Thread by @ClimateOfGavin

Every so often someone will post a link to an anonymous website that purports to show recent global temperatures. It doesn’t.

This data doesn’t look any other data set - either from weather stations, reanalyses, satellites, radiosondes, or argo floats, etc.

Nominally the data purports to be some kind of running average of weather station anomalies updated continuously as new data are posted.

While not an obviously terrible idea, there are lots of pitfalls - area weighting, etc. but you could do this properly if you wanted.

What I think they are doing though is just adding new data in and taking an average without any allowance for where the data are from or biases in that distribution. That would never work.

Of course, no one has a clue what they are actually doing; no public code, no publication.

But I think they haven’t even coded their own algorithm properly. Notice the very rapid oscillations at the beginning of the series that die away over time? No actual climate time series looks like that.

It’s what happens when you take averages over longer & longer periods tho…

My guess is that they are just putting new data into a summation and dividing by an ever larger ‘n’ - so that each new bit of data is progressively less influential.

Needless to say this gives you no useful information about ongoing temperature changes. 🤷

Feel free to judge folks who push this nonsense accordingly.

PS. I’ll happily correct any of this if someone can post a link to the code.

PPS. Code and data for GISTEMP is all public data.giss.nasa.gov/gistemp/source…