303 17 Lessons Learned in the Construction and Operation of Coral Reef Microcosms and Mesocosms Walter H. Adey CONTENTS 17.1 Introduction 303 17.2 Caribbean Model (Coral Reef Microcosm) 305 17.3 The Operational Imperative 311 17.4 Implications of Microcosm Modeling for Coral Reef Restoration 312 17.5 Summary 312 References 313 17.1 INTRODUCTION Especially since the second World War, considerable effort has been invested in attempts to maintain and display marine organisms in public aquaria and marine laboratories. In addition, in the latter part of the 20th century, numerous hobbyists joined the quest for marine display, with sometimes rather elaborate systems in their homes. The most successful of the professional organism main- tenance systems involve the use of sea water, pumped through sand beds or towers, from adjacent, high-quality marine environments (flo-through systems). In some cases, and particularly in the hobby environment, closed aquaculture methods are also employed, mostly using bacterial filtration and/or air-foamed towers (foam fractionation) to remove particulate matter. A large percentage of these aquaculture and flo-through efforts involve coral reef organisms, although usually as individuals of a relatively few species. Only rarely does current practice involve any semblance of a coral reef ecosystem, either in terms of metabolism or community structure. An introduction to this literature can be found in both research and aquarium science publications (e.g., references 1 through 5) and in numerous books of varying quality that line the bookshelves of high-quality aquarium stores. These same techniques could be employed in a shoreside laboratory to hold less sensitive organisms for timed introduction to a coral reef restoration effort, though loading and organism stress issues limit these techniques. Biodiversity in such systems, even when water quality can be maintained at high levels, is typically low. In large measure, this is due to the filtration of reproductive stages by the water quality–maintaining devices, and to the further destruction of those stages by pumping devices. However, some organisms can be reproduced vegetatively or through budding and fragmentation. More appropriate to the subject of the present volume, during the last two decades of the 20th century, a number of research and development endeavors were undertaken to model coral reef 2073_C017.fm Page 303 Friday, April 7, 2006 5:17 PM © 2006 by Taylor & Francis Group, LLC 304 Coral Reef Restoration Handbook ecosystems in a semiclosed state. These projects were carried out at a wide variety of scales, ranging from roughly a cubic meter or so up to 3500 m 3 . Numerous publications, beginning with reference 6, describe these efforts, and many are cited in this chapter. Two works 7,8 review the coral reef modeling work of the 1980s and the early ’90s and compare and contrast the various systems. The author is aware of nine microcosms and mesocosms specifically constructed to be physi- cal/ecophysiological models of coral reef ecosystems. These systems were built by four separate organizations, and provided (to date) roughly 75 years of summed operation. 7 Mostly, these units were directed to educational display, with research as a secondary feature. A few were constructed primarily to test engineering features of large ecosystem models. Only one of these models, a 1680-l Caribbean coral reef (though with a few Indo-Pacific species) operated long term and was also subjected to considerable biodiversity and ecophysiological analysis. Although a less comprehensive analysis of some of the earlier systems 7 suggested that this system was by no means unique, it is the 1680-l Caribbean system, which operated throughout the 1990s, that provides the significant data on which the conclusions of this chapter are based. The status of microcosms in restoration ecology prior to the mid-1980s was reviewed in 1987. 9 In that work, the philosophical basis for physical, living systems modeling was extensively discussed and compared to similar engineering endeavors (such as ship model towing basins and hydraulic models of bays and harbors). There is little point in repeating that discussion here, as virtually all of it is just as applicable today as it was 18 years ago. However, about the time that paper was published, research was initiated on the 1680-l Caribbean model that is the primary focus of this chapter. The essential missing ingredient of that report, long-term hard data, is now available. Hereafter, in this chapter, the unit that provided this data is referred to as the Caribbean Model. The Caribbean Model was specifically modeled after the well-developed, shallow-water, bank barrier reef off the southeast side of the island of St. Croix, U.S. Virgin Islands (for an in-depth discussion of that reef see reference 10). While teaching coral reef ecology and phycology at the West Indies Laboratory of Fairleigh Dickinson University through much of the 1970s, the author