Warning This Web page has been archived on the Web.

Archived Content

Information identified as archived on the Web is for reference, research or recordkeeping purposes. It has not been altered or updated after the date of archiving. Web pages that are archived on the Web are not subject to the Government of Canada Web Standards. As per the Communications Policy of the Government of Canada, you can request alternate formats on the Contact Us page.

Skip booklet index and go to page content

Frequently Asked Questions about the Science of Climate Change – 2008 Update

C. Detecting and Attributing Climate Change

C.1 Has the world warmed?

Response: Yes, unequivocally so. During the 100-year period between 1906 and 2005, the global average temperature at the Earth's surface has warmed by about 0.74°C. There are also many other indicators of a warming world. These include warming of the lower atmosphere and the upper layers of the world's oceans, rising temperatures in terrestrial soils, melting mountain glaciers, retreating sea ice and snow cover, rising sea levels, and shifts in distribution of many species of plants and animals (see Figure C.1).

Explanation: Experts in global temperature trend analyses report that, during the 100 years ending in 2005, global average surface temperature has warmed by at least 0.56°C, and possibly by as much as 0.92°C, with a best estimate of 0.74°C. The linear trend for the last 50 years of this record is about 0.13°C per decade, about twice that for the entire 100 years. Eleven of the 12 years between 1995 and 2006 are the warmest on record since the beginning of the global instrumental record some 150 years ago.

It is important to note, however, that these trends refer to global average surface warming. In some areas, particularly over continents, the magnitude of warming has been several times greater than the global average. In Canada, for example, there has been an increase in average annual temperature of about 1.4°C over the period 1948-2007. On the other hand, in a few areas around the world, temperatures have actually cooled.

A variety of other climate variables also provide supporting evidence of a warming climate. These include: a warming of the lower 6km of the atmosphere similar to that at the surface; a rise in average ocean temperatures down to a depth of 3000m; a 5% reduction in the extent of spring snow cover in the Northern Hemisphere since the late 1960s; a coincident reduction in Northern Hemisphere lake ice cover seasons; a reduction in Arctic annual average sea ice cover since the late 1970s of about 8% and a considerable decline in sea ice thickness; an 8cm global sea-level rise since 1960; an increase in global heat content in the Earth's soils, its cryosphere and its oceans since adequate measurements began in the 1950s; and a decrease in the frequency of cold days and nights and an increase in the frequency of hot days and nights, and heat waves, since the 1950s.

References: Trenberth et al., 2007; Lemke et al., 2007; Environment Canada, 2008.

Figure C.1. Global surface temperature records. Refer to caption beneath image for description.

Figure C.1. Global surface temperature records, adjusted for biases due to changing observing practices and influences such as urbanization, show that recent temperatures are significantly warmer than temperatures one century ago. Furthermore, the rate of warming appears to be accelerating. Sea level is also rising, while Northern Hemisphere snow cover is receding (Figure SPM.3, IPCC, 2007a).

C.2 How are the global average temperature records developed?

Response: The records for global average surface temperature are based on data collected over the past 150 or so years at thousands of weather and climate stations on land areas around the world, by ships at sea, or more recently by ocean data buoys. These data are carefully averaged over the entire globe in a manner that avoids biases towards regions with high concentrations of data. Although some areas of the world had very sparse or no coverage during the early part of the record, and there are a number of other problems associated with using these data to estimate global temperature trends,
experts have worked for many decades on addressing these concerns.

Explanation: Daily temperature data have been recorded at thousands of weather and climate stations around the world for most of the past century, with many of these station records extending back in time to 1850 and earlier. These temperatures are recorded at a height approximately one metre above the ground surface. In addition, crew members from ships navigating the world's oceans have been collecting daily temperatures for both sea surface waters and air at the level of the ship's deck. Many of these records, particularly those for the 19th century, were logged from British naval vessels. More recently, ocean data buoys have also provided additional data. There are several teams of experts within the international research community that have devoted considerable efforts at developing a comprehensive picture of the change in average surface temperatures around the world from these data. To do so, they must develop statistical techniques for properly averaging the data, addressing significant data gaps, and correcting for other problems associated with how the data have been collected. Since these groups use different techniques, the results differ slightly. However, this also helps to improve confidence in the results.

