Patterns of movement of three species of coral reef fish on the Great Barrier Reef

Davies, Campbell Robert (1995) Patterns of movement of three species of coral reef fish on the Great Barrier Reef. PhD thesis, James Cook University.

PDF (Thesis) - Submitted Version
Download (4MB)


The patterns of movement of three species of coral reef fish were investigated in two tagging studies on the Great Barrier Reef. In the first study, done within the lagoon at Lizard Island on the northern GBR, the frequency of movement of Lutjanus carponotatus, Plectropomus leopardus and Siganus doliatus within and among sites and three habitat categories were examined in a multiple capture-recapture fish trapping study spanning a period of 22 months. Rates of growth, mortality and tag loss were estimated from the capture-recapture data also. The second study was a large-scale tag-recovery program designed to estimate the extent of movement of Plectropomus leopardus within and among five individual coral reefs in the Cairns Section of the Great Barrier Reef Marine Park. The rates of loss of t-bar anchor tags and dart tags were compared as was the frequency of loss of different coloured t-bar anchor tags.

A total of 4,736 fish from 21 families and 109 species were trapped over the duration of the small-scale movement study in the Lizard Island lagoon. The catch was dominated by the Siganidae, Lutjanidae, Lethrinidae, Serranidae, Haemulidae and Acanthuridae, which collectively comprised over 88% of the catch at each site. Siganus doliatus and Lutjanus carponotatus were the two most common species and together accounted for 36 % of the total catch. P.leopardus was less common accounting for less than 5% of the catch at each site. Catch per unit effort (CPUE) of P.leopardus, L.carponotatus and S.doliatus varied among the three sites. However, the CPUE of each species was highest in the reef habitat and lowest in the sand habitat at all sites. Fish trappping proved to be an effective, but selective, technique to simultaneously sample these species of coral reef fish at many locations and across a variety of habitats with limited logistical support.

The patterns of movement of the three species within the Lizard Island lagoon were found to differ considerably. P. leopardus regularly moved among trapping positions and across habitat types while the movements of L.carponotatus and S. doliatus were considerably more restricted. The majority of individuals of L.carponotatus (68%) and S.doliatus (69%) were recaptured at the position of release while the majority (66%) of P. leopardus recaptures had moved among trapping positions. L. carponotatus exhibited a strong fidelity for the habitat of release. The frequency of movement among habitat categories for fish released in the reef habitat was considerably (although marginally significant) lower than for fish released in the patch or sand habitat categories. S. doliatus was found to show a strong fidelity for the habitat of release also.

However, the frequency of movement among the reef and patch reef habitats was higher than observed for L.carponotatus. This may be related to foraging patterns of siganids which feed over the reef flat at high tide. The results indicate that the movements of L.carponotatus and S. doliatus are generally restricted to less than a few hundred meters with the majority of fish not moving from their position of release. The frequency of movement of L.carponotatus and S.doliatus across reef-sand habitat boundaries was considerably lower than the frequency of movement across reef-patch reef or patch reef-sand habitat boundaries. This suggests that reef-sand habitat boundaries may represent less permeable management boundaries than arbitrary boundaries located within continuous sections of reef mosaic.

Estimates of survivorship for each species were made from the multiple capture-recapture data using program RELEASE version 2.6. The estimates of survivorship for L. carponotatus and S. doliatus for the reef and patch reef habitats suggested that survivorship was higher (but not significantly at α = 0.05) in the reef habitat than in the patch reef habitat. There were insufficient data to examine the effect of habitat on survivorship for P. leopardus. Estimated survivorship also appeared to vary among sampling periods. The estimates of survivorship were adjusted for tag loss and converted to estimates of annual rate of natural mortality. These were very high for each species, in comparison to estimates available in the literature, which suggest that there may have been a significant effect of capture and tagging. The results suggest that survivorship may vary among habitats and over time. Consequently, the common assumption of constant mortality within and among populations and over time requires greater scrutiny.

