Revealing long-term changes in the North Atlantic air-sea fluxes from provisionally corrected VOS observations (1900-2022)

We use Voluntary Observing Ship (VOS) observations available form the ICOADS collection for estimating surface fluxes in the North Atlantic for the period from 1900-2022. One problem of the use of VOS observations for deriving long-term air-sea flux time series is associated with inhomogeneous in space and time sampling, especially during the period prior WW2. Another problem is associated with systematic biases in a number of VOS state variables (first of all cloud cover) for the first part of 20^th century. To derive surface flux anomalies we first reconstruct turbulent heat fluxes from 1900 onwards for the whole North Atlantic from EQ to 70 N. To homogenize sampling density and obtaine more robust estimates we use the procedure of sub-sampling for the earlier decades and then integrate computed turbulent heat fluxes in the coordinates of steering parameters (vertical surface temperature and humidity gradients on one hand and wind speed on the other). Biases in cloud cover are associated with changes in the observational practices of in the early 1950s when WMO implemented new standardized coding system. These biases have the effect of systematic underestimation of total cloud cover during 1900-1940 compared to the past WW2 period ranging from 0.3 to 1 octa and imply biases in short- and long-wave radiation of up to 10 W/m² and 4 W/m² respectively. We explored all sources of these biases using direct analysis of early 20^th century log-books and performed correction of cloud cover using cloud cover probability density functions. Then short- and long-wave radiative fluxes were computed using state of the art bulk parameterizations. Thus, we obtained long-term time series of turbulent heat fluxes and radiative fluxes for 120-yr period 1900-2022. Analysis of centennial trends shows upward change in sensible plus latent flux ranging from 3 to 14 W/m² during 120 years, while the increase over the last 40 years amounts to 6-7 W/m² with the major growth during the 1990s and early 2000s. Radiative fluxes demonstrated increase in short-wave radiation (positive directed to the ocean) of 3-5 W/m² in the Atlantic subtropics and mid latitudes and weak or close to zero trends in long-wave radiation. While changes in radiative fluxes partially compensate opposite trends in turbulent fluxes, the upward tendency in ocean heat budget (atmosphere gains) remains significant with magnitude of 2-6 W/m² over 120-yr period. Interdecadal variability of surface turbulent fluxes is of an order of magnitude stronger compared to the radiative fluxes (10-20 W/m² vs 0.5-2 W/m²), thus implying the dominant role of turbulent fluxes on forming long-term changes of the ocean heat budget. Further interdecadal variability of surface heat budget is discussed in the context of the North Atlantic multidecadal variations.

Research is funded by RSF project # 23-47-00030.