1130 A Comparison of the Seasonal and Long-Term Climatology of TICOSONDE Water Vapor and Ozone Profiles for the Period 2005–17 with MLS Version 4 at Costa Rica (10°N, 84°W)

Wednesday, 10 January 2018
Exhibit Hall 3 (ACC) (Austin, Texas)
Henry B Selkirk, GSFC, Greenbelt, MD; and H. Vömel, M. R. Schoeberl, G. Taha, G. A. Morris, S. M. Davis, A. M. Thompson, and J. C. Witte

The Ticosonde Project is a joint collaboration among NASA, the Universidad de Costa Rica, the Universities Space Research Association, St. Edward’s University, NOAA and NCAR that has been making regular balloon sonde measurements of water vapor and ozone in Costa Rica [84°W, 10°N] since July 2005. Through its first 12 years, Ticosonde has contributed 550 ozone soundings to the SHADOZ database and accumulated 206 water vapor soundings to NDACC. The Ticosonde cryogenic frostpoint hygrometer (CFH) measurements comprise the longest continuous dataset of in situ water vapor in the tropics, and Costa Rica is the only sonde site within the deep tropics (10°S-10°N). These data have been used for validation of satellite instruments including Aura MLS and HIRDLS and have contributed to our understanding of the spatial and temporal variability of UT/LS H2O and ozone, the water vapor tape recorder, saturated layers, and cirrus clouds in the tropics. Here compare climatological profiles from the first twelve years of the Ticosonde water vapor and ozone to MLS v.4 profiles clustered around the Costa Rica site. Figure 1 depicts the seasonal mean profiles of water vapor for the boreal winter (DJF) and summer (JJA) seasons pairing kernel-averaged Ticosonde CFH profiles with contemporaneous local clusters of MLS profiles. Both seasons show a tendency for the MLS to be slightly wetter than CFH in the lower stratosphere and down to 121 hPa and vice-versa below. The standard deviations of the sonde observations are substantially greater in DJF. For ozone, the difference between the Ticosonde data and MLS v.4 displays an oscillatory behavior between 30 hPa and 265 hPa with differences varying back forth from close to zero to +30% (MLS higher.) We discuss these local results for both of these constituents in the context of known MLS biases as well as uncertainties in the sonde data.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner