The Comparison Among Different Radiative Transfer Models And New Radiation Scheme Application Into CAM3 And GRAPES

Atmospheric radiative transfer model can be used in the fields of climate research and remote sensing. The accuracy of radiative transfer model is closely related to the input spectral data and radiative transfer algorithm.This work includes four aspects:1, We use line-by-line model to calculate longwave radiation, and compare the difference among HITRAN96, HITRAN2K and HITRAN04 spectral data on atmospheric radiation calculation of longwave. The results from the calculation of the three versions'data show that are: upward radiation flux difference of less than 0.4W/m2; the downward radiation flux difference of less than 1.6W/m2; the net radiation flux difference of less than 1.6W/m2 and the cooling rate difference of less than 0.16K/day.2, We compare the difference of two-stream DISORT method and Eddington method from simple layer to a realistic atmospheric profile. First, we compare the difference in simple layer. The Eddington scheme is slightly more accurate in thicker optical depth region, while the two-stream DISORT is more accurate in the thin optical depth region where the solar zenith angle is less than 60 degree. Second, we compare the two two-stream schemes in a realistic atmospheric profile with gaseous transmission considered. Both of the two-stream schemes are very accurate in clear sky conditions, with an error in heating rate of less than 0.4K/day. The delta-two-stream DISORT is slightly more accurate than the delta-Eddington approximation in this case. Therefore, we conclude that the two-stream scheme is accurate for climate modeling purposes in clear sky conditions. However, for cloudy sky conditions, the relative error in cloud absorption could be as high as 12%. The incorrect cloud absorption could affect climate model behavior as cloud top heating is crucial for cloud evolution. Therefore, to obtain the accurate results for cloud absorption, a four-stream, or a higher-order stream is needed.3, We add the shortwave radiative transfer model into GRAPES and make a simple comparison. First, we compare the surface temperature. In ocean, EC(Europe Center) and new radiative transfer model are similar with the NCEP reanalysis data. In North America and Oceania, new model is better. In Antarctica, both EC and new model are not well. In Africa and South America, EC and new model are similar. In Europe, EC is better. In east China, new model and EC are rougher than the NCEP reanalysis data. Second, we compare east 120 degrees longitude temperature vertical profile. Below 400hpa and between 30 degrees and 60 degrees, the new model is better. Below 700hpa and in Antarctic, EC is better. Above 400hpa, both the new model and EC can simulate the cold center and EC is better. Also, above 300hpa and in Arctic, EC is better.4, We add the shortwave radiative transfer model into CAM3 and make a simple comparison New radiative transfer model is better in surface downward shortwave radiation flux, surface net shortwave radiation flux, top of atmosphere net shortwave radiation flux, top of atmosphere shortwave cloud force, surface shortwave cloud force and surface latent heat flux.