One could say using 90 years of data when we have 140 seems as arbitrary as using 10 years of data when we have 140. On the flip side, I can’t be the only one who thinks weather records from the 1800s are suspect due to technology or procedures or both. So it might be fair to look at 90 years. If we do that for Pittsburgh (my location), I typically only compare apples to apples and use data from 1948-present (when observations began being taken at the airport). And, considering only data from those 73 years, I think I determined that 8 of the 12 months of the year recorded their hottest average temperature on record within the past decade.