Restarting does not seem to work correctly when dealing with 4M GVO
Here is a comparison between:
- Analyses starting in 1999 using François' previous calculations (1960 - 2000) from 2000 to 2005, forecasting until 2010.
- Analyses restarting from previous analysis run from 2000 to 2020.
Calculations were tested with and without COMPOSITE data to check whether discrepancies came from this or not.
Re-analyses and forecast from previous 4M analyses does not seem to work correctly.