Project Managers Data Quality Report

START08 PM Notes

Project Manager report addendum, START-08 project

Three areas of the dataset are changing in this data release: dewpoint measurement and dependent variables; 2D-C concentrations, histograms and related variables; solar zenith angle variables. These three areas are:

1. The dewpoint measurement changes by approximately 2.9%, increasing the absolute value of the measured dewpoint temperature. This change is due to a calibration of Buck dewpoint sensors using a reference dewpoint generation system performed at EOL by T. Campos in 08/2008. The calibration does not encompass the lower end of the dewpoint range typically encountered in the stratosphere, ending at -40C. The two dewpointers did not calibrate exactly the same, either. However, there is a consistent slope of 1.025 and 1.035 demonstrated by both units that averages at 2.9% and is reflected in the new dataset. The actual calibration equation used in this release is 1.02906*x-0.0457 for both DPRS and DPLS. Note that THETA* calculations did not change in this release; this is because the current temperature calculations used at RAF are not to be used in production releases of these data. If users require precise THETA measurements they should calculate them themselves.

2. 2D-C concentrations and histograms were changed after a bug was found in processing code that caused underestimation of particle concentrations. Also, the first two bins are not output in this release because of unresolved ambiguity in determining the sample volume for small particles (<62 um). These two changes resulted typically in increase of 2D-C concentrations on the order of 5-25% depending on the number of small particles; on some occasions, the total 2D-C concentrations may decrease due to exclusion of small particles from the updated dataset.

3. Solar zenith angle and related calculations were added to this release per PI request.

Additionally, the depth of field parameter was set to the factory specification and is now being output in the netCDF header.

Thank you,


Variable list


Static pressure is available using two different systems: Research and Avionics.

Research static pressure is measured with a Paroscientific (MODEL 1000) with a stated accuracy of 0.01% of full scale. This measurement is output in the netCDF files as:

  • PSF (measured): static pressure as measured using the fuselage holes
  • PSX (reference): same as PSF. Used to choose reference variable if more than one instrument provides measurement of the same parameter.
  • PSFC (measured): static pressure corrected for airflow effects (pcor)
  • PSXC (reference): same as PSFC. Used to choose reference variable if more than one instrument provides measurement of the same parameter.

Use PSXC for the normal measure of pressure (e.g., in equation of state or hydrostatic equation).

Avionics static pressure is recorded from the GV avionics. This is slower than the Paroscientific measurement, but it has been corrected for airflow effects and it is certified for 'Reduced vertical separation minimum' (RVSM) through the calculation of pressure altitude.  RAF has no documentation on how Gulfstream and Honeywell corrected this pressure measurement, but the measurement has passed very strict FAA certification requirements.

Temperature was measured using four different sensors on the GV:

An unheated Rosemount sensor was used for fast-response measurements.  This sensor can be affected by icing, but that did not appear to be a problem in START08.  Two heated Harco sensors were used to give a slower response temperature, that would also be adequate in icing conditions.  A fourth measurement of temperature (slow and with some delay) was provided by the GV avionics instrumentation.

The Rosemount and Harco measurements were logged using serial channels and are affected by a variable recovery factor.  As a consequence, RAF recommends using the reference temperature, ATX (See below.) for all uses of the START08 data set:


Total air temperature from fast Rosemount sensor


Total air temperature from the heated HARCO sensor # 1


Total air temperature from the heated HARCO sensor # 2


Total air temperature from the avionics system


Ambient air temperature from the Rosemount system


Ambient temperature from the heated HARCO sensor # 1


Ambient temperature from the heated HARCO sensor # 2


Ambient temperature from the avionics system


Ambient temperature reference. This is usually the same as ATFR but can also be replaced by another sensor output if ATFR experiences a problem on a particular flight. In such case a flight-by-flight report will note the change.

RAF recommends using ATX for the temperature in thermodynamic equations, etc.

