README file for the CASES99 data from the CIRES Tethered Lifting System (TLS) PI: Ben Balsley Co-PI: Rod Frehlich 303-492-1055 303-492-6776 balsley@cires.colorado.edu rgf@cires.colorado.edu Project Manager: Mike Jensen 303-492-3999 mike@terra.colorado.edu Data Format: there are two netCDF files for each sensor package and flight, e.g., ncdf_f10p2_lf and ncdf_f10p2_hf Low Frequency Data: Atmospheric parameters sampled at 1 Hz Data files are denoted as ncdf_f10p2_lf f10 is flight 10 p2 is sensor package number 2 High Frequency Data: Temperature and velocity sampled at high frequency (200 Hz) Data files are denoted as ncdf_f10p2_hf f10 is flight 10 p2 is sensor package number 2 ******************** Low Frequency Data Files ********************************** The low-frequency netCDF files contain header information that includes the flight number, package number, date, and initialization time (hr:min:sec GMT). All the data consists of 180 second blocks with 9 second gaps (required to write the data to memory). Each block of data consists of 11 variables. Note that there are no place holders for the missing data in the 9 second gaps. Column netCDF Variable Units Notes ------ ------------------ ---- ------------------------------------------ 1 GMTSEC_LF sec GMT (UTC) seconds of the day 2 T_ext C temperature from solid state sensor 3 T_cw C temperature from cold-wire sensor 4 U_pt m/s wind speed from pitot tube sensor 5 U_hw m/s wind speed from hot-wire sensor 6 h km height (AGL) from basic payload 7 P mb pressure from basic payload 8 pa degrees pitch angle of package (+=nose up) 9 ra degrees roll angle of package 10 dir degrees wind direction from true north 11 dirflag no units quality flag for direction (0.6-1.4) GMTSEC_LF: The GMT time was determined by setting both PC clocks (BP data) and scientist wrist watches (LF & HF data) using the time from a GPS receiver. Therefore the time synchronization between each package has an inherent human error of under 1 sec. T_ext: The external temperature is measured using an Analog Devices model AD22100KT solid state temperature sensor. The two sensors on package numbers 1 and 2 have been calibrated in an oil bath at NCAR, and the other 3 packages calibrated to the first two using data from a side-by-side calibration. The absolute accuracy is better than 0.5 degree C and the slope is accurate to better than 3%. T_cw: The low-frequency cold-wire temperature is produced from a tungsten wire with a diameter of 5 microns and a length of 1 mm. The fine-wire was operated in constant current mode and analog electronics produced a low-pass signal and a high-pass signal with gain. The cold-wire voltage V_cw was calibrated by a spectral technique to connect the low-frequency cold-wire voltage to the low-frequency external temperature signal T_ext. The error in the best-fit straight line was less than 5% for all data sets. The cold-wire temperature T_cw is a 1 sec. average of the reconstructed high frequency signal tcwhf using the known analog filters of the electronics and the low-frequency external temperature T_ext to correctly merge the low-frequency and high-frequency signals. U_pt: The pitot tube velocity is calculated from the dynamic pressure measured by a 1-inch water column full-scale differential pressure sensor made by Data Instruments (part no. DCXL01DN). The offset voltage of the sensor is periodically (typically every 60 seconds) checked by switching a solenoid valve and measuring the pressure inside the instrument, where the second port of the differential sensor is always used as a reference. The velocity is calculated from the measured pressure difference using the equation: U=(2(dP)/rho)^(1/2), where dP is the pressure difference in Pa, and rho is the atmospheric density calculated using the corrected pressure (see P, below) and relative humidity from the BP, and the temperature from the AD22100KT. Note: The pitot tube velocity is measured with respect to the sensor package and therefore includes the velocity of the package. U_hw: The turbulent velocity measurements were produced with a tungsten wire with a diameter of 5 microns and a length of 1 mm. The fine-wire was operated in constant temperature mode with an overheat ratio of 1.8 and analog electronics produced a low-pass signal and a high-pass signal. The hot-wire voltage V_hw was calibrated using a modified version of King's Law and the pitot tube velocity as a calibration reference. The modified King's Law equation used is: U_hw= 1/(rho*d^(1/n))*(V_hw^2/(Tw-T) - c)^(1/n) c=offset, d=slope, Tw=wire