When extremes exceed historical records
In some parts of the globe such is in parts of Canada and other far north environments the extreme high temperature return periods for even a 1 in 100 year event can exceed the historical baseline most extreme 1 in 100 year event. In such cases the generation of a GEV value is not practical. In such cases we have opted to use a default 1 in 10,000 value and for the return period against the historical the obvious result is the extreme value will occur every year. The following screen captures for Kelowna, British Columbia exemplify this situation when an RCP 8.5 for extreme maximum temperature is applied for the year 2100. Note that the extreme value is still able to be charted and discerned in the charting function.
Note: When Baseline is toggled on the table function to the right of the graph after a future analysis is carried out the return period in the left column goes to 1.0 as this value will be reached every year.
When the GCM button is toggled the values in the middle column all exceed anything from the historical record therefore obviating the possibility of generating a GEC return value that makes any sense hence our decision to place the value at 1 in 10,0000. The values in the middle column are possible but returning a GEV return period is statistically impractical.
You can also add the 1 in 100 value for 2100 under RCP 8.5 to the graph (look at the top of the graph up from the 100 year return period year).