High temperature comparison of the HITRAN2012 and HITEMP2010 water vapor absorption databases to frequency comb measurements

The HITEMP2010 and HITRAN2012 databases are important tools for predicting molecular absorption under various environmental conditions. At room temperature, the databases can be quite accurate, owing to their development using room temperature absorption data. However, at elevated temperatures common to combustion and planetary research, their broadband accuracy is largely unexplored. We utilize a dual frequency comb spectrometer and a high temperature optical cell to assess the capability of these databases to accurately predict water vapor absorption for over 600 transitions at conditions up to 1300 K from 6800 cm−1 to 7200 cm−1. We demonstrate that at 1300 K, HITEMP2010 and HITRAN2012 accurately predict the existence and position (within 0.033 cm−1) of >99% of transitions with an intensity greater than 1E−28 (cm−1/molecule cm−2), which is the approximate minimum detectable intensity for our absorption spectra. Of the slightly less than 1% of transitions that are misreported, all have linestrengths below 1E−23 (cm−1/molecule cm−2). HITEMP more often predicts transitions that are not observed in the absorption spectra, while HITRAN more often fails to predict an observed line. Updated temperature-scaling coefficients for air-broadening and updated line assignments improve the high temperature performance of HITRAN compared to HITEMP for lineshapes and positions. Our analysis also shows that adding simple temperature-scaling relationships for air-pressure shift and self-broadening half width increase the predictive accuracy of simulations using both databases, and should be employed whenever transition-specific information is unavailable.