1

Understanding Relationships Between Changes in Ambient Ozone and Precursor

Concentrations and Changes in VOC and NOx Emissions from 1990 to 2004 in Central California

Envair – C. Blanchard and S. Tanenbaum

DRI – E. Fujita and D. Campbell

Alpine Geophysics – J. Wilkinson

Eight Hypothesized Explanations of Ozone TrendsJuly 1, 2008

2

Today’s Topics

I. Identify hypotheses to explain ozone trends

II. Evaluate hypotheses

III. Discuss relevance to policy questions

IV. Get feedback

3

Key Question

• Substantial and statistically significant downward trends in NOx, NMOC, CO.

• Generally downward trends in peak 8-hour ozone, but majority of sites less than 0.5 ppbv per year, only 7 of 42 statistically significant, and 10 sites have upward trends.

• Why hasn’t peak ozone decreased as much in some areas as in others?

4

Hypothesis 1

Ozone precursor emissions did not decline.

Incorrect

5

1990 – 2004 AQ Trends Summary

• No precursor trends significantly upward*

• NOx sig* down at 22 of 28 sites**

• CO sig* down at 21 of 25 sites**

• NMOC sig* down at 6 of 7 sites***

* p < 0.05** At least 10 years data. One or both metrics: morning (5 – 11 am) or midday (time of peak 8-hour ozone).*** 7 - 10 years data, morning (5 – 8 am).

6“7 x 7 Emissions” = in the 7x7 array of 4 km grid cells around a site

25 sites 7 sites 28 sites

Average Reductions of Emissions and Ambient Concentrations, 1990 - 2004

0

10

20

30

40

50

60

70

80

90

CO NMOC NOx

Pollutant

Red

uct

ion

(p

erce

nt)

Ambient 7 x 7 Emissions Regional Emissions

7

Hypothesis 2

Ozone precursor emissions did not decline enough in some areas experiencing growth and development to produce a significant ozone response in those areas.

Plausible

8

Sub-Region

N orthern S ierra Footh ills

Southern S ierra Footh ills

Sacram ento Valley

N orthern SF Bay A reaEastern SF Bay A reaSouthern SF Bay A reaN orthern San JoaquinC entra l San Joaquin

Southern San Joaquin Boxes show where both NMOC and NOx emissions in the 7x7 arrays of 4 km grid cells (28 km x 28 km) surrounding monitoring sites decreased by less than 20 percent from 1990 to 2004

9

Sites Where Local NMOC and NOx Emissions Decreased by Less Than 20 Percent From 1990 to 2004

-10

-5

0

5

10

15

20

Turlock-S MinaretStreet

Merced-S CoffeeAvenue

Madera-HealthDept #2

Madera-PumpYard

Parlier Visalia-N ChurchStreet

Site

Em

issi

on

s D

ecre

ase

1990

- 2

004

(%)

NMOC Emissions NOx Emissions

“Local” means in the 7x7 array of 4 km grid cells

10If local NMOC emissions unchanged, how much will ozone change?

Regional and Local Ozone-Season NMOC Emissions in 1990 and 2004

0

100

200

300

400

500

600

700

800

San Francisco Bay Sacramento & NSF N & C SJV & SSF Five SJV Sites: Turlockto Visalia

Region

NM

OC

Em

issi

on

s (t

on

s p

er d

ay)

NMOC 1990 NMOC 2004

11Caveat: ambient NMOC decreases > inventory decreases

NMOC Decrease (1994-2004)

0

20

40

60

80

100

120

140

160

SacramentoDel Paso

Folsom Fresno-FirstStreet

Clovis Parlier Bakersfield-Golden

Arvin

NM

OC

Dec

reas

e (p

erce

nt)

Ambient 3 x 3 Emissions 5 x 5 Emissions 7 x 7 Emissions

12

Hypothesis 3

Background* ozone concentrations increased.

* Background ozone means ozone formed outside the North American boundary layer.

Implausible

13

Increasing Sites(north to south)

No. HighlandsPittsburgConcordTurlockMerced

Fresno SkyparkClovis

Fresno DrummondParlierOildaleArvin

Maricopa

D ecreasing

Increasing

Average M axim um 8-hour O zone

Increasing Sites(north to south)

No. HighlandsPittsburgConcordTurlockMerced

Fresno SkyparkClovis

Fresno DrummondParlierOildaleArvin

Maricopa

D ecreasing

Increasing

Average M axim um 8-hour O zone

Ozone increases occurred at scattered sites in separate subregions

Mean peak 8-hour ozone in 1990-94 and in 2000-04

14

0

10

20

30

40

50M

ean

Ozo

ne (

ppb

v)

1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005Year

Crater LakePoint ReyesLava Beds

Crater Lake = -395.536 + .219 * Year; R^2 = .046Point Reyes = -1209.936 + .619 * Year; R^2 = .368Lava Beds = -1011.904 + .528 * Year; R^2 = .438

Possible upward trend in background ozone of ~0.5 ppbv per year- if true, may make overall progress more difficult

National Park Service Passive Ozone Monitoring

15

Hypothesis 4

Ozone trends were masked by changes in meteorological conditions.

