*UPDATE* Computer Models, Actual Data, and Smog
UPDATE (11/5/2012, 12:02pm ET): A piece in the Austin American-Statesman also examines the flaws in Olaguer’s report, likening it to a person’s doctor suggesting he or she is suffering from health problems despite medical tests showing the opposite — all based on a computer simulation using actuarial data and medical histories. The author then asks, “Wouldn’t you want a second opinion from someone who knows the difference between a computer model and the real world?” The piece adds an important conclusion to the whole debate about supposed smog-impacts from any process: “Citizens legitimately concerned about healthy air should ask for the real-world facts — facts more accurately revealed by monitored measurement of ambient conditions.”
—Original post, September 5, 2012—
A new study by Eduardo B. Olaguer of the Houston Advanced Research Center (HARC) purports to show that emissions from oil and gas operations will prevent nearby metropolitan areas — particularly Dallas-Fort Worth — from meeting federal ozone standards. Fortunately for people in the Metroplex, empirical data shows that such a conclusion is simply unsupported by the facts.
It’s worth noting, however, that the “ozone/smog issue” is a central plank in the anti-shale crowd’s agenda (part of its gradual shift of emphasis toward air quality). It began in earnest when Al “crucify them” Armendariz released a report in 2009 alleging that emissions of ozone precursors from oil and gas operations in the Barnett Shale were more than twice those emanating from mobile sources (i.e. cars and trucks). Given Dallas-Fort Worth’s well-known air quality problems, the study was cited far and wide.
The problem with Armendariz’s study — and other recent research suggesting harmful emissions from shale development — is that it relied on a modeling exercise that extrapolated an outlier of data into a broader trend. Proof of that has come from the Texas Commission on Environmental Quality (TCEQ), which has used state-of-the-art air quality monitors in the region to conclude not only that emissions from mobile sources are a larger contributor to smog, but that oil and gas operations have a minimal-at-worst impact (more on that below).
Against that backdrop, it’s unsurprising that the HARC study — distributed to the press by none other than Downwinders at Risk, one of the most vocal anti-drilling organizations in the area — relied on modeling in lieu of observable data to reach its conclusions, which were wholly predictable based upon the report’s stated objectives.
Below is just a sampling of the errors and oversights found in the HARC study.
Olaguer: “We used a neighborhood scale (200 m horizontal resolution) three-dimensional (3D) air dispersion model with an appropriate chemical mechanism to simulate ozone formation in the vicinity of a hypothetical natural gas processing facility, based on accepted estimates of both regular and nonroutine emissions.” (HARC study, p. 966; emphasis added)
FACT: To summarize, the author himself has described this report as a modeling exercise that simulates ozone formation from a hypothetical facility. It’s not data collected from actual air monitors, nor is it an analysis of existing infrastructure — exactly the kind of work that state regulators, most notably TCEQ, have been doing for years. In fact, available data from TCEQ directly rebuts claims about high emissions levels from oil and gas activity in the Barnett Shale. Here’s what TCEQ Chairman Bryan Shaw said about DFW area emissions:
“After several months of operation, state-of-the-art, 24-hour air monitors in the Barnett Shale area are showing no levels of concern for any chemicals. This reinforces our conclusion that there are no immediate health concerns from air quality in the area, and that when they are properly managed and maintained, oil and gas operations do not cause harmful excess air emissions.” (emphasis added)
This kind of definitive conclusion is probably why opponents have to rely on hypotheticals and modeling exercises; reality simply doesn’t support their thesis.
Olaguer: “Using these methods, Armendariz (2009) estimated peak summer emissions of ozone precursors in 2009 from all oil and gas sources in the Barnett Shale to be 307 tons per day (tpd). By comparison, he estimated on-road mobile emissions from the five counties in the DFW ozone nonattainment area with significant oil and gas production in 2009 to be 121 tpd.” (HARC study, p. 967)
FACT: If the name Armendariz sounds familiar, it should. But it’s also worth noting that it was Armendariz’s study that gave anti-drilling activists a self-proclaimed license to claim Barnett Shale development is worse for DFW area ozone than all the cars and trucks on the road. The only problem, of course, is that air sampling data debunked his modeled scenario.
TCEQ data shows quite clearly that emissions of volatile organic compounds (VOCs) from oil and gas operations are considerably less than those from mobile sources. In fact, VOCs — which Armendariz and others continue to emphasize — aren’t even the proper focus in assessing ozone formation. Instead, emissions of nitrogen oxides (NOx) are a much more significant source, and TCEQ has found that NOx emissions from mobile sources are “approximately 15 times higher” than those from oil and gas development.
As the Barnett Shale Energy Education Council (BSEEC) has pointed out: “Even if all man-made VOCs were eliminated in this region, there are more than enough naturally occurring VOCs from trees, plants and other organic sources to combine with NOx to cause smog.”
To be fair, Olaguer listed Armendariz’s paper in the context of other available research on this subject. But he also neglected to mention what TCEQ — through actual air monitors — has determined with respect to smog precursors in the region, and how that real-world data rebuts Armendariz’s computer model. Whether deliberate or accidental, that’s an omission that unfortunately speaks volumes.
