| Summary: | Marine ranching has been identified as an alternative to traditional aquacultural rearing and growing organisms for consumption. In the Eastern Cape, abalone ranching is a new and experimental industry. The aims of the research were to: first develop a GIS model to assist management in site selection for abalone seeding; and secondly to develop and standardize the sampling methodology in order to ground truth the sites, and assist in the monitoring and habitat identification of abalone. The GIS model developed in Chapter 3 was created using an unsupervised classification and fuzzy logic approach. Both vector and raster datasets were utilized to represent 7 different layers. Predominantly satellite imagery was used to classify the different substrate groups according to pixel colour signatures. The basic process was to apply a fuzzy rule set (membership) to rasters which gave an output raster (Fuzzification). The membership output rasters were overlaid which creates a single model output. It was found that model accuracy increased significantly as more layers were overlaid, due to the high variability within each of the individual layers. Model ground-truthing showed a strong and significant correlation (r2 = 0.91; p < 0.001) between the model outputs and actual site suitability based on in situ evaluation. Chapter 4 describes the investigation towards the optimal sampling methods for abalone ranching habitat assessments. Both destructive sampling methods and imagery methods were considered as methods of data collection. The study also evaluated whether quadrat and transects were going to be suitable methods to assess sites, and what size or length respectively they should be to collect the appropriate data. Transect length showed great variation according to the factor assessed. A transect of 15 metres was found to be optimal. Abalone counts showed no significant (p = 0.1) change in the Coefficent of Variance (CV) for transect lengths greater than 15m, and had a mean of 0.2 abalone per metre. Quadrat size showed a significant difference in functional group richness between quadrat sizes of 0.0625m2, and 0.25m2 but no difference between 0.25m2 and 1m2 quadrats for both scape and photographic quadrats. It was also found that between 5 and 10 replicates (p = 0.08) represents the functional groups appropriately using quadrats and that a 0.25m2 quadrat is most suitable for sampling. Chapter 5 describes the benthic community structure of Cape Recife shallow water reefs. Using the standardized methodology previously mentioned, 45 sites were assessed to identify the community structure. These sites were grouped into 5 different groups influenced by depth and substrate, as well as functional group composition according to a Wards classification. The community structure showed that depth and substrate play a significant role (p < 0.05) in the community type. There is also a significant relationship (p < 0.05) between complexity, rugosity, abalone presence and substrate. During this study the basic protocols for site selection and benthic community monitoring have been developed to support the abalone ranching initiative in the Cape Recife area. It has also provided a baseline of the benthic community in the ranching concession area which will be used as a benchmark for future monitoring efforts. The site selection, sampling, and monitoring methods developed during the course of this work have now been rolled out as Standard Operating Procedures for the ranching programme in this area.
|
|---|
