The output of the sensors was transferred to the PAM base and archived for future analysis. The data transfer was by spread spectrum direct RF linkage from the individual stations to the PAM base at the Bondville site. For Station 1, which employed both an enhanced flux-PAM and elements of the ASTER facility, data was both acquired by ADAM and transmitted directly to the PAM base, and also processed in situ by EVE and the data products transmitted to the PAM base. For Stations 2 and 3 only the latter mode was used. In addition to the direct RF linkage to the PAM base the flux-PAM's also transmitted their data products via GOES, Boulder Earth-base and then by internet to the PAM base at Bondville. A further redundancy was provided by local storage at the flux-PAMs. Table 9 and 10 outline the various data streams and data products of the FLATLAND96 program and indicate their availability for different time intervals for the three stations
TABLE 9 Data available from sensors at Stations 1, 2 and 3
----------------------------------------------------------------------------------------
Parameter Units Station 1 Station 2 and 3
0.1 s 1 s 60 s 300s 60 s 300s
TABLE 10 Data products available for FLATLAND96
---------------------------------------------------------------------------------------------
Parameter Mean Covariances Variance 3nd 4nd
moment moment
All data acquired during the deployment is archived. The application of the appropriate calibration and delay functions convert this raw data to engineering or scientific units. This processed data is available as 1, 5 or 10 minute block means, variances and covariances obtained from the initial processing of the variables measured.
TABLE 11 FLATLANDS96 data available as netcdf files
--------------------------------------------------------------------------------------------------------
Netcdf variable Measurement Time Units Station Comments
resolution
TABLE 12 Sonic anemometer boom azimuth
------------------------------------------------------- Station Angle Start Stop------------------------------------------------------- Station 1 221o 23' 00" Jun 18, 00:00 Jul 9, 18:00 224o 51' 12" Jul 9, 18:00 Aug 23, 14:00 Station 2 215o 10' 48" Jun 18, 00:00 Jul 3, 14:00 213o 58' 25" Jul 3, 14:00 Aug 23, 14:00 Station 3 230 11' 50" Jun 18, 00:00 Aug 23, 14:00 -------------------------------------------------------
TABLE 13 Sonic anemometer heights
-------------------------------------------------- Station Height, m Start Stop-------------------------------------------------- Station 1 3.20 18 Jun, 00:00 Jun 27, 22:00 4.50 Jun 27, 22:00 Jul 9, 18:00 5.70 Jul 9, 18:00 Aug 23, 14:00 Station 2 4.54 18 Jun, 00:00 Jul 3, 14:00 5.67 Jul 3, 14:00 Aug 23, 14:00 Station 3 3.21 18 Jun, 00:00 Jul 26, 16:00 4.00 Jul 26, 16:00 Aug 23, 14:00 --------------------------------------------------
Based upon this analysis, each of the sonic anemometer data sets were divided into between four and six periods. Table 14 shows the periods for the three stations and the offsets and coefficients derived for each individual period. The data was subjected to analysis and the offsets and tilt coefficients calculated. These offsets and coefficients were then use to calculate the correct turbulence for the entire deployment.
TABLE 14 Sonic anemometer data periods
---------------------------------------------------------------------------
Start Stop Interval woffset lean leanaz
Note that wind direction information obtained from the sonic anemometers is expressed in the coordinate system of the sonic anemometer. The boom azimuth angles for the three sonic anemometers re given in Table 12.
The fast temperature data is derived from the speed of sound measured by the sonic anemometer. This fast temperature is used with the turbulence data to calculate the sensible heat flux. This temperature is actually a virtual temperature, which is a linear combination of temperature and humidity. By subtracting the contribution from independently measured, humidity fluctuations the sensible heat flux is obtained.
For Station 1 the primarily fast humidity data was derived from the Campbell Scientific Krypton hygrometer. This data was used in conjunction with the turbulence data to calculate latent heat fluxes. At all three Stations the band pass covariance technique was also used to derive latent heat fluxes. This technique allows a relatively slow-response humidity sensor, when used in conjunction with fast response turbulence and thermal sensors, to derive latent heat fluxes. Due to scalar similarity, the relationship between the covariances, w'T' and w'q', is assumed to extend over the full frequency range. The covariances, w'T' and w'q' are matched at lower frequencies, and then the relationship is extended to the higher frequencies. In this manner w'q' can be defined to higher frequency than themoisture sensor can actually attain. The matching of the covariances is undertaken using a fast-Fourier transform of the time series in the frequency domain.
The wind direction data is derived by combining the sensor coordinate signal with the azimuthal alignment of the sensor boom. The wind speed data is derived from the wind tunnel calibration of the individual propeller. Post project calibration revealed an error in the prop pitch values used during FLATLAND96. This resulted in correction factors being applied to the NETCDF data set. The raw data set was not modified.
