FAR calculation

[mikelitzow]

Here's the plan going forward for calculating FAR values for regional SST anomalies:

1. Calculate GOA SST anomalies wrt 1950-1999 climatology
2. Calculate anomalies wrt 1950-1999 for hist.585 climatology for each model and preindustrial
3. Calculate probability of these anomalies (proportion of cases as large as or greater for each anomaly) for each model preindustrial simulation
4. Calculate probability of the same anomalies for 0.5, 1.0, 1.5, 2.0 warming in each model's hist.585 run
5. Calculate FAR for each model under 0.5, 1.0, 1.5, 2.0 "present"
6. Use brms to combine estimated FAR for each "present", using model weighting
7. Calculate proportion of anomalies greater than any yet observed for each "present"
8. Use brms to combine these expected proportion of greater-than-seen, using model weighting

[mikelitzow]

Step 2 should read:

Calculate anomalies wrt 1950-1999 climatology for each model, both for hist.585 and preindustrial

[mikelitzow]

Revised version:

1. Calculate GOA SST anomalies wrt 1950-1999 climatology
2. Calculate anomalies wrt 1950-1999 for hist.585 climatology for each model and preindustrial
3. Calculate FAR for each anomaly in the observed time series, for each model using years between warming = 0.5 and warming = 1.0 as present
4. Calculate FAR for each model for anomalies produced by each projection between warming = 1.0 and 1.5 as projected next "present"
5. Use brms to combine estimated FAR for each "present", using model weighting
6. Calculate proportion of anomalies greater than any yet observed for each "present"
7. Use brms to combine these expected proportion of greater-than-seen, using model weighting

[mikelitzow]

Two important updates:

• In step 2 above, I correctly note that hist.585 climatology should be used for calculating anomalies in both hist.585 and preindustrial runs. I incorrectly has used preindustrial climatology to compute preindustrial anomalies in the script, fixing that today.
• For step 3 above, I'm not sure that it's appropriate to use 0.5 to 1.0 degrees warming as the "present" for calculating FAR all the way back to 1950, when warming was below 0.5. So I'm changing to the period from 1950 to 1.0 degree warming as the "present" for calculating FAR for 1950-2021 observations.