Readme file for DISCOVER-AQ 2013 (CA) Porterville ozonesonde launches (Anne Thompson, PI)
Please contact Anne Thompson prior to use of these data
Readme file written by: Ryan Stauffer
Email: rms5539@psu.edu
Please contact Ryan Stauffer or Douglas Martins (dkm18@psu.edu) for additional information or questions regarding these data.

These data include 31 launches using ENSCI/Droplet Measurement Technologies 2Z ECC ozonesondes.
Ozonesondes used 0.5% KI sensing solutions and an attached International Met Systems radiosonde.
Profiles were taken at 1839 S. Newcomb St. Porterville, CA (36.0319, -119.0550, 136m AMSL)
at Penn State's Nittany Atmospheric Trailer and Integrated Validation Experiment (NATIVE).

All data are reported at 1 Hz intervals with missing or suspect data = -99999.

Uncertainties:

Uncertainty estimates are based on previous studies, particularly Komhyr et al. (1995, JGR) who estimated
ozone partial pressure uncertainties between -7 to +17% in the troposphere improving to ±5% in the stratosphere,
and Deshler et al. (2008, JGR) who estimated a nominal uncertainty of ±5% in the troposphere and ±5-10% in the stratosphere.
Ozone partial pressure uncertainties generally increase above 10hPa with decreasing pump efficicency.
Total integrated ozone column uncertainties have been estimated by Smit et al. (2007, JGR) at ±5%.

Pressure Smoothing performed in R0:

Extremely noisy pressure, translating to noisy ozone mixing ratios, potential temperature etc. was noticed in many of the launches and required smoothing.
Pressure data were first log-transformed, then linearly interpolated by every 5 seconds of data to obtain a smooth log pressure profile.
The pressure data were then reverted to exponential form, and all pressure-dependent variables 
including pressure altitude, ozone mixing ratio and potential temperature were recalculated.
This smoothing does not have significant effects on the measurements in whole, but reduces point to point noise observed in version RA of these profiles.
Additional information and an example of this pressure smoothing can be found in Stauffer et al. (2013), in prep for AMT.

Warning about Pressure and Pressure Altitude:

Large differences on the order of 2-3km have frequently been observed between the pressure altitude and GPS altitude (both prior to and post pressure smoothing)
on these flights, as well as many flights from other stations and radiosonde types (see Stauffer et al. (2013), in prep for AMT).
Pressure data using the GPS altitude, temperature and humidity measurements can be calculated and compared with the radiosonde pressure measurements.
Errors in the radiosonde pressure can approach 1-2 hPa by 30 km, leading to errors in ozone mixing ratio of 10% or more.
These effects are negligible in the troposphere and only become significant above 20-25 km, where the ~1 hPa offset represents a significant portion of total air pressure.
Except for the pressure smoothing noted above, nothing further (i.e. GPS pressure recalculation) was done to these data.
This information on pressure sensor errors is given only as a warning.