Anthropogenic carbon dioxide and climate change – a historical perspective
Simon Tett
Chair of Earth System Dynamics & Head of Global Change Research Institute
School of Geosciences, University of Edinburgh
[email protected]
ECG Bulletin January 2012
Chair of Earth System Dynamics & Head of Global Change Research Institute
School of Geosciences, University of Edinburgh
[email protected]
ECG Bulletin January 2012
Climate is what we expect; weather is what we get. Thus, climate change is change in the type of weather we expect. In this article, I outline the development of our understanding of how changes to CO2 levels and other greenhouse gases could affect climate. I first describe how climate change is observed, then how the understanding of CO2 as a greenhouse gas arose in the late 1950’s, before describing how climate is modelled. I finish with a short description of the evidence for a human influence on climate and what the future might hold. The material in this article is largely taken from Weart (2008), Edwards (2010), and Solomon et al. (2007). More details may be found in these three publications.
Observations of climate change
Observations of weather began in Europe in the late 17th century and had spread to most parts of the world by the 1950’s. By the late 1930’s G. S. Callendar (1898-1964), a British steam engineer, had compiled weather records, and claimed that the Earth was warming and this warming was being driven by CO2 (Callendar, 1938). His claim was not really accepted at the time. In more recent work (for example Brohan et al., 2006) many more land and marine observations, corrected for changes in observing practice and computed uncertainty estimates, have been compiled. These and other observational datasets (Solomon et al., 2007) show unequivocal evidence of warming over the 20th century.
To extend climate records back prior to the instrumental period requires the use of proxies – biological or geological records of weather over a season or longer. For example, tree rings from carefully selected trees can record the average warmth of the growing season and so can be used to reconstruct climate. The modern instrumental record suggests that climate has warmed by about 0.8 K from 1900-2010 while uncertain proxy records of the last millennium suggest that the 20th century warming is unprecedented.
Observations of weather began in Europe in the late 17th century and had spread to most parts of the world by the 1950’s. By the late 1930’s G. S. Callendar (1898-1964), a British steam engineer, had compiled weather records, and claimed that the Earth was warming and this warming was being driven by CO2 (Callendar, 1938). His claim was not really accepted at the time. In more recent work (for example Brohan et al., 2006) many more land and marine observations, corrected for changes in observing practice and computed uncertainty estimates, have been compiled. These and other observational datasets (Solomon et al., 2007) show unequivocal evidence of warming over the 20th century.
To extend climate records back prior to the instrumental period requires the use of proxies – biological or geological records of weather over a season or longer. For example, tree rings from carefully selected trees can record the average warmth of the growing season and so can be used to reconstruct climate. The modern instrumental record suggests that climate has warmed by about 0.8 K from 1900-2010 while uncertain proxy records of the last millennium suggest that the 20th century warming is unprecedented.
Carbon dioxide as a greenhouse gas
Writing in the early 19th century, the French mathematician and physicist Jean Baptiste Joseph Fourier (1768-1830) suggested that the Earth was warmer than would be expected given the radiation from the sun. By the 1850’s John Tyndall (1820-1893) had shown that water vapour, CO2, and other gases absorbed infra-red radiation and were largely transparent to incoming solar radiation. In the late 19th century, Svante Arrhenius (1859-1927) proposed that CO2 and other atmospheric gases caused the surface warming through their absorption of infra-red radiation, and he calculated how changes in CO2 might warm the Earth’s surface. However, CO2 was seen as opaque and so increases in its concentrations would not affect climate as the “CO2 effect” was saturated.
A group of scientists in 1950’s California then tackled various aspects of the CO2 problem. Gilbert Plass (1920-2004) (whose day job was researching infrared detectors for missiles) performed some calculations on the early computers to show that in the upper atmosphere CO2 did not completely absorb infra-red radiation. This implies that changes in the concentration of atmospheric CO2 could affect climate. Hans Suess (1909-1993) realised that fossil fuel carbon was depleted in 14C as 14C was produced in the atmosphere by cosmic ray bombardment and decays over a few tens of thousands of years. His early CO2 measurements, using this 14C-dating technique, suggested that most CO2 emitted by fossil fuel burning would be taken up by the oceans and so would not affect climate. Oceanographer Roger Revelle (1909-1991) considered the chemistry of sea-water and found that it would take about a decade for the upper ocean to take up CO2. But because of chemical buffering the upper ocean would then emit CO2 back into the atmosphere. This implies that the upper ocean would not take up all the CO2 emitted and some would end up in the atmosphere. Charles Keeling (1928-2005) carried out the first direct and systematic measurements of carbon dioxide in the atmosphere at Mauna Loa, Hawaii in March 1958, and these measurements have continued, despite the foolishness of funding bodies, to this day. In his early work, Keeling showed that there had been a small but persistent increase in atmospheric CO2 concentrations at the South Pole (Keeling, 1960). Since then the Mauna Loa record has found that annual average CO2 concentrations increased from 316 ppm in 1959 to 390 ppm in 2010 (Figure 1).
