Laila Mouakkir,Mohamed El hou,2,Soumia Mordane and Mohamed Chagdali

Received:12 January 2022/Accepted:03 March 2022

?Harbin Engineering University and Springer-Verlag GmbH Germany,part of Springer Nature 2022

Abstract In the last two decades,the exploitation of marine renewable energies(70%of the globe is made up of oceans),especially wave energy, has attracted great interest, not only for their high potential, but also for their high energy density. The development of wave energy is suitable for countries or regions with extensive coastline and high waves approaching the shore. This paper focuses on the study of wave potential and wave energy distribution in the Casablanca-Mohammedia nearshore area (Moroccan Atlantic coast) in order to identify prospective wave energy hotspots. To achieve this purpose,the offshore wave potential was firstly estimated from a 20 years wave data provided by ECMWF (European Center for Medium range Weather Forecasts).In the second step,a numerical modeling of the wave propagation in the study area was performed using the SWAN model jointly with WAVEWATCHIII. The performance of the model to simulate accurately the wave field was evaluated in a real situation characterized by large waves.The model then was applied to determine the patterns of wave field in the Casablanca-Mohammedia nearshore area for a typical wave conditions (winter, summer and storm).The results of this study show the abundance of wave energy in the region with an average annual wave potential of about 22 kW/m.A seasonal variability of the wave resource was demonstrated, with values five times higher in winter than in summer.In addition,a major hotspot site was identified that should be considered when studying WEC implementation.This hotspot is located at the southern edge of the Casablanca-Mohammedia coast,near the coastal area of Sidi Rahal.

Keywords Wave energy; Numerical wave modeling; SWAN model; WAVEWATCHIII model; Casablanca-Mohammedia coast(Morocco)

1 Introduction

Electricity consumption in the world is mainly due to the growth of economic activity, industrialization and pro‐duction techniques, increasing population, and welfare im‐provement. Morocco, like most countries in the word, is experiencing a significant growth in electricity consump‐tion,with an average rate of 6.9%per year(Office Nation‐al de l’Electricité. ONE 2010). This increase necessitates the exploitation of renewable and sustainable energy re‐sources.

Nowadays,the energy sector in Morocco depends heavi‐ly on imports, about 90% of its energy needs are imported(Aparicio 2014).In this respect,the Moroccan government has put in place an ambitious energy plan, called the Na‐tional Energy Strategy (NES) (Alhamwi et al. 2015). Its main objective is to increase the input of renewable energy resources into the national grid to 52%of electricity gener‐ated from renewable energy sources by 2030(IRENA 2014;Alhamwi et al.2015).

The NES has identified wind, hydro and solar as “high priority”resources to establish an ambitious renewable en‐ergy program. However, the presence of a coastline of 3 500 km, the country has a high potential of marine re‐newable energy (Alaoui 2019), which makes it a country with several sites for the exploitation of marine renewable energy(MRE).

Over the last few years, a big interest has been given to MRE (70% of the earth surface is covered by ocean).Among the potential sources of MRE,wave energy is con‐sidered as one of the energy sources with the greatest po‐tential for development over the next few years(Iglesias et al. 2009). It may be seen as a condensed form of ocean wind energy (Akpinar et al. 2017). Wave energy develop‐ment is suitable for countries or regions with extensive coastline and strong waves approaching the coast (Zhang et al.2009).

According to(Sierra et al.2016),the average wave pow‐er obtained along the Moroccan Atlantic coast is consider‐able (up to 30 kW/m, which is equivalent to annual wave energy of 262 MWh/m) and slightly lower than that of the neighboring Canary Islands.The spatial distribution of wave energy shows great variability, with a high energy stretch in the central part of this area (between latitudes 29.3°N and 34°N). The present work focuses on the anlysis of wave energy resource in the Casablanca-Mohammedia coast area (Moroccan Atlantic coast). The wave energy was characterized thoroughly in terms of wave directions,heights and periods based on numerical modeling, and us‐ing ECMWF(European Center for Medium range Weather Forecasts)ERA5 data-set.

