Project: New statistical tools for analyzing community dynamics with applications to marine zooplankton

An extended multivariate auto-regressive (MAR) modeling framework will be developed to estimate dynamics from marine time-series data and the framework will be used to analyze long-term marine plankton data sets. MAR modeling has been used extensively for freshwater plankton communities to infer the inter-species interactions, the dominant environmental drivers and the system stability and resilience. MAR modeling is well-grounded on theory concerning population and community dynamics and comparative properties of communities, such as resistance to disturbance, resilience, and return time after disturbance, are easily calculated in terms of the stability properties of the matrix of species interaction strengths. The proposed research will address four technical barriers that hinder widespread application of the MAR framework to marine data sets - observation error, lower temporal autocorrelation due to open systems and infrequent sampling, multiple spatially-distributed sampling locations, and uncertainty introduced by unmeasured species or environmental drivers. To facilitate wider use of MAR-based analysis, a statistical package for estimating MAR models from multi-site data sets with observation errors will be developed in the R language and disseminated publically. The extended MAR framework will be used do a comparative study of marine plankton community dynamics from different geographic regions using existing long-term data sets. The primary goals are 1) to identify the major drivers of plankton productivity and any directional changes in dynamics due to long-term changes in ocean conditions and 2) to compare the community dynamics -- specifically interaction strengths and community stability - to four well-studied freshwater plankton systems.

The proposed work addresses several goals of the CAMEO program. First, it develops new ecosystem modeling tools, specifically tools based on autoregressive modeling of abundance time series data. Second, it is based on a modeling approach explicitly developed to analyze the stability and resilience properties of communities, and is one that is firmly based on the theory of dynamical systems. Third, the proposed research project charts a realistic path for translating research results into decision-support tools: it is designed to use the types of time-series data that are collected in current fisheries monitoring programs and use those data to estimate the impacts of physical, biological, and anthropogenic drivers on marine ecosystems.