The corrections and adjustments that need to be made to the data are significant. For example, the abundance of good quality data is generally far greater in southern Canada, the USA and western Europe than in many other parts of the world, particularly for the 19th century. Some parts of the world, such as much of Africa and Antarctica, have very little data, even today. Methods of developing global temperature trends need to address this geographic imbalance in the data. Experts also screen the data to remove stations that are unreliable. These include stations in urban centres that have been influenced by the effect of urbanization on local temperatures. They also adjust the data, where possible, for influences caused by changes in recording methods or location of the station.

To average the selected data, the experts divide the Earth's surface into a grid of equal sized regions and develop one composite record for each of these by using advanced techniques for averaging the available data within that area when there are multiple records and for interpolating from adjacent grid areas when there are no data. Experts continue to make improvements in their methodologies, and test them against other data sources to determine how successful they are. While there are continuing uncertainties associated with the trend analyses that emerge from these data compilations, for the past century these now appear to be constrained to about +/- 0.2°C.

References: Folland et al., 2001; Trenberth et al., 2007.

C.3 Is the temperature record reliable?

Response: Yes, the data used to calculate global temperature trends provide a good indication of how our climate is changing. As required for proper use of data from all monitoring programs, the climate data used to estimate the global temperature trend are first subjected to quality control procedures and evaluated for systematic sources of error. In addition to deleting records with major errors or nonclimatic influences and correcting others where the error is readily identifiable, climate scientists also compare the instrumental climate records with those derived from other sources. To allow for any remaining non-climate factors affecting these records, experts provide a margin of error in their estimates. They state with high confidence that the global average surface temperature during the 1906-2005 period has risen by at least 0.56°C, and not more than 0.92°C.

Explanation: One method of dealing with random errors that occur at single stations is to average the temperature values over many stations. Global temperature analyses use many thousands of stations, and hence such random errors are largely removed through averaging. Systematic changes that are unrelated to climate but that can affect many or all of the records at the same time or in the same way are more difficult to remove. These include changes in observed values due to urban heat island effects, large-scale changes in instrumentation, changes in the density of recording stations, or a systematic shift in the location of instruments at weather stations. These can be at least partially addressed through careful analysis and adjustments. In undertaking the global trend analyses, climate experts have made careful allowance for a number of such systematic influences, including the heat island effect (see Figure C.3), the change in observing processes on ships, and other non-climatic influences on observations. There remains solid evidence that the warming of the recent decades is real and global. Furthermore, surface temperature records are in good agreement with the long-term trends apparent in radiosonde measurements and satellite data collected for the lower 6km of the atmosphere during recent decades. They also agree with evidence from tree rings, and with information obtained from bore holes drilled into the Earth's surface in different parts of the world. Finally, they are also consistent with concurrent trends towards reduced global snow cover, glacier retreats and other indicators of a warming world. However, because of the uneven global distribution of observation sites, climate records are still dominated by land data obtained in the Northern Hemisphere. Considering these uncertainties, the science community estimates that the Earth's surface has, on average, warmed by 0.74 ± 0.18°C over the past 100 years.

Figure C.3. Comparison between temperature trends of the full corrected land data set used in global temperature trend analysis and a subset of rural stations. Refer to caption beneath image for description.

Figure C.3. Comparison between temperature trends of the full corrected land data set used in global temperature trend analysis and a subset of rural stations suggests there is very little residual effect of urbanization remaining in the data (Peterson et al., 1999).

C.4 How unusual has recent warming been?

Response: Although the observed average global warming of 0.74°C during the past 100 years seems modest, its significance can be assessed against reconstructed climates for the Earth's surface for previous centuries and even millennia. Such comparisons indicate that, at least for the Northern Hemisphere, the second half of the 20th century was likely the warmest 50-year period of at least the past 1300 years. Furthermore, climate model studies indicate that it is very difficult to replicate the climate trends of the past millennium without including the role of anthropogenic climate forcings.

Explanation: Researchers have indirectly collected information about past climates from various indicators such as tree rings, ice cores and ocean corals. Individually, these indicators provide information on only some aspects of hemispheric climate. Tree rings, for example, are useful indicators of average temperatures during growing seasons in mid-latitudes, or of precipitation changes in arid regions. Ice cores provide information on temperatures in cold regions, while ocean corals can help reconstruct temperatures in tropical ocean climates. Analysts can use statistical techniques to combine these various information sources into a single hemispheric temperature reconstruction. There has been considerable debate in recent years about the reliability of such reconstructions (often referred to as the 'hockey stick' debate since the reconstructions show a relatively stable millennium (the shaft) followed by a rapid warming over the 20th century (the blade)). However, recent assessments by teams of experts have concluded that, at least for the Northern Hemisphere, the second half of the 20th century was very likely the warmest in the past 500 years, and likely in the past 1300 years.

