Pre-requises for deciding whether or not there are problems with leap year interpolations

[andrewphilipsmith]

Aim to decide on this on 2024/10/08. Prior to that meeting:

@andrewphilipsmith

• Determine the cause of 1982 crashing
  • if possible run pipeline for 1982 crashing

@griff-rees

• Complete plots
  • Skip 1982 if the relevant resampled data is unavailable.
  • Plot of first few days of each file, compering CPM Raw and Resampled data.
    First two days of each file (eg beginning of Dec) plot Raw vs Resampled.
    Plot leap-years and non-leap-years separately
    Hyposithesis: There may be a systematic error on the 1st Dec on leap-years, that does not occur on non-leap-years.
  • Plots of day-of-year aggregates for leap-years and non-leap-years.
    EG two plots, each with two curves.
    1. Leap-years. One curve of raw CPM and a second curve of resampled CPM
    2. Non-leap-years. One curve of raw CPM and a second curve of resampled CPM

Extra things - not required for Tuesday, but might be valuable at a later date

• standard dev for day-of-year seasonal plots