The term ecoforecasting as used on this Web site refers to the process of examining multivariate environmental data (physical, chemical, biological) from many sources (in situ observation and sensors, remote sensing, models and reanalyses, etc.), in order to forecast or "now-cast" a response to those environmental stimuli within some component of the aquatic ecosystem. An ecoforecast may also refer to the model or mechnism by which such ecoforecast products - herein referred to as ecoforecast alerts - are produced. Within the Coral Health and Monitoring Program at NOAA AOML, a variety of analysis, classificiation, and automated assessment techniques are applied in our ecoforecast models, including but not limited to expert systems (fuzzy-logic models based on semantic mapping of environment data, coupled with "if-then" rules), Artificial Neural Networks, and a variety of statistical time series analysis methods. A good summary reference for the CHAMP approach to ecoforecasting is Hendee et al., "Ecological forecasting for coral reef ecosystems" (2009).

This data-flow diagram summarizes the process by which integrated environmental data are used in ecoforecasting at CHAMP:

ICON Cloud for ICRC 2008