Closed casadoj closed 8 months ago
I have changed the code in 6_1_skill_computation_station.ipynb so that it is possible to compute hits, misses and false alarms for each station on a seasonal basis. The results of this computations are stored in the folder dada/hits/seasonal/.
Later, I have explored the results in a new version of the code 6_4_analysis___.ipynb. Unfortunately, most of the plotting functions don't work with the seasonal data. I guess that the amount of events is so low that most of the skill metrics are simply NaN.
It requires more time to explore, but I reckon there are other issues more interesting that this one. I leave it in stand by.
I close this issue because it seems that the seasonal analysis was not interesting enough. The functions are there, but I removed it from the notebooks to make them simpler.
Gwyneth Matthews (ECMWF) proposed that it would be interesting if the current skill analysis would be disaggregated in seasons. They are interested in these results because it's interesting for their post-processing algorithms.