Modeling Fused Spatial Data for Improved Public Information on Air Quality
Project Leads and Affiliations:
David M. Holland and Jim Szykman, EPA/Office of Research and Development
Chet Wayland and Phil Dickerson, EPA/Office of Air and Radiation
Ken Schere, Rohit Mahur, and Dev Roy, EPA/Office of Research and Development
Ray Hoff and Allen Chu, JCET-University of MD-Baltimore County/Goddard Space Flight Center
Jassim Al-Saadi, NASA Langley Research Center
Shobha Kondragunta, NOAA/NESDIS/ORA
Project Abstract Written: October 2005
Project Timeframe: 2006 - 2008
This project is designed to build upon recent advances in the modeling of fused spatial information to provide state-of-the-science continuous predictive maps of current day and next day (forecasts) air quality patterns in the U.S. Our primary challenge is to accelerate the use of these novel models for massive databases to provide enriched air quality information to the U.S. public and environmental health decision-makers. Accurate air quality information can offer significant health benefits, particularly for people with respiratory diseases, by leading to improved environmental decisions. National air quality forecasts and near real-time predictive spatial maps will continually be provided to the public through EPA's AIRNow web site (http://airnow.gov/). Current and next day PM2.5 air quality forecasts for over 200 cities are now provided by AIRNow on a daily basis. Initially, spatial maps of daily maximum 8-hr O3 and daily average PM2.5 concentrations will be provided based on real-time monitoring data and existing Community Multi-Scale Air Quality (CMAQ) model output. As we improve the computational efficiency of our space-time models for fused data bases, we will shift our attention to:
- providing hourly updates of O3 and PM2.5 spatial surfaces; and
- forecasting next day spatial patterns using historic and next day CMAQ forecast data and historic air monitoring data. For PM2.5, we will incorporate satellite aerosol optical depth measurements into the fused spatial prediction model.
This multi-disciplinary effort that involves scientists from several government Agencies will proceed in the following manner:
- biweekly conference calls (and meeting when possible) will be planned to choose the appropriate databases, develop statistical models, and write statistical software to implement the models and produce interpolated maps of air pollution for display on the AIRNow website; and
- quarterly milestones will be developed to ensure scientific relevance and progress. The complexity of this effort will likely require addressing new modeling challenges as the research evolves over the course of this study.