I’ve moved to WordPress. This post can now be found at Why Does GISS Use HadISST Data from 1880 to 1981?##################
This post was originally titled “Why Does GISS Use HADSST2 Data From 1880 to 1981”. The title and the content of the post were in error. GISS does not use HADSST2 data. As of this correction, GISS uses HADISST from 1880 to November 1981 and the NCDC’s OI.v2 SST data from December 1981 to present.
They clarify the timing of the switch from Hadley Centre to NCDC SST data on their map making page…
…where they write,
+ Hadl/Reyn_v2:SST 1880-present
1880-11/1981: Hadley HadISST1, ship and buoy data (Rayner 2000),
12/1981-present: oisst v2, satellite data (Reynolds-Rayner-Smith 2001)
Prior to the switch to HADISST, GISS had used an NCDC SST dataset for the pre-satellite era. In a 2005 version of their GISS Surface Temperature Analysis webpage available via Wayback…
..GISS had noted, “Our analysis includes results for a global temperature index as described by Hansen et al. (1996). The temperature index is formed by combining the meteorological station measurements over land with sea surface temperatures obtained primarily from satellite measurements (Reynolds and Smith, 1994; Smith et al. 1996). Any uses of the temperature index data, i.e., the results including sea surface temperatures, should credit Reynolds and Smith (1994) and Smith et al. (1996).”
So at one time, GISS had used NCDC SST data exclusively. Then they switched to HADISST for the longer term data, but retained the NCDC’s satellite-based OI SST data in recent decades. This is confirmed on the more recent GISTEMP Surface Temperature Analysis webpage…
…where GISS describes the GISTEMP data as, “A global temperature index, as described by Hansen et al. (1996), is obtained by combining the meteorological station measurements with sea surface temperatures based in early years on ship measurements and in recent decades on satellite measurements. Uses of this data should credit the original sources, specifically the British HadISST group (Rayner and others) and the NOAA satellite analysis group (Reynolds, Smith and others).”
Why switch to HADISST, or why not switch back to the NCDC SST data when the ERSST.v2 was released in 2004?
The linear trend of the global ERSST.v2 SST data from 1880 to 1981 was 0.024 Deg C/decade, Figure 1, while the HADISST SST anomaly data had a linear trend of 0.032 Deg C/decade. In other words, the HADISST makes the world seem a little bit warmer.
And why not simply use the HADISST-based data through the entire term of the data, from 1880 to present? In other words, why retain the OI.v2 SST data since 1982 and splice it to the HADISST data? They’re both satellite-based during that period.
The OI.v2 data has a slightly greater trend, 0.095 Deg C/decade versus 0.085 Deg C/decade. Refer to Figure 2.
HADISST, ERSST.v2 and OI.v2 SST data are available through the KNMI Climate Explorer: