Devil Is In The Details: How Realistic Are The IPCC’s Reference Emissions Scenarios?
6 April 2008
by Jack Rosebro
The four exhaustive Assessment Reports produced by the United Nations Intergovernmental Panel on Climate Change (IPCC) in 1990, 1995, 2001, and 2007 have all served in their time as the most comprehensive literature reviews and syntheses of our influence on the world’s climate as we know it.
In this as well as subsequent installments of our occasional series Devil Is In the Details, we will periodically discuss strengths and weaknesses of those reports in the context of global society and its changing economic, environmental, and social policies and pressures. We will also look at climate data that has come to light since the last components of the most recent Fourth Assessment Report were published late last year.
A Challenge To The Reference Scenarios
Writing in Nature, a team of three scientists recently challenged the foundations of reference emissions scenarios used by the UN’s Intergovernmental Panel on Climate Change (IPCC) in last year’s Fourth Assessment Report (AR4) to estimate a range of possible futures with respect to greenhouse gas production and the effect that it would have on nature and society, if no major climate policies were present.
These scenarios serve as foundations upon which the calculated effects of various mitigation strategies can be added to establish net predicted effects. The predicted effects can then be used to inform policy. Mitigation is defined by the IPCC as “an anthropogenic intervention to reduce the sources of greenhouse gases or enhance their sinks.”
In a commentary entitled Dangerous Assumptions and published online 2 April, Roger Pielke Jr., Tom Wigley, and Christopher Green argue that as much as two-thirds of the energy use reductions and decarbonizations of the energy sector that would be required to stabilize greenhouse gases (GHGs) are “already built into the IPCC reference scenarios” as if they currently existed. The article has met with immediate response, ranging from comments that such an assessment is overdue to outright dismissals of the hypothesis.
To place Dangerous Assumptions in the proper context, it is important to understand that Pielke Jr., Wigley, and Green have focused solely on reference scenarios rather than mitigation scenarios. However, since the reference scenarios affect calculations of the future effects of mitigation policies, the implications of the team’s assertions are broad.
Differentiating Between Reference Scenarios and Mitigation Scenarios
“Developing a response to climate change is characterized by decision-making under uncertainty and risk”, wrote the authors of the IPCC’s 2001 summary on the mitigation of climate change, “including the possibility of non-linear and/or irreversible changes.” But assessment is a necessary prelude to response, and a particularly difficult aspect of the assessment process has been the understanding of complex interactions between climatic, environmental, economic, political, institutional, social and technological processes, as they influence climate change.
For this reason, the IPCC developed reference scenarios sets that reflected a wide range of social and technological changes and included a variety of energy mixes, yet excluded the mitigating effects of overt climate policy. This served to separate the methodology and emissions effects of the reference scenarios from the methodology and subsequent emissions effects of mitigation scenarios, which use the reference scenarios as baselines.
The first set of reference scenarios, known as IS92, was published in 1992. The second and most recent set, which is part of the Special Report on Emissions Scenarios (SRES), was developed by the IPCC between 1997 and 2000 to provide updated scenario sets for use in the Panel’s 2001 Third Assessment Report (TAR) after a 1994 evaluation of IS92 indicated that the development of new scenario sets would be prudent. The SRES was not updated for the Fourth Assessment Report, although a discussion of post-SRES scenarios was included.
The IPCC’s 2001 Third Assessment Report and 2007 Fourth Assessment Report utilize both reference and mitigation scenario sets as follows:
Reference scenarios from 2000 to 2100 project varying business as usual (BAU) effects of population changes, economic growth, energy use, and energy mixes on global greenhouse gas production in the absence of climate change mitigation policy.
Mitigation scenarios from 2000 to 2100 project policy- and technology-driven greenhouse gas emissions reductions needed to achieve a specific peak and decline of atmospheric GHGs, with a specific peaking level and year.
Mitigation scenario sets are discussed by Working Group III in both Assessment Reports.
Although the SRES is based on scenarios that were developed or collected some time ago, it describes a wide range of emissions. Hence, the age of the data is not generally viewed as a problem when that data is used in climate modeling. However, when the SRES is used to run climate mitigation models, which project the effects of various policies, the results of those models are affected by baseline assumptions which are already embedded in the SRES.
These assumptions are at the heart of Pielke, Wigley, and Green’s concerns, who argue that the reference SRES scenarios are “optimistic at best and unachievable at worst, potentially seriously underestimating the scale of the technological challenges associated with stabilizing greenhouse gas concentrations.”
Scenarios, Scenario Groups, and Storylines: How Reference Scenarios Are Structured
But exactly what are today’s reference scenarios, and how were they developed?