For START08, processing of the temperature data was accomplished using a 1 sps AD temperature data, which is a unique case. In the past, the AD temperatures were never recorded; in the future, they will be recorded and processed at a higher data rate.

Dewpoint temperature and vapor density:
Humidity was measured using two Buck Research 1011C cooled-mirror hygrometers that are normally used for measuring tropospheric humidity.  They have a sandwich of three Peltier elements to cool the mirror, and in comparison to earlier generations of cooled-mirror hygrometers, they have a much-improved capability to measure at low temperatures.  These sensors are assumed to measure dewpoint above 0°C and frostpoint below 0°C.  The instrument has a quoted accuracy of 0.1 °C over the -75 to +50 °C; however, based on examination of the measurements RAF is not comfortable with accuracies better 0.5 °C for dewpoint and 1 °C for frostpoint.  The cooled-mirror sensors are slow, in particular at lower temperatures, and this may give considerable differences between the measurements from the two units or when comparing with faster instruments.  Their cooling rates depend in part on the airflow through the sensor, and this may depend on the angle of the external stub relative to the airflow.  The angle may differ between the two sensors, and this may contribute to response-time differences between the sensors.  At very low temperatures the sensors may jump ("rail") to even lower temperatures.  The cooled-mirror temperatures are included even when they are outside the sensor operating range; this is caused by the need to use values of water vapor in other calculations (e.g., true airspeed). However, the impact of these out of bounds conditions on derived calculations that depend on humidity correction is very small since they occur at extremely low dew points.

The chilled mirror sensors are sensitive to flooding on rapid descents of the aircraft into the humid boundary layer. This results in temporary loss of the instruments ability to measure the dewpoint, which may last from 3 to 15 minutes, depending on conditions. This problem can also be seen in the form of "ringing", or a decaying sinusoidal oscillation of the signal, that appears after altitude changes, especially those following a period of cold soaking at high altitudes. During these periods it is advised to compare the data from both chilled mirrors and choose the one that recovers faster.

Humidity was also measured using a MayComm Open-path Laser Hygrometer.  This dual-channel hygrometer detects optical absorption of water vapor at 1.37 µm.  The sensor has an estimated accuracy of 5-10% of ambient specific humidity (in ppbv).  The sensor has two spectral channels that are used to determine high and low values of humidity, and they are combined to give a single value of humidity. (See below.)

A new instrument, VCSEL Hygrometer, was deployed on START08. VCSEL hygrometer will measure atmospheric moisture content throughout the troposphere and lower stratosphere using high sensitivity optical absorption methods, using a new, near-infrared, vertical cavity surface emitting laser (VCSEL) at 1854 nm. In conjunction with a compact, multipass, open air cell and digital signal processor (DSP) electronics, this sensor consumes very low power (< 5 W), is lightweight (< 2 kg excluding the inlet housing), and occupies only the space within an aperture plate. The use of the 1854 nm VCSEL allows for a limit of detection of <1 ppmv, a precision of 3% or 0.05 ppmv max, and a minimum sampling frequency of up to 25 Hz.


Dewpoint/frostpoint for left fuselage cooled-mirror sensor


Dewpoint for left cooled-mirror sensor


Dewpoint/frostpoint for right cooled-mirror sensor


Dewpoint for right cooled-mirror sensor


Dewpoint, from either right or left cooled-mirror sensor.  The project manager has chosen the best performing of either DPLC or DPRC for a given flight.


Mixing ratio (g/kg) based on DPXC


Number density (molec/cc)


H2O mixing ratio (ppmv)

RAF recommends using DPXC as a slow 'tropospheric' variable, and MRTDL as a fast-response 'tropospheric' variable.  MRTDL is also recommended for all 'stratospheric' use.

Attack and Sideslip:
Measurements of attack and sideslip were done using the 5-hole nose cone pressure sensors, ADIFR and BDIFR.  Although sampled at 50 sps, internal filtering in the Mensor pressure sensors (model 6100) limits usefulness of high-rate analysis to about 5 Hz.


