due to its position in the Northeast Atlantic, Ireland is regularly on the path of very energetic extratropical weather systems, which can bring with them phenomenal sea-states and extreme wave conditions. For this reason, developers of ocean energy devices need to take into account that although there is a huge ocean energy resource available off the west coast of Ireland, there is a percentage of extreme events like the ones during winter 2013/2014 which may cause damage to the devices. Since the buoy data in Irish waters are sparse, both in space and in time, a hindcast properly validated against available measurements will allow developers to improve WEC design and choose adequate locations for device deployment, ensuring that device operation encounters as little down-time as possible.

The storms of winter 2013/2014 had various effects on the Atlantic coast of Europe. This winter experienced beach changes , severe gale strength winds, above average rainfall amounts and notable storm surges (Met Office: Winter storms: December, n.d.; Met Office: Winter storms: January, n.d.; Met Office: Winter 2013/14, n.d.; Met Éireann: December, n.d.; Met Éireann: January, n.d; Met Éireann: Winter, n.d.) causing a lot of coastal damage and inland flooding across the Atlantic coast of Europe. More than USD 1.8 billion worth of damage in the UK was estimated . A study by Matthews et al. (2014) showed that the winter 2013/2014 was the stormiest (they characterised storminess by taking into account the frequency and intensity of cyclones) in the last 143 years, again confirming that this winter was indeed exceptional.

For this reason, several studies have begun to examine if these extreme events are linked to anthropologically caused climate change (Masselink et al., 2016; Matthews et al., 2014; Wild et al., 2016). Masselink et al. (2016) proposed that it is tempting to assign these extreme events to climate change due to the predictions of increased storminess by climate models. A study by Wolf and Wolf (2013) concluded that the influence and impact of climate change on the northeast Atlantic waves and storms was still uncertain, while Wild et al. (2016) inferred that the large number of storms of the winter 2013/2014 could not be directly related to anthropogenic factors.

Looking into the future wave climate, a recent study using one CMIP5 climate model ensemble found that the number of low pressure systems and windstorms crossing Ireland by the end of the 21st century is projected to decrease, although not significantly, with typically 10 depressions with Mean Sea Level Pressure – MSLP < 970 hPa crossing Irish waters annually . In an additional study, focusing on extreme wave events off Ireland, Gleeson et al. (2017) found evidence of some extreme wave enhancement under the influence of a strongly positive North Atlantic Oscillation, off the west coast of Ireland. Further studies are required to try to address the current uncertainties in North Atlantic wave and storm projections and their affect on Ireland. However, it is clear that extreme wave events are likely to continue into the future and should be planned and prepared for.

This hindcast was performed with the third generation WAVEWATCH III (WW3) wave model version 4.18 with an unstructured grid . The parametrization for source terms and dissipation (Test 451) formulated in Ardhuin et al. (2010) was implemented. The bathymetry used was generated using a Digital Elevation Model (DEM) for Ireland . The DEM combined three bathymetric sources: Vector data obtained from OceanWise Ltd., the European Marine Observation and Data Network bathymetric set  and MultiBeam EchoSounders (MBES) and LIDAR INFORMAR data from the Geological Survey of Ireland (GSI) and Marine Institute (MI). EMODnet data has a resolution of approximately 500 m while MBES and LIDAR INFORMAR data has a resolution from 2 to 80 m. The unstructured triangular grid has 20 235 nodes and a varying resolution of 10 km offshore to 225 m nearshore . The limiting bottom depth for the wave model was set at 5 m.

WAVEWATCH III is driven with HARMONIE-AROME 10 m wind forcing fields and ERA-Interim wave spectra with a temporal resolution of 6 h (standard synoptic times – 00:00, 06:00, 12:00, and 18:00 UTC). The boundary feeding was set at grid nodes on segments of the open boundary (at the ERA-Interim grid points and between them) where the depth is larger than 90 m. Spectral domain is discretized in 24 directions and 30 frequencies logarithmically spaced with an increment of 1.1 from 0.0345 Hz. This coincides with the ERA-Interim wave spectra resolution. HARMONIE-AROME 10 m winds used to drive the wave model were derived from the downscaled ERA-Interim atmospheric model . The temporal resolution of the HARMONIE-AROME 10 m winds is 1 h and a horizontal grid spacing of 2.5 km.

