Elucidating the response of the ocean to the Madden Julian Oscillation

The MJO is one of the dominant intraseasonal modes of variability of the atmosphere that impacts not only the tropical Indo-Pacific basins, but also remote locations. Previous studies have demonstrated the importance of ocean processes in the simulation and prediction of the MJO, although how the MJO signal impacts the upper ocean, and feedbacks back to the atmosphere remains without consensus. Here we use an ocean-only model to elucidate the physical processes controlling the ocean response to the MJO. This requires a model with sufficient horizonal and vertical resolution together with a forcing dataset that captures the wind and precipitation fields associated with the MJO. With regard the latter we have assessed a number of reanalysis products: ERA-interim, ERA5 and JRA55. On the intraseasonal timescale, ERA5 shows the best representation of the MJO-related atmospheric properties compared to observations. While both ERA-interim and JRA55 underestimate the magnitude of some of the MJO-related atmospheric variability (OLR, precipitation, etc.) in all 8 phases of the MJO, ERA5 captures well both the spatial pattern and magnitude of the MJO-phase composites, especially precipitation. To assess the impact of the differences in the reanalysis products we have forced a high-resolution ocean model with both ERA-interim and ERA5. The ERA-interim forced simulation has a reasonable lead-lag relationship between atmospheric variability (such as precipitation) and SST on intraseasonal timescales near the Maritime Continent region. Preliminary results using ERA5 show differences in precipitation that produce significant differences in near surface salinity and temperature, compared with using ERA-interim. How much this affects the ocean response to individual MJO events, as well composites, will be determined. Examination of heat and salt budgets will be used to determine the physical processes controlling the ocean response with a view to recommending what processes need to be captured in coupled models.