The aircraft measurements of the atmospheric boundary layer during the FIFE Intensive Field Campaigns were part of a coordinated effort which also included surface measurements, balloon-borne profiles, and SODAR and LIDAR remote sensing. The chief objective of the boundary layer investigations was to describe the structure of the atmospheric boundary layer over the FIFE study area, increase knowledge of the physical processes active in the daytime boundary layer, and explore the relationship of surface properties to the time and spatial variation in the structure of the boundary layer. Three aircraft contributed to these objectives by measuring winds, humidity, temperatures, and radiation at a variety of altitudes along specially designed flight patterns.

This data set group includes the raw data and derived flux estimates from the measurements made by the NAE (Canada) Twin Otter, the NCAR King Air, and the University of Wyoming King Air. The NCAR plane flew during IFCs 1 and 5; the Twin Otter flew during IFCs 2-5; and the UW King Air was used during IFCs 3 and 4.

Most of the UW flux measurement missions during IFC 3 and IFC 4 were coordinated patterns involving the Wyoming King Air and the Twin Otter. The King Air portions of the flights were usually stacks of horizontal flight lines at various levels. For each flight line measurements were made of radiation fluxes and atmospheric temperature, pressure, and winds. Latent and sensible heat fluxes were calculated from wind gust, temperature, and momentum flux statistics.

The phenomena studied were the daytime convective boundary layer structure and physical processes. The study used airborne measurement of vertical and horizontal wind gusts, humidity, potential temperature, mean horizontal wind speed, and horizontal linear trends of temperature, humidity, and radiation. Fluxes of sensible heat, moisture, and momentum were estimated from fast-response wind gust, temperature, and humidity measurements.

The Twin Otter made measurements in the boundary layer of the fluxes of sensible and latent heat, momentum, and carbon dioxide. Supporting meteorological parameters, such as temperature, humidity, wind speed, and wind direction were also recorded, along with aircraft position and several radiometric observations.

All of the flux estimates were obtained with the eddy correlation method. The aircraft is equipped with an inertial platform, accelerometers, and a gust probe for measurement of earth-relative gusts in the X, Y, and Z directions. Gusts in these dimensions are then correlated with each other for momentum fluxes and with fluctuations in other variables to obtain the various scalar fluxes, such as temperature (for sensible heat flux) and water vapor mixing ratio (for latent heat flux).

Three sets of fluctuation and flux data were submitted for each flight line. (Thus, each aircraft and observation date has three files associated with it on the CD-ROM, distinguished by the file extension). The first set used untreated time histories in the derivation of the fluxes. The second set used linearly detrended data, and the third used time histories that were high-pass filtered with a third-order algorithm with a break point set at 0.012 Hz (5-km wavelength).

The flux aircraft flew a number of different flight patterns (grids, L-patterns, profiling stacks, soundings), either alone or coordinated with each other. In IFC 1 the NCAR King Air worked alone. Various mission profiles were explored, including L-patterns in which the aircraft flew along the north and east sides of the FIFE site. During IFC 2 the NAE Twin Otter flew numerous east-west runs and grid patterns. Most of the flux measurement missions during IFC 3 and IFC 4 involved coordinated patterns using the Wyoming King Air and the Twin Otter. One such flight included a horizontal grid pattern for the Twin Otter and a vertical "stack" of horizontal passes for the King Air. In IFC 5 the NCAR King Air worked with the Twin Otter.

Convective boundary layer budget missions were scheduled so that the average time of day was constant for all pairs of legs flown at the same level. Time variance was explored by coordinated missions in which both planes had the same objective and flight plan.

Observations of the atmospheric boundary layer over the FIFE study area were obtained by the three flux aircraft, both in individual and coordinated flight patterns, as well as several other sets of instruments (Kelly, 1992). The aircraft made eddy correlation measurements, subdivided for processing into horizontal passes, which are presented in the CD-ROM data files on a pass-by-pass basis. The information for each pass includes time, aircraft heading, and pass end-points, which allow reconstruction of the overall pattern used for that particular experiment.