Paleoclimate model simulations are broadly consistent with the reconstructed Northern Hemisphere temperatures over the past 1000 years. These simulations also indicate that the rise in surface temperatures since 1950 very likely cannot be reproduced without including anthropogenic greenhouse gases in the climate forcings5 used in the model. Furthermore, it is very unlikely that this warming was merely a recovery from a pre-20th century cold period. There is less confidence in similar conclusions for the Southern Hemisphere, since the available data for that region is as yet very sparse.

References: National Research Council, 2006; Jansen et al., 2007.

5 See A.2 for description of climate forcing.

Figure C.4. Various studies have used proxy indicators of past temperatures, including tree rings, ice cores and corals, to reconstruct Northern Hemisphere temperatures over the past several millennia. Refer to caption beneath image for description.

Figure C.4. Various studies have used proxy indicators of past temperatures, including tree rings, ice cores and corals, to reconstruct Northern Hemisphere temperatures over the past several millennia. While there is disagreement on the magnitude of hemispheric temperature variations prior to the 20th century, these studies, and other evidence, indicate that the late 20th century warmth is unprecedented in the context of at least the last millennium (Figure S.1, National Research Council, 2006).

C.5 How do scientists examine the question of what caused the warming?

Response: Scientists have studied possible causes of climate change for many years. These causes include changes in solar radiation, the effect of volcanic eruptions on climate, and the role of greenhouse gases and aerosols that humans release into the atmosphere. Researchers have been able to reconstruct, with varying degrees of confidence, how these different forces for climate change have varied over the past decades and centuries. Climate models are then used to simulate how these different climate forcing factors should have affected the global climate over space and time. By comparing the expected response of the climate to different forcings with that which actually occurred, researchers are able to attribute the causes of global changes in climate, including the recent warming, with considerable confidence. The conclusion of such work is that most of the warming of the past 50 years has been due to human influences. However, attributing changes at the regional scale, where natural climate variability becomes more significant, is much more difficult.

Explanation: Anything that causes a persistent change in the radiative balance between incoming solar and outgoing infrared radiation at the top of the atmosphere is, in essence, a force causing the Earth's climate to change, hence a 'climate forcing'. There are four broad categories of climate forcings that operate on time scales relevant to human lifetimes: i) changes in solar irradiation at the top of the atmosphere; ii) changes in the concentrations of aerosols and cloud particles within the atmosphere that reflect and scatter incoming solar radiation back to space and absorb outgoing heat radiation; iii) changes in the Earth's surface that affect both the amount of incoming solar radiation reflected back to space at the surface and the amount of heat energy released from the surface towards space; and iv) changes in the concentration of greenhouse gases that absorb and retain outgoing heat radiation.

Researchers have used tree rings, ice cores and other proxy indicators to help reconstruct past changes in most of the key climate forcings, including those due to solar radiation, changes in volcanic aerosol concentrations in the atmosphere, and changes in concentrations of greenhouse gases and of anthropogenically generated aerosols in the atmosphere. However, the impacts of these changes in forcings on climate involve many complex feedbacks within the climate system that require the most sophisticated climate models to properly simulate. Such simulations can estimate the pattern of change, vertically, horizontally and in time, that might be expected for each of the forcings individually and in combination with each other. By comparing these simulation results with observed patterns, experts can help identify which combination of factors have, for example, likely caused the global scale warming of recent decades. Results indicate that global trends of the past century cannot be explained if only natural climate forcings are considered. However, there is good agreement when anthropogenic forcings are included (see Figure C.5). In fact, the evidence supports the conclusion that most of the warming during the past 50 years was due to human influences. Similar attributions are now also available at the continental scale. However, because of the much greater variability of climate at the regional scale and the higher complexity of regional feedbacks, attribution of changes at this scale to specific global forcings is, in general, not yet feasible.

Reference: Hegerl et al., 2007.

Figure C.5. Comparison of global mean surface temperature anomalies from observations with those estimated by climate model simulations. Refer to caption beneath image for description.

Figure C.5. Comparison of global mean surface temperature anomalies (change in temperature relative to 1901-1950 period) from observations (thin black line) with those estimated by climate model simulations forced with a) natural and human factors and b) natural influences only. Results both illustrate how well models can replicate observed climate change when forced with all leading causes of change and how poorly they replicate observed changes when only natural forcings are included (Fig. 9.5, Hegerl et al., 2007).

C.6 Could changes in the cosmic radiation from outer space have caused the warming?

