We investigate the behaviour of the tropical intraseasonal variability in the Bureau of Meteorology Coupled Ocean-Atmosphere Model (BCM). This coupled model has been the basis of the Bureau's dynamical seasonal prediction system for last 3 years. In order to produce skilful seasonal predictions for the tropical regions, it is important that the BCM simulates the tropical intraseasonal variability (TISV) with realistic characteristics. However, it was found that TISV simulated by BCM is dominated by the westward propagating component, in contrast to the results of observational analyses. This is also at odds with the results obtained from the AMIP-style runs of the atmospheric component of BCM, in which eastward propagating tropical variance dominates. Therefore, this systematic error in the BCM simulated TISV was hypothesised to arise, at least in part, from the coupled model's drifted basic state. In this study, we examine the impact of the drifted basic states of the coupled model on its simulated TISV. Several AMIP-style experiments were conducted using the atmospheric component of BCM forced by the AMIP SST dataset and the drifted SST fields from previous BCM simulations. Differences in the simulated TISV characteristics were analyzed using standard techniques like Hovmoller diagram, wavenumber-frequency spectrum, and the lag-regression analysis of representative variables. Results of this analysis, along with the possible physical mechanisms involved, will be discussed.