Larvae Dispersion in Coral Reefs and Mangroves

DISPERSION IN CORAL REEFS AND MANGROVES

The complex but predictable patterns of dispersion of prawn, coral and fish larvae in mangroves and coral reefs are responsible for maintaining rich fisheries.

Eric Wolanski and Joe Sarsenski

Eric Wolanski received his Ph.D. in environmental engineering in 1973 from Johns Hopkins University in environmental engineering. For the last 18 years at the Australian Institute of Marine Science (AIMS) where he is a senior principal research scientist, he has been studying tropical coastal oceanography and its biological implication for mangroves and coral reefs. Joe Sarsenski is Program Manager of the IBM Environmental Research Program. His research interests include the use of information technology in simulating sustainable development scenarios. He received his Ph.D. in environmental engineering from the University of Connecticut in 1978. Address for Wolanski: AIMS, PMB No. 3, Townsville M.C., Qld. 4810, Australia. Internet: e.wolanski@aims.gov.au

Mangroves and coral reefs (Figure 1) are topographically very complex. For instance, Australia's Great Barrier Reef is a dense matrix of more than 2000 individual coral reefs of various sizes and shapes separated by channels and shoals, all spread along the 2600 km-long continental shelf of northeast Australia. Mangroves are heavily vegetated forests in tidal flood plains along estuaries. They are prevalent in South East Asia and along the coast of the Great Barrier Reef. Both mangroves and coral reefs interfere with the prevailing water currents. They not only block and steer the prevailing currents, but they also generate complex flows including eddies, jets and stagnation zones. As a result, understanding the links between physics and biology in mangroves and coral reefs has often be seen as too daunting a task, a problem similar to understanding the biology in individual eddies of a turbulent flow.

Understanding these links in blue water fisheries oceanography necessitates multi-disciplinary research. More often than not, however, oceanographers, chemists and biologists speak three different languages. This makes multi-disciplinary research the exception rather than the rule and little scientific cross-fertilisation results. This problem is even more difficult in the topographic maze of mangroves and coral reefs. Here the water circulation is at first glance too complex for multi-disciplinary research to be fruitful. This problem also often prevents the marine-resources managers, who also speak their own language, from extracting what is useful to them from the discipline-specific literature.

Recent advances in computer technology are helping to resolve this problem. These advances have enabled the calculation of these complex flows and their chemical and biological implications, facilitating multi-disciplinary marine research. The results of this research have shown that the interaction between the complex bathymetry and the currents controls the processes of dispersion in mangroves and coral reefs. This research also reveals that the recruitment (or process by which larvae find a permanent home) of prawn, fish and coral larvae (Figure 2), depends on the interaction between water currents and the bathymetry. Finally, this technology provides science-based tools for use by marine resources managers.

Currents through the Great Barrier Reef

The currents through the Great Barrier Reef are forced mainly by the tides and the wind while oceanic forcing generates prevailing longshore currents. These prevailing currents in turn are topographically steered and in some cases blocked by the reef matrix. The flows are thus varying spatially and temporally, and resemble turbulent flows with eddies, jets, convergence and stagnation zones. Details of these flows can only be understood through computer models that need to be carefully calibrated against field data. Computer visualisation is needed to explore the complexity of the data from the field studies and the model output.

One such field data set, the most extensive data set world-wide for flows around coral reefs, is the one we collected using 26 current meters moored in the lee of a small island, Rattray Island. These meters measured the dynamics of island "wakes"- the eddies shed by an island interfering with a swift tidal current (Figure 3). Island eddies have different physics than eddies behind blunt bodies in tank experiments in fluid mechanics laboratories. This is because of the difference in the ratio between the water depth and the width of the obstacle in the field and in the laboratory. In laboratory experiments this ratio is very much larger than 1.0 while for flows around a coral reef this ratio is much smaller than 1.0. The implication is that in the latter, bottom friction is important while in the former it is not. In addition, water tanks cannot be scaled up to mimic the intense turbulence in the separation layers behind a reef. Separation layers are the thin sheet of water separating the waters in the free stream from eddy waters. Field studies show that the separation layers are too thin to be explicitly calculated in numerical models. Their importance resides in the intense small-scale turbulence they manifest. When this turbulence is explicitly parameterised in numerical models, the models are capable of reproducing correctly the strong three-dimensional flows in the lee of a coral reef. This includes a upwelling in the bulk of the eddy and a downwelling in a thin near-circular layer along the edges of the eddy. These vertical velocities are typically 20 m per hour; values enormous compared to those found in a classical coastal upwelling, typically 5 to 20 m per day.