Due to the advances in computer science, wave energy assessment uses numerical models capable of accounting for the physics of wave propagation in the deep and shal‐low waters. Within this context, spectral wave models such as Mike 21 (Vannucchi et al. 2012; Venugopa et al.2017), SWAN (Simulating WAve Nearshore) (Bento et al. 2018; Carballo et al. 2014; Kamranzad et al. 2013;Amarouche et al. 2020), and WAVEWATCHIII (WWIII),WAM (WAve Model) (Rusu and Guedes Soares 2012;Zheng et al. 2013; Waters et al. 2009), and TOMAWAC(Dong et al.2020;Guillou 2015),have been applied to as‐sess and forecast global coastal wave resources at various locations.

To date, a very few studies explored the wave energy potential around the Moroccan waters.

As far as the authors know, the existing studies investi‐gated the wave energy availability on a large scale using a deep water models (WAM and WAVEWATCHIII) (Bouhrim and El Marjani 2019;Sierra et al.2016).The main novelty of this work is to study the wave energy on a local scale by the use of state-of-the-art coastal model SWAN. The studied area is the most important coastal stretch, in an economic view,and the most populated region of the king‐dom.The main goal of the work is to prospect the most fa‐vorable areas (hotspots) for implementing wave energy farms.

To achieve this objective, the SWAN model was imple‐mented to simulate the wave parameters during its propa‐gation in the Casablanca-Mohammedia nearshore area.The SWAN model was previously used in many research studies in order to identify the wave energy hotspots sites(Iglesias and Carballo 2010; Lin et al. 2019; Majidi et al.2020;Vannucchi and Cappietti 2016).

2 Wave models and data description

2.1 The study area

Morocco has an interesting geographical position. It is located at the crossroads between Africa and Europe(32°00'N 5°00'W).It is bordered by the Mediterranean and the Atlantic seas. The region concerned by the simulation includes the site of Casablanca-Mohammedia located at the northwest of Morocco on the Atlantic coastline. This coastal stretch has a NW-NE orientation, and is conse‐quently exposed to waves coming from NW sector. The model zone extends longitudinally from 7°2'W to 8°W,and at latitude from 33°45'N to 34°N.This area is charac‐terized by a regular bathymetry showing a slope going down to the sea. The isobath is appreciably parallel to the coast, they vary from 5 m at the coast and 269 m offshore(Figure 1).This quasi-regular isobaths configuration is per‐turbed in the extreme south by the presence of submarine ridge extending about 11km from coastline. Moreover, we notice the presence of three very marked capes (Cap Dar Bouazza, cap Elhank, cap Fedala), these latter are elongat‐ed seawards by submarine ridges (D1, D2, D3 and D4).There exist also six shoals (H1, H2, H3, H4, H5, and H6)(Mouakkir et al. 2010) where water depth is nearly 28 m(Figure 1).

2.2 Modeling tools

2.2.1 SWAN model description

In order to force SWAN with a real wave boundary conditions, WWIII model was set up for the north atlantic area(Figure 2).It is a third-generation wave model devel‐oped at NOAA/NCEP (National Oceanic and Atmospheric Administration/National Centers for Environmental Predic‐tion)(TheWAVEWATCHIIIDevelopmentGroup,2019).

The analysis of the wave energy potential in the present study was developed based on the third generation windwave model SWAN version 41.20 (Booij et al. 1999;Mouhid et al. 2020; SWAN Team 2018). It is a numerical wave model that provides realistic estimates of wave pa‐rameters in open seas, coastal areas, lakes, and estuaries from given wind, bottom, and current conditions (Booij et al.1999).The SWAN model computes the development of a sea states by means of the wave action densityN(σ,?) ,which is defined by the ratio of the spectral energy densityE(σ,?) to the intrinsic representative wave frequencyσ. In SWAN model, the evolution of the wave spectrum is de‐scribed by the spectral action balance expressed by the fol‐lowing equation:

