As with the description of the aircraft and its instrumentation, there is no place in this document for a detailed discussion of the calibration techniques or the intricacies of the steps used to obtain the final processed data set from the raw data set. This information is fairly standard for most experiments utilising the Cessna, and so its place is in a separate system description of the aircraft. Such a description is hoped to be written within the next year. A summary of those points directly relevant to the TOGA COARE experiment is given below.
Calibrations of most instruments were performed via in-flight maneouvres prior to the TOGA COARE IOP. However, final verifications of the accuracy and biases of many of the basic and derived quantities will only become available after detailed analysis of the data from the intercomparisons with other platforms.
Calibration of the fast temperature sensors was performed against the stable (but slower) Meteolab thermocouple. This is a fairly complex procedure which must be completed by in-flight "racetrack" maneouvres, in order to ascertain the dynamic pressure correction required to reduce the total temperature down to the ambient air temperature. This correction is different for each instrument, but should remain steady provided the location and housing of each instrument does not change, and that the aircraft always samples within a fairly small range of permissible airspeeds.
Calibration of the fast Lyman-alpha hygrometer has been performed against the stable (but slow) Meteolab dewpoint mirror. This was done separately from ascent/descent data of each flight, because the Lyman-alpha sensors are known to exhibit a low frequency "drift" in the calibration due to changes in the transparency of the glass "windows" (due to misting, soiling etc.). This drift appears to affect only the offset value of the calibra- tion (the slope of the curve is relatively stable), so that instantaneous deviations from the mean (required for eddy correlation) are unaffected. Calibration of the ATD humidity sensor was performed in the same manner, but does not appear to suffer from any such problems.
The calibration of the nosecone five-hole pressure-port system and the dynamic pressure port has been ascertained/checked via "racetrack", "wind-square", "reverse heading" and "pitch/roll/heading-modulating" maneouvres. Post-experiment comparisons with a windtunnel-calibrated 3m boom attached under the left wing of the Cessna, have verified the accuracy of the nosecone calibrations.
Accurate calibration of the radiometers via a sun-disk shading technique was performed at FIAMS prior to TOGA COARE using a Linke-Feussner actinometer. Basic Processing
Below is a short summary of the order of processing for the data. In Appendix A2 can be found the standardised command-files for processing of EOS flights using FIAMS' software package, RAMF. These command-files serve as detailed self-documenting records of all steps taken in the processing.
 | Decoding of the time-tagged raw binary stream data files into a number of FORTRAN unformatted binary files, each containing several columns of voltage data from one of the main data systems of the aircraft. Each of these files includes its own time channel, which varies due to differing data sampling rates between the various systems. At this stage, the raw fast data from each flight is segmented into "runs" as detailed in Chapter 4. |
| First "look" at the raw voltage channels to ascertain data quality, locate problems (eg. spikes) and check that the segmentation was performed correctly. |
| Obtain calibration coefficients for the Lyman-alpha for the current flight. |
| Application of the known calibrations to the basic (slow) meteorological channels, and removal of spurious data (eg. "spikes" caused by radio communications) where necessary. Concurrent calculation of the true air speed and speed-corrected thermocouple temperature via the method of Carl Friehe (pers. comm.). |
| Application of the known calibrations to the eddy-correlation channels (fast temperature and humidity sensors, nosecone pressures) with removal of spurious data, speed corrections to fast temperature sensors, and calculation of angles of attack and sideslip from nosecone pressures. |
| Align (transform) GPS data via linear interpolation onto the time axis of the meteorological data (20Hz for "fast" data, 1Hz for "slow" data). Convert GPS earth-centered data into latitude/longitude/u/v/w. |
| Compute aircraft velocities in the earth system from the INS accelerations, velocities and attitude-angles, and align onto the time axis of the meteorological data. |
| Application of the known calibrations to the radiometer voltages, and perform corrections according to the sun angle and the aircraft attitude angles. |
| Calculation of the 3-dimensional wind vector from aircraft velocities and relative wind vector (from true air speed, alpha and beta). Calculation of other "derived" meteorological quantities (density, potential temperature, specific humidity etc.). |
| Transformation of all processed channels onto a common distance axis (spacing: 3m for "fast" data, 50m for "slow" data) via true air speed. Computation of fluxes and other statistics of runs and segments of runs. |
The processed time series are stored in a file called "flightX.run",
where X is the Flight ID and run is the name of the run, as given in the tables of Chapter
4. These files may be read as described in Chapter 6 and Appendix A1; their contents are
set out below:
Parameter Name Description temp Air temperature (°C) altrad Radar altitude (m) altpres Pressure altitude (m) lat Latitude (decimal degrees) lon Longitude (decimal degrees) static Static pressure (hPa) tas True air speed (m/s) rho Air density (kg/m3) uair East/West wind component (m/s) vair North/South wind component (m/s) wp Vertical wind component (m/s) theta Potential temperature (C) sphum Specific humidity (g/kg) co2 CO2 concentration (g/kg)* tempsf Radiometric surface temperature (°C) ndvi Normalised-difference vegetation index* Sin Incoming short-wave radiation (W/m2) Sout Outgoing short-wave radiation (W/m2) Lin Incoming long-wave radiation (W/m2) Lout Outgoing long-wave radiation (W/m2) RNet Net radiation (W/m2) *Note: co2 and ndvi are from instrumentation under development, and cannot be guaranteed.
Statistics formed from the segmented time series will also be available with the TOGA COARE time series data set. These will comprise averages, standard deviations, minima and maxima of the above physical parameters, and also the following additional derived statistics: average wind speed and direction, fluxes of sensible heat, latent heat and momentum.
To table of contents
Next
chapter