A couple of people have been going on about “this year being warmest ever” which is probably not true (see link), however, it did start me wondering. Because 1/f type noise as present in the climate, tends to have long term trends, it is inevitable that there’s more chance of seeing the warmest or coolest at the end of the trend than would be the case for what I called “scientific noise” (white Guassian).
But I’d never attempted to put a figure on it. But now with data over 7000 runs from my 1/f noise generator, I can quickly check that out. The data is divided into “decades”. So I checked how many of the 7335 decades had the last highest. In my data that is 8.68%. By pure chance with 16 decades, we would expect any one decade to be high 6.25% of the time. So in 1/f noise, the last decade is “warmest ever” 40% more times than expected (for both 16 and 10 decades)
| 0.372926 | 16 “decades” | 10 “decades” | 5 “decades” |
| Runs with last highest | 637 | 8401 | 26930 |
| Total runs | 7335 | 58680 | 102690 |
| % of total with last “highest ever” | 8.68% | 14.31% | 26.22% |
| Expected % | 6.25% | 10% | 20% |
| Ratio to that expected | 1.39 | 1.431 | 1.311 |
That means since 1958 (when CO2 was started to be measured) we have 5 whole decades. According to normal “scientific” noise, there’s a 20% chance of the last being highest. But because it is 1/f type noise, there is a 26% chance of it being highest. If we then count “lowest” (as the period started with a global cooling scare), then there was over 50% chance that we would now be having either a global cooling or global warming scare right now.
Note: this is without fabricated global temperature data – it’s almost certain if the data is being warmed by the likes of NOAA and NASA