Rates of growth for P. leopardus, L. carponotatus and S. doliatus were estimated using growth increment data from the small-scale tagging study. Estimated von Bertalanffy growth equation parameters, L(∞) and K, were 576 mm and 0.21 for P. leopardus, 357 mm and 0.12 for L. carponotatus and 201 mm and 0.71 for S. doliatus. There was evidence of high individual growth variability for each species. This, combined with the lack of data for individuals in the lower end of the size range of each species, suggests that the estimates of L(∞) are likely to be positively biased, and hence the estimated K negatively biased. The age-based parameter estimates for L. carponotatus obtained from the same location suggested this was the case, with L(∞) and K estimated to be 312 mm and 0.31, respectively. There was considerable variation in size at age for L. calponotatus, again suggesting that there may be significant variation in growth among individuals: This highlights the need for age-based estimates of population parameters for coral reef fish as length is likely to be a poor proxy for age. The relationship between estimates of age from readings of sectioned and whole otoliths and otolith weight were examined for L. carponotatus. Readings of whole and sectioned otoliths were the same up until age 5-6 after which readings of whole otoliths tended to underestimate age relative to readings of sectioned otoliths. There was a high correlation (r² = 0.94) between otolith weight and age from sectioned otoliths of L. carponotatus, indicating otolith weight may be an objective and cost-effective alternative for obtaining age-based estimates of population parameters for some coral reef fish.

All P. leopardus and L. carponotatus were double tagged in both studies. The Bayliff and Morbrand model was used to estimate rates of tag loss. Type I tag loss was not significant for L. carponotatus. However, the instantaneous rate of loss of t-bar anchor tags was high (0.0034, ±95% CI = ±0.0021). Type I tag loss was significant for t-bar anchor tags for P. leopardus. The estimate of the proportion of tags remaining following type I tag loss for P. leopardus was 0.8927 (95% CI = 0.8140-0.9791), while L, the instantaneous rate of tag loss was 0.0010 (±95% CI = ±0.0005). The estimated proportion of t-bar anchor tags lost annually were 72% and 60% for L. carponotatus and P. leopardus, respectively. This clearly demonstrates that tag loss rates can be substantial and assuming they are negligible will result in seriously biased parameter estimates. Dart tags were found to be shed at a significantly greater frequency than t-bar anchor tags (Likelihood X²(0.05,1.) = 10.678; p < 0.005) suggesting that t-bar anchor tags are the less effective of the two tag types used. The colour of the t-bar anchor tags used didn't significantly effect their frequency of loss (Likelihood X²(0.05,5,243) = 1.902; p < 0.8625). This demonstrates that different colours of tag may be used to batch code releases of reef fish without incurring differential frequencies of tag loss due to the colour of the tags.

A total of 8,043 fish were caught from the five reefs over five trips of the large-scale tagging study. Catch was dominated by Serranidae, Lutjanidae and Lethrinidae which comprised greater than 97% of the total catch. The species composition was dominated by six species, Plectropomus leopardus (57%), Cephalopholis cyanostigma (12%), Lutjanus carponotatus (6%), L.bohar (3%), Lethrinus miniatus (3%) and L. atkinsoni (4%). The contribution of these six dominant species to the catch varied significantly among trips and reefs. Plectropomus leopardus comprised a greater proportion of the catch on the trips done during the spawning season (September 1992 and October 1993). This may indicate an increase in the catchability of P. leopardus during the spawning season. The difference among reefs was mainly due to the higher proportion of Cephalopholis cyanostigma and Lutjanus bohar and the lower proportion of Lethrinus miniatus and Lutjanus carponotatus at two reefs compared to the other three reefs studied. Catch Per Unit Effort of P. leopardus varied significantly among trips and within reefs. However, there was no significant difference in CPUE among reefs. The pattern of CPUE among trips and within reefs indicated that the observed increase in CPUE that occurs during the spawning season is likely to be the result of an increase in the catchability of P. leopardus when the fish are aggregated to spawn.

