Introduction

The oceans are characterised by random surface elevations and movements as a result of forces acting both externally and internally to balance energy transfer (Blanchette et al. 2008; Brodie and Cohn 2021), thereby making waves. The term sea state or ocean state can be described as being a cumulation of several time-evolving wave systems or the superimposition of concurrent wave systems such as swells and wind-seas (Kpogo-Nuwoklo et al. 2015). Wave generation mechanisms vary as the source of input energy varies. A common source of input energy is the wind, which is constantly blowing over the ocean surface. Meteorological events such as low pressure, hurricanes and storms can create wave systems. The randomness or chaotic nature of ocean waves made it very difficult for their statistical description, estimation and subsequent prediction. Studies of waves over the past decades have however successfully ‘decrypted’ the ‘secret code’ to the random properties and complex mechanisms of their evolution (Caetano and Innocentini 2003; Holthuijsen et al. 1998; Sverdrup and Munk 1947; Thomas and Dwarakish 2015; Tolman et al. 2002). Early studies of ocean waves begun to quantitatively describe ocean surface waves with characteristics such as mean wave height and mean wave period (Mitsuyasu 2002), stemming from the studies of Sverdrup and Munk (1947), who introduced the concept of significant wave height as statistical mean wave for the description of the random properties of ocean surface waves. During that period, simple wave models were used to generate single wave height and period at individual grid locations with a direct relationship between the wind speed, the wave height and period. Such wave models were referred to as the “Representative Wave Approach” (Tolman et al. 2002). Since then, the statistical description of ocean surface waves has advanced into the use of spectral analysis to derive important information that has been used in developing complex algorithms for the forecast and prediction of ocean surface waves. These wave models are thus referred to as spectral wave models, as they describe the distribution of wave energy over the wave frequency (f) and the wave propagation direction (θ). The spectral wave model is represented in its simplest form as indicated in Eq. 1 below and referred to as the energy balance equation:

$$\frac{DF}{Dt}=S={S}_{in}+{S}_{nl}+{S}_{ds}+\dots$$
(1)

The term to the left of the equation represents the change in the local spectrum due to the propagation of wave energy, while the terms to the right of the equation represent a combination of sources and sinks of wave energy. Typically, the sources and sinks comprise the wind input source term (Sin), dissipation due to wave breaking (Sds), and non-linear wave-wave interaction term (Snl) which transfers energy among the wave spectral components but does not change the total energy. The ellipsis represent other processes such as bottom interaction, which influence the distribution of the wave energy. The discretisation and solving of Eq. (1) results in spectral wave models.

The study of waves at various locations have helped in the improvement of wave models and in the provision of more accurate wave climate information in such areas (Caetano and Innocentini 2003; Saulter et al. 2016). Advancements in ocean wave forecasting has made it possible to accurately forecast wave properties at the global and regional levels and even up to local scales. However, in order to obtain accurate wave forecasts and hindcasts of specific local and regional areas, it is important to investigate appropriate wave propagation schemes that best suit any given location. Spectral wave models have progressed from first generation (1G) to the current third generation (3G) models, which incorporate spectra analysis and in-situ data assimilation, and propagation schemes (Lai and Kim 2020). The models continue to be improved for better accurate forecasts and hindcasts (Thomas and Dwarakish 2015).

The classification of spectral wave models has been based on the treatment of wave-wave interactions referred to as non-linear interaction (Tolman et al. 2002) pointed out earlier in this section (Eq. 1). First generation wave models were developed based on the assumption that wave components stop growing when they reached a universal saturation level (Phillips 1958). Energy components are evaluated independently of each other in 1G wave models and therefore does not or minimally consider nonlinear interaction (SWAMP Group 1985). A basic shortcoming of fist generation waves was the over estimation of the wind input and the under estimation of the non-linear transfer (WAMDI Group 1988).

The inclusion of non-linear transfer of energy during wave propagation in the wave equation revolutionized the wave propagation equation of first generation models into a second generation model (WAMDI Group 1988). Second generation (2G) models do not properly simulate complex windseas which are produced by rapidly changing wind fields (sudden change in wind direction or decrease in wind speed), such as may occur during hurricanes and cyclones (Caetano and Innocentini 2003; WAMDI Group 1988), although the low frequency portion of the energy (swells) tends to slowly align with the new wind direction (Caetano and Innocentini 2003). Second Generation models also do not properly treat the transition between windsea and swells, resulting in some observed differences (Van Vledder and Holthuijsen 1993).

In general 1G wave models consisted of simple computations and either did not include non-linear wave interaction or did so minimally. 2G wave models on the other hand fully consider non-linear interactions but do so through parameterisations. Third generation models, however, fully implemented non-linear wave interactions as a source term using the method of Hasselmann et al. (1985).