Response: No. While some scientists have hypothesized that changes in cosmic radiation could change global cloud cover, and hence surface temperatures, observational data of cloud cover in recent decades do not support this. Over the past two decades when global warming has been strong, the trend in cosmic ray intensity has been in the opposite direction to that required to explain the warming.

Explanation: Cosmic radiation consists of energetic particles, such as protons and small atomic nuclei, that originate from outer space, bombarding the Earth's atmosphere. It has been hypothesized that changes in cosmic radiation could have been a dominant influence on Earth's climate over the past 500 million years. Furthermore, it is suggested that, if such processes could affect climate on multi-million year time scales, they might also do so on century time scales. The hypothesis is that cosmic radiation can ionize aerosols within the atmosphere, and thus affect cloud formation processes. Since clouds generally have a cooling effect on climate (by reflecting solar radiation), some scientists have speculated that during periods with relatively high levels of cosmic radiation, more clouds would form and the Earth should cool. conversely, fewer cosmic rays should lead to a warmer Earth. However, studies conducted to date do not support this hypothesis. Analyses of observational data over recent decades show no linkage between fluctuations in cosmic radiation and global cloud cover. Furthermore, over the past two decades there is evidence that both cosmic ray fluxes and surface temperatures have been increasing. That is, the trend in cosmic ray intensity has been in the opposite direction to that required to explain the increasing temperature trend. In fact, variations in global temperature over the past century are well explained by changes in anthropogenic and other natural factors, namely changes in greenhouse gas and aerosol concentrations, in solar radiation and volcanic eruptions (see C.5).

References: Rahmstorff et al., 2004; Forster et al., 2007; Lockwood and Frölich, 2007.

C.7 Can solar irradiance changes have caused the warming of the past century?

Response: Changes in solar irradiance can explain part of the warming, particularly in the early part of the 20th century. However, average sunshine reaching the Earth has not changed significantly over the past 50 years, and therefore cannot explain the rapid warming of recent decades.

Explanation: Solar irradiance undergoes an approximate 11-year sunspot6 cycle, varying from minimum to maximum sunspot numbers on the Sun's surface and back again. However, while these cycles may be a contributing factor to decadal variability of climate, they do not significantly affect long-term climate trends unless the nature of the cycle itself changes. There is considerable evidence to indicate that the amplitude of the sunspot cycle slowly became larger over the past few centuries, until about 1950. Radiation experts estimate that this may have caused a small positive (warming) radiation imbalance at the top of the atmosphere over the past century of, at most, a few tenths of a W/m2, or about 10% of that due to increasing greenhouse gas concentrations (see Figures B.13 and C.8). However, although this forcing has varied over the past 50 years, first decreasing then increasing, its long-term average over that period has changed very little. Therefore, it is not considered a significant contributor to the rapid warming of recent decades.

References: Hansen et al., 2005; Forster et al., 2007.

6A sunspot is a region on the Sun’s surface that is marked by a lower temperature and appears darker than its surroundings.

C.8 What is the role of volcanoes in the recent warming?

Response: Volcanic eruptions periodically eject aerosols into the stratosphere where they can remain for several years. This can have a short-term cooling effect on the climate because these volcanic aerosols reflect sunlight. Average concentrations of these aerosols can also change over longer periods of time as the frequency and intensity of volcanic eruptions varies over time. Therefore, volcanic aerosols can also become a long-term climate forcing. This forcing causes a cooling effect when the average concentrations rise above the long-term average (because of increased reflection of sunlight) and a warming effect when they decline below the long-term average (since they now reflect less sunlight than normal). Such a decline occurred between 1900 and 1950, likely contributing to warming of the globe during the early 20th century. However, an increase in the number of large volcanic eruptions in recent decades has reversed this trend. While such eruptions have significantly affected global climates for short periods of time, they cannot explain the recent warming trend. Rather, the rise in average volcanic aerosol concentrations in recent decades should have caused a cooling trend.

Explanation: Sulphur gases released during large volcanic eruptions can cause a dramatic increase in the concentrations of sulphate aerosols in the stratosphere, where they reflect incoming sunlight back to space. This can have a large and abrupt cooling effect on the Earth's surface temperatures. For any given eruption, the cooling is short lived, since these aerosols only remain in the atmosphere for a few years before settling back to the Earth's surface. However, during periods of time with frequent large eruptions, the average concentrations of these aerosols (and hence cooling influence) are higher than during periods of time with fewer eruptions. Between 1900 and 1950, there was a decline in the frequency of large volcanic eruptions and in the mean concentration of related aerosols. This contributed to the warming of the early 20th century. However, in recent decades, such eruptions have become more frequent, once again increasing the net cooling effect (see Figure C.8).

