- Email: [email protected]
Wave transformation in the nearshore waters of Jeddah, west coast of Saudi Arabia
Ocean Engineering 163 (2018) 599–608
Contents lists available at ScienceDirect
Ocean Engineering journal homepage: www.elsevier.com/locate/oceaneng
Wave transformation in the nearshore waters of Jeddah, west coast of Saudi Arabia
T
Alaa M.A. Albarakati∗, V.M. Aboobacker Faculty of Marine Sciences, Department of Marine Physics, King Abdulaziz University, P.O. Box 80207, Jeddah, 21589, Saudi Arabia
A R T I C LE I N FO
A B S T R A C T
Keywords: Wave spectra Red sea Diurnal variability Wave transformation Jeddah coast SWAN
Many specific problems in the coastal zones require an accurate description of the wave field and knowledge of wave parameters. The studies on the spectral characteristics of wind-waves in the Red Sea are very much limited. In this study, the spectral characteristics of nearshore waves in the central Red Sea, specifically the Jeddah coast has been investigated utilizing a third generation spectral wave model, SWAN. The model results were validated against the available measured data. The seasonal and monthly characteristics and the diurnal variability of wave spectra were analysed and discussed. The wave transformation between deep, intermediate and shallow water depths were assessed at three transects – northern, central and southern regions off Jeddah. The results indicate that multi-directional swells are present in the Jeddah nearshore regions, which are propagated from the northern and southern Red Sea. The diurnal variability in the wave spectra is persistent throughout the year, although it fluctuates among the seasons according to the prevailing wind conditions. Significant attenuations in wave heights were identified in the intermediate and shallow waters, with the highest attenuation occurred in the central Jeddah coast.
1. Introduction
validated spectral wave models can be considered as an alternative to resolve the spectral transformation in the offshore and nearshore regions. The in situ wave data are very much limited in the Red Sea. NDBC provides wave parameters from a met-ocean buoy (reference number 23020) deployed in a deep water location in the central Red Sea. This data have been utilised in the previous investigations, especially for the validation of offshore wave model results (e.g., Shanas et al., 2017a; Aboobacker et al., 2016). Eventually, numerical wave models were applied to study the seasonal and long-term characteristics of the windwaves in the Red Sea. The Red Sea basin often experiences the superposition of multiple wave systems and hence the basin has been categorized into distinct regions based on the dominance of superimposed/ non-superimposed/co-existing waves (Shanas et al., 2017b). The Jeddah coast (Fig. 1) is one among the regions, where the wind seas are often superimposed over the swells. Here, the swells are predominantly from the northern Red Sea and a small contribution is from the southern Red Sea, while the local winds are usually in the form of sea breeze and land breeze. Our focus is to elaborate on the spectral wave characteristics off Jeddah using a calibrated nearshore spectral wave model. The region is particularly interesting because of the availability of measured wave spectra for model validation and due to the complex wave-wave and wave-bottom interactions. In this perspective, we have carried out numerical wave simulations for the Jeddah coast using the Simulating
The Red Sea is a semi-enclosed basin located in a narrow, elongated rift valley between Africa and the Arabian Peninsula. It is approximately 2250 km long and 350 km wide at the widest part. It has three distinct depth zones; shallow shelves of less than 50 m, deep shelves having depths between 500 and 1000 m, and the central axis with depths between 1000 and 2900 m (Rasul et al., 2015). The Large scale wind patterns in the Red sea are primarily controlled by the seasonal characteristics and the surrounding orography (Patzert, 1974; Clifford et al., 1997; Sofianos and Johns, 2003). In the northern Red Sea (north of about 20° N) the north-westerly wind blows all around the year. In the southern Red Sea the intensity and direction of winds are mainly controlled by the Arabian Sea monsoon; with dominant south-easterlies in winter (November–April) and north-westerlies in summer. There exists a convergence zone in the central Red Sea (south of Jeddah) during winter, where the north-westerlies converge to the south-easterlies that lead to very low wind speeds (Ralston et al., 2013). These peculiarities in the wind systems reflect on the wave characteristics of the Red Sea (Langodan et al., 2014; Saad, 2010; Zubier et al., 2008). Ocean wave spectra refer to the distribution of the total wave variance over frequency and direction. Continuous measurements of wave spectra are difficult due to operational constraints, whereas the
∗
Corresponding author. E-mail address: [email protected] (A.M.A. Albarakati).
https://doi.org/10.1016/j.oceaneng.2018.06.041 Received 8 October 2017; Received in revised form 28 May 2018; Accepted 13 June 2018 0029-8018/ © 2018 Published by Elsevier Ltd.
Ocean Engineering 163 (2018) 599–608
A.M.A. Albarakati, V.M. Aboobacker
Fig. 1. (a) The Red Sea and (b) the Jeddah model domain and bathymetry. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
summer) and studied the wind sea and swell characteristics. This data has been used in the present study for the validation of SWAN model results. Nonetheless, the spatial and temporal characteristics of the wave spectra in the nearshore regions of Jeddah are yet to be understood. Previous studies lack the discussion on the wave transformation from deep to intermediate and shallow regions off Jeddah coast. In this context, the present study aims to explore the spectral wave characteristics off the Jeddah coast using the spectral wave model SWAN. The monthly and seasonal features have been discussed. Diurnal variability of the wave spectra has been particularly addressed. The transformation of wave parameters from the deep to intermediate and shallow depths have been analysed considering three cross-shore transects. The paper has been organised as follows: Section 2 describes the area of study, Section 3 explains the data and methodology that consists of the description on the wave data collection, wave model setup and validation of model results, Section 4 demonstrates the results and discussions, and Section 5 summarises the important results.
Waves Nearshore (SWAN), a spectral wave model specifically designed for nearshore applications. The studies based on wave model results in the Red Sea were mainly focused on the total wave parameters rather than the component waves such as wind seas and swells. Saad (2010) used the WAM model for hindcasting the waves in the Red Sea with a relatively coarse spatial resolution. The WaveWatch III (WW3) model has been widely applied in the Red Sea for short-term and long-term wave hindcasting (e.g. Langodan et al., 2014; Aboobacker et al., 2016; Shanas et al., 2017a, 2017b). The SWAN-based models were also used for short-term wave analysis (e.g., Zubier et al., 2008; Ralston et al., 2013; Fery et al., 2012, 2015); they are capable of resolving the shallow water processes more accurately than the offshore wave models (Booij et al., 1999). Saad (2010) gives an overall idea of the wave conditions in the Red Sea. This study marked the under-estimations of wave heights, which are primarily due to the coarse resolution of model grids and input winds. Better predictions were obtained in the later hindcasts in the Red Sea with proper treatment of the source functions (Langodan et al., 2014), which leads to reliable assessment of wave power (Aboobacker et al., 2016) and understanding of short-term and long-term variability (Shanas et al., 2017 a; b). Zubier et al. (2008) customized the SWAN for the first time in the Red Sea with a focus to implement an operational wave prediction system. The follow-up study analysed the sea states along the Jeddah coast (Fery et al., 2015). A comprehensive analysis of the wave conditions in the Red Sea were made by Ralston et al. (2013), which examined the impact of Tokar winds in the central Red Sea. Although the wave conditions in the Red Sea are vastly described, localised features are yet to be well-understood. Fery et al. (2015) analysed the wave parameters measured off Jeddah and used them for the verification of a Red Sea wave model. The reported average wave heights in these locations are 0.6 m and 0.4 m, respectively, whereas the maximum wave heights are 2.2 m and 1.2 m, respectively. They identified distinct diurnal variations in the wave patterns. Pronounced diurnal variability are limited to the coastal regions, which are due to the sea breeze – land breeze systems and their influence diminishes towards the offshore regions (Ralston et al., 2013; Shanas et al., 2017b). Recently, Shanas et al. (2018) analysed the measured wave spectra off Jeddah for a limited period of time (during
2. Area of study The Jeddah coast lies in the central part of the Red Sea along the west coast of Saudi Arabia (Fig. 1). The winds are predominantly from the NW/WNW throughout the year; however, local breezes from the N to E directional sector and occasional desert winds from the SE are also accountable. The convergence zone developed in the south of Jeddah (around 18° N) during winter has several implications on the met-ocean parameters in the central Red Sea; e.g., resulting in low wind speeds and weakens the local wind seas. During summer (especially during Jul and Aug), the Tokar gap winds developed in the Tokar mountain ranges in the Sudan blow as westerlies across the Red Sea. These wind systems can generate high waves in the central Red Sea and propagate towards the Saudi coasts. A portion of these waves occasionally reaches the Jeddah region. The bathymetry off Jeddah coast is complex due to the steep gradients in water depths and by the presence of coral reefs (Fery et al., 2015). The coral and island reefs significantly reduce the wave propagation towards the coast. The orography of the Jeddah bay helps to 600
Ocean Engineering 163 (2018) 599–608
A.M.A. Albarakati, V.M. Aboobacker
calculations in the nearshore regions of Jeddah. The JONSWAP parameterization (Hasselmann et al., 1973) has been used for wave dissipation due to bottom friction, where the calibrated value of the coefficient of friction (cfjon) is 0.06. The formulation by Battjes and Janssen (1978) has been applied for the depth-induced breaking, where the calibrated values of Alpha and Gamma are 1 and 0.64, respectively. For the model computations, the frequencies were discretized logarithmically with 37 bins between 0.08 Hz and 0.6 Hz, and the directions were distributed into 36 bins with 10° intervals. The possible lower frequency cut-off for wind-waves is 0.033 Hz, however, the waves with less than 0.08 Hz are not present in the boundary spectra. The model outputs (wave spectra and integral parameters) have been derived for every 1 h. We are considering two model setups: Model 1 – with the same cut offs and frequency discretization as in the measurements and Model 2 – with relevant frequency cut offs (0.08–0.60 Hz) and logarithmic discretization. The Model 1 results have been used for the model validation, whereas the Model 2 results have been used for further analysis and discussion.