Vertical differential pressure


Attack angle. Determined from the vertical differential pressure of the radome gust probe.


Horizontal differential pressure


Sideslip angle. Determined from the horizontal differential pressure of the radome gust probe.

Both AKRD and SSLIP were calibrated using in-flight maneuvers.

True airspeed:
True airspeed was also measured primarily using a Mensor 6100 sensor, thus limiting the effective response to 5 Hz.

The radome pitot tube system uses the center hole of the 5-hole nose cone in conjunction with the research static pressure ports on the fuselage aft of the entrance door.  A standard avionics pitot tube is also mounted on the fuselage aft of the radome, and this system is also referenced to the fuselage static ports aft of the main entrance door.  It was found during empirical analysis that the fuselage pitot system gave more consistent results in reverse-heading maneuvers; it is suspected that this is due to random pressure changes at the radome center hole as has been suggested by modeling.  The fuselage system is used for the calculation of the aircraft true airspeed, as well as for attack and sideslip angles.  True airspeed is also provided from the aircraft avionics system, but this system is considered of slower response.  Measurements using the radome and fuselage pitot systems were corrected using in-flight maneuvers.


True airspeed using the radome system


True airspeed from the fuselage pitot system


True airspeed using the fuselage pitot system and adding humidity corrections to the calculations; this is mainly of benefit in tropical low-altitude flight


True airspeed from the avionics system


Reference true air speed. This is normally equal to TASF but TASR may be substituted in cases where TASF is compromised for any reason. This would be noted in the individual flight reports.

RAF recommends using TASX as the aircraft true air speed.

Position and ground speed:
The measurement of aircraft position (latitude, longitude and geometric altitude) and aircraft velocities relative to the ground are done using several sensors onboard the GV.

Garmin GPS (Reference):  These data are sampled at 10 sps and averaged to 1 sps.  This is a simple GPS unit with a serial output, and the measurements are available in real-time.  The values from this sensor start with a "G"; e.g.:


Latitude (recommended for general use)


Longitude (recommended for general use)


Geometric altitude (recommended for general use)


Ground speed


Ground speed in north direction


Ground speed in east direction

These are good values to use for cases where the highest accuracy is not needed.  These variables are subsequently used to constrain the INS drift for the calculations of the GV winds; more about this below.

Honeywell inertial reference system 1 and 2:  The GV is equipped with three inertial systems.  Data from the first two of these are logged on the main aircraft data logger, with subscripts the latter having variable names with suffix "_IRS2".  The advantage of the IRS values is that they typically have very high sample rates and very little noise from measurement to measurement.  However, since they are based on accelerometers and gyroscopes, their values may drift with time.  The drift is corrected for by filtering the INS positions towards the GPS positions with a long time-constant filter; the filtered values have a "C" added to the end.


latitude from IRS 1, no GPS filtering


latitude from IRS 1, filtered towards GPS values


latitude from IRS 2, no GPS filtering


longitude from IRS 1, no GPS filtering


longitude from IRS 1, filtered towards GPS values


longitude from IRS 2, no GPS filtering


ground speed from IRS 1, no GPS filtering


ground speed from IRS 2, no GPS filtering

The choice of variables for position analysis depends on the type of analysis; in general the Garmin GPS is sufficiently accurate.

Not all INS variables are output in the final data set. If you require more detailed INS data please contact RAF.

Attitude angles:
Aircraft attitude angles are measured by the two Honeywell IRS units.


Pitch of the aircraft


Pitch of the aircraft from the second IRS


Roll of the aircraft


Roll of the aircraft from the second IRS


True heading of the aircraft


True heading of the aircraft from the second IRS

The values of pitch angle (PITCH) have been corrected using in-flight measurements to give approximately the same values as the aircraft attack angle (AKRD) for long parts of each flight; this correction is performed to give a near-zero mean updraft over extended flight legs.  The variation from flight to flight of this offset is caused by small differences in the pre-flight alignment of the inertial navigation system.  No alignment correction has been applied to PITCH_IRS2.

