Slide 1. This (Powerpoint)
presentation was prepared by Cliff Michaelsen, Maine DEP and Rich
Poirot, VT DEC for an Air Quality Conference on "Clean Air
in the Northeast" in St. John, New Brunswick, April 30 -
May 1, 1996. See last page notes for e-mail, snail-mail and phone
connections). Feel free to pass this .ppt file along to anyone
who might be interested.
The original presentation included 3 imbedded Voyager
Movie animations (scotia3.avi, pmozsu92.avi and usflw792.avi )
which (to minimize size requirements) are not included in this
Powerpoint file. The movies may be obtained via the NEARDAT internet
website at: http://capita.wustl.edu/neardat or by contacting Rich
or Cliff (see notes on last page)
These movies may (or may not) play automatically on the Microsoft Media Player. They will play more reliably and controllably on the CAPITA Movie Program (movie4.exe), also available at http:///capita.wustl.edu, under Software Utilities.
Slide 2. Acknowledgments
Many Thanks to the Following Folks For Their Contributions to This Effort:
Paul Wishinski / State of Vermont DEC
Stefan Falke / CAPITA, Washington Univ.
Brett Schichtel / CAPITA, Washington Univ.
Rudy Husar / CAPITA, Washington Univ.
Jeff Brook / AES, Environment Canada
David Waugh / DOE, AESBED
Steve Beauchamp / DOE, AESBED
Ray Hoff / AES, Environment Canada
Slide 3. We examine 3 "transboundary" pollution events, with the following objectives:
These episodes included:
Before we started this effort, we thought:
The Movie4.exe software to play the movie can be found at:
http://capita.wustl.edu/capita/utilities/Movie/Description.html
Description: The animation
shows a dynamic flow simulation of airmasses over North America
during 7/1/1992 - 7/14/1992. The left hand side of the movie contains
the spatial location of all particles emitted at a constant rate
from a spatially uniform grid at the surface layer of the atmosphere.
The particles have been colored based upon their height. The blue
particles are at the surface, the red particles are above 3 km,
and the yellow and green particles are between these heights.
The right hand side displays the vertical position of the particles
for four longitudinal slices across the map view. The vertical
slices are evenly spaced over the map view and are 150km wide.
Purpose: The flow visualization
is created to illustrate the transport phenomena in the lower
atmosphere, such as air stagnation, convergence of airmasses,
fronts, ventilation, and general 3 dimensional motion.
Method: The transport
simulation was created using the CAPITA Monte Carlo Model. Particles
were released from 504 sources uniformly distributed over the
Continental US. Three particles were released from each source
every two hours, and were tracked for seven days. The calculation
of the transport of the particles was done using the wind fields
generated by the National Meteorological Center's Nested Grid
Model with a grid resolution of 180 km.
Interpretation: This movie
displays several of the characteristics of regional scale flow
in the lower atmosphere. First, the daily "breathing"
of the atmospheric due to the diurnal cycling of the mixing height
is clearly seen. The nocturnal mixing layer is low, 100 - 300
m, and the particles released into this layer remain at the surface.
During the day, the mixing height increases, and the particles
are mixed throughout the layer diluting the high concentrations
of particle that accumulated during the night.
Several instances are also apparent where two regional
airmasses converge. In one instance extending from 7/2 to 7/4
an airmass heading southwest from Quebec, Canada and another heading
northeast from the Midwest converge over New York state and Canada.
At the convergence zone, the flow is stagnant and there is a high
density of particles. The converging airmasses cause upward vertical
motion as seen by the large number of particles transported to
heights above 4 km. The simulation also illustrates the ventilating
effects of high wind speeds. During 7/11 meandering winds in the
Northwest and Canada cause a buildup of particles. Around 7/12,
strong westerly winds transport these particles to the east leaving
a low particle density.
Slide 5. Highest airborne
concentrations and deposition rates of sulfur (and other pollutants)
are typically observed downwind of areas of highest emissions.
Note that these NAPAP emissions data are 11 years old, and may
have changed significantly. This was, however, the most recent
emissions data we could find (in a useful format) which included
both US and Canadian emissions data.
Surely we (collectively) can do better than this!
