Role of hydrodynamics in shaping chemical habitats and modulating the responses of coastal benthic systems to ocean global change

Marine coastal zones are highly productive, and dominated by engineer species (e.g. macrophytes, molluscs, corals) that modify the chemistry of their surrounding seawater via their metabolism, causing substantial fluctuations in oxygen, dissolved inorganic carbon, pH, and nutrients. The magnitude of these biologically driven chemical fluctuations is regulated by hydrodynamics, can exceed values predicted for the future open ocean, and creates chemical patchiness in subtidal areas at various spatial (µm to meters) and temporal (minutes to months) scales. Although the role of hydrodynamics is well explored for planktonic communities, its influence as a crucial driver of benthic organism and community functioning is poorly addressed, particularly in the context of ocean global change. Hydrodynamics can directly modulate organismal physiological activity or indirectly influence an organism's performance by modifying its habitat. This review addresses recent developments in (i) the influence of hydrodynamics on the biological activity of engineer species, (ii) the description of chemical habitats resulting from the interaction between hydrodynamics and biological activity, (iii) the role of these chemical habitat as refugia against ocean acidification and deoxygenation, and (iv) how species living in such chemical habitats may respond to ocean global change. Recommendations are provided to integrate the effect of hydrodynamics and environmental fluctuations in future research, to better predict the responses of coastal benthic ecosystems to ongoing ocean global change.

knowledge gap needs to be filled because the magnitude of water velocity can strongly modify the outcome of experiments testing the impacts of ocean global change on benthic marine organisms (Comeau et al., 2014(Comeau et al., , 2019Cornwall et al., 2014).
Temporal and spatial fluctuations in chemical conditions in shallow waters mainly result from the interaction between biological activity and hydrodynamic forcing Kapsenberg & Hofmann, 2016;Wahl et al., 2018). Biological activity is the main driver of chemical variations in highly productive systems that are dominated by 'engineer species' (see Box 1) such as seagrass meadows, kelp forests, coral reefs or mollusc beds (Falkenberg et al., 2021). At high population densities, engineer species have the capacity to strongly modulate the chemical and physical conditions of their surrounding waters (Gutiérrez et al., 2003(Gutiérrez et al., , 2011Hurd et al., 2014;Kregting et al., 2013;Pujol et al., 2019;Turk et al., 2015;Unsworth et al., 2012). Physiological processes such as respiration, photosynthesis, calcification or excretion shape the chemistry of coastal waters via their influence on inorganic carbon fluxes (e.g. Bordeyne et al., 2015;Chauvaud et al., 2000Chauvaud et al., , 2003Martin et al., 2007a), nitrogen exchange (e.g. Martin et al., 2006Martin et al., , 2007b, seawater oxygen content (e.g. Smith et al., 2013) and carbonate chemistry variables including pH and alkalinity (e.g. Ninokawa et al., 2020;Takeshita et al., 2018). The intensity of biological activity and the temporal and spatial scale of effects are mainly dependent on the size, biomass, morphology and metabolic rate of organisms and communities in relation to the water volume surrounding them ( Figure 2; Lesser et al., 1994;Lowe & Falter, 2015;Stewart & Carpenter, 2003). Biologically driven fluctuations ( Figure 2) occur on spatial scales ranging from small (µm to cm) to large (m to hundreds of m), and temporal scales of variation are generally faster at small (seconds) compared with large (hours to days) spatial scales (Boyd et al., 2016;Kapsenberg & Cyronak, 2019;Wahl et al., 2016).
Metabolic processes of the organisms are in turn influenced by their physiology, which is regulated by the amount of energy available in the habitat in terms of light, inorganic or organic sources, as well as by other physical factors such as temperature, salinity and hydrodynamics.
An increasing number of publications emphasizes the importance of incorporating fluctuating regimes into experiments testing the responses of organisms and communities to ocean global change, in particular to ocean warming, acidification and deoxygenation (Boyd et al., 2016;Britton et al., 2019;Rivest et al., 2017;Wahl et al., 2016). Furthermore, at this stage, we do not understand how alterations in the metabolism of engineer species due to ocean global change will feedback to shape their surrounding water chemistry, in relation to hydrodynamics (e.g. Hirsh et al., 2020;Noisette & Hurd, 2018;Wahl et al., 2018).
In this review, we focus on (i) the influence of hydrodynamics on the biological activity and performances of engineer species, especially under ocean global change. We then describe (ii) specific chemical habitats resulting from the interaction between F I G U R E 1 Hydrodynamics in subtidal coastal zones. Different hydrodynamic processes occur in coastal areas (left of this sketch). In subtidal zones, hydrodynamic processes generally result in an overall unidirectional flow, which influences benthic organisms and communities. When the flow touches a solid surface, shear stress makes the flow laminar at the surface of the organism and a diffusive boundary layer (DBL) can build up, creating specific chemical conditions at the surface of the organism, different from the bulk seawater (upper right sketch). Engineer organisms (e.g. seaweeds, seagrasses, corals, molluscs) can attenuate the flow depending on their complexity and density. By aggregation of layer on layer of DBLs, a canopy-or community-boundary layer can form (right of this sketch). More details on flow structuring by seagrass and coral communities can be found in Nepf (2012) and Lowe and Falter (2015), respectively engineer species metabolism and hydrodynamics, where environmental fluctuations are different than in the bulk seawater. We highlight (iii) the importance of these specific chemical habitats in providing refugia against ocean acidification and deoxygenation but also (iv) point out how they can facilitate the resistance of populations or decrease species' fitness, depending on species' capacities to respond to large abiotic fluctuations.
We deliberately do not address in this review indirect hydrodynamic effects on shifts in species' interactions and interactive impacts of multiple drivers fluctuating asynchronically (see Wahl et al., 2021 for instance).

