NASA Goddard Institute
GISS Homepage
About GISS
Datasets & Images
More Resources

Earth Science Portal


NASA Goddard Institute for Space Studies

Global Temperature Trends:
1998 Global Surface Temperature Smashes Record

Figure 1
Figure 1: Global mean temperature in degrees Celsius relative to the mean temperature for the period 1951-1980 based on measurements at meteorological stations (the mean for 1951-80 is about 14°C). The vertical lines at several dates indicate the estimated uncertainty in the annual-mean temperature due to the incomplete coverage of stations. (Click any figure to see the full-size image.)

Global surface temperatures in 1998 set a new record by a wide margin for the period of instrumental measurements, report researchers at the NASA Goddard Institute for Space Studies who analyze data collected from several thousand meteorological stations around the world. The global temperature (Fig. 1) exceeds that of the previous record year, 1995, by about 0.2°C (0.4°F). And unlike many recent years, the warmth is beginning to hit home; the United States experienced in 1998 its warmest year in the past several decades.

The temperature rise is practically identical for calendar 1998 and for the 1998 meteorological year. (A meteorological year runs from the beginning of winter to the end of autumn, and so the 1998 meteorological year started on Dec. 1, 1997, and ended on Nov. 30, 1998.) Numerical results for both calendar and meteorological years are provided in the GISS Surface Air Temperature Analyses datasets.

The NASA scientists, James Hansen, Reto Ruedy, Jay Glascoe and Makiko Sato, use data that is collected at the NOAA National Climate Data Center in Asheville, North Carolina and made available to all researchers (see Hansen et al. 1996, Peterson and Vose 1997). In their analysis, the NASA scientists use data from rural stations to correct for urban influences on reported temperatures. They also use satellite measurements of ocean temperature to obtain a second measure of global temperature change that is more uniformly spaced over the world (Hansen et al. 1996). The results reported today will change slightly as late-reporting
Figure 2
Figure 2: Surface temperature anomaly for the 1998 meteorological year (December 1997-November 1998) derived from measurements at several thousand meteorological stations and satellite measurements of the ocean surface. (Hansen et al. 1996)

station data are included, but late data will not alter the conclusion that both analyses, with and without ocean data, easily set new records in 1998.

The global warming is believed by many researchers, including the NASA scientists, to be at least partly a consequence of increasing human-made gases in the atmosphere, especially carbon dioxide, which arises from the burning of coal, oil and gas. The 1998 warmth was associated partly with a strong El Niño that warmed the air over the eastern tropical Pacific Ocean in the first half of the year and in turn affected weather around the world.

The largest temperature anomalies in 1998 were in North America (Fig. 2), in a pattern that commonly occurs in El Niño years. But almost the entire world was warmer than normal in 1998. The El Niño, by itself, cannot account for either the observed long-term global warming trend or the extreme warmth of 1998. Because the Pacific Ocean temperature has returned to a more normal level, it is anticipated that the global temperature in 1999 will be less warm than 1998 but will remain well above the long-term mean for the period of climatology, 1951-1980.

The rapid global warming since the mid 1970s exceeds that of any previous period of equal length in the time of instrumental data (Fig. 1). The increase of the 5-year mean temperature since 1975 is about 1/2°C or almost 1°F. This warming occurs during the time when greenhouse gases increased most rapidly, and thus, according to Dr. Hansen "the rapid warming of the past 25 years undercuts the argument of 'greenhouse
Figure 3
Figure 3: Temperature anomaly (deviation from the 1951-1980 average) in degrees Fahrenheit for the contiguous United States.

skeptics', who have maintained that most of the global warming occurred early this century while greenhouse gases were increasing more slowly -- in fact, the fastest warming is occurring just when it is expected".

Temperatures in the United States were also warm in the meteorological year 1998 (Fig. 3). In this preliminary analysis the 1998 United States mean temperature is the warmest for at least 40 years. Exact comparison will depend on more complete station reporting, but it is clear that 1998 did not match the record warmth of 1934, which occurred during the Dust Bowl era. Fig. 3 also illustrates that the chaotic year-to-year fluctuations of temperature are much larger for the contiguous United States than for the global mean temperature. This is related to the fact that the United States covers only about 2 percent of the world area, and thus its temperature is influenced more by regional fluctuations.

Is the global warming of 1/2°C or 1°F, which is much smaller than day to day temperature changes, large enough to be noticeable to the lay person? Some insight into that question is provided by Fig. 4, which shows the temperature anomalies in the United States for each of the past four seasons. Even though 1998 was remarkably warm, in every season there was some part of the country that was cooler than "normal", that is, cooler than the 1951-1980 average. This is a result of the average warming being smaller than natural fluctuations of regional temperature.

Figure 4
Figure 4: Seasonal temperature anomalies for the United States. The temperature anomalies in this figure are smoothed over a 500 km distance, while for the global map (Fig. 2) the temperature is smoothed over a 1200 km distance.

Yet global warming may be approaching a level where the astute observer can notice its effects. It is apparent in Fig. 4 that the area with warm temperature anomalies exceeds that with cold anomalies. The issue of whether the more extensive warming is noticeable to people can be investigated using a "Common Sense Climate Index" defined earlier this year by the NASA researchers.

One of the elements in the "common sense" index is the frequency of unusually warm seasons, where the local temperature required to qualify as a warm season is chosen such that one-third of the seasons in the period 1951-80 were in this category. Climate models have predicted that the global warming would cause the frequency of such warm seasons to increase from 33% to 50-70% by the 1990s. In other words, although global warming is too small to prevent some seasons from being cooler than normal, the perceptive observer may notice that the "climate dice" are becoming loaded (Fig. 5).

Figure 5
Figure 5: The dice on the left represent the period of climatology, 1951-1980: the cut-off for "warm" seasons (red) is chosen such that the warmest 1/3 of the seasons qualify as "warm". With the same cut-off, a "warm" season now occurs almost 2/3 of the time at a typical location.

Fig. 6 (updated from Hansen 1997) shows the average frequency of warm seasons for all meteorological stations located at middle latitudes in the Northern Hemisphere and for all stations globally. The result confirms that there is an increasing tendency for warm seasons. However, perception of change by people is a sociological matter, and it is not proven that this amount of change, i.e., this degree of loading of the dice, is sufficient yet to be noticeable to most people (Hansen et al. 1998).

The temperatures analyzed by the NASA researchers are measured at a height of about two meters over land areas, while the temperatures refer to the sea surface in ocean regions sampled by satellite measurements (Hansen et al. 1996). Temperature changes are not necessarily the same at higher levels in the atmosphere sampled by weather balloons and by satellite microwave observations. However, the warming in 1998 is so large that the NASA researchers believe that record warm temperatures should be found by all observing systems measuring from the surface through the troposphere (lowest 5-10 miles of atmosphere). According to Dr. Hansen "There should no longer be an issue about whether global warming is occurring, but what is the rate of warming, what is its practical significance, and what should be done about it."

Figure 6
Figure 6: Fraction of meteorological stations having unusually warm seasons for (left) stations at latitudes 30-60°N, and (right) all stations averaged into six latitude zones with each zone weighted by area.

Further Information

Other GISS pages related to this research are:


Please address all inquiries regarding GISS surface temperature trends analysis to James E. Hansen. Other GISS scientists involved in this research are Reto A. Ruedy, Makiko Sato, and Jay Glascoe.


Research ] [ Modeling ] [ Observations ] [ Impacts ] [ Planets ] [ Paleoclimate ] [ Radiation ] [ Chemistry ] [ Astrophysics ]
Written: 1998-12-16, jeh
Last updated: 1999:04:23, rbs