Water circulation through the 2000 obstacles in the Great Barrier Reef is thus akin to extremely complex, three-dimensional eddies or jets, each with the horizontal dimensions of the individual reefs, typically a few km in diameter but 30-60 m deep. Similar complex tidal flows prevail through mangrove forests, where the scale of the eddies is similar to the vegetation obstructing the flow, typically centimetres to tens of centimetres.

Biological implications

Water currents are important because they determine the destination of water-born particles. Particles of interest include natural ones such as coral eggs and the larvae of prawn and fish, as well as pollutants such as mud from runoff and dredging, pesticides and nutrients from farm runoff, and sewage from treatment plants.

The upwelling (Figure 3) in topographically steered currents around coral reefs results in removing the finest mud particles from the sea floor of the island wake, thereby providing a different substrate for benthic organisms. The downwelling results in aggregating floating matter along the outer edges of the eddy. This aggregation can readily be observed following mass spawning of the corals, an event that occurs annually on the Great Barrier Reef. After release in the water column, the coral eggs float to the surface and are initially distributed widely on the surface all over the submerged coral reef (Figure 4). A day later however the eggs and larvae are found aggregated in long lines that can be several hundreds of meters long but only a few meters wide (Figure 5). The distribution of coral eggs is thus very patchy within a few hours after mass spawning. Aggregation and patchiness prevail even though there are generally negligible vertical gradients of salinity and temperature gradients on the Great Barrier Reef.

This aggregation is contrary to what oceanic dispersion models predict, that the size of the patch should continuously increase as a result of diffusion. The three-dimensional flows should be calculated to account for aggregation which is akin to negative diffusion. Since the three-dimensional flows are small-scale processes, either a very small mesh size must be used in the models, at great cost, or the processes must be explicitly parameterised in dispersion models.

The next day these aggregations are not visible, thus reliable measurements of coral larvae concentration are difficult. Over- or under-estimating concentration is possible using current field sampling techniques which rely on towing nets underwater at a number of sites. Accurate estimates thus require a large number of sites and repeated net tows. The only data set world-wide which address this issue was collected at Bowden Reef. The Australian Institute of Marine Science (AIMS) enjoys access to this data base which shows that the total numbers of coral larvae decrease daily and that the bulk of the coral larvae are flushed away from their natal reefs in a few days.

Numerical models of currents and advection-dispersion of coral eggs can be calibrated against such data. Such models predict that a plume of coral larvae forms in the lee of the coral reef (Figure 4). The plume shows great spatial and temporal patchiness. Patterns are visible in the patchiness, but these change constantly with time, making field calibration very difficult. Computer animation facilitates this step.

For a coral reef thus the recruitment of larvae depends on oceanic dispersion of larvae emanating from reefs further upstream. Which is an upstream reef and which is downstream depends on the prevailing currents during the dispersal period (lasting two to three weeks). Because of varying wind and oceanic forcing, these prevailing currents can vary from year to year during the spawning season. Computer modeling, calibrated on historical data, can be used to determine this inter-annual variability. This enables a calculation of source reefs and sink reefs for individual years.

Fish larvae also drift with the currents at early stages of their life, making it possible to use them as a tracer to measure their recruitment on reefs. Once again AIMS enjoys access to the most extensive field data base on the concentration of pre-recruitment fish larvae around two coral reefs, Helix and Bowden reefs (Figure 6). The data were obtained using fish traps equipped with a light attractor (Doherty, 1987). They display significant complex temporal and spatial variability which requires computer animation for interpretation. Based on an analysis of the computer animation, the fish larvae are not initially resident around a particular coral reef. Arriving from upstream on water currents, the smallest fish larvae are concentrated in hydrodynamic shadow zones on the lee side of coral reefs while the largest fish larvae aggregate in front.