had the opportunity to direct and participate in numerous class projects and student papers that related to that very well-developed reef system. At that time, there was little direct tourist or industrial effect on the southeast reef, and fishing was minimal and artisanal. Also, there had not been significant hurricane impact on those reefs for roughly the prior 50 years. Late in the 1970s, based on about 6 years of continuous research on this reef complex, a year-long, upstream/down- stream analysis of ecosystem metabolism was carried out. 11 That study correlated to reef structure, and geological development and analyses were carried out at several stations stretched down the 25-km length of the reef. The extensive coral reef complex after which the Caribbean Model was patterned faces south- east, largely open to the Caribbean trade winds and seas. Yet, it is somewhat protected from the destructive winter “rollers” (large swells out of the North Atlantic) and had developed a roughly 10- to 20-m thick, carbonate reef structure over a period of about 5000 years. Based on extensive coring and carbon 14 (C14) dating, portions of the reef were shown to have upward growth rates greater than 10 m/1000 yr. Generally, growth rates slowed down as reef flat levels were attained and flat widths further developed. Considering the entire pan-tropic coral reef environment, this reef is certainly a strong “performer” (at least it was in the 1970s). Nevertheless, many similarly well-developed coral reefs exist in the Caribbean Sea and West Indies. 12 At the time when the metabolism and organism cover studies were carried out on the southeast St. Croix reef, all reef/time stages from submature, with very rapid growth rates, to fully mature, with broad reef flats, were present. All major stages were included in the studies that preceded the development of the coral reef models cited in the references for this chapter. In this chapter, “St. Croix reefs” specifically refers to the bank barrier reef on the southeast side of St. Croix during the 1970s. For specific details on the wild-type reefs that relate to the microcosm discussion, consult references 10 and 11. Although physiologically, especially with regard to basic physical–chemical parameters, particularly light and metabolism, the microcosm system was matched to the 5-m-deep 2073_C017.fm Page 304 Friday, April 7, 2006 5:17 PM © 2006 by Taylor & Francis Group, LLC Lessons Learned in the Construction and Operation of Coral Reef Microcosms and Mesocosms 305 fore reef location on a midsection of the St. Croix south reef, community structure was more generally related to a mean of the entire reef. 13 17.2 CARIBBEAN MODEL (CORAL REEF MICROCOSM) In this chapter, to provide a ready reference, I briefly describe the Caribbean Model that is our primary source of experimental data. However, for a detailed treatment of background and exper- imental protocol, readers are referred to references 8, 13, and 14. A diagram of the physical layout of the Caribbean Model and its basic “closure state” is shown in Figure 17.1. As noted above, the 400-l or metabolic unit is ecophysiologically the microcosm match to a specific point on the St. Croix reef. Water movement, current, and wave action in this primary system are driven by slow-moving, low-shear bellows pumps. These prevent significant damage to swimming and floating reproductive stages. These pumps also supply water to the Algal Turf Scrubber (ATS), which is a nonfiltration, managed algal community that allows the algal turf species resident in the model reef to colonize the ATS, where photosynthesis and growth conditions are optimal. The ATS system is lighted during the dark cycle of the coral reef and thus provides the ability to control the primary chemistry of the water column on a diurnal basis. The 1280-l “refugium,” with human operator interactions, provides the broader scale of organism interactions that keep the model from being a constantly changing patch of that reef. This function is particularly important in a nearly closed model. Here, normal population fluctuations could provide a dead end in community structure without the readily available larger reef surface that provides for the immigration and emigration of organisms and/or their reproductive stages. Table 17.1 provides the primary matching physiological conditions of model to wild reef ecosystem in a more static, long-term mode. The experimentation that determined the more diurnal, dynamic characterization of the unit, as briefly given below, was carried out for 10 months in 1998. At that time, the 400-l microcosm unit had been operating for about 10 years. Both the 400-l microcosm and the 1280-l refugium were in operation and essentially closed (except for a few experimental species) from 1991 to the time of implementation of the experimentation (i.e., about FIGURE 17.1 Diagram of the microcosm. Lights not shown on 400 L metabolic unit; see Adey and Loveland 8 for a detailed mechanical description