In addition to the aircraft, two different LIDARs were present and extensive radiosonde profiles were regularly obtained. These measurements provide the basis for both verifying and extending the aircraft measurements. For more details on the information that is available, the reader is referred to the summary paper on Vertical Atmospheric Soundings and Profiles (Lutz and Strebel).

The aircraft flux estimates can also be extrapolated to the surface and compared with the measurements by the surface flux investigators. The surface flux data is described in another summary paper in this series from the CD-ROM (Field and Heiser) and is compared to the aircraft flux data in Kelly, Smith, and MacPherson (1992).

The airborne flux measurement program was technically complex, involving measuring earth-relative fluxes of heat and momentum from a moving platform. Much of the work in FIFE was experimental; for example, testing a variety of flight patterns, averaging lengths, and filtering procedures. The individual detailed documents for each data set list a number of cautions and potential sources of error that should be considered by someone anticipating independent use of the data.

Several specific confidence and error issues have been analyzed by the investigators themselves. The 15-km run length was found to be just adequate for low-level flux estimates (as long as the runs were cross-wind), but too short nearer the inversion level (Grossman, 1992). It is estimated that the 15-km flight paths prevent agreeing with the surface flux measurements to within even 10% (Kelly, et al., 1992). This is partially due to undersampling of high frequency, short wavelength fluxes. Overall, the surface flux estimates made from the airborne measurements were about 30% less than the averages of the values determined by the surface stations for sensible heat and 10% less than the latent heat averages.

There is also a question about how well the measurements compare between aircraft. Wing-to-wing comparison flights were flown on four days in 1987 and three days in 1989 (MacPherson, et al., 1992). Sensible heat fluxes agreed to within 15 [W][m^-2] and moisture fluxes were within 21 [W][m^-2]. Other measurements were equally reliable: mean wind component, 1.0 [m][s^-1]; air temperature, 0.3 [deg C]; mixing ratio, 2 [g][kg^-1]. In general, it was felt that there is little need for corrections to adjust for instrument biases, but that researchers combining sets of boundary layer data from different aircraft should consider the comparison results carefully.

Note that three sets of measurements and flux data are provided (designated by different filename extensions on the CD-ROM). The first uses untreated time histories in the derivation of the fluxes using the eddy correlation technique, the second uses linearly detrended data, the third uses time histories that were high-pass filtered with a third-order algorithm with a break point set at 0.012 Hz (5-km wave length). The investigators express the opinion that most scientists working with flux and correlation coefficient data would prefer to use the linearly detrended data.

The aircraft flux measurement investigators in FIFE produced results both in measurement techniques and in scientific understanding of the atmospheric boundary layer. Results in the measurement category have been mentioned already in Sections VII and VIII above. Betts, et al. (1992) used the grid flight data from the Twin Otter to analyze the boundary layer budget using a mixed-layer model. The fluxes through the top of the boundary layer were found to be twice as large as had been the conventional wisdom for modeling studies.

Kelly, Robert D., Eric A. Smith, and J. Ian MacPherson 1992. A Comparison of Surface Sensible and Latent Heat Fluxes from Aircraft and Surface Measurements in FIFE 1987. J. Geophys. Res. 97(D17):18445-18453.

Kelly, Robert D. 1992. Atmospheric Boundary Layer Studies in FIFE: Challenges and Advances. J. Geophys. Res. 97(D17):18373-18376.

Grossman, Robert L. 1992. Sampling Errors in the Vertical Fluxes of Potential Temperature and Moisture Measured by Aircraft During FIFE. J. Geophys. Res. 97(D17)18439-18443.

MacPherson, J.I., R.L. Grossman, and R.D. Kelly, 1992. Intercomparison Results for FIFE Flux Aircraft. J. Geophys. Res. 97(D17):18499-18514.

Betts, A.K., R.L. Desjardins, and J.I. MacPherson. 1992 Budget Analysis of the Boundary Layer Grid Flights During FIFE 1987. J. Geophys. Res. 97(D17)18533-18546.