| |||||||||||||||||||||||||||||||||||
|
|
| ||||||||||||||||||||||||||||||||||
BAND NAME |
SRC |
SRB |
|---|---|---|
| SPECTRA | 125-175nm | 175-200nm |
| DATA | WMO 1985 | Minschwaner LBL model |
| No. of ESFT | 3 | 7 |
| FITTING | Delta log(k) | Gaussian |
| ABSPT Error | ~within 2% | ~within 3% |
Line-by-line (GENLN2) (HITRAN 92) calculations of monochromatic absorption coefficients for the CO2 15micron band have been carried out for 31 pressures (between 1097 and 0.00034 hPa) and 10 temperatures (between 140 and 320 K).
Various methods to fit the k-distribution were tried (Gaussian, Modified Gaussian and Delta log(k)) and it was concluded that the best approach, in terms of a compromise between accuracy and computational demands, was to use 14 c-k coefficients for the band centre (590-750 cm-1) and 17 c-k coefficients for the band wings (500-590 + 750-800 cm-1). An efficient and accurate linear interpolation method is used to find the scaling approximation used to represent the pressure and temperature-dependence of the absorption.
LBL (GENLN2) (HITRAN 92) calculations of monochromatic absorption coefficients for the H2O strong bands (0-550, 1500-1900 cm-1) have been done for 26 pressures (between 1096 and 0.004 hPa) and 10 temperatures (between 140 and 320 K). The E-S interval 1500-3000 cm-1 is split into two intervals (see Table 2). There is no change for the window region and band wings because their contributions to heating rates in the middle atmosphere are negligible.
Band region (cm-1) |
Change (Y/N) |
c-k terms |
|---|---|---|
| 0-350 | YES | 25 |
| 350-550 | YES | 16 |
| 1500-1900 | YES | 14 |
| 1900-3000 | YES | 2 |
Figure 2 shows the cooling rates of LBL(GFDL solid line), this work (dash-dotted line) and E-S (dotted line) for a mid-latitude atmosphere. Significant improvements in the middle atmosphere can be seen, remaining differences near the stratopause are due to the omission of minor CO2 bands and residual problems with the H2O fitting and other approximations.
Figure 3 shows the difference in net heating rate between the revised and former versions of E-S as a function of latitude and height for an atmosphere consisting of January CIRA temperature, UGAMP ozone climatology and GEDEX water vapour. Differences are very small in the troposphere and lower stratosphere but significant (of order 1 K/day) around the stratopause and very large in the upper mesosphere and lower thermosphere.
This improvement is achieved at very little cost in terms of computational requirement. We now intend to implement the scheme in the TSM version of the UKMO Unified Model.
Wenyi Zhong and Joanna Haigh
Imperial College
W.Zhong@ic.ac.uk
![]()