In January 1997, the IPCC Working Group III selected a writing team to produce new reference scenarios, and placed advertisements in scientific journals in an effort to solicit wide participation in the process. Global modeling teams as well as regional modelers from both developing and developed regions were encouraged to participate. To facilitate participation, an open-process web site was created. The SRES approach and process is described online, and the open process website is archived online, as well..
The scenario-building process divided possible futures, as represented by contemporary academic literature, into storylines that represented available combinations from a double axis of global-to-regional as well as economic-to- environmental development. Catastrophic scenarios were ignored, as it was assumed that any chain of events which led to societal or economic collapse would render the immediate mitigation of climate change to be a low priority.
|Figure 1. Conceptual diagram of the interaction of social, economic, environmental, and technological forces that shaped the four SRES scenario storylines. Click to enlarge.|
By the end of 1997, four preliminary storylines—also referred to as “marker scenarios”—had been developed by the writing team in an iterative process. The storylines provide alternate “future histories” of how global regions interrelate, how new technologies diffuse, how regional economic activities evolve, how protection of local and regional environments is implemented, and how demographic structure changes.
The four storylines are:
A1 Global-Economic: “rapid and successful economic development” worldwide;
A2 Regional-Economic: “differentiated” world, rich and poor, with energy type determined by available resources;
B1 Global-Environmental: a “globally coherent approach to a more sustainable development”; and
B2 Regional-Environmental: community-based efforts toward environmental and social sustainability.
|Figure 2. Four storyline families, six scenario groups, and 40 individual scenarios comprise the basis for the IPCC’s 2000 Special Emissions Report on Scenarios. 35 of the scenarios were used in the Fourth Assessment Report. Click to enlarge.|
Each storyline is based on six to seventeen scenarios, depending on availability at the time the SRES was developed.
As the A1 storyline was at the time represented the most by academic literature, it was further developed into a family of six scenario sub-types:
A1B “no single source of energy is overly dominant”;
A1C a primarily coal-based "“synfuel society”;
A1FI “fossil-intensive”; conventional oil, gas, and coal is dominant;
A1G a “massive development of unconventional oil and gas resources”; and
A1T a “technology-intensive” society, moving away from fossil fuels, with low energy demands
|Figure 3. Qualitative directions of each scenario group’s primary indicators. Note that in this matrix, each of the six SRES scenario groups is referred to as “scenario”. Click to enlarge.|
Scenario types A1C and A1G were subsequently discarded, leaving a family of three A1 variations: A1FI, A1T, and A1B, as well as A2, B1, and B2. A matrix of the qualitative directions of each scenario group’s primary indicators is reproduced in Figure 3.
The Hypothesis of “Dangerous Assumptions” And Its Critics
The SRES employs four key metrics: population; GDP per capita; energy intensity, measured in primary energy use per capita; and carbon intensity, measured in carbon-dioxide emissions per unit of energy.
In the SRES, it is assumed that technological advances will reduce both energy intensity and carbon intensity. However, the energy and carbon intensities of developing countries can increase sharply during periods of intense growth, declining only after the economies of those countries have matured.
In the IPCC’s most recent report (AR4), the authors noted that “the projected emissions of energy-related CO2 in 2030 are 40-110% higher than in 2000.”
While the authors of “Dangerous Assumptions” do not question the chosen metrics, they state that a “frozen technology baseline” should be used in place of scenarios that assume a society which will move, albeit slowly, toward a more efficient, decarbonized society even without significant climate change policy.
As the SRES projections run from 2000 to 2100, the first seven years of those projections can be compared to actual data. Pielke Jr., Wigley, and Green did just that for the 2000-2005 period, and argue that the IPCC assumptions for decarbonization in the short term (2000-2010) are already inconsistent with socio-economic development.
|Figure 4. Comparison of actual 2000-2005 data for global energy and carbon intensity with the 35 scenarios used in the Fourth Assessment Report. Click to enlarge.|
All scenarios predict declines in energy intensity, and in most cases carbon intensity, from 2000 to 2010. However, the study found that energy intensity, as measured in energy use per unit of GDP, actually increased from 2000 to 2005. The carbon intensity of energy use also increased during that time. Each metric naturally intensifies the other with regard to the production of greenhouse gases, as seen in figure 4.
Bert Metz, who was one of the lead authors of the IPCC’s Working Group III and contributed to the SRES reference scenarios, forcefully responded:
The claim that the IPCC has underestimated the technological challenge of stabilization is unwarranted and must be rejected... The assumptions about the rate of technological change in these scenarios have been thoroughly reviewed and are accepted by the community of technological-change experts. They confirm well-known facts about, for instance, the enormous improvements in computers over much shorter time-frames than expected. The assumptions also reflect that high economic growth normally goes hand in hand with high rates of technological change.
