8.31-P

The Cryogenic Frost point Hygrometer (CFH): A reference instrument for tropospheric and stratospheric water vapor measurements Holger Vömel1, Don David, Ken Smith Cooperative Institute for Research in Environmental Sciences, University of Colorado, Boulder

1. Introduction Accurate measurements of water vapor between the middle troposphere and the middle stratosphere are highly important due to the strong impact of water vapor on our climate system and its important role in atmospheric chemistry. The urgent need for validation of remote sensing observations from space borne or ground based remote sensors requires in situ measurements, which remain a technical challenge. Radiosondes, depending on the manufacturer, generally provide useful data only in the lower and middle troposphere. Only a few techniques currently exist that allow frequent observations on small meteorological balloons and only one has been in use for an extended period of time. The Cryogenic Frost point Hygrometer (CFH), which has been developed at the University of Colorado (Vömel et al., 2006a), is based on the NOAA/CMDL (now Global Monitoring Division of the NOAA/Earth System Science Laboratory) frost point hygrometer (e.g. Vömel et al., 1995), which has been launched at Boulder, CO for 26 years. The CFH has a significantly reduced power consumption and weight, and requires less skill to operate. It can be launched using smaller balloons, which is of great importance for the network of stratospheric and upper tropospheric water vapor sounding sites currently in preparation.

2. Instrument description The instrument is based on the chilled-mirror principle, using a cryogenic liquid, which is able to cool the mirror to between 30°C and 100°C below any ambient frost point temperature. The mirror temperature is controlled by heating against this cold sink, which allows both fast heating and fast cooling rates. This cooling capacity avoids the low humidity limits faced by Peltier-cooled frost point instruments (e.g. Vömel et al., 2003). The condensate detector was implemented using a phase-sensitive detection scheme, eliminating solar contributions to the photodiode signal. This allows daytime use with an open flow system, reducing the risk of contamination in low humidity regions, in particular in the stratosphere. For clean sampling of ambient air the CFH uses stainless steel inlet tubes at the top and bottom, which extend well beyond the instrument in both directions. The thermistor imbedded in the mirror is individually calibrated to a NIST traceable standard over the temperature range +25°C to -100°C. The uncertainty of the mirror temperature measurement due to the calibration standard, the calibration setup and the calibration fit curve that is applied in the measurements is less then 0.05°C over the entire relevant temperature range. The liquid/ice uncertainty, which is inherent in all chilled mirror hygrometers, is avoided by force freezing the mirror condensate at a predetermined temperature. The feedback controller uncertainty is the largest contribution to the water vapor measurement accuracy and is estimated to be less then 0.5°C over the entire temperature range. This translates to an uncertainty in the water vapor partial pressure between 3% at warm temperatures (tropical surface) and about 9% at cold temperatures (tropical tropopause region). The Ticosonde 2005 campaign in July 2005 1

Holger Vömel Cooperative Institute for Research in Environmental Sciences University of Colorado Campus Box 216 Boulder, CO 80309 [email protected] ph.: (303) 497 6192 fax: (303) 497 5590

launched 24 CFH instruments, of which 9 reached above 25 km. The variability of these measurements, which combines both instrumental and atmospheric variability, lies between 11% at the top of the tropical tropopause layer and 6% at 25 km. A second campaign at Biak, Indonesia in January 2006, reproduced this result with 6 soundings. In several tropical soundings a small number of sondes have failed after being launched into precipitating or very wet clouds, caused by wetting of the detector lens, either through liquid water deposition or in situ condensation on the unheated lens.

3. Observations The CFH has participated in a significant number of scientific and comparison campaigns. The comparison between the CFH and the older NOAA/CMDL hygrometer shows a very good agreement in the stratosphere and upper troposphere, indicating that this instrument can be used to continue trend observations of stratospheric water vapor at Boulder, CO and other locations. Comparisons with the Vaisala RS92 radiosonde during three extensive field experiments at the Oklahoma ARM/CART site (Miloshevich et al., 2005), at Sodankylä, Finland, and at Alajuela, Costa Rica show a good agreement down to -60°C in night time observations. The tropical campaign was also able to identify a slight wet bias of the RS92 of up to 13% at -70°C, which is most likely related to the early termination of the heating cycle. This campaign also conducted extensive daytime comparisons between the RS92 and the CFH and found a very large dry bias, which ranges from 9% at the surface to in excess of 50% at 15 km. Indications of this dry bias had been found in the microwave scaling of radiosondes at the ARM sites, but the vertical profile of this dry bias had not been previously determined. Since this dry bias is highly reproducible, a correction scheme was developed (Vömel et al., 2006b). Simultaneous observations of water vapor using the CFH and of cirrus clouds using a backscatter lidar provide important information about the dehydration processes in the tropical tropopause region. One such observation shows that regions of high supersaturation may contain ice clouds, therefore actively contributing to dehydration in the tropical tropopause layer, but also that large values of supersaturation may be achieved without the formation of ice particles (Shibata et al., 2006). The comparison between the AURA/MLS water vapor observations and the CFH at several sites including high-, mid-, and low- latitudes, show a very good agreement between MLS stratospheric water vapor and CFH measurements. Only tropical observations during the boreal winter do not adequately reproduce the ‘tape-recorder’ signature of the vertical water vapor profile, with both high and low biases depending on the retrieval level. Upper tropospheric observations show a significant dry bias with is largest at the lowest retrieval level at 316 hPa. The variability of upper tropospheric measurements by MLS is significantly larger than that by CFH during the extensive campaigns at Costa Rica during July 2005 and January through March 2006. The comparison of CFH water vapor measurements and Tunable Diode Laser (TDL) observations on small balloon payloads at Costa Rica in January 2006 show good agreement throughout the troposphere and reasonable agreement in the tropical tropopause region and up to 20km. The TDL observations show a larger variability in the lower stratosphere compared to the CFH. The comparison of these observations with those onboard the NASA- WB-57F high altitude research aircraft show significantly higher values in the aircraft observations. This difference has not yet been resolved.

