Patrick W. Robinson
O.S. 101
Winter 2003
Thursday 10am


Theoretical and Observed Mechanisms for Flow-Related

Larval Retention on Oceanic Islands


Some of the largest and most dominant processes in the ocean have only been found to exist recently, as evident by the youth of physical oceanography as a research field.  Oceanographic work has focused mainly on the near-shore coastal processes of continental landmasses as well as ocean basin-scale dynamics.  Seasonal upwelling of nutrient-rich water, for example, has been studied, along with its associated biological implications, in great detail.  Understanding coastal current systems and the effects of topography has also lead to important discoveries.  Many of these advances, both intellectually and technologically, have allowed for the study of oceanography to extend off of continental landmasses to smaller, and often more complex, systems.  The relatively small-scale oceanographic processes occurring around oceanic islands have recently developed an interest among researchers.  Literally hundreds of scientific papers have been published regarding the physics, oceanography, and biology related to “island mass effects” and “island wakes,” most enumerating the importance of these phenomena to the retention of many benthic and pelagic species in geographically and genetically isolated regions.  This pattern is of marked difference when compared to the advantages of dispersal with little retention on mainland coasts.  Physicists have conducted laboratory experiments to understand the fluid dynamics that make these complex effects possible.  Osborne Reynolds pioneered much of the mathematics behind the study of fluids and described several fundamental relationships that have direct ocean-island applications:  The Rossby number, Ekman number, and Reynolds number are all valuable tools for predicting flow regimes around islands, given observation of several measurable physical variables.  Relatively simple field studies followed this lab-based theoretical work.  Boden (1952) and Boden and Kampa (1953) studied the flow around Bermuda and the larval retention implications of seasonal patterns.  Chiswell and Roemmich (1998) took their research a step further by integrating advancing technology to yield more powerful quantitative results.

Miguel Caldeira (2002) recently completed his doctoral thesis containing a review of the history of island wake studies.  He states that the initial interest of oceanic island-effects was from biologists.  They were interested in the so-called “island wake” effect in which an increased nutrient concentration was found as one approached an island.  Although in some cases this phenomenon could simply be due to runoff of nutrient-rich freshwater from the island, evidence from some islands pointed to an oceanographic explanation.  The upwelling of nutrient-rich water was thought to be the cause.  Studies, such as finding the rate of carbon fixation, were conducted to determine regions around islands that were indeed nutrient enriched as well as high in primary productivity.  Differences were found when comparing the windward versus the lee-side of islands.  Early in these studies, it was assumed that an island in the midst of an oceanic current would produce some degree of flow instability and, therefore, a small amount of mixing; however, the degree of difference in primary productivity precluded this from being the sole origin of the observed pattern.  It was obvious that a much more complicated system was at play.

Initially, three different hypotheses were posed to explain the elevated plankton concentration, or biomass, on the lee side of oceanic islands.  First, the transport of these organisms from regions closer to the island via advection was a possibility.  That is, currents could easily move planktonic organisms to the lee of an island, just as wind would disperse seeds from a tree.  Secondly, the pattern could be due to the presence of hydrodynamic processes known as eddies and turbulence zones.  Lastly, a buildup of nutrients via an elongation of residence time may allow an otherwise small population of plankton to grow and reproduce into higher densities.  This would require a combination of upwelling, a nutrient source, and recycling current patterns, an elongation of residence time, to be successful.

Patterns of aggregation in phytoplankton and zooplankton were studied in great detail at a variety of locations.  Local patchiness in plankton concentrations were found to be correlated to eddies and convergence zones.  Eddies are water masses moving in a circular motion, thus dissipating the energy associated with a convergence zone with additional friction.  These can form on either side of the turbulence zone off the lee of an island.  Also, eddies are a particularly attractive solution because they will tend to draw plankton toward the center and form higher concentrations.  Eddies, upwelling, convergence zones, and other oceanographic events are all thought to make up the process of the island mass effect.  The various components and the island mass effect itself all have different periods of temporal variability (i.e. longevity).  This is summarized with a three-dimensional surface plot in figure 1.  The plot shows the approximate size, duration, and magnitude of effect (z-axis) of these different components.  Of particular interest are the upwelling zones, eddies, and island-mass effects.  Upwelling zones and eddies are included in the annual cycle region and have a relatively large-scale intense effect, as evident by the high z-axis values.  Island mass effects were found to be much more unpredictable with respect to all three variables.  That is, the effects can range in time from weeks to centuries, in space from one to one hundred kilometers, and can even vary significantly with respect to intensity.

