Laurent M. Chérubin

Reserve design for uncertain responses of coral reefs to climate change

Rising sea temperatures cause mass coral bleaching and threaten reefs worldwide. We show how maps of variations in thermal stress can be used to help manage reefs for climate change. We map proxies of chronic and acute thermal stress and develop evidence-based hypotheses for the future response of corals to each stress regime. We then incorporate spatially realistic predictions of larval connectivity among reefs of the Bahamas and apply novel reserve design algorithms to create reserve networks for a changing climate. We show that scales of larval dispersal are large enough to connect reefs from desirable thermal stress regimes into a reserve network. Critically, we find that reserve designs differ according to the anticipated scope for phenotypic and genetic adaptation in corals, which remains uncertain. Attempts to provide a complete reserve design that hedged against different evolutionary outcomes achieved limited success, which emphasises the importance of considering the scope for adaptation explicitly. Nonetheless, 15% of reserve locations were selected under all evolutionary scenarios, making them a high priority for early designation. Our approach allows new insights into coral holobiont adaptation to be integrated directly into an adaptive approach to management.

Meddy coupling with a deep cyclone in the Gulf of Cadiz

Hydrological measurements made in July 1999 in the Gulf of Cadiz reveal the presence of two meddies, one south of Cape Saint Vincent (meddy Christine, centered at 9°45′W–35°30′N), the other one near the Moroccan shelf (meddy Isabelle, centered at 8°30′W–34°25′N), and their interaction with a deep cyclone (centered at 8°30′W–35°15′N). These meddies are medium-scale features with thermohaline anomalies of 60–70 km diameter concentrated between 750 and 1500 m depths; these anomalies reach 2 °C and 0.5 in salinity. Meddy Christine exhibits unusually strong thermohaline gradients at its periphery. Their maximum azimuthal velocities are 0.12 m/s (meddy Isabelle) and 0.10 m/s (meddy Christine). They both have a (vertically) tripolar structure in potential vorticity anomaly. The cyclone is shallower (between 600 and 1300 m) with a weaker signature of Mediterranean water (12 °C at 800 m, 36.2 in salinity between 800 and 1200 m). Its velocity maximum (0.16 m/s) is also centered near 800 m. This cyclone appears coupled with meddy Isabelle as a baroclinic dipole. This dipole tears a long warm and salty filament away from meddy Christine at 1200 m. The long-term evolution of these eddies is described by means of deep-drogued drifting buoy trajectories: while meddy Christine drifts southwestward, the two other eddies remain close to each other and move together along the continental shelf towards Cape Saint Vincent where they separate. A three-layer quasi-geostrophic model is used to evaluate the possible origin of meddy Isabelle and of the cyclone near the Portuguese coast, and to assess if their interactions, both mutual, with the domain boundaries, and with bottom topography, can account for their motion. The interaction of this dipole with meddy Christine is also quantified in this model.

Loop current ring shedding : The formation of cyclones and the effect of topography

Abstract The formation of cyclones in the vicinity of the Loop Current ring during the separation stage is analyzed in the frame of a high-resolution ECMWF daily wind-forced Miami Isopycnic Coordinate Ocean Model (MICOM) simulation. Mesoscale cyclones, observed in sea surface height maps in the vicinity of the Loop Current in a necking-down position, are found to contribute to the separation of the ring from the Loop Current as they grow between the Loop Current and the ring in the MICOM simulation. To understand the origin of the cyclones, the instability of the Loop Current idealized as an isolated vortex is studied. After noticing the cyclonic vorticity belt around the Loop Current, and based on the vertical distribution of potential vorticity in a Loop Current ring in the MICOM simulation, the linear stability of a shielded vortex is studied in the quasigeostrophic formalism. To simulate the effects of the planetary vorticity gradient and topography on the Loop Current, the nonlinear states of the ide...

Instability of the Mediterranean Water undercurrents southwest of Portugal: effects of baroclinicity and of topography

Abstract Southwest of Portugal, in situ data show that part of the mesoscale variability of the Mediterranean Water undercurrents is triggered by topographic effects (near capes and submarine canyons), and is driven by a dominantly baroclinic instability. The vertical alignment of these undercurrents observed at sites of eddy formation is indeed favorable to baroclinic instability. This instability is materialized by the formation of filaments and of small eddies in the Portimao canyon and of meddies near Cape Saint Vincent and near the Estramadura Promontory. Hydrological data and float trajectories reveal that these eddies have a baroclinic structure. Near Portimao canyon, the variation of the potential vorticity of the undercurrents again proves their sensitivity to baroclinic and also to barotropic instabilities, enhanced by the canyon. Finally, a model of stationary coastal current explains the variations of its horizontal structure over a canyon.

