pcmdi_metrics
pcmdi_metrics copied to clipboard
PCMDI
Update central arctic domain
This PR expands the central arctic domain to remove the masking for the pole hole. On closer reading of the Ivanova (2016) paper, the pole hole was only masked for the ice edge analysis but not for the total ice area analysis.
@acordonez Thank you for the PR. If I understood correctly, before this PR, the sea-ice metric removes the pole hole by masking it out because there is no observation. With this PR, the metric will include the sea-ice area in the pole hole. Is that right? Is there any chance that this PR could make the model looking like overestimating the sea-ice extent in case model does good job, because the model has data over the pole hole (that is counted for sea-ice area) while observation does not (thus not counted for sea-ice area)?
I think it is okay to improve metrics beyond what Ivanova (2016) did. I am tagging @gleckler1 and @durack1 for their insights.
@lee1043 That's correct. The OSI-SAF data we're using does have data over the hole, not sure how that's accomplished.
@lee1043 Also looking at this code we've been including the sea ice data over the pole in the whole Arctic metrics, but not the central arctic metrics. If we do decide to mask the hole then I need to update the total Arctic region.
@acordonez Do you think this exclusion of pole hole can be set as optional capability with a parameter accompanying with?