of this unit. Impeller pumps (not shown) deliver water from the right- hand end of the refugium unit to the Algal Turf Scrubber and from the left side of the refugium unit to the wave bucket. Distilled water is added to compensate for evaporation. Experimental coral reef microcosm (5.0 m 2 ; 1,680 L) Import mixed feed (0.41 g/day) Bellows pumps 8 liter refugium from fish Algal turf scrubber Export of dried algae Manual transfer of organisms Lighting: Six 400 W metal halides Wave bucket Export water: 2 L/day Manual transfer of exchange water: 2 L/day Import seawater: 2 L/day Patch sand Reef Wave surge Algal turf scrubber Unit for metabolic work (0.757 m 2 ; 400 L) Refugium unit (4.29 m 2 ; 1,280 L) 2073_C017.fm Page 305 Friday, April 7, 2006 5:17 PM © 2006 by Taylor & Francis Group, LLC 306 Coral Reef Restoration Handbook 7 years). The model and its refugium had been constructed and fully stocked, from the wild, from about 1988 to 1990. Figure 17.2, Figure 17.3, and Figure 17.4 provide the most essential diurnal and long-term dynamic characteristics of the model, graphically showing diurnal oxygen cycling, long-term diurnal respiration and primary productivity, and diurnal carbonate cycling. Those diagrams also show the relationship of those basic metabolic characteristics to wild coral reefs, and especially the St. Croix–type reef. The long-term, whole-system (400-l unit) calcification rate of the model was 4.0 ± 0.2 kg CaCO 3 /m 2 /yr. Although the calcification rate follows from the oxygen/carbonate cycling, it provides a more direct and summary measure of the ecophysiological “success” of the model. This matches the “top” 2 to 4% of reefs worldwide. 13 This number cannot be directly related to the St. Croix reefs because those specific data were not collected at that reef. However, the upward growth rates of the St. Croix reef, as determined by core-drilling and C14 analysis, provide equivalent calcification rates, and these are among the higher rates described for coral reefs worldwide. Also, calcification rates of several individual species of stony corals were experimentally obtained by short-term isolation within the model. The highest rates, for a branching Acropora species, were 8.1 ± 0.7 kg CaCO 3 /m 2 /yr. This rate is about as high as the highest rates reported from the wild. High calcification rate is among the most important characteristics that allow coral reefs to develop their unique structure and topography and consequently their highly diverse community structure. It is clear from the metabolism and calcification rates briefly described above, for both wild and micro- cosm coral reefs, that it is possible to match basic reef function in a controllable model. Although many physical/chemical variables of the St. Croix reefs were considered in the development of the Caribbean Model design (e.g., key nutrients, diurnal oxygen concentrations, diurnal carbonate cycling and carbon import and export, temperatures, wave action, lighting, etc.), in summary, ecophysiological success was determined by whole-system calcification rates. Primary TABLE 17.1 Comparisons between Microcosm and St. Croix Reefs (annual mean or mean daily range with standard error) Microcosm St. Croix Reefs (fore reef) a Temperature ( ° C) (am-pm) 26.5 ± 0.03 ( n = 365) − 27.4 ± 0.02 ( n = 362) 24.0 − 28.5 Salinities (ppt) 35.8 ± 0.02 ( n = 365) 35.5 a pH (am-pm) 7.96 ± 0.01 ( n = 62) − 8.29 ± 0.02 ( n = 39) 8.05 − 8.35 a Oxygen concentration (mg l − 1 ) (am-pm) 5.7 ± 0.1 ( n = 14) − 8.7 ± 0.2 ( n = 11) 5.8 − 8.5 GPP(g O 2 m − 2 day − 1 ); (mmol O 2 m − 2 day − 1 ) 14.2 ± 1.0 ( n = 4); 444 ± 3 ( n = 4) 15.7; 491 Daytime NPP (g O 2 m − 2 day − 1 ); (mmol O 2 m − 2 day − 1 ) 7.3 ± 0.3 ( n = 4); 228 ± 9 ( n = 4) 8.9; 278 Respiration (g O 2 m − 2 h − 1 ); (mmol O 2 m − 2 h − 1 ) 0.49 ± 0.04 ( n = 4); 15.3 ± 1.3 ( n = 4) 0.67; 20.9 N-NO − 2 + NO 2 3 ( µ mol) 0.56 ± 0.07 ( n = 6) 0.28 Calcium (mg l − 1 ); (mmol l − 1 ) 491 ± 6 ( n = 33); 12.3 ± 0.2 ( n = 33) 417.2 a ; 10.4 Alkalinity (meq l − 1 ) 2.88 ± 0.04 ( n = 59) 2.47 a Light a ( Langleys day − 1 ) 220 430 (surface); 220 (5 m deep in fore-reef) a Data from Small and Adey, 2001. 13 2073_C017.fm Page 306 Friday, April 7, 2006 5:17 PM © 2006 by Taylor & Francis Group, LLC Lessons Learned in the Construction and Operation of Coral Reef Microcosms and Mesocosms 307 control of all of these biogeochemical processes was attained through the use of ATS processes. In its essentials, ATS is an algal control system that simulates a large body of water (open-ocean tropical surface water in this case) through manipulation of an externally sited, coral reef algal turf community. 8 ATS has a 25-year history of development and application and is now used in commercial-scale aquaculture and landscape-scale surface water purification. 