The use of a Roving Creel Survey to monitor exploited coastal fish species in the Goukamma Marine Ptrotected Area, South Africa

A fishery-dependant monitoring method of the recreational shore-based fishery was undertaken in the Goukamma Marine Protected Area (MPA) on the south coast of South Africa for a period of 17 months. The method used was a roving creel survey (RCS), with dates, times and starting locations chosen by stratified random sampling. The MPA was divided into two sections, Buffalo Bay and Groenvlei, and all anglers encountered were interviewed. Catch and effort data were collected and catch per unit effort (CPUE) was calculated from this. The spatial distribution of anglers was also mapped. A generalized linear model (GLM) was fitted to the effort data to determine the effects of month and day type on the variability of effort in each section. Fitted values showed that effort was significantly higher on weekends than on week days, in both sections. A total average of 3662 anglers fishing 21 428 hours annually is estimated within the reserve with a mean trip length of 5.85 hours. Angler numbers were higher per unit coastline length in Buffalo Bay than Groenvlei, but fishing effort (angler hours) was higher in Groenvlei. Density distributions showed that anglers were clumped in easily accessible areas and that they favored rocky areas and mixed shores over sandy shores. Catch documented between October 2008 and December 2009 included a total of 361 fish, of 27 species from 12 families. Sparidae had the highest contribution (12 species). A Shannon-Weiner diversity index showed that diversity was higher in Buffalo Bay (0.81) than Groenvlei (0.57). Catch composition of retained fish (336 individuals) showed that the six numerically most important species were blacktail (Diplodus sargus capensis) (66 percent of catch), followed by galjoen (Dichistius capensis) at 11 percent, Cape stumpnose (Rhabdosargus holubi), belman (Umbrina robinsonii) and strepie at 3 percent, and elf (Pomatomus saltatrix) at 2 percent. Catch composition of an earlier study in Goukamma (Pradervand and Hiseman 2006) was compared with the present study, as well as data from the De Hoop MPA, which is closed to fishing. A multi-dimensional scaling plot of catch composition showed tight clustering of the De Hoop samples, and high variability among the Goukamma samples. A bray-curtis similarity index and dendrogram of similarity between study sites and study periods showed that there was an 83 percent similarity among De Hoop samples and a 75 percent similarity among Goukamma samples (ignoring the two outliers). The two sites are different with respect to species composition, but this is expected because they are different areas. Differences between time periods in Goukamma (i.e. the previous study versus the present study) were not significant. The most significant result from the catch composition analyses is the high variability among the Goukamma samples. This can be explained by the variable fishing methods used by anglers in Goukamma, compared with the standardized fishing methods used by researchers in De Hoop, and the fact that fish are more abundant and populations are more stable in De Hoop – giving higher sample sizes which reduce the variability in the statistics. Species-specific CPUE was calculated for the six numerically most important species. In both sections, CPUE was highest for blacktail, with an average of 0.133 fish per hour for Groenvlei, and 0.060 fish per hour for Buffalo Bay, over the 12 months. The second highest CPUE values per section were 0.030 for galjoen in Groenvlei and 0.039 for strepie in Buffalo Bay. Remaining CPUE values ranged from 0.014 (belman in Groenvlei) to the lowest value of 0.001 (strepie in Groenvlei). Total estimated CPUE for these six species in the MPA using the estimated effort and catch results amounted to 0.018 fish per hour. An annual estimated 3897 fish were landed in the reserve during 2009. Most fish (n=2481, 64 percent) were caught in the Groenvlei section. Numbers of blacktail were the highest of all species, within both sections (2353 fish). Strepie was the next most common (561 fish), but was caught almost entirely within the Buffalo Bay section (97 percent of individuals), followed by galjoen (548 fish) caught mostly within the Groenvlei section (92 percent of individuals). Size comparisons of the six species between the Goukamma and De Hoop MPAs showed that ranges in size are similar, but there are substantial differences in mean sizes between the two MPAs. Sample sizes of all species from the Goukamma MPA were too small to draw conclusions about stock status, except for blacktail. The Goukamma MPA is a popular fishing destination and angler effort is high. It can be considered a node of exploitation for surf zone fish, for which it provides no protection. Even though the MPA allows shore angling, sustainable fishing practices should be incorporated in management plans if the MPA is expected to protect and conserve its stocks. Of noteworthy concern is the occurrence of illegal night fishing (the public may not enter the reserve between sunrise and sunset) which leads to underestimates of catch and effort (night surveys were not conducted because of safety concerns). It is recommended that more communication should take place between the angling community and the reserve management. Sign boards giving information on species which are under pressure, and why they are under pressure, with a short explanation on their life cycles, is advised. The roving creel survey method was suitable for the study area and delivered statistically rigorous results. I thus recommend that it is continued in the future by management. I make some recommendations for reducing costs of future surveys, as well as for altering the survey design if funds are very limited