HITRAN Spectroscopy Evaluation using Solar Occultation FTIR Spectra Geoffrey C. Toon 1, Jean-Francois Blavier 1, Keeyoon Sung 1, Laurence S. Rothman 2, Iouli Gordon 2 1 Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA 2 Harvard-Smithsonian Center for Astrophysics, Cambridge, MA The HITRAN 2012 linelist contained several preventable errors. Late delivery of key linelist components left too little time for a thorough validation, and too little time to remedy the problems that were found. We have tried to automate part of the validation process by fitting atmospheric spectra measured by the MkIV balloon interferometer. By comparing rms fitting residuals with previous HITRAN linelists, errors are quickly identified. This work, applied to HITRAN 2016, will hopefully reduce the number of preventable errors. Results also provides guidance for future spectroscopic work. Copyright 2016 Caltech. Government sponsorship acknowledged

Approach Defined 113 contiguous windows covering 700 to 5600 cm -1 Fitted an occultation of MkIV balloon spectra (57 cm OPD) with 20 different tangent altitudes from 9 km to 38 km Large range of T/P/vmr seen in balloon spectra exposes spectroscopic errors Used four different linelists: HITRAN 2000, 2004, 2008, 2012 So a total of 113 x 20 x 4 = 9,040 spectral fits were performed with GFIT Compared rms fitting residuals from HITRAN 2012 linelist with its predecessors Investigate cases where HITRAN 2012 rms residual is not the best, or is too large Mainly tests HITRAN line positions, E”, and relative intensities (and completeness)

MkIV Balloon Spectra From May 1997 balloon flight from Fairbanks, Alaska, part of POLARIS campaign. The NASA ER-2 aircraft made dozens of flights over Alaska. There were also frequent ground-based remote-sensing observations. Thus, the atmosphere over Alaska was exceptionally well characterized during POLARIS. MkIV spectra balloon cover cm -1 at 0.01 cm -1 resolution (57 cm MOPD) Spectra were ratioed against a high-sun solar spectrum to remove solar and instrumental features. Observed residuals are therefore predominantly telluric. ILS was determined by fits to high altitude (38 km) spectra of CO 2. In any case, the same ILS is used for all linelists, and so does not cause residuals to be different

RMS fitting residuals at 31 km At 31 km, the smallest rms residuals occur in window regions, e.g cm -1, with weak absorptions. Wavenumber /cm -1

RMS Fitting Residuals at 10 km Altitude At 10 km, the smallest rms residuals occur in saturated regions, e.g cm -1.

RMS Residuals versus altitude (1/2)

RMS Residuals versus Altitude (2/2)

Fit to cm -1 window at 20 km altitude This is the worst window for HITRAN All 4 HITRAN linelists produce similar residuals, peaking at 20 km altitude. This and the 0.4 cm -1 spacing of residuals indicate HNO 3 as the culprit. Wavenumber /cm -1

Fit to cm -1 window at 31 km altitude HITRAN 2012 worse than 2008 due to a cm -1 position error on CH 4 line at cm -1 Residual at cm -1 due to cm -1 position error in resonance O 3 line, a perennial problem Wavenumber /cm -1

Fit to cm -1 window at 31 km altitude HITRAN 2012 fits are worse than those using HITRAN 2004 or 2008 due to a worsening of the line positions, esp. in ν 3 +v 4 Q-branch with peak residuals of 6% versus 4% with HITRAN 2004/8 Wavenumber /cm -1

Fit to cm -1 window at 23 km altitude Wavenumber /cm -1 RMS has reduced with each successive HITRAN version (improved CO 2 ), but residuals at 20 km are now dominated by the missing v 1 HNO 3 band, whose P- and Q-branches are clearly visible.

Fit to cm -1 window at 10 km altitude HITRAN 2012 is better than HITRAN 2004/8 despite the huge residual at cm -1, due to a missing CH 4 line which was present in HITRAN HITRAN 2000 is still best at 10 km

Fit to cm -1 window at 12 km altitude Large positive residual near cm -1, due to a factor ~4 intensity over-estimate of H 2 O line at cm -1, was not present in fits using HITRAN 2004 or 2008.

Average RMS Fitting Residual Table (1/4) Center Width Fitted gases HIT2000 HIT2004 HIT2008 HIT h2o co2 o3 hcn h2o co2 o3 clno3 hcn h2o co2 o3 clno3 hcn n2o5 hno h2o co2 o3 clno3 hcn n2o5 hno3 ccl h2o co2 o3 clno3 hcn hno3 ccl4 cof2 chclf h2o co2 o3 clno3 hno3 ccl4 chclf2 ho2no h2o co2 o3 clno3 hno3 ccl4 chclf h2o co2 o3 hno3 ccl3f cocl2 ocs h2o co2 hno3 ccl2f2 ocs h2o co2 hno3 ccl2f h2o co2 o3 sf h2o co2 o h2o co2 o h2o co2 o3 ccl2f2 ccl3f chclf2 hcooh ch h2o o3 n2o ccl2f2 chclf2 ch4 chf h2o o3 n2o ch4 hno3 hdo h2o co2 o3 n2o ch4 hno3 hdo n2o5 cof h2o co2 n2o ch4 hno3 hdo clno3 cf h2o co2 n2o ch4 hno3 hdo chclf h2o co2 n2o ch4 hdo h2o co2 ch4 hdo h2o ch4 hdo h2o ch4 hdo h2o ch4 hdo h2o ch4 hdo h2o ch4 hdo h2o ch4 hdo h2o ch4 hdo h2o ch4 hdo h2o ch4 hdo h2o ch4 hdo no