15 It is therefore appli- cable to any scale of coral reef restoration, whether for the development of the actual site, as a heuristic tool to help the restorationists understand the dynamics of the project, or simply as an FIGURE 17.2 Comparison between oxygen concentrations of St. Croix and the microcosm. Standard error bars shown where n varies for each point; starting from the left: n = 8, 3, 11, 3, 3, 11, 3, 3, 11, 3, 3, 4, 9, 3, 3. St. Croix data from reference 11. The first set of tank lights comes on at 06:00 h and go off at 18:00 h; the second set comes on at 08:00 h and goes off at 20:00 h. The scrubber lights come on at 19:00 h and go off at 07:00 h. 13 FIGURE 17.3 Gross primary production vs. respiration. Modified after reference 18. Line shows 1:1 relationship. Microcosm Microcosm vs. St. Croix 9.5 9.0 8.5 8.0 7.5 7.0 Oxygen (mg l −1 ) 6.5 6.0 5.5 5.0 0600 0800 1200 1600 2000 2400 0400 1000 1400 1800 Time 2200 0200 0600 St. Croix 400 300 200 GPP (mol Cm −2 yr −1 ) 100 0 0 100 Respiration (mol Cm −2 yr −1 ) 200 St. Croix fore-reef Microcosm 300 2073_C017.fm Page 307 Friday, April 7, 2006 5:17 PM © 2006 by Taylor & Francis Group, LLC 308 Coral Reef Restoration Handbook organism maintenance tool. Note that at small scale, the refugium unit in the Caribbean Model was employed as just such a management tool for the smaller “metabolic” reef. As described above, in the design and construction of the model, considerable efforts have been expended to duplicate flow rates (currents) and wave surge and period (though not wave magnitudes (see Reference 13). In their review of coral reef biogeochemistry, Atkinson and Falter 20 conclude that these are indeed critical factors. In this chapter, as in all precedent studies, the author has emphasized the role of low nutrients, as have most workers in the field. Atkinson and Falter 20 conclude that the nutrient factor in reef biochemistry is poorly understood, and perhaps under some conditions not a significant factor (in the author’s view, this may well be so when light levels are very low, either in deep reefs or in poorly lighted models). As earlier demonstrated by Adey 21 and repeated by Atkinson and Falter, 20 C:N:P ratios are high in macroalgae from wild reef systems with oligotrophic nutrient levels, but this does not reduce high primary productivity. 21 In the same environments, physical energy (wave surge and current) allows algal turfs to achieve high uptake rates and normal C:N:P ratios (and thereby achieve the high rates of productivity seen). This same process is repeated in the large-scale ATS water-cleaning systems now in use to achieve oligotrophic water quality in degraded fresh waters. 15 Carbonate skeletal construction (i.e., framework-building) by coral reef organisms, especially corals, is the key to the development and maintenance of controlled coral reef ecosystems. While it is therefore appropriate to measure the success of coral reef microcosm modeling in part by measuring whole reef calcification rates, it is unlikely that anyone would consider coral reef restoration exclusively for the production of carbonate. The essential secondary consideration would certainly be biodiversity. Some might argue that fish biodiversity alone is the key element. However, except for the specialty artificial reefs built for anglers, where the fish populations are accepted to be based on plankton populations and the reef mostly provides structural habitat, broad-band restored coral reef fish populations can only be maintained with an even greater broad band of algal and invertebrate biodiversity. Hopefully, considerable efforts are currently underway to define biodiversity in at least a few, scattered wild reefs. However, at this time, the only direct measure of coral reef biodiversity is the Caribbean Model that is the focal point of this chapter. 16 In the 5-m 2 , 1680-l volume of that model, after 7 years of semi-closure, 534 species were tallied by a team of 24 specialists (Table 17.2). FIGURE 17.4 Daytime carbonate cycle in Caribbean Model, as calculated by nomograph from total alkalinity and pH. 19 Microcosm Total Alkalinity 95 93 91 89 Total Alk. & HCO 3 mg/l 87 85 83 81 79 77 75 600 800 1000 1200 1400 1600 1800 2000 8.02 Time pH 7.97 8.17 8.29 8.24 0 2 4 6 8 10 12 14 16 18 20 CO 3 & CO 2 mg/l CO 3 = CO 2 HCO 3 − 2073_C017.fm Page 308 Friday, April 7, 2006 5:17 PM © 2006 by Taylor & Francis Group, LLC Lessons Learned in the Construction and Operation of Coral Reef Microcosms and Mesocosms 309 TABLE 17.2 Families, Species, and Genera Tallied in the Caribbean Model Plants, Algae, and Cyanobacteria Division Cyanophota Chroococcaceae 6/5 Pleurocapsaceae 4/2 UID family 4/4 Oscillatoriaceae 8/6 Rivulariaceae 4/1 Scytonemataceae 1/1 Phylum Rhodophyta Goniotrichaceae 2/2 Acrochaetiaceae 2/2 Gelidiaceae 1/1 Wurdemanniaceae 1/1 Peysonneliaceae 3/1 Corallinaceae 11/8 Hypneaceae 1/1 Rhodymeniaceae 3/2 Champiaceae 1/1 Ceramiaceae 3/3 Delesseriaceae 1/1 Rhodomelaceae 7/6 Phylum Chromophycota Cryptomonadaceae 2/2 Hemidiscaceae 1/1 Diatomaceae 6/4 Naviculaceae 9/4 Cymbellaceae 3/1 Entomoneidaceae 1/1 Nitzchiaceae 6/4 Epithemiaceae 3/1 Mastogloiaceae 1/1 Achnanthaceae 9/3 Gymnodiniaceae 6/4 or 5 