|
Preliminary Calculation of Radiative Flux for SHEBA Using Observed Detailed Cloud Information and New NASA GISS Model |
Columbia University, New York City
NASA GISS, New York City
NOAA, Boulder, Colorado
NOAA, Boulder, Colorado
SHEBA's intensive in-situ measurements of very detailed vertical distributions of cloud properties (March to June, 1998) and radiation for the first time offers the opportunity to realistically study the effects of clouds on radiation as well as test radiative transfer model's capability of dealing with all macro- and microphysical properties of clouds generally, and in particular, for Arctic regions. After years work, the NASA GISS radiative transfer model has been added to many features that can handle such detailed cloud information.
In this preliminary calculation, three cases of ice, liquid and mixed cloud phases are employed to calculate radiative fluxes. Each case has duration of 2 hours and is sampled every 5 minutes in the calculation. Their cloud information includes vertical structure (single/multiple cloud layers with observed cloud top and base positions), phases, particle sizes and optical thickness (derived from water content). Supplementary rawinsonde data, approximately at the center of the 2-hour duration, supply high-resolution temperature and humidity profiles. Surface total albedos are adopted from Curry et al., 1999 (fig. 8, measured by C-130Q), but the spectral dependence is not yet correctly specified. Aerosols, ozone and trace gases are from GISS GCM's climatology. Cloud fraction is specified as 100% since all the observations have clouds. These specifications leave many uncertainties for SW calculation. By comparing with observed surface fluxes of two 1-hour averages whose 2-hour periods are exactly the same as the 2-hour calculations' for all the case, the mean differences over all 3 cases between the calculated and the observed are -54.76 and -63.63 W/m2 for downwelling and upwelling SW respectively, and -1.9 and 2.7 W/m2 for downwelling and upwelling LW respectively.
These results show that LW fluxes are very well modeled by GISS model compared with that LW modeling has more problems in Curry et al. 1999 while our SW calculation seems more problematic partly because our inaccurate specifications of surface albedo, aerosol and ozone. These cases with high vertical and time resolution will be used to explore the importance of small-scale cloud variation on surface fluxes.
