There is an extensive amount of evidence indicating that the Earth's climate has warmed during the past century (see Table 1.1). Foremost among this evidence are compilations of the variation in global mean sea surface temperature and in surface air temperature over land and sea. Supplementing these indicators of surface temperature change is a global network of balloon-based measurements of atmospheric temperature since 1958. As well, there are several indirect or proxy indications of temperature change, including satellite observations (since 1979) of microwave emissions from the atmosphere, and records of the width and density of tree rings. The combination of surface-, balloon-, and satellite-based indicators provides a more complete picture than could be obtained from any given indicator alone, while proxy records from tree rings and other indicators allow the temperature record at selected locations to be extended back for a thousand years. Apart from temperature, changes in the extent of alpine glaciers, sea ice, seasonal snow cover, and the length of the growing season have been documented that are consistent with the evidence that the climate is warming (e.g., IPCC, 1996a; Vaughn and Doake, 1996; Johannessen et al., 1999). Less certain, but also consistent, changes appear to have occurred in precipitation, cloudiness, and interannual temperature and rainfall variability.
As a starting point, paleoclimatic records of past climate changes should be a useful guide as to what one might expect if the climate is warming. During warmer climates in the past, high latitudes have warmed more than lower latitudes (Hoffert and Covey, 1992). Mountain glaciers should retreat. Sea level should rise. The current climate change is showing all of these features (Haeberli, 1990; Diaz and Graham, 1996).
Thermometer-based measurements of air temperature have been systematically recorded at a number of sites in Europe and North America as far back as 1760. However, the set of observing sites did not attain sufficient geographic coverage to permit a rough computation of the global average land temperature until the mid-nineteenth century. Land-based, marine air, and sea surface temperature datasets all require rather involved corrections to account for changing conditions and measurement techniques. Analyses of these records indicates a global mean warming from 1851 to 1995 of about 0.65±0.05°C (Jones et al., 1997a, b).
As shown in Figure 1.1, the increase in temperature has occurred in two distinct periods. The first occurred from roughly 1910-1945, while the second is since 1976. Recent warming has been about 0.2°C per decade. Very large changes have occurred in the last decade, with 1998 being the warmest year in the global temperature record. The highest ten years in global surface temperature have been since 1980, with eight of them occurring in the last eleven years.
In addition to limited sampling of temperature with altitude through balloon-borne instruments, satellite-based sensors, known as microwave sounding units (MSUs), are being used to examine global temperature changes in the middle troposphere (mainly the 850-300 HPa layer), and in the lower stratosphere (—50-100 Hpa). None of the channels sample at the ground. The MSU measurements have been controversial because some earlier versions of the satellite dataset had indicated a cooling in the lower troposphere in contrast to the warming from the ground-based instruments. However, several errors and problems (e.g., due to decay in the orbit of the satellite) with the MSU data have been found, and the latest analyses of MSU corrected for these problems show a warming (about 0.1 °C per decade), albeit somewhat smaller than that found at the ground (NRC, 2000). These analyses also suggest that the cooling effect of decreasing ozone in the lower stratosphere (as a result of chlorine and
Figure 1.1 Variations of the Earth's surface temperature for the last 1000 years. The top panel shows the combined annual land-surface and sea-surface temperature anomalies for 1861 to 1999, relative to the average of the 1961 to 1990 period. This figure is an update by P. D. Jones of the analysis previously done for IPCC (1996a). The bottom panel shows the Northern Hemispheric temperature reconstruction over the last 1000 years from proxy data in combination with the instrumental data record (Mann et al.,
bromine from human-related emissions of chlorofluorocarbons and other halocarbons) may have led to the difference in upper tropospheric and ground-level temperature trends.
The 1910-1945 warming primarily occurred in the Northern Atlantic. In contrast, the most recent warming has primarily occurred at middle and high latitudes of the Northern Hemisphere continents in winter and spring, while the northwest portion of the Northern Atlantic and the central North Pacific Oceans have shown year-around cooling. Significant regional cooling occurred in the Northern Hemisphere during the period from 1946 to 1975.
Proxy temperature indicators, such as tree ring width and density, the chemical composition and annual growth rate in corals, and characteristics of annual layers in ice cores, are being used at a number of locations to extend temperature records back as much as a thousand years (Jones et al., 1998; Mann et al., 1999; Bradley, 2000). As seen in Figure 1.1, the reconstruction indicates the decade of the 1990s has been warmer than any time during this millennium and that 1998 was the warmest year in the 1000-year record (Mann et al., 1999). Using a different approach, based on underground temperature measurements from boreholes, Huang et al. (2000) found temperature changes over the last 500 years that are very similar to the trend in Mann et al. (1999). The basic conclusion is the same, that the late-twentieth century warming is unprecedented in the last 500 to 1000 years.
Recent studies (for example, Boer et al., 2000; Delworth and Knutson, 2000; Wigley, 1999) with state-of-the-art numerical models of the climate system have been able to match the observed temperature record well, but only if they include the effects of greenhouse gases and aerosols. These studies indicate that natural variability of the climate system and solar variations are not sufficient to explain the increasing temperatures in the 1980s and 1990s. However, natural variability and variations in the solar flux are important in fully explaining the increase in temperature in the 1910-1945 period. Emissions from large volcanic eruptions resulting in sulfate aerosols and other aerosols in the lower stratosphere are also important in explaining some of the short-term variations in the climate record.
Levitus et al. (2000) have used more than five million measurements of the temperature of the world ocean at various depths and locations to show that the temperatures have increased in the middle depths by an average of about 0.06 oC between the 1950s and the mid-1990s. Watts and Morantine (1991) had previously suggested, based on data of mid-depth Atlantic Ocean temperature changes reported by Roemmich and Wunsch (1984), that much of the global warming temperature signature lay in the deep ocean.
Any changes in climate associated with increasing levels of carbon dioxide would also be expected to result in cooling stratospheric temperatures. The stratosphere indeed is cooling (Angell, 1999). While part of the cooling in the lower stratosphere can be explained by the observed decrease in stratospheric ozone, such changes in ozone can only explain part of the observed temperature change. The increase in CO2 is also necessary to explain the changes in lower stratospheric temperatures (Miller et al., 1992).
Was this article helpful?