Unified Forecast System Model Prediction of the Madden-Julian Oscillation and East Pacific Teleconnections During Boreal Summer

This study evaluates the subseasonal to seasonal (S2S) prediction skill of the Madden-Julian Oscillation (MJO) and its teleconnected impacts on the eastern Pacific (EP) in the Unified Forecast System (UFS) model during boreal summer (May-October). Utilizing four experimental versions, Prototypes 5-8, the study finds that although the UFS model generally captures the MJO's propagation characteristics near the initializtaion time, it encounters diffculty in accurately predicting the propagation speed and decay rate of the MJO beyond 15 days. Specifically, the model simulates slower-moving MJOs that do not transition in phase quickly enough, although behavior is improved in Prototype 8. Teleconnected effects on the EP are delayed and too strong during the S2S forecast period, potentially leading to less accurate predictions of tropical cyclone (TC) genesis. The UFS produces too strong negative anomalies in intraseasonal genesis potential during MJO phases when east Pacific convection is suppressed, and too large positive anomalies during enhanced convection phase. Analysis of the vertically integrated moist static energy (MSE) budget reveals that all four UFS prototypes underestimate the damping effect of vertical MSE advection and the amplifying effect of longwave radiative heating, indicating weaknesses in tropical convective parameterization and cloud radiative feedback, although these biases are somewhat improved in Prototype 8. These deficiencies result in less efficient vertical MSE export weaker damping of MJO convection, slow MJO propagation, and delayed MJO remote impacts on EP. Thus, improving the UFS model's accuracy in capturing MJO propagation and maintenance processes is crucial for better predicting MJO's teleconnected effects on east EP TC genesis and enhancing S2S forecast capabilities for such extreme events.