| Hydrodynamics in coastal subtidal environments
Compared with air in terrestrial systems, the higher density and viscosity of seawater causes lower rates of molecular diffusion of essential nutrients and gases. The result is that the fluid medium in which marine organisms grow exerts greater regulation over biological and F I G U R E 2 Environmental fluctuations and chemical environments at different scales. Organisms modulate their chemical environment at the small-scale up to larger ecosystem scales of hours to days (rights panels). The conceptual basis underlying these fluctuations are illustrated in the left panels. We outline how photosynthetic organisms (i.e. primary producers) modulate their chemical environment within the diffusive boundary layer (DBL, upper panels) up to the community scale (cBL, lower panels). These changes can be large and are usually linked to biomass and/or flow conditions. The amplitude of environmental fluctuations is positively related to increases in biomass (green gradient in bottom left graph shows increasing biomass to the right) and negatively to the intensity of water flow (blue gradient in bottom left graph illustrates increasing flow to the left with constant biomass). Small-scale variation of pH (upper right panel) was measured within the DBL of calcifying red algae as a function of flow (data are adapted from Cornwall et al., 2013, occurring within minutes). At larger scales (lower right panel), these variations are generally slower and less intense (data are adapted from  changes measured proceeded at this scale over a full day) community performance (Denny, 2015;Sand-Jensen & Krause-Jensen, 1997;Vogel, 1999). Therefore, water motion, which counteracts reduced diffusion rates, is a fundamental regulator of biological activity, at scales ranging from that of the organism to that of entire ecosystems (Nowell & Jumars, 1984). This is particularly the case in coastal habitats, where hydrodynamic processes encompass waves, tidal cycles, storm surges, up-and downwelling. In some coastal ecosystems such as the intertidal zone, wind-driven flows create massive and powerful waves and oscillatory currents (Hurd, 2000). However, the majority of subtidal environments are under the influence of a dominant flow, resulting from cumulative hydrodynamic processes (Nowell & Jumars, 1984; Figure 1).

| Water flow shapes velocity and DBLs at the organism's surface
The limitation of physiological processes by flow is closely related to what happens at the surface of the organism (De Beer et al., 2000;Hurd, 2015;Kühl et al., 1995;Noisette & Hurd, 2018). The theoretical background of the behavior of a fluid when it touches a solid surface has been reviewed previously (Denny, 2015;Hurd, 2000;Vogel, 1999). Briefly, water flowing over marine organisms generates frictional drag at their surface (Hurd, 2000;Vogel, 1999). A velocity gradient (VBL) forms, with zero velocity at the solid surface due to the no slip condition, and a maximal thickness at a velocity of 99% of the bulk seawater flow overhead. In the viscous sublayer of the VBL, where the flow is laminar, molecules and gases move by the slow process of 'diffusion' from high to low concentrations. This sublayer is termed the 'DBL' (Box 1; Figure 1; Vogel, 1999). The movement of all metabolically important dissolved gases (e.g. CO 2 , O 2 ), dissolved inorganic nutrients (e.g. nitrate, ammonium, phosphorous, bicarbonate), metabolites (e.g. reactive oxygen species, refractory carbon), and metabolic waste (e.g. urea) to and from an organism surface, across the DLB, is via diffusion. Although the seawater flow velocity controls the formation of a DBL and its thickness, the concentration gradient in the DBL is driven by the organism's metabolic activity, the concentration of dissolved substances in the bulk seawater at the outer edge of the DBL, and the thickness of the VBL (Denny, 2015;Stevens et al., 2001). Temperature also changes the transport of dissolved materials across the DBL as it affects fluid viscosity, gas solubility and the diffusion coefficient (Denny, 2015;Vogel, 1999). Increasing temperature usually decreases the DBL thickness and increases diffusion rates, leading to a steeper concentration gradient in the DBL (Wahl et al., 2016). The effects of temperature on the physical properties of the DBL are superimposed to direct effects on organism's metabolism (Brown et al., 2004) and result in specific chemical conditions surrounding the organism.

