10.07.2017: FB1-Seminar

11 Uhr, Hörsaal West, Düsternbrooker Weg 20

Abstract

We consider the potential of Voluntary Observing Ship (VOS) observations for estimating ocean surface heat budget at different time scales. VOS provide the longest coverage of the World Ocean by in-situ meteorological observations going back to the mid 19^th century. However, being collected primarily along the major shipping routes, VOS observations have spatially and temporally inhomogeneous sampling density, especially for the period prior 1950. In order to minimize sampling impact on the means and variability patterns derived from VOS-based surface fluxes, we propose a concept employing probability distributions of fluxes and flux-related variables. For turbulent heat fluxes we suggest the integration of fluxes (computed using COARE-3 algorithm) in the coordinates of steering parameters (vertical surface temperature and humidity gradients and wind speed) for which theoretical probability distributions are known (Weibull distribution and Modified Fisher-Tippet distribution). This approach allows for the switch from flux integration over the geography domain to the integration in the parameter space and, thus, obtaining less biased and more robust large-scale estimates. For short-wave radiative fluxes sampling uncertainties were minimized by "rotating local observation time around the clock" and using new probability density functions for approximating cloud cover fraction distributions. Then new algorithm for computing short wave radiation was developed and applied to VOS data. Similar approach was used for estimating long-wave radiation.

Analysis was performed for the North Atlantic latitudinal band from 25 N to 60 N, for which also estimates of the meridional heat transport are available from the ocean cross-sections. Over the last 35 years turbulent fluxes within the region analysed increase by about 6 W/m² with the major growth during the 1990s and early 2000s. Decreasing incoming short wave radiation during the same time (about 1 W/m²) implies upward change of the ocean surface heat loss by about 7-8 W/m². We further consider application of the proposed concept to the changes in surface fluxes on longer time scales analyzing time series going back to 1880s for turbulent fluxes and to 1920s for radiative fluxes. This allows for estimating full surface heat budget over the North Atlantic starting from 1920s. Magnitudes of multidecadal changes in radiative fluxes are considerably smaller compared to those implied by turbulent fluxes. Our analysis of radiative fluxes does not change the earlier conclusion that surface heat flux on multidecadal scale is closely correlated with SST-based AMO index, thus implying the active role of the ocean in air-sea interaction on multidecadal scale.

Liste aller FB1-Seminare