NASA GHCNv4 land weather station locations in 1880 (almost none = lot's of wild guessing "infilling" of temperature numbers needed to compile a global average) versus 2020 (still sparse coverage in many locations)