Average RMS Fitting Residuals Table (2/4) Center Width Fitted gases HIT2000 HIT2004 HIT2008 HIT h2o hno3 ch4 n2o o3 clno3 hdo h2o hno3 ch4 o3 clno3 hdo h2o hno3 ch4 n2o o3 clno3 hdo h2o hno3 ch4 n2o o3 clno3 hdo h2o hno3 ch4 n2o o3 clno3 hdo h2o ch4 n2o o3 clno3 hdo h2o ch4 n2o o3 clno3 hdo h2o ch4 n2o o3 clno3 hdo h2o co2 ch4 n2o o3 hdo no h2o co2 ch4 n2o o3 hdo no h2o co2 ch4 n2o o3 hdo no h2o co2 n2o o3 no h2o co2 n2o o3 no cof h2o co2 o3 no cof h2o co2 o3 ocs h2o co2 o3 co ocs h2o co2 o3 co ocs h2o co2 o3 co n2o ocs h2o co2 o3 co n2o h2o co2 o3 co n2o h2o co2 o3 n2o h2o co2 ch h2o co2 o3 n2o ch h2o co2 o3 n2o ch h2o co2 o3 n2o ch h2o co2 o3 n2o ch h2o co2 o3 n2o ch h2o co2 o3 n2o ch h2o co2 o3 n2o ch

Average RMS Fitting Residuals Table (3/4) Center Width Fitted gases HIT2000 HIT2004 HIT2008 HIT h2o co2 o3 n2o ch4 hdo hcl h2o o3 n2o ch4 hdo hcl h2o o3 n2o ch4 hdo hcl h2o o3 n2o ch4 hdo no2 hcl ocs h2o o3 ch4 hdo hcl c2h h2o o3 ch4 hcl c2h h2o o3 ch h2o co2 o3 ch h2o co2 o3 ch4 hcn h2o co2 n2o hcn h2o co2 o3 n2o h2o co2 n2o hno3 hdo h2o co2 o3 n2o hno3 hdo h2o co2 o3 n2o hf hdo h2o co2 o3 n2o hf ch4 hdo h2o co2 o3 n2o ch4 hdo h2o co2 o3 n2o hf ch4 hdo h2o co2 o3 n2o ch4 hdo h2o co2 o3 n2o ch4 hdo h2o co2 o3 n2o ch4 hdo h2o co2 o3 n2o hf ch4 hdo h2o co2 o3 n2o ch4 hdo h2o ch4 hdo h2o n2o hf ch4 hdo h2o co2 o3 ch4 hdo

Average RMS Fitting Residuals Table (4/4) Center Width Fitted gases HIT2000 HIT2004 HIT2008 HIT h2o co2 o3 ch4 hf h2o o3 n2o ch4 hf h2o o3 ch4 co hf h2o o3 ch4 co h2o o3 ch4 co h2o n2o ch h2o ch h2o co2 ch h2o co2 o3 n2o ch h2o co2 n2o ch h2o co h2o co h2o co2 o h2o co2 n2o h2o co2 n2o ch4 hdo h2o co2 n2o ch4 hdo h2o co2 ch h2o co2 ch h2o co2 ch h2o co2 ch h2o ch h2o ch h2o ch

Summary/Conclusions Comparing rms fitting residuals (new linelist versus previous linelists) provides a quick method of validation and for identifying problems. The HITRAN 2000 linelist produces the best fits in 2 cases; the 2004 linelist in 25 cases; the 2008 linelist in 20 cases, and the 2012 linelist in 65 cases. Atmospheric spectra are used only to assess the HITRAN linelist and to identify problems. To actually fix the problems requires laboratory spectra. Results also show regions/gases for which further spectroscopic work would be beneficial

Future Work Analyze lower altitude spectra where H 2 O plays a much more important role and where errors in line widths and shifts become more important. This will require inclusion of line mixing effects which above 10 km are small. Sub-dividing some of the windows, such that we would end up with more the 113 windows with a narrower average width, would provide more specificity in the identification of regions where the rms fits worsen. For example, in a wide window improvements in the spectroscopy of gas A can mask a worsening of gas B. Examine the band-to-band consistency of the retrieved gas amounts. Evaluate HITRAN 2016 when it comes out later this year. This can now be done quickly (a few days) since the windows and spectra are already defined and the scripts for running the GFIT code and making the plots already exist.

ATM 2016 linelist ATM 2016 linelist is a “greatest hits” compilation based on HITRAN, but including empirical fixes of obvious problems. It is used by the TCCON network. Previously described paper did not include ATM 2016 because it is not documented in a publication and it would not be a fair comparison (ATM is 4 years later) When HITRAN 2012 linelist came out, ATM linelist was updated for gases/region that were improved. In regions that didn’t improve, no update. So it should be no surprise that in terms of the rms residuals ATM2016 linelist is as good as the best HITRAN linelist in nearly every window.

Comparisons Including ATM 2016 Linelist

Average RMS Fit versus Altitude

RMS fits at 31 km altitude

RMS fits at 10 km Altitude

Acknowledgements NASA's Upper Atmosphere Research Program (UARP) who funded this work. The Columbia Scientific Balloon Facility (CSBF) who launched the balloon and recovered the MkIV payload. The HITRAN team gratefully acknowledges support from the NASA AURA program Grant NNX14AI55G. Worldwide spectroscopy community whose work is encapsulated in the HITRAN linelists.