These results use a spherical triangulation of land (GHCN) and ocean (HadSST3) temperature data with a baseline of 1961-1990.
Since August I have been forced to update to GHCN-V4 land temperatures since GHCN-V3 has been discontinued. September shows a small fall in the global temperature from August (0.832C to 0.793C). Fig 1 shows the September spatial temperature dependence.
Fig 1. Temperature field in 3D. Shaded triangles have measurements at each vertex.
Recent monthly temperatures since 1998 are shown in Fig 2. below for comparison
Fig2. Monthly averaged temperatures.
The annual averages updated using the first 9 months of 2019 are shown in Fig 3 below. The difference between using the V3 data and the new V4 data is large and worrying. V4 and V3 have different heritage. Although V4 has far more stations (~17000 with normalisation data compared to ~7400), there are still unresolved linkage of components to similar V3 composite stations, showing different temperature values. The graph below shows the annual temperature comparisons for V3 (up to August) and V4.
Fig 3. Yearly averaged temperatures
2019 looks almost certain to become the second warmest year. 2016 s the warmest.
Note how V4/HadSST3 temperatures since 2005 lie consistently ~0.03C higher than V3/HadSST3, although calculated in exactly the same way. The information that V3 had been discontinued was hidden in their status file (shown below). When I had previously emailed them asking why V3 was not been updated they told me there was just an IT problem.
********************************************************************************
09/19/2019
GHCNM version 3 is no longer operational (TAVG, TMAX amd TMIN). Users should use
GHCNM version 4 for TAVG, found here: ftp://ftp.ncdc.noaa.gov/pub/data/ghcn/v4/
In addition Version 4 TMAX and TMIN is under development and will be available in
the future.
********************************************************************************
Interesting that the UAH satellite reported a HUGE jump (0.6 C) in September, the highest in 40 years of monitoring. Is there much month-to-month correlation between GHCN V3 or V4 and the satellite measurements?
http://www.drroyspencer.com/2019/10/uah-global-temperature-update-for-september-2019-0-61-deg-c/
The ground based temperature measurements are normally in good agreement with each other at least as far as trends go. So V4 and V3 agree but V4 shows a warming offset of about +0.03C, which I find a bit suspicious. Normally there is fair agreement with satellite data.
However the UAH results for September are way above normal. Satellites measure temperature across a range of heights in the atmosphere. Roy Spencer now blames this jump in the September temperature on a sudden stratospheric warming over Antarctica and knock on effects.
https://www.drroyspencer.com/2019/10/record-antarctic-stratospheric-warming-causes-sept-2019-global-temperature-update-confusion/
I guess that NH summer is the coolest global season, as the Earth points its bald white polar head towards the sun. A challenge for spectrum analysis gurus?
Bizarre, but par for the course, that only TAVG is now available, not a quantity that I lie awake at night worrying about, but one guaranteed to give headlines. Bizarre also because the computer runs that generate TAVG from daily data are probably also generating Tmax and Tmin.