Some time ago, I got curious about the sources of data in the period where humans were measuring temperature using thermometers. When we talk about predictions

Some time ago, I got curious about the sources of data in the period where humans were measuring temperature using thermometers. When we talk about predictions from fifty years ago matching today (also I read the linked article and I didn't really see proof that a model that was specified at that time being explicitly benchmarked), it matters how the distribution of sources of surface temperature data varied since then. I haven't been motivated enough to do another one of these animations, but this video[1] shows locations of all temperature stations in GHCNv3[2] (I see that GHCNv4[3] is out). Notice how where humans measure temperature depends so much on what living conditions humans seek or find acceptable. I find this visualization interesting in terms of the number of modeling questions it poses. If you are interested, it shouldn't take much to replicate something similar with current hardware/software. As another note, it makes little sense to say model predictions are not statistically significantly different from each other. What is the population from which these models are being drawn? What is the measure of variability among models? The article[4] is behind a paywall, so I can't see what they did. [1]: https://www.youtube.com/watch?v=h95uvT67bNg https://www.youtube.com/watch?v=h95uvT67bNg [2]: https://www.ncdc.noaa.gov/data-access/land-based-station-data/land-based-datasets/global-historical-climatology-network-monthly-version-3 https://www.ncdc.noaa.gov/data-access/land-based-station-dat... [3]: https://www.ncdc.noaa.gov/data-access/land-based-station-data/land-based-datasets/global-historical-climatology-network-monthly-version-4 https://www.ncdc.noaa.gov/data-access/land-based-station-dat... [4]: https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/2019GL085378 https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/2019...