Introduction To The IMPROVE Data
The data on fine aerosol generated by the IMPROVE network
over the entire U.S. during nearly two decades represent a potential bonanza of
information. I have recently decided to use a graphical approach to investigate
the large-scale sources of this aerosol, and the initial results are very
promising. The initial goal was to see just how much crustal aerosol was
actually reaching the West Coast of North America from Asia (or better stated,
from over the Pacific Ocean). This goal quickly broadened to include pollution
aerosol from the Pacific, crustal and pollution aerosol coming down from the
Arctic, marine and crustal aerosol coming eastward from the Atlantic and Africa,
and pollution aerosol moving northward from the Southeast into the
Northeast.
The main graphical approach shown here is to plot selected
elemental concentrations and ratios from a site against Julian date (called DOY
by IMPROVE), with all years superimposed. "All years" can range from
one or two for the newer sites to 15 or so for the older sites. These plots have
revealed a wealth of information, most of which is not yet amenable to the use
of conventional statistical techniques. To display these plots most efficiently,
we have taken the base map from IMPROVE's web site and fixed it so that the plot
for elements or ratios at a given site can be displayed (superimposed in the
base map) simply by clicking on the point for that site. We are developing
separate slide shows for several elements and ratios in PowerPoint and
converting them to HTML form for use here. [Warning--it appears that Version 6 of MicroSoft's Internet Explorer is required to view these slide shows
properly. Netscape at least through v. 6 does not work properly. Sorry!]
To view a show, first click on its link below. You will be
able to work it just like a PowerPoint show, but without having to load the
whole thing to your site (you get each new slide in turn from my site). You can
even click on the lower right and expand the slides to full screen. You start
with the first slide, which is a reference map. Click on any of the green sites
to get its plot vs. DOY, then click again anywhere on that plot to return to the
reference map. The only exception to this rule is for Hawaii, where repeated
clicking cycles through the three Hawaiian sites of Mauna Loa, Haleakala, and
Hawaiian Volcanoes before returning to the base map. When one of these sites is
used as reference in a series of slides, it is omitted from the Hawaiian
sequence.
To learn the most from these slides, you must take your time
and look through all the sites with green dots. (The sites with red dots do not
have enough data to plot.) I recommend working your way around the US and noting
the spatial groupings of the temporal patterns. Given the large number of green
sites (about 90) and the large mass of data per site, this job may take many
hours. But the rewards should be worth the effort. Please feel free to write to
me at krahn@uri.edu with questions or
comments. I will post all appropriate comments for all to see. At the beginning
I am particularly interested in any difficulties anyone has with displaying the
maps with the embedded plots.
I first discussed the idea for these slide shows with Ms.
Jinghua Guo about a month ago. Just before she returned to her Ph.D. studies at
Beijing Normal University in late August of 2002, she created the basic
PowerPoint structure for the shows, which was a very creative achievement for
such a short period. I then added the plots vs. DOY and transformed the shows
into HTML. The basic idea seems to be working every bit as well as we had hoped
it would.
Some editorial comments before the analysis: The IMPROVE
network standardized in fine aerosol because it was primarily concerned with
reductions in visibility. That is all well and good, for fine aerosol degrades
visibility far more per unit mass than coarse aerosol does. In most situations
that means that fine aerosol degrades visibility more than coarse aerosol does,
for the mass of coarse aerosol is usually the same or less than that of fine
aerosol. But more importantly, fine aerosol is not *the* aerosol. That term is
reserved for total aerosol. Near the sea and near arid regions, the
concentrations of coarse-particle elements like Na (sea plus deserts) and Si
(deserts) may be up to an order-of-magnitude higher in the coarse range than in
the fine range. IMPROVE's samples will miss this important component. To be
sure, the process of aging works primarily on the coarse fraction, so that
aerosols far from their sources (in places such as elevated locations and the
polar regions) tend toward the fine fraction as a natural limit, but this does
not nullify the major point that the IMPROVE samples are capturing only part of
the true aerosol.
