2012-08-012024-05-18https://scholars.lib.ntu.edu.tw/handle/123456789/698239Abstract: Incorporation of ecosystem knowledge into the management of ocean resources comes from a growing recognition that management based on classical single-stock assessments is insufficient. However, how to operationally implement ecosystem-based management remain elusive. Our objectives are 1) to develop new methods to provide fishery science with fundamental predictive understanding of ecosystem dynamics and, 2) to provide tools to identify objectively “which main variables to measure” in order to achieve a reasonable forecasting skill and thus achieve ecosystem-based management. First, we develop a global database by compiling data, including catch, CPUE, spawning biomass, recruitment, age (size) structure, and fishing mortality from existing stock assessment reports and literatures. Secondly, we link these data to environmental variables. Thirdly, using nonlinear methods, we determine dimensionality (system complexity), nonlinearity (vulnerability to regime shift driven by climate change), and predictivity (uncertainty) of each time series. Finally, we use multivariate embedding to obtain mechanistic understandings and forecasting dynamics of exploited fishes. Such information allows us to inversely construct a dynamics model based on data and investigate the system behavior. In addition, our prescriptive approach allows one to investigate potential future effects of fishing efforts and climate changes. This predictive and fundamental organizational information combined with subsequent scenario exploration models will serve as a basis for scientific advice for policymakers. Goals: Identify objectively key variables (including biological and environmental variables) that determine the dynamics of target species and provide mechanistic prediction for stock fluctuation that is critical in management. Activities: 1. Construct a global fisheries database, including catch, CPUE, spawning biomass, recruitment, age (size) structure, and fishing mortality from existing stock assessment reports and literatures, and link these data to environmental variables. 2. Investigate the dimensionality (system complexity) and nonlinearity (vulnerability to regime shift) of the target species. 3. Use multivariate embedding to identify objectively key variables that determine the dynamics of target species. 4. Provide proscriptive short-term forecasts as a function of fishing effort and environmental changes. Outcomes: 1. We provide a protocol to identify objectively “which main variables to measure” in order to achieve a reasonable forecasting skill, and thus achieve ecosystem-based management. 2. Our prescriptive approach allows one to investigate the potential future effects of fishing efforts and climate changes.