Data in this set come from two studies: "Ability of protected reefs to resist alien algae" and
"How many fish does it take to keep the alien algae out?" Both are authored by Ms. Danielle Jayewardene and
Dr. Charles Birkeland of University of Hawaii Manoa, in 2005 and 2006, respectively. For the former, the goal
was to objectively and quantitatively assess the ability of marine protected areas (MPAs), to maintain ecosystem
health, and to thereby resist invasion of alien algae. For the latter, specific objectives involved a) determining
whether there are differences in cropping ability by larger herbivorous fishes compared to smaller herbivorous
fishes, and b) whether there is a threshold level of algal biomass above which even an increased number of
herbivorous fishes simply cannot crop down. The field work took place at eight coral reef sites located on the
island of Oahu and the island of Hawaii.
To objectively and quantitatively assess the ability of MPAs to restore ecosystem health, and thereby
resist invasion of alien algae. The primary research objectives of this project were to experimentally determine
and measure the differences between coral reefs inside and outside MPAs. Also, to determine the level of grazing
pressure needed to crop down a standing biomass of algae on the reef.
These sites were surveyed using SCUBA equipment and an Olympus digital camera. Data was analyzed
in the laboratory. The project was launched by the Hawaii Coral Reef Intiative under the Department of Zoology, at
the University of Hawaii at Manoa.
Dataset credit required
Edmonton 164, University of Hawaii at Manoa
Hawaii Coral Reef Initiative, Dept. of Zoology, University of Hawaii, US Geological Survey
see lineage, process step
I. From 2005 Report: For the purpose of categorizing our study sites based on ranging biomass levels of herbivorous fish we selected study sites where fish biomass data were already available. Along the Kona coast, we chose sites and used current data collected in 2003/04 by the West Hawaii Aquarium Project, WHAP (www.coralreefnetwork.com/kona/default.htm). On Oahu, we used data collected in 2000 by the Hawaii Coral Reef Assessment and Monitoring Program, CRAMP (cramp.wcc.hawaii.edu). In order to link the observations and experiments gained from our fieldwork with the herbivorous fish biomass data, it was necessary that our sites were compatible with the WHAP and CRAMP sites. At the Kona sites, WHAP has collected their fish data bimonthly since 1998 by running 2 x (2 x 25m) parallel permanent fish transects at a depth between 9-13m. At Hanauma Bay on Oahu, CRAMP has run 2 x 50m parallel transects at a depth of 10m between years 1998-2001. At Portlock, biomass data has been assessed from rapid fish transects. Based on these methods, our field work at all sites has been focused around the sites of the fish transects at depths near 10m.However, at Hanauma Bay we also set out experiments at a shallower depth of 5m and at different locations within the Bay to enable within site comparisons. In order to make quantitative assessments of our objectives, the following standardized methods were designed to determine and compare each of the ecological factors between study sites. 1. Grazing pressure Grazing pressure was assessed by comparing percent cover of filamentous and macroalgae between sites. This is an indirect method of assessing grazing intensity. It is based on the assumption that increased grazing leads to a lower prevalence of macroalgae and also filamentous algae. Percent cover was determined by analyzing photoquadrats that were taken at each site. 2. Prevalence of crustose coralline algae (CCA) We assessed the prevalence of CCA by quantifying the percent cover of CCA present at each of the sites. This was also done by taking photoquadrats in the field. CCA was not identified down to genus or species level, but was lumped as one overall CCA category. This was to avoid error due to incorrect identification. 3. Prevalence of coral colonies We have assessed the prevalence of coral colonies by quantifying the percent cover of coral at each of the sites. This has also been done by means of photoquadrats. Corals have been identified down to species level. 4. Success of juvenile coral survival We examined success of juvenile coral survival by transplanting juveniles into the field using tiles and monitoring their survival over time. Setting out terracotta tiles rather than direct transplantation onto the substrata standardized the measurements as well as avoided manipulation of coral within protected areas. Survival of three size classes of juvenile corals was determined, involving two different methods of preparation. The smallest size class tested was newly settled larvae of Pocillopora damicornis. The larvae were attracted to terracotta tiles in the lab at the Hawaii Institute of Marine Biology (HIMB) before being set out in the field. This was done by holding coral heads of P. damicornis (10- 20 cm in size) in microcosm tanks (approx. 