Summary
To allow SP calculation before we have another CR to interpolate to, the last point of the GLOWS L3d should be repeated, with a time 1CR out, which effectively "freezes" the L3e SP at the last valid value.
Ultimately this should result in a quality flag in L3d, L3e, and the map, but that might not be doable before the May 22 date.
GLOWS is implementing this in their pipeline, so we will just have to integrate it.
- After the last interpolated record of L3d, create a new, non-interpolated record solely from the L3bc inputs of the current CR, and timestamp it with the midpoint of the data (close to the middle of the CR)
- duplicate this record and timestamp to the middle of the next CR
See this diagram from Izabela.
Summary
To allow SP calculation before we have another CR to interpolate to, the last point of the GLOWS L3d should be repeated, with a time 1CR out, which effectively "freezes" the L3e SP at the last valid value.
Ultimately this should result in a quality flag in L3d, L3e, and the map, but that might not be doable before the May 22 date.
GLOWS is implementing this in their pipeline, so we will just have to integrate it.
See this diagram from Izabela.