After revising the MJO index calculation method using the retrieve_MJO_projection_data function provided in the updated AI_WQ_package, the MJO index calculated from the ERA5 data (U200, U850, and daily mean TTR, 1979–2026) aligns perfectly with the updated retrieve_annual_training_data under the “Training Data — AI_Weather_Quest 3.9 documentation”.
However, it still does not align with the retrieve_daily_MJO_obs data found under the “Forecast Evaluation — AI_Weather_Quest 3.9 documentation”.
What could be the cause of this? It is quite puzzling why the results derived from the exact same dataset and (revised) methodology match one set of official data but fail to match the other. Could this discrepancy be due to a specific calculation detail, or might there be an issue with the files provided on the official website?
Could you please look into this and verify? I would be extremely grateful if you could help clarify this or share any missing details.
