Better reporting on computer models could dispel some of the mysteries of climate change

Now that climate topics have been allowed back in the public arena, it’s time for the media to fill some serious gaps in the coverage of climate science. A good place to start would be to explain how computer models work. While a story on the intricacies of algorithms might seem to be a “yawner,” if told from the point of view of a brilliant scientist, complete with compelling graphics, or, better yet, with the immersive technology of new media, stories on climate models could provide ways for non-scientists to evaluate the reliability of these tools as predictors of the future.

Equally important, social media and the virtual communities that websites are capable of forming can help to overcome a major barrier to the public’s understanding of risk perception: The tendency of citizens to conform their own beliefs about societal risks from climate change to those that predominate among their peers. This derails rational deliberation, and the herd instinct creates an opening for persuasion — if not deliberate disinformation — by the fossil fuel industry. Online communities can provide a counter-voice to corporations. They are populated by diverse and credible thought leaders who can influence peers to not just accept ideas but to seek out confirming evidence and then take action. Because social networks enable the rapid discovery, highlighting and sharing of information, they can generate instant grassroots activist movements and crowd-sourced demonstrations.

Studies show that a major cause of public skepticism over climate stems from ignorance of the reliability of climate models. Beyond their susceptibility to garbage in, garbage out, algorithms on which models are based have long lacked the transparency needed to promote public trust in computer decisions systems. The complexity and politicization of climate science models have made it difficult for the public and decision makers to put faith in them. But studies also show that the media plays a big role in why the public tends to be skeptical of models. An article in the September issue of Nature Climate Change written by Karen Akerlof et al slammed the media for failing to address the science of models and their relevance to political debate:

Little information on climate models has appeared in US newspapers over more than a decade. Indeed, we show it is declining relative to climate change. When models do appear, it is often within sceptic discourses. Using a media index from 2007, we find that model projections were frequently portrayed as likely to be inaccurate. Political opinion outlets provided more explanation than many news sources.

In other words, blogs and science websites have done a better job of explaining climate science than traditional media, as visitors to, and other science blogs can attest. But the reach of these sites and their impact on the broader public are debatable. Websites such as the U.S. Department of Energy’s Office of Science have a trove of information on climate modeling but, with the exception of NASA’s laboratories, most government sites on science make little effective use of data visualization. This void offers mainstream journalists an opportunity to be powerful agents in the climate learning process, to tell dramatic multimedia stories about how weather forecasts can literally save our lives and, by extension, why climate forecasts can be trusted.

Two recent events can be thought of as whetting the public’s appetite for stories about computer-generated versions of reality. The prediction that Hurricane Sandy would eventually turn hard left out in the Atlantic and pound the northeastern shore of the United States was made almost a week in advance by weather forecasters.

This technology-driven prediction no doubt saved countless lives. In addition, some media coverage of Hurricane Sandy did much to enable non-scientists to understand why it is tricky to attribute specific storms to climate change but still gave the public the big picture of how warmer ocean waters provide storms with more moisture and therefore make them bigger and more damaging.

Simultaneously, in a different domain but using the same tools of analysis and prediction, Nate Silver’s FiveThirtyEight computer model, results of which were published in his blog at The New York Times, out-performed traditional political experts by nailing the November national election outcomes. How did he pull that off? A story about his statistical methods, complete with graphics, could reveal how risk analysts create spaces between the real world and theory to calculate probabilities. This would help the public to become familiar with models as a source of knowledge.

Some reporters have produced text stories on climate models that are examples of clarity. Andrew Revkin, while as an environment writer for The New York Times and now as the author of his Dot Earth blog at’s opinion section, has for many years covered how climate models relate to a large body of science, including a posting on Oct. 30 that placed Hurricane Sandy in the context of superstorms of the past.

David A. Fahrenthold at The Washington Post wrote how “Scientists’ use of computer models to predict climate change is under attack,” which opens with a baseball statistics analogy and keeps the reader going. Holger Dambeck at SpiegelOnline did a thorough assessment of climate model accuracy in non-science language, “Modeling the Future: The Difficulties of Predicting Climate Change.” But these stories are rare and often one-dimensional.

Effort is now being spent on making scientists into better communicators, but more might be accomplished if mainstream journalists, including those who publish on news websites with heavy traffic, made themselves better acquainted with satellite technology and its impact on science. Information specialist Paul Edwards explains in his book, “A Vast Machine: Computer Models, Climate Data and the Politics of Global Warming,” how climate modeling, far from being purely theoretical, is a method that combines theory with data to meet “practical here-and-now needs.” Computer models operate within a logical framework that uses many approximations from data that — unlike weather models — can be “conspicuously sparse” but still constituting sound science, much as a reliable statistical sample can be drawn from a large population. How statistics guide risk analysis requires better explanation for a public that must make judgments but is seldom provided context by news stories. The debate over cap-and-trade policy might be Exhibit A.

Depicting model-data symbiosis in such diverse fields as baseball performance, hurricane forecasts and long-range warming predictions would be ideally suited to web technology. Not only can climate models be reproduced on PCs and laptops, showing atmospheric changes over the past and into the future, but also the models’ variables can be made accessible to the web user, who could then take control of the model and game the display by practicing “what ifs” — how many degrees of heat by year 2100 could be avoided by a selected energy policy, how many people would be forced into migrations if this amount of food supplies were lost, how big would a tidal barrier need to be to protect New York City from another Sandy disaster? (If this sounds a bit like SimCity, the new version of the game due in 2013 includes climate change as part of the simulated experience.)

This narrative approach to news, including personal diaries and anecdotes of everyday lived experience, is what Richard Sambrook, former director of BBC Global News and now a journalism professor at Cardiff University, has termed “360 degree storytelling.” Mike Hulme, a professor of climate change at East Anglia University, provides this description of the new public stance toward science in his book, “Why We Disagree About Climate Change”:

Citizens, far from being passive receivers of expert science, now have the capability through media communication “to actively challenge and reshape science, or even to constitute the very process of scientific communication through mass participation in simulation experiments such as ‘’. New media developments are fragmenting audiences and diluting the authority of the traditional institutions of science and politics, creating many new spaces in the twenty-first century ‘agora’ … where disputation and disagreement are aired.”

Today’s media is about participation and argumentation. A new rhetoric of visualization is making science more comprehensible in our daily lives. What goes around, comes around. One of the pioneer online journalism experiments in making the public aware of how technology, risk assessment and human fallibility can cross over was a project by known as the “baggage screening game.” Players could look into a simulated radar screen and control the speed of a conveyor line of airline passenger baggage — some of which harbored lethal weapons. Assuming you were at the controls, the program would monitor your speed and accuracy in detection and keep score, later making you painfully aware of missed knives and bombs. Adding to your misery was a soundtrack of passengers standing in line and complaining about your excessive scrutinizing, with calls of “Come on! Get this thing moving! We’re late!” It was hard to be impatient with the TSA scanners after that.

About Larry Pryor

I am an associate professor at the Annenberg School of Journalism and am a former editor of OJR. I left online journalism to work full-time at teaching environmental journalism. I had been an environment writer at the Los Angeles Times before getting into new media.
I'm attempting to combine my work in visual journalism with environmental coverage. Digital models can help us connect data points into more understandable patterns. Mash-ups are great tools.