Classical oceanographic fisheries models assume the larvae are carried by the water currents only. These models predict unrealistic larvae distribution patterns with no aggregation and patchiness when applied to the Bowden Reef site (Figure 6). Direct SCUBA observations by Dr. J. Leis at the Australian Museum reveal the fish larvae actually swim horizontally towards a reef. It is possible to quantitatively reproduce observed aggregations (Figure 6) if swimming is incorporated within a 2000m radius of the reef. The importance of this behavioural pattern has never been suspected.

Most larvae will recruit on coral reefs and not be lost at sea when the active swimming hypothesis is incorporated in the AIMS model for the entire Great Barrier Reef. In contrast, the classical 'no swimming' hypothesis results in most larvae dying at sea. Thus for the Great Barrier Reef the evidence suggests a direct link between adult fish stock and larval recruitment. This makes the Great Barrier Reef sensitive to over-fishing.

Mangroves fisheries

Mangroves are vegetated, inter-tidal wetlands, common in tropical and sub-tropical coastal environments. They play a vital role as producers of nutrients, forest resources and animal species of economic value. Less well known is the link between the prawn fisheries and the health of the mangrove ecosystem. Mangrove-fringed, shallow coastal waters support rich prawn fisheries. The spawning grounds are located in coastal waters, but the nursery grounds are in the mangroves. On their way to the mangroves, the drifting planktonic prawn larvae traverse shallow coastal waters. Field data collected by Dr. V.C. Chong show great spatial and temporal patchiness of the larvae during this phase. This phenomenon is best understood by exploring the data through computer visualisation, which reveals seasonal and spatial gradients as well as inter-species variability.

Hydrodynamic models of mangrove-fringed coastal waters have been formulated and calibrated against field data. Model runs incorporating two-weeks old, pre-recruitment larvae at concentration and sites determined from the field data reveal that mangroves sustain prawn fisheries by providing a hydrodynamic trap (Figure 7). Recruitment is enhanced at spring tides and the trapping at neap tides.

Conclusion: Information technology for resource management

Another important use of computer technology for mangroves and coral reefs is to provide a science-based tool for managers. The problems are pressing because these ecosystems are rapidly disappearing world-wide. The Great Barrier Reef may be more sensitive to over-fishing than previously was believed. This has important management implications as it now appears possible to scientifically determine the key source reefs. These reefs need higher protection levels to conserve the adult fish stock.

For the case of mangroves, computer models and visualisation of the prawn fisheries recruitment processes offer marine resources managers a science-based tools to assess strategies for managing mangroves and their fisheries in view of pressure for converting mangrove land to other uses.

Computer technology is essential for advancing multi-disciplinary research in marine science. It offers a practical way for physicists, chemists and biologists to merge their data sets from point measurements of currents, concentration of larvae of shrimp, fish and corals, suspended sediment and chemical matters (eg hydrocarbons). Computer-visualisation of the data is necessary because the data are patchy in time and in space and practically unusable using standard statistical techniques. The visualisation enables the scientist to understand which are the dominant physical, chemical and biological processes that must be taken into account in models of coral reefs and mangroves. These processes can then be incorporated in mathematical models of processes such as coral larvae dispersion, fisheries recruitment, siltation and pollution. The computer visualisation of the model output provides a science-based tool for marine resources managers to estimate the environmental impact of strategic decisions.

Acknowledgments

The authors thank the Australian Institute of Marine Science, the IBM International Foundation, Japan's PHRI and KEPCO-KEEC, the CRC Reef Research, and Drs. Brian King, Peter Doherty, Jamie Oliver, Ving Chin Chong and Ian Gardner.

Bibliography

Doherty, P. 1987. Light traps: selective but useful devices for quantifying the distribution and abundances of larval fishes. Bulletin of Marine Science, 41: 423-431.

Chong, V.C., Sasekumar, A. and Wolanski, E. 1996. Tidal hydrodynamics and larval prawn advection in the Klang Strait, Malaysia. Mangroves and Salt Marshes (in press).

Galloway, D., Wolanski, E. and King, B. 1996. Modelling eddy formation in coastal waters: a comparison between model capabilities. In M. Spaulding and R. Cheng (Editors):"Estuarine and coastal modeling", American Society of Civil Engineers, New York, pp. 13-25.

Wolanski, E. 1994. Physical Oceanographic Processes of the Great Barrier Reef. CRC Press, Boca Raton, Florida, 194 pp.

Wolanski, E., Asaeda, T. , Tanaka, A. and Deleersnijder, E.1996. Three-dimensional island wakes in the field, laboratory and numerical models. Continental Shelf Research, 16, 1437-1452.

Figures

Figure 1. Abundant and diverse marine resources prevail on the Great Barrier Reef. and the ecologically inter-dependent mangroves.

Figure 2. One common characteristic of coral eggs (left), fish larvae (middle) and prawn larvae (right) is all are displaced and dispersed during their early life cycles.

Figure 3. The top right picture shows bathymetry of Rattray Island, a 1500 m wide island located in 20-30 m water depth in the Great Barrier Reef. The picture on the top left shows a snapshot of the observed distribution of currents in the lee of the island, an eddy is apparent. The thick arrows are measured currents and the thin arrows are from linear interpolation. The area is colour-coded for speed from blue to red to indicate slow to fast currents. The picture on the bottom right shows the predicted current speeds, the arrows show the horizontal motions and the colours the vertical motions. There is an upwelling in the center of the eddy and a downwelling all along the outer edges of the eddy.

Figure 4. Top left view is a snapshot of predicted currents in the central region of the Great Barrier Reef The white lines represent coral reefs. The land is shaded in green. The view covers about 200 km north-south. The coastline (top) is rugged with many headlands. The reefs (white) are located mainly on the mid- to outer-shelf. The shelf width is about 100 km. Note the complex currents through the matrix of coral reefs with eddies, jets and zones of stagnation. The white frame is centred around Bowden Reef. The top right picture is a colour-shaded three-dimensional view of the bathymetry of Bowden Reef, a 5000 m long coral reef in the central region of the Great Barrier Reef in 50 m depth, , a key study site over 10 years for multi-disciplinary studies by the Australian Institute of Marine Science in oceanography, coral spawning and fish larvae recruitment. The pins indicate measurement sites. The bottom left picture shows as colour-coded vertical the number of coral larvae found the day after mass spawning in December 1987, at the surface (left) and at 6 m depth (right). The bottom bars indicate the mean number of coral eggs and the top bars the standard deviation. The bottom right plot is the predicted concentration of coral larvae three days after mass spawning. The data show export in a few days of the bulk of the larvae away from the natal reef. The plume shows much patchiness attributed to complex topographical currents.

Figure 5. Aerial photograph of lines of aggregation of coral eggs the day after mass spawning. These aggregation lines are several hundreds of meters long and only a few meters wide. The research ship shown near the junction of the two lines is 33 m long.

Figure 6. The top view shows a 3-D colour-coded view of the bathymetry of Bowden Reef and the white pins show fish larvae collection sites. This image also shows a snapshot of observed pre-settlement, reef fish larvae distribution three days after arrival of the cloud of larvae from reefs upstream, showing that the reef fish were aggregated downstream. The middle view shows the predicted distribution of reef fish, making the classical assumption that fish larvae are moved passively the water currents, somewhat like a dye cloud. The model is clearly unable to reproduce the observations. However if the model also includes the horizontal swimming behaviour of the larvae, the model predicts (bottom picture) aggregations patterns which resemble the observations. The model also predicts that contrary to the case of passive dispersion, most of the larvae are trapped by the reef.

Figure 7: Three-dimensional visualisation of the colour-coded bathymetry of the mangrove-fringed Klang Strait waters on the west coast of Malaysia. The area is about 60 km long. The mangroves are shaded green. A 10-15 m deep channel separates the mangroves from Angsa Bank, a large intertidal shoal further offshore. In the left view, a patch of pre-settlement, penaeid prawn larvae was released in the model at sites where field data reveal they are abundant. The middle view shows the distribution of these larvae after 135, and the right view after 300 hours. About half of the larvae remain trapped in that period in mangrove-fringed waters where they can mature, while the rest of the larvae are essentially lost at sea.