Necessity of Sarah SIS and SID Interpolation about atlite HOT 7 CLOSED

Interpolation is necessary for older years in sarah, i think there was an analysis in Pfenninger 2016

from atlite.

Could you determine whether the interpolation is necessary?

from atlite.

Yes, it's not the case, the interplolation function is fast.

from atlite.

Are you sure about this?
I dug out the profiling I did last week (small UK cutout):

Based on that

• I would agree that a single interpolation is fast (time percall < 1s)
• The number of calls is extremly high due to the way the data is decomposited such that we get ncalls = 6221 in the example and the majority of the cumulated time somewhere in the interpolation step (~ 4800 s).

from atlite.

Thanks for the clarification!

They mention as a reason for interpolating (pg. 1253):

For the analysis performed here, missing periods of 6 h or shorter are interpolated from neighboring values. Longer periods are filled by taking data for the same dates from the preceding year (or the subsequent year in the first year ofdata), and adjusting by the between-year difference in mean of the 7 days before and after the missing period. The amount of missing data and the difference between raw and filled SARAH data is shown later in Fig. 11.

The relevant area is the gray one.

What do you think @FabianHofmann ?

from atlite.

And some more fuzzy tidbit from me, either from 2000 or from 2010, there weren't any periods of missing data longer than 6h, that's why i implemented only the first half of what stefan did.

from atlite.

closing this, the interpolation can now be turned off in #20

from atlite.