The Need for More Extensive Air Quality Monitoring to Resolve the Air Quality Health Issue

Transcription

1 The Need for More Extensive Air Quality Monitoring to Resolve the Air Quality Health Issue Ronald E. Wyzga, Sc.D. Senior Technical Executive May 13, 2010 Xi an, China

2 Current Regulatory Monitoring USEPA, States Criteria Air Pollutants NO 2 CO SO 2 PM Not all measured with same intensity O 3 Pb 2

3 Current Monitoring Speciation Trends Network EC/OC Sulfate Ammonium Elements XRF Nitrate Measured every 3 rd, 6 th day 3

4 U.S. PM 2.5 Compliance and Chemical Speciation Network 4

5 Southeastern Aerosol Research and Characterization Study (SEARCH) Supersites (Hansen et al., 2003, JAWMA, p. 1460) 5

6 Hazardous Air Pollutants Monitoring is not mandated 188 substances Sporadic monitoring Emissions data available EPA modeling to estimate concentrations 6

7 Air Quality Health Studies Epidemiology Need for exposure data Monitoring data often used as surrogate for exposure Two types of studies Time series (requires daily data) Cohort (requires average measures) 7

8 Time Series Studies Sufficient data Measurements every 3 rd and 6 th day; data limit power of studies Make distributed lag analyses difficult, impossible Measuring correct pollutants 8

9 Cohort Studies Inform annual standard Need long-term average concentration Issue of bias if record incomplete Issue of correct measure Lags Correct time average Importance of peak exposures Require data for several locations 9

10 Cohort Study Major Issue: Are we measuring correct pollutants 10

11 Example: The Veterans Cohort 67,000 males, recruited in , with indications of hypertension 35% African-American, 81% current or former smokers Average age at recruitment: 51 years Nationally representative (44% in Northeastern US) 66% died through 2001 Followed since mid 1970 s Hypertensive diagnosis, but Homogeneous socioeconomic status 11

12 Previous Results: Veterans Study Examined criteria pollutants:? ozone signal Examined traffic density Traffic density dominated criteria pollutants Examined STN variables 12

13 PM 2.5 Constituents; STN Data Single Pollutant Model Constituent t-statistic Mean effect EC 2.829* NO * V 2.612* Ni 2.143* OC SO *statistically significant 13

14 Two Pollutant Models t-statistic mean effect ln (traffic density) EC ln (traffic density) 2.723* NO ln (traffic density) 2.043* V ln (traffic density) 2.390* Ni *statistically significant 14

15 Extension of Analyses: HAPs Data 188 substances; subset chosen EPA modeled ambient concentrations at county (sub-county level) Use of 50 th percentile for each county Metals: As, Pb, Mn, Hg, Ni Traffic-related compounds: benzene, diesel particulate matter Others: formaldehyde, POMs, HCl, polypropylene Also added sulfates, NO x 15

16 Example: Cohort Study Air Quality Variables Extended to Estimated HAPs Concentrations Significant in single pollutant models: * Remains significant with traffic density in model Benzene* Formaldehyde* Diesel PM* NOx* EC* Nickel* Polycyclic organics* Hg Pb SO 2 Cl* Arsenic* 16

17 Conclusions Current epidemiological studies examine only a few pollutants Limited by available monitoring data Results to date suggest other pollutants important Indicator species or Possible replication of results limited Statistical power limits analyses in time series methods Additional monitoring is needed to resolve issue: greater frequency, more pollutants Toxicological support needed as well 17

18 Leapfrogging Opportunity Monitor beyond the usual suspects! 18