Use of sensitivity and comparative analyses in constructing plausible trophic mass-balance models of a data-limited marine ecosystem - The KwaZulu-Natal Bight, South Africa

Ecosystem modelling allows for an understanding of the structure and functioning of data-limited ecosystems provided that models undergo extensive sensitivity analyses to explore the levels of uncertainty. We explored one such data-limited system, the KwaZulu-Natal (KZN) Bight, a river-influenced bight on the east coast of South Africa. Potential system states of the KZN Bight were created by constructing multiple models in Ecopath with Ecosim, carrying out sensitivity analyses and comparing outputs. Sensitivity analyses showed that models were most sensitive to apex predator parameters and a comparison of outputs showed the important influence of riverine detritus on system functioning. To demonstrate the KZN Bight models could reproduce known differences to other ecosystems a comparison of the nutrient-poor KZN Bight to the nutrient-rich Southern Benguela was carried out. This confirmed that the KZN Bight was considerably smaller in biomass, productivity and fishery landings than the Southern Benguela with the systems being detritus-driven and phytoplankton-driven respectively. The KZN Bight relied on large detritus imports from rivers and had higher cycling through the system. The reliance on detritus import from rivers has riverine and fishery management implications as a decrease in riverine detritus caused a decrease in biomass of biotic groups, including targeted fishery species. (C) 2011 Elsevier B.V. All rights reserved.