Slide 6. Here, monthly
mean wet sulfate deposition data from the NADP network are compared
with fine particle sulfate data from the IMPROVE and NEPART networks
for the month of July, 1990. Collective US and Canadian efforts
to understand "acid rain" over the past decade have
led to a reasonably firm scientific understanding of the patterns,
origins and effects of sulfur pollution. Recently initiated, bilateral
emissions control programs should result in significant reductions
in sulfur emissions in the next 10 years.
Slide 7. While US (or
North American) patterns of many air pollutants are similar to
those for sulfur deposition or sulfate concentration, other pollutants
exhibit very different spatial patterns. In this case we see monthly
mean Arsenic concentrations from the combined IMPROVE and NEPART
networks for December, 1989. Note that attempted measurements
were conducted at all sites (small white circles), but values
at many sites were below detection limits. Note also, that these
networks include only rural/remote "background" sites,
and that the spatial coverage of measurement sites is sparse in
some regions. The highest concentrations of fine particle arsenic
were observed in an area of the Southwest - particularly at Chiracahua
National Park in New Mexico. High concentrations were also observed
in the Northeast - particularly at Underhill Vermont, on the side
of Mt. Mansfield.
These high Northeastern concentrations were of particular
concern in Vermont - from a human health perspective, as arsenic
is a suspected carcinogen - and also from a regulatory perspective,
as the concentrations exceeded the State's ambient standards for
arsenic at a remote background site. Vermont's hazardous air contaminant
regulations required a determination that the impact from any
new source, combined with background levels, would not exceed
standards.
Slide 8. The temporal
pattern of arsenic at the Vermont site was extremely "episodic".
More than half the attempted measurements were below the detection
limit for arsenic (about 0.2 nanograms/m3). But occasional "spikes"
were recorded at nearly 50 times the MDL level. Relatively high
concurrent levels of arsenic in VT and other northeastern sites
suggested source influence at some distance from the region. The
episodic exposure regime, suggests possible influence of an intermittent
"plume hit" from a distant stationary source (or few
sources).
Slide 9. We recently came
across a small but very useful batch of Canadian data that (someone
who deserves a medal) had submitted to the EPA AIRS database.
From AIRS, it has become quite easy for us to extract any air
quality data in the highly useful "Voyager" format -
using the AIRS/Voyager Data Delivery System, developed at the
Center for Air Pollution Impact and Trend Analysis (CAPITA), with
funding support from EPA OAQPS, NESCAUM and EPA-New England..
Such data can then be explored, merged with other data sets, and
analyzed with the (freeware) Voyager Data Visualization software,
available via Internet at http:03/31/97/capita.wustl.edu under
he "Software Utilities" page. SO2 data from a site in
the northwest corner of Montreal was of particular interest here.
While the timing of VT As and Montreal SO2 was typically offset
by a day or 2, nearly every VT As "hit was accompanied by
a SO2 "hit" in Montreal a day before or after.
So what kinds of large, distant, northern point sources
emit SO2 and arsenic?
Slide 10. Backward air
trajectory calculations using the NOAA HY-SPLIT trajectory model
and NGM meteorological data suggest flows arriving at Underhill,
VT at 3 PM on days with highest Arsenic concentrations had passed
over an area where large Canadian smelters at Noranda and Sudbury
are presumed to emit relatively high levels of Arsenic (and SO2).
For these events, it appeared more likely that Noranda,
rather than Sudbury, was the most important source. Although conceivably
there may be other unconsidered Arsenic sources along this path.
Also, the estimated uncertainty of the trajectory model (20 to
30 % of trajectory distance) may not allow a clear distinction
between the large smelters at this distance and direction.
Slide 11. During a similar
time period, Ray Hoff from environment Canada, AES had also noted
high Arsenic concentrations at several southern Ontario sites.
Ray used the ratio of Arsenic (emitted by both Coal combustion
and Smelters) to Selenium (emitted in much higher quantities by
Coal burning) to focus on the Smelter influences. His back trajectories
(calculated by the AES Trajectory model) tend to implicate Sudbury
rather than Noranda for these high arsenic events. Hum?
