A new spectroscopic database for water in the range 1250-1750
cm-1 was generated. All data were derived from measurements carried out
This is a beta version. Updates regarding line positions
and linestrengths from quantum mechanical fits will replace current entries.
Since the Voigt line profile was found to be inadequate under certain
conditions, it is planned to reanalyze the data with a more sophisticated
lineshape model (may take a while). Care should be taken when modelling
opaque atmospheric lines since they especially show the deviations from
the Lorentzian. In case of using a low resolution ("spectrally integrating")
instrument the observed radiance may be slightly wrong. This is especially
the case for Lorentz width range 0.0004 to 0.004 cm-1.
Line positions from 9 ambient temperature low pressure (<5mb)
pure water measurements were averaged on a single line basis.
Linestrengths from 9 ambient temperature pure water measurements
and 14 ambient temperature air-broadened measurements were averaged on
a single line basis.
Air and self broadening parameters and temperature dependence
were fitted from 33 air-broadened measurements on a single line basis.
In cases where not all 4 parameters (gamma_air, n_air, gamma_self, n_self)
could be fitted values from HITRAN2004 were used in the fit. This was
accounted for in the error budget.
In order to include an appropriate statistical error bar at temperatures different from 296K the fit was
carried out for two reference temperatures (296K and 220K) yielding statistical
error bars for the broadening parameters at 296K and 220K, repectively.
Uncertainties for intermediate temperatures can be approximated by interpolation.
(Propagating uncertainties from gamma(296K) and the temperature exponent
gives wrong numbers due to the high off-diagonal elements in the covariance-variance
matrix of the fit, which are not included when considering statistical
uncertainties of gamma(296K) and the temperature exponent only).
Error bars are not symmetric due to propagation of systematic errors and thus
positive and negative error bars are given.
Pressure-induced lineshifts and temperature dependence were
fitted on a single line basis from 43 air-broadened measurements and the
low pressure limit reference line positions mentioned above.
Self shifts were found to be negligible within the experimental accuracy.
The temperature dependence of the shift is defined in the following equation:
where delta_p is the temperature dependence and delta_0 the shift parameter
at 296K, p is given in atm (1 atm=1013.25 mb).
For similar reasons as explained for the broadening parameters the uncertainties of the lineshift
parameters are given at two reference temperatures (296 and 220K), and
positive and negative errors are supplied.
The HITRAN2004 format was not sufficient for representing
the data and had to be extended. In order to show significant changes
to HITRAN2004 differences to all HITRAN2004 parameters are given.
The database was generated from HITRAN2004 by either replacing
existing entries (line position, linestrength, air broadening parameter,
self broadening parameter, air broadening temperature exponent, pressure
induced lineshift) or adding entries after the original HITRAN2004 format
(self broadening temperature exponent, pressure induced lineshift temperature
dependence, flags indicating that data in the line have been updated,
uncertainties for parameters, differences to HITRAN2004), whereever data
from our work was available. Parameters not updated are left unchanged.
Columns after the regular HITRAN2004 format (160 columns):
1-6: Temperature exponent for self broadening (0 when not determined)
9-19: Temperature dependence of pressure induced lineshift in cm-1/(atm K) (0 when not determined)
The following entries are only present when the respective parameter of
the respective line has been updated. If not updated the corresponding
space is left blank. If all spaces in the following columns are blank
no parameters have been updated and the original HITRAN2004 line is given.
21: Flag V indicates that line position has been updated
23-30: uncertainty of line position in cm-1
32-40: difference HITRAN2004 - updated line position in cm-1
42: Flag S indicates that linestrength has been updated
45-47: Uncertainty of linestrength in %
49-53: difference HITRAN2004 - updated linestrength in % of updated linestrength
55: Flag P indicates that air broadening parameter has been updated
56-59: positive air broadening parameter uncertainty at 296K in %
61-64: negative air broadening parameter uncertainty at 296K in %
65-69: positive air broadening parameter uncertainty at 220K in %
70-74: negative air broadening parameter uncertainty at 220K in %
76-81: difference HITRAN2004 - updated air broadening parameter at 296K in % of updated air broadening parameter 83-87: difference HITRAN2004
- updated air broadening parameter at 220K in % of updated air broadening parameter
89: Flag p (perturbed) indicates that at least one of the measurements included in the fit of air broadening had a self broadening contribution
larger then 2% and that the self broadening parameter could not be fitted (and was thus taken from HITRAN2004). Flag u (unperturbed) indicates that
only measurements with negligible unfitted self broadening contribution were in the fit
90: Flag b indicates a "best of" line at 296K. A line is "best of" when flag u in column 89 is set, n_air was fitted, the total
uncertainty of the air broadening parameter is better than 5%, chi from the fit is better than 2, and at least 4 air-broadened measurements have
contributed to the fit.
91: Flag b indicates a "best of" line at 220K. A line is "best of" when flag u in column 89 is set, n_air was fitted,
the total uncertainty of the air broadening parameter is better than 5%, chi from the fit is better than 2, and at least 4 air-broadened measurements
have contributed to the fit.
93-95: number of foreign broadened linewidths in fit
97-101: chi-squared of fit
102: Flag n indicates that the temperature exponent for air broadening has been determined and updated. When an error
for 220K is given and this flag is not set the temperature exponent from HITRAN2004 was used in the fit.
104: Flag G indicates that a self broadening parameter at 296K has been determined
105-108: positive self broadening parameter uncertainty in % 109-113: negative self broadening parameter
uncertainty in %
115-120: difference HITRAN2004 - updated self broadening parameter at 296K in % of updated self broadening parameter
122: Flag D indicates a pressure-induced lineshift at 296K has been determined
124-133: positive pressure-induced lineshift parameter uncertainty at 296K in cm-1/atm
133-140: negative pressure-induced lineshift parameter uncertainty at 296K in cm-1/atm
142-149: positive pressure-induced lineshift parameter uncertainty at 220K in cm-1/atm
151-158: negative pressure-induced lineshift parameter uncertainty at 220K in cm-1/atm
160-168: difference HITRAN2004 - updated pressure-induced lineshift parameter at 296K in cm-1/atm