Olaguer: “Figure 1 shows 1-hr ambient monitoring data collected at a pipeline compressor station in Lake Arlington, Fort Worth, during the BSEEC study. Note the very large short-term concentrations of formaldehyde (HCHO) approaching or exceeding 100 ppb around the site. Such large concentrations may be cause for concern, not only because of short-term health impacts such as nosebleeds, vomiting, and skin irritation, but also because of formaldehyde’s capacity to release radicals and thus contribute to rapid ozone formation (Olaguer et al., 2009).” (HARC study, p. 967; emphasis added)
FACT: Large concentrations of formaldehyde could indeed be cause for concern, so let’s see what the above-mentioned study (conducted by TITAN Engineering) actually said about those HCHO concentrations:
“Formaldehyde concentrations exceeded the short-term AMCV in several samples at the Lake Arlington compressor station. TITAN determined that formaldehyde at this site originated from off-site sources.” (emphasis added)
This is important for a couple of reasons. First, Olaguer is clearly misrepresenting the findings in that study by suggesting the concentrations are from the compressor station analyzed, even though the authors of the study came to a completely different conclusion. Second, Olaguer concludes his paper by saying the oil and gas industry needs more regulation, including the “control of formaldehyde emissions” (HARC study, p. 976). In other words, based partially upon a clear misreading of a separate study, Olaguer is trying to make the case for costly new regulations — hardly the kind of sound scientific basis that should guide policymakers (take note, Dallas City Hall).
Olaguer: The author states the study set out to answer a series of questions, including: “How far from the source are significant ozone impacts likely to be seen?” (HARC study, p. 968; emphasis added)
FACT: It’s probably not a great idea to use leading questions to guide your research — unless, perhaps, you’ve already determined what the answers are before you conduct the study. For example, Olaguer asks “how far” away from the source will “significant” impacts be seen, instead of asking the obvious question: “Are there significant ozone impacts from oil and gas activities?” Clearly, Olaguer had already decided what the answer to that question would be, and his decision clearly did not mesh with what the available data actually show. So, instead of even trying to explain that away, he simply leaped from hypothesis to supposition.
Olaguer: “To answer these questions, we conducted a schematic modeling exercise that was not intended to implicate any actual operational facility, but only to provide reasonable quantitative bounds.” (HARC study, p. 968; emphasis added)
FACT: The language above says it all — this isn’t an assessment of what’s actually in the field, and it’s certainly not an aggregation of real-world data. And yet, Downwinders at Risk director Jim Schermbeck said the report is “proof” that industry facilities are improperly managed, and that “our air is not getting cleaner because gas pollution is still under-regulated.” Again, it’s a hypothetical situation shaped by carefully constructed inputs in a computer model…which is fitting, because that’s perhaps the only realm where opponents can possibly validate their claims.
Olaguer: “Given the possible impact of large single facilities, it is all the more conceivable that aggregations of oil and gas sites may act in concert so that they contribute several parts per billion to 8-hr ozone during actual exceedances.” (HARC study, p. 976; emphasis added)
FACT: In addition to the lack of definitive conclusions about actual air quality — to be expected from a modeling exercise — data from the past few years actually contradicts Olaguer’s “maybe, possibly” conclusion. For example, here’s a graph showing 8-hour ozone levels in the Dallas-Fort Worth area between 1999 and 2009, along with the substantial increase in natural gas production.
If natural gas production were a significant contributor to area ozone — which the HARC paper tried to establish essentially by decree — then ozone concentrations would increase along with an expansion in natural gas production. But that’s simply not what the data show — inconvenient as that may be for those claiming otherwise.
Olaguer: “Our findings suggest that improved regulation of the upstream oil and gas industry in nonattainment areas should include reporting of emission events, and more aggressive deployment of control strategies, such as vapor recovery to avoid flaring, and the use of oxidation catalysts on stationary engines.” (HARC study, p. 976; emphasis added)
FACT: It’s well known that natural gas in much of the Barnett Shale — especially in Dallas County — is what’s known as “dry gas” since it contains little if any associated liquids. Mandating the use of vapor recovery technology on gas that isn’t producing vapor reflects a poor understanding of the subject matter itself. Even if liquids were present, strict regulations are in place for the Barnett Shale (details can be found here) that mandate the use of vapor-recovery systems.
As Ed Ireland from BSEEC has noted:
“[M]ost of the natural gas produced in and around the 9‐county DFW NAA [non-attainment area] is very “dry” gas. The reason, geophysicists say, is that this part of the Barnett Shale is “thermally mature”, meaning that these natural gas wells produce no associated oil or other liquids. This means that the vast majority of Barnett Shale wells in the 9‐county DFW NAA do not require tanks for condensate storage. Little or no VOC is emitted from these gas wells.”
Olaguer said that his findings constitute “a severe challenge to oil and gas producers in the DFW area,” and Downwinders at Risk said the Dallas city council specifically “has a chance to react positively to this new evidence.” But what does it say that at least one of the recommendations behind those claims is something that literally makes no sense for Dallas County?
Some might argue that TCEQ is only one source of information, so there’s plenty of room for additional research. That’s certainly a valid point. But in the absence of other meaningful assessments that rely on empirical data instead of computers models, TCEQ’s findings have to be seen as the benchmark.
It’s also worth stressing that this isn’t meant to demonize the use of computer models in research. Indeed, when gaps in available data exist for a given subject, modeling exercises can play an important role in estimating impacts (or the lack thereof). But in this case, that data actually do exist, and it’s unfortunate that people choose to believe hypothetical emissions estimates are more valid than easily-accessible readings from actual air monitors.