Slide 12. We took another
look at our HY-SPLIT trajectories, in this case calculating them
for arrival times at 3 AM (instead of 3 PM) at Underhill, VT on
days when 24 hour levels of Arsenic (and As/Se ratios) were high.
At this point, were pretty confused about whether Noranda or Sudbury
(or both) was (were) the influential source(s).
If only we had some emissions data....
Slide 13. In any event,
there appears to be a happy ending to the Vermont Arsenic story.
Concentrations dropped substantially after the end of 1989, and
have now fallen below the level of Vermont's State toxic standard.
Our understanding (not well confirmed) is that there were major
revisions to control equipment at both Noranda and Sudbury at
about this time period (for SO2/ Acid Rain control purposes).
It would appear that there were also serendipitous
benefits in terms of Arsenic emissions reductions. In any event,
the Northeast States are certainly grateful (although now we aren't
sure who to thank).
Too often, in both countries, we tend to focus on
a single air pollutant at a time - in our air quality measurements,
emissions inventories and control strategies. How many other potential
multi-pollutant effects and opportunities for "serendipitous"
management strategies go unrecognized because we neglect to routinely
exchange and compare data across our common border?
Slide 14. One area of
transboundary data exchange where we have seen some recent progress
is in the form of meteorological calculations routinely calculated
by the Canadian AES and shared with the Northeast States.
Back Trajectories supplied to the State of Maine
by AES Bedford via the internet have been extremely helpful in
assisting the Bureau of Air Quality in generating the smog forecasts
48 hour in advance. 36 hour back trajectories are generated at
the standard 12Z and 0Z time intervals.... and are calculated
for multiple locations throughout the Provinces and the Ozone
Transport Region.
Three levels are calculated from specific locations
at: 1000mb, 925mb, and 850mb's. The Canadian Met Model that drives
these calculations, the RFM model, has been very accurate in identifying
"looping " that occurs off the coast of Maine and transports
the 'Boston Plume' inland.
Slide 15. Routine ozone measurements alone can often provide strong empirical evidence of transport - for example along the coast of Maine during the July 20-22, 1994 episode. Concentrations in excess of the 80 ppb Maine State Standard were recorded at all of these coastal sites, four of which also recorded exceedances of the federal standard. With persistent southwesterly winds throughout the day, the hour of maximum concentration occurs progressively later in the day as one moves to the northeast, ranging from 12 noon at Lynn, MA to as late as 9pm at Jonesport, ME.
It seems that there are times during the ozone season
where one could actually set their watch to the hourly progression
of these transported plumes migrating down the coast of Maine.
The NE-ward movement is not the only force at hand; coastal sea
breezes that seem to elude many of the advanced meteorological
models play a very important role in re-introducing off-shore
plumes into coastal locations in Maine. Many of Maine's most significant
episodes are driven by these E/SE to W/NW flow patterns that become
active in the afternoon when the onshore winds become active.
Note: Several of the following images are reproduced from, and are discussed in additional detail in a NESCAUM report entitled "Preview of 1994 Ozone Precursors in the Northeastern US", available at:
http://capita.wustl.edu/nescaum/Reports/PAMS94/nepams_c.htm
Slide 16. The ozone levels
from these North Atlantic Coastal sites can also displayed as
8-hour running averages and provide a clear impression of ozone
transport along the coast. In addition to the smooth, northeasterly
time progression and broadening of the plume over time, the 8-hour
averages also show the chronic nature of the transported ozone
exposures. An 8-hour standard of 0.09 ppm would have been exceeded
at 5 of these sites; 0.08 ppm would have been exceeded at 7 sites;
and 0.07 ppm exceeded at all of these coastal sites.
Slide 17. When monitored
ozone data from New Brunswick and Nova Scotia are added to this
8 hour moving average, distinct differences are observed. Observed
peaks are considerably smaller in scale at those Canadian Maritime
sites the farther one enters the Bay of Fundy. As has been suggested
by comparisons of acid fog monitoring in the Gulf of Maine with
ozone monitoring in the US and in the Maritimes, ozone is likely
to be precipitated out by the large amount of AM fog that is representative
along the coast of Maine, but of more suspect is the division
and degradation of of the off-shore plumes as they are transported
into the spillway of the St. Lawrence. (Note the Forest Hills
and Fundy Park moving average cycles).