Table 1 Details of the wave data used. Instruments
Locations
Directional Wave Recorder (DWR) Non-directional Wave and Tide Recorder (WTR)
39° 21° 39° 21°
06.479′ 34.509′ 05.130′ 42.265′
E N E N
Water depths (m)
Data duration
Data intervals (hour)
4.5
15–28 Jul 2006 21 Jun – 03 Jul 2005
1
9.5
3
reduce impact of large waves in the vicinity of the Jeddah Port. 3. Data and methodology 3.1. Measured wave data The Valeport Midas Directional Wave Recorder (DWR) and nondirectional and Wave and Tide Recorder (WTR) were deployed at 4.5 m and 9.5 m depths, respectively for the wave data collection (Fig. 1b). The directional and non-directional frequency spectra were sampled over a period of 20 min during every 1 h at DWR, while only the nondirectional frequency spectra during every 3 h were collected at WTR (Table 1). The instruments were configured to process the raw data in a way that includes 39 frequencies having 0.0078125 Hz intervals within the frequency limit of 0.0078–0.3047 Hz. In DWR, 180 bins of directions with 2° intervals were sorted out. The accuracies of wave heights in the instruments DWR and WTR are ± 2.0 cm and ± 0.5 cm, respectively (Valeport Limited, 2008). The integral wave parameters were estimated from the spectra as follows: Significant wave height, Hm0 = 4 m 0 , mean wave period,
Tm02 =
m0 m2
Mean wave direction, θm = arctan
b1 a1
3.3. Model validation Fig. 2a shows the snapshots of measured and modelled wave spectra at DWR. The models were nearly predicted the peak and shape of the measured wave spectra within the available frequency range. The small inconsistencies of the model spectra in the high frequency region are due to the coarse spatial resolution of the input winds applied in the offshore wave model, which was carried over to the nearshore wave model boundaries. Nonetheless, the wind sea part of the spectrum is evident in the model spectra (Model 2), which is obvious in the Red Sea coasts. The energies in these lower frequencies generally fluctuate according to the local wind conditions, which often leads to diurnal variability. However, they were partially neglected in the measured data due to the frequency cut off of around 0.30 Hz. Fig. 2b and c shows the time series comparison of measured and Model 1 wave parameters. The model derived significant wave height and mean wave period are in a reasonable agreement with the measurements. Diurnal patterns are partially apparent in the DWR data, especially when the wind sea energy shifts to relatively lower frequencies. However, they are almost negligible in the WTR data. This does not imply that the wind sea energy is not sufficiently available in this region. The Model 2 spectra clearly indicate the presence of wind seas in the higher frequencies. In the real scenario, diurnal variability is expected due to these wind sea components. The error statistics estimated between the measured and Model 1 wave parameters and spectra are listed in Table 2. The fit is reasonable with correlation coefficients of 0.84–0.89 [0.65–0.78] and r.m.s.errors of 0.03–0.07 m [0.19–0.48 s] for the significant wave heights [mean wave periods]. The Model 1 significant wave heights were slightly overestimated during low wave conditions and under-estimated during relatively high wave conditions (Fig. 2d). The Model 1 mean wave periods were slightly over-estimated during high wave conditions. These inconsistencies can be ascribed to the differences in the frequency spectra (Fig. 2a), which are due to the possible inaccuracies in the input winds. The very low energy densities in the measured spectra around 0.28–0.30 Hz are also leading to the mismatch in the comparison. In addition, the irregularities in the coastal bathymetry data due to coarse resolution may lead to errors in the estimation of energy densities, especially on a region where the bathymetry slopes are steep. Nevertheless, the accuracy of model predictions attained here is consistent with Fery et al. (2012).
, where m0, m1 and
m2 are the zero-order, first and second order spectral moments respectively. The a1 and b1 are the Fourier coefficients. 3.2. Wave modelling The spectral wave model SWAN (WL Delft Hydraulics, 2011) has been applied to simulate the waves along the Jeddah coast during Jun 2005–Aug 2006 (14 months). The model takes into account of the processes like wave generation due to wind input, non-linear quadruplet wave-wave interactions, dissipation due to white-capping and the shallow water effects such as bottom dissipation, refraction and shoaling. We refer to Booij et al. (1999) for a full description of the wave model. Fig. 1b shows the wave model domain and bathymetry along the Jeddah coast. The shoreline data were obtained from the latest Google Earth. The water depth data were obtained from the MIKE-CMAP, a digital database provided by DHI Water & Environment, and from a survey data (anonymous source). The model resolution is 100 m × 100 m in longitude and latitude with a total number of 134400 grids. The bathymetry data were linearly interpolated to each grid element in the model domain. The wave model has been initialized using JONSWAP spectrum and forced with the hourly CFSR wind velocities, which are available at every 0.312° × 0.312° spatial resolution (Saha et al., 2010). We refer to Shanas et al. (2017a) for the verification of CFSR winds in the Red Sea. Three open boundaries were chosen; in the south, west and north of the model domain. The open boundary conditions are the wave energy spectra derived from a Red Sea WW3 model hindcast (Shanas et al., 2017a, 2017b; Aboobacker et al., 2016). Fery et al. (2012, 2015) tested different wave growth formulations for the offshore and nearshore regions in the Red Sea, and identified that the formulations by Janssen (1991) and Komen et al. (1984) are better reproduced in offshore and nearshore regions, respectively. We applied the formulations by Komen et al. (1984) for the wave growth 601
Ocean Engineering 163 (2018) 599–608
A.M.A. Albarakati, V.M. Aboobacker
Fig. 2. Comparison between measured and modelled wave parameters and spectra: (a) Snapshots of measured and model wave spectra at DWR, (b) time series Hs and Tm at DWR, (c) time series Hs and Tm at WTR and (d) Scatters of Hs and Tm combined at DWR and WTR.
et al. (2017b). Clearly, there are two distinct directional sectors (W/WNW and S/ SSW) in all the seasons; however, dominated by the W/WNW waves during summer, pre-winter and pre-summer. This is aligned with the previous investigations in the offshore regions of the Red Sea, where the waves are predominantly in the NW/NNW direction, except during winter (e.g., Langodan et al., 2014; Shanas et al., 2017b). The directional shifts of these offshore waves, when they approach the Jeddah coast, are mainly due to the wave refraction. During winter, the energy densities are almost equally distributed in the above directional sectors. This indicates that the coastal winds along the Saudi coast and the NE monsoon wind generated waves propagated from the southern Red Sea (Langodan et al., 2014) are sufficiently contributing to the waves in the central Red Sea during winter. Fig. 4 shows the monthly mean normalized 2D wave spectra at DWR. The spectral features are nearly the same at location WTR, but slightly different in magnitudes due to the depth variations. Well-defined multi-directional spectra are present during the months Jul (in summer) and Dec–Feb (in winter). In Jul, the secondary peak is from the S/SSW direction, which is due to the influence of Tokar gap winds, consistent with the earlier observations (Ralston et al., 2013; Langodan et al., 2014; Aboobacker et al., 2016). In Dec–Feb, the primary peaks in the S/SSW directions are associated with the swells propagated from the southern Red Sea due to the prevailing NE monsoon wind conditions. Although multi-directional waves (S/SSW and W/WNW) are present in the monthly mean spectra during Dec–Feb, their coexistence are reasonably low in the real time spectra (Shanas et al., 2017b). This indicates the prevalence of one of the individual components at some point of time. The wide frequency spectra during the other months (Mar–Nov) reveals the coexistence of distant swells propagated from the northern Red Sea and the waves generated within the central Red Sea. However, both the waves are in similar directions as the forcing winds in the central and northern Red Sea have similar patterns.
