Refining Suitability Modelling for Sea Cucumber (Holothuria scabra) Using Fully Raster-Based Data

Abstract

Geographical Information System (GIS) modelling using vector data is a commonly used method of modelling offering simple data input and analysis. However, the vector-data model assumes homogeneity in mapping units based on subjectively applied classification and simplification, and this may lead to over-simplification and consequent reduction in the variety of information obtained and uncertainty in results. This research aimed at refining the suitability modelling for sea cucumber (Holothuria scabra) using fully raster-based data for the waters of Kiowa Bay, Kahyapu village in the district of Enggano, North Bengkulu, Indonesia. Using a GIS, all parameters affecting suitability for sea cucumber were rasterised to improve compatibility. The relevant data includes nine parameters of sea water namely acidity, depth, current velocity, temperature, salinity, brightness, dissolved oxygen concentration, condition of the sea floor, and coastal protection of the area. These parameters were surveyed in the field at 51 stations and each parameter was then digitized and interpolated (using Kriging method) to create a continuous raster-dataset. Correlation analysis was then conducted to check parameter correlation. Parameters with a correlation coefficient of > 0.75 were excluded from further analysis since results could be derived from the remaining parameter set. Principal component analysis (PCA) was then applied to ascertain the weight of each component. Furthermore, scree plotting was employed to choose which principal components were relevant for insertion into the formula of suitability. The final result was then compared to the map of suitability from the analysis of vector-based data as the reference data set. The research results showed that this method can be used to locate areas that are suitable for sea cucumber farming. The suitability map for sea cucumber generated from the analysis using fully raster-based data displayed less uncertainty than the suitability map generated using vector-based data.