·  Wind speeds:
Wind speeds are derived from the 5-hole radome gust probe, other pressure measurements, temperature and inertial measurements supported by
GPS data. The use of the Mensor 6100 pressure sensors for ADIFR, BDIFR and QCF results in the following limitations on the wind data: these pressure measurements were sampled at 50 sps and thus resulting in power spectra to 25 Hz.  Examination of power spectra and specifications from Mensor indicate that the sensors have internal filters with a -3dB (half-power) cutoff at 12 Hz, resulting in a noticeable roll-off in the spectra beginning approximately at 6 to 7 Hz.  Users of wind data should be aware that contributions to covariances and dissipation calculations will be affected at and above these frequencies.

The following lists the most commonly used wind variables:


Wind vector, east component


Wind vector, east component, GPS corrected for INS drift


Wind vector, north component


Wind vector, north component, GPS corrected



Wind vector, longitudinal component


Wind vector, longitudinal component, GPS corrected


Wind vector, lateral component


Wind vector, lateral component, GPS corrected



Wind vector, vertical gust component


Wind vector, vertical gust component, GPS corrected



Wind speed, horizontal component


Wind speed, horizontal component, GPS corrected


Horizontal wind direction


Horizontal wind direction, GPS corrected

RAF recommends using the GPS corrected wind components, i.e., the variables ending in "C".

Liquid water content:
A PMS-King type liquid water content sensor was installed on the GV.  The probe did not operate properly until RF11 due to a hardware malfunction. From RF11 on the probe delivered usable data.

Icing rate indicator

Rosemount Model 871FA Icing Rate Detector - Right Wing pylon Access Plate (RICE). The instrument was operational but reported little data due to infrequent penetrations of supercooled liquid water clouds.

SID-2H (Small Ice Detector, version 2)

SID-2H measures the intensity and pattern of near-forward light scattering to determine the size, shape, and concentration of cloud particles.  The estimate of particle size is based on the integrated scattering intensity, not the peak.  Scattered light falls on multi-anode photomultiplier detector so that the scattering pattern of each particle is measured in 28 pie-shaped wedges.  Spherical particles will produce symmetric patterns.  Asymmetry in the scattering pattern indicates non-sphericity (snow crystals).  This is a single-particle instrument. The scattering intensity patterns and time of arrival are logged for each particle. SID‑2H was custom-built for use on HIAPER by having very fast electronics to keep up with the fast airspeeds.

size range

~1 - 60 nm diameter

concentration range

0 - 200 cm-3

sample area

~0.3 mm2

volume sample rate

~50 cm3/s

data interface

custom, ten serial pairs

data rate

max ~9,000 particles per second

more information:

The Microwave Temperature Profiler (MTP) was deployed on START08. The MTP is a passive microwave radiometer, which measures the natural thermal emission from oxygen molecules in the earth-s atmosphere for a selection of elevation angles between zenith and nadir by scanning through an arc in the flight direction. The MTP observing frequencies are located on oxygen absorption lines at 56.363, 57.612 and 58.363 GHz. The measured brightness temperatures versus elevation angle are converted to air temperature versus altitude using a quasi-Bayesian statistical retrieval procedure. An altitude temperature profile (ATP) is produced in this manner every 18 seconds along the flight track. Temperature accuracy is approximately 1 K within 3 km of flight altitude, and < 2K within 6 km of flight altitude. Vertical resolution is approximately 150 meters at flight level, and approximately half the distance from the aircraft away from flight level. Data are provided in the form of altitude temperature profiles and color-coded temperature "curtains", and may be converted to isentropic cross sections.

