Comments on 'What is the origin of high ozone concentrations?'

Back Trajectories Don't Indicate Origin of Emissions Causing High Ozone

Comparing the two plots, it is obvious that the average trajectory length to the 22 stations is much longer for those that start in Canada. Longer trajectories equal higher wind speeds. All other factors being equal, concentration is inversly dependent on wind speed. Double the wind and you halve the concentration. A further problem with the interpretation given is that other meteorological conditions associated with high ozone tend to have lower wind speeds, amplifying the inverse corellation.


Submitted by Robert Imhoff on 12/17/96 RecID: RobertImhoff

Re: O3 Lifetime from UAM

Robert, hope I am not disrupting this nice dialog you have with Rich & Paul on the interpretation of trajectories with a side question to about the UAM model-derived lifetime.

First on the lifetime definition. Evidently there are two definitions here: your lifetime starts when the ozone is formed. For impact calculations, we have been including the formation time as well. The two differ quite a bit since half of the day is night, and at night the chemistry is almost dead but the transport takes off. Even if we say that the ozone forms instantly during the day, we have to add about 12H to the yet unborn ozone. So, the 12H incubation time at night + your 12-18H mature ozone time adds up to 24-30H of 'lifetime' as we think of it. This is pretty close to the day and a half people, including us here, have been using for ozone lifetime. The actual calculation is a bit more tricky, but the above illustrates my point. What do you think?

Now some nit-picking. Is it fair to say that the formed, mature afternoon ozone will have shorter lifetime than any ozone from any other time? [Probably not]. For instance, part of the morning ozone will survive the day as ozone and enters the long night’s journey virtually unchanged. My actual point is that it is worth thinking of chemistry, throughout the 24 hours, not only when we see the action taking place. The same holds for transport where we know for a fact that night is the time of real horizontal transport.

Another thing. It is great that more stuff is trickling out of the modeling work than the scenario alphabet soup. Your lifetime estimates, the region of influence estimates of STRao & JJensen are model outputs that are in the same language as the air quality work. I consider that part of our (AQWG) role is to provide constructive support to modeling group. However, we on the AQ side can give very little support to the modeling group on scenario runs, except for the base-case. I know that all those scenarios runs are probably needed, but I think that the policy folks (or for that matter any other thinkers) would benefit equally from more general, reusable knowledge that will be usable even after OTAG. So, all of you AQ-minded ‘modelers’, keep it up. Uuups, did I call you a modeler :)?

One last thing (boy, am I picking out Rich Poirot’s habits here?) any ideas on how to open up some kind of dialog with the modeling group? Like, have someone from the group jumping in here? I have a few questions that would be nice to discuss. Maybe you could even volunteer as spokes-person for the group? Would Iron Mike let you? Now I know I am in BIG trouble! :).



Submitted by Rudolf Husar on 1/5/97 RecID: RudolfHusar

---

Re: Back Trajectories Don't Indicate Origin of Emissions Causing High Ozone

Robert Imhoff's preceding comment suggests that the referenced residence-time probability plots provide no information on ozone source regions, and only indicate areas characterized by high (left) and low (right) wind speeds. Although we disagree with this interpretation (for reasons discussed below), we do feel that Robert has raised an important point that has not yet been (but should be) articulated in an OTAG Air Quality Analysis summary. Specifically, some degree of local stagnation over source areas is an important prerequisite to the accumulation of high ozone concentrations which may then be subject to subsequent transport to downwind receptor areas (although more rapid nocturnal transport of poorly mixed, un-reacted, elevated point source NOx emissions is an exception).

