1 HEALTH IMPACTS OF AIR QUALITY ON THE BISHOP PAIUTE RESERVATION FOCUS ON PARTICULATE MATTER TONI RICHARDS, Ph.D., AIR QUALITY SPECIALIST ENVIRONMENTAL.

1

HEALTH IMPACTS OF AIR QUALITY

ON THE BISHOP PAIUTE RESERVATION

FOCUS ON PARTICULATE MATTER

TONI RICHARDS, Ph.D., AIR QUALITY SPECIALISTENVIRONMENTAL MANAGEMENT OFFICEBISHOP PAIUTE [email protected]

THE BISHOP PAIUTE RESERVATION IS 60 MILES FROM THE LARGEST SOURCE OF PM-10

IN THE NATION

THE OWENS DRY LAKE!

dustdust

3

WHERE IS THE BISHOP PAIUTE TRIBE?

On the California-Nevada BorderIn the “deepest valley” at 4,000ft between the Sierra Nevada and White Mountains200 miles South of Reno270 miles East of Las Vegas300 miles North of Los Angeles

Bishop Paiute Tribe

Los Angeles 300 miles

Las Vegas 270 miles

Reno 200 miles

Dry Lake

4

Separated high PM-10 days:dry lake probably had an impact on Bishop air quality

did not have an impact

Half the time, high PM-10 concentrations in Bishop were associated with dry lake activity

Conclusion: the dry lake has a significant Conclusion: the dry lake has a significant impact on Bishop Reservation air qualityimpact on Bishop Reservation air quality

EARLIER STUDY:Days with the highest PM-10

concentration on the Bishop Paiute Reservation 2003-2006

5

PM-10: DustOwens dry lake

Other barren lands

Dirt roads and parking lots

PM-2.5: SmokeWood smoke for home heating

Wildfires

Controlled burns

BISHOP’S PARTICULATE MATTER SOURCES

6

HISTORICAL PARTICULATE CONCENTRATIONS

2009 593 Apr 7

2008 1,064 Feb 13

2007 698 Sept 27

2006 370 Nov 11

2005 550 Jun 25

2004 1,220 Oct 19

PM-10 HOURLY MAX

2009 74 Jan 1

2008 94 Dec 94

2007 156 Jul 18

2006 96 Jan 28

2005 96 Dec 20

2004 109 Nov 24

PM-2.5 HOURLY MAX*

* Excludes July 4

AQI: PM-10 above 154 is unhealthyAQI: PM-10 above 154 is unhealthy AQI: PM-2.5 above 65.4 is unhealthyAQI: PM-2.5 above 65.4 is unhealthy

7

HEALTH STUDY: BACKGROUND

Few or no studies of the impacts of the dry lake

No studies of the impact of particulate matter on Reservation populations

WHY?

Can’t use standard methods (mortality / hospitalizations) on sparse rural populations