CDP (Cloud Droplet Probe)

The CDP is a commercial instrument from Droplet Measuring Technologies (DMT). It measures the intensity of forward light scattering (4 - 12°) to determine the sizes of individual cloud droplets. An internal multi-channel analyzer assigns to individual particles to bins, and the data interface outputs a histogram of particle size and concentration. On the NSF/NCAR C­‑130 and HIAPER, it is mounted in a PMS canister. 

size range

2 - 50 µm diameter

concentration range

0 - 5,000 cm-3

number of size bins

10, 20, 30, or 40

sample area

200µm x 1.5mm

volume sample rate

30 cm3/s at airspeed 100 m/s

airspeed range

10 - 200 m/s

data interface

serial RS-232 or RS-422

data rate

10 histograms per second

more information:

2D-C (Two-Dimensional Optical Array Probe)

RAF-s 2DC probe is a highly modified version of the original Particle Measuring Systems (PMS) instrument.  It detects shadow images of cloud particles that pass through a laser beam. The beam illuminates a linear diode array, and each diode state changes to shadowed when a particle passes through the arms of the probe and interrupts its part of the beam. The diode array is sampled at a rate proportional to the airspeed, and this allows the shadow image to be reconstructed.

The recent (2007) modifications include using a laser diode instead of gas laser, changing from a 32‑element diode array to a 64‑element array, faster electronics, and a high speed USB‑2 data interface. From the shadow image records, RAF software derives the particle concentration and size distribution.

size range

25 - 1600 µm diameter

concentration range

0 - ~5,000 L-1

number of size bins

64 diodes @ 25 µm spacing

sample area

1600 µm x 6.1 cm

volume sample rate

10 L/s at airspeed 100 m/s

airspeed range

10 - 240 m/s

data interface


Data logging and averaging:

Analog data were logged at 500 sps and averaged to 1 sps. Serial data (e.g., RS-232), ARINC data (IRS units), etc. were recorded at the instrument-specific output rate.

The recordings listed for a given second contains measurements logged at e.g., 12:00:00.000 and until 12:00:01.  The value of "Time" corresponding to this interval is given a 12:00:00 in the released data set.

All measurements are "time-tagged" at the time of logging.  Subsequently these measurements are interpolated onto a regular grid and averaged.

RAF staff have reviewed the data set for instrumentation problems.  When an instrument has been found to be malfunctioning, specific time intervals are noted.  In those instances the bad data intervals have been filled in the netCDF data files with the missing data code of -32767.  In some cases a system may be out for an entire flight.


The following table identifies the sensors serial numbers and calibrations that were used on the GV during START-08:























































































Data Quality Control


  • There are interruptions in the data system recording when the GV crosses the UTC midnight. These interruptions may be as long as 5 minutes.
  • DPRS, DPLS, DPRC, DPLC: dewpointers tend to overshoot and oscillate after rapid temperature increase on aircraft descents. Using best judgment, these overshoots are removed from DPXC, which is the recommended reference dewpoint variable, but are left in DPLC, which is the source variable for DPXC, for comparison. The operating range for both DPLC and DPRC is down to approximately -70C and values below that should not be used for quantitative analyses.
  • LONC, LATC deviate where there is the 5 min interruption in data after midnight on all flights crossing UTC midnight. It is recommended that direct GPS position is used instead (GGLAT, GGLON).
  • Vertical wind speed (WIC) can deviate on steep climbs and descents. This is due to imperfections in the calculation algorithm.
  • PLWC (King probe liquid water content) was part of the project but did not operate properly due to a hardware problem. The instrument did not respond correctly to changing airflow and did not output proper background signal. Although there is response from the instrument to the liquid water in clouds, it is not quantitative and there are random spurious spikes in the signal. Compare PLWCC with the CDP liquid water (PLWCD_LWI). The instrument was repaired after RF11 but current algorithms do not account properly for the GV high air speed. PLWCC may be updated in the future as algorithms improve.
  • Exhaust line pressures (PDUMP*) and PCAB exhibit occasional narrow (1-3 s) spikes. These spikes are believed to be a recording system artifact, not actual pressure spikes, and were removed from the final data set.
  • Water-based CN counter experienced repeated problems with drying out at high altitudes. Sometimes the instrument would recover after being manually primed and some times it would not. Therefore, absence of WCN data (CONCN_WCN) at high altitudes does not always mean actual absence of particles.