The objective of this paper is to present a review of ocean state modelling activities in relation to the West African marine environment including some information on general oceanography, climate and meteorology of the region, which are interlinked with the current operating agenda of Blue Economy and the United Nations Decade of the Ocean and their relevance with monitoring and projecting ocean state parameters. This review can serve as a guide and reference material for coastal and marine environment studies of West Africa and other regions, as well as may be applicable to the use and conservation of the marine resources of the region. It is particularly related to the studies of Foli et al. (2021a, b) and Foli et al. (2022), which address ocean wave modelling in the West Africa region. The remainder of this paper is organised as follows: Section 2 – The relevance of ocean state monitoring and modelling to the Blue Economy; 3 – Global ocean wave forecasting; 4 – Errors in global wave models; 5 – Brief comparisons of major third-generation wave models; 6 – Ocean wave studies off West Africa; 7 – Summary of general oceanography of the West Africa region; 8 – The climate and meteorology of West Africa with related impacts; and 9 – Conclusions.

The relevance of ocean state monitoring and modelling to the blue economy

The global Blue Economy agenda has its core philosophy to promote the sustainable and equitable use of coastal and marine resources such that future generations can benefit from the same shared resources available today. The Blue Economy concept sees it best to decouple socioeconomic development in relation to the ocean and related activities from the degradation of ecosystems and the environment as a whole, while the sustainable use of coastal and marine resources are encouraged (World Bank and United Nations Department of Economic and Social Affairs 2017). The Blue Economy agenda is not complete without the consideration of issues related to safety at sea, climate change and ocean renewable energy. The achievement of blue growth and the Blue Economy agenda is inextricably linked to the safety of actors involved in the day-to-day activities in the ocean environment. No successful economic activity on the ocean is sustainable if the safety of the people and infrastructure involved is not ensured. The application of ocean state information is therefore paramount in achieving blue growth and this is why it is important to provide and access information on ocean conditions for decision making to protect life and property, especially for West Africa which has a young growing economy. Similarly, the sustainable exploitation of ocean resources and services for blue growth (Garland et al. 2019; World Bank and United Nations Department of Economic and Social Affairs 2017), especially in the wake of advocacy for ocean renewable energy, requires in-depth knowledge of ocean physical variables such as waves and currents that rely primarily on data. The ocean controls the global climate, which impacts several activities and as such requires attention in all climate change related issues.

Although the Blue Economy concept is promoted as a global agenda, sovereign countries contribute significantly in their actions to achieving blue growth, as they are responsible for the sustainable exploitation of their resources including the safety of the actors, with the support and collaboration of regional and global efforts. National, regional and global community actions can thus contribute to improving issues relating to safety in the marine environment, climate change and sustainable resource exploitation. These cannot be completely achieved without the acquisition, analysis and use of ocean state data. Physical ocean variables affect activities both offshore and onshore and their role in achieving blue growth cannot be overlooked.

Some global policy initiatives have been instigated in order to assist countries in addressing issues relating to safety at sea, climate change and ocean resource exploitations. Ocean development is a key component of the seventeen (17) United Nations’ Sustainable Goals (UN SDGs) that have been adopted by the United Nations (UN) (United Nations General Assembly 2015). Although agenda 12 and 14 of the UN SDGs are mainly dedicated to addressing issues relating to the marine and coastal environment, almost all the SDGs are cross-cutting and interlinked in addressing the several activities that affect this environment (Fig. 1) (Foli et al. 2021a, b; Hossain et al. 2017).

Fig. 1
figure 1

Contribution of the UN SDGs to Blue Economy and their inter-relationship (source: Hossain et al. 2017)

Full size image

The policies of the UN SDGs that are laid down in addressing marine environment issues, especially relating to sustainable exploitation of renewable energy, climate change and safety at sea towards achieving a blue economy, are mainly engraved in UN SDG 7, 11, 13 and 14 (Neumann et al. 2017; United Nations General Assembly 2015). SDG 7 touches on ensuring “access to affordable, reliable, sustainable and modern energy for all”. Specifically, Agenda 7.2 and 7a stresses on use of renewable energy and promoting clean energy infrastructure by the year 2030 (United Nations 2015; United Nations General Assembly 2015). The ocean provides a huge source of renewable energy that can be tapped from ocean waves, tides and winds (Kilcher et al. 2021; Musial et al. 2020; Rusu and Venugopal 2019). These options provide cleaner, greener and climate friendly energy sources that are already being explored and exploited, thus contributing to advancing the Blue Economy Agenda and blue growth.

SDG 11 (11b) and 13 on the other hand advocate combating the impacts of climate change in the marine and coastal domain as well as its mitigations and adaptions. This is also in line with the Sendai Framework for Disaster Risk Reduction 2015–2030, which addresses disaster risk management and its understanding in a holistic manner at all levels (United Nations Office for Disaster Risk Reduction 2015).

In order to sustainably manage and protect marine and coastal ecosystems, the UN SDG 14 (especially SDG 14.2) addresses the conservation and sustainable use of the oceans, seas and marine resources for sustainable development. Here, acquisition and use of scientific knowledge, research capacity development and marine technology transfer, as well as sustainable blue economy form part of a range of multi-stakeholder thematic Communities of Action (COA) that were launched by the UN at a high-level conference to support SDG 14 (UNDESA & WRI 2019).

