GISS Surface Temperature Analysis

Fraction of "Warm" Stations

Using 1951-1980 as a base period, we compute the fraction of the stations (in a given latitude band and globally) that fall in the "warm" category. The "warm" category is defined at each station location from the climatology for the period 1951-1980; the 10 warmest years (33%) in that 30-year period are categorized as "warm", which is essentially the way that the National Weather Service defines the categories "warm", "normal" or "average", and "cool". We choose 1951-1980 as the base period because that was the base period at the time that "global warming" began to be a public issue. Also it is the time that "baby boomers" grew up, so this choice for base period allows those people to relate today's climate to that which they remember.

Of the 6013 NOAA/NCDC GHCN (Global Historical Climatology Network) stations in the GISS temperature analysis (map), we use for this warm station analysis only the stations with at least 300 months of data during the period 1951-1980. These stations are shown in the following map:

Global map of stations

The following graphs show the fraction of "warm" stations for six different latitude zones from 1950 to 2005:

6 line graphs of temperature anomalies in different latitude zones

The dashed line at 33% is the average number of "warm" seasons that occured during the period of climatology, 1951-1980. When global warming is large enough (in the average over several years) that the number of warm stations is 67% (the other dashed line in the figures), this will mean that the "climate dice" are loaded to the extent that four of the six sides of the dice are red (warm) as compared with the two sides of the dice that were red during the period of climatology. We have argued that such a degree of global warming may be sufficient for the "person-in-the-street" to notice that climate is indeed becoming warmer.

Note that the number of warm stations fluctuates a lot from year to year, especially in the winter season. For the latitude interval 30-60°N (shown above) the year-to-year fluctuations are magnified by the fact that a large portion of the stations are located in a single region (the United States). Because the longwave atmospheric weather patterns (Rossby waves) have a scale of several thousand kilometers, it is not unusual for the temperature of a region the size of the United States to be substantially warmer or colder during a single season than the zonal mean temperature. These fluctuations are a useful reminder that "global warming" as yet remains smaller than natural temperature fluctuations on regional and seasonal time scales.

In a 1988 paper (Hansen et al. 1988) and in congressional testimony that year, the GISS climate model was used to predict that the frequency of "warm" summers would increase to about 60% by the end of the century and it would continue to increase in the 21st century. The rate of increase was calculated for three scenarios (A, B, C) of greenhouse gas growth rates (representing fast, slow and no growth of greenhouse gases after 2000). The actual growth rate (show above) has fallen close to but slightly below scenario B (Hansen et al. 1998, Hansen 2002).