This page's content is no longer actively maintained, but the material has been kept on-line for historical purposes.
The page may contain broken links or outdated information, and parts may not function in current web browsers.

Notice: This page describes GISTEMP v2 updates.
As of June 2019, the current GISTEMP version is v4 and may be accessed at data.giss.nasa.gov/gistemp.

GISS Surface Temperature Analysis (v2)

Updates to Analysis (2003-2011)

Note 1: This webpage describes updates to the GISS analysis made between August 2003 and June 2011, i.e. before switching to GHCN v3. Updates made between June 2011 and May 2019, the "GHCN v3 period", are detailed here. Updates made beginning June 2019 are detailed here.

Note 2: In spring 2015, NOAA's NCDC (mentioned in various entries below) has been folded into NOAA's NCEI (National Center for Environmental Information), a new entity combining the three centers NCDC, NGDC, and NODC (National Climate, Geophysical, and Oceanographic Data Center).

Graphs and tables are updated around the 10th of every month using the current GHCN and SCAR files. The new files incorporate reports for the previous month and late reports and corrections for earlier months. NOAA updates the USHCN data at a slower, less regular frequency. We will switch to a later version, as soon as a new complete year is available.

Several minor updates to the analysis have been made since its last published description by Hansen et al. (2001). After a testing period they were incorporated at the time of the next routine update. The only change having a detectable influence on analyzed temperature was the 7 August 2007 change to correct a discontinuity in 2000 at many stations in the United States. This flaw affected temperatures in 2000 and later years by ~0.15°C averaged over the United States and ~0.003°C on global average. Contrary to reports in the media, this minor flaw did not alter the years of record temperature, as shown by comparison here of results with the data flaw ('old analysis') and with the correction ('new analysis').

August 2003: A longer version of Hohenpeissenberg station data was made available to GISS and added to the GHCN record. This had no noticeable impact on the global analyses.

March 2005: SCAR data were added to the analysis. This increased data coverage over Antarctica, as evident in the global maps of temperature anomalies.

April 2006: HadISST ocean temperatures are now used only for regions that are identified as ice-free in both the NOAA and HadISST records. This change affects a small number of gridboxes in which HadISST has sea ice while NOAA has open water. The prior approach damped temperature change at these gridboxes because of specification of a fixed temperature in sea ice regions. The new approach still yields a conservative estimate of surface air temperature change, as surface air temperature usually changes markedly when sea ice is replaced by open water or vice versa. Because of the small area of these gridboxes the effect on global temperature change was negligible.

Aug. 7, 2007: A discontinuity in station records in the U.S. was discovered and corrected (GHCN data for 2000 and later years were inadvertently appended to USHCN data for prior years without including the adjustments at these stations that had been defined by the NOAA National Climate Data Center). This had a small impact on the U.S. average temperature, about 0.15°C, for the period 2000-2006 (since NOAA stopped updating the GHCN data used in USHCN in March 2006), and a negligible effect on global temperature, as is shown here. (Added later: this correction stopped having any effect whatsoever as soon as the USHCN data were extended beyond March 2006)

This August 2007 change received international attention via discussions on various blogs and repetition by some other media, with no graphs provided to show the insignificance of the effect. Further discussions of the curious misinformation are provided by Dr. Hansen on his personal webpage (e.g., his post on "The Real Deal: Usufruct & the Gorilla").

Sep. 10, 2007: The year 2000 version of USHCN data was replaced by the current version (with data through 2005). In this newer version, NOAA removed or corrected a number of station records before year 2000. Since these changes included most of the records that failed our quality control checks, we no longer remove any USHCN records. The effect of station removal on analyzed global temperature is very small.

Mar. 1, 2008: Starting with our next update, USHCN data will be taken from NOAA's ftp site — the original source for that file — rather than from CDIAC's web site; this way we get the most recent publicly available version. Whereas CDIAC's copy currently ends in 12/2005, NOAA's file extends through 5/2007. Note: New updates usually also include changes to data from previous years. Whereas the GHCN and SCAR data are updated every month, updates to the USHCN data occur at irregular intervals.