Figure 1 Study area location

Figure 2 WAVEWATCHIII computational domain with ETOPO1 bathymetry used to force the model,water depths are in meter

wherexandyare horizontal Cartesian coordinates,tis the time,θis the propagation direction of each wave compo‐nent,Cx,Cy,Cσ, andCθare the propagation velocity inxspace,y-space,σ-space, andθ-space respectively, andStotrepresents the energy source.The first term on the lefthand side of Equation (1) is the rate of change of action density in time,the second and third terms are the propaga‐tion of action in physical space. Term 4 represents the shifting of the relative frequency due to the variation of depth and current. Term (5) represents depth-induced and current-induced refraction.

The itemStoton the right side of this equation includes dissipations, quadruplet interaction, and triad interaction caused by wind input, white-capping, bottom friction, and depth-induced wave breaking. Details of these processes can be found in the SWAN technical documentation manu‐al(SWAN Team 2018).

2.2.2 Wave energy flux

The deformation of the surface water caused by wind and the transmitted water particles, give the swell a me‐chanical Energy. This energy is the sum of the potential and kinetic energy;it is considered as the total energy Con‐tained in a wavelength per linear meter of Crest (Bon‐nefille 1992;Holthuijsen 2007).This energy is expressed by:

The wave power, flux of energy per unit crest, in terms of wave spectrum can be defined using the spectral output of the numerical wave model as energy flux per wave crest width unit in kW/m (Cornett et al. 2008; Sheng et al.2017)as:

wherexandyrepresent the problem coordinate system,CxandCythe propagation velocities of wave energy in the geographical space.The absolute value of energy transport(denoted also as wave powerPw)will be:

In deep water condition (d>0.5L), the Equation (2) be‐comes(Cornett et al.2008;Randi et al.2012):

The unit of wave power in Equation (6) is Kilowatts(kW) per meter (m) of wave front length as the significant wave height is given in meters and the wave period is in seconds.In this study,the Equation(6)is used to calculate the wave power.

ρis the seawater density taken as 1 027 kg/m3,gis the ac‐celeration of gravity,Hsis the significant wave height andTeis the wave energy period.HsandTeare defined as:

The mean wave direction may be obtained from the di‐rectional wave energy spectrum through:

whereE(ω,θ) is the variance density spectrum andωis the absolute radian frequency equaling to the relative fre‐quencyσwhen no ambient currents are present, it can be determined by the Doppler-shifted dispersion relation in the presence of the ambient current.

2.3 Forcing data and models setup

The model bathymetry is based on the ETOPO (www.ngdc.noaa.gov/mgg/global) global bathymetric data. ETO‐PO are the digital terrain elevation data released by the United States NGDC, including data about land elevation and sea bottom topography.

Figure 1 shows also the ETOPO bathymetry of the con‐sidered area.The wave climate in the general area of Cas‐ablanca–Mohammedia was drawn up using ECMWF(Eu‐ropean Center for Medium range Weather Forecasts)ERA5 data-set (Hersbach et al. 2020). This global dataset provides an accurate and reliable time series describ‐ing the past wave climate and used in many studies for wave energy assessment on the global and regional scales(Amrutha et al. 2020; Mahmoodi et al. 2019; Rusu and Rusu 2021).

The wave time series spans twenty years and covers the period 2001?2020. The wave climate analysis was based on a standard wave parameter (significant wave heightHs,peak wave periodTe,and mean directionθm).

A Datawell buoy observations provided by Meteoro‐logical General Direction (MGD Morocco), were exploit‐ed to test the performance of SWAN model to simulate the propagation of waves in the nearshore area.The loca‐tion of the buoy is indicated in the Figure 1 with a red di‐amond symbol.

In order to force SWAN with a real wave boundary con‐ditions,WWIII model was set up for the north Atlantic ar‐ea presented in the Figure 2.The source term physics pack‐age used in the present study is ST2(Tolman and Chalikov 1996). Table 1 summarizes the main setting parameters of WWIII model.The weather forcing data of WWIII are ex‐tracted from ERA5 data-set with 0.25°×0.25° spatial reso‐lution and 1 hour time step.

