Record high temps are often used as evidence that AGW is happening. The assertion being that somehow record hot days today proves the planet is heating up.
This post will simulate one 30 day span (a month) for temperatures between 30 and 40C so we can see how these record temps get filled in.
Record highs have nothing to do with temperature trends. It has everything to do with accounting and the span of time records have been kept.
For the first year of record keeping, every day is a record breaker regardless of the temp, because there are no records to compare to. As the years go on, fewer and fewer records will be broken until a point is reached when every possible temperature that can be achieved will be filled and no more records can be broken.
This simulation was simple enough. A range of randomly selected temperatures in 1/10C was chosen between 30 and 40C for 30 day span (one month) for as many years as was necessary to fill all 1200 possible spots.
Each randomly picked temp for any of the 30 days was evenly possible, no weighting of any temps.
Each run of a selected temp for any given day, if highest for that day, would be considered a record breaker for that day and recorded as such. Soon as all days hit 40C (1200) the looping stops.
The results are rather interesting. The last year in which all days became 40C s 2155. By 2010, 21 of the days were full, 13 of which were before 1955. Thus 2/3s of the record breaking days were before 1955.
The simulation was paused at 2010 to see what the record temps were for each day.
Notice again that years before 1955 dominate the records.
Now the flaw in this simulation is that each temp picked has equal value. That is, 40C is just as likely picked as 30C. We know this is not the case in the real world. Lower temps are more likely than higher temps. This will force the fill of all possible temps much further into the future.
This table shows actual percent of temps for Melbourne Australia:
So using these numbers we can allow for the random numbers to be within that percent range. This is done by selecting a random temp, then selecting a random percent. If the random percent is at or below the percent in this table for that temp, then the temp is used, if not a new temp is randomly selected. This is looped until a temp under its percent is selected.
The slots are finally filled by the year 3000.
So there you have it, it would take at least an additional 2000 years to fill all possible temps and no longer have record breaking days. Thus our puny 100 years is a small percent of time. Record setting days has nothing to do with trends in temperature.