Assessing the Impacts of Local Deposition of Mercury Associated

Brookhaven Science AssociatesU.S. Department of Energy

Assessing Local Deposition of Mercury Associated with

Coal-Fired Power Plants

T. Sullivan1, B. Bowerman1, J. Adams1, F. Lipfert2,

1 Brookhaven National Laboratory2 Independent Consultant

Presented at theDOE/NETL Mercury Control Technology R&D Program Review

Pittsburgh, PA July, 15, 2004


BackgroundBackgroundMercury regulation is on its way.EPA has suggested a Cap-and-Trade ProgramImpacts of local deposition are perceived to be an issue of concern for Cap-and-Trade• Secretary McGinty (PA – DEP) “Unlike most pollutants,

mercury is highly toxic and does not disperse easily, creating “hot spots” of contamination.” July 04

• Secretary Cipriano (IL – EPA) “Specifically, we are concerned that local "hot spots" of elevated mercury may result or worsen, especially if the required reduction levels are not sufficiently strict.” Feb 04


BackgroundBackground

Modeling studies suggest some increased local deposition, however, few measurement studies.Local deposition of reactive gaseous mercury and any particulate mercury is expected during precipitation events. Elemental mercury enters the global cycle.Data are needed to asses local impacts.


ApproachApproach

Select a coal-fired power plant for analysis.Perform deposition modeling based on plant specific parameters and meteorological conditions.Design soil and vegetation sampling program to look for excess deposition within 10 Km of the plant and determine the validity of modeling.Examine data for correlation with model predictions and look for hot spots.


Coal-Fired Power PlantCoal-Fired Power PlantA large sized power plant in the Midwest was

selected for study.

Annual emissions of Hg 340 kg/yr

Fraction of Hg(+2) = 0.18 (61kg/yr of Hg(+2))

Large stack height (200 m), lower deposition locally.


Deposition ModelingDeposition Modeling

• Use plant specific parameters (emission rates, stack height, etc.).• Examine meteorological records over 10 year period. Select most representative year.• Model deposition over the course of 1 year.


Wind and Rain RoseWind and Rain RoseRain RoseWind Rose


Deposition Modeling ResultsDeposition Modeling Results

Power PlantN

Deposition unitug/m2/yr

Background wetdeposition of 5 – 10 ug/m2/yr


Final Sampling GridFinal Sampling Grid

8 km


SamplingSamplingCover an area of 64 km2 primarily south west of plantAt each of the 51 grid locations collect:• 3 soil samples from top 5 cm (3 meter spacing)• 1 deep soil sample from 5 – 10 cm• 1 vegetation sample


AnalysisAnalysisAll samples shipped back to BNLSamples analyzed on Direct Mercury Analyzer (DMA-80). (Moisture determined separately.)Each sample measured 3 times to verify homogeneity.Blank analysis every 10th sampleNIST Standard every 10th

sample


ResultsResults51 Sample Sites Average 28.7 ng/g• Median 27.4 ng/g• Standard Deviation – 7

ng/g• Maximum – 55 ng/g• Minimum – 11.6 ng/g

Probability of being less than soil Hg level

00.10.20.30.40.50.60.70.80.9

1

0 10 20 30 40 50 60Soil Hg (ng/g)

Pro

babi

lity


Soil ResultsSoil ResultsEffects of deposition are subtle and there is variability in the spatial distribution of Hg in soils.

Statistically significant agreement between model and data not found.

No clear definition of background.


Results (deep vs. surface samples)Results (deep vs. surface samples)

0

10

20

30

40

50

60

70

80

0 10 20 30 40 50Locxation (1=A6, 48=J3)

Mer

cury

(ng/

g)

DeepSamplesSurface


Results (deep vs. surface samples)Results (deep vs. surface samples)Difference in Surface soil versus Deep soil Concentration

-20

-15

-10

-5

0

5

10

15

20

25

30A

6B

2B

3B

4.1

B4.

2B

5.1

B5.

1B

6B

7C

.a3

C2

C3. C3

C4. C4

C5. C5

C6

C7

D.a

3D

3D

6D

7E.

a3E.

a4E.

a5E3

.1E3

.2 E3E4

.2 E4E5

.1 E5 E6F2

.1 F2 F3 F6 F7 G4

G5

H3

H4

H5

H6 I6 I7 J3

Location

Diff

eren

ce in

con

cent

ratio

n (n

g/g)


Vegetation ResultsVegetation ResultsEffects of deposition are similar and there is much larger variability than in the soil samples.

Same general pattern asSoil samples.

-8000 -6000 -4000 -2000 0 2000 4000 6000-6000

-4000

-2000

0

2000

10.70 to 22.90 22.90 to 31.50 31.50 to 84.40 84.40 to 691.10

Vegetation mercury levels (ng/g) versus predicted deposition


Comparison of deposition modeling and data on mass depositionComparison of deposition modeling and data on mass deposition

Three estimates obtained:• Mass deposited = Excess in surface soils versus

deep soils (0.7 ng/g)• Mass deposited = Surface soil average minus

average of the lowest 1/3 of the samples (8.4 ng/g). Lowest 1/3 taken as pre-operational background.

• ISCT model annual deposition prediction multiplied by plant life.


Estimates of Mass DepositedEstimates of Mass Deposited

0.70.129400Deposition Modeling (23 years)

Lowest 1/3 as background (upper bound)

Deep vs. Surface Soils

Case

30000

2700

Total Mass Required over plant life time of 23 years (g)

0.4

0.04

Percentage of total Hg emission

2.1

0. 2

Percentage of reactive Hg emissions


ConclusionsConclusions

At the power plant under study a small fraction (< 0.5%) of mercury appears to be deposited within 5 miles of the plant. No evidence for ‘hot spots’ in soil.General agreement between estimates of total mass deposited between data and model.Agreement between data and model in terms of spatial pattern is not statistically significant.Next step: Repeat at another plant

Documents

Assessing the Impacts of Local Deposition of Mercury Associated