Comparing the historically exponential increase in computing power to anticipated changes in society’s energy supply system, however, is an “apples and oranges” exercise, at best.
“There is a great deal of inertia in the modern energy system, given its vast complexity and scale”, notes Jeremy Bentham of Shell International. “The often lengthy timescales required for planning and constructing new energy infrastructure mean that strains within the system cannot be resolved easily or quickly, if at all.”
Richard Tol, the energy economist at the Economic and Social Research Institute in Dublin, who famously criticized the UK’s 2006 Stern Review Report on Economic Change, commented:
The IPCC scenarios developed in 2000 don’t match historical observations. For instance, nobody foresaw the rapid growth in recent years of China’s economy, or the launch of the USD 2500 people’s car in India... We need new scenarios—not just any, but scenarios that are in line with the real development of global energy systems.
Tol’s concerns come on the heels of last month’s report Forecasting the Path of China’s CO2 Emissions Using Province Information (earlier post) by Maximilan Auffhammer and Richard T. Carson, which used a unique provincial-level panel data set from the Chinese Environmental Protection Agency to calculate China’s CO2 emissions. Auffhammer and Carson found that China’s average annual CO2 growth rate from 2000 to 2004 was 14.53%, and estimated the average annual CO2 growth rate from 2000 to 2010 to be in excess of 11%.
By contrast, the SRES had assumed that China’s average CO2 growth rate from 2000 to 2010 to be between 2.53 and 4.82% per annum. This estimate is used in all six of the IPCC’s scenario groups.
Detlef van Vuuren, an emission scenario developer at the Netherlands Environmental Assessment Agency, countered:
It is not very helpful to discredit emission scenarios on the sole basis of their being at odds with the most recent economic trends in China. Chinese statistics are not always reliable. Moreover, the period in question is too short to signify a global trend-break. Rapid economic growth, in combination with the high price of oil, might spur long-term developments in renewable-energy technologies, for instance.
Alternate Views: Scramble and Blueprints
|Figure 5: Coal has an ascending role 2000-2050 energy mixes in Shell’s Scramble scenario, as compared to Blueprints. Click to enlarge.|
The SRES, while assuming spontaneous advances in technology, makes little mention of the timing of social and technological change. This is often regarded to be a crucial factor in scenario planning, For example, Shell’s recently released pair of energy scenarios for 2050, Scramble and Blueprints (earlier post), envision near-identical futures by 2050, separated by at least two pivotal differences: grass-roots awareness and political will.
Both of Shell’s scenarios assume that the future will bring with it “three hard truths”:
Step changes (unexpected spurts of growth) in energy demand;
Energy supply will struggle to keep pace with demand; and
Environmental stresses will increase.
In Scramble, “policymakers pay little attention to more efficient energy use until supplies are tight”, while in Blueprints, “local actions begin to address the challenges of economic development, energy security and environmental pollution”. Put another way, Scramble starts out as an A1FI wold, while Blueprints is closer to a B2 world. In both scenarios, society eventually transforms into an A1T world, but Scramble ends up with more GHG emissions, and is thereby exposed to more risk.
Land Use, Land Use Change, and Forestry
|Figure 6. SRES projected effects on global CO2 emissions from land use changes from 2000 forward. Negative numbers indicate a mitigating effect on global CO2 emissions. Click to enlarge.|
All six SRES scenarios groups calculate the effects of human-induced land-use changes on greenhouse gas production. The difficulty of accurately assessing land-use change is reflected by the broad range of estimated effects by 1990, as reported in the Second Assessment Report (SAR) of 1995, and seen in Figure 6.
However, almost all scenarios project steady improvements in the effects of land use on CO2 emissions. Only the most pessimistic scenario group, as seen in Figure 6, depicts a future in which land-use changes maintain a net increase in greenhouse gases from 2080 forward.
The remaining five scenario groups project reductions starting between 2020 and 2070. In other words, the effects of global afforestation are generally projected by the SRES to outstrip the detrimental effects of deforestation in the absence of any meaningful policy. The scenarios attribute this to slowing (and in some cases declining) population growth, increased agricultural productivity, and—oddly—“increasing scarcity of forest land.”
But global land-use change since 2000 has proven to be somewhat different so far; the scenarios of 1997-1998 did not forecast the biofuels boom of recent years, nor the increasing demand for meat (and subsequently for cattle, raised on land that could otherwise support much more efficient calorie production from crops) in developing countries. The UN Food and Agriculture Organization State of the World’s Forests 2007 biennial assessment estimates that the world’s forested areas declined by about three percent between 1990 and 2005.