Gonyaulacaceae 1/1 Prorocentraceae 2/1 Zooxanthellaceae 1/1 Ectocarpaceae 2/2 Phylum Chlorophycota Ulvaceae 1/1 Cladophoraceae 4/2 Valoniaceae 2/2 Derbesiaceae 3/1 Caulerpaceae 3/1 Codiaceae 6/2 Colochaetaceae 1/1 Phylum Magnoliophyta Hydrocharitaceae 1/1 Kingdom Protista Phylum Percolozoa Vahlkampfiidae 2/1 UID family 2/2 Stephanopogonidae 2/1 Phylum Euglenozoa UID family 4/3 Bondonidae 7/1 Phylum Choanozoa Codosigidae 2/2 Salpingoecidae 1/1 Phylum Rhizopoda Acanthamoebidae 1/1 Hartmannellidae1/1 Hyalodiscidae 1/1 Mayorellidae 2/2 Reticulosidae 2/2 Saccamoebidae 1/1 Thecamoebidae 1/1 Trichosphaeridae 1/1 Vampyrellidae 1/1 Allogromiidae 1/1 Ammodiscidae 1/1 Astrorhizidae 1/1 Ataxophragmiidae 1/1 Bolivinitidae 3/1 Cibicidiidae 1/1 Cymbaloporidae1/1 Discorbidae 5/2 Homotremidae 1/1 Peneroplidae 1/1 Miliolidae 10/2 Planorbulinidae 2/2 Siphonidae 1/1 Soritidae 4/4 Textulariidae 1/1 Phylum Ciliophora Kentrophoridae 1/1 Blepharismidae 2/2 Condylostomatidae 1/1 Folliculinidae 4/3 Peritromidae 2/1 Protocruziidae 2/1 Aspidiscidae 7/1 Chaetospiridae 1/1 Discocephalidae 1/1 Euplotidae 11/3 Keronidae 7/2 Oxytrichidae 1/1 Psilotrichidae 1/1 Ptycocyclidae 2/1 Spirofilidae 1/1 Strombidiidae 1/1 Uronychiidae 2/1 Urostylidae 4/2 Cinetochilidae 1/1 Cyclidiidae 3/1 Pleuronematidae 3/1 Uronematidae 1/1 Vaginicolidae 1/1 Vorticellidae 2/1 Parameciidae 1/1 Colepidae 2/1 Metacystidae 3/2 Prorodontidae 1/1 Amphileptidae 3/3 Enchelyidae 1/1 Lacrymariidae 4/1 Phylum Heliozoa Actinophyridae 2/1 Phylum Placozoa UID Family 5 Phylum Porifera Plakinidae 2/1 Geodiidae 5/2 Pachastrellidae 1/1 Tetillidae 1/1 Suberitidae 1/1 Spirastrellidae 2/2 Clionidae 4/2 Tethyidae 2/1 Chonrdrosiidae 1/1 Axinellidae 1/1 Agelasidae 1/1 Haliclonidae 4/1 Oceanapiidae 1/1 Mycalidae 1/1 Dexmoxyidae 1/1 Halichondridae 2/1 Clathrinidae 1/1 Leucettidae 1/1 UID family 2/? Eumetazoa Phylum Cnidaria UID family 3/? Eudendridae 1/1 Olindiiae 1/1 Plexauridae 1/1 Anthothelidae 1/1 Briareidae 1/1 Alcyoniidae 2/2 Actiniidae 3/2 Aiptasiidae 1/1 Stichodactylidae 1/1 Actinodiscidae 4/3 Corallimorphidae 3/2 Acroporidae 2/2 Caryophylliidae 1/1 Faviidae 3/2 Mussidae 1/1 Poritidae 3/1 Zoanthidae 3/2 Cerianthidae 1/1 Phylum Platyhelminthes UID family 1/1 Anaperidae 3/2 Nemertodermatidae 1/1 Kalyptorychidae1/1 Phylum Nemertea UID family 2/2 Micruridae 1/1 Lineidae 1/1 Phylum Gastrotricha Chaetonotidae 3/1 Phylum Rotifera UID family 2/? Phylum Tardigrada Batillipedidae 1/1 Phylum Nemata Draconematidae 3/1 Phylum Mollusca Acanthochitonidae 1/1 Fissurellidae 2/2 Acmaeidae 1/1 Trochidae 1/1 Turbinidae 1/1 Phasianellidae 1/1 Neritidae 1/1 Rissoidae 1/1 Rissoellidae 1/1 Vitrinellidae 1/1 Vermetidae 1/1 Phyramidellidae 1/1 Fasciolariidae 2/2 Olividae 1/1 Marginellidae 1/1 Mitridae 1/1 Bullidae 1/1 UID family 4/? Mytilidae 2/1 Arcidae 2/1 Glycymerididae 1/1 Isognomonidae 1/1 Limidae 1/1 Pectinidae 1/1 Chamidae 1/1 Lucinidae 2/2 Carditidae 1/1 Tridacnidae 2/1 Tellinidae 1/1 (continued) 2073_C017.fm Page 309 Friday, April 7, 2006 5:17 PM © 2006 by Taylor & Francis Group, LLC 310 Coral Reef Restoration Handbook Since some groups of algae and invertebrates were omitted from the tally due to the lack of a specialist, the authors estimated that the system actually contained about 800 species. Except for a dozen species undergoing manipulation and a few long-lived fish species, this microcosm had been closed to organism import for 7 years prior to analysis; therefore, most of the 500+ species were reproductively maintaining their populations. Based on the biodiversity of this model, and using the relationship S = kA z , it has been estimated that the pantropic biodiversity of wild coral reefs must be at least 3 million species. One must conclude therefore that unless wild coral reefs have a considerably higher biodiversity (the previous maximum estimate being 1 million species 17 ), microcosm models can be successful in biodiversity as well as ecophysio- logical terms. A glance at the family list in Table 17.2 shows that the biodiversity of this system was very widely based. A few of the reported 230+ families tallied in the Caribbean Model have a half- dozen species in two to five genera. However, most families have only a single species. Thus, while the system is oriented toward the relative success of smaller organisms (e.g., protists and annelids), and at least for the larger organisms, the lower-to-middle levels of food webs, it does not have an ecophysiology that is selecting for a limited range of food web and community structure charac- teristics. The number of functionally closely related species is likely limited, simply by space (i.e., the small size of the model). There is no room for “pulsing” between adjacent patches in this small model, and species that occupy the same ecological niches cannot coexist because of distance, as many species will be able to do in a large wild coral reef. While species diversity is high by any measure, 16 family diversity for such a small system is likely to be truly extraordinary. The