Transcript PERCH_Symposium_Mercury

Atmospheric Deposition of
Mercury, Trace Metals and Major
Ions in the Pensacola Bay
Watershed
Jane Caffrey
Center for Environmental Diagnostics and Bioremedation
University of West Florida
and William M. Landing
Department of Oceanography
Florida State University
Acknowledgements
• EPA PERCH project
• Melissa Overton, Autumn Dunn, Nathaniel Davila,
Tanner Martin, Fran Aftanas, Elizabeth Gaige , Brad
Kuykendall for their dedication in the field
• Sara Cleveland, Kati Gosnell, and Nishanth
Krishnamurthy at FSU
• Dr. Sikha Bagui and Jessie Brown at UWF Computer
Science for database development
• Dr. Subhash Bagui and Arup Sinha at UWF Mathematics
and Statistics
Escambia County 9th in total toxic
emissions nationwide, 16th in air
emissions
• Coal fired power plant
• Papermill (coal for some
power generation)
• Other industrial activities
(Solutia, Air Products, etc.)
• 8 Superfund sites
• Partnership for
Environmental Research and
Community Health (PERCH)
– EPA funded
– Human health effects
– Environmental effects
Mercury emissions sources in the Gulf of Mexico region,
based on the 1999 National Emissions Inventory (U.S. EPA)
From Mark Cohen, NOAA ARL
Generally higher mercury deposition in
the SE, most driven by higher rainfall
along the Gulf Coast
Mercury deposition is an international problem
Worldwide emission estimates
US coal fired power plants represent about 1% of Hg emissions globally
53% of emissions come from Asia, 18% from Africa, 11% from Europe, 9% from North America
What is the atmospheric wet deposition
of mercury, trace metals and major ions
to the lower Pensacola Bay
Watershed?
• Are there seasonal patterns in deposition?
• How do prior weather conditions affect
deposition?
• Are there hot spots?
• How important are Local sources?
Sampling
Locations
Sample collection began in
November 2004 and will continue
through March 2010
Precipitation
16
Hurricane
Dennis
Ellyson
Molino
Pace
cm
12
8
4
0
Nov
2004
May
2005
Nov
May
2006
Nov
~ 565 samples from 225 rain events to Feb 2008
May
2007
Nov
2008
Analyses
UWF -WRL
• pH
• Sulfate
• Nitrate+nitrite
• Chloride
• Ammonium
• Sodium
• Calcium
• Phosphate
FSU - Oceanography
• Mercury
• Trace metals
– mineral/crustal elements:
Al, Si, Sc, Ti, Mn, Fe, Rb,
Y, Nb, Cs, La and all rare
earth elements, Th, U
– Sea Salt aerosols: Li, Na,
Mg, Sr
– urban pollution, fossil fuel
combustion: V, Ga, Sb,
Pb, Bi, P, Cu, Zn
– As, Se, Sn
Rain is generally acidic, sometimes
highly acidic
7
Ellyson
Molino
Pace
6.0
5
5.6
4.8
Nov 4.4
2004
May
2005
May
Nov
2006
Nov
May
2007
Nov
4.0
FL
23
FL
14
Pa
ce
lys
on
El
M
ol
in
o
3.6
AL
02
3
5.2
AL
24
4
VWM pH
pH
6
2008
Sulfate Flux
Ellyson
2-
SO4 flux
mg SO42-/m2/event
1000
Molino
Pace
100
10
1
0.1
11/04
5/05
11/05
5/06
11/06
5/07
11/07
Ellyson
Molino
Pace
Sulfate fluxes are higher in
Pensacola and AL sites
2005
3000
SO4 flux mg SO4/m2/y
2006
2500
2007
2000
1500
1000
500
0
LA12
LA30
AL02
AL24
M
E
Highest fluxes at Ellyson
Lowest at FL14 and FL23
P
FL14
FL23
Counties with higher sulfate emissions
have higher sulfate deposition
Sulfate
Deposition mg/m2/y .
2000
r = 0.68
SR
AL24
Esc AL02
1500
FL23
FL14
1000
500
100
1000
10000
Emissions 2002 lbs/yr
Based on EPA TRI estimates for 2002
100000
Nitrate flux
Ellyson
Molino
1000
-
NO3 flux
mg NO3-/m2/event
Pace
100
10
1
0.1
11/04
5/05
11/05
5/06
11/06
5/07
11/07
Ellyson
Fluxes were similar at Pensacola Bay sites
Higher fluxes in spring and summer
Molino
Pace
Regional Nitrate Flux
2500
2005
NO3 flux mg NO3/m2/y
2006
2007
2000
1500
1000
500
0
LA12
LA30
AL02
AL24
M
E
P
FL14
Higher nitrate fluxes at Pensacola sites than at any NADP sites
FL23
Counties with higher nitrate emissions
had higher nitrate deposition
Nitrate
Deposition mg/m2/y .
