Recruitment patterns and processes aod the connec{il'ity of rock)' shores in southern Africa

ABSTRACT This thesis addresses the recruitment patterns of barnacles and mussels at a range of spatial and temporal scales and attempts to relate them to larval and adult abundance, as well as to local productivity and oceanographic conditions that would influence larval transport and the connectivity of adult populations. In Chapter 1, I investigated the spatial and temporal patterns of mussel recruitment and adult bed dynamics offour mussel species (Aulacomya ater, Mytilus galloprovincialis, Perna perna and Semimytilus a/gosus) across 2500 km of the southern African coast. A coast-wise gradient of recruitment intensity was evident, with high recruitment rates on the West Coast and lower rates on the South and East Coasts. At the largest scale (100s ofkm), populations appear closed, as adult biomass was significantly and positively correlated with recruitment density. Although neither nearshore productivity nor meso-scale oceanic features were significantly correlated with among-location variability in recruitment intensity, an index that combined the two explained 62% of the variability, indicating that, taken in conjunction, both of these factors are important in understanding recruitment processes of benthic organisms with meroplanktonic larval stages. High inter-annual variability in recruitment was recorded, with low recruitment on the West Coast in years with Benguela Nifio or Nifio-like warming events. At most locations and for most species, sites were more variable in the density than in the timing of recruitment events. This study represents a first critical step in identifYing and comparing large and meso-scale recruitment patterns and processes in southern Africa, and created a framework for the detailed investigations in the following three chapters. Chapters 2-4 examine patterns and processes of the recruitment of mussels and barnacles along a 25-km stretch of shoreline in the Maputaland Marine Reserve (MMR) in northern KwaZulu-Natal (KZN), South Africa. The investigations in these chapters not only extend our knowledge of recruitment in South Africa into an under-studied bioregion, but also address marine reserve management-related questions of benthic-pelagic coupling and the connectivity of rocky shores in this important conservation area. Knowledge of the shape, patchiness and density of the meroplanktonic larval pool, and how these are controlled by local processes such as productivity and onshore transport mechanisms, is critical to interpret both recruitment rates and connectivity between areas. Chapter 2 examines the spatial and temporal patchiness of the pool ofmytilid, bivalve and cyprid meroplankton in the MMR. This is the first study of this kind in this region. The size and density of the patches were analyzed in relationship to the in situ biomass of zooplankton and phytoplankton, several physical parameters and the adult density of the mussel Perna perna on adjacent headlands. There is a strong sea-breeze effect and wind-driven upwelling, although weak, was present at this site; nearshore phytoplankton concentrations responded positively to decreases in water temperature. 1 Bivalves and mytilids were vertically unstratified, positively correlated with water temperature and negatively correlated with phytoplankton concentration, indicating that they may be closer to shore during periods of upwelling-neutral or downwelling-positive winds. Cyprids aggregated near the bottom of the water column and were negatively correlated with temperature and positively correlated with phytoplankton, indicating that upwelling may be an important onshore transport mechanism for cyprids. Larvae of barnacle and mussel groups were, on average, denser at the inshore than the offshore stations, but very few nauplii were found, indicating that competent larvae may be using behavioral or swimming adaptations to aggregate nearer the shore. Cyprids were not denser at rocky headland points, where the adults occur, than in the bays between. Mytilids were found in highest densities proximate to two rocky headlands, however, suggesting that both accumulating mechanisms and associations with adult biomass are important in explaining the spatial structure of the mytilid larval pool. Chapter 3 reports the recruitment rates of two mussels (Brachidontes semistriatus and Perna perna) and three barnacles (Chthamalus dentatus, Tetraclita serrata and T. squamosa rufotincta) onto natural and artificial substrates. These were recorded monthly for 18 months at each of five sites in the MMR, simultaneous to the collection of larval data presented in Chapter 2. Recruitment rates of mussel species in northern KZN were lower than those recorded anywhere else in southern Africa. Recruitment was strongly seasonal for all species, with B. semistriatus, P. perna and C. dentatus recruiting in spring/summer and Tetraclita spp. in autumn/winter. Locality, site and intertidal zone were less important than season in explaining recruitment variability, although locality was also often a significant factor for barnacles. Mean monthly temperature was generally the most important predictor for the recruitment variability of all species, which was expected due to the seasonal pattern of both water temperatures and recruitment in this region. Nearshore larval density was also a significant predictor for barnacles and for B. semistriatus, establishing benthicpelagic coupling. The relationship between recruitment rate and adult density was found to be weak and non-significant for P. perna, implying that the population may be open at the scale of the study area (approx. 25 km), although it may reflect harvesting pressure on adult stocks. To explore connectivity of the rocky shore communities, Chapter 4 investigates nearshore hydrography in the MMR. Dominant meso-scale patterns in SST were found to be more closely linked to offshore forcing at the confluence of currents joining to form the Agulhas Current than to local wind events. These dominant patterns were found to explain 10% and 15% of recruitment variability of the mussels P. perna and B. semistriatus, and 16% and 48% of the recruitment variability of the barnacles C. dentatus and Tetraclita spp., respectively. Analyses of the dominant current structures found this coast to be driven by coastal boundary, wave direction, wind and tidal forcing mechanisms. While the across- and alongshore currents were strongly sheared, the principal components of their patterns were not related to the patterns revealed by empirical orthogonal function analysis of SST. Alongshore velocity data and knowledge of across-shore transport and 2 processes were then combined to estimate larval dispersal distances for mussels and barnacles, as well as other organisms, within the MMR. A semi-stochastic Lagrangian simulation was constructed to predict dispersal kernels of passive drifters during the recruitment season for mussels and C. dentatus. In all cases, the dominant direction of travel was southward, and mean dispersal distances for the various models ranged from 30-122 km south along the coast. Using these results, together from those in the previous chapters, three alternate examples of the spacing of no-take areas using dispersal distances were proposed, with recommendations for improvement of the management of the rocky shores within the MMR.