linking of several separate systems through piping and nonstressing pumps would likely proportionally increase species diversity by simulating adjacent patches. In a more sophisticated system, organism cross-access to those patches could be controlled by gates or organism-specific filters; in short, they could function as controlled refugia. This bodes well for the reef restorer who would invest the considerable effort to work with whole coral reef ecosystems and not a limited subset of the more visible or spectacular species. TABLE 17.2 Families, Species, and Genera Tallied in the Caribbean Model (Continued) Phylum Annelida Syllidae 3/2 Amphinomidae 1/1 Eunicidae 3/1 Lumbrineridae 1/1 Dorvilleidae 1/1 Orbiniidae 1/1 Spionidae 1/1 Chaetopteridae 1/1 Paraonidae 1/1 Cirratulidae 4/3 Ctenodrilidae 4/3 Capitellidae 3/3 Muldanidae 1/1 Oweniidae 1/1 Terebellidae 2/1 Sabellidae 14/4 Serpulidae 6/6 Spirorbidae 2/2 Dinophilidae 1/1 Phylum Sipuncula Golfingiidae 1/1 Phascolosomatidae 3/2 Phascolionidae 1/1 Aspidosiphonidae 3/2 Phylum Arthropoda Halacaridae 1/1 UID family 2/? Cyprididae 2/2 Bairdiidae 1/1 Paradoxostomatidae 1/1 Pseudocyclopidae 1/1 Ridgewayiidae 2/1 Ambunguipedidae 1/1 Argestidae 1/1 Diosaccidae 1/1 Harpacticidae 1/1 Louriniidae 1/1 Thalestridae 1/1 Tisbidae 1/1 Mysidae 1/1 Apseudidae 2/1 Paratanaidae 1/1 Tanaidae 1/1 Paranthuridae 1/1 Sphaeromatidae 1/1 Stenetriidae 1/1 Juniridae 1/1 Lysianassidae 1/1 Gammaridae 4/4 Leucothoidae 1/1 Anamixidae 1/1 Corophiidae 1/1 Amphithoidae 2/2 Alpheridae 2/2 Hippolytidae 2/1 Nephropidae 1/1 Diogenidae 1/1 Xanthidae 2/? Phylum Echinodermata Ophiocomidae 1/1 Ophiactidae 1/1 Cidaroidae 1/1 Toxopneustidae 1/1 Holothuriidae 1/1 Chirotidae 1/1 Phylum Chordata Ascidiidae UID species 1/1 Grammidae 1/1 Chaetodontidae 1/1 Pomacentridae 5/4 Acanthuridae 1/1 Note: Ceramiaceae 3/3 = three species in three genera in the family Ceramiaceae. UID = unidentified. 2073_C017.fm Page 310 Friday, April 7, 2006 5:17 PM © 2006 by Taylor & Francis Group, LLC Lessons Learned in the Construction and Operation of Coral Reef Microcosms and Mesocosms 311 17.3 THE OPERATIONAL IMPERATIVE Successful microcosm and mesocosm operation requires the monitoring of a large number of physical and chemical factors. To a large extent, this can be automated with electronic sensors, and the data can be logged and the system computer controlled. Some chemical parameters, such as the low-level nutrients that are characteristic of coral reef systems, still require wet chemistry, though a once-a-week analysis is usually sufficient in a well-run system (even nutrient sensing and control can probably be automated, but the cost could well be prohibitive for restoration programs). Like any piece of complex laboratory equipment (a scanning electron microscope, for example) a dedicated and highly trained technician is needed to manage the monitoring equipment, though in a well-tuned system, considerable time can be available for other duties. ATS management is effectively manual, though this is typically not a large consumer of technician time until a system exceeds hundreds of cubic meters. The ATS system of the Caribbean Model typically required about half an hour per week for physical/chemical maintenance. An operational feature that is rarely discussed, and in practice is mostly anecdotal in expres- sion, is that of population instability. A microcosm, in effect, is a few-square-meter patch of a larger coral reef ecosystem. In the wild, reef patches of a few square meters can be subject to considerable short-term variability, though stability is achieved to some extent by the smoothing effect of the larger local island or coastal reef that may be measured in square kilometers. On the other hand, even large geographic areas can be subject to population explosions. Coral reef scientists learned this in a very vivid way in the 1970s and 1980s when the Diadema populations in the Caribbean slowly built up to very high densities and then suddenly collapsed. Thus, whether the restorer is dealing with an open-water project of many tens of thousands of square meters, a research model, or a holding system for organisms destined for a coral reef being restored, the principles are the same. The type microcosm of this chapter operated in a closed mode, in essentially the same ecological state, for 7 years (nearly 9 years including the period of intense analysis). Nothing that could remotely be described as a “crash” ever occurred in this coral reef model, in spite of occasional physical or electronic failures. However, on the scale of several months, single populations of the system (e.g., Caulerpa spp., various dinoflagellate species, and a few species of polychaete worms) would undergo a population explosion. Usually, this tendency, for which the observant operator typically had weeks of warning time, would last a few months, and then the subject population would reduce to “background” levels. Sometimes, the explosion would recur years later; in other cases it was a one-time experience. It has been my experience that microcosms and mesocosms require an ecologist, fully acquainted with “normal” community structure of the “wild-type” system. Effectively, that ecolo- gist/operator performs as the highest, and most omnivorous, predator. This “variable” predator (if not on vacation or otherwise detained) is instantly available at “full population” to limit short- term population inbalance. In a 10-year study of scaling effects on Chesapeake Bay mesocosms, researchers often were able to solve scaling problems by just such ecological manipulation. 