temperature(K), n=exponent, T=ambient temperate (K) from T_ext, and rho=is the atmospheric density The King's Law parameters are determined by minimizing the mean-square-error between the hot-wire velocity U_hw and the pitot-tube velocity U_pt using the Powell routine in IDL. The high frequency hot-wire voltage v_hw is produced by using the known analog filters of the electronics to merge the low-frequency and high-frequency voltage signals. U_hw is a 1 sec. average of the calibrated high-frequency velocity uhwhf using King's Law. The quality of the velocity fit can be determined by comparing the pitot tube velocity, U_pt, with the hot-wire velocity, U_hw. The absolute accuracy is better than 1 m/s and the velocity fluctuations have an accuracy better than 5% for the data with good fits to King's Law. Note: The hot-wire velocity is measured with respect to the sensor package and therefore includes the velocity of the package. h: height above ground of each package is calculated from a linear interpolation of the 10-sec BP data to 1-sec. The BP height is calculated from the Vaisala RS-80 radiosonde pressure, temperature, and relative humidity by integrating the hydrostatic equation while accounting for water vapor content of the air. The error in the height will increase with time. P: pressure at each package is calculated from a linear interpolation of the 10-sec BP data to 1-sec. The pressure is then corrected for the package package distance from the BP using a 2nd order polynomial fit of the BP altitude and pressure, where the package distance from the BP is known. pa & ra: The pitch and roll angles are measured using an Advanced Orientation Systems (AOSI) EZ-TILT 2000 dual axis tilt sensor. These sensors were calibrated in the laboratory with manual angle sensors. dir: The wind direction is calculated from the outputs of a two-axis magnetometer. As the package vanes in the wind the direction is measured with respect to magnetic north. The true direction referenced to true north is given by magnetic direction + 5.322 degrees. The magnetometer sensors were calibrated using the best-fit of all the CASES data. A small offset error may exist due to ferrous objects (batteries, sensors, etc.) around the sensors. dirflag: This flag measures the quality of the direction. A value of unity is a good reading. Typically, this flag is between (0.6,1.4). ******************* High Frequency Data Files ********************************** The high-frequency netCDF files contain header information that includes the flight number, package number, date, and initialization time (hr:min:sec GMT). The high-frequency data is produced from the calibrated cold-wire and hot-wire signals discussed in the low-frequency calibration section (T_cw and U_hw). All the data consists of continuous 180 second blocks with 9 second gaps (required to write the data to memory) that are synchronized with the low-frequency data. Each block of data consists of 4 variables (GMTSEC_HF, tcwhf, uhwhf, and tcwac) typically sampled at 200 Hz. Note that there are no place holders for the missing data in the 9 second gaps. Column netCDF Variable Units Notes ------ ------------------ --------- ------------------------------------------ 1 GMTSEC_HF millisec GMT (UTC) milli-seconds of the day 2 tcwhf C High frequency temperature 3 uhwhf m/s High frequency velocity 4 tcwac C High frequency temperature fluctuations GMTSEC_HF: The GMT time is interpolated from the low-frequency time to the high-frequency sampling interval which is typically 5 millisec (200 Hz). The time in milli-seconds is stored as 4 Byte integers. tcwhf: the calibrated reconstructed high-frequency temperature using the known analog filters to correctly merge the high-frequency cold-wire signal with the low-frequency external temperature T_ext. uhwhf: the calibrated reconstructed high-frequency velocity using the known analog filters to correctly merge the high-frequency and low-frequency hot-wire voltages. The merged high-frequency voltage v_hw was converted to high-frequency velocity uhwhf using Kings Law obtained from the low-frequency hot-wire signal and the pitot-tube wind speed. The pitot-tube and hot-wire wind speed are measured with respect to the sensor package. tcwac: The fluctuating component of the high-frequency temperature obtained from the high-pass cold-wire voltage and the temperature calibration constant of the cold-wire. This temperature is ideal for spectral analysis because the low-frequency ramps and gradients have been removed by the analog high-pass filter.