References Vömel, H., D. David, and K. Smith, Accuracy of tropospheric and stratospheric water vapor measurements by the Cryogenic Frost point Hygrometer (CFH): Instrumental details and observations, J. Geophys. Res., submitted, 2006a. Vömel, H., H. Selkirk, L. Miloshevich, J. Valverde, J. Valdés, E. Kyrö, R.Kivi, W. Stolz, G. Peng, and J. A. Diaz, Radiation dry bias of the Vaisala RS92 humidity sensor, J. Atmos. Oceanic Technol., submitted, 2006b. Shibata, T., H. Vömel, S. Hamdi, S. Kaloka, F. Hasebe, M. Fujiwara, and M. Shiotani, Cirrus clouds in tropical tropopause layer observed under supersaturated condition, J. Geophys. Res., submitted, 2006. Miloshevich, L. M., H. Vömel, D. N. Whiteman, B. M. Lesht, F. J. Schmidlin, and F. Russo (2006), Absolute accuracy of water vapor measurements from six operational radiosonde types launched during AWEX-G and implications for AIRS validation, J. Geophys. Res., 111, D09S10, doi:10.1029/2005JD006083. Vömel, H., M. Fujiwara, M. Shiotani, F. Hasebe, S. J. Oltmans, and J. E. Barnes, The behavior of the Snow White chilled-mirror hygrometer in extremely dry conditions, J. Atmos. Oceanic Technol, 20, 1560-1567, 2003. Vömel, H., S. J. Oltmans, D. J. Hofmann, T. Deshler, and J. M. Rosen, The evolution of the dehydration in the Antarctic stratospheric vortex, J. Geophys. Res., 100, 13919-13926, 1995.

Appendix Figure 1:

Cryogen

Frost coverage Thermistor Detector lens

Heater coil

Detector IR LED Mirror

µ Controller

Air flow

Schematic of the CFH sonde. The dashed lines indicate the vertical inlet tubes. The air flow inside the tubes is downward during balloon ascent upward during descent. Only the lens and the mirror are exposed to the airflow inside the tube. The microprocessor controller regulates the mirror temperature such that the bulk reflectivity of the frost covered mirror remains constant.

Figure 2:

Comparison of the Vaisala RS92 and the CFH in 24 soundings at Alajuela, Costa Rica during July 2005. Left panel: The nighttime comparison based on 4 dual soundings reveals a slight altitude dependent dry bias below 13 km (-55°C) that changes to a wet bias a colder temperatures. Data above the tropopause (about 16 km) are meaningless. These results are consistent with previous observations. Right panel: The daytime comparison based on 19 dual soundings reveals a dramatic altitude dependent dry bias, which ranges between 9% at the surface and 50% at 15 km.

Figure 3:

Comparison of CFH and AURA/MLS water vapor measurements at Alajuela Costa Rica in July 2005. Left panel: CFH water vapor profiles between 8 July and 17 August 2006 and all MLS profiles in this region (±5° latitude, ±20° longitude). The tropopause is at around 100 hPa. Right panel: Difference between the MLS and CFH. The matched pair comparison (solid red) is based on coincident measurements less than 6 hours and less then 300 km apart, whereas the comparison of the average profiles (solid blue) considers all profiles during the observation period. Both difference calculations show the same results: Above 100 hPa there is good agreement between both measurements, in the upper troposphere MLS shows a significant dry bias.

Figure 4:

Similar comparison of CFH and AURA/MLS water vapor measurements at Costa Rica in January through March 2006. Left panel: same figure 3. The tropopause is at around 80 hPa. Right panel: Same Figure 3. Here both difference calculations show on alternating pattern of wet bias and dry bias in the stratosphere, which is a result of the inadequate representation of the ‘tape-recorder’ structure of the seasonal profile of water vapor.