1m x 1m x 0.5m), and leaning 6 x 6 inch terracotta tiles against the walls of the tanks providing substrata for settling planulae. P. damicornis, which is a brooding coral species with internal gamete fertilization, releases planulae once a month according to the lunar cycle. These traits are the justification for using this particular species, as spawning coral species show external fertilization and release gametes only once a year. As soon as the planulae were released and settled on the tiles, they were counted and mapped on underwater paper. We managed to attract up to 60 juveniles per tile. Eight of the tiles holding the highest number of spat were set out at each site. The juveniles were monitored at one-week intervals for up to 5 weeks. Survival was defined as the number of spat remaining alive per tile. Unfortunately, we did not find appropriate facilities such as HIMB in Kona to allow for the attraction of spat onto tiles, so tiles with spat were only set out at the Oahu sites. To determine whether factors affecting survival differed depending on the size of the juveniles, replicates of two larger size classes were also exposed to natural mortality factors in the field. The intermediate size class consisted of approx. 1 inch high Porites compressa nubbins and the large size class consisted of approx. 21/2 inch high Porites compressa nubbins. P. damicornis nubbins were not used since they were too readily eaten by fish in the field, invalidating comparison between sites. The two size classes of P. Compressa nubbins were attached to tiles using Sea Goin Poxy Putty and transplanted into the field similar to the spat. Nubbins were monitored at one week intervals for 4 or more weeks. Survival was defined as the maintenance of nubbins in their initial size and healthy state. If a fish bit off half the nubbin, the nubbin was considered to have 50 percent survival. A total of 8 replicate tiles were set out per study site. The tiles were set out at two substations holding 4 tiles each. 5. The balance between reef accretion and bioerosion We determined bioerosion by experimentally measuring the level of bioerosion of calcium carbonate blocks set out on the reef, and accretion by determining the calcification budgets of the calcifying community characteristic to sites. Following the methodology of Hibino & Van Woesik (2000), replicate calcium carbonate blocks were attached to dead reef substrate and left out for 12 months at each of the sites. The uniform blocks were cut from dead coral tissue of Porites lobata. Dead P. lobata was collected from the north shore where large blocks are washed up on shore and dried out following the winter storms. Using a rock saw, any bioeroded material was cut away from the dead coral blocks, leaving a clean, non-impacted calcium carbonate core. This raw material was cut squarely into 1 cm x 5 cm x 5 cm blocks. Once the blocks were cut, they were soaked in freshwater for 24 hours then dried in an oven at 60 degrees Celsius for another 24 hours, and weighed. This was to eliminate any microorganisms that may be residing in the calcium carbonate as well as standardize the dry-weighing of the blocks. Thereafter numbers were engraved into each of the blocks, the bases siliconized and attached to the reef using Z-spar epoxy putty. After 12 months the blocks will be removed from the reef, and the rate of structural bioerosion for the blocks determined from the total measure of lost (or gained) weight for replicates per site (blocks remain in the field and will not be retrieved until the summer of 2005. Thus the following methods have yet to be applied). Before re-measurement, the blocks are prepared by a bleaching, soaking and drying process according to the method by Hibino & Van Woesik (2000). Thin cross sections may also be cut through the blocks (using a diamond geological blade) to quantify the loss of area due to internal bioerosion. A total estimate of structural bioerosion for each site can be calculated from the bioerosion rates determined from the blocks and the percent cover of non-coral substrate prone to bioerosion. Percent cover of non-coral substrate is determined through the photoquadrat methodology. At each site, 8 replicate blocks were set out at 2 substations holding 4 blocks each. At the Kona sites, the blocks were set out around the WHAP fish transects at a depth pf approx. 