Table 1 Main settings of WWIII implemented in the present study

In this study, the SWAN cycle III version 41.20 was used for wave simulations. The model was executed in both stationary and non-stationary modes. The spectral di‐rections cover the full circle. The resolution inθ-space is 10° for the variableθand 32 frequencies. The lowest fre‐quency is fixed at 0.052 1 Hz and the highest at 1 Hz.The non-stationary application of SWAN was forced by winds retrieved from ERA5 database. While in the stationary mode, it was implemented without considering the effect of winds on the incident waves as in (El Mekadem et al.2011)and(Hemdane et al.2016).The SWAN physical pro‐cesses during the wave simulation are wave generation,quadruplet nonlinear interaction, dissipation by friction on the bottom. The parameterization of the friction used for our simulations is that of Madsen. The simulations also take into account the effects of diffraction as well as wavewave interactions.

3 Wave climate and wave simulations

3.1 Offshore wave climate and wave energy

The study of data on the wave climate in the general ar‐ea of Casablanca was drawn up using ECMWF (European Center for Medium range Weather Forecasts) ERA5 dataset. The wave climate analysis was based on a standard wave parameter (significant wave height, peak wave peri‐od, and mean wave direction) covering the period 2001?2020.

Figure 3 depicts the wave-rose off Casablanca-Moham‐media coast.The deep-water waves reaching the area of in‐terest come mainly from NW and NNW directions. The dominant heights are in the interval [1.0–2.0 m] with about 62% of probability.As far as wave periods concern‐ing, the most frequent are in the interval [12–15 s] with about 35%of frequency,followed by those in the intervals[10–12 s] and [8–10 s] respectively (frequencies 27% and 20%). Table 2 summarizes the yearly and seasonal mean wave parameters in the study area. Both observed mean significant wave height and wave energy period are signifi‐cantly higher during winter months (Hs= 2.3 m,Te= 11.8 s)than in summer period(Hs=1.3 m,Te=8 s).

Figure 3 Wave-rose offshore the studied zone

Table 2 Annual and seasonal mean wave characteristics calculated at Casablanca-Mohammedia offshore area

Based on the ERA5 wave data and using the Equation (6), the wave potential was calculated for the point (33.95°N8°W). The estimated wave energy poten‐tial in Casablanca provides a good resource for wave con‐version with an estimated mean annual wave power of 22.4 kW/m with a maximum value of 37.1 kW/m in winter.The previous potential is characterized by a noticeable sea‐sonal variability culminating at winter months (Table 2).The observed wave energy during winter months is five times higher than the quantity occurring during summer period. The corresponding theoretical annual wave energy is about 196 MWh/y.The results obtained in this study are in good agreement with those presented by Bouhrim and El Marjani(2019)and Sierra et al.(2016).

To design of a suitable WEC for a selected site, the socalled scatter diagram defined by wave heightHsand ener‐gy periodTebins is a fundamental tool.Figure 4 illustrates the wave energy scatter diagram of Casablanca-Mohamme‐dia coast. Wave height bins have 0.5 m width and energy period bins 0.5 s width.We note the fact that wave energy is concentrated in wave withHsbetween 2.0 and 3.0 m,andTebetween 10.5 s and 12.0 s.The most energetic waves are those with 2.5 m ofHsand 11 s ofTe.

Figure 4 Wave energy Scatter diagram in terms of Hs and Te for Casablanca-Mohammedia coast

3.2 SWAN model simulations

In order to evaluate the SWAN model, a wave buoy is installed in the study area (33°44'81"N–7°23'2"W)(Figure 1).For this purpose,a simulation covering a peri‐od of 8 days period was conducted. The selected period spans from 01/01/2008 to 08/01/2008 and includes a heavy storm on 04/01/2008, with a significant wave height culminating at 8 m. The numerical models were forced by the ERA5 reanalyzed wind field corresponding to 10 m wind components(U10 and V10).A six-day peri‐od, starting from 26/12/2007, to spin-up the wave models was firstly performed.