Flight specific notes:

RF01: Water CN did not perform well for most of the flight. DPRS and DPLS were pegged at all high altitude legs and responded briefly at lower altitudes. PLWC was not operational, no quantitative data.

RF02: DPRS and DPLS were pegged at all high altitude legs.

RF03: DPLS flooded on the last descent at 20:07 and did not recover through the end of flight. The data had been set to NAN.

RF04: Both dewpointers flooded at 18:39 and recovered around 19:11. PLWC data are useless for the entire flight and are not reported.

RF05: The DSM collecting temperature data experienced a spiking problem and about 11 4-s intervals of temperature and all related data were lost throughout the flight.

RF06: DPXC overshot and was removed from 22:56 to 23:02. DPRC and DPLC are available for reference during this interval. Aircraft encountered light icing at 21:36:30 leading to obstruction of the QCR orifice and later, to water ingestion and slow sensor response. Wind data are unaffected by this because QCF is used for all calculations..

RF07: Due to overheating of the nose DSM, no temperature data were acquired until 16:38, resulting in loss of all dependent data as well. Dewpointers overshot on descents. DPXC was set to NAN for the most apparent instances, but several other regions are also present where the dewpointer readings are not realistic and demonstrate a characteristic "hump" and ringing. QCR ingested water at 16:42 and cleared up by 17:09; no other measurements were affected.

RF08: The flight included three extremely rapid descents. Dewpointers could not control under these conditions and lost nearly all data in the boundary layer. The bad sections were set to NAN in DPXC; the source variables remained untouched for reference.

RF09: Both dewpointers exhibited moderate overshooting and pronounced ringing on rapid descents. PLWC was not on aircraft for this flight.

RF10: PLWC was not on aircraft for this flight. DPXC was set to NAN during periods of overshooting and when the dewpointer power was cycled at 16:34, which caused them to balance and lose data. QCR iced up, then ingested water at 23:52:30 and did not recover until the end of the flight.

RF11: DPLC recovered better from flooding on the first rapid descent and using it instead of DPRC is recommended. DPXC is set to NAN in areas where overshooting is most prominent. Water CN counter lost flow at the beginning of flight, then recovered at 18:01. PLWC was repaired and reinstalled, and provided qualitative data.

RF12: If PLWCC is left in, blank out 16:12 - 16:17. WIC for this flight varies quite a lot, averaging at a non-zero value. This was likely caused by water ingestion into the radome ports at the beginning of the flight around 16:03.

RF13: DPXC was set to NAN during periods of overshooting. DPRC and DPLC are available to help visualize the removed areas.

RF14: The DSM collecting temperature data experienced a spiking problem and about 9 4-s intervals of temperature and all related data were lost throughout the flight. At  21:02 the radome QCR port ingested liquid water and QCR measurement diverged from QCF for the rest of the flight. A lot of maneuvers in the first half of the flight cause WIC to vary more than realistic. PLWCC and PLWCD_LWI are showing abnormal behavior from 21:02:30 to 21:08:30.

RF15: A lot of maneuvers in the first half of the flight cause WIC to vary more than realistic.

RF16: DPXC was set to NAN during periods of overshooting. DPRC and DPLC are available to help visualize the removed areas. Radome gust probe ingested water at 21:51:30 and QCR did not recover until the end of the flight. A lot of maneuvers in the second half of the flight cause WIC to vary more than realistic.

RF17: DPXC was set to NAN during periods of overshooting. DPRC and DPLC are available to help visualize the removed areas. A lot of maneuvers in the middle part of the flight cause WIC to vary more than realistic.

RF18: DPXC was set to NAN during periods of overshooting. DPRC and DPLC are available to help visualize the removed areas. Four small spikes caused by the A/D cards were removed in all temperature data.