2000
r = 0.80
1500
Esc
SR
AL24
1000 FL23
AL02
FL14
500
0
10000
20000
30000
40000
Emissions 2002 lbs/yr
Based on EPA TRI estimates for 2002
50000
60000
Chloride Flux
10000
Ellyson
Molino
Pace
Cl- flux
mg Cl-/m2/event
1000
100
10
1
0.1
11/04
5/05
11/05
5/06
11/06
5/07
11/07
Ellyson
Molino
Pace
Sodium and Chloride fluxes are
high relative to other sites
P
2005
2500
2006
2007
Na flux mg/m2/y
2000
1500
1000
500
0
LA12
LA30
AL02
AL24
M
E
P
FL14
FL23
Higher sodium and chloride fluxes
when closer to Gulf of Mexico
Average Na+ flux mg/m2/y .
1400
1200
R2 = 0.70
1000
800
600
400
200
0
0
20
40
60
80
100
Distance from Gulf, km
120
140
160
Conclusions
• Low pH in some rain events (pH usually < 5,
sometimes less than 4)
– Ellyson and Pace sites had significantly higher H+
fluxes than NADP sites
• Sulfate and nitrate fluxes higher at Pensacola
Bay sites than most NADP sites
• Sulfate and Nitrate fluxes are highest in counties
that have high SO2 or NOx emissions
• Sodium and Chloride fluxes higher at Pensacola
Bay sites. Sea salt aerosols are important
component of rain in the region
Mercury and trace elements in rainfall
from the Pensacola airshed: local,
regional, and distant sources
William M. Landing
Department of Oceanography
Florida State University
Jane Caffrey
Center for Environmental Diagnostics and
Bioremedation
University of West Florida
Acknowledgements
• EPA PERCH project (2005-2008)
• EPRI funding for 2008-present.
• FSU Graduate students: Sara Cleveland, Kati Gosnell,
Nishanth Krishnamurthy
• UWF students: Nathaniel Davila, Tanner Martin, Brad
Kuykendall, Fran Aftanas, Elizabeth Gaige
Geographic Distribution of Largest Anthropogenic
Mercury Emissions Sources in the U.S. (1999) and
Canada (2000); from Mark Cohen NOAA/ARL
Gulf coast is not loaded with Hg point sources
Generally higher mercury deposition in the Southeastern US
and along the Gulf coast.
Is this due to long-range transport of GEM and GEM-> RGM
conversion during summer months? (Yes)
Escambia County:
• Coal fired power plant
• International Paper mill (coal
for some power generation)
• Other industrial activities
(Solutia, Air Products,
Sacred Heart medical waste
incinerator, etc.)
EPA National Emissions Inventory shows that Plant Crist is
the largest air emitter of mercury in the region.
-- Need reliable mercury speciation profile for each source
(RGM, GEM, Hg-p)
Plant Crist
International Paper Company
Sacred Heart Hospital
Perdido Landfill & Mrf
Naval Air Station Pensacola
Baker & Son Construction Co. C&D Lf
Panhandle Paving & Grading (Long Lf)
Beulah Landfill
Camp Five Landfill
Auto Shred Industries Dump
Solutia Inc
1999
Mercury
Emissions
(lb/yr)
2002
Mercury
Emissions
(lb/yr)
2005
Mercury
Emissions
(lb/yr)
220.0000
46.5000
2.5600
0.0395
0.0017
0.0017
0.0010
0.0010
0.0008
0.0004
183.0000
191.0000
1.8600
0.0020
0.0395
0.0000
0.0017
0.0010
0.0010
0.0008
0.0004
0.0103
0.0395
0.0017
0.0010
0.0010
0.0008
0.0004
0.1820
RGM
GEM
Hg-p
68%
25%
6%
What is the atmospheric wet deposition of
mercury, trace metals and major ions to the lower
Pensacola Bay Watershed?
Rainfall chemistry in the Pensacola region is impacted by
multiple local and regional emission sources.
Rainfall mercury deposition in the southeastern US and
along the Gulf of Mexico coast is already elevated due to
long-range transport and transformation of Gaseous
Elemental Mercury (GEM) to Reactive Gaseous Mercury
(RGM); can we reliably quantify the impacts from individual
local and regional sources of RGM (and particulate Hg-p)?