In addition to the UN SDGs, which is a global policy initiative, African Union (AU) as part of its Agenda 2063 also incorporates maritime safety and security into its continental African Integrated Maritime Strategy (AIMS), which provides guidance to regulating and managing maritime issues by making available a framework for the protection and exploitation of the marine resources of Africa, with the aim of increasing wealth creation from the African maritime domain (African Union 2012; Foli et al. 2021a, b). Similarly from a regional perspective, the Economic Community of West African States (ECOWAS) also introduced the ECOWAS Integrate Maritime Strategy (EIMS) with the mandate of protecting the natural environment, which is integrated into the AIMS.

Other policy initiatives that ensure the safety of actors in the marine domain as well as sustainable exploitation of the marine resources include the Paris Agreement on climate change (Britannica 2021; UNFCC 2015) and the UN Decade of Ocean Science for Sustainable Development (2021–2030) (Ryabinin et al. 2019), which all aim at promoting the Blue Economy agenda and achieving blue growth.

In order to effectively mitigate against the impacts of climate change, ensure safety at sea, and sustainably exploit the vast ocean resources in the West Africa region, including renewable energy, there is therefore the need to obtain adequate knowledge on the ocean and its properties through data collection via in-situ and remote observations as well as modelling the marine environment. This provides the impetus for investigating physical properties such as waves that will provide information for aiding marine constructions, navigation and early-warning information on ocean conditions for safety at sea, as well as past data for climate analysis and projections. Projecting and making available accurate ocean state information on a timely basis and integrating it into an ocean climate models and early-warning system is therefore key to the Blue Economy agenda and achieving blue growth.

Global ocean wave forecasting

The fundamental quantity to predict in wave forecasting is the action density spectrum, N. Wave forecasting is primarily based on Eq. (2):

$${{N}}\boldsymbol{ }=\boldsymbol{ }{{g}}{{F}}/\sigma ,$$
(2)

with σ = (gk tan (kd))½ being the dispersion relation. F is the wave number spectrum, σ is the relative frequency; g is acceleration due gravity; k is wave number; d is finite water depth.

The wave number spectrum gives the distribution of wave variance over wavenumber k, i.e. F(k;x,t).

The energy E of the waves is given by E = σN and the wave momentum P, is given by P = kN.

Numerical equations used in forecasting ocean waves have been developed over the years and still continue to undergo refinements. Currently, the basic equation describing the evolution of surface wave field in space and time is governed by the transport or energy balance equation for ocean wave modelling as represented in Eqs. (1) and (3) (Liu et al. 2019; WMO 1998).

$$\frac{\partial E}{\partial t}+\nabla \bullet \left({c}_{g}E\right)=S={S}_{in}+{S}_{nl}+{S}_{ds},$$
(3)

where E = Spectral Energy; Cg = group velocity; S = Cumulative Source function (i.e. sources and sinks); t is the time. The S (source term) represents the physics of wind-wave generation (Sin), non-linear four-wave interactions (Snl) and dissipation by wave breaking and other sources (Sds) (Bidlot et al. 2007).

From Eq. (3), wave propagation is therefore obtained by using Eq. (4) as incorporated into the WAVEWATCH III (WW3) wave model by the National Oceanic and Atmospheric Administration (NOAA) (WAVEWATCH III Development Group 2016).

$$\frac{DN}{Dt}= \frac{S}{\sigma },$$
(4)

where N is the wave density spectrum.

While Eq. 3 relates the source terms to change in energy density spectrum (E), Eq. 4 does so in terms of action density (N), with Eq. 3 providing an extension of the sources and sinks of energy, which are also applicable to Eq. 4.

Errors in global wave models

The evolution of wave models has aimed primarily at achieving better understanding of the drivers of changing wave fields with respect to space and time that may lead to better prediction of ocean wave parameters. The progression from 1G through to 3G ocean wave models has the primary objective of significantly reducing the resulting errors in wave model outputs (Thomas and Dwarakish 2015).

Despite the improvements in wave models, they still exhibit uncertainties at various degrees. Sources of errors in wave models are well documented (Abdolali et al. 2021; Caires and Sterl 2003; Cavaleri 2009; Gracia et al. 2021; Thomas and Dwarakish 2015). These sources of errors are mainly categorised into three, namely lack of complete understanding of the governing physical processes, discretization of continuous fields, and the degree of knowledge of the drivers or controllers of the circulation (Blumberg and Georgas 2008). The underlying bathymetry, state of overlying atmospheric forcing, etc. form part of drivers or controllers of circulation in a model.

As indicated in the previous sections, the physics of wave modelling is progressive (Lai and Kim 2020), leading to the current 3G models because of incomplete understanding of phenomena and representation of such phenomena that contribute to wave generation, propagation, interactions and dissipation. The inaccurate representation of such contributing phenomena, physically and numerically, is bound to reflect in the wave outputs with respect to the error levels. In WW3 for instance different parameterization schemes exist (WAVEWATCH III Development Group 2016) that can be applied to the same location to yield different results, and thus makes it important for the right physics to be implemented to represent a particular region or location. Despite the improvements in source term parameterizations, some degrees of errors still remains to be resolved.