A variety of OTAG AQA analyses, including Source Regions of Influence (Schichtel and Husar), backward trajectory analyses and time-lagged intra-regional ozone correlations (Poirot and Wishinski), and various SOS studies (Meagher, Edgerton) illuminate the importance of local stagnation for accumulation of high ozone concentrations in the Southern and/or Midwestern sections of the OTAG domain. This geographical pattern of stagnation potential along the western edge of the Appalachian Mountains is hardly a new discovery, and was characterized in detail during the 1960s and 70s by Hosler (1961), Korshover (1967), Holzworth (1964, 1967, 1971, 1972) Gross (1970) and others, based on long-term (5 to 30 year) climatological evaluations of surface and upper level wind speeds, mixing depth, prevalence and duration of subsiding high pressure systems and/or predicted and observed episodes of various air pollutants. It would be informative to include some references to these earlier climatological evaluations in the OTAG AQA summary. As Robert Imhoff has observed, " all other factors being equal, concentration is inversely dependent on wind speed. Double the wind speed and you halve the concentration" (and vice versa). Consequently an Eastern US ozone control strategy which seeks to minimize the influence of ozone transport to areas throughout OTAG should consider some focus on sources located in areas most prone to local stagnation (for example the "TVA area"), as these areas are chronically prone to accumulation of high concentrations necessary for subsequent transport to downwind receptors.

All other factors being equal, emissions in areas central to any geographic domain have the greatest potential to contribute to ozone concentrations over the broadest range of receptor locations throughout the domain. Of course, all other factors are not equal in the Eastern US. Emission densities are not uniform, and for any given combination of mixing height and wind speed, reducing the emissions will also reduce the concentration (and vice versa). The industrial Midwest - Ohio & Tennessee River Valley region - is spatially large, compared to any individual urban area, and characterized by central location, frequent/persistent stagnation and very high NOx emission densities. This is the "source area" most persistently "upwind" of highest ozone at many receptors by the trajectory residence-time analysis (and by several other OTAG AQA analyses).

It is logical that low ozone concentrations will often tend to be associated with good ventilation and rapidly moving air parcels (see Schichtel and Husar: "Source Regions of influence"). Air masses moving into the Eastern United States from the Canadian plains are especially likely to be moving rapidly (and to have passed for considerable distance over areas of low emissions density). The low ozone probability plot also includes offshore areas all along the north, central and south Atlantic and Gulf Coasts, which are also obviously areas of low emissions density, but which are not obviously areas of high wind speeds. It is likely that air masses moving to inland receptors from these marine areas will often be associated with clouds, fog or precipitation. It is illogical to assume that these Canadian and marine areas of minimal emissions density are the "true" ozone source areas, but that their influence is somehow masked by meteorological conditions non-conducive to ozone accumulation. Nor is it logical that high emission areas are "unfairly" implicated if they also tend to be subject to frequent and persistent local stagnation.

We certainly don't believe that low wind speed provides a justification for high emissions, or a distortion of the perceived influence of emissions contributing to high ozone concentrations. We don't believe much information on trajectory speed is "obvious" from simply eye balling the referenced incremental probability plots, but will attempt a further evaluation space/time characteristics of trajectory velocity in association with high and low ozone levels, and will report results back to the AGA website. Meanwhile, we suggest some overview of the above-referenced climatological studies would be a useful section to include in the AQA summary.

Rich Poirot and Paul Wishinski, VT DEC

Submitted by Richard Poirot on 12/31/96 RecID: RichardPoirot

Re: Re: Back Trajectories Don't Indicate Origin of Emissions Causing High Ozone

Summary.
Short trajectories mean lower wind speeds near the receptor. This increases the potential for ozone accumulation near the receptor. Low wind speeds are also correlated with other parameters which are associated with high ozone concentrations.

Details.
The short trajectories are markers of stagnation relatively near the point of ozone impact, not of
stagnation farther upwind. Since ozone has a finite lifetime in the atmosphere, the farther upwind in time that ozone is formed, the less proportional impact that that ozone will have on the downwind receptor. UAM-V model estimates for ozone lifetime in the atmosphere (for ozone existing at midafternoon) are between 12 and 18 hours. Most of the ozone formed near the beginning of a 3-day trajectory has already been deposited or reacted with VOC’s. Even if areas at the start of a 3-day trajectory have experienced high ozone concentrations due to stagnation conditions, little will remain. With a short path length, this implies that areas not that far from the receptor have relatively little impact and that most of the ozone is formed near the receptor. Thus the accumulation of most concern is that near the region of impact and not that which occurred far away. Formation of high ozone concentrations upwind does not necessarily imply significant impact at a far downwind location. What we have observed from reactive-tracer simulations with the UAM-V model is the following rule of thumb:

If transported ozone concentrations are high, they have not moved far. If transported ozone concentrations have moved far, they are not high.

