Climate Change Opinion

Should AGU Have Fly-in Meetings Anymore?

Should members of the American Geophysical Union “walk their talk” by cutting carbon emissions related to meeting travel?


The American Geophysical Union’s (AGU) Fall Meeting is the world’s largest international gathering of Earth and space scientists. Last year, nearly 24,000 people gathered in San Francisco, Calif., to attend. This year, about 22,500 people attended Fall Meeting in New Orleans, La.

To satisfy my curiosity, I undertook a study in 2012, a year I happened to attend the meeting, to estimate the carbon contribution of all this air travel to and from the meeting. In 2012, 21,702 people gathered for the meeting. I expected that the contribution would be vanishingly small, given that the anthropogenic production of carbon dioxide (CO2) from fossil fuels in 2010 was approximately 30 gigatonnes [Intergovernmental Panel on Climate Change (IPCC), 2014]. The contribution was more significant than I expected.

Attendees’ air travel to and from this meeting was responsible for more than 5 millionths of total global carbon emissions, even though these attendees accounted for 3 millionths of the global population. How can AGU “walk its talk” on environmental stewardship when participation in its annual meetings requires such carbon-intensive travel?

Context and Caveats

AGU’s leadership in discussions of anthropogenic warming is evident from published research by its members and the organization’s position statements on climate change. The impact of the meeting on CO2 emissions has never been estimated.

I estimated CO2 emissions for each conference presenter’s airline travel to and from San Francisco (SFO) for a sample of conference presenters. (The data I used are here.) This travel is the only part of each attendee’s travel emissions that I estimated. No attempt was made to estimate CO2 emitted during other activities, such as transportation to and from airports, use of restaurants, hotels, and so on.

Gathering the Data

The AGU 2012 Fall Meeting had 1,739 technical sessions, excluding Union and town hall sessions. I sampled 37 of these sessions (2.1%) as follows: Each session was assigned a unique number from 1 to 1,739. I used a random number generator, based on atmospheric noise, to choose 37 numbers. The session corresponding to each number was included in the sample set.

The random sample did not include sections on Atmospheric and Space Electricity (AE), Cryosphere (C), Earth and Planetary Surface Processes (EP), Earth and Space Science Informatics (IN), and Public Affairs (PA). These five sections held 13% of all sessions. As measured by the deviation of the percentage sampled from the percentage of total papers, Biogeosciences (B) was the most undersampled (1 session sampled out of 135) section, and Geodesy (G) was the most oversampled (4 out of 39).

Each presentation had a designated presenter. I assumed that the designated presenter gave the presentation, and the presenter traveled directly between SFO and an airport close to her or his home institution. The effect of this assumption is to underestimate the total travel because it is likely some traveled elsewhere in the same trip, and airline connections were not considered. Other simplifying assumptions included that presenters traveled to SFO from the largest city near their home institutions via a great circle route and that presenters from northern California did not fly. I used various online resources to obtain airport information and determine distances.

Calculating Emissions

Carbon dioxide emissions were based on fuel consumption for different types of aircraft and distances flown, and the “share” of consumption by each passenger was based on the passenger capacity of the aircraft. A number of online emissions calculators for airline flights are available, but they give inconsistent results and do not explain the derivations, hence the methods used here.

I made simplifications that included the following, with values derived from a variety of industry websites:

  • Middle values of fuel consumption for each type of aircraft were used, with the following assumptions:
  • Long-range jets (Airbus 380 and Boeing 747) were used for flights from destinations outside North America and from Alaska.
  • Middle-range jets (Boeing 757) were used for flights from eastern and central North America and Mexico.
  • Short-range jets (Boeing 737 and Airbus 320) were used for flights within western North America except Alaska.
  • I used middle values for cruise speeds of long-, medium-, and short-range jets.
  • I used middle values for passenger capacity for each type of aircraft.

Much more detailed information is available from a variety of sources. For example, the European Environment Agency’s report on aviation distinguishes the landing and takeoff phases of flight from the cruise phases; I used consumption rates only for the cruise phases.

A value of 2.55 kilograms of CO2 emitted per liter of jet fuel (kg/L) was used. Fuel consumption is reported in pounds per hour, and jet fuel expands and contracts substantially with temperature and is subject to a wide range of temperatures during flight. Therefore, fuel consumption volume must be converted to weight. For this study, volume was converted to weight at 15°C (standard temperature); jet fuel weighs 0.8 kg/L.

Presenters from 36 countries are represented in the sample. Of these, 24.4% were from Europe (including western Russia), 13.9% were from Australia and Asia, 0.6% were from the Middle East, 0.6% were from Africa, 1.4% were from South America, and the remaining 59.1% were from North America (including Hawaii). Of the latter, approximately 4.7% were from Canada, 0.3% were from Mexico, 52.8% were from the lower 48 U.S. states, and 1.4% were from Alaska and Hawaii.