The average size (mean length to caudal fork) of P. leopardus decreased significantly over the five trips, with a monotonic reduction in average size from April 1992 to February 1994. Mean size of P. leopardus varied significantly among reefs and blocks (1.5-2.5 km strip of reef perimeter) also, with Taylor Reef having a significantly greater average size than the other reefs and Beaver Reef having a significantly smaller average size than all other reefs. Although the overall reduction in mean size of P. leopardus across all reefs is indicative of fishing and cause for concern, in the absence of size-at-age information it is not possible to accurately interpret these effects in terms of differences in the population dynamics of P. leopardus among reef or over time. The significant effect of block on mean length of P. leopardus suggests that there may be significant differences in either age-structure or growth rates within reefs also. These results highlight the need for rigorous and powerful sampling programmes, which include within reef strata, for monitoring changes in relative abundance and size and age-structure- of exploited populations of coral reef fishes.

A total of 4,627 P. leopardus were tagged and released on the five reefs with a total of 443 recaptured; 300 from the public and 143 from the four tag-recovery exercises. Ninety-nine percent of the research returns of P. leopardus were returned from the reef of release. One inter-reef movement was recorded from Taylor to Beaver Reef. These results indicate that the extent of inter-reef movement was negligible. In contrast, 36% of the public returns were returned from reefs other than the one on which they were released. The majority of inter-reef movement from the public returns was from Beaver (Closed) to Taylor reefs and from Potter Reef to other reefs in the cluster. The disparity in the extent of inter-reef movement of P. leopardus from Beaver Reef (Closed to fishing) between the public and research returns appears to be the result of infringement and misreporting of location of capture by the public. It is suggested that the level of fishing effort on Beaver Reef (Closed to fishing) indicated by the tag returns may be sufficient to negate the potential effects of protection from fishing. This was supported circumstantially by the CPUE and length frequency data for P. leopardus. In contrast to the negligible level of inter-reef movement by P. leopardus, there was considerable movement within reefs. On average 35% of the P. leopardus returned had moved out of the 1.5-2.5 km block in which they were released. The extent of movement varied among reefs and appeared to be related to movement to, or from, spawning aggregations. The results of the large-scale movement study suggest that partial reef closures may not effectively protect the populations of more mobile reef fish such as P. leopardus, due to their relatively high frequency of movement within reefs, and that it would be more effective to use individual reefs as the minimum spatial unit for reserve design.

Item ID: 27169
Item Type: Thesis (PhD)
Keywords: Plectropomus leopardus; Cephalopholis cyanostigma; Lutjanus carponotatus; L. bohar; Lethrinus miniatus; L. atkinsoni; Siganus doliatus; Lizard island; Cairns section; Taylor Reef; Beaver Reef; Potter Reef; Great Barrier Reef; GBR; t-bar anchor tags; dart tags; tag loss; site fidelity; habitat fidelity; foraging patterns; survivorship; growth estimates; age estimates; otolith weight; catchability; spawning aggregations; reserve design
Date Deposited: 03 Jun 2013 02:22
FoR Codes: 05 ENVIRONMENTAL SCIENCES > 0502 Environmental Science and Management > 050202 Conservation and Biodiversity @ 33%
06 BIOLOGICAL SCIENCES > 0608 Zoology > 060801 Animal Behaviour @ 33%
05 ENVIRONMENTAL SCIENCES > 0502 Environmental Science and Management > 050205 Environmental Management @ 34%
SEO Codes: 96 ENVIRONMENT > 9605 Ecosystem Assessment and Management > 960507 Ecosystem Assessment and Management of Marine Environments @ 34%
97 EXPANDING KNOWLEDGE > 970105 Expanding Knowledge in the Environmental Sciences @ 33%
96 ENVIRONMENT > 9608 Flora, Fauna and Biodiversity > 960808 Marine Flora, Fauna and Biodiversity @ 33%
Downloads: Total: 382
Last 12 Months: 19
More Statistics

Actions (Repository Staff Only)

Item Control Page Item Control Page