Social Channels


Additional Options

Date Range

Receive a Daily or Weekly summary of the most important articles direct to your inbox, just enter your email below:

Roz Pidcock

Roz Pidcock

14.11.2013 | 4:15pm
ScienceTemperatureNot so slow “slowdown”? New paper says warming in last 15 years may be double what scientists thought
SCIENCE | TEMPERATURE | November 14. 2013. 16:15
Not so slow “slowdown”? New paper says warming in last 15 years may be double what scientists thought

New research challenges the idea put forward in the most recent IPCC report that there has been a so called “slowdown” in surface warming recently, saying temperature rise in the last decade and a half may be nothing unusual after all.

The paper – which is an updated look at one the major temperature data sets the IPCC uses – suggests the rate of surface warming since 1997 is more than twice previously thought. But scientists tell us when you look a bit closer at the new data and the IPCC’s numbers, the two aren’t inconsistent.

IPCC on the “slowdown”

A big discussion point surrounding the latest climate report from the Intergovernmental Panel on Climate Change (IPCC) is how surface temperatures – that’s the air above the land and oceans – have been rising relatively slowly over the past decade and a half.

The IPCC report put the rate of warming between 1998 and 2012 at 0.05 degrees Celsius per decade. That’s quite a lot slower than the average of 0.12 degrees per decade since 1951.

In the report, this is given as an example of how natural climate variability can cause temperatures to rise faster or slower than than the long term average from one decade to the next.

But a new paper, just published in the journal Quarterly Journal of the Royal Meteorological Society, takes a fresh look at one of the main datasets the IPCC uses, the Met Office’s HadCrut4 temperature record.

The authors say correcting for well-known gaps in the dataset – to do with how much of the globe it covers – brings the rate of warming since 1997 up to 0.12 degrees Celsius. In other words, right in line with the decadal average.  

Met Office and its “cooling bias”

The HadCrut4 dataset, jointly run by the Met Office and the University of East Anglia, is one of four global temperature records. Compared to the other three, HadCrut4 is generally considered the most complete.

The new paper addresses a particular feature of the temperature record. Continuous and accurate temperature measurements are sparse in remote and difficult to reach places, especially in the poles. This means the dataset actually covers only 85 per cent of the globe.

As this RealClimate post discusses, this wouldn’t be a problem if the whole globe was warming at the same rate. But temperatures in the Arctic are rising faster – at about twice the speed of global average. It follows that missing out measurements from this region could lead to an underestimation of how much global temperatures are changing.

Climate scientist James Hansen first pointed out this so-called “cooling bias” in the HadCrut4 data back in 2006. A Met Office spokesperson says of the new research:

“This paper acknowledges something that has been known about for some time, which is that incomplete global coverage of observations adds uncertainty in global mean temperature estimates â?¦ If you use methods to fill in those blanks [in the Arctic] you can try to capture what is going on in those sparsely observed regions and potentially reduce uncertainty in estimates of global temperature.”

Gap-filling exercise

The new paper essentially tries to plug the gaps in the HadCrut4 dataset using satellite data. Satellite data measures the temperature high up in the atmosphere rather than at the surface, but by comparing pairs of air and ground measurements, the team worked out what the corresponding temperature at the earth’s surface should be where data was missing.

The process is pretty complicated but the authors give a good explanation here. Kevin Cowtan from the University of York and lead author on the paper tells us:

“What we do is essentially to take the satellite data and ‘anchor’ it to the nearest surface observations â?¦ I don’t think anyone has attempted this.”

But the gap-filling method has its drawbacks and getting the method right has taken some time, Cowtan continues:

“[S]atellite data measures lower atmosphere temperature, not surface temperature, so the two are not directly comparable. As a result most of our work has been devoted to testing whether what we are doing is valid. One way we do this by leaving out surface observations and seeing if we can predict what they would have been. This tells us that we can use the satellite data, at least over land.”