Implausible

16Four meteorological types identified by PCA and K-means clustering

Number of Days in Each Meteorological Type, By Month

0

50

100

150

200

250

300

350

400

May June July August Sep Oct

Month

Nu

mb

er o

f D

ays

Type 1 Type 2 Type 3 Type 4

17

Frequency of Days In Each Meteorological Type (May - October)

0

20

40

60

80

100

120

140

160

180

200

1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Year

Nu

mb

er o

f D

ays

Type 1 Type 2 Type 3 Type 4

Year-to-year variations occurred in frequency of met types ….

18

Central San Joaquin High-Ozone Days By Meteorological Type

0

10

20

30

40

50

60

70

1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Year

Nu

mb

er o

f D

ays

Type 1 Type 2 Type 3 Type 4

…. And the distributions of high-ozone days varied ….

19…. But downward ozone trends did not depend upon met type ….

Number of Sites With Downward Ozone Trends, 1994-2004, by Meteorological Type

0

10

20

30

40

50

60

70

Type 1 Type 2 Type 3 Type 4 Total Sites With 5+Years Data (All

Trends)

Meteorological Type

Nu

mb

er o

f S

ites

20

Number of Sites With Upward Trends

0

2

4

6

8

10

12

Annual Mean Annual Mean Annual Mean Annual Mean Annual Mean Annual

4th-Highest Peak 8-Hour Peak 8-Hour Peak 8-Hour Peak 8-Hour Peak 8-Hour

Peak 8-Hour on Top 60 Days Met Type 1 Met Type 2 Met Type 3 Met Type 4

Trend Test

Nu

mb

er o

f S

ites

…. And upward ozone trends were consistent across met types

21

Hypothesis 5

Decreases in VOC emissions and reactivity slowed the rate of ozone formation but did not reduce the ultimate amount of ozone formed. Therefore, reductions of peak ozone occurred in urban cores and at near-downwind sites, but not at far downwind locations.

Plausible

22

Measurements show ~30 – 120% decreases in ambient NMOC concentrations

Sum of Species, Reactivity-Weighted Sum of Species, and Total NMOC Decreases (1994-2004)

0

20

40

60

80

100

120

140

160

180

200

Sacramento DelPaso

Folsom Fresno-FirstStreet

Clovis Parlier Bakersfield-Golden

Arvin

Dec

reas

e (p

erce

nt)

Sum of Species Reactivity-Weighted Sum of Species NMOC

23

Average reactivity* shows some declines(but mass decrease is larger and more important)

*kOH weighted by mean annual concentrations (units are ppbC-1 sec-1)

24

0

10000

20000

30000

40000

50000

60000

70000

0 20 40 60 80 100Distance (km)

NMOC 1997-2004NMOC 1990-1996

Fremont

San Jose- Piedmont

San Jose-4th

Los Gatos

San Martin

Gilroy

Southern San Francisco Bay Area

NM

OC

Em

issi

on

s (t

on

s/yr

)

0

10000

20000

30000

40000

50000

60000

70000

0 20 40 60 80 100Distance (km)

NMOC 1997-2004NMOC 1990-1996

Fremont

San Jose- Piedmont

San Jose-4th

Los Gatos

San Martin

Gilroy

Southern San Francisco Bay Area

NM

OC

Em

issi

on

s (t

on

s/yr

)

40

50

60

70

80

90

100

110

Ozo

ne (

ppb

v)0 20 40 60 80 100

Distance (km)

Peak 8-hour Ozone 1997-2004

Peak 8-hour Ozone 1990-1996

Southern San Francisco Bay Area

Fremont

San Jose-Piedmont

San Jose-4th

Los Gatos

San Martin

Gilroy

40

50

60

70

80

90

100

110

Ozo

ne (

ppb

v)0 20 40 60 80 100

Distance (km)

Peak 8-hour Ozone 1997-2004

Peak 8-hour Ozone 1990-1996

Southern San Francisco Bay Area

Fremont

San Jose-Piedmont

San Jose-4th

Los Gatos

San Martin

Gilroy

In S Bay area, ozone declined at both urban and downwind sites

25

40

50

60

70

80

90

100

110

Ozo

ne (

ppb

v)

