Someone suggested that the divergence between measured US temperatures and NCDC published US temperatures is due to Time Of Observation Bias.
I tested that out by comparing only versus stations which took morning or night readings on July 15, 1936. These stations should be biased cold in the 1930s, and produce more warming than the full set of stations. It doesn’t make much difference. The frequency of 90 degree days is still very low in 2014.

There is still a hockey stick of divergence between the measured and reported US temperatures.

Over the last 20 years, the TOBS adjustment should be a downwards adjustment. The 1936 afternoon stations are warming faster than the 1936 morning stations. USHCN TOBS adjustments do the exact opposite and make the problem worse.

TOBS does not account for the discrepancy between measured and reported US temperatures. Here is the list of morning/night stations
Continue reading →