The physical and numerical discretization of ocean wave propagation with respect to space and time as well as the treatment of flux conservation through the discretization of velocity affects wave behaviour, most especially in shallow areas (Zijlema 2021b). Several discretization types are employed in numerical wave models. These include volume, time, space, flux and velocity discretization schemes (Zijlema 2021a). These discretization types are executed through different discretization methods such as the finite difference, finite volume, or finite element method (Zijlema 2021a), which are aimed at obtaining higher accuracy levels in wave projections. Discretization errors therefore directly translates into uncertainties in ocean wave model outputs. Time stepping methods are mostly used in discretising wave equations and these are also affected by the type of grid used (structured or unstructured grid) in the wave model (Burman et al. 2021; Griesmaier and Monk 2014; Nguyen et al. 2011). The type of time stepping method (discontinuous Galerkin method; conforming method; continuous time Galerkin method; hybridizable discontinuous Galerkin methods) has also been indicated to have an influence on the accuracy of wave projections (Nguyen et al. 2011). In effect, the type, method and level of discretization of source term parameters and wave model physics play a significant role in the accuracy of wave properties from numerical wave models.

Errors, or uncertainties in wave models have long been traditionally known to be attributed to drivers such as the accuracy of the bottom topography, which in many regions is often inaccurately known (Wang et al. 1990). The sources of bathymetry input into wave models vary. Often times, the bottom topography is not well represented due to lack of real measurements. Available bathymetry data is largely interpolated and this contributes to errors as a result of unresolved bathymetric features such as submerged banks, bars and ridges that may impact wave propagation, especially in coastal environments. Wave models also rely heavily on wind inputs as forcing for deriving wave hindcasts and forecasts. This usually comes from reanalysis data in the case of hindcasts, and wind forecasts from numerical weather prediction (NWP) in the case of forecast waves. This suggests that the wind forcing in itself already presents some level of errors or uncertainties as they mainly come from atmospheric models and thus are not perfect. These uncertainties are propagated to the outputs of the wave model. Errors in wind forcing has been cited to be the major source of errors in operational wave forecasting (Rogers et al. 2005). Several possible sources of errors in the wave model can therefore lead to uncertainties and inaccuracies in wave model outputs.

Brief comparison of major third-generation wave models

There are a number of 3G wave models. These include the WAM, SWAN, WW3, AUSWAVE, AUSWAM, STWAVE and MIKE 21 (BOM 2010; Gonçalves et al. 2015; Strauss et al. 2007; Umesh et al. 2018). These are used for global as well as regional ocean state forecasts. Three of these modern 3G models that are most widely used are the WAM (WAve Modeling, WAMDI Group 1988), SWAN (Simulating Waves Nearshore, Booij et al. 1999) and WW3 wave models (Padilla-Hernández et al. 2007). These three have similarities as well as differences that make them applicable to various settings and applications. All three models solve for the evolution of wave fields using the energy balance equation. However, whiles WW3 solves the energy balance equation in terms of spectral wave action (see Eq. 4), WAM does so in terms of spectral energy (see Eq. 3) (Padilla-Hernández et al. 2007). Computation of source terms and propagation in WAM is done using different methods and time steps to explicitly solve the energy balance equation, whiles WW3 on the other hand does so using fractional time steps methods and with configurable propagation schemes (Baordo et al. 2020).

While the WAM and WW3 are designed for projecting wave properties on global and regional scales, the SWAN is more configured towards wave propagation on coastal scales (Umesh et al. 2018). The SWAN model can however be nested in both WAM and WW3 to produce improved results (Padilla-Hernández et al. 2007). It is actually dependent on boundary conditions outputted from a global model such as the WAM and WW3 (Saulnier et al. 2013). SWAN also applies the action balance equation to solve for wave evolution similar to WW3 (Rogers et al. 2007). WW3 is reported to respond better to fast changes in winds compared to the SWAN (Ortiz-Royero and Mercado-Irizarry 2008). However, the SWAN has an advantage of being implemented in both deep and shallow water due to its superior shallow water physics over WW3. The SWAN therefore tend to favour coastal engineers for its ease of use and application to coastal environments.

Several studies (Baordo et al. 2020; Brus et al. 2021; Foli et al. 2022; Puscasu 2014; Saulnier et al. 2013; Sheng et al. 2019; Umesh and Behera 2020; Wojtysiak et al. 2018) have used WW3, which seems to be preferred over the WAM or SWAN. The choice of WW3 over SWAN can probably be related to the findings of Padilla-Hernández et al. (2007) and Ortiz-Royero and Mercado-Irizarry (2008), who found that a comparison between buoy observations and model output tends to favour WW3 over SWAN, although the SWAN model has a better ease of use. Similarly, considering WAM and WW3, the latter in its current versions (i.e. v.5.16 and v.6.07) employs superior physics, wave-current interaction and air-sea temperature difference as an additional input; and its numerical approaches are much superior in terms of growth and dissipation physics to WAM (Fradon et al. 2000). The WW3 model is a third generation model developed at NOAA/NCEP and has undergone several improvements for ocean wave forecasting and hindcasting. It has evolved from just being a wave model into a wave modelling framework, allowing for easy development of additional physical and numerical approaches to wave modelling. The current 3G wave models make improvements upon previous models by use of high frequency wave spectra and utilisation of data assimilation techniques to enhance the accuracy of wave forecasts.

