Use of in situ observations to verify the diurnal cycle of sea surface temperature in ECMWF coupled model forecasts

At the European Centre for Medium-Range Weather Forecasts (ECMWF), coupling of the atmospheric model to an ocean model has recently been extended to the High-Resolution (HRES) forecast model. This move is driven by: (i) encouraging results in coupled Ensemble medium-range and extended-range forecasts; (ii) studies reporting a positive impact of coupling on the prediction of atmospheric events; and (iii) the need to better represent coupling feedbacks. As the lower atmospheric boundary condition over the ocean for the atmospheric model, the sea surface temperature (SST) field is key to this coupling. It is known that sub-daily variation in SST—its diurnal cycle— modulates air-sea exchange at the interface, and that resolving this variation can have a positive impact on the forecasting of both the upper ocean and the atmospheric state. In this work, we explore the diurnal cycle of SST in global coupled simulations with the ECMWF Integrated Forecast System (IFS). A year’s worth (March 1, 2015–March 1, 2016) of 10-day forecasts are run under coupled and then uncoupled model configurations: in the uncoupled runs, the (at depth) SST is based on the daily Operational Sea Surface Temperature and Sea Ice Analysis (OSTIA) field, with use of a prognostic scheme for the skin surface temperature; in the coupled runs the atmospheric model is coupled to the NEMO ocean model with a one-hour coupling frequency. The amplitude and phase of the diurnal cycle are validated against estimates from in situ oceanographic observations. We also explore the dependencies of the modelled diurnal cycle on key meteorological forcings (10-m wind speed and total cloud cover), and compare with the dependencies exhibited by the observations.