Table 2 Error statistics of the model results. Location
Spectra/ Parameters
Correlation coefficient
Bias (ModelMeasurement)
RMS error
Scatter Index
DWR
Hs (m) Tm (s) Mean energy density (m2/ Hz) Hs (m) Tm (s) Mean energy density (m2/ Hz)
0.84 0.78 0.84
0.00 0.07 0.02
0.03 0.19 0.37
0.46 0.13 0.61
0.89 0.65 0.86
−0.02 −0.14 0.02
0.07 0.48 0.46
0.39 0.10 0.88
WTR
4. Results and discussion 4.1. Seasonal and monthly variability in the nearshore wave spectra The modelled energy densities have been averaged for seasonal and monthly analysis. Fig. 3 shows the seasonal mean normalized 2D wave spectra at DWR and WTR. Here four seasons are considered: Summer (Jun–Sep), Pre-winter (Oct–Nov), Winter (Dec–Mar) and Pre-summer (Apr–May). The spectral distribution patterns are nearly the same in both the locations, although the magnitudes are different due to the differences in water depths. In the frequency domain, the spectra are relatively narrow during winter, whereas they are wider during the other seasons. This is because the swells observed in the Jeddah coast during winter are mainly propagated from the far south and the far north regions of the Red Sea, whereas those propagated from the middle regions of the northern Red Sea and the local wind seas co-exist during the other seasons. The multi-frequency/multi-directional components are present in the annual mean wave spectra in the nearshore regions of Jeddah. This is consistent with the observations of Shanas 602
Ocean Engineering 163 (2018) 599–608
A.M.A. Albarakati, V.M. Aboobacker
Fig. 3. Seasonal mean normalized 2D spectra (m2/Hz) at (a) DWR and (b) WTR.
same in both the locations. The peak spectral densities during the early morning hours are in a narrow frequency band representing swells, which are nearly consistent in all the seasons. However, the peak spectra during the evening hours vary according to season. The peak frequency bands during summer, pre-winter, winter and pre-summer are 0.14–0.3 Hz, 0.15–0.35 Hz, 0.12–0.25 Hz and 0.13–0.3 Hz, respectively. The winter peak frequency bands are relatively narrow indicating the dominance of swells over wind seas, while the pre-winter bands are wide enough to accommodate swell and winds sea energy densities. Here, the diurnal spectra derived during the summer are consistent with that presented by Shanas et al. (2018). Fig. 7 shows the diurnal variability in the monthly mean wave spectra at DWR. The spectral distributions are nearly the same at WTR. The spectral shifting in different bands of frequencies, especially during the evening hours, are the notable features. Spectral spreading is relatively small during Dec–Feb, while larger during the other months. Swell contribution during the early morning hours is higher during Nov–Jan. Well defined wind sea spectra are observed during Aug and Oct. There exist relatively longer swells with sufficient energy in the hours 06–12 during Feb, which are not present during the other months. This could be associated with the differences in potential swell generation regions in the northern Red Sea, which will reflect on the arrival time of swells in the Jeddah coast. A detailed investigation on
4.2. Diurnal variability in the nearshore wave spectra Fig. 5 shows the diurnal variability in the annual mean wave spectra at DWR and WTR. The maximum spectral density is during 18–20 h, and the minimum is during 8–10 h. During the early morning hours, the peak spectral density is on a narrow banded low frequency interval (0.14–0.20 Hz), during which the swells are dominated, while the wind seas are almost negligible. During the evening hours, the spread of the spectra is wide – in a combination of swell and wind sea components, hence the peak spectral density is in the frequency band of 0.12–0.30 Hz. Hence, superimposition occurs between the wind seas and swells during the evening hours, and revert back to the swell conditions during the early morning hours. The swell energies are low during the late morning and noon hours. Similar variability was found in the long-term means of significant wave heights off Jeddah (Fery et al., 2015) and in the annual mean wave parameters (significant wave height and mean wave period) derived at a nearshore location north of Jeddah (Shanas et al., 2017b). The regular diurnal variations in winds are due to circulations owing to large-scale heat differences between land and sea, which creates a corresponding variability in the nearshore waves of the Jeddah coast. Fig. 6 shows the diurnal variability in the seasonal mean wave spectra at DWR and WTR. The spectral distributions are more or less the 603
Ocean Engineering 163 (2018) 599–608
A.M.A. Albarakati, V.M. Aboobacker
Fig. 4. Monthly mean normalized 2D spectra (m2/Hz) at DWR.
Here we considered three transects; northern, central and southern regions to analyse the transformation of wave parameters from deep to shallow waters (Fig. 8). Fig. 9 shows the significant wave heights at the selected deep, intermediate and shallow water locations. Due to the effect of bottom friction the wave height reduces as it propagates from deep to shallow waters. On average, the reductions from deep water to intermediate and shallow water locations are around 20% and 30%, respectively at the northern transect; 28% and 51%, respectively at the central transect, and 21% and 31%, respectively at the southern transect. The maximum attenuation in mean Hs is in the central region due to the relatively
the potential swell generation regions in the Red Sea and on the estimation of swell arrival times at different parts of the Red Sea coasts has been proposed for a future study. 4.3. Nearshore wave transformation The Jeddah coast is one among the steeper nearshore regions, as is the case of the Red Sea coasts in general, and compared to the coasts elsewhere. Considering a maximum wavelength of approx. 70 m in this region, the deep, shallow and intermediate waters can be defined with respect to the water depths as > 35 m, < 3.5 m, 3.5–35 m, respectively.
Fig. 5. Diurnal variability in the annual mean wave spectra at (a) DWR and (b) WTR. 604
Ocean Engineering 163 (2018) 599–608
A.M.A. Albarakati, V.M. Aboobacker
Fig. 6. Diurnal variability in the seasonal mean wave spectra at (a) DWR and (b) WTR.
instance, the attenuation of significant wave heights from the deep to shallow waters along the Goa and Ratnagiri coasts is less than 10%, whereas that along the Dwarka coast (relatively rough sea bottom) is up to 22% (Aboobacker et al., 2013). Compared to the large ocean basins, the wave attenuations in the Red Sea coasts are unique, due to the absence of large and long-period swells. Ardhuin et al. (2003) identified strong attenuation of large swells and relatively weaker attenuation of smaller waves over a wider shelf in the North Carolina coast. The changes in mean wave period usually occur only when there is an interaction between the swells and the local wind seas. Within the short distance in consideration, there are no significant changes in mean wave periods. The systematic transformation due to shoaling and refraction is clearly visible in the mean wave direction (Fig. 10). In the north, the predominant wave direction at deep water location was around 290°, while they were shifted to 275° and 265°, respectively at the intermediate and shallow water locations. Similar variations have been found in the central and southern regions. Although large variations in wave directions have been observed in the deeper locations in
wide shallow regions dominated by denser coral reefs. The reductions of maximum Hs in deep water when propagate to intermediate and shallow water locations are 24.6% and 59.7%, respectively at the northern transect, 30% and 49.8%, respectively at the central transect, and 23.3% and 53%, respectively at the southern transect. The above figures indicate that the percentage of wave attenuation at the northern and southern regions (steep slope regions) are higher for the higher waves and smaller for the smaller waves, while the percentage of wave attenuation in the central region (gentle slope and rough bottom) is nearly the same for higher and smaller waves. This is one of the implications of the coral reefs, where the waves attenuate significantly in any wave conditions. The estimated reductions in wave heights for the Jeddah coast can be considered as a typical scenario for the Red Sea coasts, however, field based investigations are required to substantiate this quantification. The wave attenuations occurred along the Jeddah coast is higher than that observed in most part of the world coasts; however, consistent with those observed in regions where the bathymetry is rough. For 605
Ocean Engineering 163 (2018) 599–608
A.M.A. Albarakati, V.M. Aboobacker
Fig. 7. Diurnal variability in the monthly mean wave spectra at DWR.