| Flow speed impacts metabolic rates and physiology
Physiological and biological rates follow a quadratic response curve as a function of flow speed: they improve with increasing mainstream speed until a maximal rate is reached, when enzymes or organs (like feeding appendices) start being saturated (Hurd et al., 1996;Pujol et al., 2019;Stewart & Carpenter, 2003;Wheeler, 1980). Related to thick DBLs, slow flows can restrict biological performance by impairing the supply of essential dissolved inorganic and organic substances needed for physiological processes. When flow speed increases, convection and advection improve mass transport and decrease DBL thickness, which enhances the supply of essential dissolved substances and the removal of harmful ones, allowing an improvement of an organism's performance. When flow exceeds the optimal velocity, it can be detrimental for the organism, triggering new energetic trade-offs for maintenance or repair (see figure 4 in Pujol et al., 2019). It is important to remember that the concept of fast and slow flows is relative to the organism studied and the hydrodynamics they experience naturally in their habitat.
Most studies reporting the effects of hydrodynamics on the metabolism of engineer species are related to photosynthetic primary producers, hereafter called "primary producers" (Hurd, 2017).
Seawater velocity has been shown to positively impact primary production rates and nutrient uptake up to the optimum in turf (e.g. Carpenter & Williams, 2007), larger 'macrophytes' such as seagrasses and seaweeds (e.g. Barr et al., 2008;Carpenter et al., 1991;Kregting et al., 2011;Mass et al., 2010) and corals (e.g. Atkinson & Bilger, 1992;Lesser et al., 1994). Indeed, in slow flow, 'mass transfer limitation' (see Box 1) can affect the ability of macrophytes to acquire and use essential dissolved substances, limiting their metabolism and likely causing a reduction in production and growth (Carpenter & Williams, 2007;Pujol et al., 2019). In corals, slow flow can reduce production and calcification rates, when oxygen accumulates in coral tissues at the peak of photosynthesis around midday (Dennison & Barnes, 1988;Jokiel, 1978). The increase in flow speed, thus, turns out beneficial until it becomes too fast and detrimental.
For instance, the productivity of marine macrophytes that use an external carbonic anhydrase in their carbon-acquisition mechanism may decline in fast flows as this enzyme could be washed away from the seaweed's surface (Enríquez & Rodríguez-Román, 2006). In corals, fast flow could increase the cost of energetic expenses of light acclimated corals (Patterson et al., 1991) by affecting their carbon translocation process (Edmunds & Davies, 1986). Flow speed can also influence respiration, calcification, and/or the feeding efficiency of heterotrophic engineer species such as cnidarians, (Dennison & Barnes, 1988;Patterson et al., 1991;Rex et al., 1995;Stambler et al., 1991), bryozoans (Hermansen et al., 2001;Okamura, 1985), or molluscs (Saurel et al., 2007;Wildish et al., 1987), affecting their growth (Jokiel, 1978), and likely the local conditions in their surrounding seawater (De Beer et al., 2000). To increase their performance, organisms including seaweeds, mussels, and barnacles, can be morphologically adapted to the prevalent flow regime of their habitat (Hurd & Stevens, 1997;Marchinko & Palmer, 2003). However, the impact of flow speed on morphological changes (e.g. body size, surface topography) is not developed here as it is considered out of the scope of this review.