Writing in the early 19th century, the French mathematician and physicist Jean Baptiste Joseph Fourier (1768-1830) suggested that the Earth was warmer than would be expected given the radiation from the sun. By the 1850’s John Tyndall (1820-1893) had shown that water vapour, CO2, and other gases absorbed infra-red radiation and were largely transparent to incoming solar radiation. In the late 19th century, Svante Arrhenius (1859-1927) proposed that CO2 and other atmospheric gases caused the surface warming through their absorption of infra-red radiation, and he calculated how changes in CO2 might warm the Earth’s surface. However, CO2 was seen as opaque and so increases in its concentrations would not affect climate as the “CO2 effect” was saturated.
A group of scientists in 1950’s California then tackled various aspects of the CO2 problem. Gilbert Plass (1920-2004) (whose day job was researching infrared detectors for missiles) performed some calculations on the early computers to show that in the upper atmosphere CO2 did not completely absorb infra-red radiation. This implies that changes in the concentration of atmospheric CO2 could affect climate. Hans Suess (1909-1993) realised that fossil fuel carbon was depleted in 14C as 14C was produced in the atmosphere by cosmic ray bombardment and decays over a few tens of thousands of years. His early CO2 measurements, using this 14C-dating technique, suggested that most CO2 emitted by fossil fuel burning would be taken up by the oceans and so would not affect climate. Oceanographer Roger Revelle (1909-1991) considered the chemistry of sea-water and found that it would take about a decade for the upper ocean to take up CO2. But because of chemical buffering the upper ocean would then emit CO2 back into the atmosphere. This implies that the upper ocean would not take up all the CO2 emitted and some would end up in the atmosphere. Charles Keeling (1928-2005) carried out the first direct and systematic measurements of carbon dioxide in the atmosphere at Mauna Loa, Hawaii in March 1958, and these measurements have continued, despite the foolishness of funding bodies, to this day. In his early work, Keeling showed that there had been a small but persistent increase in atmospheric CO2 concentrations at the South Pole (Keeling, 1960). Since then the Mauna Loa record has found that annual average CO2 concentrations increased from 316 ppm in 1959 to 390 ppm in 2010 (Figure 1).
Figure 1: Annual atmospheric CO2 concentrations recorded at Mauna Loa, 1958-2010
|
Charles Keeling
The Mauna Loa Observatory, Hawaii
|
By the early 1960’s evidence indicated that when CO2 is emitted into the atmosphere, the atmospheric concentration of this gas increases and could cause climate change. This led to the first report, in 1965, suggesting that CO2 might be a problem, though it was considered unlikely that it would be a problem in the near future. Atmospheric concentrations of other greenhouse gases such as methane, nitrous oxide and the chlorofluorocarbons have also increased over the last 50 years.
As the Greenland and Antarctica ice-caps form, tiny bubbles of air are encapsulated within the ice. The contents of these bubbles can be analysed in ice core samples and tell us how atmospheric concentrations of greenhouse gases have changed over the last 800,000 years. Apparent in these records of past climates are the great swings associated with extensive northern hemisphere glaciations. At times of peak glaciation, atmospheric CO2 levels are about 180 ppm while in inter-glacial periods, they reach values of about 280 ppm. By comparison, current atmospheric CO2 and other greenhouse gas concentrations are unprecedented.
Modelling the global climate
In this section, I briefly outline two different approaches to modelling climate. The work described above had concluded that fossil fuel burning had the potential to affect climate but what was unclear was by how much. The amount of climate change depends on the “feedbacks” in the climate system. For example, water vapour is a greenhouse gas whose atmospheric concentration depends on the temperature. So if the temperature increases then the amount of water vapour in the atmosphere increases. This would then increase the greenhouse effect and thus the surface temperature. Early developments used energy balance models, which represented the fluxes of energy into and out of the Earth. The key ideas are that the outgoing energy flux depends on the greenhouse effect and the surface temperature. The effects of feedbacks are then represented through modifying the energy fluxes via a relationship with surface temperature.