that multi-directional swells are present in the Jeddah nearshore regions. The multi-directionalities are more evident in the seasonal and monthly 2D spectra, with high intensities during winter, especially during Dec–Feb. Intense multi-directional swells were observed during Jul, during which the S/SSW swells were associated with the Tokar gap winds. Diurnal variability is evident in the annual, seasonal and monthly mean diurnal 2D wave spectra. The maximum spectral density is at around 18–20 h, and the minimum is around 8–10 h. During the early morning hours, the peak spectral density is over a narrow banded low frequency interval (0.14–0.2 Hz), during which the swells are dominant and the wind seas are almost absent. During the evening hours, the spread of the spectra is wide (over a frequency band 0.12–0.3) due to the superimposition of wind seas and swells. The peak frequency bands
annual cycle, the waves at the shallow water locations were significantly refracted to make shore normal propagation directions. 5. Summary and conclusions The wave transformations in the nearshore waters of Jeddah were studied by applying a third generation spectral wave model, SWAN for the period Jun 2005–May 2006. The model results were validated against available measurements, and the verification is consistent with earlier studies. The issue with high frequencies that are not measured by the in-situ measurements and are directly related to the wind data, whose accuracy is unknown, is an interesting consideration for modelling the wave conditions. The seasonal, monthly and diurnal variability of wave spectra were analysed and discussed. The results indicate 606
Ocean Engineering 163 (2018) 599–608
A.M.A. Albarakati, V.M. Aboobacker
Fig. 8. Deep, intermediate and shallow water locations at the north, central and south transects off Jeddah.
Fig. 9. Significant wave heights at deep, intermediate and shallow water locations along the north, central and south transects off Jeddah during Jun 2005–May 2006.
respectively) as the waves approach from the deep water locations (600 m depth), with the highest reduction in the central regions due to the presence of dense coral reefs. The wave periods are not significantly altered during the wave transformation. The wave direction shifts to make the shore normal propagations.
during summer, pre-winter, winter and pre-summer are 0.14–0.3 Hz, 0.15–0.35 Hz, 0.12–0.25 Hz and 0.13–0.3 Hz, respectively. The narrow winter bands indicate the dominance of swells over wind seas, while the pre-winter bands are wide enough to accommodate swells and winds seas with sufficient contribution. Significant attenuations in the wave heights were identified in the intermediate and shallow locations off Jeddah (20–28% and 30–51%, 607
Ocean Engineering 163 (2018) 599–608
A.M.A. Albarakati, V.M. Aboobacker
Fig. 10. Mean wave direction at deep, intermediate and shallow water locations along the north, central and south transects off Jeddah during Jun 2005–May 2006. D.J., Richter, K., Sell, W., Walden, H., 1973. “Measurements of wind-wave growth and swell decay during the Joint North Sea Wave Project (JONSWAP). Ergaenzungsheft Dtsch Hydrographischen Z. Reihe A 8 (12), 95. Janssen, P.A., 1991. Quasi-linear theory of wind-wave generation applied to wave forecasting. J. Phys. Oceanogr. 21, 1631–1642. Komen, G.J., Hasselmann, S., Hasselmann, K., 1984. On the existence of a fully developed wind-sea spectrum. J. Phys. Oceanogr. 14, 1271–1285. Langodan, S., Cavaleri, L., Viswanadhapalli, Y., Hoteit, I., 2014. The Red Sea: a natural laboratory for wind and wave modeling. J. Phys. Oceanogr. 44, 3139–3159. Patzert, W.C., 1974. Wind-induced reversal in Red sea circulation. Deep Sea Res. 21, 109–121. Ralston, D.K., Jiang, H., Ferrar, T.F., 2013. Waves in Red Sea: response to monsoonal and mountain gap winds. Continent. Shelf Res. 65, 1–13. Rasul, N.M.A., Stewart, I.C.F., Nawab, Z.A., 2015. Introduction to the Red sea: its origin, structure, and environment. In: Rasul, N.M.A., Stewart, I.C.F. (Eds.), The Red Sea the Formation, Morphology, Oceanography and Environment of a Young Ocean Basin. Springer Earth System Sciences, pp. 1–24. Saad, A.M.E., 2010. Wave and Wind Conditions in the Red Sea: a Numerical Study Using a Third Generation Wave Model. Masters Thesis. Geophysical Institute University of Bergen, pp. 88. Saha, S., Moorthi, S., Pan, H.-L., et al., 2010. The NCEP climate forecast system reanalysis. Bull. Am. Meteorol. Soc. 91, 1015–1057. http://dx.doi.org/10.1175/ 2010BAMS3001.1. Shanas, P.R., Aboobacker, V.M., Zubier, Khalid M., Albarakati, Alaa M.A., 2018. Spectral wave characteristics along the central coast of eastern Red Sea. Article in Press. Arab. J. Geosci. 11, 90. Shanas, P.R., Aboobacker, V.M., Albarakati, Alaa M.A., Zubier, Khalid M., 2017a. Climate driven variability of wind-waves in the Red Sea. Ocean Model. 119, 105–117. Shanas, P.R., Aboobacker, V.M., Albarakati, Alaa M.A., Zubier, Khalid M., 2017b. Superimposed wind-waves in the Red sea. Ocean Eng. 138, 9–22. Sofianos, S.S., Johns, W.E., 2003. An oceanic general circulation model (OGCM) investigation of the Red sea circulation: 2. Three dimensional circulation in the Red sea. J. Geophys. Res. 108 (C3), 3066. Valeport Limited Midas DWR & WTR Wave Recorders Operation Manual, 2008. Valeport Limited, St. Peter's Quay, Totnes. WL Delft Hydraulics, 2011. Delft3d User Manual. Delft. The Netherlands. Zubier, K.M., Abulnaja, Y.O., Al-Subhi, A.M., 2008. Development of an operational wave prediction system for the Red Sea: experimental phase. In: Offshore Arabia 2008 Conference.Dubai: United Arab Emirates.
Acknowledgements This project was funded by the Deanship of Scientific Research (DSR), King Abdulaziz University, under the grant no. G-168-150-37. The authors, therefore, acknowledges with thanks DSR's technical and financial support. The authors extend the gratitudes to Dr. Khalid M Zubier for providing the wave data and Mr. Shanas PR for the help in setting up the SWAN model. The simulations in this work were performed at the King Abdulaziz University's High Performance Computing Center (Azizi Supercomputer) (http://hpc.kau.edu.sa). References Aboobacker, V.M., Vethamony, P., Samiksha, S.V., Rashmi, R., Jyoti, K., 2013. Wave transformation and attenuation along the west coast of India: measurements and numerical simulations. Coast Eng. J. 55 (1350001), 1–21. Aboobacker, V.M., Shanas, P.R., Alsaafani, M.A., Albarakaati, Alaa M.A., 2016. Wave energy resource assessment for the Red Sea. Renew. Energy 114, 46–58. Ardhuin, F., O'Reilly, W.C., Herbers, T.H.C., Jessen, P.F., 2003. Swell transformation across the continental shelf. Part I: attenuation and directional broadening. J. Phys. Oceanogr. 33 (9), 1921–1939. Battjes, J., Janssen, J., 1978. Energy loss and set-up due to breaking of random waves. In: Proceedings 16th International Conference Coastal Engineering. ASCE, pp. 569–587. Booij, N., Ris, R.C., Holthuijsen, L.H., 1999. A third-generation wave model for coastal regions: 1. model description and validation. J. Geophys. Res. 104, 7649–7666. Clifford, M., Horton, C., Schmitz, J., Kantha, L.H., 1997. An oceanographic nowcast/ forecast system for the Red Sea. J. Geophys. Res. 102, 25101–25122. Fery, N., Bruss, G., Al-Subhi, A.M., Mayerle, R., 2012. Numerical study of wind generated waves in the Red Sea. In: University of Ghent (Ed.), Book of Proceeding 4th International Conference Coastlab12. Belgium: Ghent. Fery, N., Al-Subhi, A.M., Zubier, K.M., Bruss, G., 2015. Evaluation of the sea state near Jeddah based on recent observations and model results. J. Oper. Oceanogr. 8 (1), 1–10. Hasselmann, K., Barnett, T.P., Bouws, E., Carlson, H., Cartwright, D.E., Enke, K., Ewing, J.A., Gienapp, H., Hasselmann, D.E., Kruseman, P., Meerburg, A., Müller, P., Olbers,
608
Contents lists available at ScienceDirect
Ocean Engineering journal homepage: www.elsevier.com/locate/oceaneng
Wave transformation in the nearshore waters of Jeddah, west coast of Saudi Arabia
T
Alaa M.A. Albarakati∗, V.M. Aboobacker Faculty of Marine Sciences, Department of Marine Physics, King Abdulaziz University, P.O. Box 80207, Jeddah, 21589, Saudi Arabia
A R T I C LE I N FO
A B S T R A C T
Keywords: Wave spectra Red sea Diurnal variability Wave transformation Jeddah coast SWAN
Many specific problems in the coastal zones require an accurate description of the wave field and knowledge of wave parameters. The studies on the spectral characteristics of wind-waves in the Red Sea are very much limited. In this study, the spectral characteristics of nearshore waves in the central Red Sea, specifically the Jeddah coast has been investigated utilizing a third generation spectral wave model, SWAN. The model results were validated against the available measured data. The seasonal and monthly characteristics and the diurnal variability of wave spectra were analysed and discussed. The wave transformation between deep, intermediate and shallow water depths were assessed at three transects – northern, central and southern regions off Jeddah. The results indicate that multi-directional swells are present in the Jeddah nearshore regions, which are propagated from the northern and southern Red Sea. The diurnal variability in the wave spectra is persistent throughout the year, although it fluctuates among the seasons according to the prevailing wind conditions. Significant attenuations in wave heights were identified in the intermediate and shallow waters, with the highest attenuation occurred in the central Jeddah coast.