| Hydrodynamics can modulate organism responses to ocean global change
An increasing number of studies highlight how prevalent hydrodynamics are in modulating an organisms' responses to ocean global change, including warming, hypoxia or deoxygenation and ocean acidification (Comeau et al., 2014(Comeau et al., , 2019Egea et al., 2018;Ho & Carpenter, 2017;Hurd, 2017). For example, Nakamura and Yamasaki (2005) showed that two branched coral species bleached less and grew better under fast (20 cm s −1 ) compared with slow flow (3 cm s −1 ), during high sea surface temperature events, which lasted for several weeks. By increasing the diffusion rate and reducing the oxidative stress, fast flow may enable these corals to recover from thermal stress (Finelli et al., 2006;Nakamura & Van Woesik, 2001).
Under ocean acidification scenarios, corals and coralline algae have been shown to maintain positive levels of net calcification regardless of the bulk seawater pH decrease, in flows increasing from 2 to 10 cm s −1 . The authors hypothesized that this positive effect of fast flow is likely related to an alleviation of mass transfer limitation, which facilitates the uptake of carbonate/bicarbonate ions as well as the export of protons to and from calcification sites (Comeau et al., 2014).  showed that coralline algae and sea urchins living in slow flows had greater pH variations than in fast flows, and hypothesized that this would support greater physiological flexibility to adjust to environmental changes. This study also suggested that slow flow could create favorable conditions for calcification, a hypothesis later supported by Cornwall et al. (2013Cornwall et al. ( , 2015

| Metabolic processes that drive DBL chemical fluctuations
The nature, rate and rhythm (e.g. diel cycles) of metabolic processes of an organism are the main cause of the chemical fluctuations that occur in the DBL (e.g. Figure 1). Photosynthesis and respiration affect oxygen, CO 2 , bicarbonate (HCO 3 ), hydroxide (OH − ), and proton (H + , linked to pH) concentrations within the DBL. In light, photosynthesis by primary producers strongly increases oxygen concentration at photosynthetically active surfaces (e.g. Brodersen et al., 2015;Kaspar, 1992;Lichtenberg et al., 2017;Mass et al., 2010;Shashar et al., 1993). Additionally, changes in CO 2 : bicarbonate ratio are observed within the DBL (De Beer et al., 2000;Hofmann et al., 2016). These ratio changes depend on the dissolved inorganic carbon uptake strategy of the organism, that is, CO 2 -only users versus species that additionally use bicarbonates Raven, 1997;Van Der Loos et al., 2019). Around most photosynthetic organisms, surrounding seawater pH usually increases over the light period due to (i) CO 2 uptake, which alters the seawater carbonate system, (ii) the accumulation of OH − ions within the DBL, as a biproduct of bicarbonate uptake used for photosynthesis (Fernández et al., 2014), and/or (iii) light-induced proton pumps . Conversely, respiration releases CO 2 , which accumulates within the DBL at night for photosynthetic organisms . The release of CO 2 influences the seawater carbonate system, leading to a pH decrease within the DBL (Hurd, 2000). For heterotrophs, respiration tends to almost exclusively lower pH in the DBL compared with the bulk seawater, regardless of day and night.
Calcification and carbonate-dissolution processes at the surface of calcareous species also change the local chemical environment of the DBL by influencing calcium content, carbonate chemistry equilibrium and total alkalinity (De Beer et al., 2000;Ninokawa et al., 2020;Roleda et al., 2012;Wolf-Gladrow et al., 2007).
Nutrient concentrations, nitrogen in particular, can also fluctuate within the DBL, driven by nitrogen uptake in primary producers (Fernández et al., 2017), and ammonium or urea excretion in invertebrates (Hurd et al., 1994). Fluctuations in dissolved nitrogen species can also affect the carbonate chemistry within the DBL, as ammonium and nitrate concentrations influence the alkalinity and the pH of seawater (Wolf-Gladrow et al., 2007). Nitrate assimilation results in a pH increase at the cell surface, whereas ammonium assimilation leads to a pH drop due to influx and efflux of protons, respectively (Raven, 1981;Raven & Michelis, 1980). However, Fernández et al. (2017) showed that nitrogen assimilation did not change pH at the blade surface of the giant kelp Macrocystis pyrifera, emphasizing that the co-occurrence of different processes affecting the pH can repress their individual effects.

