The Crucial Role of the Initial State in MJO Prediction

Abstract The Madden Julian Oscillation (MJO) is a key driver of intraseasonal predictability. Accurate prediction of the MJO is challenging partly due to its complex interactions with the background state and sensitivity to initial conditions and unresolved processes. Using NOAA's Unified Forecast System, we explore the impact of subtle differences in initial conditions on MJO forecasts over the Indo‐Pacific. Model runs initialized with two independent reanalyses show significant and systematic differences in MJO circulation amplitude that emerge rapidly and persist over the 15‐day forecast period. Additional analysis suggests the reason is due to differences in the initial atmospheric static stability. Specifically, the reanalysis that is initially less stable is associated with enhanced large‐scale vertical motion and divergent winds throughout the forecast. Notably, a convection‐based MJO index shows less sensitivity to initial conditions, suggesting that the differences in MJO circulation amplitude arise through mechanisms other than changes in convection or diabatic heating.