I think editing the function below to remove the low conditional and to always summarize the three largest values will yield correct design values for both freezing and thawing index values.
def avg_extremes(arr, low=True):
# if first element is np.nan, then all should be
if np.isnan(arr[0]):
return np.nan
# otherwise, sort and return avg
# based on `low` param
arr = np.sort(arr)
if low:
return arr[:3].mean()
else:
return arr[-3:].mean()
XREF https://github.com/ua-snap/arctic-eds/issues/181
I think editing the function below to remove the
lowconditional and to always summarize the three largest values will yield correct design values for both freezing and thawing index values.