Trend setting

You can see the difference the reanalysis makes to global temperature record in the graph below. The black line is the original HadCrut4 data, the new analysis is in blue.

Cowtan &Way _yearcomparison

For an individual year, the difference isn’t particularly large. In fact, it’s well within Met Office uncertainty limits, shown by the grey shading.

But put all the corrected measurements together and it does make a difference to the recent temperature trend. The image below shows the trend after 1997 calculated from the original HadCrut4 data (thin red line) compared to the new, updated data (thick red line).

Cowtan &Way _trendcomparison

The warming trend in the last decade or so is considerably steeper than without the corrections. The authors explain in their University of York blog post:

“Our global record suggests that surface temperatures have been warming two and a half times faster than Met Office estimates over the past 16 years.”

Cowtan tells us why temperature trends after 1997 seem particularly affected. He says:

The reason why it makes a big difference post 1997 is that that is when the Arctic warming really kicked into high gear. Look at this figure from NASA/GISS. Notice how the light blue line turns steeply upward after 1997.”


Gistemp -coverage -zonal

Source: NASA/Goddard Institute for Space Science (NASA/GISS)

So where does this leave the “slowdown”?

In its estimate of the trend in temperature rise since 1998, the IPCC didn’t just provide a single figure of 0.05 degrees Celsius. To take into account the uncertainty in the measurements, it gave a range of – 0.05 to 0.15 degrees Celsius. The new estimate falls within those limits, at the upper end. So the two assessments are not necessarily contradictory.

Another way to look at it is that the 0.12 degrees Celsius figure in the new study is the authors’ best estimate, but it still allows a good chance the trend could be less than that. Indeed, the authors say the “most probable case is that there has been a slight slowdown in the rate of warming”.

A Met Office spokesperson tells us the there’s strong evidence of a slowdown in some parts of the world:

“Areas of the world with better data coverage still show slower warming in the last decade and a half, so there is still evidence of a slowdown to some extent in the data record.”

Last year, the Met Office released three reports looking in detail at what could be behind slower recent warming. It came to much the same conclusion as the IPCC, which said in its latest report:

“The observed reduction in surface warming trend over the period 1998 to 2012 as compared to the period 1951 to 2012, is due in roughly equal measure to a reduced trend in radiative forcing and a cooling contribution from natural internal variability, which includes a possible redistribution of heat within the ocean”

The Met Office tells us these proposed explanations are still likely to be valid, but the extent to which they’re affecting global temperature remains uncertain. A spokesperson explains:

“The three Met Office papers from last year looked at the ranges of potential contributions to slower temperature rise from natural variability and changes in external forcing â?¦ It’s not a question of whether the slowdown hasn’t happened at all â?¦ This new paper would suggest that the slowdown may not be as great as indicated by previous estimates based on HadCRUT4, however.”

The problem with short-term trends

One important conclusion you could draw from this paper echoes a point the IPCC makes in its latest report:

“Due to natural variability, trends based on short records are very sensitive to the beginning and end dates and do not in general reflect long-term climate trends.”

This new paper highlights another reason why interpreting short term trends in temperature rise is difficult. With uncertainty in temperature measurements that can be bigger than the trend, this paper might indicate that there’s even less chance your short-term trend will reflect what’s happening in the long term.

It remains to be seen if other analyses – including perhaps one from the Met Office itself as part of ongoing improvements to the dataset – come up with similar findings. As author of the paper Dr Cowtan says, we should be wary of attaching too much importance to one bit of research:

“No difficult scientific problem is ever solved in a single paper. I don’t expect our work to be the last word on this, but I hope we have advanced the discussion.” 

Related Articles


Expert analysis directly to your inbox.

Get a Daily or Weekly round-up of all the important articles and papers selected by Carbon Brief by email.


Expert analysis directly to your inbox.

Get a Daily or Weekly round-up of all the important articles and papers selected by Carbon Brief by email.