Have model, need forcings. Towards the development of accurate radiative forcing fields for sea ice model experiments (INVITED)

Ice-ocean models are the tools of the trade as researcher are trying to understand the recently observed changes in Arctic sea ice. Lacking an atmosphere, experiments with these models depend on externally specified "forcing" fields. The answers these models provide, are sensitive to the accuracy of these forcing fields. In this paper we will concentrate on the particularly problematic aspect of providing radiative fluxes for such experiments. Radiative fluxes dominate the surface energy balance over sea ice for most of the year and sea ice thermodynamics are therefore very sensitive to errors in the radiative forcing fields. Traditionally sea ice model experiments have relied on radiative fluxes which were parameterized in terms of observations of temperature, cloud amount and water vapor. Unfortunately, very little has been known about the spatial and temporal variability of some of these parameters so that the role of radiative fluxes on modelled sea ice variability is masked. Alternatively, researchers have used radiative forcings adopted from the reanalysis data sets produced by NCEP and ECMWF. However, these fields suffer from substantial uncertainties and contain significant biases. In this paper we discuss the benefits and pitfalls associated with these approaches. We present a new approach based on the combination of surface observations and satellite measurements from the TOVS Path-P data set to construct a 20-year time series of downwelling long and shortwave fluxes for sea ice model experiments. We present results from model studies using these data sets and discuss the impact of variability in radiative forcing fields on sea-ice model output parameters.