Overall, from 1880 to the present, approximately 99% of the temperature data in the USHCN homogenized output has been estimated (differs from the original raw data).
And didn’t they “lose” a lot of the original data?