We should also not forget that the IMPROVE data, with their
focus on the long-lived, fine aerosol, will tend to magnify the effects of
distant sources over that shown by the total aerosol. As above, that does not
nullify the conclusions for the fine component, but it creates an impression of
longevity and transport that is not shared by the total aerosol. The effects of
transport shown by IMPROVE hold for its kind of aerosol only.
A personal note as the last remark before the analysis. This
is my first large-scale experience with fine aerosol. For 30 years I have held
that the fundamental aerosol is the total aerosol, or TSP. The IMPROVE data have
given me no reason to alter that view. But I must admit that there is a great
deal of usefulness in revealing loud and clear the long-range transport in the
way that IMPROVE does. I have found this experience refreshing and invigorating,
and it ain't over yet.
Summary of basic findings
- Most of the western United States is exposed to pollution aerosol from
over the Pacific Ocean during winter and spring. I am reluctant to call it a
plume because it can be seen extending from the northernmost to the
southernmost portions. Perhaps "air-mass effect" would be better.
It peaks during March, April, and May, and drops off rapidly after that. Its
maximum concentrations of Pb are 2 to 3 ng m-3, or 1000 times below the
U.S. standard. It
may be higher before the peak than after it. The Pacific Pb becomes undetectable as it crosses the
Rockies and enters regions with higher local backgrounds. This makes it
difficult to determine how much it is diluted and removed versus how much it is just
masked.
- A spring peak of Pb is also seen in the East wherever backgrounds are low
enough. There are indications of something similar as far away as southern
Florida and the Virgin Islands. These peaks are likely more than just transported
Pacific aerosol.
- As indicated by Pb, concentrations in the spring peak are low enough to be only a
curiosity, albeit one with high tutorial value.
- Zn also has a spring peak in the West, with the same timing and
concentrations as Pb. Its background is higher and noisier, however, which
tends to obscure the peak at more sites. Zn also comes to the north-central
states from the north during winter, and is enriched in that aerosol by a
factor of two relative to the Pacific aerosol. Zn has summer maxima in the
West that remind one of S.
- At some sites there are also indications of a fall peak in pollution, but
considerably weaker than the spring peak. I am reminded of a similar
situation that Suilou Huang and I found some years ago for Bermuda (regular
spring peak in pollution, plus weaker fall peak).
- A spring peak is also found for S. It has the same timing as for Pb, but
is harder to discern because of its higher background. It stands out best in
and around Idaho. Its maximum concentrations are 0.2
to 0.4 µg m-3.
- Although it is tempting to call the spring peak Asian, it may well be
more than that, for it is found over a very long north-south transect. It
may be a large-scale
effect of the cold, dry northern air mass of winter, a sort of
generalization of Arctic haze.
- Along with the pollution from the Pacific comes crustal aerosol. Much of
this is probably just the crustal component of polluted air masses, but in
spring it appears to be supplemented by the same additional crustal material
seen, say, in and around Beijing in spring. This material appears to have
two general sources there, dust storms and general dustiness.
- At sites in western North America it is very hard to distinguish the
transported (Pacific) crustal material from local crustal material, for they
both peak in spring. In fact, crustal aerosol over much of the world peaks
in spring. I am provisionally using two criteria, both of which must be
satisfied in order to call the crustal material Pacific: (1) The
concentrations must be right, where "right" is defined as those
found over the Pacific (Hawaii, in practical terms); and (2) The
compositions must also be right (same definition as for the first
criterion). Together, the two tests seem to show that much of the crustal
material of the higher elevations in the West is Pacific during winter and
early spring, but local in summer and fall. Concentrations of transported Si
are usually 0.1 to 0.6 µg m-3, but can occasionally reach or
exceed 1 µg m-3. That corresponds roughly to 0.5 to 5 µg m-3
of transported crustal aerosol. (By contrast, fine crustal aerosol in a big
dust storm in Beijing can be 6 to 15 mg m-3, or 4000 times
greater.) The sources switch quite
rapidly between transported and local modes. (A similar seasonal relation is shown for
pollution aerosol by Pb, although the switch between modes is more
subtle.) Local concentrations generally exceed transported
concentrations, and sometimes by large factors. The best indicator of
crustal concentration is Si, while the best test of crustal composition is
Ca/Si.