Notice how large sources of NOx produced by utility
stationary sources just to the SW of Pt. Lapreau scavenge ozone
during the time when other stations are monitoring their highest
readings. Utility demands, wind direction and NOx output are consistent
with this assumption. At the farthest downwind site, North Cape
PEI, long range transport from the entire Eastern Seaboard allows
O3 readings to peak very late at night.
The diurnal pattern of Halifax more closely resembles
its counterparts in Maine.
Slide 18. At the moment,
there is no easy way for Canadian and US scientists to examine
combined US/Canadian air quality data. US analysis stops at the
US border and vice versa. Information collected by both governments
fall into their respective repositories, and to date, only an
exhaustive procedure involving numerous manipulations of several
data sets with multiple software applications is available to
view the collective monitored data.
There are many significant reasons to establish a mechanism that allows the sharing of this wealth of isolated information. Here are three:
Each government is in the process of updating their
main data archive center. It is hoped that a more user-friendly
system can be created to utilize the monitored data that is collected,
or only half of the story will be told.
Vector wind speeds and directions are are run in
concert with color segregated ground level ozone readings (from
Long Island Sound to Prince Edward Island) and at the bottom,
a profile of the five day air quality episode.
As the movie is run in animation or single step mode,
concentrations are seen to maximize on the 21st of July, 1994.
Values indicated by the Scotia Prince three hours off-shore the
coast of Maine exceed the NAAQS standard of .124 ppm as the boat
literally runs into a plume (which then migrates toward the coast,
mixes with fresh continental emissions/ozone, and cumulatively
impacts inland air quality).
Slide 20. Originally,
we thought that by analyzing the 36 hour back trajectory (created
with NOAA's HY-SPLIT model), we would be able to ascertain the
source origin of this July 21st episode. Both coastal locations
in Maine were highly impacted by transport up the eastern seaboard,
while North Cape PEI, with it's signature of long range night-time
transport, derives its origins from a more predominantly westerly
flow.
Apparently this only half of the story
Slide 21. By creating
the 3 PM back trajectory for July 20th, 1994, there is the strong
indication that the magnitude of the episode on the 21st may have
been amplified by synoptic-scale "background" contributions
from more westerly transport on the preceding day.
Note: PM-10 Sulfate concentrations are not routinely
measured at all sites. During this time period they are limited
primarily to the States of Illinois, New York, Vermont New Hampshire
and Maine.
Slide 23. The intervening
time period - August 22 to 28, 1992 represents a slowly moving
episode of multi-pollutant transport from the Midwestern US into
the Northeast States and Eastern Canadian Provinces.
Slide 24. The temporal
pattern of maximum 24-hour moving average ozone across the northern
US shows a peak at Toledo Ohio on the evening of August 22. Erie,
PA peaks on the following day. Underhill, VT peaks 2 days later
on August 25. While the most northeasterly site - Jonesport, ME
peaks on August 27.
Slide 25. A similar West
to East progression can be observed in fine particle (< 2.5
microns) mass and sulfate measurements provided by Jeff Brook
of Environment Canada, AES.
Note that the highest fine particle mass concentrations
are in excess of 50 micrograms per cubic meter - a level recently
proposed by the US EPA as a new 24-hour National Ambient Air Quality
Standard for Fine Particle Mass Concentration. Note also that
half or more of the fine mass concentrations are typically composed
of sulfate during peak days of this episode. At this point, we
appear to have a relatively clear-cut example of a classic ozone
and sulfate transport event from the Midwestern US.
Note however, that fine mass and sulfate levels peak
at the most easterly Kejimkujik, Nova Scotia site on August 25
and 26. Keji fine mass concentrations drop considerably on August
27 - the day when ozone peaked at the (not-so-far-away) Jonesport
Maine site.
Slide 26. Jeff Brook's
measurements also include concentrations of strong acidity associated
with the fine particle concentrations. Note that very high acidity
levels were observed at the remote Keji site on August 25 and
26, even though sulfate levels were only moderately high (about
15 micrograms per cubic meter) at Keji on these dates.