The Polar jet stream marked the tracks for 12 major consecutive storms in winter 2013/2014: 5/14/18/24/26/27 December 2013, 3/25/26 January 2014, and 1/8/12 February 2014, according to Met Éireann . Inland and coastal flooding in Ireland and UK was a result of the combination of strong winds, tidal surges, low pressure and heavy outbreaks of rain (Met Office, 2014, 2015; Met Éireann, 2013, 2014). UK suffered land and air transport disruption, loss of power and even fatalities (Met Office, 2014, 2015). The majority of December monthly rainfall totals in Ireland were well above average except at a few stations in the West, East and North. Co. Kerry reported its wettest December on record since records began in 1949 (69 years) . Valentia Observatory recorded the lowest minimum monthly pressure values of 947.1 hPa on 27 December due to a depression of 944 hPa that was passing over the north and northwest coast. Britain, Ireland and France were under long-lasting and heavy rainfall with strong winds in the early 2014 . All the rainfall totals in Ireland for January were above their long term average rainfall. In January 2014 river and coastal flooding occurred with strong winds reaching up to 150 km h-1. Windstorm Christina (3rd to the 10th) with a low pressure centre of 940 hPa caused substantial damage to infrastructure and coastal homes in Ireland.

For brevity, we have chosen four of the storms that affected Ireland during the winter of 2013/2014: (i) “Xaver” on 5 December 2013; (ii) “Christina” on the 3rd; (iii) “Hercules” on the 5–6th; and finally (iv) a record breaking storm on 26–27 January 2014. Each of these storms is described and discussed in detail in separate sections below. We give an overview of the storms importance and weather and wave conditions on the day of the storm. We analysed time series in three locations (M4 off the northwest coast, Killard off Co. Clare, and Kinsale Energy Gas Platform off the south coast – See Fig. ) with maps of the significant wave height (Hs), peak wave period (Tp), and peak wave direction (Dp). We examined wind speeds and directions in three locations taken from The Unified Climate and Synoptic Observation Network (TUCSON): Malin Head, Mace Head, and Sherkin Island. These are Automatic Weather Stations (AWS) which are part of Met Éireann's synoptic weather observation network. These stations were chosen because they are both close to the coastline and the wave measurement locations.

We validated the model by comparing it to the buoy data available. In Figs. , , and we show the timeseries comparison at Killard, Co. Clare, the Kinsale Energy Gas Platform, and at the location of the the Irish Weather Buoy Network M4 buoy. The values of the statistical parameters for Hs at different locations are summarised in Table . These parameters are defined as following:

Pearson correlation coefficient (CC): ρ(A,B)=1n-1∑i=1nAi-μAσABi-μBσB where μA, σA, μB, and σB are the mean and standard deviation of A (model values) and B (measured values) respectively and n is the number of records.

It can be seen that the model mostly underestimates the values of Hs at the Killard wave buoy (bias = 0.34 m) and M4 (bias = 0.39 m) while slightly overestimating at Kinsale (bias = -0.13 m), indicating that the model settings could be improved to try to reduce the bias. As seen in Fig.  only 25 % of hourly records have been available for the observed period. Using the values of the Pearson correlation coefficient, we can see that we have a strong relationship between the measured and modelled Hs with a slightly more pronounced correlation at Kinsale and M4 (0.97) than at Killard (0.96). For Killard, Kinsale, and M4 we have a positive relationship with Kinsale and M4 having a higher slope closer to 1. The agreement could be improved by feeding the model with 2-D spectra from the ECMWF operation archive, and by refining the wind input and dissipation scheme used in the model. The HARMONIE-AROME 10 m wind input used to force the wave model for this study was previously validated thoroughly by comparison to over 120 land and buoy stations and altimeter data in Gallagher et al. (2016a).

This study used the WAVEWATCH III model to reproduce the events of winter 2013/2014. We simulated December and January of this winter and analysed four storms: 5 December 2013, 3, 5–6, and 26–27 January 2014. The maximum Hs height of 13.1 m was reached by the model on 26 January at 19:00 UTC, keeping in mind that the model underestimated the value of Hs by 1.5 m when compared to the observations. The model did reproduce the extreme wave events as can be seen in the validation, but underestimated the maximum values somewhat. Phenomenal sea-states occurred off the west coast of Ireland during the winter of 2013/2014, and these extreme wave events need to be taken in account not only in the stage of WEC design, deployment, and operation but also to improve our understanding and forecasting of these extreme coastal hazards. The future work will include carrying out this simulation with higher resolution winds, ECMWF operational archive 2-D wave spectra and refining of model wind input and dissipation scheme to match more closely the peaks of these powerful storms.