| From DBLs (µm-mm) to the formation of larger (cm-m) chemical habitats
The biologically-driven chemical fluctuations occurring in the organism's DBL range from µm to mm depending on the flow velocity (e.g. Hendriks et al., 2017;Noisette & Hurd, 2018;Noisette et al., 2020).
These micro-habitats created at the direct surface of organisms represent specific niches with characteristics completely different from the bulk seawater. They provide habitat for microorganisms, hold complex animal assemblages, and may constitute a preferred settlement habitat for bacteria, small larvae and spores (Irwin & Davenport, 2010;Wahl et al., 2016). In favorable conditions, that is, at very slow flow and high metabolic activity, the thickness of the boundary layer at the surface of a macrophyte can even reach a few centimeters (Raven & Hurd, 2012;Wahl et al., 2016), being effective as habitat for also larger epibionts (Saderne & Wahl, 2013).
In dense aggregations of engineering species such as algal canopies or mollusc beds (Ninokawa et al., 2020;Rosman et al., 2010;Shi et al., 1995), organisms can attenuate the flow speed making the VBLs coalesce (Kregting et al., 2021). In there, biological processes can drive notable chemico-physical differences between seawater in-and outside the formed habitat, leading to the build-up of larger canopy-or community boundary layers from centimeter to meter ( Figure 3; Cornwall et al., 2013;Hurd, 2015;Ninokawa et al., 2020).
The formation of such thick boundary layers depends on the engineer species' morphology and the complexity of the habitat formed Kregting et al., 2021;Stewart & Carpenter, 2003). They preferentially build up when the engineer species are large and complex enough to sufficiently slow down the water flow within the habitat (Figure 1; Gutiérrez & Jones, 2006;Jackson & Winant, 1983;Lowe & Falter, 2015).

| Canopy-boundary layers
To date, research has focused on macrophytes as engineer organisms able to greatly attenuate the flow inside the canopy and changing the hydrodynamic regime for fleshy and calcareous algae in the understory (Cornwall et al., 2015;Jackson & Winant, 1983;Nepf, 2012;Nishihara et al., 2015;Unsworth et al., 2012). In and below the canopy, the seawater residence time increases, while mixing with the water column decreases. The formation of a canopy-boundary layer can reduce the exchange rate of gases, the replenishment of ions and the removal of metabolic by-products, influencing the amplitude and the frequency of the chemical fluctuations within (James et al., 2019;Koweek et al., 2017). Chemical conditions within these habitats are thus defined by the balance between hydrodynamics and biological processes (Ninokawa et al., 2020). For instance, within dense macrophyte communities, pH variability may exceed one pH unit per day (Middelboe & Hansen, 2007;. These chemical fluctuations can be beneficial or detrimental to the physiological performance of the associated organisms living on the engineer species or in understory (Bergstrom et al., 2019;Cornwall et al., 2013;Garrard et al., 2014;Semesi et al., 2009). Experiencing these fluctuations can therefore modulate an organism's efficiency in coping with ongoing ocean global change such as ocean acidification (Falkenberg et al., 2021;Kapsenberg & Cyronak, 2019;Wahl et al., 2018). This has recently generated much interest in characterizing these chemical habitats and in assessing the responses of "associated species" (see Box 1).

| Ecosystems influenced by upstream engineer species
Water masses transformed by the intense metabolic activity of some dense populations of engineer species can be transported by advection to downstream adjacent ecosystems ,

F I G U R E 3
Refuge habitats created by engineer species. The interaction between hydrodynamics and biologically driven fluctuations creates chemical micro-habitats of sizes from µm to m, which can provide refuge in the context of global change (particularly ocean deoxygenation and acidification). The chemical conditions can be mitigated around the engineer organism (individual diffusive boundary layer, DBL and canopy or community boundary layer, cBL), also referenced to by Cyronak et al. (2018) as the within-system buffering action. The advection of water masses from main producers' communities to influenced areas located downstream, represent the downstreamsystem buffering , which could benefit other communities, yet largely depending on the water to biomass ratio hereafter called influenced-areas (Figure 3; Box 1). This advective transport was shown to affect the local carbonate chemistry of influenced-areas (Krause-Jensen et al., 2015;, increasing for instance the seawater saturation state of carbonate and likely mitigating ocean acidification effects (Anthony et al., 2013;Unsworth et al., 2012). This phenomenon has been evidenced for ecosystems adjacent to seagrass meadows (Camp et al., 2016;Unsworth et al., 2012), seaweed beds (Delille et al., 2009;Krause-Jensen et al., 2015), coral reefs (James et al., 2019), and mangroves (Yates et al., 2014). Compared with boundary-layer chemical habitats, the local chemistry of influenced-areas is even more dependent on biomass to water ratios and water exchange as current speed and direction (Hirsh et al., 2020;Wahl et al., 2018). Slow flow reduces the exchange rate and dilution of a water body, which permits a higher amplitude in fluctuating variables. Conversely, fast currents result in much lower excursions from the mean (Frieder et al., 2012;Koweek et al., 2017).
In tropical ecosystems, models and in situ studies showed that the calcification of coral communities could be maintained or increased due to the biologically-driven increase of the mean pH provided by upstream macrophyte beds (Camp et al., 2016;Mongin et al., 2016;Unsworth et al., 2012). It has also been shown that dense oyster beds could modify the alkalinity of an entire bay, likely affecting influencedareas adjacent to these mollusc beds . In addition, the capacities of macrophytes in buffering the negative effects of hypoxic events and/or long-term ocean acidification on economi-  areas can be greater than mean changes projected for the bulk seawater by the end of the century . Hence chemical refugia can provide mitigation and temporal buffer, likely alleviating stress relative to conditions without fluctuations (Wahl et al., 2015).

