A San Diego TV station’s mid-January one-hour broadcast reporting that two key federal climate research centers deliberately manipulated temperature data appears to have been based on a fundamental misunderstanding of the nature of the key climatology network used in calculating global temperatures.

Independent TV news station KUSI in San Diego aired a story challenging current scientific understanding of climate science and offering “breaking news” of government wrongdoing based on work of Joseph D’Aleo, a meteorologist, and E.M. Smith, a computer programmer.

The two maintained that the National Oceanic and Atmospheric Administration, NOAA, “is seriously complicit in data manipulation and fraud” by “creating a strong bias toward warmer temperatures through a system that dramatically trimmed the number and cherry-picked the locations of weather observation stations they use to produce the data set on which temperature record reports are based.”

The program’s host, KUSI meteorologist John Coleman, accused NOAA and NASA climate research laboratories of “lying” to the American public, a charge NOAA and NASA spokespersons have both rejected. These are extraordinary claims which, as Carl Sagan was fond of saying, should require “extraordinary evidence.”

The broadcast accusations appear to have resulted from a misunderstanding or misrepresentation of the nature of the Global Historical Climatology Network (GHCN) and methods used in calculating global temperatures.

D’Aleo and Smith charged that NOAA “systematically eliminated 75 percent of the world’s stations with a clear bias towards removing higher latitude, high altitude and rural locations, all of which had a tendency to be cooler.” They also said NOAA used a “slight of hand” to make 2005 appear to be the warmest year on record; that “the National Data Climate Center [NCDC] deleted actual temperatures at thousands of locations throughout the world as it evolved to a system of global grid boxes”; and that NOAA and NASA arbitrarily adjusted individual station temperature data in order to exaggerate warming.

When glancing at the chart showing the number of temperature stations used over time, it does appear rather odd to see the number of stations used in the GHCN network drop dramatically between the 1970s and present. D’Aleo and Smith point to purposeful elimination of those stations.

However, as Thomas Peterson and Russell Vose, the researchers who assembled much of GHCN, have explained:

The reasons why the number of stations in GHCN drop off in recent years are because some of GHCN’s source datasets are retroactive data compilations (e.g., World Weather Records) and other data sources were created or exchanged years ago. Only three data sources are available in near-real time.

It’s common to think of temperature stations as modern Internet-linked operations that instantly report temperature readings to readily accessible databases, but that is not particularly accurate for stations outside of the United States and Western Europe. For many of the world’s stations, observations are still taken and recorded by hand, and assembling and digitizing records from thousands of stations worldwide is burdensome.

During that spike in station counts in the 1970s, those stations were not actively reporting to some central repository. Rather, those records were collected years and decades later through painstaking work by researchers. It is quite likely that, a decade or two from now, the number of stations available for the 1990s and 2000s will exceed the 6,000-station peak reached in the 1970s.

View larger image
Figure taken from Peterson and Vose (1997), showing the change in temperature stations over time for daily mean temperatures (solid line) and min/max temperatures (dotted line).

There actually is a fairly easy way to test if the absence of more recent data from a number of stations has a significant effect on temperature records. If stations were purposefully dropped in favor of those with greater warming trends, one would expect to see cooler temperatures in the stations that do not have temperature records available in the last few decades than in those stations with a continuous record up to the present.

The chart below shows this analysis for all stations with continuous records between 1960 and 1970. Of the 1,419 temperature stations containing data for this period, available at the National Center for Atmospheric Research, 1,017 continue up to at least the year 2000, and 402 stop providing data at some point between 1970 and 2000.

There is no significant difference between the temperature from discontinuous and continuous stations, suggesting that there was no purposeful or selective “dropping” of stations to bias the data. If anything, discontinuous stations have a slightly higher trend over the century than continuous stations. This result strongly suggests that the discontinuity in station data results from having inadequate resources to gather those records, rather than from some pernicious plot to exaggerate warming trends.

View larger image
Based on an analysis raw temperatures from all station records in the NCAR database with duration of at least 10 years (the minimum required for inclusion in GHCN). Continuous stations are defined as those that at a minimum cover the period from 1960 to 2000. Discontinuous stations are those that cover the period from 1960 to 1970 but stop contributing data to GHCN between 1970 and 2000. Discontinuous station data is not plotted post-1997 due to the availability of fewer then 25 stations remaining in the group. Raw data and source code for this analysis can be found here. Note that changes in the spatial distribution of stations in both groups will impact trends in ways unrelated to real global temperatures, particularly in the discontinuous group when the number of stations available becomes small.

