Global surface air temperatures

List of contents

 

Estimates of recent global temperature change

All temperature diagrams shown below have 1979 as starting year. This roughly marks the beginning of the recent period of global warming, after termination of the previous period of global cooling (from 1940). In addition, the year 1979 also represents the starting date for most satellite-based global temperature estimates. For the three surface air temperature estimates shown below (HadCRUT3, NCDC and GISS) the reference period (the normal period) differs. HadCRUT3 refers to the official WMO period 1961-1990, while NCDC and GISS instead uses the periods 1901-2000 and 1951-1980, respectively, which results in higher positive temperature anomalies for these two estimates. For all three surface air temperature estimates shown, small differences in anomalies for past years may occur from time to time due to the addition of recently collected station data or the incorporation of more recent versions of the base data sets.

In the text found below each diagram you will, among other links, find a possibility for downloading the data yourself.

Global monthly average lower troposphere temperature since 1979 according to University of Alabama at Huntsville, USA. This graph uses data obtained by the National Oceanographic and Atmospheric Administration (NOAA) TIROS-N satellite, interpreted by Dr. Roy Spencer and Dr. John Christy, both at Global Hydrology and Climate Center, University of Alabama at Huntsville, USA. This temperature record show good agreement with the independent radiosonde temperature record. The thick line is the simple running 37 month average, nearly corresponding to a running 3 yr average. Click here to read about data smoothing. The cooling and warming periods directly influenced by the 1991 Mt. Pinatubo volcanic eruption and the 1998 El Niño, respectively, are clearly visible. Click here to download the entire series of UAH MSU global monthly lower troposphere temperatures since December 1978. IPCC = foundation of the International Panel on Climate Change (November 1988). Last month shown: May 2008 (update 5 June 2008).  

Global monthly average lower troposphere temperature since 1979 according to Remote Sensing Systems (RSS). This graph uses data obtained by the National Oceanographic and Atmospheric Administration (NOAA) TIROS-N satellite, and interpreted by Dr. Carl Mears (RSS). The thick line is the simple running 37 month average, nearly corresponding to a running 3 yr average. Click here to read about data smoothing. The cooling and warming periods directly influenced by the 1991 Mt. Pinatubo volcanic eruption and the 1998 El Niño, respectively, are clearly visible. Click here to download the entire series of RSS MSU global monthly lower troposphere temperatures since January 1979. Click here for a description of RSS MSU data products. IPCC = foundation of the International Panel on Climate Change (November 1988). Last month shown: May 2008 (update 8 June 2008).   

Global monthly average surface air temperature since 1979 according to Hadley CRUT, a cooperative effort between the Hadley Centre for Climate Prediction and Research and the University of East Anglia's Climatic Research Unit (CRU), UK. The thick line is the simple running 37 month average, nearly corresponding to a running 3 yr average. Click here to read about data smoothing. The cooling and warming periods directly influenced by the 1991 Mt. Pinatubo volcanic eruption and the 1998 El Niño, respectively, are visible, but not as clear as in the two satellite records. Click here to download the entire series of estimated HadCRUT3 global monthly surface air temperatures since 1850. Click here to read a description of the data file format. An introduction to the dataset has been published by Brohan et al. (2005). Lower down the present page you will find a graph showing the entire series since 1850. IPCC = foundation of the International Panel on Climate Change (November 1988). Base period: 1961-1990. Last month shown: May 2008 (update 19 June 2008).

Global monthly average surface air temperature since 1979 according to the National Climatic Data Center (NCDC), USA. This time series is calculated using land surface data from the Global Historical Climatology Network (Version 2) and sea surface temperature anomalies from the United Kingdom MOHSST data set and the NCEP Optimum Interpolated SSTs (Version2). The thick line is the simple running 37 month average, nearly corresponding to a running 3 yr average. Click here to read about data smoothing. The cooling and warming periods directly influenced by the 1991 Mt. Pinatubo volcanic eruption and the 1998 El Niño, respectively, are visible, but not as clear as in the two satellite records and by the HadCRUT3 series. Click here to download the entire series of the NCDC global monthly surface air temperatures since 1850. IPCC = foundation of the International Panel on Climate Change (November 1988). Base period: 1901-2000. Last month shown: May 2008 (update 17 June 2008).

