Fingerprinting a Climate

Climate simulation models are usually 250 years in scope – 150 years into the past and 100 years into the future.  The models have time intervals of 20 minutes.  Many millions of calculations have to be undertaken for every 20 minute time-step, so huge super-computers are required to do the job.  Even with such super-computers, as at the UK’s Met Office, it still takes 3 months to run a complete simulation.

To test a climate model, a simulation is undertaken of the observed 0.8 degree Celsius warming of the globe over the past 150 years – a process called fingerprinting.  In this way, the simulation’s predictions can be tested against observed data.  The models are built using the factors that climate scientists think have affected climate in that time:  natural factors such as volcanoes and variations in the output of the sun; and human factors, most particularly the increase of carbon dioxide resulting from burning fossil fuels and deforestation. 

When the models are run using only only natural factors, they can reproduce the observed climate data, but only until about 1970.  After 1970 the models and the observed data diverge, and in fact the models tend to show a period of planetary cooling.  When the human factors, particularly carbon dioxide emissions, are included in the models, then the real, observed global warming and that predicted by the simulation, coincide.

This ability to pull out the effects of natural and human factors not only tests the models, but also allows scientists to accurately attribute the effects that human activity is having on the Earth’s climate.  This process of fingerprinting therefore allows the Intergovernmental Panel on Climate Change to make such definitive statements about the effect we are having on our climate.  Their fourth report states that “There is at least a 90% chance that the observed increase in temperature globally is due to man-made greenhouse gases”. 

The challenge then is to work out how to project climate change into the future, which is a far more challenging process as there are so many variables, including:

  • How will the land and sea sinks behave when concentrations of atmospheric carbon dioxide increase?
  • How will human population grow?
  • How will human behaviour change, regarding levels of carbon dioxide output?

To model future climate change, scientists create a number of different scenarios for these unknown factors.  These seem to result in a spread of possible future global warming over the next 100 years from 2-6 C.  The current worst case of 6C constitutes a warming of more than that between the last ice age and now, but 100 times faster.  And this has the ability to tip us over the edge into a series of horrific scenarios, such as the slowing of the Gulf Stream, and the melting of the Antarctic and Greenland ice sheets, the latter resulting in a sea level rise of more than 10m. 

Decoding the Keeling Curve from Mauna Loa

Since 1958, Charles David Keeling, then his son, Ralph Keeling, have made a continuous recording of carbon dioxide levels in the atmosphere at the Mauna Loa Observatory in Hawaii.  This has produced the now iconic Keeling Curve, showing a steady increase in concentrations of levels of atmospheric carbon dioxide from the 1950s to the present day.  This increase roughly correlates with increases in burning of fossil fuels, and also with levels of deforestation.  Currently 90% atmospheric carbon dioxide increase comes from fossil fuels and 10% from deforestation.

Key to the history of the graph is a tale of individual determination in the face of funding cuts and general disinterest at the start of the CO2 recording process.  Keeling had also been recording CO2 levels at the North Pole, but these recordings were abandoned when funding was cut.  In the face of these cuts, Keeling kept going with the Mauna Loa record, despite having few resources.

The graph is not a steady line.  It curves upwards with regularly spaced wiggles and with irregularly flattened sections on the curve. Scientists have spent much time decoding these readings, and are continuing to do so.

For example, it has been found that the regularly spaced wiggles on the graph relate to the annual growth rate of vegetation in the northern hemisphere.  In the summer the vegetation sucks down CO2 and releases it in the winter when it decomposes.  This results in a regular wiggle on the curve.

There are also irregularly spaced flat areas on the curve, showing variation in the year to year growth rate of CO2.  These can sometimes be tied to specific global events, such as a decrease in the level of CO2 after the PInatubo eruption.  In the main, however, this annual variation in growth rate has been found to relate to how tropical lands, particularly tropical forests, are responding to anomalies in tropical weather and temperature.  Spikes in the annual growth rate of carbon dioxide in 2005 and 2010 can be seen to relate to drought in Amazonia.  And plotting the annual growth rate of CO2 against tropical temperature anomalies produces a very tight correlation, reflecting how tropical temperature affects how much carbon is absorbed by tropical forests.

This week, scientists from Peking University published a paper in Nature (Wang et al, 2014), analysing how flat spots in the annual mean line have changed sensitivity through time in the period 1970 to 2013.  In the 1970s, spikes in tropical temperature were producing lower spikes in CO2 than they do today – in 2013, temperature spikes caused a doubling of CO2 levels compared to the rate in the early 1970s.  This paper does not propose an explanation as to why this is happening.  Professor Peter Cox suggests a couple of potential reasons, such as a long term trend of drying in the Amazon, or do with variation in the ocean?

The Mauna Loa record is a key climate change record, which scientists are working hard to decode, comparing it with other data sets to reveal new questions, and sometimes new answers.  We need to be very grateful to the tenacity of one family, the Keelings, who worked so hard to keep the record going.  Without it, our understanding of how we are affecting the carbon cycle, and how in turn that is affecting our lives, would be so much less.

Every time look at Mauna Loa record and compare with other data sets, finding new questions, and sometimes new answers.