The values of significant wave height computed by the SWAN model for the location of the buoy are compared with the actual buoy measurements in Figure 5. We no‐tice that there is a good agreement between the model re‐sults and measurements throughout the simulation peri‐od, including the build-up of the storm, its peak (around noon of 04/01/2008). The correlation coefficient between the buoy and the model data is 0.94 for the wave height and 0.90 for the wave energy flux. The Figure 5 shows a comparison graphs depicting the model wave height, and energy versus observed data. The graphs demonstrate that, in general, the large wave event is well captured by the simulation. However, the peak of the episode was slightly underestimated by SWAN by less than 1m with three hours delay.

Figure 5 Significant wave height Hs and wave power Pw obtained with the SWAN model, and to buoy data observation from 01/01/08 at 0000GMT to 08/01/2008 at 0900GMT

The modeled wave parameters (significant height, peri‐od, direction) and wave energy flux fields were provided at thirty minutes time step.The figure 6 illustrates the spa‐tial distribution of the significant height wave energy flux relative to the conditions at 15:00 GMT,on 04/01/2008 i.e.the peak of a strong storm recorded during 2008.The fore‐casted wave heights are between 8.0?8.5 m offshore and in the range 6.5?7.5 m nearshore (Figure 6(a)). The ob‐served waves come mainly from NW direction.

In term of wave energy flux(Figure 6(b)),the values are more pronounced in deep waters and decrease towards the shore. The maximum energy flux is around 480 kW/m in the northern edge of the studied domain. We note a near‐shore high energy area situated north of Cap Fedala with values between 400?440 kW/m. South of Cap Fedala, the wave energy is generally low except in Sidi Rehal near‐shore area, where wave energy arises locally and reaches 440 kW/m.This area,as mentioned in the section 2.1,con‐tains a pronounced submarine ridge that probably led to a local wave energy concentration.

Figure 6 Spatial distribution of Hs and Pw simulated by SWAN on 04/01/08 at 1500UTC

Having tested the coastal wave model in real conditions,it was applied to identify the location of the nearshore hot‐spots in the Casablanca-Mohammedia coast. For this pur‐pose, the nearshore wave patterns were computed in three case studies.The first two are representative of a character‐istic winter and summer situations, and the third, of a heavy storm conditions. These synthetic wave conditions were selected based on the wave climate analysis done in the section 3.1.

For the simulation results, the first case study I, repre‐sents a typical winter wave condition (Hs=2.3 m,Te=12.4 s,θ0=304°). The Figures 7(a) and 7(b), show the wave field,and wave energy flux simulated by SWAN in Mohamme‐dia-Casablanca for this first case study. The wave heights are generally in the interval 2?2.2 m.In the extreme south,nearshore Sidi Rahal, significant height exceeds 2.3 m.As far as the wave energy flux, in the nearshore area, the val‐ues are in the range 20?25 kW/m. We note a remarkable area of energy concentration located in the southern part of the coast in front of Sidi Rahal area, where wave energy rises significantly to 30?35 kW/m.This increase is equiva‐lent to about 17%of the offshore wave energy.

Figure 7 Spatial distribution of Hs and Pw for the case study I(Hs=2.3 m,Te=12.4 s,θ0=304°)

The second case study II represents a typical summer wave condition (Hs=1.3 m,Te=8.5 s,θ0=321°).The wave height as well as wave flux in the nearshore region are sig‐nificantly low(Figures 8(a)and 8(b))withHsbetween 1?1.2 m(4?6 kW/m), except in the southern extremity whereHsarises slightly above 1.2 m(about 7 kW/m).

