Bayesian data-model synthesis for biological conservation and management in Antarctica

This paper introduces a proposed data integration and assimilation tool to assist the CCAMLR Ecosystem Monitoring Program in obtaining policy-relevant summaries of Adélie penguin abundance and distribution. The engine of this decision-support tool is a physically-based algorithm for retrieval of continent-wide Adélie penguin distribution and abundance from satellite remote sensing imagery. An ecologically-based Dynamic Bayesian Network (DBN) model assimilates remote sensing results streaming in from a multitude of sensors with other sources of information such as field counts and predictions from state-space models of population change. The DBN model synthesizes this data flow into policy ready metrics of Adélie penguin abundance at any user defined spatial or temporal scale. The results will route through a browser based geospatial application custom designed to address the needs and concerns of the Antarctic research and management community. In sum, we propose to develop the data-to-knowledge pipeline required to fully harness the power of remote sensing for effective resource management in the Antarctic. This paper serves as an introduction to our proposed development work (as recently submitted in response to NASA Research Announcement NNH12ZDA001N-ECOF) and a request for input on the design of the associated user interface.