What is the seasonal pattern in mercury and trace element
deposition? (Completed)
What are the relationships to local and regional
meteorology? (On-going)
Are there hot spots? (Not apparent)
What is the atmospheric wet deposition of
mercury, trace metals and major ions to the lower
Pensacola Bay Watershed?
Can we use other trace elements to “fingerprint” specific
emissions sources?
Installation of Hg emission control technology on the local
CFPP (Plant Crist) in late 2009 may change local rainfall
chemistry and trace element deposition.
Project is monitoring rain events at multiple sites for
multiple years to obtain a statistically-significant number of
such events both before and after new Hg emission control
technology is installed.
Rainfall impact from a point-source plume requires
simultaneous presence of the plume and rainfall.
Sampling
Locations:
1. Ellyson
2. Pace
3. Molino
a.
b.
Modified AerochemMetrics Wet/Dry samplers:
1. Plexiglas splash guard on leading edge of roof to eliminate
splash contamination.
2. Foam seal inside FEP Teflon film bag.
3. Three replicate “nested” funnel/bottle sets (one for pH, Nspecies and major ions; two for Hg and trace elements).
Analyses: Samples collected on an “event” basis
(within 24 hours).
UWF -WRL
• pH
• Sulfate
• Nitrate+nitrite
• Chloride
• Ammonium
• Sodium
• Calcium
• Phosphate
FSU - Oceanography
• Mercury
• Trace metals
– mineral/crustal elements: Al, Si,
Sc, Ti, Mn, Fe, Rb, Y, Nb, Cs,
La and all rare earth elements,
Th, U
– Sea Salt aerosols: Li, Na, Mg,
Sr
– urban pollution, fossil fuel
combustion: V, Ga, Sb, Pb, Bi,
P, Cu, Zn, As, Se, Sn
Duplicate receiving bottles allows evaluation of analytical reliability. As
concentrations approach the detection limit, scatter increases (Ba).
0.4
60
50
As-B (ng/L)
Hg-B (ng/L)
0.3
40
30
0.2
20
0.1
10
0
0.0
0
10
20
30
40
50
60
0.0
0.1
Hg-A (ng/L)
0.2
0.3
0.4
As-A (ng/L)
2.0
10000
1.8
1.6
8000
Ba-B (ng/L)
Se-B (ng/L)
1.4
1.2
1.0
0.8
6000
4000
0.6
0.4
2000
0.2
0.0
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Se-A (ng/L)
0
2000
4000
6000
Ba-A (ng/L)
8000
10000
Sample collection:
November 2004 through December 2007; ~ 500 samples
from 175 rain events at three sites.
Hg concentrations consistent with regional MDN sites.
Generally higher in summer; small-volume winter events
have concentrations up to 55 ng/L.
60
Mercury ng/L
50
Ellyson
Molino
Pace
40
30
20
10
0
Nov
Mar
2005
Jul
Nov
Mar
2006
Jul
Nov
Mar
2007
Jul
Mercury deposition per month:
summertime deposition dominates due mostly to
higher rainfall Hg concentrations.
Monthly Mercury Deposition
Ellyson
Molino
Pace
ng/cm2/week
500
400
300
200
100
0
N-04
M-05
O-05
M-06
O-06
Month beginning Nov 2004
A-07
O-07
Pensacola sites are not significantly different from
each other, and not significantly higher than
regional MDN sites
Rainfall Hg Flux (ng/m^2)
25,000
2005
2006
2007
20,000
15,000
10,000
5,000
LA28
MS22
AL24
AL02
Molino Ellyson
Pace
Plant Crist
Units and In-Service Dates:
94 MW (1959), 94 MW (1961), 370 MW (1970), 578 MW (1973)
The increase in Hg sedimentation since 1965 (-45 years) also coincides with
increased industrialization and coal-fired electricity generation throughout the
southeast, nationally, and globally.
Need better local coring sites to define deposition history over the past 150 years.
Should we expect to be able to measure significant differences
in Hg rainfall deposition due to Plant Crist?
1. “Background” rainfall Hg deposition is already elevated
along the gulf coast (15-20 ug/m2/year).
2. RGM from CFPP may convert to GEM in near-field plume.
3. Individual CFPP impact from RGM emissions typically <15%
of existing rainfall Hg deposition within 50 km in the
southeastern US (Mark Cohen at NOAA/ARL).
