Title: Investigation of statistical property in residual series after detrending
Authors: Hiroko Solvang - Institute of Marine Research (Norway) [presenting]
Abstract: During the last decade, global warming has led to changes in spatial distribution of plankton and fish in marine ecosystem. There is an urgent need for a holistic Ecosystem Based management of the ecosystem, taking the rapid climate driven changes in the natural systems. A simulation-based study for future scenario is necessary to evaluate climate change impact on biological ecosystem component, which especially affects fish stock dynamics under different management strategies. To generate realistic simulation data, the information extracted from time series observation recorded by annual survey should be involved. The data present long-run movements caused by temperature and short/middle cyclic terms in temporal changes in ecosystem. The stationary statistical property of the data is investigated by applying another statistical analysis to short/middle cyclic terms. We propose a systematic statistical procedure, integrating time series decomposition and robust statistical testing procedures. This procedure estimates the trend component corresponding to long-run movements and identifies the statistical property for the detrended component. Based on the simulation data, future scenarios of complex temperature variation and cyclic fluctuation may be predicted.