This study formed part of a larger provincial marine systematic conservation plan for KwaZulu-Natal (KZN), South Africa, called SeaPLAN. Owing to budget and time constraints, not all ± 1640 fish species that occur in the region were considered. A method to prioritise species was therefore developed to identify those species which were most at most risk of being excluded by a conservation plan based primarily on habitat representation (i.e. SeaPLAN). The method was based on three underlying principles: (i) species with limited conservation options; (ii) threatened species; and (iii) inherently vulnerable species. From these three principles, seven criteria were defined (e.g. endemic or rare species). Sixtyseven species met the qualifying conditions for these criteria and were consequently included in this study (FishPLAN). In order to map the distributions of these 67 fish species, the spatial and temporal accuracy of existing marine fish data for KZN was investigated. Only 17 percent of the data evaluated met the spatial resolution requirements of 1 km2, while temporal resolution was high: >99 percent of the data were collected at daily resolution. A resulting recommendation is that future data collection employ handheld data recording devices (with GPS capability), in order to increase the spatial accuracy of data, minimise human error and improve the efficiency of data flow. Species life cycle envelopes (SLICES) were developed to capture spatial differences in areas occupied during three life-cycle phases (reproductive, juvenile and feeding). Two distribution modelling techniques were used: Maxent, which uses quantitative data, and CHARMS (cartographic habitat association range models), which uses qualitative range data. A combination of statistical and biological criteria was used to determine the most informative and appropriate model for each species. Species distribution models (SDMs) were constructed for three temporal partitions of the data: annual, summer and winter. Patterns of species richness developed from the seasonal models showed seasonal differences in patterns that conformed to known seasonal distributions of fish assemblages: richness was higher in southern KZN during winter, while it was higher in northern KZN during summer. The resulting SDMs were used to develop a conservation plan for fish: conservation targets were set using the minimum recommended baseline of 20 percent of a species’ range, to which biological retention targets (additional proportion of the range) were added, in an attempt to ensure species persistence. The conservation targets were then adjusted using catch per unit effort (CPUE) data to match seasonal abundance of a given species. Within the existing network of marine protected areas (MPAs), none of the species’ targets are met by MPA sanctuary zones (zone As) alone, and all species require greater areas of protection. Three areas, namely offshore of the Tugela River mouth, the reefs offshore of Durban, and Aliwal Shoal, were consistently identified as being important in addition to existing MPAs for conservation of the fish species investigated. The greater efficiency of a seasonal MPA network to protect seasonally varying distributions of biodiversity, suggests that this may be a useful tool to consider in conservation management. The outcome of a conservation plan from this study (FishPLAN) was finally compared with the broader, more inclusive conservation plan, SeaPLAN. This comparison demonstrated how conservation plans based on a single group of species run the risk of identifying areas that are appropriate only for the relevant species, and might fail to conserve biodiversity as a whole.

Although the marine environment has generally been considered to be less in need of attention from conservationists, politicians and managers than the terrestrial environment, there has been an increasing appreciation that there is little left in the marine environment that can be considered unaltered by humans. Conservation action is therefore urgently needed to protect marine populations, species and environments from anthropogenic pressures. Since overharvesting seems to be the main cause of the collapse in coastal ecosystems, areas need to be protected so that stocks can recover. However, the proclamation of such marine protected areas (MPAs) need to be well planned to ensure that species are well represented and that their biological requirements are met in the long term. While terrestrial conservation planners have started to incorporate biological, social and economic information into so-called systematic conservation plans, marine planners have been more hesitant to use these approaches due to the greater openness of the ocean (with the associated greater dispersal of marine organisms and propogules, lower levels of endemism, and lack of ownership) and the paucity of available information. The current study aimed to investigate the applicability of systematic approaches to the process of selecting sites for a network of potential replenishment zones for seabream species that could supplement protection already afforded by existing no-fishing reserves. Since many of these seabream species are popular targets in both the recreational and commercial linefishery sectors, they have been overexploited and are in need of additional protection. Information on these species and the diversity of their associated reefs were obtained through consultations with scientists, managers and resource-users (namely SCUBA divers, spearfishers as well as recreational and commercial fishers). These data were obtained during workshops where Participatory Rural Appraisal (PRA) scoring and mapping methods were applied. Although ordinal-scale, rather than interval-or ratio-scale data were obtained, these data could be used in an interactive selection process. Because these methods require relatively cheap materials and few personnel, they could be used to collect information where limited budgets are available. Quantitative targets for the protection of individual species were set on the basis of their status in the Marine Living Resource Act, since this reflected their different requirements for protection. A preliminary investigation indicated that the current no-fishing reserves effectively achieved most of these goals for seabream species. The selection of sites for the replenishment network was based on irreplaceability analyses, while decisions between sites with lower irreplaceability values were based on explicit design rules. A total of 26 sites were required to achieve targets for all species. These replenishment zones could potentially be established, managed and monitored by voluntary community groups. This study therefore indicated that systematic conservation planning can successfully be applied to the marine environment, while the involvement of stakeholders is crucial to ensure that conservation action is supported.

Freshwater fish are the most imperilled vertebrate group with a high projected extinction rate. In general, the world's freshwater ichthyofauna is in a poor state and is threatened by habitat destruction, pollution, the effects of alien species, damming, water extraction and overfishing. Within South Africa, legislation supporting the conservation of aquatic habitats and its associated fauna is strong, but there is a lack of capacity and poor co-ordination between agencies responsible for the conservation and management. The Greater Addo Elephant National Park (GAENP) includes the upper catchments of five of the Sundays River tributaries. Since the optimal method of conserving freshwater aquatic biota is to conserve the headwaters of either the river or its tributaries, there is an opportunity to conserve the native ichthyofauna of the Sundays River system. The principal objective of this thesis was, therefore, to provide information pertinent to the conservation and management of the fishes in the Sundays River system. Sixteen species of freshwater and estuarine fish were sampled, of which eight were alien. Species composition in each of the rivers, with the exception of the Kabouga River, had changed significantly from previous studies. The distribution ranges of several indigenous species had reduced, while those of some alien species was increased. Two species not previously recorded, Labeo capensis and Gambusia affinis, were sampled. Alien species were sampled from all the tributaries. Diversity indices and community assemblage models were identified as approaches for monitoring the ichthyofauna in this species depauperate system. Diversity indices were not recommended for use as a monitoring tool, as they provided insufficient detail on community composition and structure. A modified depletion estimator was applied to multiple pass electrofishing data to determine the effects of various physico-chemical variables on the absolute abundance of Pseudobarbus afer, an endangered species. Pseudobarbus afer density was found to be positively correlated to the presence of boulders. Where P. afer were present, they were abundant. Of the total sampled catch, 21.8% was comprised of alien fishes. Clarias gariepinus was the most successful invader having penetrated all but one of the tributaries. Micropterus salmoides changed the species composition of the Wit River significantly through predation, and is potentially the most dangerous of the alien species in the Sundays River system, capable of extirpating a population of P. afer in as little as three days. Only one man-made barrier to the upstream migration of alien fishes was present on the tributaries, and several addition barriers are required to safeguard the populations of indigenous fish in the headwaters of these tributaries. Several management and conservation tools are presented and discussed, including river rehabilitation, translocation, eradication of alien species, erection of barrier weirs and captive breeding programmes. Public awareness and education is stressed because without support from the communities surrounding the GAENP, management plans for conservation of the indigenous ichthyofauna will, in all likelihood, fail.