A decision tree framework for assessing status of exploited marine ecosystems under changing environmental conditions

• Main Author: Lockerbie, Emma Margaret
• Other Authors: Shannon, Lynne J
• Format: Doctoral Thesis
• Language: English
• Published: University of Cape Town 2018
• Subjects: Marine Ecosystems
• Online Access: http://hdl.handle.net/11427/27996

The removal of marine species through fishing has impacted marine ecosystems for thousands of years. The pressure of exploitation on marine ecosystems has now reached a point at which there is serious concern over ecosystem well-being on a global scale. There has, therefore, been a global move towards an ecosystem approach to fisheries management. The objective of this study was to develop a decision tree framework to assess the status of exploited marine ecosystems, which could be successfully applied to numerous ecosystems and guide decision support under changing conditions. This work was based on that of the IndiSeas project, which makes use of indicators designed to detect the impacts of fishing on marine ecosystem around the world. A suite of indicators, selected from those utilised in the IndiSeas project, was divided into ecological and fishing pressure indicators. Ecosystem specific suites of environmental indicators were also included, allowing the framework to ascertain the impacts of environmental variability on ecosystem components. This is an important addition as currently many assessments of the impacts of fisheries do not account for the influence of the environment. The framework was developed for the Southern Benguela ecosystem and then applied, with minor adjustments to account for ecosystem-specific characteristics, to the South Catalan Sea and North Sea. Indicator time series were analysed making use of linear regressions, resulting in the assignment of a score between one and five, depending on the direction and significance of trends. Data series were divided into distinct periods based on known environmental changes or shifts within ecosystems. Careful consideration was given as to whether fishing and environmental indicator trends could explain the observed trends in ecological indicators. A method of score adjustment was then developed to account for the impacts of both fishing and environmental variability on ecological indicators. Correlations were conducted to detect potential redundancies of ecological indicators and weightings were applied to decrease the contribution of correlated indicators to overall ecosystem trends. However, as correlations differed between indicators and amongst ecosystems, it was necessary to adjust the applied weightings for individual ecosystems. Results for the Southern Benguela classified the ecosystem as neither improving nor deteriorating during Period 1 (1978-1993) and Period 2 (1994-2003). During Period 3 (2004-2010) the ecosystem was classified as possibly improving. The South Catalan Sea was classified as possibly deteriorating during Period 1 (1978-1990) and neither improving nor deteriorating during Period 2 (1991-2010). The North Sea ecosystem was classified as neither improving nor deteriorating during Period 1 (1983-1992). During the second (1993-2003) and third (2004-2010) periods the ecosystem was categorised as possibly improving. When assessing fisheries impacts at an ecosystem scale there are typically high levels of uncertainty. However, this thesisoncluded that the development of a scoring and weighting system, alongside the addition of environmental drivers and the inclusion of expert knowledge throughout the applications of this framework, has allowed the developed decision tree framework to successfully categorise the three ecosystems. It is anticipated that the knowledge that this framework will add to current methods of generating advice for fisheries management will aid future decision support within these ecosystems.