Ocean wave studies off West Africa

Limited studies on ocean wave projections exist for the western Africa region. Existing studies (Toualy et al. 2015) have relied on either limited short-term in-situ measurements from very few locations or on satellite altimetry data which is also limited on time scales. Other studies such as the West African Swell Project (Forristall et al. 2013; Olagnon et al. 2004; Ondoa et al. 2017; Prevosto et al. 2013), have also used model hindcast data generated on global grids using global wave parameterisation schemes. Results from such studies have been used to make inferences on ocean wave impacts on the coast as well as coastal and offshore structures. Although data obtained from hindcast models presents tremendous advantages for analysis with respect to spatial and temporal scale availability, care must however be taken to ensure the accuracy of such datasets when used on regional and local scales as the accuracy may be impacted by the set of source terms used in their generation.

As part of the few existing studies, Toualy et al. (2015) investigated wave climate and origin of ocean swells on the West African coast using a 4.5-year satellite data supported with WW3 hindcast data from NOAA generated on a global grid for the same period. This expanse of data is inadequate on temporal scales for such studies since climate studies need a minimum of a decade of data or more for such analysis to be conducted to facilitate deriving meaningful inferences. The unavailability of observed data covering large time scales of more than a decade leaves the opportunity for use of hindcast data from models, which must be appropriately generated for specific locations.

Although Toualy et al. (2015) indicated the fact that the investigations of the West African Swell Project (WASP) of the wave climate on the West African coast was flawed as a result of the limited sources of measured data, their investigation also did not cover large enough temporal data coverage warranting wave climate investigations. This is likely due to the limited altimetry data, which could be augmented with model hindcast data into the past. For such climate studies to be relevant to coastal and marine engineering, Olagnon et al. (2014) makes it clear that “wave conditions need to be characterised by a history of directional spectra continuously provided for long durations of several years or decades”. Olagnon et al. (2014) provided instances of how such studies could be applied to fatigue assessment of offshore structures, prediction of profitability from the power output of a wave marine energy extraction device, studies of coastal erosion of a shoreline, etc. This is the challenge of the moment and although the WASP project (Olagnon et al. 2004) sought to provide solutions to this challenge, there still remains a significant lacuna that need to be filled since wave parameterisation schemes for the West African sub-region is lacking for the generation of hindcast data on a regional scale for long term applications, as well as forecasting for application to safety at sea.

In a related study, Ondoa et al. (2017) investigated beach response to wave forcing from event to inter-annual time scales in the coastal area of Grand Popo Beach, Benin. Their study made use of hindcast data to investigate the beach’s response to wave forcing from event to inter-annual time scales. Ondoa et al. (2017) compared result of wave measurements using a video camera with WW3 hindcast data from a global grid to make inferences on the wave climate on coastal erosion. Although being aware that the model data they used in their study had errors, and by comparing data from the video camera to the model data, Ondoa et al. (2017) concluded that the video data had more errors, which was estimated in reference with the model hindcast data. A more accurate source of data, such as obtained from investigated source terms on a regional level, however, will enhance the reliance on hindcast and forecast information for better decision making. Accurate information is needed for wave climate studies to assist in the decisions concerning mitigation of coastal erosion as well as construction of coastal and offshore structures. The investigations of Ondoa et al. (2017) was aimed at using an alternate means of deriving ocean wave information in the absence of conventional measurement techniques such as buoys and ADCPs. This may introduce uncertainties in the decisions that emanate from such studies. This is why it is important to investigate wave source term parameterisation schemes for generation of waves on local and regional levels.

The use of global hindcast data on regional and local scales present the challenge of accuracy, which needs to be addressed. Although data generated on a global grid may be useful for several purposes, when used on a regional or local scale, for different applications, the results may not differ because of the inherent systematic errors. It is therefore not surprising for the authors such as Ondoa et al. (2017) to obtain results which they indicated were consistent with wave climate observations of Laïbi et al. (2014) and Almar et al. (2015), who also used the same hindcast datasets for their investigations.

Perhaps the most intensive studies on wave climate in West Africa was conducted by the West Africa Swell Project (Forristall et al. 2013; Olagnon et al. 2004). The West African Swell Joint Industry Project (WASP JIP) conducted by Olagnon et al. (2004), investigated the characteristics of swells off West Africa using in-situ wave measurements from eight locations along the West African coast as well as hindcast data for ten locations from Côte d’Ivoire to Namibia, provided by Oceanweather Inc. and from the NOAA WAVEWATCH III model. Hindcast data used for their investigations was obtained from global datasets which were generated using global wave parameterisation schemes. Among their findings, the WASP indicated that northwest swells have very low amplitudes (Prevosto et al. 2013). Although they also found parameters describing spectral width of swells in the region between measured data and the hindcast data to have ‘good agreement’, a better comparison could be achieved from hindcast data generated using local or regional wave source term parameterisation schemes.

