Improved diurnal variability forecast of ocean surface temperature through community model development

Abstract

During daytime, under low winds and due to solar heating, the skin and sub-skin temperatures, typically measured by satellites can increase by several degrees compared to the foundation temperature. Diurnal variability has been observed in the Mediterranean [5], western North Atlantic [1] and the global ocean [2,3] from in situ and satellite observations. Diurnal heating has been reported at higher latitudes [4] and an extended study to characterise the regional patterns of diurnal SST variability over the Atlantic Ocean and the European Seas [5], showed frequent occurrences of diurnal warming events reaching several degrees, in the Baltic Sea. Nonetheless, diurnal SST variability it is not fully resolved by ocean and coupled oceanatmosphere models. Although some of the important diurnal variability and cool skin effects [6] have been shown to be reproducible [7], the vertical grid resolution of the models is of meter-scale. In addition, regional CMEMS ocean forecasting systems only assimilate a single SST field per day, representative of night-time conditions when the water column is well mixed and thus, no diurnal signal is present. Such simplification of the SST has been reported to cause biases in the estimated surface fluxes [8,9]. The implications associated with the lack of a properly resolved SST daily cycle in atmospheric, oceanic and climate models have been quantified in terms of heat budget errors mostly in the Tropics. Heat flux errors associated with the warm layer development were reported in [9] to range between 10 and 50 Wm-2. In regions with diurnal warm layer formation, [10] reported an annual mean surface flux out of the ocean that reached up to 9 Wm-2. In addition, strong SST diurnal signals can complicate the assimilation of SST fields in ocean and atmospheric models, the derivation of atmospheric correction algorithms for satellite radiometers and the merging of satellite SST from different sensors [11]. Not accounting for the daily SST variability can cause biases in the prediction and modelling of algal blooms, especially as cyanobacteria blooms in the Baltic Sea are promoted by high SST values [12] - and the estimated net flux of CO2, as the outflux of oceanic CO2 is positively correlated with the increase of SST [13]. Various models exist for the description of the diurnal cycle and their complexity varies from empirical parameterisations, based on various input parameters such as the surface winds and heat fluxes, to turbulent closure models. Parameterisation models are typically developed based on observational evidence at specific locations and depths, thus carry the uncertainty of their parameters and forcing fields, typically from Numerical Weather Prediction (NWP) models not resolving the SST diurnal cycle. Such parameterisations were compared to SEVIRI SST derived signals in the North Sea and the Baltic Sea [14] with moderate results. More sophisticated models such as turbulence closure models can resolve the vertical extend of the diurnal signal but are computationally expensive. The one-dimensional General Ocean Turbulence Model [15] was shown in [16] to perform very well in reproducing the vertical temperature structure as described by satellite SST and buoy measurements. The success of such modelling attempts highly depends on the accuracy of the input fields, typically obtained from atmospheric models. Consequently, there is a need to evaluate the impact of properly resolving the daily variability of SST in atmospheric models as well. When examining very strong diurnal warming cases, it was found that updating the SST every 6hours in the meso-scale model WRF, as opposed to using one daily value, resulted in average day-time differences of up to 20% for the 10 m winds and up to 40% for the surface heat flux [17]. The “Improved Diurnal Variability Forecast of Ocean Surface Temperature through Community Model development (DIVOST-COM)” project will “improve the representation of diurnal variability and cool skin layer in forced ocean and coupled ocean-atmosphere models” and the aim of this report is to provide an overview of the project. Section 2 describes the project methodology, section 3 describes the background and expected outcomes while the main conclusions are presented in Section 4.