Title: BALTEX_Cabauw Flux

CONTACT(S):

Dr. Fred C. Bosveld
Royal Netherlands Meteorlogical Institute
Wilhelminalaan 10
P.O.Box 201
3730AE, De Bilt
The Netherlands
tel. +31 (0)30 2206911 (787)
fax +31 (0)30 2210407
e-mail: fred.bosveld@knmi.nl

1.0 DATA SET OVERVIEW

Contains the surface flux observations at the BALTEX anchor station Cabauw, lat. 51.97N lon. 4.93E, height (m.s.l.) -0.7 m.

This data set includes soil heat flux at 0, -0.05, and -0.10 m heights and latent heat flux, sensible heat flux, and CO2 flux at 5.37 m height.

2.0 INSTRUMENTATION DESCRIPTION:

2.1 Turbulent surface fluxes

Turbulent fluctuations of wind, temperature, humidity and CO2 are measured with a combination of a sonic anemometer/thermometer (wind vector and sonical temperature) and an optical open path sensor (H2O and CO2). From this the vertical fluxes of momentum, sensible heat, latent heat and CO2 are derived by means of the eddy correlation technique. Let wi be a sampled timeseries of the vertical wind speed and ci the quantity for which the vertical flux is to be determined. The average turbulent flux (Fc)for a given time periode (T) in which N samples are acquired is then given by:

Fc = 1/N*SUM((wi-< w >)*(ci-< c >)

Where < w > and < c > denotes the average values over the period T.

Corrections on the sonical temperature are applied for moisture and lateral wind. Humidity and co2 fluctuations are corrected for density fluctuation induced by temperature and humidity fluctuations (Webb-correction). Wind fluctuations are corrected for streamline tilt due to flow obstruction around the supporting mast and instruments. Low frequency losses are corrected according to Bosveld (1999), this method is verified by Schalkwijk et al. (2010). No corrections for sensor seperation and path averaging is applied untill now. Comparing co-spectra of wt and wq for the current 3 m configuration shows a typical loss of 6% in wq due to the sensor seperation between the sonic and the licor. Spectral losses in wt will be significantly smaller since the paths of the temperature and vertical wind observations coincide. (ongoing work)

Details and limitations of the eddy correlation technique have been described extensively in the open literature. A good introduction can be found in Xuhui Lee, W. J. Massman, Beverly E. Law (2004)

Instruments

20000801

The sonic anemometer is a Kaijo-Denki, probe type TR60-A, electronic unit DAT-300 or DAT-600. The sonic path is 0.2 m. Resolution is 0.1 K. The H2O/CO2-sensor is a KNMI Infrared Fluctuation Meter (IFM). Pathlength is 0.3 m. Resolution is 0.003 g/m3 H2O and 0.15 ppm CO2.

20051101

IFM is replaced by LICOR 7500 Open path H2O/CO2 sensor

20060913 - now

Kaijo Denki is replace by Gill R3 sonic anemometer/thermometer. Configuration

20000801

Turbulent surface fluxes are measured at 5 m height approximately 200 m South of the main tower . A sonic anemometer/thermometer is used to measure turbulent fluctuations of the three wind components and (sonical) temperature. The sonical temperature is measured along the vertical transducer pair. An open path infrared fluctuation meter is used to measure turbulent fluctuations of humidity and carbondioxide. The sonic anemometer has an azimuthal opening angle of 120o for horizontal wind measurements. The open path sensor is positioned vertically just behind the sonic probe at a distance of 0.3 m from the vertical sonical path. The instruments are mounted on a 1 m thin vertical cylinder to avoid a too strong flow obstruction due to the supporting mast. The vertical cylinder is supported by a rotator which is controlled by a wind direction tracking system and automatically turned into the mean wind direction each 2 hours. An inclinometer is positioned between the rotator and the supporting cylinder.

20060913 - now

To avoid further interference with the maize field to the West the surface flux equipment is moved to the North of the Cabauw site (Energy balance terrain). To get mainly grass in the footprint the instrument level is lowered to 3 m. To diminish maintenance the rotator device has been removed and a fixed configuration is made. The Licor is now below the sonic anemometer to diminish flow obstruction. This has consequences for the flux loss due to sensor seperation (not yet treated)

Data logging and data reduction

Data are logged at a rate of 10 Hz through a PC-MicrostarA/D combination. Fluxes are calculated on a 10 minute time basis.

Field conditions

The orientation of the ditches is in the direction 157o - 337o. The position of the 5m turbulence mast is circa 25 m from the Western border of the KNMI terrain. Since 2003 the neighbouring farmer grows maize at the bordering field. Typically the field is bare soil till March. Growing of the maize occurs from April till September. The maize grows up to a height of 2 m. Harvesting is at the end of September or beginning of October. In September 2006 the turbulence mast has been moved to avoid further interference.

Calibration

Calibration of the sonic anemometer is done, approximately each second year, in a wind tunnel of TNO-Apeldoorn. Occasionally temperature calibrations has been performed at KNMI. H2O and CO2 calibration is performed in the field each year according to the calibration procedure provided by the manufacturer. Currently a calibration procedure for temperature, humidity and co2 calibration is setup in the calibration laboratory at KNMI

Known Problems

Sonic temperatures were compared with the standard temperature observations at Cabauw. By regression for each day an estimate of the sensitivity of the sonic temperature was obtained. It was found that sensitivity was 0.97 before februari 2003 after that sensitivity started to deviate till june 2004. After that the instrument was replaced and sensitivity was back to 1.00. Covariances with temperature are corrected according to the table below. For intermediate dates sensitivity values were interpolated.