1. Introduction
validated spectral wave models can be considered as an alternative to resolve the spectral transformation in the offshore and nearshore regions. The in situ wave data are very much limited in the Red Sea. NDBC provides wave parameters from a met-ocean buoy (reference number 23020) deployed in a deep water location in the central Red Sea. This data have been utilised in the previous investigations, especially for the validation of offshore wave model results (e.g., Shanas et al., 2017a; Aboobacker et al., 2016). Eventually, numerical wave models were applied to study the seasonal and long-term characteristics of the windwaves in the Red Sea. The Red Sea basin often experiences the superposition of multiple wave systems and hence the basin has been categorized into distinct regions based on the dominance of superimposed/ non-superimposed/co-existing waves (Shanas et al., 2017b). The Jeddah coast (Fig. 1) is one among the regions, where the wind seas are often superimposed over the swells. Here, the swells are predominantly from the northern Red Sea and a small contribution is from the southern Red Sea, while the local winds are usually in the form of sea breeze and land breeze. Our focus is to elaborate on the spectral wave characteristics off Jeddah using a calibrated nearshore spectral wave model. The region is particularly interesting because of the availability of measured wave spectra for model validation and due to the complex wave-wave and wave-bottom interactions. In this perspective, we have carried out numerical wave simulations for the Jeddah coast using the Simulating
The Red Sea is a semi-enclosed basin located in a narrow, elongated rift valley between Africa and the Arabian Peninsula. It is approximately 2250 km long and 350 km wide at the widest part. It has three distinct depth zones; shallow shelves of less than 50 m, deep shelves having depths between 500 and 1000 m, and the central axis with depths between 1000 and 2900 m (Rasul et al., 2015). The Large scale wind patterns in the Red sea are primarily controlled by the seasonal characteristics and the surrounding orography (Patzert, 1974; Clifford et al., 1997; Sofianos and Johns, 2003). In the northern Red Sea (north of about 20° N) the north-westerly wind blows all around the year. In the southern Red Sea the intensity and direction of winds are mainly controlled by the Arabian Sea monsoon; with dominant south-easterlies in winter (November–April) and north-westerlies in summer. There exists a convergence zone in the central Red Sea (south of Jeddah) during winter, where the north-westerlies converge to the south-easterlies that lead to very low wind speeds (Ralston et al., 2013). These peculiarities in the wind systems reflect on the wave characteristics of the Red Sea (Langodan et al., 2014; Saad, 2010; Zubier et al., 2008). Ocean wave spectra refer to the distribution of the total wave variance over frequency and direction. Continuous measurements of wave spectra are difficult due to operational constraints, whereas the
∗
Corresponding author. E-mail address: [email protected] (A.M.A. Albarakati).
https://doi.org/10.1016/j.oceaneng.2018.06.041 Received 8 October 2017; Received in revised form 28 May 2018; Accepted 13 June 2018 0029-8018/ © 2018 Published by Elsevier Ltd.
Ocean Engineering 163 (2018) 599–608
A.M.A. Albarakati, V.M. Aboobacker
Fig. 1. (a) The Red Sea and (b) the Jeddah model domain and bathymetry. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
summer) and studied the wind sea and swell characteristics. This data has been used in the present study for the validation of SWAN model results. Nonetheless, the spatial and temporal characteristics of the wave spectra in the nearshore regions of Jeddah are yet to be understood. Previous studies lack the discussion on the wave transformation from deep to intermediate and shallow regions off Jeddah coast. In this context, the present study aims to explore the spectral wave characteristics off the Jeddah coast using the spectral wave model SWAN. The monthly and seasonal features have been discussed. Diurnal variability of the wave spectra has been particularly addressed. The transformation of wave parameters from the deep to intermediate and shallow depths have been analysed considering three cross-shore transects. The paper has been organised as follows: Section 2 describes the area of study, Section 3 explains the data and methodology that consists of the description on the wave data collection, wave model setup and validation of model results, Section 4 demonstrates the results and discussions, and Section 5 summarises the important results.
Waves Nearshore (SWAN), a spectral wave model specifically designed for nearshore applications. The studies based on wave model results in the Red Sea were mainly focused on the total wave parameters rather than the component waves such as wind seas and swells. Saad (2010) used the WAM model for hindcasting the waves in the Red Sea with a relatively coarse spatial resolution. The WaveWatch III (WW3) model has been widely applied in the Red Sea for short-term and long-term wave hindcasting (e.g. Langodan et al., 2014; Aboobacker et al., 2016; Shanas et al., 2017a, 2017b). The SWAN-based models were also used for short-term wave analysis (e.g., Zubier et al., 2008; Ralston et al., 2013; Fery et al., 2012, 2015); they are capable of resolving the shallow water processes more accurately than the offshore wave models (Booij et al., 1999). Saad (2010) gives an overall idea of the wave conditions in the Red Sea. This study marked the under-estimations of wave heights, which are primarily due to the coarse resolution of model grids and input winds. Better predictions were obtained in the later hindcasts in the Red Sea with proper treatment of the source functions (Langodan et al., 2014), which leads to reliable assessment of wave power (Aboobacker et al., 2016) and understanding of short-term and long-term variability (Shanas et al., 2017 a; b). Zubier et al. (2008) customized the SWAN for the first time in the Red Sea with a focus to implement an operational wave prediction system. The follow-up study analysed the sea states along the Jeddah coast (Fery et al., 2015). A comprehensive analysis of the wave conditions in the Red Sea were made by Ralston et al. (2013), which examined the impact of Tokar winds in the central Red Sea. Although the wave conditions in the Red Sea are vastly described, localised features are yet to be well-understood. Fery et al. (2015) analysed the wave parameters measured off Jeddah and used them for the verification of a Red Sea wave model. The reported average wave heights in these locations are 0.6 m and 0.4 m, respectively, whereas the maximum wave heights are 2.2 m and 1.2 m, respectively. They identified distinct diurnal variations in the wave patterns. Pronounced diurnal variability are limited to the coastal regions, which are due to the sea breeze – land breeze systems and their influence diminishes towards the offshore regions (Ralston et al., 2013; Shanas et al., 2017b). Recently, Shanas et al. (2018) analysed the measured wave spectra off Jeddah for a limited period of time (during
2. Area of study The Jeddah coast lies in the central part of the Red Sea along the west coast of Saudi Arabia (Fig. 1). The winds are predominantly from the NW/WNW throughout the year; however, local breezes from the N to E directional sector and occasional desert winds from the SE are also accountable. The convergence zone developed in the south of Jeddah (around 18° N) during winter has several implications on the met-ocean parameters in the central Red Sea; e.g., resulting in low wind speeds and weakens the local wind seas. During summer (especially during Jul and Aug), the Tokar gap winds developed in the Tokar mountain ranges in the Sudan blow as westerlies across the Red Sea. These wind systems can generate high waves in the central Red Sea and propagate towards the Saudi coasts. A portion of these waves occasionally reaches the Jeddah region. The bathymetry off Jeddah coast is complex due to the steep gradients in water depths and by the presence of coral reefs (Fery et al., 2015). The coral and island reefs significantly reduce the wave propagation towards the coast. The orography of the Jeddah bay helps to 600
Ocean Engineering 163 (2018) 599–608
A.M.A. Albarakati, V.M. Aboobacker
calculations in the nearshore regions of Jeddah. The JONSWAP parameterization (Hasselmann et al., 1973) has been used for wave dissipation due to bottom friction, where the calibrated value of the coefficient of friction (cfjon) is 0.06. The formulation by Battjes and Janssen (1978) has been applied for the depth-induced breaking, where the calibrated values of Alpha and Gamma are 1 and 0.64, respectively. For the model computations, the frequencies were discretized logarithmically with 37 bins between 0.08 Hz and 0.6 Hz, and the directions were distributed into 36 bins with 10° intervals. The possible lower frequency cut-off for wind-waves is 0.033 Hz, however, the waves with less than 0.08 Hz are not present in the boundary spectra. The model outputs (wave spectra and integral parameters) have been derived for every 1 h. We are considering two model setups: Model 1 – with the same cut offs and frequency discretization as in the measurements and Model 2 – with relevant frequency cut offs (0.08–0.60 Hz) and logarithmic discretization. The Model 1 results have been used for the model validation, whereas the Model 2 results have been used for further analysis and discussion.