Fig 1.  Three-dimensional surface indicating the special scale, length of time, and intensity variation in several island-related oceanographic events.  Important events include:  C-upwelling, D-eddies, and E-island mass effects (Haury et al., 1978).

In addition to being described qualitatively, the effects that an island has on flow patterns can be described mathematically.  Reynolds (1883) was a leader in the field of fluid dynamics in his time, despite the absence of current technology regarded as necessity.  Through his work in the laboratory with very controlled conditions, he was able to discover several important physical relationships that, with the later help of oceanographers, were applied to much larger systems.  The Rossby’s number, Ekman number, and Reynold’s number are all tools that, when used together, can demonstrate the hydrodynamic effects of island wakes.  The Rossby number is defined as:
Ro = U / ( f * 2L )         Where: U = Velocity of flow

                                                            F = The coriolis  parameter

                                                            L = The horizontal length of the land mass (radius)

This results in a dimensionless number that describes the ability for the Earth to induce vorticity on a water mass.  That is, a larger Rossby number indicates that a rotation event is more likely.

            The Ekman number is defined as:

            Ek = v / [ f * (2L)^2 ]   Where: v = kinematic viscocity

                                                            F = The coriolis parameter

                                                            L = The horizontal length of the land mass (radius)

This describes the balance between the size of the ekman layer and the frictional forces that could overpower it.  A large ekman number indicates that viscosity due to friction will cause dissipation of a rotation event.

            Lastly, the Reynold’s number is defined as:

                        Re = Ro / Ek = {U / ( f * L )}* [ f * (L^2) ] / v = ( U * L ) / v

This value is also described as the inertial force divided by the viscous force.  A Reynolds number of less than one indicates no flow separation.  This would occur, by definition, under three circumstances:  a low velocity of flow, a small horizontal length of landmass, or a large viscosity value.  A Reynolds number greater than one indicates the formation of two stable eddies.  Figure 2a shows a simulation of this stable pattern on the lee of an island.  As the Reynolds number increases, instabilities in the flow, known as “meanders” begin to appear (figure 2b).  Finally, as the Reynolds number increases further, large rotating water masses known as Van Karman vortices are formed (figure 2c).  Contrary to the “no flow separation” model, this would occur under three circumstances:  a high velocity of flow, a large horizontal length, or a small viscosity value.  These relationships are compared in figure 3 with respect to the Rossby number and the Ekman number.


Figure 2:  Flow patterns around simulated islands for Re ~ 1, Re ~ 20, and Re >> 20, respectively (Caldeira 2002).

Figure 3:  Theoretical island lee-effects as determined by the Rossby number and Ekman number (Caldeira 2002).

The work of Reynolds is pivotal to the basic understanding of flow dynamics around stationary objects, but his work lacks real oceanographic examples.  His experiments are limited to very controlled conditions and cannot, therefore, be used alone to describe the complex systems found in the natural world.  Field studies are, therefore, a necessity to understand all of the possible implications of Reynolds’ work as well as other mechanisms that may work to conserve larvae in genetically and geographically closed island populations.

Boden (1952) was one of the first studies to look at the retention of insular, i.e. island dwelling, plankton via water flow mechanisms.  His work brought him to the summer season in Bermuda:  a relatively isolated deep-water oceanic island.  The motivation for this work began with the observation that a seemingly permanent plankton community exists near the island, despite an intermittent current that could potentially advect these organisms.  Also, it was observed that many of the benthic organisms spawned only during the summer months, possibly indicating a preferential time for larval retention.  The first step in discovering this retention mechanism involved the collection of oceanographic data.  Vertical profiles of the water were taken during cruises both on and off of the small shelf of the island.  Measures of temperature, salinity, and density were made.  Throughout the summer, vertical plankton hauls were also completed.  Plankton, by definition, are at the mercy of the currents, thus providing an excellent proxy for information about convergence and divergence of water masses.  Using this information, graphs of isopycnals and isotherms were created.  The results of this work indicated that the oceanic currents flowing around the island were weak and not influencing the local flow of water near the island.  Essentially, a convection cell was set up over the reef, trapping water in this region.  Associated higher temperature and salinity were found within the convection region.  Additionally, the plankton tows showed significantly higher concentrations above the shelf when compared to neighboring oceanic water and that, as expected, a convergence zone was set up at the interface.