The response of the three ozone sensors had first to be synchronized to the sonic anemometer. Sample air for Kokod was drawn from the immediate vicinity of the sonic anemometer, along 10 m of quarter inch teflon tubing, through a mass flow controller and into the low pressure reaction chamber. The time offset was invariant and was calculated by reference to the cross correlation between Kokod and the vertical component of turbulence. A value of 0.3 seconds time delay was determined. Seclod was mounted on the boom 70 cm behind the sonic anemometer. The time offset was again calculated by reference to the cross correlation with the vertical component of turbulence. In this case the time offset varied with wind speed and direction. Sample air for Tecod was also drawn from the immediate vicinity of the sonic anemometer, along 10 m of quarter inch teflon tubing. Tecod operates at near ambient pressure and the flow rate through the intake was much slower than for Kokod. Again the time offset was invariant, but calculation by cross correlation between Tecod and the vertical component of turbulence gave only a crude estimate of 20 seconds delay. This was due to the ten second resolution of the Tecod data and the manner in which Tecod reports data. Tecod has two absorption cells through which sample air is drawn. Alternately the air is switched to pass through a manganese oxide scrubber to remove ozone. Air flow for 7 seconds flushes the cell and then a 3 seconds measurement period allows the ultraviolet light intensity to be determined. The value for the ozone mixing ratio is calculated and then reported for the duration of the following 10 seconds. In the post-project data analysis the data point at the center of the 10 second reporting period was assigned to the midpoint of the previous cell flushing period, a delay of 11.5 seconds An additional delay of 6 seconds due to inlet flow time was added resulting in a total time offset of 17.5 seconds.
The data from the Tecod/Kokod combination is available but treatment of the corresponding Tecod/Seclod data is still ongoing.
The Q7 net radiometers used at all three stations were calibrated, prior to the deployment, by REBS. Different calibration coefficients were provided to be used for periods of net up-welling and for periods of net down-welling radiation. The net radiometric fluxes reported here are generated using these coefficients applied to the net radiometer output. REBS has announced that these coefficients will need to be revised in light of their subsequent appreciation that further corrections need to be applied to the Eppley pyranometer and pyrgeometer data which they used during all previous Q7 calibrations.
The values for incoming solar, visible radiation fluxes reported at all three stations were generate from the Licor pyranometer output using calibration coefficients supplied by the manufacturer at the last calibration in August, 1984.
Station 1 1.54 g cm-3
Station 2 1.64 g cm-3
Station 3 1.49 g cm-3
The continuously-output data for soil moisture given by the soil moisture sensors were first corrected to reduce the effect of the temperature response and were then normalized to best fit the periodic gravimetric soil sample points. There is a discernible diurnal variation in the soil moisture trace at all three sites. This variation is most extreme early in the deployment, particularly for the soybean site and diminishes as the canopy develops. This indicates the effect is driven by solar heating of the bare soil. The effect could be due to imperfect temperature compensation or it could be due to daytime desiccation of the upper layers followed by subsequent nocturnal capillary refill by moisture from deeper in the soil.
Surface heat fluxes can be calculated using the value of the flux at 5 cm below the soil surface corrected by a term for the thermal capacity of the top 5 cm of soil. This thermal capacity was calculated from the temperature and the heat capacity of the top 5 cm of soil. The thermal capacity is calculated from the bulk density and the water content of the soil.
The four day weather plots show values determined at Station 1 for:
The values plotted for temperature, humidity, pressure, rainfall, wind speed and direction, net radiation and soil heat flux are 5 minute averages, while the values plotted for the sensible heat flux, the latent heat flux, Z/L, u*, Bowen ratio are 20 minute averages.
Weather plots for Station 1 are available as postscript files:
steam:/netaster/projects/FLATLAND96/results/plots/
Table 15 Station 1 weather plots wx960615.ps wx960619.ps wx960623.ps wx960627.ps wx960701.ps wx960705.ps wx960709.ps wx960713.ps wx960717.ps wx960721.ps wx960725.ps wx960729.ps wx960802.ps wx960806.ps wx960810.ps wx960814.ps wx960818.ps wx960822.ps
An electronic logbook was maintained during the field program. A total of ~500 entries were made during the FLATLAND96 deployment in 21 different categories:
TABLE 16 Logbook categories
------------------------------------------------
ADAMS/NETWORK BAROMETERS BAND PASS HYGROMETER
DATA EVE GOES/RF
HYGRO LOG LOGISTICS
OZONE PROPS RADIATION
RAIN SOFTWARE SOIL
SONICS STATUS TAPE_ARCHIVE
UNIX VISIT_LOG WEATHER
------------------------------------------------
The electronic logbook is available to complement this report.