In this section, I briefly outline two different approaches to modelling climate. The work described above had concluded that fossil fuel burning had the potential to affect climate but what was unclear was by how much. The amount of climate change depends on the “feedbacks” in the climate system. For example, water vapour is a greenhouse gas whose atmospheric concentration depends on the temperature. So if the temperature increases then the amount of water vapour in the atmosphere increases. This would then increase the greenhouse effect and thus the surface temperature. Early developments used energy balance models, which represented the fluxes of energy into and out of the Earth. The key ideas are that the outgoing energy flux depends on the greenhouse effect and the surface temperature. The effects of feedbacks are then represented through modifying the energy fluxes via a relationship with surface temperature.
The first attempt to numerically simulate the atmospheric flow for weather forecasting was made by L. F. Richardson (1881-1953), who carried out many of the calculations between ambulance shifts in the first world war (Richardson, 1922). Though the attempt failed, subsequent work built on Richardson’s project. Following the second world war, electronic computers became available and numerical models of the atmosphere were developed. The first numerical simulations were carried out in the USA (Charney et al., 1950). The UK Meteorological Office developed this USA work and instigated numerical methods in the early 1950’s to forecast the weather using the LEO-1 computer (Lyons Electronic Office 1) (Bushby and Hinds, 1954). The Japanese meteorologist Syukuro Manabe (1931- ), working at the Geophysical Fluid Dynamics Laboratory in the mid 1960’s, was one of the first scientists to apply models of the atmosphere and ocean, which work by simulating flows using appropriate and approximate forms of the Navier-Stokes equations on a rotating Earth. One problem for these General Circulation Models (GCMs) is that many phenomena occur on scales which are not explicitly resolved and so their effects on the large-scale flow need to be parameterised. This parameterisation leads to uncertainty in climate prediction. Current GCMs simulate the atmospheric, oceanic and land surface flows on a grid of O(100x100) km. By the late 1970’s two groups had constructed working GCMs, which included representations of the atmosphere, ocean and land-surface. The National Academy of Sciences commissioned a study on the possible effect of CO2 on climate. This study reported that in response to the doubling of CO2, the range of global-mean warming from the two models was 2-3.5 K with more warming at high latitudes. It also concluded, based on expert judgement, that the most probable warming in response to the doubling of CO2 was 3 ± 1.5 K (Charney et al., 1979). These figures were largely supported by Solomon et al. who concluded that it was likely that the response to the doubling of CO2 was in the range 2.5-4.5 K (Solomon et al., 2007).
Human influence on climate and future scenarios
Carrying out controlled experiments on the Earth’s climate is not possible. However, by using GCMs with various different drivers one can compare these models with observations and determine the relative importance of human and natural drivers. Several different GCM’s were constructed either with natural drivers or with both natural (changes in solar irradiance and volcanic effects) and human drivers (CO2, other greenhouse gases and other human drivers). These models were then compared with observations of surface temperature change. Simulations with only natural drivers were inconsistent with the observations while those with natural and human drivers were consistent with observations. This work and other evidence led the IPCC to state “Most of the observed increase in global average temperatures since the mid-20th century is very likely due to the observed increase in anthropogenic greenhouse concentrations” (Solomon et al., 2007).
Using GCM’s, various modelling centres have simulated the possible response of the Earth’s climate to different future emissions of CO2 and other greenhouse gases. These different “scenarios” represent different future pathways of human development with no attempt to reduce CO2 emissions to mitigate climate change. (Two examples are shown on p 36, based on Solomon et al., 2007). Global-mean surface warming by 2100 is dependent on the choice of model and the scenario, but there is agreement that warming this century will warm the world to more than 2K above pre-industrial conditions – a level that the 2009 Copenhagen meeting deemed would represent dangerous climate change. In the scenario with the largest CO2 emissions, models suggest that global-mean warming, relative to pre-industrial conditions, could reach 5K.
Carrying out controlled experiments on the Earth’s climate is not possible. However, by using GCMs with various different drivers one can compare these models with observations and determine the relative importance of human and natural drivers. Several different GCM’s were constructed either with natural drivers or with both natural (changes in solar irradiance and volcanic effects) and human drivers (CO2, other greenhouse gases and other human drivers). These models were then compared with observations of surface temperature change. Simulations with only natural drivers were inconsistent with the observations while those with natural and human drivers were consistent with observations. This work and other evidence led the IPCC to state “Most of the observed increase in global average temperatures since the mid-20th century is very likely due to the observed increase in anthropogenic greenhouse concentrations” (Solomon et al., 2007).