0 20 40 60 80 100Distance (km)

Peak 8-hour Ozone 1997-2004Peak 8-hour Ozone 1990-1996

Sacramento and the Northern Sierra Foothills

Sac-T St

Sac-Del Paso

N. Highlands

Roseville

Elk Grove

Folsom

Auburn

Placerville

Grass Valley

40

50

60

70

80

90

100

110

Ozo

ne (

ppb

v)

0 20 40 60 80 100Distance (km)

Peak 8-hour Ozone 1997-2004Peak 8-hour Ozone 1990-1996

Sacramento and the Northern Sierra Foothills

Sac-T St

Sac-Del Paso

N. Highlands

Roseville

Elk Grove

Folsom

Auburn

Placerville

Grass Valley

0

10000

20000

30000

40000

50000

60000

70000

0 20 40 60 80 100Distance (km)

NMOC 1997 -2004NMOC 1990 -1996

Sac-T St

Sac-Del Paso

N. Highlands

Roseville

Elk Grove

Folsom

Auburn

PlacervilleGrass Valley

Sacramento and the Northern Sierra Foothills

NM

OC

Em

issi

on

s (t

on

s/yr

)

0

10000

20000

30000

40000

50000

60000

70000

0 20 40 60 80 100Distance (km)

NMOC 1997 -2004NMOC 1990 -1996

Sac-T St

Sac-Del Paso

N. Highlands

Roseville

Elk Grove

Folsom

Auburn

PlacervilleGrass Valley

Sacramento and the Northern Sierra Foothills

NM

OC

Em

issi

on

s (t

on

s/yr

)

In Sacramento, ozone declines varied among sites.

26

10000

20000

30000

40000

50000

60000

70000

0 20 40 60 80 100 120 140Distance (km)

NMOC 1997 -2004NMOC 1990 -1996

NM

OC

Em

issi

ons

(ton

s/yr

)

Northern San Joaquin and Southern Sierra Foothills

Stockton-Hazelton

Stockton-Mariposa

Modesto

San Andreas Jackson

Turlock

Sonora

Merced

Yosemite-Turtleback

10000

20000

30000

40000

50000

60000

70000

0 20 40 60 80 100 120 140Distance (km)

NMOC 1997 -2004NMOC 1990 -1996

NM

OC

Em

issi

ons

(ton

s/yr

)

Northern San Joaquin and Southern Sierra Foothills

Stockton-Hazelton

Stockton-Mariposa

Modesto

San Andreas Jackson

Turlock

Sonora

Merced

Yosemite-Turtleback

40

50

60

70

80

90

100

110

Ozo

ne

(pp

bv)

0 20 40 60 80 100 120 140Distance (km)

Peak 8-hour Ozone 1997-2004Peak 8-hour Ozone 1990-1996

Northern San Joaquin and Southern Sierra Foothills

Stockton-Hazelton

Stockton-Mariposa

Modesto

San Andreas

Jackson

Turlock

Sonora

Merced

Yosemite-Turtleback

40

50

60

70

80

90

100

110

Ozo

ne

(pp

bv)

0 20 40 60 80 100 120 140Distance (km)

Peak 8-hour Ozone 1997-2004Peak 8-hour Ozone 1990-1996

40

50

60

70

80

90

100

110

Ozo

ne

(pp

bv)

0 20 40 60 80 100 120 140Distance (km)

Peak 8-hour Ozone 1997-2004Peak 8-hour Ozone 1990-1996

Northern San Joaquin and Southern Sierra Foothills

Stockton-Hazelton

Stockton-Mariposa

Modesto

San Andreas

Jackson

Turlock

Sonora

Merced

Yosemite-Turtleback

Ozone

NMOC Emissions

NMOC emissions did not decline in some parts of the N SJV and SSF. This makes it difficult to assess the argument that NMOC-control was not effective at the far downwind sites.

27

Hypothesis 6

Decreases in NOx emissions and concentrations resulted in less titration of ozone. Therefore, ozone concentrations increased in NOx-rich areas, but peak ozone declined in NOx-limited areas.

Plausible

28

Mean Hourly NOx at Merced

0

5

10

15

20

25

30

35

40

45

50

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Hour

NO

x (p

pb

v)

NOx 1990-94 NOx 1995-99 NOx 2000-04

Mean Hourly Ozone at Merced

0

10

20

30

40

50

60

70

80

90

100

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Hour

Ozo

ne

(pp

bv)

O3 1990-94 O3 1995-99 O3 2000-04

Ozone increases at night can be linked to lower NOx ….

