NOAA makes a large adjustment to US temperatures known as the Time of Observation Bias Adjustment (TOB.) Here is their official explanation.
They claim that the adjustment has little effect until 1970, then causes 0.3F warming from 1970 to 1990, and none after 1990.
Applying the Time of Observation adjustment (black line) resulted in approximately a 0.3F warming from the late 1960’s to the 1990’s
Their actual adjustment looks nothing like what their documentation claims. It cools the past in a hockey stick by 0.7F, and it gets steeper after 1990, not flat. Their documentation (like the adjustment) is fake.
The TOBS adjustment turns a hundred year cooling trend into a warming trend, by massively cooling the past.
The theory behind the adjustment is that in the past people tended to reset their min/max thermometers only once per day in the afternoon, causing double counting of hot temperatures from the previous day. Now people tend to reset in the morning, causing double counting of cold days. If it were true, it would be a real problem and would indicate that station operators are complete morons.
Suppose day one is 90 degrees at 4 PM, and a cold front comes through at 5 PM and it starts snowing. If the thermometer was reset at 4 PM, then the maximum temperature for day 2 would also be recorded as 90 degrees, even though the temperature never got out of the 30s. This is not a subtle problem. I had a min/max thermometer when I was eight years old, and I recognized the problem after about two days. So I reset the max in the morning, and reset the min in the afternoon – problem solved. I assume that anyone operating a min/max thermometer would come up with a similar solution.
But NOAA assumes that their station operators are idiots and applies their fake TOB adjustment, which is nothing even remotely like their documentation claims it is.
It is simple enough to test if the adjustment is necessary, by simply eliminating any stations which might be at risk. NOAA keeps station histories, and I identified which stations reset their thermometers during the afternoon in 1936, and which ones reset their thermometers in the morning in 1936.
According to NOAA documents, there were 273 stations which reset their thermometers at morning or night during July 1936 and 861 stations which reset their thermometers in the afternoon. I plotted the trends for both groups of stations and found that the trends were essentially identical. the only difference is that morning stations are warmer because they are located at lower latitudes than afternoon stations. People tend to be more active in the morning where it is warmer.
What this shows us is that the TOB adjustment is junk science. If TOB was a real problem, the slope of the morning stations would be more positive than the slope of the afternoon stations. This is because morning stations would have been recorded too cold in the past, and afternoon stations would have recorded too hot in the past.
But we don’t see that. The slope for morning and afternoon stations is essentially identical. Instead of doing a simple check like I did, NOAA simply applies their fake adjustment, and then Zeke and Nick write about it proudly on skeptics’ blogs.
Read more at Real Climate Science