Later, Boden and Kampa (1953) conducted a very similar study during the winter months to learn about any seasonal fluctuations in the previously observed patterns.  The methods were nearly identical with the exception of the removal of the plankton hauls, due to adverse weather conditions.  When similar graphs were made with the new winter data, a surprising result was obtained; a complete reversal of the local current structure had taken place.  Instead of water being trapped in a convection cell over the shelf, a “cyclonic, convergent movement about a sinking center” was found to exist.  Essentially, planktonic retention, with the exception of vertically migrating organisms, was non-existent.  This confirmed the reasoning behind the summer spawning of the benthic organisms.

The work of Boden and Boden and Kampa, however, was not without weakness.  The methods lacked direct measurement via current meters and used extrapolations based on relatively few forms of data.  Despite these limitations, the work does show that effects unrelated to current-driven lee-effects are also important in the retention of larvae.  More recent studies have utilized technological advances to increase the power of the results.

Chiswell and Roemmich (1998) studied the currents off of the north island of New Zealand.  This system is much larger than that of Bermuda, but the importance of larval retention remains.  The system consists of two large, relatively stable eddies off of the east coast.  The motivation for this work was the retention of Rock Lobster larvae.  These larvae have a planktonic stage that lasts up to a year, not weeks or months like many species.  This places added importance on not only the presence, but the stability of larval retention mechanisms.  This study focused on the East Cape Eddie and the Wairarapa Eddie.  Four moorings were deployed for at least one year to measure flow rates.  Additionally, oceanographic cruises were done to collect CTD, conductivity-temperature-depth, data for a vertical profile.  Lastly, TOPEX / Poseidon satellite altimetry data was used.  The data was compiled into a model that attempted to predict the location of simulated “drifters” over time.  After running the simulation for a sufficient number of drifters, a probability plot was created to show the likelihood that a drifter released between the two eddies would remain in the system for an entire year (figure 4).  Amazingly, the model predicted a retention likelihood of up to 30%, indicating that the eddies are indeed a potential source of larval retention.  The abundance of data used to create the model added to the power of the predictions and was a definite strength of the paper.  However, predictions without validation, such as the release and subsequent tracking of real drifters, prevents other hypotheses from being excluded.

Larval retention mechanisms are of clear importance in genetically isolated closed populations, specifically those located on oceanic islands.  The work of Reynolds shows that retention via lee effects such as eddies is at least theoretically possible.  Boden and Boden and Kampa demonstrated that current-driven lee effects are not the only mechanism that can lead to larval retention.  Also, they found an oceanographic-biological relationship due to the ephemeral nature of the retention processes on Bermuda.  Lastly, Chiswell and Roemich show that permanent eddies may be an important oceanographic event with respect to long-lived and slow-growing organisms.  Many studies have aimed to study the various lee effects and larval retention mechanisms with limited success and often questionable results.  Further study of island lee effects are needed to assess their importance in the maintenance of isolated biological communities.

Figure 4:  “Contours of probability of passive drifters with retention times of 365 days or longer (Chiswell and Roemmich 1998).


Boden, B. (1952). Natural conservation of insular plankton. Nature, 169, 697-699.

Boden, B., & Kampa, E. (1953). Winter cascading from an oceanic island and its biological implications. Nature, 171, 426-427.

Caldeira, M. (2002).  Multidisciplinary Studies of the Island Mass Effect Phenomena Around Santa Catalina Island, Southern California Bight.  Ph.D. Dissertation, UCLA.

Chiswell, S. M.  and D. Roemmich (1998).  The East Cape Current and two eddies:  a mechanism for larval retention?  New Zealand Journal of Marine and Freshwater Research.  32:  385-397.

Haury, L. R., McGowan, J. A., & Wiebe, P. H. (1978). Patterns and processes in the time-space scales of plankton distributions. In: Spatial patterns in plankton communities.  Plenum Press, pp. 277-327.

Reynolds, O. (1883).  An experimental investigation of the circumstances which determine whether the motion of water shall be direct or sinuous, and of the law of resistance in parallel channels.  Philosophical Transaction so of the Royal Society of London.  174:  935-982.