| Chemical habitats as refugia
The size of the chemical habitat has to be considered to fully assess its refuge capacity. Some authors suggest that refugia must be large enough to manage a small population (Kapsenberg & Cyronak, 2019;Morelli et al., 2016), while we consider that refugia associated with DBLs are also of importance to very small species (e.g. bryozoans) and some early life stages of larger species (e.g. mollusc larvae or macrophyte spores), which are critical for species persistence.
Larger-scale community/canopy-boundary layer habitats can provide wider refugia allowing larger associated species to live, reproduce and interact within these chemical habitats.

| Ocean acidification refugia provided by primary producers
Biologically-driven processes creating refugia have recently become relevant in understanding species' responses to ocean acidification, focusing mainly on photosynthetic activity by macrophytes  James et al., 2019;. The net effect of macrophytes is, thus, an overall rise in the diel mean pH Wahl et al., 2018) and a reduction of the duration of low pH exposure . In addition, the range of pH fluctuations driven by macrophytes metabolism along diel cycles could be narrowed in a global change scenario. For instance, Noisette and Hurd (2018) showed that oxygen and pH fluctuations within the DBL of kelp blades would be less broad and shifted towards higher means under ocean acidification scenarios compared with current conditions. Field studies in temperate and tropical ecosystems dominated by macrophytes have shown that the increase in pH was correlated to an increase in the carbonate saturation state (e.g. Koweek et al., 2017;Pacella et al., 2018;Semesi et al., 2009). Macrophytes could then mitigate the negative impacts of ocean acidification on marine calcifiers in their vicinity by providing buffer from corrosive conditions (Short et al., 2015;Silbiger & Sorte, 2018), as demonstrated on corals (Kleypas et al., 2011;Unsworth et al., 2012), molluscs (Young & Gobler, 2018) and coralline algae (Bergstrom et al., 2019;Cornwall et al., 2014;Short et al., 2015). These local positive effects of macrophyte communities against ocean acidification have been rarely modelled (Pacella et al., 2018) and studied in the field  and have to be considered with caution. Local natural pH variations have to be carefully characterized (Van Dam et al., 2021a, 2021b to understand if this buffering effect might help associated species to endure stressful future pH shifts, as experimentally shown in seagrass communities (Cox et al., 2017;Guilini et al., 2017;. For more details about the buffering effects of marine macrophytes, see Falkenberg and collaborators (2021) who reviewed the role of seagrass and seaweeds habitats as refugia from ocean acidification and their implementation in management plans. A list of case-studies reporting the effect of engineer species on the chemistry of their surroundings in the context of ocean acidification and deoxygenation is reported in Table 1.