The publicly available source codes were modified to automatically adjust if new years are added.

June 9, 2008: Effective June 9, 2008, our analysis moved from a 15-year-old machine (soon to be decommissioned) to a newer machine. This will affect some results, though insignificantly. Some sorting routines were modified to minimize such machine dependence in the future.

A typo was discovered and corrected in the program that dealt with a potential discontinuity in the Lihue station record and some errors were noticed on www.antarctica.ac.uk/met/READER/temperature.html (set of stations not included in Met READER) that were not present before August 2007. We replaced those outliers with the originally reported values.

Those two changes had about the same impact on the results as switching machines (in each case the 1880-2007 change was affected by 0.002°C). See graph and maps.

Aug. 11, 2008: Nick Barnes and staff at Ravenbrook Limited have generously offered to reprogram the GISTEMP analysis using Python only, to make it clearer to a general audience. In the process, they have discovered in the routine that converts USHCN data from hundredths of °F to tenths of °C an unintended dropping of the hundredths of °F before the conversion and rounding to the nearest tenth of °C. This did not significantly change any results since the final rounding dominated the unintended truncation. The corrected code has been used for the current update and is now part of the publicly available source.

Sep. 10, 2008: Comments were added to the homogenization program (PApars.f) and a line was changed to make the code compiler independent. (Thanks to Nick Barnes who noticed that some compilers may create an infinite loop due to roundoff errors.) These changes had no effect on any results.

Feb. 11, 2009: Two bugs in STEP0 programs were corrected before they had a chance to affect any results. Thanks to Mr. Peter O'Neill for discovering and reporting them to us.

SCAR corrected some errors in their data files; this had no effect on our work since we noticed and corrected most of them a few months ago (see 6/9/2008) and the others concerned stations whose records were too short to be included in our analysis. The analysis description was extended to describe the place where stations with short records are dropped.

May 2009: The sea ice mask adopted in April 2006 was slightly extended to include all ocean northward of 75N, since in that region in the winter months ice was present, particularly at the beginning of our data period, making water temperatures a bad proxy for air temperatures. This has no effect on our analysis, but it removes some odd discontinuities in some trend maps. The gridding tool was changed correspondingly. In addition, the last argument in that tool (mkTsMap.f) was changed to make it easier to use.

September 11, 2009: NOAA NCDC provided an updated file on 9 September of the GHCN data used in our analysis. The new file has increased data quality checks in the tropics. Beginning 11 September the GISS analysis uses the new NOAA data set. The change affects mainly part of South Africa.

November 13, 2009: NOAA is no longer updating the original version of the USHCN data; it ended in May 2007. The new version 2 currently extends to July 2009. Starting today, these newer data will be used in our analysis. Documentation and programs will be updated correspondingly.

December 3, 2009: Nick Barnes and staff at Ravenbrook Limited, while continuing reprogramming the whole GISS analysis, discovered a bug in a program used in STEP5; it was fixed and a rerun of the analysis showed that not a single number posted on this web site was affected by that correction. The public source code was modified correspondingly.

January 16, 2010: The urban adjustment, previously based on satellite-observed nightlight radiance in the contiguous United States and population in the rest of the world (Hansen et al., 2001), is now based on nightlight radiances everywhere, as described in an upcoming publication. The effect on the global temperature trend is small: Based on the 1900-2009 period, that change reduces it by about 0.005 °C per century.

March 12, 2010: NOAA updated USHCN version 2. The name of the file depends now on the date of the update, hence some programs had to be modified. The downloadable sources were updated. In order for them to work, the original USHCN v2 file has to be renamed to USHCNv2.avg.

April 12, 2010: Reports have been coming in from Nitchequon, a ghost town in Quebec, after a long gap from 1986-2006; however those data don't seem to be consistent with the older records. Hence they are disregarded until further notice.