Reference: Hegerl et al., 2007.

Figure C.8. Changes in natural and anthropogenic climate forcings since 1850. Refer to caption beneath image for description.

Figure C.8. Changes in natural and anthropogenic climate forcings since 1850. The cooling influence (negative forcing) of volcanic eruptions over the second half of the 20th century is shown clearly here, as is the strong warming influence (positive forcing) from long-lived greenhouse gases (LLGHG) (Fig 2.23 upper panel, Forster et al., 2007).

C.9 Why do scientists point to greenhouse gases and anthropogenic aerosols as the reason for recent warming?

Response: While climate model simulations that include only the natural forcings of the climate system due to variations in solar radiation and volcanic aerosol concentrations project that the Earth should have cooled in recent decades, simulations that include changing concentrations of greenhouse gases and human-induced atmospheric aerosols replicate the recent warming remarkably well. Furthermore, the vertical, horizontal and temporal patterns of observed changes in temperature agree with the pattern expected due to these human factors, and not with those due to natural forcings. Finally, the unusualness of the recent warming within the past millennia also suggests that human factors are a likely cause for the recent warming.

Explanation: There are a variety of indicators that have led scientists to agree that most of the warming during the past 50 years is very likely due to the effects of rising greenhouse gas concentrations, partially masked by concurrent influences of anthropogenic aerosol emissions (which have a cooling effect). First, model simulations using natural forcings due to solar and volcanic activities only, cannot explain the recent warming (see Figure C.5). In fact, all else being equal, these should have caused a cooling. Second, the unusual nature of the recent trends within the context of climate variability of the past two millennia indicate that they are also unlikely due to the combination of natural forcings and natural variability. Third, the recent trends are remarkably similar to those simulated by climate models when forced by human-induced changes in greenhouse gas and aerosol concentrations. Finally, the spatial pattern (or 'fingerprint') of change is very similar to that expected due to the human forcings, but not the natural ones. This pattern is a composite of a relatively uniform warming due to rising greenhouse gas concentrations and a cooling influence from industrial aerosol emissions that is highly variable from region to region, particularly in the Northern Hemisphere.

Reference: Hegerl et al., 2007.

C.10 A large increase in temperature occurred in the early part of this century when emissions of CO2 and other greenhouse gases were still relatively low. However, temperatures actually cooled in the 1950s and 1960s when these emissions began to increase rapidly. Doesn’t this contradict the idea that increased greenhouse gas emissions will cause warmer climates?

Response: Rising concentrations of greenhouse gases is only one of a number of factors that affect the climate system. Other factors important on decadal and longer time scales include natural forcings due to solar and volcanic activity, human emissions of aerosols (particularly sulphates), and natural internal variability of the climate system. The short period of slight cooling in mid-century appears to be at least partially linked to a period of rapidly rising aerosol concentrations that coincided with a cool phase in natural decadal climate variability. Since then, acid rain control measures have helped to reduce aerosol concentrations in the Western Hemisphere, while greenhouse gas concentrations continue to increase rapidly. Hence, the role of greenhouse gas forcing has become increasingly dominant relative to that due to aerosols. As shown in Figure C.5, models can now replicate that pattern of change with time remarkably well.

Explanation: During the first half of the century, increasing solar radiation intensity, declining volcanic aerosol concentrations and human activities (particularly those causing a rise in greenhouse gas concentrations) were all contributors to the modest rise in observed global average surface temperature (see Figure C.5). However, from the mid-20th century to about 1980, anthropogenic emissions of sulphates into the atmosphere over North America and Europe rose rapidly. These aerosols reflect sunlight and their increased concentrations caused rapid regional cooling at the surface that affected global temperatures and offset much of the warming caused by rising greenhouse gas concentrations. Greenhouse gas emissions also increased more rapidly during the same period, but climate system response times to changes in greenhouse gas concentrations are slower than for aerosols. Furthermore, the North Pacific and North Atlantic Oceans underwent a period of regional surface cooling in the 1950s and 1960s as part of a long-term multi-decadal natural climate oscillation. This combination of natural variability and sulphate cooling appears to have been enough to offset the enhanced greenhouse effect - which was still relatively modest at the time. During the 1980s, most of the industrialized world began stringent acid rain control programs that helped to decrease sulphate aerosol concentrations over North America and Europe. Since aerosols have very short atmospheric residence time, these controls quickly reduced their concentrations and hence their cooling influence on regional climates. Furthermore, the cool phase of the ocean cycles ended, and greenhouse gas concentrations continued to increase rapidly. When these factors, together with solar and volcanic forcings, are incorporated into climate model simulations, results confirm that they can explain the decadal pattern over temperature change during the past century - including the slight cooling in mid-century and the rapid rise since.

