How do we explain what we see?

Global CO2 and temperature records show that CO2 in the atmosphere has increased dramatically with coinciding temperature increases, sea level rise, and other warming signals. Intuitively, it makes sense that these are related; as the greenhouse gas blanket gets thicker, the surface warms up. However, as with the cycles between ice ages and warm periods, to estimate the relative effects of humans and other factors on recent climate change we have to consider the full range of climate drivers.


The major human drivers of climate change include emissions of greenhouse gases, changes in land use affecting the local land surface (warm or cool), and the effect of sulfate-producing particulate pollution (cooling) and soot (warming) from burning fossil fuels.The most significant of these is SO2, from burning of fossil fuels, which forms sulfate in the atmosphere. The white haze of sulfate makes the atmosphere more reflective and exerts a cooling effect.

Over time periods of decades to a few centuries, the natural drivers meriting consideration include changes in the brightness of the Sun (e.g., due to sunspots) and volcanic eruptions, which loft small particles into the sky.


Current research shows that natural drivers cannot adequately explain the timing, amount, or pattern of warming that has been observed over the second half of the 20th century. This figure illustrates the expected effects of natural drivers, and natural and human drivers combined, on global temperature in the 20th century.

The contribution to 20th century climate change from natural and human factors. The green band shows how model simulations estimate temperature when only natural factors work to alter the climate. The blue band shows how model simulations estimate temperature when both natural and human factors are in play. The width of each band shows the range in climate due to different model characteristics and internal variability. The black line shows one estimate of observed surface temperature.

As in the chart above, note that only when human influences are included do the temperature increases calculated by climate models reasonably represent the amount of warming in the observed record, as well as its timing. More sophisticated studies of the relative weighting of human and natural climate drivers take into account regional variations in temperature as well as the vertical distribution of temperature change to separate human and natural signals.


Individual studies that calculate how much of the recent change is due to human activity find that nearly all of the warming observed over the late 20th century was the result of human influence. Taking a cautious estimate of error in such calculations, scientists writing the summary for the last IPCC report concluded that at least half of the temperature increase since the mid-20th century (total 0.6° C) was the result of human influence. Their best estimate was that all of the recent temperature increase was human-caused. Without the cooling effect of particulate aerosol pollution, we probably would have experienced greater warming from increases in the CO2 and other greenhouse gas concentrations.


Disentangling the combined effects of natural variability, external factors, and human influence on the temperature changes before 1950 is more difficult, because the observations themselves are more uncertain, data coverage is less complete and conclusions are necessarily less confident. However, there is little doubt that, since the 18th century, the largest changes to agents that warm and cool the climate have come from human influences, and they favor warming.


Other climate changes that are consistent with greenhouse gas warming include sea level rise, warming oceans, retreating glaciers and sea ice, and cooling of the upper atmosphere (stratosphere). All of which have been observed to some degree already. This consilience of evidence is cited by the IPCC and others to confidently assert not just the reality of global warming, but its human causes. However, the conclusion about human influence for each type of phenomena weakens at regional scales and for specific phenomena (such as drought and other extreme weather events), as data and physical understanding get noisier, subject to more variability, and the amount of data coverage declines.