". . . Another thing: according to Christy (personal communication through Spencer), the margin of error for the estimates of annual global average temperature for the satellite estimates is 0.1˚C.
With that in mind, the difference between any given year and the next-warmest in the satellite record exceeded the margin of error in only one case: 1998 (second-warmest in the record) was 0.107˚C warmer than 2017 (third-warmest).
The difference between 2016 (the warmest year) and 1998 was only 0.028˚C, or about three-tenths of the margin of error. In other words, we don’t know whether 2016 or 1998 was warmer. The fourth- and fifth-warmest years (2010 and 2015) are also within the margin of error from each other.
One has to go from the sixth-warmest year (2002) to the twelfth-warmest (2001) to get a gap that exceeds the margin of error again; i.e., we don’t know which of 2002, 2005, 2003, 2014, 2007, 2013, or 2001 was actually the sixth—or the twelfth—warmest year, or anything in between.
All that makes it pretty clear that global temperature has plateaued over the last twenty years. We simply don’t know whether “Earth’s long-term warming trend” stopped in 1998, will resume sometime, or will reverse and turn into a cooling trend.
This isn’t even to broach the question of what caused the warming from 1880 the present—or, rather, as shown in this graph by NOAA of global land and ocean surface temperature anomalies (which, unlike satellite data, are subject to great doubt because of spatial distribution, measuring station dropouts, homogenization methods, and other problems), the cooling from about 1880–1910, the warming from about 1910–1945, the cooling from about 1945–1975, the warming from about 1975–1998, and the plateau from about 1998–2015. (We do have a pretty good idea what caused the warming of 2015–2016 and into 2017: an extraordinarily strong El Niño, similar to the one that made 1998 so warm.) . . . "