Ocean circulation and terrestrial runoff dynamics in the Mesoamerican region from spectral optimization of SeaWiFS data and a high resolution simulation

The evolution in time and space of terrestrial runoff in waters of the Mesoamerican region was examined using remote sensing techniques combined with river discharge and numerical ocean circulation models. Ocean color SeaWiFS images were processed using a new Spectral Optimization Algorithm for atmospheric correction and ocean property retrieval in Case-2 waters. A total of 157 SeaWiFS images were collected between 1997 and 2006 and processed to produce Colored Detrital Material images of the Mesoamerican waters. Monthly terrestrial runoff load and river discharge computed with a land-elevation model were used as input to a numerical model, which simulated the transport of buoyant matter from terrestrial runoff. Based on land cover for years 2003–2004, modeling results showed that the river discharge seasonality was correlated with the image averaged CDM, and the simulated plume reproduces the spatial patterns and temporal evolution of the observed CDM plume. River discharge peaked in August and CDM peaked from September to January. The buoyant matter concentration was high from October to January, and was at its lowest from March to April. Between October and December the plume was transported out of the Mesoamerican waters by a cyclonic gyre located north of Honduras. Part of the runoff from Honduras was transported towards Chinchorro Banks and the Yucatan Channel, part re-circulated into the Gulf of Honduras, and part taken toward the outside of the Mesoamerican Barrier Reef System. This study shows that all the reefs of the MBRS, including the most offshore atolls of the region, are under the influence of terrestrial runoff on a seasonal basis, with maximum effect during October to January, and minimum from March to April. Furthermore, what is seen as a giant plume in satellite images is in fact composed of runoffs of different ages.

Regional-scale scenario modeling for coral reefs: a decision support tool to inform management of a complex system

The worldwide decline of coral reefs threatens the livelihoods of coastal communities and puts at risk valuable ecosystem services provided by reefs. There is a pressing need for robust predictions of potential futures of coral reef and associated human systems under alternative management scenarios. Understanding and predicting the dynamics of coral reef systems at regional scales of tens to hundreds of kilometers is imperative, because reef systems are connected by physical and socioeconomic processes across regions and often across international boundaries. We present a spatially explicit regional-scale model of ecological dynamics for a general coral reef system. In designing our model as a tool for decision support, we gave precedence to portability and accessibility; the model can be parameterized for dissimilar coral reef systems in different parts of the world, and the model components and outputs are understandable for nonexperts. The model simulates local-scale dynamics, which are coupled across regions through larval connectivity between reefs. We validate our model using an instantiation for the Meso-American Reef system. The model realistically captures local and regional ecological dynamics and responds to external forcings in the form of harvesting, pollution, and physical damage (e.g., hurricanes, coral bleaching) to produce trajectories that largely fall within limits observed in the real system. Moreover, the model demonstrates behaviors that have relevance for management considerations. In particular, differences in larval supply between reef localities drive spatial variability in modeled reef community structure. Reef tracts for which recruitment is low are more vulnerable to natural disturbance and synergistic effects of anthropogenic stressors. Our approach provides a framework for projecting the likelihood of different reef futures at local to regional scales, with important applications for the management of complex coral reef systems.

Adaptive significance of the formation of multi-species fish spawning aggregations near submerged capes.

Background Many fishes are known to spawn at distinct geomorphological features such as submerged capes or “promontories,” and the widespread use of these sites for spawning must imply some evolutionary advantage. Spawning at these capes is thought to result in rapid offshore transport of eggs, thereby reducing predation levels and facilitating dispersal to areas of suitable habitat. Methodology/Principal Findings To test this “off-reef transport” hypothesis, we use a hydrodynamic model and explore the effects of topography on currents at submerged capes where spawning occurs and at similar capes where spawning does not occur, along the Mesoamerican Barrier Reef. All capes modeled in this study produced eddy-shedding regimes, but specific eddy attributes differed between spawning and non-spawning sites. Eddies at spawning sites were significantly stronger than those at non-spawning sites, and upwelling and fronts were the products of the eddy formation process. Frontal zones, present particularly at the edges of eddies near the shelf, may serve to retain larvae and nutrients. Spawning site eddies were also more predictable in terms of diameter and longevity. Passive particles released at spawning and control sites were dispersed from the release site at similar rates, but particles from spawning sites were more highly aggregated in their distributions than those from control sites, and remained closer to shore at all times. Conclusions/Significance Our findings contradict previous hypotheses that cape spawning leads to high egg dispersion due to offshore transport, and that they are attractive for spawning due to high, variable currents. Rather, we show that current regimes at spawning sites are more predictable, concentrate the eggs, and keep larvae closer to shore. These attributes would confer evolutionary advantages by maintaining relatively similar recruitment patterns year after year.

River-reef connectivity in the Meso-American Region

The accumulation and seasonal impact of riverine discharge on coral reefs of the Meso-American Region (MAR) were estimated using a numerical simulation of river runoff dispersion. River-reef connectivity, or source-sink dynamics of terrestrial runoff was further assessed from more than 400 watersheds of the region onto discrete coral reef areas. Using land use for 2003 and 2004 in the MAR, this work builds upon a Regional Ocean Modeling System simulation of the MAR validated by ocean color satellite data, and on the monthly river nutrient and sediment load and discharge provided by the World Resources Institute using the N-SPECT model. Analysis of the variability of simulated runoff transport to the reefs showed that reefs of the Mesoamerican Barrier Reef System (MBRS) were mostly impacted from June to September, following the peak time of river discharge. At that time, the coastal and oceanic circulations contribute quickly to expel the runoff from the MBRS. High runoff concentration waters leaving the eastern coast of Honduras during the months of October to December return to the southwestern MAR in March as they are entrained in a cyclonic gyre. Coral reefs of the MAR are thus impacted twice, first from the coastal side with runoff of local origin and later from the oceanic side with runoff from mixed origin. High probability of connectivity between rivers and remote reefs is established as this study revealed that river runoff from the north shore of Honduras has a wide-spread impact on most of the coral reefs of southern Belize, while watersheds on the Gulf of Honduras are mostly connected to coral reefs in the northern shore of Honduras. Although the level of remote influence (or runoff concentration reaching the reef) is lower than the local, the cumulative effect of numerous remote river-reef connections remains significant.

Evidence of Mediterranean Water dipole collision in the Gulf of Cadiz

A collision of Mediterranean Water dipoles in the Gulf of Cadiz is studied here, using data from the MedTop and Semane experiments. First, a Mediterranean Water eddy (meddy) was surveyed hydrologically in November 2000 southwest of Cape Saint Vincent. Then, this meddy drifted northeastward from this position, accompanied by a cyclone (detected only via altimetry), thus forming a first dipole. In February 2001, a dipole of Mediterranean Water was measured hydrologically just after its formation near Portimao Canyon. This second dipole drifted southwestward. The western and eastern meddies had hydrological radii of about 22 and 25 km, respectively, with corresponding temperature and salinity maxima of (13.45°C, 36.78) and (11.40°C, 36.40). Rafos float trajectories and satellite altimetry indicate that these two dipoles collided early April 2001, south of Cape Saint Vincent, near 35°30′N, 10°15′W. More precisely, the eastern meddy wrapped around the western one. This merger resulted in an anticyclone (a meddy) which drifted southeastward, coupled with the eastern cyclone. Hydrological sections across this final third resulting dipole, performed in July 2001 in the southern Gulf of Cadiz, confirm this interaction: the thermohaline characteristics of the final meddy can be tracked back to the original structures. The subsequent evolution of this dipole was analyzed with Rafos float trajectories. A numerical simulation of the interaction between the two earlier dipoles is also presented. We suggest that these dipole collisions at the Mediterranean Water level may represent a mechanism of generation of the larger meddies that finally leave the Gulf of Cadiz.

Vortex Dipole Formation by Baroclinic Instability of Boundary Currents

In situ data of the Mediterranean Water undercurrents and eddies south of Portugal indicate that the undercurrents have a tubelike structure in potential vorticity and that dipole formation can occur when the lower undercurrent extends seaward below an offshore upper countercurrent. A two-layer quasigeostrophic model is used to determine the dynamical conditions under which dipole formation is possible. With piecewise-constant potential vorticity, the flow exhibits two linear modes of instability comparable to those found in the Phillips model with topography. Weakly nonlinear analysis and fully nonlinear simulations of the flow evolution agree on the regimes of either finite-amplitude perturbation saturation, corresponding to filamentation, or amplification, corresponding to vortex or dipole formation. This latter regime is more specifically studied: vortex dipole formation and ejection from the coast is obtained for long waves, with opposite-signed but similar amplitude layer potential vorticities. A simple point vortex model reproduces this phenomenon under the same conditions. It is then shown that dipole formation occurs for minimal wave dispersion, and hence for weak horizontal velocity shears. As observed at sea, dipoles are formed when the lower potential vorticity core extends seaward below a countercurrent.