The chart below shows a map of all 1,200 temperature stations that provide updated temperature data on a monthly basis. While certain places have much better spatial coverage than others, there is a good distribution of stations across all major landmasses, with the possible exception of parts of Africa and the Polar Regions, particularly Antarctica (though Antarctic temperature records are supplemented by a number of temporary stations located in the interior of the continent).

View larger image
Figure taken from Peterson and Vose (1997) and shows all stations in GHCN v2 with regularly updating temperature records.

In addition, the accuracy of the surface temperature record can be independently validated against satellite records. Over the period from 1979 to present where satellite lower-tropospheric temperature data is available, satellite and surface temperatures track quite well as shown in the chart below. One analysis of the satellite data by the Remote Sensing Systems (RSS) group has a slope (0.15 C per decade) virtually identical to that of the GISS and NCDC (0.16 C per decade) temperature records, while another from University of Alabama, Huntsville has a slightly lower slope (0.13 C per decade).

If stations had intentionally been dropped to maximize the warming trend, one would expect to see more divergence between surface and satellite records over time as the number of stations used by GHCN decreases. However, a close examination of the residuals when satellite records are subtracted from surface station records shows no significant divergence over time compared to either UAH or RSS.

View larger image
Figure One: Based on monthly data from 1979 through December 2009 from GISS, UAH, RSS, and NCDC.

In addition to arguing that NCDC manipulated the stations used in the temperature record, D’Aleo and Smith said “the National Data Climate Center deleted actual temperatures at thousands of locations throughout the world” when they switched to a method that averages data from individual temperature stations into discrete grid boxes that cover the globe to better account for the spatial location of individual station measurements. But it is hard to reconcile that view with the vast amount of raw station data available in various places from NCDC and others.

D’Aleo and Smith also said that NCDC and NASA adjusted the raw temperatures from stations in a number of ways in the process of producing their temperature records. They identified a number of stations where these adjustments appear to turn a cooling trend into a warming trend.

However, many of those adjustments made to raw temperature data are completely justified. Temperature stations have undergone many changes over their long lifetimes, including moves to different locations, readings at differing times of day, changes in the type of screens used to house thermometers, changes in the environment surrounding the station, and other factors. NCDC and NASA correct for these changes in slightly different ways, but one of the primary methods involves comparing individual temperature station records to those of the nearest stations to identify significant discontinuities (e.g., large persistent upward or downward step changes in the data) and correct for them.

Adjustments will be negligible for most stations, but a few stations will experience large positive or negative adjustments that can have a significant effect on the long-term temperature trends. To determine the net effect of these adjustments, one would have to examine the adjustments across all stations, rather than highlighting a few outliers as D’Aleo and Smith did.

The image below, from an analysis by an Italian molecular biologist, shows a histogram of the effect on the slope over the record of each temperature station for all adjustments made in GHCN. As expected, most adjustments are quite small, but there are positive and negative outliers on both sides. This allows parties interested in criticizing adjustments to pick individual stations that show either a large increase in warming or cooling trends due to adjustments.

View larger image
Analysis of the trend effects of 6737 adjustments in GHCN over the full record of each station. The median adjustment is 0 and the mean adjustment is 0.017 degrees C per decade. Note that these trend effects may be larger or smaller if a shorter timeframe is examined. From Giorgio Gilestro. Note that Realclimate has a similar analysis.

After examining the evidence, there seems little indication that either the discontinuities in recent records from many GHCN stations or the adjustments made to the raw data have any substantive effects on global temperature trends. The accusations by D’Aleo and Smith aired as part of the KUSI “The Other Side” broadcast seem to be mostly unfounded, and certainly do not justify the seriousness of their allegations.

Creating global temperature records is no simple task, and the process might not always be pretty. But there is no evidence of major methodological problems that would compromise the validity of the records, and certainly no evidence of deliberate manipulation.

Avatar photo

Zeke Hausfather

Zeke is an energy systems analyst and environmental economist with a strong interest in conservation and efficiency. He was previously the chief scientist at C3, an energy management and efficiency company,...