Global monthly average surface air temperature since 1979 according to the Goddard Institute for Space Studies (GISS), at Columbia University, New York City, USA. GISS is a laboratory of the Earth-Sun Exploration Division of NASA's Goddard Space Flight Center and a unit of the Columbia University Earth Institute. The thick line is the simple running 37 month average, nearly corresponding to a running 3 yr average. Click here to read about data smoothing. The cooling and warming periods directly influenced by the 1991 Mt. Pinatubo volcanic eruption and the 1998 El Niño, respectively, are visible, but not as clear as in all other records shown above, especially the two satellite records. The GISS and the NCDC surface air temperature estimate visually deviate somewhat from the three other estimates shown here. Discussions of reasons why the GISS temperature estimate differs from other estimates can be read by clicking here, here and here. Click here to download the entire series of the GISS global monthly surface air temperatures since 1880. IPCC = foundation of the International Panel on Climate Change (November 1988). Base period: 1951-1980. Last month shown: May 2008 (update 11 June 2008).  

 

The difference between the individual diagrams shown above demonstrate the difficulties associated with calculating a correct average global temperature. Essex et al. (2006) have an interesting discussion of the whole concept of calculating an average global temperature. In addition, global surface air temperatures should only be considered a poor indicator of global climate heat changes, as air has relatively little mass associated with it. Ocean heat changes remain the dominant factor for global heat changes. Global air temperatures, however, continues to attract widespread interest, and many scientist assume that the air temperature at least may be considered a useful proxy for the present state of the global climate system.

In order to enable a visual comparison of the five different global temperature estimates shown above, the diagram below show all series superimposed. Before inclusion in this diagram, all data series were normalised by setting their starting value in January 1979 = 0.

Superimposed plot of all five global monthly temperature estimates shown above, after setting January 1979 = 0. The numbers given in the upper left of the diagram represent the average temperature anomaly recorded since January 1979. The two satellite-based temperature estimates (RSS MSU and UAH MSU) deviate somewhat from each other, with RSS MSU generally being the warmer. The satellite-based range for the average monthly temperature anomaly since January 1979 is 0.22 - 0.35 oC. The three surface-based temperature estimates (HadCRUT3, GISS and NCDC) also show differences, but the average monthly temperature anomaly range since January 1979 is small, 0.26 - 0.27 oC. The numbers shown in the lower right represent the anomaly since January 1979 for the last month with data for all five series (May 2008). Values are rounded off to the nearest two decimals, even though some of the the original data series come with more than two decimals. Last figure update: 19 June 2008.    

 

All five global temperature estimates presently show stagnation, at least since 2002. There has been no increase in global air temperature since 1998, which was affected by the oceanographic El Niño event. This does not exclude the possibility that global temperatures soon will begin to increase again. On the other hand, it also remain a possibility that Earth just now is passing a temperature peak, and that global temperatures will begin to decrease within the coming 5-10 years. Only time will show which of these possibilities is the correct. Click here to read a few additional reflections on the recent period of global temperature stagnation.

The HadCRUT3 global surface air temperature estimate takes an intermediate position among the five global temperature estimates, but is it impossible to say which of the series shown is the correct one, if any. The intermediate position taken by the HadCRUT3 series, however, is the practical background for displaying this specific temperature estimate below and elsewhere on this website. Among the satellite-based global temperature estimates the UAH MSU shows the highest degree of agreement with the independent radiosonde record. At the moment, all surface based temperature estimates are higher than the sattellite bases estimates.   

Usually modern surface air temperatures are compared to the so-called normal temperature, representing the likewise so-called normal climate. This normal temperature is calculated as the average for values recorded during a 30-year period. The period 1961-1990 is the official World Meteorological Organisation (WMO) normal period, and is therefore often referred to. Another 30-year period used as reference for comparisons is 1951-1980. This is partly because the total number of meteorological stations during this period reached a maximum, and since has undergone a marked reduction in number. 

Unfortunately, both these periods are dominated by the cold period 1945-1980, and almost any comparison with such a low average value will therefore appear as high or warm. This makes it difficult to decide if surface air temperatures at present are increasing or decreasing? The only thing that will be clear is that modern temperatures are higher than back in this cold period. Click here to see a diagram showing the entire global temperature series (HadCRUT3) since 1850 with the 1945-1980 cold period and the WMO normal period indicated.

Click here to jump back to list of contents.

 

Records of atmospheric carbon dioxide (CO2) and Methane (CH4)

Most current climate models assume greenhouse gasses like especially CO2 but also CH4 to influence significantly upon global temperature. It might therefore be useful to compare the temperature records shown above with current measurements of these two gasses. Any comparison, however, should not be made on a monthly or annual basis, but for a longer time period, as other effects (oceanographic, etc.) may well override the potential influence of CO2 and CH4 on short time scales such as just a few years. It is of cause equally inappropriate to present new meteorological record values, whether daily, monthly or annual, as support for the hypothesis ascribing high importance of atmospheric CO2 and CH4 for global temperatures. Any such short-period meteorological record value may well be the result of other phenomena.

