Several people on both sides of the debate continue to criticize my averaging of the entire measured USHCN data set.
The graph below shows the average of three different USHCN groups of data since 1990, which was the year they started exponentially losing data.
The raw data is green. It shows a small warming since 1990, all of which occurred before 1998. The final adjusted data is blue, and shows much stronger warming. The fabricated data (temperatures marked with an “E”) shows a very strong warming, and is the component of the final data which creates almost all of the difference between final and raw.
According to the USHCN V1 documentation, there is no additional adjusting needed after 1990, and according to the USHCN V2 documents they use the same TOBS algorithm as V1. So other than infilling, final should match raw after 1990.
I don’t see how this could be any clearer. Infilling of fabricated temperatures is causing the vast majority of reported warming since 1990. The reason I see this and others don’t – is because I use the actual data reported by USHCN exactly as it is reported.
Note how the divergence coincides with the beginning of wide scale station data loss.