22 In the case of Caulerpa and Prorocentrum “explosions” in the Caribbean Model, the operator’s function was obvious, a once-a-week “grazing” (i.e., hand harvest) of the Caulerpa or similarly a once-a-week scraping and collection of mostly glass surfaces for Prorocentrum until the explosion tendency subsided. In other cases, the short-term introduction of a fish or invertebrate predator (such as angelfish for fire worms) could be quite successful. In some cases, it has even been valuable to maintain such “managed predators” in the refugium unit where they are readily available for such service. It is the view of some coral reef ecologists, struggling to understand why wild reefs are now apparently rapidly degrading, that extensive (albeit local) fishing (effectively large-scale species manipulation) is the primary factor responsible for reef decline. This demon- strates that human interaction with coral reefs can occur at any scale. This aspect of microcosm operation has considerable application to coral reef restoration, as will be discussed further below. 2073_C017.fm Page 311 Friday, April 7, 2006 5:17 PM © 2006 by Taylor & Francis Group, LLC 312 Coral Reef Restoration Handbook 17.4 IMPLICATIONS OF MICROCOSM MODELING FOR CORAL REEF RESTORATION An extensive ecophysiological as well as biodiversity and community structure understanding of the wild system to be modeled is the primary key to the restoration of coral reef ecosystems. The success achieved with the microcosm briefly described herein was based, more than any other single factor, on having an existing, functioning wild ecosystem (and not an elusive ideal) that had to be matched. Most critically, if negative factors of water quality, such as turbidity and nutrients, cannot be returned to the state of the wild system to be emulated, success is unlikely. For example, although the direct role of elevated nutrients has not been fully defined (indeed, it is probably quite complex), it appears certain that most coral reef ecosystems require that the primary nutrients, phosphorus and nitrogen, be measured in single digits of parts per billion. In the most successful of the coral reef microcosms discussed in this chapter, primary nutrients were maintained at extremely low levels for many years at a time. Nevertheless, if the coral reef restorationist felt that a particular reef ecosystem could maintain itself in the face of moderate nutrient levels, once its community structure was restored during a period of high water quality, nutrient scrubbing, very large volumes of water (tens of millions of gallons/day) can be achieved if adjacent land surface is available. 15 Alternatively, if coral reefs of restoration interest have been degraded by the low water quality of incoming fresh waters from human activities in nearby terrestrial environments, ATS fresh water systems can scrub rivers, streams, and sewage outfalls of nutrients, sediments, and toxic compounds. Also, in situ culture methods have been devised that can re-establish reefs destroyed by hurricanes, low water quality, or excessive temperatures. 23 However, if return of low water quality or high temperatures cannot be avoided, considerable effort can be expended that will be of little value to long-term reef restoration. Development and operation of the kind of laboratory ecosystem model described in this chapter can be an extremely valuable heuristic tool and can provide ecological experience that could be achieved only over long periods of time on a wild reef. Effectively, a microcosm or mesocosm coral reef can be a pilot project carried out prior to the initiation of a full restoration operation. While the system described above was operated for a decade, such long-term operation is not essential to success. Short-term stability in coral reef models is usually attainable on the scale of months. Also, such microcosms are useful in the developing and testing of many experimental protocols. Microcosms are often capable of producing experimental results in a shorter time frame than in the wild because variables are more controllable. Finally, restoration of a wild reef will generally require considerable organism manipulation by a wide range of specialists working as a team. Particularly if a high-quality source reef and its organisms are at a considerable distance from the reef to be restored, a moderately sized microcosm (probably a mesocosm in most