Using GCM’s, various modelling centres have simulated the possible response of the Earth’s climate to different future emissions of CO2 and other greenhouse gases. These different “scenarios” represent different future pathways of human development with no attempt to reduce CO2 emissions to mitigate climate change. (Two examples are shown on p 36, based on Solomon et al., 2007). Global-mean surface warming by 2100 is dependent on the choice of model and the scenario, but there is agreement that warming this century will warm the world to more than 2K above pre-industrial conditions – a level that the 2009 Copenhagen meeting deemed would represent dangerous climate change. In the scenario with the largest CO2 emissions, models suggest that global-mean warming, relative to pre-industrial conditions, could reach 5K.
References
P. Brohan, J. Kennedy, I. Harris, S. F. B. Tett, and P. D. Jones (2006). Uncertainty estimates in regional and global observed temperature changes: a new dataset from 1850. J. Geophys. Res., 111, D12106.
F. H. Bushby and M. K. Hinds (1954). The computation of forecast charts by application of the Sawyer-Bushby two-parameter model. Q. J. R. Meteorol. Soc., 80, 165-173.
G. S. Callendar (1938). The artificial production of carbon dioxide. Q. J. R. Meteorol. Soc., 64, 223-240. (Available from http://www.rmets.org/pdf/qjcallender38.pdf).
J. G. Charney, R. Fjörtoft and J. von Neumann (1950). Numerical Integration of the Barotropic Vorticity Equation. Tellus, 2, 237-254
J. G. Charney [+ co-workers] (1979). Carbon Dioxide and Climate: A Scientific Assessment: Report of an Ad Hoc Study Group on Carbon Dioxide and Climate, National Academies Press, Washington D.C.
P. N. Edwards (2010). A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming, MIT Press, Massachusetts.
C. Keeling (1960). The concentration and isotopic abundances of carbon dioxide in the atmosphere. Tellus, 12, 200-203.
L. F. Richardson (1922). Weather Prediction by Numerical
Process, Cambridge University Press, Cambridge.
S. Solomon [+ co-authors] (2007). Climate Change 2007 – The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge.
S. R. Weart (2008). The Discovery of Global Warming, (revised edn.), Harvard University Press, Cambridge, Massachusetts.
Other sources of historical information on climate change
J. R. Fleming (1998). Historical Perspectives on Climate Change, OUP, New York.
Discovery of Global Warming http://www.aip.org/history/climate/bibdate.htm.
This article is based on a presentation by Professor Tett at the joint ECG/Historical Group Symposium ‘Environmental Chemistry: A Historical Perspective’ held at Burlington House on 26th October 2011.
P. Brohan, J. Kennedy, I. Harris, S. F. B. Tett, and P. D. Jones (2006). Uncertainty estimates in regional and global observed temperature changes: a new dataset from 1850. J. Geophys. Res., 111, D12106.
F. H. Bushby and M. K. Hinds (1954). The computation of forecast charts by application of the Sawyer-Bushby two-parameter model. Q. J. R. Meteorol. Soc., 80, 165-173.
G. S. Callendar (1938). The artificial production of carbon dioxide. Q. J. R. Meteorol. Soc., 64, 223-240. (Available from http://www.rmets.org/pdf/qjcallender38.pdf).
J. G. Charney, R. Fjörtoft and J. von Neumann (1950). Numerical Integration of the Barotropic Vorticity Equation. Tellus, 2, 237-254
J. G. Charney [+ co-workers] (1979). Carbon Dioxide and Climate: A Scientific Assessment: Report of an Ad Hoc Study Group on Carbon Dioxide and Climate, National Academies Press, Washington D.C.
P. N. Edwards (2010). A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming, MIT Press, Massachusetts.
C. Keeling (1960). The concentration and isotopic abundances of carbon dioxide in the atmosphere. Tellus, 12, 200-203.
L. F. Richardson (1922). Weather Prediction by Numerical
Process, Cambridge University Press, Cambridge.
S. Solomon [+ co-authors] (2007). Climate Change 2007 – The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge.
S. R. Weart (2008). The Discovery of Global Warming, (revised edn.), Harvard University Press, Cambridge, Massachusetts.
Other sources of historical information on climate change
J. R. Fleming (1998). Historical Perspectives on Climate Change, OUP, New York.
Discovery of Global Warming http://www.aip.org/history/climate/bibdate.htm.
This article is based on a presentation by Professor Tett at the joint ECG/Historical Group Symposium ‘Environmental Chemistry: A Historical Perspective’ held at Burlington House on 26th October 2011.