What exactly defines the critical length of a relevant time period to consider for evaluating the alleged importance of CO2 and CH4 remains elusive, and is still a topic for discussions. The critical length must, however, be inversely proportional to the temperature sensitivity of CO2 and CH4 assumed by the climate models considered. In 1988, when IPCC was established, about 10 years of global temperature increase was apparently seen as being sufficient by many climate scientists.

Below you will find links to CO2 and CH4 data and diagrams:

Atmospheric carbon dioxide:

Atmospheric methane:

Click here to jump back to list of contents.

 

 

Monthly surface air temperature anomalies versus average 1998-2006 in areas between 72oN and 60oN

The diagram table below contains clickable monthly spatial temperature diagrams since 2005, to illustrate the changeable geographical pattern of surface air temperature variations, integrated by graphs like that above. These diagrams are geographical asymmetrical to cover most of the planets land areas, ranging from 72oN to 60oS.  

YEAR JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC ANNUAL
2005
2006
2007
2008                

Spatial distribution of monthly surface air temperature deviation between 72oN and 60oS in relation to the average for the period 1998-2006. Warm colours indicates areas with higher temperature than the 1998-2006 average, while blue colours indicate lower than average temperatures. In the individual diagrams the month is indicated by a number: 1 = January, 2 = February, etc. Click on the individual small diagrams to open full-size diagrams. Similar spatial temperature diagrams showing the polar regions can be seen by clicking here. Data source: NASA Goddard Institute for Space Studies (GISS).

 

It is important to note that the map projection used above is of the type Mercator. This is a useful cylindrical map projection that preserves angles at all locations, but scale varies from place to place, distorting the size of land areas. In particular, areas closer to the poles are more affected, making land areas of similar size looking increasingly oversized towards the poles. To exemplify this effect, the areas of Mexico (1,972,550 km2) and Greenland (2,166,086 km2) are comparable in size. Greenland, however, in the map looks very much bigger than Mexico, even though only the southern half of Greenland is shown. The visual effect of this is to overstate the importance of temperature variations near the poles, compared to equatorial regions. To avoid the worst effects of this cartographic distortion of areas, the two Polar Regions are therefore shown in separate, polar projections. Click here to go to the polar spatial temperature diagrams.

To monitor the present global temperature trend, up or down, it is not efficient to compare with some past period like, e.g., 1961-1990, even though this is exactly what is frequently done. This will not inform about the current temperature trend. It clearly makes more sense to compare with a more recent period, e.g., the average of the last 10 years. This is why the diagrams in the table above all use 1998-2006 as reference period.

All the diagrams in the table above were prepared using gridded data downloaded from the public domain NASA Goddard Institute for Space Studies (GISS) web page. For surface interpolation of the gridded data a kriging algorithm was used, plotting all data in a polar projection map. The kriging procedure attempts to express trends and is widely considered one of the more flexible interpolation methods, producing a smooth map with few ‘bull eyes’. It is usually recommended for gridding almost any type of data set, especially data sets with a heterogeneous point distribution, such as characterising the present data set. It should be noted that the observation network within the two regions considered is not of equal density or quality all over the geographical regions covered by the diagrams.

Click here to jump back to list of contents.

 

 

Global and hemispheric annual surface air temperatures since 1850

As supplement to the graph showing the estimated monthly global surface air temperature since 1985 at the top of the present page, three graphs below show the estimated annual surface air temperature since 1850, global and hemispherical, respectively. The graphs are all prepared using HadCRUT3 temperature data from the Climatic Research Unit (CRU). Also shown in the diagrams is the WMO normal reference period. Click here to se an explanation of the background for this reference period.

Global annual surface air temperature anomaly since 1850 as estimated by Hadley CRUT. HadCRUT3 temperature data from the Climatic Research Unit (CRU) has been used to prepare the diagram. A description of these data can be seen by clicking here. The HadCRUT3 data can be downloaded from the same website. Latest update: 25 January 2008.

 

 

Northern hemisphere annual surface air temperature anomaly since 1850 as estimated by Hadley CRUT. HadCRUT3 temperature data from the Climatic Research Unit (CRU) has been used to prepare the diagram. A description of these data can be seen by clicking here. The HadCRUT3 data can be downloaded from the same website. Latest update: 25 January 2008.

 

 

Southern hemisphere annual surface air temperature anomaly since 1850 as estimated by Hadley CRUT. HadCRUT3 temperature data from the Climatic Research Unit (CRU) has been used to prepare the diagram. A description of these data can be seen by clicking here. The HadCRUT3 data can be downloaded from the same website. Latest update: 25 January 2008.

 

Click here to jump back to list of contents.

Click here to see a spatial analysis of monthly variations of surface air temperature in areas north of 50oN and south of 50oS since 2005.

Click here to see decadal variations of precipitation and surface air temperature in the northern hemisphere polar region during the 20th century.

Click here for an update on present global, Arctic or Antarctic meteorological conditions.