| Deoxygenation refugia provided by primary producers
Oxygen enrichment as a result of photosynthesis by primary producers such as seaweeds (Irwin & Davenport, 2002, seagrass meadows (Koopmans et al., 2018;Saderne et al., 2015) and around corals (Shashar et al., 1993;Smith et al., 2013), could also provide refugia in the context of the ocean deoxygenation (Altieri et al., 2021;Keeling et al., 2010;Laffoley & Baxter, 2019). The net daytime increase in oxygen concentration may be beneficial for heterotrophs as it would increase the partial pressure of oxygen (pO 2 ). Increasing pO 2 facilitates gas exchanges and, thus, also decreases the energy consumption caused by anti-stress mechanisms (Ramajo et al., 2016). For some primary producers, this increase in oxygen level might cause an increase in photorespiration: high pO 2 may decrease the affinity of the Rubisco for carbon dioxide and, thus, reduce photosynthesis (Mass et al., 2010;Nishihara et al., 2015). Nevertheless, numerous coastal regions are predicted to face a future drop in oxygen levels and/or encounter upwelling events that regularly shoal deoxygenated waters (Grantham et al., 2004;Omstedt et al., 2014;Wei et al., 2019). In these regions, the net and, in particular, the daytime increase in oxygen concentration mediated by macrophytes may be particularly beneficial for heterotrophs and might provide important refugia. In tropical ecosystems, which face important deoxygenation episodes (Altieri et al., 2021), an improved understanding of the oxygen variation in boundary to canopy layer habitats could help to better understand coral responses to accelerating future deoxygenation (Hughes et al., 2020).

| Boundary layer fluctuations as an additional stress
The refuge capacity of boundary layer chemical habitats must be evaluated by balancing any benefits of periodic relief from environmental stress, with any potential deleterious effect due to intensifying exposures to harmful conditions (alternating phases of high stress and recovery; see figure 1 in Wahl et al., 2015). For instance, conditions around macrophytes can be improved compared with the surrounding bulk seawater during day (light) time but also become unfavorable at night Koweek et al., 2017;Pacella et al., 2018). Even though daily net increases in pH and oxygen levels are recorded around macrophytes, often steep and strong changes over relatively short time scales may not always provide mitigation from mean stress but may challenge an organisms' capacity for physiological adjustments (Middelboe & Hansen, 2007;Wahl et al., 2018). These fluctuating conditions (see Figure 2b) may actually add an additional component to the stress portfolio of the environment, strongly driven by the amplitudes of fluctuations, but also their rates of change, in diurnal to weekly to seasonal patterns of shifts (Sabine, 2018;Wahl et al., 2016).
Intensity and duration of such deviations from the mean would, therefore, determine whether fluctuations alleviate or aggravate stress in boundary layer chemical habitats, relative to conditions without variability. Overall, fluctuation effects may be determined by different aspects: (i) Jensen's inequality, mathematically differentiating the impacts from variable versus stable environmental conditions (Ruel & Ayres, 1999), (ii) relaxation from stress might provide temporal release (refuge) from stress allowing physiological or community recovery (Wahl et al., 2016), and (iii) intense peak stress that might drive organisms to their limits, leading to selection, mortality, and/or detrimental population collapse (Wernberg et al., 2016). Typical performance curves in response to environmental variables such as temperature or oxygen concentration have mono-modal quadratic or modified Gaussian shapes (Angilletta, 2006;Woodin et al., 2013). Fluctuations may, therefore, enhance or decrease an organisms' performance depending on whether fluctuations are based within the ascending or the descending part of the curve (Ruel & Ayres, 1999). In the stress range of the environmental variables (outside the optimum), the net effect of the fluctuations depends on the position of the mean relative to the energetic break-even point, and the proportional time the variable resides below or above this point (Camp et al., 2016;Woodin et al., 2013). To understand the effects of fluctuating regimes on organisms living within chemical habitats, it is important to determine whether their biological responses are more dependent on the mean condition of exposure or the time they spend below a critical threshold (Sabine, 2018).