Table 1 Details of the wave data used. Instruments
Locations
Directional Wave Recorder (DWR) Non-directional Wave and Tide Recorder (WTR)
39° 21° 39° 21°
06.479′ 34.509′ 05.130′ 42.265′
E N E N
Water depths (m)
Data duration
Data intervals (hour)
4.5
15–28 Jul 2006 21 Jun – 03 Jul 2005
1
9.5
3
reduce impact of large waves in the vicinity of the Jeddah Port. 3. Data and methodology 3.1. Measured wave data The Valeport Midas Directional Wave Recorder (DWR) and nondirectional and Wave and Tide Recorder (WTR) were deployed at 4.5 m and 9.5 m depths, respectively for the wave data collection (Fig. 1b). The directional and non-directional frequency spectra were sampled over a period of 20 min during every 1 h at DWR, while only the nondirectional frequency spectra during every 3 h were collected at WTR (Table 1). The instruments were configured to process the raw data in a way that includes 39 frequencies having 0.0078125 Hz intervals within the frequency limit of 0.0078–0.3047 Hz. In DWR, 180 bins of directions with 2° intervals were sorted out. The accuracies of wave heights in the instruments DWR and WTR are ± 2.0 cm and ± 0.5 cm, respectively (Valeport Limited, 2008). The integral wave parameters were estimated from the spectra as follows: Significant wave height, Hm0 = 4 m 0 , mean wave period,
Tm02 =
m0 m2
Mean wave direction, θm = arctan
b1 a1
3.3. Model validation Fig. 2a shows the snapshots of measured and modelled wave spectra at DWR. The models were nearly predicted the peak and shape of the measured wave spectra within the available frequency range. The small inconsistencies of the model spectra in the high frequency region are due to the coarse spatial resolution of the input winds applied in the offshore wave model, which was carried over to the nearshore wave model boundaries. Nonetheless, the wind sea part of the spectrum is evident in the model spectra (Model 2), which is obvious in the Red Sea coasts. The energies in these lower frequencies generally fluctuate according to the local wind conditions, which often leads to diurnal variability. However, they were partially neglected in the measured data due to the frequency cut off of around 0.30 Hz. Fig. 2b and c shows the time series comparison of measured and Model 1 wave parameters. The model derived significant wave height and mean wave period are in a reasonable agreement with the measurements. Diurnal patterns are partially apparent in the DWR data, especially when the wind sea energy shifts to relatively lower frequencies. However, they are almost negligible in the WTR data. This does not imply that the wind sea energy is not sufficiently available in this region. The Model 2 spectra clearly indicate the presence of wind seas in the higher frequencies. In the real scenario, diurnal variability is expected due to these wind sea components. The error statistics estimated between the measured and Model 1 wave parameters and spectra are listed in Table 2. The fit is reasonable with correlation coefficients of 0.84–0.89 [0.65–0.78] and r.m.s.errors of 0.03–0.07 m [0.19–0.48 s] for the significant wave heights [mean wave periods]. The Model 1 significant wave heights were slightly overestimated during low wave conditions and under-estimated during relatively high wave conditions (Fig. 2d). The Model 1 mean wave periods were slightly over-estimated during high wave conditions. These inconsistencies can be ascribed to the differences in the frequency spectra (Fig. 2a), which are due to the possible inaccuracies in the input winds. The very low energy densities in the measured spectra around 0.28–0.30 Hz are also leading to the mismatch in the comparison. In addition, the irregularities in the coastal bathymetry data due to coarse resolution may lead to errors in the estimation of energy densities, especially on a region where the bathymetry slopes are steep. Nevertheless, the accuracy of model predictions attained here is consistent with Fery et al. (2012).
, where m0, m1 and
m2 are the zero-order, first and second order spectral moments respectively. The a1 and b1 are the Fourier coefficients. 3.2. Wave modelling The spectral wave model SWAN (WL Delft Hydraulics, 2011) has been applied to simulate the waves along the Jeddah coast during Jun 2005–Aug 2006 (14 months). The model takes into account of the processes like wave generation due to wind input, non-linear quadruplet wave-wave interactions, dissipation due to white-capping and the shallow water effects such as bottom dissipation, refraction and shoaling. We refer to Booij et al. (1999) for a full description of the wave model. Fig. 1b shows the wave model domain and bathymetry along the Jeddah coast. The shoreline data were obtained from the latest Google Earth. The water depth data were obtained from the MIKE-CMAP, a digital database provided by DHI Water & Environment, and from a survey data (anonymous source). The model resolution is 100 m × 100 m in longitude and latitude with a total number of 134400 grids. The bathymetry data were linearly interpolated to each grid element in the model domain. The wave model has been initialized using JONSWAP spectrum and forced with the hourly CFSR wind velocities, which are available at every 0.312° × 0.312° spatial resolution (Saha et al., 2010). We refer to Shanas et al. (2017a) for the verification of CFSR winds in the Red Sea. Three open boundaries were chosen; in the south, west and north of the model domain. The open boundary conditions are the wave energy spectra derived from a Red Sea WW3 model hindcast (Shanas et al., 2017a, 2017b; Aboobacker et al., 2016). Fery et al. (2012, 2015) tested different wave growth formulations for the offshore and nearshore regions in the Red Sea, and identified that the formulations by Janssen (1991) and Komen et al. (1984) are better reproduced in offshore and nearshore regions, respectively. We applied the formulations by Komen et al. (1984) for the wave growth 601
Ocean Engineering 163 (2018) 599–608
A.M.A. Albarakati, V.M. Aboobacker
Fig. 2. Comparison between measured and modelled wave parameters and spectra: (a) Snapshots of measured and model wave spectra at DWR, (b) time series Hs and Tm at DWR, (c) time series Hs and Tm at WTR and (d) Scatters of Hs and Tm combined at DWR and WTR.
et al. (2017b). Clearly, there are two distinct directional sectors (W/WNW and S/ SSW) in all the seasons; however, dominated by the W/WNW waves during summer, pre-winter and pre-summer. This is aligned with the previous investigations in the offshore regions of the Red Sea, where the waves are predominantly in the NW/NNW direction, except during winter (e.g., Langodan et al., 2014; Shanas et al., 2017b). The directional shifts of these offshore waves, when they approach the Jeddah coast, are mainly due to the wave refraction. During winter, the energy densities are almost equally distributed in the above directional sectors. This indicates that the coastal winds along the Saudi coast and the NE monsoon wind generated waves propagated from the southern Red Sea (Langodan et al., 2014) are sufficiently contributing to the waves in the central Red Sea during winter. Fig. 4 shows the monthly mean normalized 2D wave spectra at DWR. The spectral features are nearly the same at location WTR, but slightly different in magnitudes due to the depth variations. Well-defined multi-directional spectra are present during the months Jul (in summer) and Dec–Feb (in winter). In Jul, the secondary peak is from the S/SSW direction, which is due to the influence of Tokar gap winds, consistent with the earlier observations (Ralston et al., 2013; Langodan et al., 2014; Aboobacker et al., 2016). In Dec–Feb, the primary peaks in the S/SSW directions are associated with the swells propagated from the southern Red Sea due to the prevailing NE monsoon wind conditions. Although multi-directional waves (S/SSW and W/WNW) are present in the monthly mean spectra during Dec–Feb, their coexistence are reasonably low in the real time spectra (Shanas et al., 2017b). This indicates the prevalence of one of the individual components at some point of time. The wide frequency spectra during the other months (Mar–Nov) reveals the coexistence of distant swells propagated from the northern Red Sea and the waves generated within the central Red Sea. However, both the waves are in similar directions as the forcing winds in the central and northern Red Sea have similar patterns.
