Carter does pay lip service to the idea of a trend. But he disdains the trend for the entire data set — in fact he seems to disdain trends altogether — instead showing these two straight lines, one ending at 1997, the other beginning at 1999:
He then tells a fable about the 1998 el Nino causing some kind of “shift” in the fundamental state of the climate system, after which it “settled in” to a different basic temperature level. There’s a name for his kind of theory.
More to the point, the two lines he draws aren’t trend lines. He just drew two flat lines to give the impression of no change. There’s a name for that too.
Here are actual trend lines up to 1997, and after 1999:
I note that even his cherry picked horizontal lines don’t fit the data as well (i.e. have a higher chi squared statistic) as the trend line through the whole data set.