The impact of Argo observations in a global weakly coupled ocean-atmosphere data assimilation and short-range prediction system

The Argo array is a large component of the ocean observing system on which operational ocean forecasts rely. Global observations of the sub-surface temperature and salinity from Argo enables the accurate initialisation of ocean forecasts, improving the position of currents and the overall energy available for air-sea interactions. Such constraints on the sub-surface ocean are important for coupled initialisation for seasonal forecasting where the atmosphere is coupled to the upper ocean. As operational centres move toward using coupled atmosphere-ocean models for numerical weather prediction (NWP), it is useful to understand how ocean observations contribute to short-range coupled forecasts. Using the Met Office Weakly Coupled Data Assimilation system, we have run Observing System Experiments (OSEs) to investigate the effect of withholding Argo observations on short-range coupled ocean-atmosphere analyses and forecasts. Withholding Argo significantly increases the sub-surface temperature and salinity RMS errors with consequent changes in the upper ocean heat content and positions of currents. Overall, there is little systematic impact of these ocean differences on global atmosphere analyses. However, in a case-study over the period of hurricane Sandy (October 2012), we do see that Argo observations have some impact on atmospheric forecasts. The assimilation of Argo observations affects the position of the Gulf Stream leading to a small effect on the position and intensity of hurricane Sandy as it made landfall. As the resolution of the system is quite low compared to operational NWP systems, seeing even a small impact is encouraging. We might expect with future increased-resolution coupled models that this influence of ocean observations on the atmosphere would only increase.