Reference: Hegerl et al., 2007.

C.11 Has natural climate variability over time scales of several decades contributed to the recent warming trend?

Response: Climate oscillations that occur every few decades or so as a result of natural variability internal to the climate system are an important factor in short-term regional climate trends. However, when averaged over multiple decades, and over continental to global scales, variations caused by these oscillations largely average out. Therefore, while oscillations are likely a factor in enhanced warming in regions like the Arctic, they cannot adequately explain the observed magnitude of multi-decadal warming at continental and larger scales.

Explanation: The Earth's atmosphere and oceans constantly undergo oscillations that can significantly affect regional climates on decadal and multi-decadal time scales. These oscillations represent variability internal to the climate system. The best known of these is the El Niño/La Niña cycle, officially called the El Niño Southern Oscillation (ENSO). The trend in regional surface temperatures over, for example, a 20-year period can be significantly influenced by such oscillations. The recent warming in the Arctic, for example, is likely attributable in part to a change in the Arctic Oscillation, which is also linked to the North Atlantic Oscillation. Likewise, recent warming in the North Pacific also appears to have been influenced by a change in phase of the Multi-Decadal Pacific Oscillation. However, when averaged over hemispheric and multi-decadal time scales, many of these natural variations average out. Thus, the observed global warming over the past 50 years appears to be primarily due to forcings external to the climate system, not internal climate system variability. The unusualness of recent warming when compared to variability over the past millennium further supports this conclusion.

Reference: Hegerl et al., 2007.

C.12 Despite the overall global warming during the 20th century, some argue that current average temperatures are still lower than during warm periods experienced in the past, such as the Medieval Warm Period. Doesn't this suggest that current increases are likely due to natural causes, and therefore of no real concern?

Response: As noted in C.4, although the observed warming of 0.74°C during the past 100 years seems modest, comparisons with reconstructed climates for the Northern Hemisphere for previous centuries show that the second half of the 20th century was likely the warmest 50-year period of at least the past 1300 years, at least in the Northern hemisphere. This period includes the interval often referred to as the Medieval Warm Period. Furthermore, climate model studies indicate that it is very difficult to replicate the climate trends of the past millennium without including the role of anthropogenic climate forcings. Experts therefore conclude that the trends of the past 50 years are indeed very significant and very unlikely to be due to natural causes.

Explanation: Researchers have indirectly collected information about past climates from various indicators such as tree rings, ice cores and ocean corals. These indicate that, for at least the Northern Hemisphere, the second half of the 20th century was likely the warmest 50 year period in at least the past 1300 years. Furthermore, the 1990s was the warmest single decade. By comparison, the Medieval Warm Period of about 1000 years ago appears to have been warm in regions surrounding the North Atlantic but not in other parts of the Northern Hemisphere. Average temperatures for the entire hemisphere during that period were cooler than that for the past century (see Figure C.4). Proxy data for the Southern Hemisphere are as yet too sparse to make similar conclusive comparisons in that region. However, paleoclimate scientists have also made some approximations of global temperatures further back in time. These suggest that temperatures experienced during the peak of the current interglacial period some 6,000 to 8,000 years ago were about 1°C warmer than today, and that temperature variations within this range have occurred on thousand-year time scales since then. This suggests that some of the recent warmings could be due to natural causes.

As shown in Figure C.4, climate model studies indicate that, during the first half of the 20th century, a significant part of the warming is, in fact, likely due to a combination of increased solar radiation, decreased volcanic dust in the atmosphere and rising greenhouse gas concentrations. However, during the past 50 years, solar intensity has not shown a significant long-term trend and more frequent major volcanic eruptions have, on average, increased the level of volcanic dust in the atmosphere with time. Thus, the combined effects of the natural causes for change, by themselves, would have caused cooling during that period. In contrast, the observed climate record shows a rapid warming in recent decades consistent with that expected due to human influences. Therefore, while temperature changes during the whole of the past century are due to a combination of natural and human factors, that for the past 50 years is very likely due primarily to human influences.

References: Hegerl et al., 2007; Jansen et al., 2007.