Indeed, Prevosto et al. (2013) in their findings as part of the WASP Project accepted that there was poor overlapping between the hindcast data used and the in-situ measurement database, although “comparisons between in-situ measurements and hindcast models permitted identification of the limitations of the different numerical models available at the time of the project”. This is a clear indication that the use of model outputs generated on a global grid with global source term parameterisation schemes is inaccurate and requires further investigations on the use of hindcast data generated with local or regional source term parameterisation schemes.

Studies by Toualy et al. (2015) and the WASP project (Forristall et al. 2013; Olagnon et al. 2004; Prevosto et al. 2013) found different results with respect to the origin of swells arriving at the coast of West Africa. Whiles findings of the WASP project from the hindcast data indicated that swells originate from two geographic zones, i.e. south swells from the South Atlantic mainly between latitudes 40oS and 60oS, and north swells from the North Atlantic off North America, Toualy et al. (2015) found that swells arriving on the West African coast are generated in the Southern Ocean and then propagate from south to north in the South Atlantic Ocean, before turning south-west to north-east close to the coast. Although findings from these studies share some similarities, the ‘agreements’ are not definite. These are as a result of several inconsistencies which could have affected the outcomes of the findings of both studies and these include the following:

  1. i.

    Investigations of Toualy et al. (2015) did not span a long enough time although they used both satellite and hindcast data, and therefore could not record some observations recorded by the WASP Project.

  2. ii.

    The period of investigation by Toualy et al. (2015) did not overlap with that of the WASP Project (periods were separate).

  3. iii.

    Hindcast model data used in both studies were obtained from global datasets or data generated using global parameterisation schemes.

To address these concerns, it is important, first of all, for the hindcast data to be obtained from a regional model and the period of the data must overlap in order for the results to be more comparable.

Summary of general oceanography of the West Africa Region

The marine environment of the West Africa domain (Fig. 2), as defined in this review, is primarily made of two main current regimes. These are the Guinea Current System (GCS) located in the southern section and the Canary Current System (CCS) in the North-western section of the study area. Figure 2 shows the major currents in the West Africa region. The CCS is situated in the Canary Basin and it comprises of the Canary Current (CC) and the Canary Upwelling Current (CUC) (Hailegeorgis et al. 2021; Mason et al. 2011). The Canary Current, which covers an area between 10o N to 40o N and offshore to about 20o W, is an eastern boundary current which flows equatorwards (southward) with speeds of 10–15 cm s−1 (Zhou et al. 2000). It forms part of the North Atlantic subtropical gyre that extends from the Iberian Peninsula whose velocities have also been reported to exceed 75 cm s−1 (Gyory et al. 2005b; Vazquez et al. 2021). It is indicated to be among the four main eastern boundary upwelling regions of the global oceans that supports buoyant fishery; namely, the Canary Current Upwelling System, Benguela Current Upwelling System, Humboldt Current Upwelling System and California Current Upwelling System (Arístegui et al. 1994; Vazquez et al. 2021). The upwelling regime of the CCS is near-permanent, occurring as a result of the north-easterly alongshore Trade winds that blow over the ocean, which creates the eastern boundary current responsible for the cold upwelled North Atlantic Central Water (Lathuilière et al. 2008). The CC and a Poleward Undercurrent (PUC), which is an intense northward flow of subsurface current, forms the basic components of the Canary Current Large Marine Ecosystem (CCLME) (Pelegrí and Peña‐Izquierdo 2015). The location of the Canary Islands, forming the Canary archipelagos in the north-western part of West Africa within the CCS influence several meso-scale activities or features such as eddies and filaments, which contribute to localized export of organic carbon and nutrients offshore (Hailegeorgis et al. 2021). This archipelago consist of several volcanic islands that rise from the deep ocean at depths that exceed 2000 m (Barton et al. 1998).

Fig. 2
figure 2

Distribution of ocean currents in the West Africa region showing the Canary Current (CC), North Equatorial Current (NEC), North Equatorial Countercurrent (NECC) and the Guinea Current (GC). The map is a modification from Djakouré et al. (2014), Lathuilière et al. (2008) and Philander (2001)

Full size image

The north-western section of West Africa also comprise of the North Equatorial Current (NEC) and the North Equatorial Countercurrent (NECC). The CC empties into the NEC which also contributes to the sources of the Guinea Current (Agbakwuru and Nwaoha 2015). A recirculation gyre is developed between the NEC and the NECC. The position of the NECC and the recirculation gyre varies seasonally as with the Inter-Tropical Convergence Zone (ITCZ) (Lathuilière et al. 2008), which is a transition zone between the northerly winds and the southerly winds migrating from about latitude 9° N in January and 20° N in August (Padi et al. 2021).

The Guinea Current (GC) derives its water mass from the NECC as well as from waters of the Canary Current depending on seasonal fluctuations (Agbakwuru and Nwaoha 2015; Gyory et al. 2005a). The GC flows from Senegal down to Nigeria. The GCS forms part of the north-western Gulf of Guinea and comprise mostly of the eastward flowing Guinea Current, Equatorial Undercurrent and the South Equatorial Current (Nyadjro et al. 2021). A westward counter flowing Guinea Undercurrent is said to develop during summer underneath the GC (Giarolla et al. 2005; Nyadjro et al. 2021). The GC also features areas of upwelling with high biological production similar to those of eastern boundary currents and flows eastwards along the western coast of West Africa at about 3° N and reaching velocities of 100 cm s−1 at about 5° W (Gyory et al. 2005a). The upwelling phenomenon experienced along the northern boundary of the Gulf of Guinea is unique among other upwelling regions, in that its occurrence does not correlate sea surface temperatures with wind patterns (Gyory et al. 2005a) according to the classical offshore Ekman transport induced by winds (Djakouré et al. 2014). The phenomenon has been attributed to several factors such as Kelvin waves (Adamec and O’brien 1978), geostrophic divergence of isotherms as a result of seasonal changes (Ingham 1977), cyclonic turbulent eddies that create vertical motions (Marchal and Picaut 1977), and wind stress induced Ekman pumping (Colin 1991). This upwelling phenomenon in the Gulf of Guinea is seasonal, with a major occurrence reported to be between July and September while a minor season occurs between December and March, resulting in increased plankton production (Wiafe et al. 2008) that supports a vibrant fishery.

The West Africa marine environment is characterised by varying levels of wave heights or sea state. The wave regime varies considerably from the north-west (Canary Current region) to the south-east (Guinea Current region). Ocean waves in the CC region that reach the north-west African coast have been indicated to come from swells during winter and from locally generated wind-waves during the summer (Semedo 2018). The swells originate from the central mid- to high latitudinal North Atlantic and subsequently propagate to the coast. The wind-waves on the other hand as reported by Semedo (2018) are generated by high wind speeds that blow along the Morocco and Western Sahara coasts and triggering extreme waves that can reach heights of about 2.5 to 3 m in the summer. Waves occurring in the Gulf of Guinea within the Guinea Current Large Marine Ecosystem (GCLME) region are rather dominated by swells (Cardone et al. 1994; Prevosto et al. 2013). These swells originate from within the southern ocean (Laïbi et al. 2014; Toualy et al. 2015), although some reports also indicate of some north-west swells from the North Atlantic (Prevosto et al. 2013).

Generally, sea-state with respect to wind waves and swells affects air-sea exchange and interaction processes such as momentum, heat and mass (Semedo 2018). The seasonal variability of the air-sea interactions and oceanic processes in the CCS and the GCS thus varies and affects the distribution of heat, and precipitation in the West Africa region differently. The impact of ocean waves in the region is huge. This is evident as a contributing factor in the widespread coastal erosion experienced across the coast of West Africa, which is a major problem to which several strategies and solutions have been proposed (Adeaga et al. 2021; Alves et al. 2020; Aman et al. 2019; Appeaning Addo et al. 2020; Appeaning Addo 2014; Foli et al. 2021a, b).

The climate and meteorology of West Africa with related impacts

The climate and meteorology over West Africa is defined by the prevailing winds and pressure systems. Two high pressure systems affect West African weather and climate. These are the Azores high and the St. Helena high located in the North and South Atlantic respectively. These pressure systems mainly regulate the wind dynamics in the region, leading to the various observed seasons and impacts on ocean waves in the region.

The wind regimes over West Africa is dominated by two major air masses with different moisture characteristics that interact to determine the prevailing weather and climate. These are the south-eastern maritime (moisture-laden) air mass that originates from the Atlantic Ocean, and the north-eastern continental (dry) air mass or harmattan winds. These are also termed the trade winds. The region also experiences monsoons. The seasonal variation in the intensities of the trade winds and their interaction with the monsoon results in the seasonal displacement of the Inter-Tropical Convergence Zone (ITCZ), where the trade winds meet (Philander 2001). The ITCZ is another important feature within the atmospheric regime of West Africa that impacts on oceanic phenomena. This affects major oceanic and atmospheric processes such as the currents, albedo, humidity, precipitation and evapotranspiration with further resultant effects on ocean properties, e.g. salinity and temperature. During the northern summer, the south-east trade winds become intense, causing it to cross the equator and penetrate into the northern hemisphere, shifting the ITCZ northwards to about 10° to 15° N (Hagen 2001; Philander 2001). The result of this is an intensification in surface currents, especially the NECC which feeds into the GCS. This is also coupled with increased rainfall. Upwelling along the West African coast is also impacted during this period where lowest temperatures are recorded (Hagen 2001; Soares et al. 2019).

The south-east trade winds weaken while the north-east trade winds are strengthened during the southern hemisphere summer, resulting in an equatorwards shift of the ITCZ. This causes the weakening and disappearing of the eastward NECC from the surface currents (Philander 2001). The Guinea Current is thus reported to experience minimal transport in boreal winter and maximum transport in the boreal summer with velocities of up to about 100 cm s−1, which also corresponds with the major upwelling season (Djakouré et al. 2017). The region occupied by the ITCZ, also known as the doldrums, is characterized by very calm conditions, slow currents, low wave heights and lots of rain and lightning (Collier and Hughes 2011; Liu et al. 2020). Less turbulence is experienced in this area and thus may present safe conditions for fishing expeditions but may be less productive. This area is mostly located west of Liberia and Sierra Leone but varies with the seasons as discussed.

Remote atmospheric phenomena also impact the climate and meteorology of West Africa. Seasonal atmospheric teleconnections from the Pacific El-Niño Southern Oscillation (ENSO) warm events affect the tropical North Atlantic between the periods of November and January, leading to a basin-wide increase in sea surface temperatures (Okumura and Xie 2006). Similarly, the Atlantic Niño, which develops as a result of the weakening of the trade winds over the equatorial Atlantic in boreal summer, impacts the ITCZ resulting in enhancement of seasonal rainfall along the Gulf of Guinea coast and other areas (Tokinaga et al. 2019; Zhang and Han 2021). Remote teleconnections also play an important role in the seasonal and climatic variability of physical atmospheric, meteorological and oceanographic anomalies, which in turn impact energy transfer, biological production and the fisheries in entirety.

One issue of concern with respect to the marine environment of West Africa is pollution. The issue of marine pollution is a global one and very wide-spread. In West Africa, one of the major culprits of marine pollution is single use plastics (SUPs). Plastic waste finds its way into the ocean via several means such as direct human deposits, through waterways (i.e. rivers, streams etc.) and vehicular transport (Hardesty et al. 2016; UNEP and CSIR 2018). These are mostly as a result of indiscriminate and illegal waste dumping, poor waste management practices and inadequate infrastructure for recycling and treatment. Adam et al. (2020) provide a collection of policy interventions in West African countries addressed at curbing marine pollution as a result of single use plastics. The prevailing lack of knowledge on the impacts of marine litter on the environment and marine ecosystems by the general public is also a contributing factor to the continued pollution of the oceans with plastics (UNEP 2009). The solution to marine pollution in the region relies largely on the willingness and readiness of West African states to implement strategies and policies that have already been put in place to reduce marine pollution.

An important phenomenon worth mentioning is the occurrence of marine heat waves (MHWs) in the West Africa region. MHWs are a rare phenomenon in the region, however, they may occur as high sea surface temperatures persist for longer periods. The influence of ENSO and the occurrence of the Atlantic Niño may lead to MHWs. Possible areas of occurrence of MHWs in the region could be found off the coast of Senegal as well as areas within the Benguela Upwelling System (BUS). These events have been referred to as the Dakar Niño (Holbrook et al. 2019; Oettli et al. 2016) and the Benguela Niño respectively (Imbol Koungue et al. 2021; Rouault et al. 2007; Shannon et al. 1986). The occurrence of the Benguela Niño may influence ecological activities within the Gulf of Guinea region as heat is advected from the Benguela region to nearby areas.

The Dakar Niño, which is a coastal marine heatwave that develops in the west coast of North Africa (Holbrook et al. 2019) occurs as a result of weakening of alongshore winds which reduce costal upwelling and surface evaporation, which is further impacted by net surface heat flux by solar radiation, leading to anomalously high sea surface temperatures (Oettli et al. 2021). The occurrence of MHWs off the coast of Senegal (Dakar Nino) was first identified by Oettli et al. (2016) from data collected between 1982 and 2011 where six anomalously warm events were identified. MHWs in the region may impact upwelling phenomena as well as precipitation, resulting in the persistence of low-nutrient, low oxygen and reduction in biological productivity in the ocean.

Conclusions

Ocean waves affect several atmospheric and oceanic phenomena just as they impact human activities both offshore and onshore. Ocean state is largely influenced by wave phenomena which is a characteristic of a chaotic system, although not unpredictable. The projection of ocean surface waves has progressed through several stages over time. Current generation of wave model physics provide better and accurate projections of ocean waves, providing more reliable information for decision making on construction of marine infrastructure, climate predictions, as well as for safety at sea. Despite some levels of inaccuracies in sea state or ocean wave projections, the need for such information has become paramount, considering the increasing levels of developments and reliance on the marine and coastal environment.

The West Africa marine environment is diverse and unique as compared to other regions. The unique oceanography, meteorology and climate of this region requires extensive knowledge and understanding through research to enable the sustainable exploitation of the numerous resources in this region for blue growth towards the attainment of the Blue Economy agenda that is being pursued globally, regionally and by individual nations through several policy initiatives. The region has relied mostly on ocean wave products generated from global parameterization physics for research and operational use, which can present some challenges of inaccuracies on a regional or local level. It is therefore important to investigate appropriate regional and local wave propagation physics or schemes for the generation of wave data and products for the region so as to minimize the errors associated with the use of global data or data generated using global parameterization physics for the region.