Table 2 Error statistics of the model results. Location
Spectra/ Parameters
Correlation coefficient
Bias (ModelMeasurement)
RMS error
Scatter Index
DWR
Hs (m) Tm (s) Mean energy density (m2/ Hz) Hs (m) Tm (s) Mean energy density (m2/ Hz)
0.84 0.78 0.84
0.00 0.07 0.02
0.03 0.19 0.37
0.46 0.13 0.61
0.89 0.65 0.86
−0.02 −0.14 0.02
0.07 0.48 0.46
0.39 0.10 0.88
WTR
4. Results and discussion 4.1. Seasonal and monthly variability in the nearshore wave spectra The modelled energy densities have been averaged for seasonal and monthly analysis. Fig. 3 shows the seasonal mean normalized 2D wave spectra at DWR and WTR. Here four seasons are considered: Summer (Jun–Sep), Pre-winter (Oct–Nov), Winter (Dec–Mar) and Pre-summer (Apr–May). The spectral distribution patterns are nearly the same in both the locations, although the magnitudes are different due to the differences in water depths. In the frequency domain, the spectra are relatively narrow during winter, whereas they are wider during the other seasons. This is because the swells observed in the Jeddah coast during winter are mainly propagated from the far south and the far north regions of the Red Sea, whereas those propagated from the middle regions of the northern Red Sea and the local wind seas co-exist during the other seasons. The multi-frequency/multi-directional components are present in the annual mean wave spectra in the nearshore regions of Jeddah. This is consistent with the observations of Shanas 602
Ocean Engineering 163 (2018) 599–608
A.M.A. Albarakati, V.M. Aboobacker
Fig. 3. Seasonal mean normalized 2D spectra (m2/Hz) at (a) DWR and (b) WTR.
same in both the locations. The peak spectral densities during the early morning hours are in a narrow frequency band representing swells, which are nearly consistent in all the seasons. However, the peak spectra during the evening hours vary according to season. The peak frequency bands during summer, pre-winter, winter and pre-summer are 0.14–0.3 Hz, 0.15–0.35 Hz, 0.12–0.25 Hz and 0.13–0.3 Hz, respectively. The winter peak frequency bands are relatively narrow indicating the dominance of swells over wind seas, while the pre-winter bands are wide enough to accommodate swell and winds sea energy densities. Here, the diurnal spectra derived during the summer are consistent with that presented by Shanas et al. (2018). Fig. 7 shows the diurnal variability in the monthly mean wave spectra at DWR. The spectral distributions are nearly the same at WTR. The spectral shifting in different bands of frequencies, especially during the evening hours, are the notable features. Spectral spreading is relatively small during Dec–Feb, while larger during the other months. Swell contribution during the early morning hours is higher during Nov–Jan. Well defined wind sea spectra are observed during Aug and Oct. There exist relatively longer swells with sufficient energy in the hours 06–12 during Feb, which are not present during the other months. This could be associated with the differences in potential swell generation regions in the northern Red Sea, which will reflect on the arrival time of swells in the Jeddah coast. A detailed investigation on
4.2. Diurnal variability in the nearshore wave spectra Fig. 5 shows the diurnal variability in the annual mean wave spectra at DWR and WTR. The maximum spectral density is during 18–20 h, and the minimum is during 8–10 h. During the early morning hours, the peak spectral density is on a narrow banded low frequency interval (0.14–0.20 Hz), during which the swells are dominated, while the wind seas are almost negligible. During the evening hours, the spread of the spectra is wide – in a combination of swell and wind sea components, hence the peak spectral density is in the frequency band of 0.12–0.30 Hz. Hence, superimposition occurs between the wind seas and swells during the evening hours, and revert back to the swell conditions during the early morning hours. The swell energies are low during the late morning and noon hours. Similar variability was found in the long-term means of significant wave heights off Jeddah (Fery et al., 2015) and in the annual mean wave parameters (significant wave height and mean wave period) derived at a nearshore location north of Jeddah (Shanas et al., 2017b). The regular diurnal variations in winds are due to circulations owing to large-scale heat differences between land and sea, which creates a corresponding variability in the nearshore waves of the Jeddah coast. Fig. 6 shows the diurnal variability in the seasonal mean wave spectra at DWR and WTR. The spectral distributions are more or less the 603
Ocean Engineering 163 (2018) 599–608
A.M.A. Albarakati, V.M. Aboobacker
Fig. 4. Monthly mean normalized 2D spectra (m2/Hz) at DWR.
Here we considered three transects; northern, central and southern regions to analyse the transformation of wave parameters from deep to shallow waters (Fig. 8). Fig. 9 shows the significant wave heights at the selected deep, intermediate and shallow water locations. Due to the effect of bottom friction the wave height reduces as it propagates from deep to shallow waters. On average, the reductions from deep water to intermediate and shallow water locations are around 20% and 30%, respectively at the northern transect; 28% and 51%, respectively at the central transect, and 21% and 31%, respectively at the southern transect. The maximum attenuation in mean Hs is in the central region due to the relatively
the potential swell generation regions in the Red Sea and on the estimation of swell arrival times at different parts of the Red Sea coasts has been proposed for a future study. 4.3. Nearshore wave transformation The Jeddah coast is one among the steeper nearshore regions, as is the case of the Red Sea coasts in general, and compared to the coasts elsewhere. Considering a maximum wavelength of approx. 70 m in this region, the deep, shallow and intermediate waters can be defined with respect to the water depths as > 35 m, < 3.5 m, 3.5–35 m, respectively.
Fig. 5. Diurnal variability in the annual mean wave spectra at (a) DWR and (b) WTR. 604
Ocean Engineering 163 (2018) 599–608
A.M.A. Albarakati, V.M. Aboobacker
Fig. 6. Diurnal variability in the seasonal mean wave spectra at (a) DWR and (b) WTR.
instance, the attenuation of significant wave heights from the deep to shallow waters along the Goa and Ratnagiri coasts is less than 10%, whereas that along the Dwarka coast (relatively rough sea bottom) is up to 22% (Aboobacker et al., 2013). Compared to the large ocean basins, the wave attenuations in the Red Sea coasts are unique, due to the absence of large and long-period swells. Ardhuin et al. (2003) identified strong attenuation of large swells and relatively weaker attenuation of smaller waves over a wider shelf in the North Carolina coast. The changes in mean wave period usually occur only when there is an interaction between the swells and the local wind seas. Within the short distance in consideration, there are no significant changes in mean wave periods. The systematic transformation due to shoaling and refraction is clearly visible in the mean wave direction (Fig. 10). In the north, the predominant wave direction at deep water location was around 290°, while they were shifted to 275° and 265°, respectively at the intermediate and shallow water locations. Similar variations have been found in the central and southern regions. Although large variations in wave directions have been observed in the deeper locations in
wide shallow regions dominated by denser coral reefs. The reductions of maximum Hs in deep water when propagate to intermediate and shallow water locations are 24.6% and 59.7%, respectively at the northern transect, 30% and 49.8%, respectively at the central transect, and 23.3% and 53%, respectively at the southern transect. The above figures indicate that the percentage of wave attenuation at the northern and southern regions (steep slope regions) are higher for the higher waves and smaller for the smaller waves, while the percentage of wave attenuation in the central region (gentle slope and rough bottom) is nearly the same for higher and smaller waves. This is one of the implications of the coral reefs, where the waves attenuate significantly in any wave conditions. The estimated reductions in wave heights for the Jeddah coast can be considered as a typical scenario for the Red Sea coasts, however, field based investigations are required to substantiate this quantification. The wave attenuations occurred along the Jeddah coast is higher than that observed in most part of the world coasts; however, consistent with those observed in regions where the bathymetry is rough. For 605
Ocean Engineering 163 (2018) 599–608
A.M.A. Albarakati, V.M. Aboobacker
Fig. 7. Diurnal variability in the monthly mean wave spectra at DWR.