8

HEALTH STUDY: OUR APPROACH

Short term impacts: 3-5 daysDaily clinic visits as a measure of health

All visitsUnder age 5 / age 65 and overRespiratory / circulatory

Link visits to daily PM-10 and PM-2.5 concentrations

Hourly maximum in a 24-hour period24-hour average

Data: October 2006 to September 2007

9

HEALTH STUDY: METHODS

Descriptive statisticsExplore data structureVerify data quality

Time series correlationsVerify relationships among health variables and PM

All visitsUnder age 5 / age 65 and overRespiratory / circulatory

ModelingStandard time series methodsPoisson regression

10

DESCRIPTIVE STATISTICS: VISITSNumber of Visits Per Day

3 3 2

39

11

0

10

20

30

40

50

60

70

80

90

Total Under age 5 Over age 65 Respiratory Circulatory

Patient Type

Nu

mb

er o

f vi

sits

Max Num VisMin Num VisAvg Daily Visits

11

DESCRIPTIVE STATISTICS: PM-10Particulate Matter Less than 10 Microns

0

100

200

300

400

500

600

700

800

Oct-06

Nov-06

Dec-06

Jan-07

Feb-07

Mar-07

Apr-07

May-07

Jun-07

Jul-07

Aug-07

Sep-07

Month

PM

in m

icro

gra

ms

per

cu

bic

met

er

Average 24-Hr MCMax 24-Hr MCMax Hourly

12

DESCRIPTIVE STATISTICS: PM-2.5

Particulate Matter Less than 2.5 Microns

0.0

20.0

40.0

60.0

80.0

100.0

120.0

140.0

160.0

180.0

Oct-06

Nov-0

6

Dec-0

6

Jan-

07

Feb-0

7

Mar

-07

Apr-0

7

May

-07

Jun-

07

Jul-0

7

Aug-0

7

Sep-0

7

Month

PM

in m

icro

gra

ms

per

cu

bic

met

er

Average 24-Hr MCMax 24-Hr MCMax Hourly

13

DESCRIPTIVE STATISTICS: autocorrelations for visits

Autocorrelation Functions for Visits

-0.1

-0.05

0

0.05

0.1

0.15

0.2

0.25

0.3

Lag 1 Lag 2 Lag 3 Lag 4

Lag

Co

rrel

atio

n Total VisitsUnder Age 5Age 65 and overRespiratoryCirculatory

14

DESCRIPTIVE STATISTICS:cross correlations

Cross Correlation of Visits and Average PM-10

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

lag 0 lag 1 lag 2 lag 3 lag 4

Lag in days

Co

rre

lati

on

Total Visits

Visits under age 5

Visits 65 and over

Respiratory visits

Circulatory visits

Cross Correlations between Visits and Maximum PM-10

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

lag 0 lag 1 lag 2 lag 3 lag 4

Lag in days

Co

rre

lati

on

Total Visits

Visits under age 5

Visits 65 and over

Respiratory visits

Circulatory visits

Cross Correlation of Visits and Average PM-2.5

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

lag 0 lag 1 lag 2 lag 3 lag 4

Lag in days

Co

rre

lati

on

Total Visits

Under 5

65 and over

Respiratory

Circulatory

Cross Correlation of Visits and Maximum PM-2.5

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

lag 0 lag 1 lag 2 lag 3 lag 4

Lag in days

Co

rre

lati

on

total x PM2.5max

under5 x PM2.5max

over65 x PM2.5max

resp x PM2.5max

circ x PM2.5max

PM

-10

PM

-2.5

24-hour average Daily hourly maximum

15

PRELIMINARY MODELING:time series methods

Data Structure: 2nd order autoregressive process with indicator for days following a weekend or holiday

Visitst = α + β weekend/holiday + µt

Where µt = ρ1 µt-1 + ρ2 µt-2 + εt

Distributed Lag ModelVisitst = α + β PMt + β1 PMt-1 + β2 PMt-2 + β3 PMt-3

+ β4 PMt-4 + β5 weekend/holiday + µt

Where µt = ρ1 µt-1 + ρ2 µt-2 + εt

And t indexes days

16

TIME SERIES RESULTS:response of visits to PMResponse of Visits to Average PM-10

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

lag 0 lag 1 lag 2 lag 3 lag 4

Lag in days

Co

eff

icie

nt

Total Visits

Under age 5

65 and over

Respiratory

Circulatory

Response of Visits to Maximum PM-10

-0.015

-0.01

-0.005

0

0.005

0.01

0.015

lag 0 lag 1 lag 2 lag 3 lag 4

Lag in days

Co

eff

icie

nt Total Visits

Under age 5

65 and over

Respiratory

Circulatory

Response of Visits to Average PM-2.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

lag 0 lag 1 lag 2 lag 3 lag 4

Lag in days

Co

eff

icie

nt Total Visits

Under age 5

65 and over

Respiratory

Circulatory

Response of Visits to Maximum PM-2.5

-0.08

-0.06

-0.04

-0.02

0

0.02

0.04

0.06

0.08

lag 0 lag 1 lag 2 lag 3 lag 4

Lag in days

Co

eff

icie

nt Total Visits

Under age 5

65 and over

Respiratory

Circulatory

PM

-10

PM

-2.5

24-hour average Daily hourly maximum

17

FINAL MODELING:Poisson Regression

The number of visits follows a Poisson distribution with clustering within weeks

Visitswt = exposuret (exp (β PMt + β1 PMt-1

+ β2 PMt-2 + β3 PMt-3 + β4 PMt-4 + µ wt

where corr (µwt , µvs) = ρ if v=w = 0 otherwise

t indexes days and w indexes weeks

and exposuret = 1 (unknown) by assumption

The coefficients expβ compare the ratio of visits on days where PM increased by 1 microgram to those where it did not. Values >1 indicate a positive effect.

18

POISSON MODEL RESULTS:response of visits to PM

Incidence Rate Ratios of Visits to Average PM-10

0.98

0.99

1.00

1.01

1.02

lag 0 lag 1 lag 2 lag 3 lag 4

Lag in days

Inci

den

ce r

ate

rati

o total visits

under age 5

65 and over

respiratory

circulatory

Incidence Rate Ratios of Visits to Maximum PM-2.5

0.985

0.990

0.995

1.000

1.005

1.010

1.015

lag 0 lag 1 lag 2 lag 3 lag 4

Lag in days

Inci

den

ce r

ate

rati

o total visits

under 5

65 and over

respiratory

circulatory

Incidence Rate Ratio of Visits to Average PM-2.5

0.97

0.98

0.99

1.00

1.01

1.02

1.03

lag 0 lag 1 lag 2 lag 3 lag 4

Lag in days

Inci

den

ce r

ate

rati

o total visits

under age 5

65 and over

respiratory

circulatory

Incidence Rate Ratio of Visits to Maximum PM-10

0.997

0.998

0.999

1.000

1.001

1.002

1.003

lag 0 lag 1 lag 2 lag 3 lag 4

Lag in days

Inci

den

ce R

ate

Rat

io total visits

under age 5

65 and over

respiratory

circulatory

PM

-10

PM

-2.5

24-hour average Daily hourly maximum

19

Pilot study are clinic visits a measure of health that responds to air quality?Two approaches: time series and Poisson modelsResults broadly consistent across modelsSome evidence of impact of PM-10 for Some evidence of impact of PM-10 for circulatory visitscirculatory visitsModest evidence of impact of PM-2.5 for Modest evidence of impact of PM-2.5 for pediatric visitspediatric visits

LESSONS LEARNED

20

Small sample size due to clinic closures and missing PM data (about 180 observations)

Next step: obtain another year of dataNext step: obtain another year of data

NEXT STEPS