- The north-central U.S. is swept over by aerosol from the north during
winter. This aerosol contains both pollution and crustal components, and
occasionally a marine component as well (indicated by Na/Si). The best
indicator for the pollution component seems to be S. I suspect that this
northern aerosol is just diluted and transported Arctic haze, which is known
for its high relative concentrations of S, particularly during spring. This
northern aerosol can be seen extending from the eastern edge of the Rockies to the
northern East Coast (Acadia, Maine, in particular), although it seems to
peak around Minnesota. This finding confirms what I had suspected about
Arctic haze for a very long time, but had otherwise been unable to document.
- Crustal aerosol from the Sahara sweeps over the Southwest, the Southeast,
and the Northeast during summer. It is shown by high Si and by low Ca/Si. As
one might expect, the effect is most pronounced in the Southeast, where it
can be dramatic. I was surprised to find it extending clearly into Texas,
and apparently also into New Mexico and Arizona, although more weakly. Near
the coast, the appearance of high Si is accompanied by a sudden drop in
pollution elements, which shows that the Saharan aerosol is embedded in
clean Atlantic air.
- Pulses of summertime S can be seen emanating from broad sources in the
Southeast and mid-Atlantic states (up to the Ohio River Valley) and reaching points throughout the
Northeast.
- Even though marine aerosol is largely restricted to the coarse particles,
it can be found at coastal and near-coastal sites. In general, marine
aerosol in the U.S. forms a U-shaped pattern, with higher amounts along the
coasts and lower amounts inland. But residue of marine aerosol from the
Pacific can be seen penetrating well into the West, in some cases into Idaho
and Colorado. Marine aerosol from the Atlantic is generally restricted to
areas closer to the coast, but can reach well into the Southeast. One surprising result was
the marine aerosol found along the northern tier of states (the
north-central states) during winter. It appears to be associated with the
transported Arctic haze found in the same area at the same time. Another
surprise was the marine aerosol at Denali, at the base of Mt. McKinley in the middle of
Alaska. But it was clearly to be seen, in association with the northern
aerosol of winter. Marine aerosol is also detectable at the two lower
Hawaiian sites (Haleakala and Hawaiian Volcanoes, both at about 1500 m). There are
indications of it at the much higher Mauna Loa, but at much lower concentrations.
- The plots reveal analytical problems with some of the elements,
particularly those at lower concentrations. The problems appear as noise,
outliers, and too-high concentrations. The latter case is best displayed in
X-Y scatter plots or in plots of ratios vs. DOY. It is a pity that the PIXE
was not more sensitive. For example, even the low concentrations of Fe seem
to be artifactually high at many of the sites, and a major element like Ti
is nearly useless.
The following slide shows are discussed on separate pages.
Pb
Pb vs. Pb at Haleakala, HI
Zn
Zn vs. Zn at Haleakala
Pb/Zn
S
S vs. S at Mauna Loa
Si
Si vs. Si at Haleakala
Ca/Si
Ca/Si vs. Ca/Si at Mauna Loa
Spatial patterns of marine aerosol
Temporal patterns of marine aerosol
You can jump directly to the PowerPoint slide shows (web version) from the
links below. Note: You do NOT need PowerPoint to do this. It is all done through
HTML and Internet Explorer.
Pb
Pb vs. Pb at Haleakala, HI
Zn
Zn vs. Zn at Haleakala
Pb/Zn
S
S vs. S at Mauna Loa
Si
Si vs. Si at Haleakala
Ca/Si
Ca/Si vs. Ca/Si at Mauna Loa
Ca/Si vs. Na/Si (Effect of
marine aerosol on Ca/Si)
Back to Home Page