… But effects on peak ozone are difficult to interpret

29

Yosemite National Park - Turtleback Dome

0

20

40

60

80

100

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Ozo

ne

(pp

bv)

Jerseydale

0

20

40

60

80

100

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Ozo

ne

(pp

bv)

Five Mile Learning Center

0

20

40

60

80

100

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Ozo

ne

(pp

bv)

Yosemite National Park - Turtleback Dome

0

20

40

60

80

100

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Ozo

ne

(pp

bv)

Jerseydale

0

20

40

60

80

100

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Ozo

ne

(pp

bv)

Five Mile Learning Center

0

20

40

60

80

100

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Ozo

ne

(pp

bv)

0

10

20

30

40

50

60

70

80

90

Ozo

ne (

ppbv

)

0 4 8 12 16 20 24HOUR

2000_20041995_19991990_1994

Period

Yosemite – Turtleback Dome

0

10

20

30

40

50

60

70

80

90

Ozo

ne (

ppbv

)0 4 8 12 16 20 24

HOUR

2000_20041995_19991990_1994

Jerseydale

Period

Mid-elevation Sierra Nevada sites show peak ozone improvements compared to early 1990s (but less change since the mid-1990s)

Mean Peak 8-Hour Ozone

30

Hypothesis 7

Greater or lesser decreases of VOC emissions compared to NOx emissions resulted in changes in VOC/NOx ratios, which led to changes in the efficiency of ozone production per unit of precursor mass.

Plausible

31

HighVOC/NOx

Low VOC/NOx

Termination Reactionat low VOC/NOx,

removes NOx.

Termination Reactionat high VOC/NOx, removes radicals.

Significance of VOC/NOx Ratios in Ozone Formation

Ozone formation is most efficient at VOC/NOx

(ppbC/ppb) ratios between 10-12.

32

Morning (5 – 8 a.m.) NMOC/NOx ratios declined over time

(Emission ratios average about 30% lower than ambient ratios)

33

NMOC/NOx - Parlier

0

5

10

15

20

25

30

1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Year

NM

OC

/NO

x (p

pb

C/p

pb

v)

5:00 AM 12 Noon 4:00 PM 11:00 PM

Differences between morning and afternoon NMOC/NOx ratios declined

34

Other ratios suggest changes in air mass ages or aging

5 – 8 a.m 12 Noon – 2 p.m.

35

Hypothesis 8

The validity of hypotheses 5, 6, and 7 varied during the trend period, with hypothesis 5 (VOC reductions ) having greater relevance early and hypotheses 6 (NOx reductions) and 7 (VOC/NOx ratios) increasing in importance toward the end.

Plausible

36Somewhat larger NMOC declines from early to mid ’90s

Ozone-Season Emission TrendsNineteen Counties in CCOS Domain

0

200

400

600

800

1000

1200

1400

1600

1800

2000

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Year

Em

issi

on

s (t

on

s p

er d

ay)

NMOC NOx NMOC Trend NOx Trend

37

Plausible Hypotheses

H2: Emissions did not decline enough in some areas to affect ozone concentrations

• Applicable in portions of N SJV & C SJV

• NMOC emissions in affected area are ~20 percent of N & C SJV total

38

Plausible Hypotheses (continued)

H5: VOC emission reductions were effective in urban centers and near-downwind areas but not far downwind

• Ambient NMOC reductions are evident,

but spatial patterns of ozone more complex than hypothesis

• Difficult to infer causal link to ozone

39

Plausible Hypotheses (continued)

H6: NOx emission reductions increased ozone in NOx-rich areas but reduced peak ozone at NOx-limited sites

• Higher ozone during non-peak hours is probably related to decreased NO

• Effects of NOx reductions on peak ozone are unclear

40

Plausible Hypotheses (continued)

H7: Changes in VOC/NOx made ozone production more (or less) efficient

• Declines in NMOC/NOx ratios occurred in measurements and the inventory

• Differences between afternoon and morning NMOC/NOx ratios declined - may indicate that air mass ages and aging changed

41

Plausible Hypotheses (continued)

H8: H5, H6, and H7 are all applicable, but differed in importance at different times

• Consistent with data but difficult to demonstrate

42

Preliminary Implications

• Observed ozone trends reflect emission changes and atmospheric chemistry

• Ongoing emission reductions needed – including local reductions in some areas

• NMOC emissions: some areas became more NMOC-sensitive. NMOC mass changes have probably been more important than reductions of reactivity