The actual decline is difficult to track. Illegal logging is on the rise in Central Africa, the Amazon, Russia and Indonesia, among other areas, and has already infiltrated legitimate markets. It is estimated to provide, for example, about half of Europe’s lumber imports. Last month, the European Union vowed again to tighten controls on the import and sale of illegally cut timber, as it had in 2003, 2005, 2006, and 2007.
Although the annual rate at which the Amazon rainforest is being cleared declined to an estimated 9,600 km2 (3,000 square miles) last year, the lowest rate of destruction since the 1970s, the rainforests of Indonesia, much of which grows on peat carbon sinks, have been much harder hit by deforestation, primarily because of logging and the development of palm oil plantations. A 2006 Dutch study estimated that Asia’s peatland CO2 emissions already equal almost 8% of the world’s emissions from fossil fuel burning, and that peatland burning has made Indonesia into the number three emitter of greenhouse gases, behind the US and China. A later study increased the estimated emissions to 10%.
Rainforests in Indonesia are traditionally cleared with fire, which is extinguished by seasonal rains. However, a strong El Niño-Southern Oscillation (ENSO) in 1997 disrupted the traditional monsoonal season. Fires burned out of control well into 1998, consuming an estimated 9,655,000 hectares (37,000 square miles), according to the Asian Development Bank Project.
Of Inventories and Feedbacks
Attempts to inventory greenhouse gases are often rough guesses at best, although methods are continually being improved. IPCC emissions scenarios have assumed, for example, an annual output of 400 million tonnes of CO2 from the global shipping sector, declining in both energy and carbon intensities as years go by. However, The Guardian (UK) reported in February that a study commissioned by the International Maritime Organization puts the actual amount at 1.12 billion tonnes in 2007.
When contacted by The Guardian, Dr. Rajendra Pachauri, chair of the IPCC, said:
This is a clear failure of the system. The shipping industry has so far escaped publicity. It has been left out of the climate change discussion. I hope [shipping emissions] will be included in the next UN agreement. It would be a cop-out if it was not. It tells me that we have been ineffective at tackling climate change so far.
But the source of the most significant and persistent emissions not yet counted by the SRES may well turn out to be nature itself. On page 6 of the SRES Summary for Policymakers, a footnote observes that in the 2000 scenarios “no feedback effect of future climate change on emissions from the biosphere has been assumed.”
Potential tipping elements in the biosphere such as the collapse of the Atlantic thermohaline circulation (popularly called the “ocean conveyor belt”), dieback of the Amazon rainforest, and the decay of the Greenland ice sheet had all been identified by 2000. However, such data was preliminary, and the IPCC tends to view new data conservatively.
Data collected since then has to a large part confirmed the increasing fragility of these as well as other subsystems of nature, such as the Indian summer monsoon, boreal forests, and the El Niño/Southern Oscillation (ENSO). The collapse or decay of these subsystems would contribute to additional warming (earlier post), although corresponding declines in economic activity could lead to a trend towards lower emissions.
In addition, evidence began to emerge last year that the Earth’s natural carbon sinks are weakening. The first paper to deal with this possibility addressed the planet’s largest carbon sink, the Southern Ocean, and projected that at the present rate of deterioration, the Southern Ocean will have lost two-thirds of its ability to store carbon by 2050. (Earlier post.)
The IPCC will consider next week at its 28th Session in Budapest, Hungary whether to continue the five-to-six year interval between Assessment Reports or to lengthen the interval between reports. Governments generally support the current interval, while authors generally support a longer interval to provide enough time to assess the data. If the current interval is retained, the Fifth Assessment Report (5AR), upon which work has already begun, will likely be completed in 2012 or 2013. Work began on new reference scenarios in 2006, and the Panel has already signaled that those scenarios are likely to be completed in time for 5AR. If so, they will represent the first new look at the IPCC’s view of business as usual GHG emissions, upon which all mitigation scenarios rest, in more than a decade.
 Maximilian Auffhammer and Richard T. Carson: Forecasting the Path of China’s CO2 Emissions Using Province Information
 Delft Hydraulics: PEAT-CO2 assessment of CO2 emissions from drained peatlands in SE Asia (2006)
 John Vidal, The Guardian: True Scale of CO2 Emissions from Shipping Revealed
 Le Quéré et al: Saturation of the Southern Ocean CO2 Sink Due to Recent Climate Change (2007)
TrackBack URL for this entry:
Listed below are links to weblogs that reference Devil Is In The Details: How Realistic Are The IPCC’s Reference Emissions Scenarios?: