Tuesday, July 12, 2016

Charles River Analytics, developer of intelligent systems solutions, announces a follow-on contract awarded by the Office of Naval Research (ONR) to develop a seasonal weather forecasting tool. The US Navy is seeking tools to generate forecasts of the probability that conditions are different from the long-term average climatology or mission-specific thresholds. Charles River Analytics is developing machine learning software that uses big data to support an extended-range weather prediction tool under the Climatological Observations for Maritime Prediction and Analysis Support Service effort, or COMPASS.

“In COMPASS, we’re using newly available extended-range forecasts to produce a unified and more accurate, long range, probabilistic forecast for seasonal and sub-seasonal periods. We are also accessing suitability of locations for specific missions in a location over a period of time,” explained Joe Gorman, Division Software Engineer at Charles River and Principal Investigator on COMPASS.

COMPASS will integrate a number of forecasting systems already in use, including the North American Multi-Model Ensemble (NMME). It will produce a single unified and improved forecast for a specific time period and location up to twelve-months in advance. Also, it will provide a mission-specific forecast display to visualize weather and climate predictions, representing the probabilistic nature and uncertainty of the predictions.

Charles River is partnering with Dr. Benjamin Kirtman of the University of Miami Rosenstiel School of Marine and Atmospheric Science for the COMPASS effort. Dr. Kirtman is the lead Principal Investigator on the National Oceanic and Atmospheric Administration’s (NOAA) NMME program. Charles River is also partnering with Mr. Bruce Ford (USN Retired) of Clear Science, Inc., to assist in integrating COMPASS into an existing forecasting system. Mr. Ford is the primary developer of the Advanced Climate Analysis and Forecast (ACAF) system.

Charles River will apply machine learning techniques and its Figaro probabilistic programming language to learn how to combine multiple models to better predict environmental conditions (e.g., temperature, wind stress, and cloud cover). Figaro is a free, open-source probabilistic programming language for probabilistic modeling.

“We will continue to develop the machine learning and inference approaches introduced during the initial effort to produce a single, improved forecast using Figaro,” added Mr. Gorman. “We will use Figaro’s built-in machine learning and inference capabilities to produce a coherent unified forecast from individual time, location predictions.”

This material is based upon work supported by the US Navy under Contract No. N00014-15-P-1067. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the US Navy.