Table 1.

Features of the two types of air quality data being proposed

Factor	Ambient monitor data	Statistically combined data
Timeliness of data	3–6 months after the monitor year	Each year of statistically combined data will be available within 2 years of the model year. There is a large computational and technical burden to producing the CMAQ estimates. Then it requires 3–4 months to compute and check the statistical predictions, and statistical expertise to ensure that proper modeling assumptions and procedures have been used, and the results are reasonable.
	State and local agencies are required to submit their air quality monitoring data into AQS by the end of the quarter following the quarter in which the data were collected. These data must be certified by these agencies by June 30 each year (for the previous year)—within 1.5 years of the model year.
Accuracy	The most accurate characterization of the concentration of a given pollutant at a given time and location. Measurements are based on nationally consistent methods including State precision and accuracy evaluations.	Improved estimates of pollutant concentrations (and uncertainties) at times and locations where they are not measured compared to CMAQ model. Accuracy near monitors is better than accuracy where there are no monitors.
Spatial coverage	Spatial gaps, especially for rural areas, since the monitoring network is mostly population based.	Data will be provided on grid: 12 km in Eastern US and 36 km in Western US.
		No spatial gaps.
Temporal resolution	Varies by pollutant and location. Ozone is monitored daily, but for most locations only during the ozone season (approx. April through October). PM2.5 is often monitored year round using the Federal Reference Method (FRM). PM2.5 FRM daily measures are often only available for every third day. Some continuous PM2.5 monitors report daily PM2.5 measurements on an hourly basis and have been converted to the FRM-like measures within AQS.	Daily estimates with no temporal gaps except for the first and last day of the year (end effects of the model).
Ease of use	Medium—users must deal with missing values in space and time. EPHT must use consistent methods for handling missing values and developing exposure regions around monitors.	Easy—there are no missing values in space or time so the data will be used consistently in a national analysis.
Quality control	The data are supported by a comprehensive quality assurance program, ensuring data of known quality.	The quality of the predictions may not be consistent depending on the quality of the CMAQ estimates across states and may change as the data and methodology are improved over time. This information needs to be clearly documented for data users.