that multi-directional swells are present in the Jeddah nearshore regions. The multi-directionalities are more evident in the seasonal and monthly 2D spectra, with high intensities during winter, especially during Dec–Feb. Intense multi-directional swells were observed during Jul, during which the S/SSW swells were associated with the Tokar gap winds. Diurnal variability is evident in the annual, seasonal and monthly mean diurnal 2D wave spectra. The maximum spectral density is at around 18–20 h, and the minimum is around 8–10 h. During the early morning hours, the peak spectral density is over a narrow banded low frequency interval (0.14–0.2 Hz), during which the swells are dominant and the wind seas are almost absent. During the evening hours, the spread of the spectra is wide (over a frequency band 0.12–0.3) due to the superimposition of wind seas and swells. The peak frequency bands
annual cycle, the waves at the shallow water locations were significantly refracted to make shore normal propagation directions. 5. Summary and conclusions The wave transformations in the nearshore waters of Jeddah were studied by applying a third generation spectral wave model, SWAN for the period Jun 2005–May 2006. The model results were validated against available measurements, and the verification is consistent with earlier studies. The issue with high frequencies that are not measured by the in-situ measurements and are directly related to the wind data, whose accuracy is unknown, is an interesting consideration for modelling the wave conditions. The seasonal, monthly and diurnal variability of wave spectra were analysed and discussed. The results indicate 606
Ocean Engineering 163 (2018) 599–608
A.M.A. Albarakati, V.M. Aboobacker
Fig. 8. Deep, intermediate and shallow water locations at the north, central and south transects off Jeddah.
Fig. 9. Significant wave heights at deep, intermediate and shallow water locations along the north, central and south transects off Jeddah during Jun 2005–May 2006.
respectively) as the waves approach from the deep water locations (600 m depth), with the highest reduction in the central regions due to the presence of dense coral reefs. The wave periods are not significantly altered during the wave transformation. The wave direction shifts to make the shore normal propagations.
during summer, pre-winter, winter and pre-summer are 0.14–0.3 Hz, 0.15–0.35 Hz, 0.12–0.25 Hz and 0.13–0.3 Hz, respectively. The narrow winter bands indicate the dominance of swells over wind seas, while the pre-winter bands are wide enough to accommodate swells and winds seas with sufficient contribution. Significant attenuations in the wave heights were identified in the intermediate and shallow locations off Jeddah (20–28% and 30–51%, 607
Ocean Engineering 163 (2018) 599–608
A.M.A. Albarakati, V.M. Aboobacker
Fig. 10. Mean wave direction at deep, intermediate and shallow water locations along the north, central and south transects off Jeddah during Jun 2005–May 2006. D.J., Richter, K., Sell, W., Walden, H., 1973. “Measurements of wind-wave growth and swell decay during the Joint North Sea Wave Project (JONSWAP). Ergaenzungsheft Dtsch Hydrographischen Z. Reihe A 8 (12), 95. Janssen, P.A., 1991. Quasi-linear theory of wind-wave generation applied to wave forecasting. J. Phys. Oceanogr. 21, 1631–1642. Komen, G.J., Hasselmann, S., Hasselmann, K., 1984. On the existence of a fully developed wind-sea spectrum. J. Phys. Oceanogr. 14, 1271–1285. Langodan, S., Cavaleri, L., Viswanadhapalli, Y., Hoteit, I., 2014. The Red Sea: a natural laboratory for wind and wave modeling. J. Phys. Oceanogr. 44, 3139–3159. Patzert, W.C., 1974. Wind-induced reversal in Red sea circulation. Deep Sea Res. 21, 109–121. Ralston, D.K., Jiang, H., Ferrar, T.F., 2013. Waves in Red Sea: response to monsoonal and mountain gap winds. Continent. Shelf Res. 65, 1–13. Rasul, N.M.A., Stewart, I.C.F., Nawab, Z.A., 2015. Introduction to the Red sea: its origin, structure, and environment. In: Rasul, N.M.A., Stewart, I.C.F. (Eds.), The Red Sea the Formation, Morphology, Oceanography and Environment of a Young Ocean Basin. Springer Earth System Sciences, pp. 1–24. Saad, A.M.E., 2010. Wave and Wind Conditions in the Red Sea: a Numerical Study Using a Third Generation Wave Model. Masters Thesis. Geophysical Institute University of Bergen, pp. 88. Saha, S., Moorthi, S., Pan, H.-L., et al., 2010. The NCEP climate forecast system reanalysis. Bull. Am. Meteorol. Soc. 91, 1015–1057. http://dx.doi.org/10.1175/ 2010BAMS3001.1. Shanas, P.R., Aboobacker, V.M., Zubier, Khalid M., Albarakati, Alaa M.A., 2018. Spectral wave characteristics along the central coast of eastern Red Sea. Article in Press. Arab. J. Geosci. 11, 90. Shanas, P.R., Aboobacker, V.M., Albarakati, Alaa M.A., Zubier, Khalid M., 2017a. Climate driven variability of wind-waves in the Red Sea. Ocean Model. 119, 105–117. Shanas, P.R., Aboobacker, V.M., Albarakati, Alaa M.A., Zubier, Khalid M., 2017b. Superimposed wind-waves in the Red sea. Ocean Eng. 138, 9–22. Sofianos, S.S., Johns, W.E., 2003. An oceanic general circulation model (OGCM) investigation of the Red sea circulation: 2. Three dimensional circulation in the Red sea. J. Geophys. Res. 108 (C3), 3066. Valeport Limited Midas DWR & WTR Wave Recorders Operation Manual, 2008. Valeport Limited, St. Peter's Quay, Totnes. WL Delft Hydraulics, 2011. Delft3d User Manual. Delft. The Netherlands. Zubier, K.M., Abulnaja, Y.O., Al-Subhi, A.M., 2008. Development of an operational wave prediction system for the Red Sea: experimental phase. In: Offshore Arabia 2008 Conference.Dubai: United Arab Emirates.
Acknowledgements This project was funded by the Deanship of Scientific Research (DSR), King Abdulaziz University, under the grant no. G-168-150-37. The authors, therefore, acknowledges with thanks DSR's technical and financial support. The authors extend the gratitudes to Dr. Khalid M Zubier for providing the wave data and Mr. Shanas PR for the help in setting up the SWAN model. The simulations in this work were performed at the King Abdulaziz University's High Performance Computing Center (Azizi Supercomputer) (http://hpc.kau.edu.sa). References Aboobacker, V.M., Vethamony, P., Samiksha, S.V., Rashmi, R., Jyoti, K., 2013. Wave transformation and attenuation along the west coast of India: measurements and numerical simulations. Coast Eng. J. 55 (1350001), 1–21. Aboobacker, V.M., Shanas, P.R., Alsaafani, M.A., Albarakaati, Alaa M.A., 2016. Wave energy resource assessment for the Red Sea. Renew. Energy 114, 46–58. Ardhuin, F., O'Reilly, W.C., Herbers, T.H.C., Jessen, P.F., 2003. Swell transformation across the continental shelf. Part I: attenuation and directional broadening. J. Phys. Oceanogr. 33 (9), 1921–1939. Battjes, J., Janssen, J., 1978. Energy loss and set-up due to breaking of random waves. In: Proceedings 16th International Conference Coastal Engineering. ASCE, pp. 569–587. Booij, N., Ris, R.C., Holthuijsen, L.H., 1999. A third-generation wave model for coastal regions: 1. model description and validation. J. Geophys. Res. 104, 7649–7666. Clifford, M., Horton, C., Schmitz, J., Kantha, L.H., 1997. An oceanographic nowcast/ forecast system for the Red Sea. J. Geophys. Res. 102, 25101–25122. Fery, N., Bruss, G., Al-Subhi, A.M., Mayerle, R., 2012. Numerical study of wind generated waves in the Red Sea. In: University of Ghent (Ed.), Book of Proceeding 4th International Conference Coastlab12. Belgium: Ghent. Fery, N., Al-Subhi, A.M., Zubier, K.M., Bruss, G., 2015. Evaluation of the sea state near Jeddah based on recent observations and model results. J. Oper. Oceanogr. 8 (1), 1–10. Hasselmann, K., Barnett, T.P., Bouws, E., Carlson, H., Cartwright, D.E., Enke, K., Ewing, J.A., Gienapp, H., Hasselmann, D.E., Kruseman, P., Meerburg, A., Müller, P., Olbers,
608