Content Block


Big data versus the denialists

US presidential hopefuls may be in denial over climate science, but the Research Data Alliance is working on new ways to improve our knowledge

Every year sea ice in the Arctic expands to its greatest extent in March then shrinks to a minimum in September. These millions of square miles of brilliant ice play an important role in the temperature of our world. The more there is, the more sunlight gets reflected back into space and the cooler the planet stays.

September 2012 saw a record low not just for the year, as is normal, but since modern measurement began. This February the extent was the greatest for the year, as is normal, but again a record minimum.

Evidence of global warming? Without a doubt, says Mark Parsons, formerly a data scientist at the US National Snow and Ice Data Center in Colorado. But comments by at least half a dozen of the prospective candidates for the US presidency indicate they aren’t worried by this kind of data. “There’s always going to be either cooling or warming going on,” says Ben Carson. Mike Huckabee, when asked if man contributes to global warming, said: “He probably does, but a volcano in one blast will contribute more than a hundred years of human activity.”

Parsons is reluctant to get drawn into battle. He says the politicians are appealing to voters’ emotion. As an expert in data science, Parsons isn’t professionally predisposed to argue on emotion. His job, and that of a growing number like him, is to study and improve policy and protocols on research data so that they can be interrogated transparently and used more effectively.

Working out where the edge is

When it comes to sea ice, there are enough problems to deal with before getting in the ring with Carson, Huckabee et al. The first is quite simply working out where the edge is. If you can’t define the edge, you can’t measure the area. “Sea ice does not end neatly,” says Parsons. “It is jagged and bumps against itself.” Moreover, as those records suggest, Arctic sea ice contracts and expands throughout the year.

With a size in spring greater than any country other than Russia – and a harsher climate – it is not possible to get close to the perimeter of the sea ice to conduct checks on what’s solid and what’s liquid. There are plenty of hunters, fishermen and fisherwomen, following cracks, known locally as “leads”, to their prey. But satellites in space, travelling over the Arctic several times a day, are the prime source of estimation.

The complexity of this hostile terrain still makes authoritative numbers hard to reach. For example, clouds and darkness hinder one kind of observation from space. The challenges don’t always bring sympathy from customers: commercial shippers are notorious for ribbing the scientists about their mapping of Arctic ice - just as the general population mock weather forecasters (As Nate Silver well documented in his 2012 book The Signal and the Noise, however, meteorologists have made far greater strides in accuracy over the past twenty-five years than forecasters in other branches of human endeavour.).

Weather is all about the short term: what happens in the next 72 hours really matters to a ship trapped in ice. But climate is all about the long term, typically defined as periods of thirty years or more. Organisations such as the NSIDC serve communities with different timeframes, and there has been more interest this century in the long term due to phenomena such as record minima of sea ice.

Part of Parsons’ job was to work out how to serve these new communities, especially when different sets of long-run data were compiled using different criteria and technology, including algorithms. If this sounds like a matter exclusively for science nerds and IT departments, think again. Back in 2009 emails from the University of East Anglia’s Climatic Research Unit were hacked and leaked. Emerging just one month ahead of the United Nations Climate Change Conference in Copenhagen, the leaks gave much fuel to the arguments of climate-change sceptics: CRU was responsible for compiling data to be discussed at the Conference.

The exposed emails were selected to give the impression that scientists at East Anglia and elsewhere were fiddling numbers to make climate change seem worse than it is. The spat got as far as US Congress, while Sarah Palin weighed in with an opinion piece in the Washington Post of the kind that would not satisfy Parsons’ criteria for factual analysis.


His own contribution to "Climategate" was a scholarly article pointing out some of the barriers, or rather holes, to any external researcher trying to understand exactly what has been measured and how. For example, the recommended citations for the CRU datasets are a variety of scientific papers that do not contain the actual gridded data. The CRU data also continue to evolve and change in ways not documented in the original articles.

Parsons makes some excellent points that even among scientists and science publications in the academic community, more value is placed upon minor articles in minor journals than really good data. One of the reasons is that no formal, objective overarching process for data review exists.

Given that data underpins scientific observation, checking, interrogating and re-evaluating the facts deserves more attention. It is widely agreed that what needs to be addressed are not the technological issues: the very computers which have facilitated the increase of data are capable of managing it. It is rather the culture of universities and research centres that need to accept the central function of data management in learning. These points should appeal not just to scientists but journalists and anyone else wanting the best possible information on scientific endeavour.

One body practically advocating change is the Research Data Alliance, a initiative backed by government agencies of Australia, the US and the European Union. How to share, challenge, curate and interrogate research is the RDA’s specialism. Its committees are packed with data scientists – in simple terms the people who give the rest of us the opportunity to make sense of Big Data.

Data scientists don’t necessarily generate or input the information themselves: their job is to facilitate others’ crunching, which is why the RDA is truly super-disciplinary, welcoming new members in the humanities as well as science.

Although the RDA is open to all disciplines of research, climate change will be the theme of its international meeting in Paris between 22-25 September.

Perhaps Ben Carson should have received an invitation to attend. The former neurosurgeon may be ambivalent about climate change but his remarks on his own scientific discipline would make him a reasonable advocate for the power of shared data. In the same speech last November as his gnomic remarks on global warming, he said: "The wonderful thing about medicine, when you show data, it changes people's minds."

Moreover, as a published medical author, Carson’s own work will likely be one of millions of future beneficiaries of the efforts of the Research Data Alliance.


Brendan specialises in explaining public and private welfare systems. He writes about how people save and the business of saving, for the Financial Times and Investments & Pensions Europe among other publications. He is currently investigating the commercialisation of Intellectual Property within universities. He began his journalist career with the FT Group.

Related Posts

  1. Arthur I. Miller is emeritus professor of history and philosophy of science at University College London. He is the author of several acclaimed books, including Einstein, Picasso, which was nominated for the Pulitzer Prize, Empire of the Stars, which was shortlisted for the 2006 Aventis Prize for Science Books, and 137, which we’re discussed on a previous Little Atoms. An experienced broadcaster, lecturer and biographer, he is particularly interested in the relationship between science and creativity, and his latest book is Colliding Worlds: How Cutting-Edge Science is Redefining Contemporary Art.