Ten-year global distribution of downwelling longwave radiation

[摘要] Downwelling longwave fluxes,DLFs, have been derived for each month over a ten year period (1984-1993), on a global scale with a spatial resolution of2.5x2.5 degrees and a monthly temporal resolution. The fluxes were computed using adeterministic model for atmospheric radiation transfer, along with satellite andreanalysis data for the key atmospheric input parameters, i.e. cloud properties, and specifichumidity and temperature profiles. The cloud climatologies were taken from the latestreleased and improved International Satellite Climatology Project D2 series. Specifichumidity and temperature vertical profiles were taken from three different reanalysisdatasets; NCEP/NCAR, GEOS, and ECMWF (acronyms explained in main text). DLFs werecomputed for each reanalysis dataset, with differences reaching values as high as30 Wm^-2 in specific regions, particularly over high altitude areas anddeserts. However, globally, the agreement is good, with the rms of the difference between theDLFs derived from the different reanalysis datasets ranging from 5 to 7 Wm^-2.The results are presented as geographical distributions and as time series ofhemispheric and global averages. The DLF time series based on the different reanalysisdatasets show similar seasonal and inter-annual variations, and similar anomalies relatedto the 86/87 El Niño and 89/90 La Niña events. The global ten-year averageof the DLF was found to be between 342.2 Wm^-2 and 344.3 Wm^-2, depending on thedataset. We also conducted a detailed sensitivity analysis of the calculated DLFs to the keyinput data. Plots are given that can be used to obtain a quick assessment of thesensitivity of the DLF to each of the three key climatic quantities, for specific climaticconditions corresponding to different regions of the globe. Our model downwellingfluxes are validated against available data from ground-based stations distributed overthe globe, as given by the Baseline Surface Radiation Network. There is a negative bias of the model fluxes when compared against BSRN fluxes, ranging from-7 to -9 Wm^-2, mostly caused by low cloud amount differences between the station andsatellite measurements, particularly in cold climates. Finally, we compare our modelresults with those of other deterministic models and general circulation models.

[学科分类] 大气科学