Figure 8 Spatial distribution of Hs and Pw for the case study II(Hs=1.3 m,Te=8.5 s,θ0=321°)

Finally, the case study III representing a typical storm waves (Hs=4.2 m,Te=17 s,θ0=259°) was considered.These conditions, extracted from ERA5 reanalysis, corre‐spond to a real situation occurred on 28/02/18 at 15:00 GMT. This situation is characterized by West-South West‐erly (WSW) waves and was intentionally selected to study the WSW wave’s transformation due to the local seabed.The simulation results of the last case study are given in Figures 9(a)and 9(b).The wave height,close to the shore,is between 3?3.6 m. We observe in the vicinity of the capes an increase of wave heights reaching 3.6?4 m.An ar‐ea of wave accentuation is visible in the southernmost part of the coast.

Figure 9 Spatial distribution of Hs and Pw for the case study Ⅲ(Hs=4.2 m,Te=17 s,θ0=259°)

The wave energy flux is mainly in the range 80?110 kW/m,except in the aforementioned area. In this zone, the wave potential amounts to 100?140 kW/m. This demonstrates that even WSW waves are affected by the submarine ridge D1. Thus, the wave transformation leads to wave energy concentration even in WSW wave condition.

In conclusion, it is apparent that as the waves propa‐gate from the sea to the coast, the wave height gradually decreases, however, this general trend is not verified in some coastal areas where the bathymetry leads to a con‐centration of wave energy and therefore an increase in wave height. This remark is observed in the vicinity of the Cape Dar Bouazza, precisely around the ridges D1 and D2 as well as the shoals H1 and H2.This observation is clearly visible on the longitudinal sections that pass through the three graphical capes,i.e.,an increase in height as it moves from the open sea to the coast (Figure 10).(Idrissi et al. 2004) previously studied the evolution of wave in the studied region using a computer code devel‐oped by (Nicholas and Germain 1993). Based on this code, which considers the refraction of the waves, the wave orthogonals were drawn up in the coastal area. The study revealed that the convergence of the wave orthogo‐nals, producing a concentration of wave energy, is ob‐served behind the submarine ridges (D) and at the shoals(H). This finding may explain the remarkable increase of wave energy in the aforementioned hotspot located in front of Sidi Rehal due to the presence of the large sub‐marine ridge(D2).

Figure 10 Wave power simulated by SWAN along five longitudinal sections for the first case study(winter)

4 Conclusion

Currently, only 10% of energy needs in Morocco are satisfied by renewable energy, mainly hydraulic, wind and solar energy. However, with its geographical posi‐tion, bordered by two seas: the Atlantic Ocean and the Mediterranean Sea, Morocco has a great potential in ma‐rine energy.

The present study presents an assessment of the poten‐tial wave energy resource in Casablanca-Mohammedia nearshore area (Moroccan Atlantic coast), considered as the most populated region of the country. The wave cli‐mate has been presented based on 20 years wave data pro‐vided by ECMWF. Afterwards, the SWAN model was used to compute the wave propagation and transforma‐tion in the studied area under three typical wave scenari‐os.The modeling tool was validated with a buoy data,in‐dicating a good agreement between the measurements and the simulations with regard to significant wave heights and wave energy.

The spatial distribution of wave energy in the study ar‐ea is examined under different conditions (winter, sum‐mer, and storm conditions). The results of this study shown in term of wave energy power analysis confirm the abundance of wave energy in the area with an annual mean potential of about 22 kW/m. Moreover, the study demonstrated a clear seasonal variability of the wave re‐source and that wave power levels are five times higher in winter than in summer.The application of SWAN dem‐onstrated that the wave energy is unequally distributed within the studied area and presents areas with energy concentration.

In addition, an important hotspot site was identified to be taken in consideration when investigating WEC imple‐mentation.The hotspot site is located in the southern limit of Casablanca-Mohammedia coast, nearshore Sidi Rahal area.

AcknowledegmentsThe authors are thankful to General Direction of Meteorology (GDM Morocco) for providing in-situ Buoy wave data in Mohammedia coast.The authors would like to thank the EC‐MWF (European Center for Medium range Weather Forecasts)ERA5 data-set, for providing the data on the wave climate, and NG‐DC/NOAA for providing the bathymetry data ETOPO 1 arc-minute.The authors also would like to express their gratitude to the SWAN group at Delft University of Technology for making the model freely available.