It was discovered via the EDS beta test that erroneous values existed in the design freezing index computations.
The root cause was that the three warmest (least) values were used, rather than the coldest (greatest) three values in the summary time period (e.g. 2070-2099).
This PR edits the avg_extremes function such that it always selects three greatest values from the sorted array of possible freezing (or thawing) index values.
This isn't in our current quarterly theme of data projects so I think a single reviewer is OK here.
It was discovered via the EDS beta test that erroneous values existed in the design freezing index computations. The root cause was that the three warmest (least) values were used, rather than the coldest (greatest) three values in the summary time period (e.g. 2070-2099).
XREF https://github.com/ua-snap/arctic-eds/issues/181
This PR edits the
avg_extremesfunction such that it always selects three greatest values from the sorted array of possible freezing (or thawing) index values.This isn't in our current quarterly theme of data projects so I think a single reviewer is OK here.
Closes #47