It seems that USHCN is being replaced by an updated and renamed version every month.

May 14, 2010: The April 2010 report from Uigi, Mongolia, contained in GHCN seems incompatible with nearby stations and also with the Weather Underground report. It appears to be about 7°C too high and is therefore disregarded in our analysis until further notice.

June 4, 2010: The strange reports from Finland for March 2010 reappeared in the GHCN data set. In order to exclude them from our analysis, those data were added to the list of data that are disregarded. Starting with the May update, those data will be dropped unless they are replaced by corrected data.

June 24, 2010: Mid February, an experimental version of the programs that extract data from the basic sets to create the tables and line graphs was tested. The purpose was to create line graph data for additional zonal means. A few bugs in a preliminary version were found and a corrected version was saved under the old name. Unfortunately, the operational program still pointed to the preliminary version, so that the hemispheric values in our tables were slightly off. The graphs were not visibly affected, except for a few new figures on Dr. Sato's Columbia web page. Thanks to Dr. Peter O'Neill for bringing this oversight to our attention. The affected tables and graphs were replaced yesterday. The source codes made available to the public were not affected.

July 21, 2010: Robert Davies from the Utah Climate Center reported to us that the Utah station data were different from last month. It turned out, that a mishap occurred when we downloaded NOAA/NCDC's USHCN v2 data; the last third of the stations were missing and the unmodified GHCN reports were used in our analysis for these stations. The analysis is being redone based on a complete version of USHCN (last update 7/20/2010).

September 10, 2010: Three more stations were removed in our cleaning step: Zaragoza Airport (misidentified as Tenkodogo in the inventory file), Albacete and Sintra/Granja; there seems to be a potential mismatch between station ID and station name and location. The effect of this change on the estimates presented on this web site is well with the margin of error of those estimates.

Inventory file and list of discarded data have been modified on the site where the programs can be downloaded. We also made available the original night light satellite data that we obtained from Marc Hofmann.

October 14, 2010: A closer inspection of the correspondence table ushcn2.tbl, that pairs USHCN and GHCN IDs, showed that in USHCN v2 a station (Addison) was added that is not part of GHCN. This caused two misidentifications which were corrected by assigning this new station an unused GHCN ID. The inventory file v2.inv was changed accordingly; at the same time, a location of a station that differed from the location of the same station in the USHCN inventory file was corrected. The program dif.ushcn.ghcn.f had to be modified to handle the situation of a USHCN station that is not part of GHCN. The ccc version may or may not be able to handle this case properly without modifications. The impact of those corrections on our analysis was about ten times smaller than the margin of error. The downloadable source package was changed correspondingly.

April 13, 2011: Three GHCN v2 reports were found inconsistent with other sources and reports from neighboring stations, hence removed from our analysis:

  • Feb 1989 Base Orcadas (was corrected in GHCN v3)
  • Feb 2011 Gassim
  • Mar 2011 Ostrov Dikson

The reported values may still be seen in the "raw" station data, but become invisible if any of the other two options is chosen. The combined effect of these omissions is negligible.

May 12, 2011: The reports for all stations in Libya for April 2010 were changed last August to unrealistically low values. In this update, they were reset to the original reports as was also done by NOAA/NCDC in GHCN v3.

It was brought to our attention that the GISS analysis slightly underestimates the recent trend over Australia due to an inhomogeneity in v2.mean. The problem will automatically disappear as soon as GHCN v3 will become the basis for the GISS analysis. This should happen in the near future after some final testing.

June 17, 2011: The list of unacceptable reports was extended to eliminate outliers that were also either replaced or dropped in GHCN v3. This new list is now part of the downloadable source package. A corresponding much shorter list was created for GHCN v3. After some further discussions with the providers of GHCN v3, the switch to that data set should occur next month.

Return to GISTEMP homepage