cases) would be an extremely valuable resource to hold the organisms being manipulated. As described above, for the Caribbean Model, success in achieving organism stability in the metabolism unit was assured by having a “refugium” that allowed more-or-less instant manipulation of many organisms. Another aspect of ATS-controlled units is the ability to modify such systems, either in the main microcosm or mesocosm, or perhaps more appropriate in a restoration effort, in a side loop, for aquaculture purposes. This has been demonstrated in a large- scale commercial aquaculture for fin fish, and could be modified relatively easily to produce large numbers of a particular species, algal, invertebrate, or vertebrate, that are needed for introduction but are not readily available at reasonable shipping distances. 17.5 SUMMARY Twenty-five years of intensive and repeated ecosystem modeling of coral reefs, as microcosms and mesocosms, has demonstrated this approach to be a viable experimental tool. Such a system or systems can provide prior, experimental understanding of the problems to be overcome in a specific 2073_C017.fm Page 312 Friday, April 7, 2006 5:17 PM © 2006 by Taylor & Francis Group, LLC [...]...2073_C 017. fm Page 313 Friday, April 7, 2006 5 :17 PM Lessons Learned in the Construction and Operation of Coral Reef Microcosms and Mesocosms 313 coral reef restoration project Also, in a situation where healthy source reefs lie at a considerable distance from the reef to be restored or would be depleted by the required removal of one... of corals cultured at the Waikiki Aquarium, Honolulu, Hawaii Coral Reefs 14: 215–223 Borneman, E 2001 Aquarium Corals: Selection, Husbandry, and Natural Selection TF Neptune Publications, Charlotte, VT Adey, W 1983 The microcosm, a new tool for reef research Coral Reefs 1: 193–201 Luckett, C., W Adey, J Morrissey, and D Spoon 1996 Coral reef mesocosms and microcosms— successes, problems, and the future... Adey, W 2001 Reef corals, zooxanthellae, and free-living algae: a microcosm study that demonstrates synergy between calcification and primary production Ecol Eng 16: 443–457 McConnaughey, T., W Adey, and A Small 2000 Community and environmental influences on coral reef calcification Limnol Oceanogr 45: 1667–1671 www.hydromentia.com Small, M., W Adey, and D Spoon 1998 Are current estimates of coral reef biodiversity... Gilpin, and J Alber, Eds Restoration Ecology Cambridge Univ Press, Cambridge, U.K pp 134–149 Adey, W 1975 The algal ridges and coral reefs of St Croix Atoll Res Bull 187: 1– 67 Adey, W and R Steneck 1985 Highly productive eastern Caribbean reefs: synergistic effects of biological, chemical and geological factors NOAA Symposium Series for Undersea Research 3: 163–187 Adey, W 1978 Coral reef morphogenesis:... depleted by the required removal of one or several species, local model systems can be an invaluable organism culture and manipulative asset during the restoration process Control of global degrading factors may be critical to long-term coral reef restoration We now have the tools to achieve global water quality control, although whether sociopolitical factors will allow that to happen remains to be... The view through the window of a microcosm Atoll Res Bull 458: 1–18 Reaka-Kudla, M 1997 The global biodiversity of coral reefs: a comparison with rain forests In Reaka-Kudla, M., D Wilson, and E.O Wilson, Eds Biodiversity II: Understanding and Protecting our Biological Resources Joseph Henry Press, Washington, D.C Gattuso, J.-P., M Frankignoulle, and R Wollast 1998 Carbon and carbonate metabolism in... 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Carlson, B 1987 Aquarium systems for living corals Int Zool Yb 26: 1–9 Moe, M 1989 The Marine Aquarium Reference Green Turtle Publications Plantation, FL Spotte, S 1995 Captive Seawater Fishes, Science and Technology Wiley New York Atkinson, M., B Carlson, and G Crow 1995 Coral growth in high nutrient, low pH sea water: a case study of corals cultured at the... Atkinson, M and J Falter 2003 Coral reefs In Black K and G Shimmield Biogeochemistry of Marine Systems CRC Press, Oxford, U.K 372 pp Adey,W 1987 Food production in nutrient poor seas: bringing tropical ocean deserts to life Bioscience 37: 340–348 Petersen, J E et al 2003 Multiscale experiments in coastal ecology: improving realism and advancing theory Bioscience 53: 1181–1197 www.reefball.com © 2006 by Taylor . below. 2073_C 017. fm Page 311 Friday, April 7, 2006 5 :17 PM © 2006 by Taylor & Francis Group, LLC 312 Coral Reef Restoration Handbook 17. 4 IMPLICATIONS OF MICROCOSM MODELING FOR CORAL REEF RESTORATION An. yr −1 ) 200 St. Croix fore -reef Microcosm 300 2073_C 017. fm Page 307 Friday, April 7, 2006 5 :17 PM © 2006 by Taylor & Francis Group, LLC 308 Coral Reef Restoration Handbook organism maintenance. endeavors were undertaken to model coral reef 2073_C 017. fm Page 303 Friday, April 7, 2006 5 :17 PM © 2006 by Taylor & Francis Group, LLC 304 Coral Reef Restoration Handbook ecosystems in a