In total, presenters in this sample accounted for the emission of 257 tonnes of CO2 during their flights to and from SFO. These presenters represented 1.7% of all presenters at the meeting. If the geographic distribution of attendees is the same as that of the sample, all attendees (21,702) accounted for the emission of 15,493 tonnes of CO2. The assumption of geographic distribution probably results in an overestimate because it is likely that a higher proportion of attendees who were not presenters were from northern California.

Annual total anthropogenic emissions of CO2 (not CO2 equivalents) is about 30 gigatonnes [e.g., IPCC, 2014]. At that level, AGU presenters accounted for 5 millionths of total global anthropogenic emissions from fossil fuels in 2012.

Reasons for Meetings

Attendees at the American Geophysical Union’s 2017 Fall Meeting, this year in New Orleans, La., crowd the poster hall.
Attendees at the American Geophysical Union’s 2017 Fall Meeting in New Orleans, La., crowd the poster hall. Roughly 22,500 people attended this year’s meeting. Credit:

Other attempts to calculate the carbon footprints of large conferences (e.g., United Nations Climate Change Newsroom and Programming Languages Enthusiast) reveal a similarly large footprint compared to the population in attendance. In light of this, some debate exists whether fly-in conferences should be held at all [Parrish, 2009; Environmental Humanities Center, 2017]. The discussion goes to the question of the purpose of a scientific meeting.

Scientific meetings are not held solely to present scientific findings. Indeed, scientific meetings can be argued to be the least efficient means of disseminating scientific results because attendees cannot possibly see all presentations, even in their fields of study.

However, scientific meetings do provide opportunities for face-to-face interaction, which enhances the exchange of ideas. An informal poll of my colleagues revealed that the social interactions—handshakes, hugs, friendly competition, and brainstorming—are the main thing they would miss if they could not attend meetings in person.

The value of these interactions has been assessed for one meeting as a part of the development of a carbon-neutral meeting model. Although the loss of face-to-face contact was the primary concern of the study, only 20% of the participants regarded it as a significant shortcoming.

For large meetings like AGU’s, face-to-face contact is limited by the sheer size of the meetings and the tendency for attendees to gravitate toward people they already know and with whom they already efficiently collaborate by electronic means. Large meetings are inefficient for maintaining social connections for the same reason they are inefficient for networking: There are simply too many people.

Other considerations include whether vacating seats on scheduled flights would make a difference. Greenhouse gas emissions by air transport have been estimated at 1.6% of the world total [Herzog et al., 2005]. Business travelers represent about 12% of air travelers (although, interestingly, they provide 75% of airline profits). An even smaller fraction of passengers would be travelers to AGU meetings. Typical load factors (percentage of passenger seats occupied) for airline flights are currently 82%–84% according to the U.S. Department of Transportation, which is high enough that flights would likely be profitable even if AGU members stopped traveling to Fall Meetings.

Beyond Flying In

Although 5 millionths do not sound like a big contribution to the world’s carbon footprint, we must bear in mind the very narrow focus of this estimate—air transportation to and from a single scientific meeting in a single year. AGU prides itself on being a leader on the issue of global warming, and many sessions are devoted to the consequences of anthropogenic CO2 emissions and related topics. Does the current form of the Fall Meeting detract from this message?

Whether dispensing with fly-in AGU meetings would have a significant effect on carbon emissions is almost beside the point. AGU’s leadership role compels it to act as an example to the community. By leading the way, the initial small effect of its actions could be multiplied, encouraging enough other organizations to dispense with large meetings and perhaps reducing business travel enough to influence airline scheduling.

Thus, I suggest that AGU consider moving with appropriate haste toward using current and developing technologies for its meetings and eventually end fly-in meetings. AGU already provides video access to a few sessions, but the presenters in most cases are at the meeting, and the video is for people who cannot attend.

AGU puts a lot of effort into warning society about the dangers of anthropogenic warming. Given the gravity of the problem, are the social connections for 3 millionths of the world’s population really so important? Shouldn’t we be willing to give up these meetings for the good of society? Paraphrasing Rabbi Hillel the Elder, “If not us, who? If not now, when?”


I am grateful to two anonymous reviewers for constructive comments and for calling my attention to some of the cited websites.


Environmental Humanities Center (2017), A nearly carbon-neutral conference model, white paper and practical guide, Univ. of Calif., Santa Barbara,

Herzog, T., J. Pershing, and K. A. Baumert (2005), Navigating the Numbers: Greenhouse Gas Data and International Climate Policy, 8 pp., World Resour. Inst., Washington, D. C.

Intergovernmental Panel on Climate Change (IPCC) (2014), Climate Change 2014: Synthesis Report, Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Geneva, Switzerland,

Parrish, J. T. (2009), A field geologist looks at a digital world, GSA Today, 19(1), 4–6,

—Judith Totman Parrish (email: [email protected]), Department of Geological Sciences, University of Idaho, Moscow

Citation: Parrish, J. T. (2017), Should AGU have fly-in meetings anymore?, Eos, 98, Published on 21 December 2017.
© 2017. The authors. CC BY 3.0