Dealing with discontinuous meteorological forcing in operational ocean modelling: A case study using ECMWF-IFS and GETM (v2.5)

Abstract

Meteorological data providers release updated forecasts several times per day-at the forecast epochs. The first time step (t D 0) of each forecast, the so-called analysis step, is updated by a data-assimilation process so that the meteorological fields at this time in general do not match the fields from the previous forecast. Seen from the perspective of oceanographic modelling, the analysis step represents a possible discontinuity in the model forcing. Unless care is taken, this "meteorological discontinuity" may generate spurious waves in the ocean model. The problem is examined and quantified for a single meteorological model: the European Centre for Medium-Range Weather Forecasts (ECMWF) Integrated Forecasting System (IFS). A simple straightforward solution is suggested to overcome the forcing discontinuity and the effect on two particular ocean models is examined: the FCOO NA3 (North Atlantic 3 nm) stormsurge model and the NS1C (North Sea-Baltic Sea 1 nm) circulation model. 1 Introduction Numerical ocean models are in operational use at many institutions around the world. In the present paper, a particular issue related to an inherent discontinuity of operational meteorological forcing data will be examined. It will be shown that a naïve implementation of an operational forecast process may lead to discontinuous forcing data, which excite spurious waves in the oceanographic model. Further, unless care is taken, there could be a difference in the forcing data applied in, respectively, forecast/nowcast and hindcast modes. This difference may result in under-prediction of forecast/nowcast errors of ocean models. The discontinuity problem is examined and a straightforward solution is proposed. The effects of the discontinuity on a particular ocean model are quantified and results with and without use of the outlined solution are examined.