| Fluctuations can promote resistance to ocean global change
Recent reviews, out of the context of boundary layer chemical habitats, compiled the effects of fluctuations on the physiology of benthic organisms in the context of ocean global change (Boyd et al., 2016;Rivest et al., 2017). With respect to ocean acidification scenarios, pH fluctuations may improve survival  or increase the growth rate of fleshy algae (Britton et al., 2016, 2019). These variable regimes might also lead to an increased resilience of calcareous species because of an acclimation to a wider pH range than normally experienced (Semesi et al., 2009;Short et al., 2015). In the context of fluctuations based on a diel light cycle, it has been observed that species might develop mechanisms to endure phases of stress and relaxation by shifting their metabolic rates (Price et al., 2012). A study from TA B L E 1 Case-studies illustrating the effects of engineer organisms in shaping the chemistry of their surrounding environment with specific reference to ocean acidification and/or deoxygenation context. Environmental parameters related to ocean acidification and/or deoxygenation mitigation or worsening are reported (Ω, calcium carbonate saturation state; DIC, dissolved inorganic carbon; DO: dissolved oxygen; pCO 2 , carbon dioxide partial pressure; TA, total alkalinity). References in italic specifically refer to studies mentioning deoxygenation   Wahl et al. (2018) has suggested that mussels were able to shift the majority of their costly physiological processes such as calcification to day times, when the surrounding chemical conditions were the most favorable, driven by the biological activity of the habitat-forming bladder wrack (Fucus vesiculosus).
Thus, chemical habitats created by biologically driven fluctuations may alter the intensity and direction of the effects caused by superimposed ocean global change like ocean acidification and deoxygenation (Gunderson et al., 2016). It could in the long run facilitate a "hardening" of populations throughout processes of individual and transgenerational acclimation and adaptation (Bulleri et al., 2018;Eriander et al., 2016;Pansch et al., 2014;Rivest et al., 2017). Fluctuating regimes may select for physiological generalists' that can endure a wide range of environmental variables through acclimatization processes, that is, the change in physiological performances in response to environmental changes driven by phenotypic plasticity (Kroeker et al., 2020). In fact, organisms from environments characterized by greater heterogeneity have more phenotypic plasticity, that is, an ability to express different phenotypes based on the same genotype (Boyd et al., 2016;Pansch et al., 2014). However, the transient transgressions of tolerance thresholds followed by relaxation periods may not only selectively favor high phenotypic plasticity but may also select for more robust genotypes

| Can heterotrophs provide refugia against ocean global changes?
Although photosynthetic organisms are now recognized to provide potential refugia in the context of ocean acidification and deoxygenation, the role of heterotroph engineer species is barely studied (Table 1). Their contribution to the local raise of pCO 2 is even considered detrimental (Ninokawa et al., 2020). However, changes in carbonate chemistry driven by dense populations of calcareous heterotrophs such as bivalves and reef-forming corals may induce specific conditions around them such as alkalinity regeneration (Waldbusser et al., 2011). These local changes in chemistry can be perceived as favorable cues for larval recruitment in zones submitted to intense pH decrease or drastic changes in alkalinity (Green et al., 2013;. In addition, shell debris enrichment was shown to increase local pH and aragonite saturation state likely providing a chemical refuge that would promote a better recruitment of mollusc larvae (Greiner et al., 2018). Nevertheless, measurements of vertical gradients within and above mussel beds over a range of several flow velocities showed that reduction in pH and oxygen concentration inside dense mussel beds may negatively impact the species associated (Ninokawa et al., 2020). This latest study counteracts the few previous outcomes supporting that heterotroph engineer species may provide refugia to face ocean global change. Better exploring the frequency and the amplitude of the chemical fluctuations generated by invertebrate reefs and their beneficial and deleterious effects for the associated species is, therefore, an avenue worth to be investigated in the future.  (Falkenberg et al., 2021): seasonality of the variations is rarely assessed. The spatial extent of chemical refugia in situ is complex to assess  and often prevented by technical and logistical limitations. To overcome these identified gaps and to provide accurate data for better understanding the future of coastal benthic ecosystems, we advise the scientific community to pay attention to the following main points:

| CON CLUS ION
1. Accurately report the intensity and type of hydrodynamics (e.g. laminar vs. turbulent, unidirectional, oscillatory, wave-induced) in an experiment or in a field study, as water motion controls individuals' physiology and might be partly accountable for the observed wide variations in responses of organisms to ocean global change. 3. Better characterize the temporal and spatial abiotic fluctuations occurring in the field  and incorporate them into experimental designs of studies addressing the effect of ocean global change. Especially for perennial species, more attention should be paid to seasonal variations because both environmental baselines (i.e. means) and their distance to specific optima change with seasons Wahl et al., 2021). These measurements require new technical developments in autonomous loggers and infrastructure (Pansch & Hiebenthal, 2019), as well as publication of best practice guides to standardize procedures (Lorenzoni et al., 2017).
4. Dedicate more attention to the chemical habitats resulting from the interaction of flow and metabolism (e.g. James et al., 2019) in order to better characterize the magnitude of buffering/amplification, the area impacted, and the frequency of the fluctuations generated.
Addressing these different points more carefully will be key to understand the complicated feedback loops within coastal benthic communities, often dominated by engineer species with high metabolic rates, under ocean global change. It could also permit to elucidate whether or not biologically-driven chemical habitats (boundary-layer associated and influenced areas) can fulfill the role of refugia or merely provide additional stress for organisms that leads to strong selection pressure in the ocean of tomorrow.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data sharing not applicable to this article as no datasets were generated or analysed during the current study.