The lower wind speeds implicit in the shorter trajectories may also be a marker for a
covarying parameter, such as temperature. Statistical analyses done for several years
of ozone and meteorological data for Atlanta (1992)and Nashville (1994) have shown that
the highest correlation with peak 1-hour ozone concentration is for temperature. In fact, quite good estimates of peak ozone could be made with this parameter alone. Wind speed was inversely correlated with temperature. Increasing temperature, both in the BEIS-2 and real world, promotes increased emissions of biogenic VOC’s, notably isoprene.

Wind directions from the Canadian and maritime areas may be associated with meteorological patterns which are not conducive to ozone formation near the receptor. In the presentation by Rudy Husar at the OTAG meeting in Cherry Hill N.J., he was able to demonstrate a strong link between meteorological patterns and transport direction. I certainly do not mean to imply that high ozone concentrations form over the Atlantic or over Canada, but that these concentrations are masked by meteorological conditions. Rich Poirot is correct that this would be illogical. I do think that a further investigation is warranted of the linkage between meteorological patterns and transport direction.


Submitted by Robert Imhoff on 1/3/97 RecID: RobertImhoff1

Re: Re: Re: Back Trajectories Don't Indicate Origin of Emissions Causing High Ozone

As Rich has suggested in his previous response to the comment, we will be presenting summary plots of average "wind speed" implicit in the time trajectories have spent over each of the grid squares. The spatial plots of average "wind speed" over the gridded area for the entire 7 summer time period or subsets of the entire period should resolve the issue of what the relative pattern of higher or lower "wind speeds" looks like (based on the NGM met fields used to develop the back-trajectories).
Robert Imhoff's comments are well taken, but I must infer that there is a mis-interpretation of what the referenced incremental probability plots are capable of showing with respect to "wind speed". There is no data presented on the plots which allows any estimate of "wind speed" anywhere in the domain, except if one assumes that each and every back-trajectory of which the ensemble plot is composed represents the same time period from beginning to end. Nominally, back-trajectories were computed back 108 hrs.
However, as pointed out in previous presentations of this work, trajectories run off the edges of the domain or occassionly have been terminated early due to missing met data. The pattern depicted in the ensemble plot does not retain information relative to the implicit speed along the path of the back-trajectory into the monitor location. This point is made only for clarification of what may be concluded about "wind speed" by simply eye-balling the plots themselves.

In order to represent a "spatial distribution" or "pattern" of average wind speeds implied by the individual back-trajectories from which the ensemble plots were derived, it is necessary to save the wind speeds implied by the time and distance relationship of trajectory segments which are over the individual grid squares. This is currently being done and as indicated plots will be forthcoming.

We agree that low wind speeds near a receptor would increase the potential for ozone accumulation near the receptor (accumulation from local sources if any exist near the receptor). I am not sure I fully agree that "short trajectories mean lower wind speeds near a receptor", as the commentor has indicated. Short trajectories may be short because they have only been followed for a short distance even though the implicit "wind speed" along the trajectory may be very high, there is no way of knowing the wind speed unless something in addition to the trajectory length is given.
If one accepts the proposition that ozone and ozone precursors can be transported (above the mixed layer at night for instance) then lower wind speeds in the vicinity of large sources of ozone precursors also allow more of these precursors and the ozone formed in daylight hours to be injected into air masses that eventually arrive at receptors which are not necessarily near the location of the stagnation (low wind speed) condition.



Submitted by Paul Wishinski on 1/3/97 RecID: PaulWishinski

---

---

---

---