The effect of uncertain river forcing on the thermohaline properties of the North West European Shelf Seas

Modeling studies and observations show that the thermohaline properties of the North West European Shelf Seas (NWESS) are sensitive to surface wind and heat flux forcing, as well as river outflows that transport fresh water from land to the ocean. In previous studies, it was assumed that the variability of the thermohaline properties in response to river outflow could be adequately sampled with a high-resolution, submesoscale permitting, long-term (i.e., 30-year) deterministic hindcast. In this study, we assume that the statistical distribution of the river forcing, rather than the time series of forcing itself, is adequately constrained by a 28-year history (1991 to 2018) of river forcing created specifically for our domain. In this way, we created an ensemble of 10 lower-resolution (approximate to 7-km), short-term (i.e., 2.5 years) hindcast models that are forced with randomly perturbed river outflows and an ensemble of surface fluxes from the 10-member ECMWF ERA5 reanalysis (the 'Test' ensemble) as well with a companion ensemble that is forced with the ERA5 surface forcing fluxes but unperturbed river outflows (the 'Base' ensemble) for the June 2016 through December 2018 time period. In both ensembles, the modeled evolution of 25-hour averaged (to partially filter out tides) temperature and salinity is realistic with peaks in summer for sea surface temperature and in winter for salinity, and annual amplitudes that are comparable to those found in other studies of the NWESS. The increased mean and standard deviation of the sea surface and bottom salinity in the Test ensemble are partly an artifact of the assumption that the errors in river forcing have a log-normal distribution that mimics the episodic nature of river outflow with a positive mean and an asymmetrical shape with a long tail toward large values. For surface density, the standard deviation in the Test ensemble was below 0.5 kg/m3, covering an areal extent larger than that for the Base ensemble throughout the year. The annual cycle of the areal extent of density in that range had a peak in summer and minima in winter, in phase with that of the river outflow forcing. Overall, the effect of uncertain river forcing on the thermohaline properties in this study is small. In order to understand the true impact of river forcings, better temporal and spatial observations of river outflow are needed.