The relatively high ratio of acidity to sulfate suggests
a relatively low degree of neutralization (by ammonium) of the
sulfuric acid aerosol. Since ammonia emissions are relatively
ubiquitous over land, a poorly neutralized sulfate aerosol is
not consistent with the concept of very long-range transport of
acidity (for example from the Midwestern US to Maritime Canada).
Slide 27. The backward
HY-SPLIT trajectories displayed here are calculated (provided
by Paul Wishinski, VT DEC) for 3 PM arrival times at several of
Jeff Brooks sites on the days when peak fine mass and sulfate
concentrations were recorded at each site. For Windsor, Toronto
and Montreal, their highest concentrations are associated with
(markedly similar) flows which had previously circulated over
an area of the eastern US including both the Central Atlantic
Coast and Ohio River Valley.
The Kejimkujic trajectory on 8/26/92 is more easterly
and has not passed over the high sulfur-emitting areas of the
Midwestern US. For more than a day prior to arrival at Keji, the
trajectory traveled over water (where sources of sulfur and ammonia
are minimal. Perhaps these factors can help account for the unique
combination of relatively lower sulfate, but higher acidity at
Keji compared to the other sites.
Slide 28. By August 27
- when fine mass and acidity levels have dropped substantially
at Jeff Brook's monitoring sites - the trajectory flows to these
Canadian sites indicate strong ventilation from the North.
Yet just to the South, Ozone, PM-10 and Sulfate are
approaching their maximum levels in Northern New England.
Slide 29. Limited US fine
particle mass and sulfate data are available from several Northeastern
US sites (from the NESCAUM NEPART network) on August 26 and 28.
While concentrations are much higher than average at these sites
on August 26, there are much higher yet on August 28. A peak fine
mass concentration of greater than 80 micrograms per cubic meter
(half of which was sulfate) was observed at the remote Quabbin,
Massachusetts on the 28th!
Slide 30. Back trajectories
for several of the NEPART sites on 8/28/92 were provided by Bret
Schichtel of CAPITA using the CAPITA Monte Carlo model. The NEPART
network did not measure aerosol acidity directly; however an indirect
estimate of acidity (shown as black bars) is occasionally possible
through comparison of elemental sulfur and hydrogen measurements
from the NEPART filters. The estimated acidity levels at several
NEPART sites on 8/28/92 were the highest ever observed in the
1989-93 period of NEPART operation.
Unlike the southern Canadian sites, New England was
still experiencing flows fro the Midwestern US at this time. Note
however that the (blue) trajectories - arriving at the Quabbin,
MA site (with maximum fine mass sulfate and acidity) have passed
over Midwestern source areas, but have also stagnated more recently
over over Northeast coast metropolitan areas. Here, as with the
Kejimkujic site a few days earlier, we appear to have a combined
influence from distant and relatively nearby sources.
Effective management of transboundary air resources
requires a better understanding of the complex influences of local,
regional and more distant sources. This understanding and subsequent
development of management strategies will continue to be compromised
so long as our data (unlike our pollution) does not flow freely
across our common border.
Slide 31. We Examined 3 Episodes of Transboundary
Pollution Transport
We Assumed We Understood These Episodes in Advance.
Assumed it would be Easy To Merge US & Canadian
Data.
Both Assumptions Were Wrong!
We Efficiently Exchange Air Pollution Across our
Border, Where it is Merged with Local Source Influences, but is
Too Often Unobserved, Undocumented, and Uncontrolled.
We Need More Efficient Mechanisms to Exchange and
Merge our Collective Air Quality Data!
For additional information, please feel free to contact:
Rich Poirot (richp@qtm.anr.state.vt.us)
Vermont Department of Environmental Conservation
Air Pollution Control Division
Building 3 South
103 South Main Street
Waterbury, VT 05671-0402
phone: (802) 241-3840
or
Cliff Michaelsen (Clifton.W.Michaelse@STATE.ME.US)
Maine Department of Environmental Protection
Bureau of Air Quality
State House Station #17
Augusta, ME 04333
phone: (207) 287-2437