I am considered a heretic by some on both sides of the global warming debate, because I refuse to tamper with the raw US data.
I have already shown why you can’t detect baseline tampering using anomalies, and thus should not use them. I have also shown how infilling missing data massively biases the US data towards warming over the past 20 years.
Now I am going to show why gridding is unimportant. The graph below shows my ungridded USHCN final averages in blue, and gridded NCDC temperatures in red.
The slope is almost identical, and the offset is probably mainly due to NCDC’s extra co-op stations. Gridding is unimportant for US data, because the stations are fairly evenly spaced, by design.
On the other hand, the data tampering being done by USHCN is huge. People on both sides of the debate should quit focusing on meaningless precision in tampered data, and instead focus on the huge inaccuracy created by tampering.
With large data sets with a random distribution of error (like missing data) real scientists do not attempt to make corrections. As soon as corrections start being made, the door is opened to confirmation bias at best – and it gets worse from there.
Leave the data alone. You are falling for an old NOAA mind trick.