10m. At Hanauma Bay, 2 sets of 8 blocks were set out at 10m depth. One by the CRAMP fish transects on the right hand side of the bay and one set of blocks on the left hand side of the Bay. At Portlock 8 blocks have were set out at 10 m depth by the underwater amphitheatre. Coral accretion is assessed by determining the percent cover of different calcifying species present at a site, finding the growth rate of each species in the literature and using this to calculate the accretion rate of the calcifying community at each site. Percent cover of calcifying species has been determined using the photoquadrat methodology. Budgets of accretion versus bioerosion will be calculated by subtracting structural bioerosion from the total growth rate for each site (Hubbard et al. 1990, Hubbard 1992, 1997). Photoquadrat methodology We chose to take photoquadrats instead of using visual estimates or point quadrats in order to identify and quantify the benthos at study sites (Foster et al. 1991, Meese & Tomich 1993, Dethier et al 1993). Taking photoquadrats was more time efficient in the field, since more replicate quadrats were gathered with less time spent underwater. Furthermore, since photoquadrats provide a permanent database of the benthos, it is always possible to go back and re-analyze data if necessary, or to make temporal comparisons with future data. The drawback of the method was the extra time required analyzing the images in the laboratory. We based our photoquadrat methodology largely on the methodology used by WHAP in Kona. In the field: Each photoquad was taken using an Olympus 5050 camera. The camera was mounted onto a 75 cm high rod, with resulting images covering an area of approx. 60x40cm of substrate. For study sites in Kona where benthic cover was correlated with WHAPs fish data, photoquads were taken along a theoretical grid set around their 2 (2x 30m) parallel fish transects marked by permanent pins at each site. To get an appropriate representation of the area, 10 x individual 20m length transect lines were placed within this grid. The transect lines were laid out randomly by generating 10 random numbers between 1 and 50, and starting the transect at the point of each of these numbers in the grid. Along each transect, one image/meter was taken, totaling 20 images/transect and 200 images/site. At Hanauma Bay, photoquadrats were taken at multiple sites within the bay enabling between sites as well as within site comparisons of percent cover. Sites included the 10 m depth CRAMP site and a 5 m depth site on the right hand side of the bay, and 10 m depth site on the left hand side of the bay. In the lab: The images taken were analyzed using an image analysis program called PhotoGrid designed by UH Botany graduate student Chris Bird. PhotoGrid generated random points on each of the images. The substrate that fell directly below each point was identified and classified into categories including coral species, macroalgae, turf, and CCA. When all images for a site were analyzed in this manner, an overall percent cover was determined for each category or species. This percent cover was then compared between sites. To determine how many images/transect and points/image actually needed to be analyzed in order to get an accurate representation of percent cover at a site (whilst being time efficient), we ran a test for one of our more heterogeneous study sites, the Wawaloli FMA. All 200 images taken at this site were analyzed, and percent cover of the main categories compared when analyzing 10 points/image vs. 20 and 30 points/image. We also determined whether there was a point at which analyzing more images/transect was not worthwhile time-wise. The results indicated that the number of points/image really had little effect on the overall estimate of main categories such as coral, turf and CCA cover. Based on this, we decided to analyze 20 points/image. Although the graphs also indicated that it was adequate to analyze as few as 5 images/transect, we decided to analyze as many as 15 images/transect to allow for the accurate percent cover estimation of coral on a species level (we also ran a species level analysis indicating this). These two choices made analysis more time efficient whilst maintaining our ability to estimate an accurate cover. In total, 150 images/site, i.e. 3000 points/site compared to 6000 points/site, were analyzed and used to determine percent cover for all study sites. II. From 2006 Report Instead of manipulating the biomass levels of fish exposed to macro-algal communities to test grazing intensity in the field, algae were manipulated.This was done by cultivating high biomass algal communities in-situ on selected reefs and subsequently exposing these communities to the resident herbivorous fishes at these sites. This was conducted by setting out fish exclusion cages, preventing the herbivorous fishes from grazing the algae within, allowing the uninhibited growth and succession of algae in theprotected plots. Once the algal plots reached a high biomass level, the mesh on the cages was removed and the algae within the cages exposed to the grazing fishes. Grazing intensity was determined by monitoring the efficiency in which the fishes cropped down the algal plots. In order to compare cropping efficiencies by different herbivorous fish communities, experimental cages were set out on different reefs with ranging