So, rainfall Hg deposition may not be significantly elevated
from Plant Crist. What other tools do we have?
Can we use multi-element analysis to “fingerprint”
various sources of mercury and other trace elements in
Pensacola Bay rainfall?
Simple correlation and multi-variate statistical analysis
used to examine relationships among mercury and trace
metals data (Factor Analysis and Positive Matrix
Factorization).
Important to convert to “deposition” (Conc. x Rain depth)
since small volume events have high concentrations of
all tracers and skew regression analysis.
Crustal Factor: alumino-silicate aerosols (mineral dust)
Al, Si, Mn, Fe, Co, REE, Rb, Cs, Th, U
Pensacola Rainfall (3 sites combined)
Pensacola Rainfall (3 sites combined)
10000
Fe deposition (ug/m2)
Si deposition (ug/m2)
100000
10000
1000
100
1000
100
10
10
10
100
1000
Al Deposition (ug/m2)
10000
10
100
1000
Al Deposition (ug/m2)
10000
Cd/Zn Factor: P, Cr, Zn, Cd
Sea Salt Factor: Na, Mg, Sr
Pensacola Rainfall (3 sites combined)
Pensacola Rainfall (3 sites combined)
1000
Sr deposition (ug/m2)
P deposition (ug/m2)
10000
1000
100
10
100
10
1
0.1
1
1
10
100
Zn Deposition (ug/m2)
1000
1
100
10000
Mg Deposition (ug/m2)
1000000
“Pollution” Factor: Hg vs. As, Sn, Se, Sb (volatiles in coal)
Pensacola Rainfall (3 sites combined)
Pensacola Rainfall (3 sites combined)
R2 = 0.27
10
10.00
Sn deposition (ug/m2)
Sb deposition (ug/m2)
R2 = 0.46
1.00
0.10
0.01
0.01
0.10
1.00
1
0.1
0.01
10.00
0.1
Hg Deposition (ug/m2)
Pensacola Rainfall (3 sites combined)
= 0.54
100
Se deposition (ug/m2)
As deposition (ug/m2)
100
10
1
0.1
0.01
0.1
1
Hg Deposition (ug/m2)
10
Hg Deposition (ug/m2)
Pensacola Rainfall (3 sites combined)
R2
1
10
R2 = 0.45
10
1
0.1
0.01
0.01
0.1
1
Hg Deposition (ug/m2)
10
Using volatile trace element concentrations to estimate Hg input
from regional coal combustion (assumes volatile TE comes only
from coal combustion):
1. Use Hg/TE vaporization ratios from CFPPs and “excess” rainfall
TE and Hg deposition (adjust for RGM+Hg-p fraction (74%);
assumes Hg/TE ratio is maintained until deposition):
%Hg from coal =
Annual XS-TE deposition x (Hg/TE)coal x 100
Annual XS-Hg deposition
2. Or, use minimum observed XS-Hg/XS-TE ratios for volatile
elements in rain samples to approximate Hg/TE from coal burning
(average of 10-14 lowest ratios for each element).
%Hg in rainfall from regional coal combustion:
TE
As
Se
Sn
Sb
CFPP Hg/TE ratios
29%
38%
30%
40%
Observed min. Hg/TE ratios
23%
22%
14%
14%
Future Goals:
Use of meteorological data to
understand the chemistry of individual
rain events
43
Tags in Florida, updated
simulation
35
Total Hg deposition (2001)= 25 ug/m2/yr
Wet Deposition (2005-2007)= 14-19 ug/m2/yr
Inferred Dry Deposition = 6-11 ug/m2/yr
78% of total Hg deposition from “background”
7.2% Plant Crist
8.2% from CFPP in Florida
44
REMSAD results courtesy of Dwight Atkinson (EPA) and Tom Myers (ICF)
45
Conclusions and Future research
• Rainfall mercury deposition in the Pensacola Bay watershed is similar
to deposition across the northeastern Gulf of Mexico.
• Factor Analysis and other statistical tests can be used to identify
source “types”, but not individual point sources. Pensacola rainfall Hg
deposition appears to be impacted by coal combustion sources (1440%). Is this true along the entire Gulf coast? Need trace element
analyses at MDN sites. New Gulf Breeze site will help.
• We will conduct detailed meteorological analysis of individual storms
affecting all four monitoring sites within the region (includes
EPRI/OLF site).
– Back trajectories
– Cloud-top heights (indicator of tall convection)
– Prior meteorological history
– Hg isotopes in large-volume samples (Summer 2010)