Analysis of the FASCODE model and its H(2)O continuum based on long-path atmospheric transmission measurements in the...

An experimental study performed to evaluate the atmospheric transmission model FASCODE and its water vapor continuum [Clough, Kneizys, and Davies (CKD) model, Atmos. Res. 23, 229-241 (1989)] in the 850-2250-cm-(1) spectral region is presented. The analysis is based on a comparison between model calculations and transmission measurements carried out at the Defence Research Establishment Valcartier over a 5.7-km horizontal path for a wide range of ambient temperature (from -8.6 to 29.4°C) and humidity (from 1.16 to 14.2 g/m(3)) conditions. The agreement between measurements and calculations is good on the average. However, there are three specific spectral intervals where the differences cannot be explained by experimental errors. For summer conditions, it is shown that FASCODE overestimates the transmittance by approximately 3-6% (absolute terms) in the 850-950-cm(-1) region. For winter conditions, measurements are higher than calculations by as much as a factor of 2 at the edges of the 6.3µm absorption band of water vapor, namely near 1250-1380 cm(-1) and 1800-2000 cm(-1). The continuous nature of these differences is interpreted as anomalies that are due to the broadening coefficients of the water vapor continuum (CKD model). A set of coefficients is derived from experimental spectra and compared with coefficients from the CKD model. The results suggest that first the self-broadening coefficients at high temperature, C¯(s)(ν, 296), need to be increased by 10-16% near 850-950 cm(-1) and second the foreign broadening coefficients, C¯(ν), need to be decreased by approximately a factor of 2 near 1250-1380 cm-(1) and 1800-2000 cm(-1) to recover a good model-measurement agreement in these three spectral intervals. A modified continuum (based on coefficients derived from transmittances) has been implemented in FASCODE and used to analyze emission spectra from the High-Resolution Interferometer Sounder instrument. The modified continuum reduces the discrepancy by almost a factor of 5 near 1350 cm(-1).