Information literacy in our technocratic age
A new paper provides a window into the climate industrial complex
“Information literacy” is often a core pedagogical concept in university curriculums. The University of North Carolina Wilmington, where I am a professor, describes success in this area as:
Information literate citizens examine the socio-cultural production of information and knowledge, and they approach all information production and consumption in a critical manner.
In my classes, rather than creating arbitrary rules around ‘reliable sources’ I advise that whether students are viewing Tik Tok videos or papers in Nature, they should always ask themselves, “Who is speaking and to what end?” (This phrase should be credited to someone but I don’t recall who.)
A more in depth and classic framework for analyzing information is:
Who, said what, in which channel, to whom, with what effect?
With this in mind, let’s turn to a recent academic article.
A new paper in Nature Communications seeks to resuscitate the outdated and extreme RCP8.5/SSP5-8.5 emission scenarios to legitimize its continued use. To be more specific though the paper aims to provide a post hoc rationalization for the continued use of the 8.5 W/m2 radiative forcing that is an outcome of the outdated assumptions underpinning emission scenario.
The paper is a collaboration between the consulting firm Abt Global, the Environmental Protection Agency, and a UK researcher. Abt is a major government contractor.
According to the Abt website, the company contributed to the development of the EPA’s Climate Impacts and Risk Assessment (CIRA) modeling project by estimating damages in a range sectors and in many instances Abt used RCP8.5 in its methods (I did not review all cited papers).
So, the new paper is best understood in the context of the EPA’s climate change damage estimating regime.
A few observations…
The EPA is the keeper of the US Social Cost of Carbon modeling process in which it uses emission scenarios developed by Resources for the Future (RFF). So in the paper, the authors write:
RFF-SP is recent, open-source, fully probabilistic, extends through the year 2300, and is used by the U.S. SC-GHG process, making it an ideal dataset for emission projections.
This is self referential: We, the EPA, use these scenarios in this paper because the EPA uses these scenarios. Therefore all is good.
The authors conclude:
Despite the low probability of exceeding 8.5 W m−2 by the end of the century, there are at least three applications where it remains crucial to consider these high-forcing, low-probability scenarios: (1) development of damage functions or “by-degree” climate impact analyses, (2) as an analog for post-2100 climate change, and (3) consideration of low probability but high impact possibilities.
The decision making contexts invoked by the EPA- contractor team relates to three executive level policy matters listed here in order of release:
The Office of Science and Technology Policy legitimized RCP8.5/SSP5-8.5 for continued use in climate risk analysis by agencies
President’s Council of Advisors on Science and Technology relied heavily on government and peer-reviewed literature using RCP8.5 to recommended that the United States invest in advancing climate change risk modeling. Several members of the recommending team work closely with private risk modeling firms that work with government agencies and financial interests.
Of course, the paper does not mention any COI. However, if we are to apply a critical lens to the production and consumption of information we can see that Abt has a business interest in this and there are significant policy implications with interest from on high.
The Nature Communications article provides access to the peer review comments providing insight into the technical discourse among experts assessing the strengths/weaknesses of the argument and methods applied.
Some interesting points brought out from the exchange:
Reviewer comments found the article lacked novelty because it only served to reiterate earlier arguments for rejecting the scenario. Reviewer #1 took this farther questioning the relevance of the work in the context of timing for real world outcomes as the paper was focused on probabilities for 2100, 2150, and 2300.
In response, the authors directly speak to the desire to decouple climate risk from time thereby bringing theoretical futures many decades or centuries out into the pricing and policy structure of the present:
The use of by-degree/global warming level/time-shift/pattern-scaling approaches mean that the impact modeler can disassociate the climate model output from the climate model time period. E.g., if a policymaker cares about a 2 degree stabilization scenario, then the 8.5 W/m2 data from the decade where 2 degrees is reached can be used as a proxy for the 2 degree outcome in 2100
In other words, the consulting business and its granting agency authors believe that a theoretical extreme possible sometime by 2100, 2150, or 2300 should be included in today’s financial regulations and government cost-benefit analyses.
Effort to continue legitimizing a methods that the broader climate change research practice is moving away from is shown in responses to a comment by Reviewer #4 and a similar one made by Reviewer #1 to which the authors respond:
Because of the upcoming decision in ScenarioMIP about the highest scenario to include for CMIP7, the question of whether to include 8.5 is very immediately relevant.
This is an attempt to anchor current industry/agency modeling practices and views of risk regardless of how things proceed among international researchers and IPCC reporting.
Quick Wrap up
Who is speaking? Government contractor and climate risk consultant, Abt, and it’s granting agency, EPA (IDK what’s in it for the UK researcher)
Said what? The authors create a post hoc justification for continued use of a modeled outcome (ie 8.5W/m2) in order to decouple underling assumptions and continue using existing damage functions embedded in commercial practices and government regulations
In which channel? A prestigious academic journal
To whom? A technically savvy and litigious audience that are in one of two camps: 1) highly critical of existing practices or 2) highly dependent on these practices for business and politics
With what effect? 1) Legitimizing common practices used by climate risk modeling consultants, the EPA, and climate advocates
You might wish to add to your analysis and overview of modelling scenarios. The result of this modelling is described in the mainstream media as scientific because it is done by scientists. However, any model whether considered probable or improbable is not science or scientific. It is merely an if-then statement, as an act of the human imagination.
"All models are wrong, but some are useful.", George Box