Sorry about missing the update for March. I waited patiently for the HadCRUT4 data to update on the Met Office’s old webpage (recently deleted), but hadn’t realized they’d posted the March data on a new webpage because of a new revision level. I also made up my mind to present the monthly global surface temperature data in two phases. The first will include the GISS and NCDC data, because they’re both updated around mid-month. The second phase will include the HadCRUT data, which is posted toward the end of the month. Unfortunately, the NCDC data was late this month. They normally update on the 15th.
Note: If the topic of global surface temperature anomaly data is new to you, refer to the general discussion at the end of the post.
GISS LOTI DOWN IN APRIL 2013
The GISS Land-Ocean Temperature Index dropped about 0.08 deg C between March and April 2013—from +0.58 deg C to +0.5 deg C (base years of 1951-1980).
NCDC LAND+OCEAN TEMPERATURE INDEX
The NCDC’s Monthly Global Land+Ocean Temperature Anomaly Index also cooled from March to April 2013—from about 0.58 deg C to 0.52 deg C. No surprise there, since GISS and NCDC use the same sea surface temperature dataset (ERSST.v3b) and share the same source land surface air temperature dataset (GHCN). The NCDC uses the period of 1901 to 2000 as the base years for anomalies.
The GISS and NCDC global surface temperature anomalies (land air and sea surface) are compared in the following two graphs. For the comparisons, the base years for anomalies have been changed to the WMO-recommended period of 1981-2010. Changing the base years alters where the data intersects with zero, and it alters the monthly wiggles slightly—but changing the base years does not impact the relationships of the long-term trends. The first comparison starts in 1979, which is a commonly used start year for global temperature presentations. The second one starts in 2003 to provide a closer look at the last decade plus.
GISS and NCDC Comparison Starting in 1979
GISS and NCDC Comparison Starting in 2003
GENERAL OVERVIEW OF GLOBAL TEMPERATURE PRODUCTS
For those new to the presentation of global surface temperature anomalies, the following are a few notes about the datasets.
1. Global surface temperature datasets are a combination of sea surface temperature anomalies for about 70% of the surface of the globe and land surface air temperature anomalies for the other 30%.
2. Land surface air temperatures mimic and exaggerate the variations in sea surface temperatures.
3. Land surface air temperature measurements are not available for all parts of the globe. (Examples: There are few temperature sampling stations in the interiors of Africa and South America.) And the sampling gets poorer as we travel back in time. GISS and NCDC infill grids with missing data using statistical methods, but even so, there are large areas without even infilled data toward the early years of the datasets. The UKMO does not infill grids with missing land surface air temperature data.
4. For sea surface temperature data, the absence of source data is even more of a problem. The temperature sampling of the surfaces of the global oceans depended on ship locations during the decades before the recent stationary or free-floating buoys. In fact, temperature samples from ship inlets are still used today. So the locations of sea surface temperatures measurements depended on where ships traveled, and, unfortunately, ships typically traveled the same paths called shipping lanes. To complicate matters more, sea surface temperatures were sampled using different methods. In addition to ship inlets, buckets of different materials (canvas and wood) were tossed over the sides of ships, then hauled back on board and a thermometer was placed in the bucket of water—not too high tech.
5. GISS and NCDC both use the NCDC’s reconstructed sea surface temperature data called ERSST.v3b, which has been adjusted for the different sampling methods and then infilled. The UKMO uses the HADSST3 dataset in its HadCRUT4 global temperature product. The HADSST3 data have also been adjusted but grids with missing data have not been infilled, other than expanding the 2 x 2 degree longitude and latitude grids of the source data to 5 x 5 grids.
6. Both ERSST.v3b and HADSST3 sea surface temperature datasets are based on source data from ICOADS.
7. For the source of land surface air temperatures, GISS, NCDC and UKMO use the GHCN dataset.
8. Because GISS, NCDC and UKMO all use the same source datasets, they are not independent measures of the changes in global temperatures, as is often claimed.
9. GISS uses data from other sources to capture variations in the Arctic and Antarctic.
10. GISS also masks sea surface temperature data wherever sea ice has existed, and they extend land surface temperature data, with its higher variability, out over the surface of the oceans regardless of whether sea surface temperature data is present during a given month.
11. All of the global temperature suppliers use different reference periods, or base years, for temperature anomalies. GISS uses 1951 to 1980. NCDC uses 1901 to 2000. UKMO uses 1961 to 1990.
12. There are a number of reasons why the suppliers of global temperature products present their data as anomalies—as deviations from a base period. There are significant annual cycles in absolute land surface air temperatures (see example here) and in sea surface temperatures, though the variations in sea surface temperatures are not as great (example here).
13. There is no evidence in the sea surface temperature data to indicate that manmade greenhouse gases were responsible for their warming. Refer to the illustrated essay “The Manmade Global Warming Challenge” [42MB].