abundances of herbivorous fishes. Two sites were known to have a high biomass of herbivorous fishes including a significant proportion of large parrotfishes, (Ke'ei and Hanauma Bay), and three sites known to have a low biomass of herbivorous fishes with few large parrotfishes (Wawaloli, Portlock and Ko'olina). To test for threshold levels of algal biomass above which even a high intensity of grazing may not result in cropping down of the algae, the grazing fish communities at these sites were exposed to experimental plots with a lower biomass level of algae, compared to ones with a higher biomass level of algae. These different biomass levels were simulated in the plots by leaving half the experimental cages out for a shorter time than the others, thus exposing the algal plots to the grazing community at an earlier algal successional stage. To test the different cropping effects that large herbivorous fish compared to small herbivorous fish have on the simulated algal communities, we excluded only the large herbivorous fishes (most likely the larger parrotfishes) from cropping down a random selection of the opened plots. This was done by replacing the small mesh on the cages used to exclude all fishes with large mesh which would exclusively exclude the larger sized fishes. 16 cages were set out per site. At each site, the cages were placed randomly within a grid set around permanent fish transects run by WHAP and CRAMP at 10m depth (Figure 2). 8 plots where high biomass levels were simulated were out for 8 months while 8 plots intended to simulate a lower biomass were out for 5 months. The second set of cages were set out 3 months after the first set to allow all experimental plots, whether high or low biomass, to be released to grazing simultaneously. This eliminated potential differences in cropping ability resulting from temporal differences in grazing. Exclusion cages were constructed from weather resistant clothes line and a fine polypropylene aqua mesh. Mesh used to exclude all herbivores was 1/2 inch across in size, while mesh used to exclude only the larger fishes was 4 inches across in size. The plot area protected by each cage was approximately 900 cm2. The mesh did not manipulate water motion or light levels (pers.com Linda Preskitt). The cages were attached directly to the natural benthic substrate (non-live) using cable ties. A skirt of mesh surrounding the bottom of each cage prevented fishes from entering through gaps under the rim of the cage. Monitoring of all algal plots when released to grazing was conducted daily for up to 9 days. The actual biomass grazed upon was to be determined by taking small random samples of algae from within each of the replicate plots when the cages were opened, then at the end of the monitoring period. Digital images from above and from an angle of each of the plots were also taken on a daily basis. In order to appropriately compare relative algal cropping rates between sites,the background level of algae at each site was determined by analyzing 150 benthic photoquadrats per site. Dethier M.N., Graham E.S, Cohen S. and Tear L.M. 1993 Visual versus random-point percent cover estimations: objective is not always better. Marine Ecology Progress Series, vol.96, p.93-100 Foster M.S, Harold C. and Hardin D.D. 1991 Point vs. photo quadrat estimates of the cover of sessile marine organisms. Journal of Experimental Marine Biology and Ecology. Vol 146, P. 193-203 Hibino K. and van Woesik R. 2000 Spatial differences and seasonal changes of net carbonate accumulation on some coral reefs of the Ryukyu Islands, Japan, Journal of Experimental Marine Biology and Ecology, vol. 252 p. 1-14 Hubbard, D.K. 1992. Hurricane-induced sediment transport in open-shelf tropical systems: an example from St. Croix, U.S. Virgin Islands. J. Sedimentary Petrology 62: 946-960 Hubbard, D.K. 1997. Reefs as dynamic systems. Pages 43-67 In C. Birkeland (ed.), Life and death of coral reefs. Kluwer Academic Publishers, City, The Netherlands. 536 p. Hubbard, D.K., A.I. Miller, and D. Scaturo. 1990. Production and cycling of calcium carbonate in a shelf-edge reef system (St. Croix, U.S. Virgin Islands): applications to the nature of reef systems in the fossil record. J. Sedimentary Petrology 60: 335-360 Meese R.J. and Tomich P.A. 1992 Dots on rocks: a comparison pf percent cover estimation methods. Journal of Experimental Marine Biology and Ecology. Vol165 p 59-73
National Coastal Data Development Center, Building 1100
NOAA makes no warranty regarding these data,expressed or implied, nor does the fact of distribution constitute such a warranty. NOAA, NESDIS, NODC and NCDDC cannot assume liability for any damages caused by any errors or omissions in these data, nor as a result of the failure of these data to function on a particular system.
1000 Pope Road, MSB 316
Dept. of Oceanography
University of Hawaii at Manoa
check email@example.com if not available