DateSensitivity
200001010.97 200302010.97 200306090.81 200403010.81 200407190.65 200408111.00 200912311.00

Quality Checks and processing

Data are archived near real-time in 10 minute intervals. This makes up the unvalidated dataset. After each month data are checked for completeness. Data are checked and flagged for mall functioning by visualisation of time series. This make up the the validated dataset Data rejection mostly takes place under unfavourable conditions like rain and dew formation.

References Bosveld F.C. (1999).The KNMI Garderen experiment, micro-meteorological observations 1988-1989: corrections. KNMI publication: WR-99-03. Schalkwijk, J., F.C. Bosveld and A.P. Siebesma (2010). Timescales and structures in vertical transport in the atmospheric boundary layer. KNMI publication: WR-2010-02. Xuhui Lee, W. J. Massman, Beverly E. Law (2004). Handbook of micrometeorology: a guide for surface flux measurement and analysis

2.2 Soil heat flux

Soil heat flux is measured at the soil-terrain, 100 m South of the main tower, with six soil heat flux plates. The six plates are burried at the three vertices of an equilateral triangle with sides of 2 m at depths of 0.05 and 0.10 m.

The instruments are manufactured by TNO-Delft. Type: WS31S, principle: thermo-pile, diameter 0.11 m, thickness 5 mm, sensitive surface: central square of 25*25 mm2. Calibration is done by TNO-Delft.

3.0 DATA COLLECTION AND PROCESSING:

3.1 Data collection

Turbulence data are logged at a rate of 10 Hz through a PC-MicrostarA/D combination. Fluxes are calculated on a 10 minute time basis. Soil heatflux plates are logged is with a Campbell Scientific CR21X datalogger. Measurements are taken 5 times in a 10 minute interval. Averages over 10 minutes are saved.

3.2 Data processing

Corrections on the sonical temperature are applied for moisture and lateral wind. Humidity and CO2 fluctuations are corrected for density fluctuation induced by temperature and humidity fluctuations (Webb-correction). No corrections for high frequency loss are applied since they are estimated to be very small. Corrections for low-frequency loss are applied (Bosveld, 1999). All these corrections are applied on a 10 minute basis. No corrections for streamline tilt are performed. The soil heat flux measurements are averaged over the three plates at each depth. To obtain the surface soil heat flux a Fourrier decompostion method is used.

Data are in the CEOP data format agreed to by the CEOP Scientific Steering Committee. This format is described in detail as part of the CEOP Reference Site Data Set Procedures Report which is available at the following URL: http://www.joss.ucar.edu/ghp/ceopdm/refdata_report/ceop_flux_format.html

4.0 QUALITY CONTROL PROCEDURES

4.1 Cabauw Quality Control Procedures

After the data are stored in the database a manual (on-eye) check is performed and together with information from the logbook suspect data are rejected.

4.2 UCAR/JOSS Quality Control Procedures

UCAR/JOSS conducted two primary quality assurrance/control procedures on the reference site data. First the data has been evaluated by a detailed QA algorithm that verifies the format is correct, examines any QC flags, and conducts basic checks on data values. Second, JOSS conducts a manual inspection of time series plots of each parameter.

5.0 GAP FILLING PROCEDURES

No gap filling is applied

6.0 DATA REMARKS:

6.1 Sensible heat flux correction

New version of BALTEX_Cabauw 1st part of EOP3 flux data:

A problem in the tubulent temperature sensor at the BALTEX_Cabauw location was detected which gives an error in the sensible heat flux observation. The error occurred from 1-Feb-2003 till 19-Jul-2004. Correction are applied. The error was detected after the flux data of the 1st part of EOP3 was send in to the CEOP data centre. The flux data for the 1st part of EOP3 were resubmitted to the CEOP data centre since 27-apr-2005. Flux data from BALTEX_Cabauw for the later EOP's were already corrected at the time of their first submission.

Nature of the problem and correction procedure:

A drift in sensitivity of the turbulent temperature sensor was detected by comparing the diurnal variation of turbulent temperature observations with the standard temperature observations at Cabauw. Corrections had to be applied for the 1st part of the EOP3 ranging from 3% at the beginning to 10% at the end of the period. In the following EOP's sensitivity drifted further down to ultimately a value of 35% in July 2004. Corrections were applied also for these periods. In August 2004 the sensor was replaced and sensitivity was back to 1.00.

6.2 Other Remarks

The main reason for missing data is wet conditions when the turbulence instruments give unreliable data. The surface energy budget which can be made up of the data in this file and the net radiation in the acompanying .sfc file does not sum up to zero. This is a well known problem in micro-meteological science and not specific for the Cabauw site. Typical imbalances are 15 % during daytime and even larger during night time. A poormans approach to might be to split the available energy (net radiation minus surface soil heat flux) with the observed Bowen ratio (ratio between sensible and latent heat flux) into the two turbulent heat fluxes.

7.0 REFERENCE REQUIREMENTS:

Users of the data are required to follow the CEOP data policy. Above this users of this data are kindly requested to acknowledge:

"The Royal Netherlands Meteorological Institute"

in any publication in which the data are used. Questions about the data can be directed to the e-mail adress given above. The user is kindly asked to communicate any problem in the data set to the e-mail adress given below. Users are also kindly asked to send a copy of any publication in which the data are used to the adress given above.

8.0 REFERENCES:

Bosveld F. C. (1999). The KNMI Garderen experiment: micro-meteorological observations 1988-1989, corrections. Published in the scientific reports series of KNMI, P.O.Box 201, 3730AE, De Bilt, The Netherlands. Ref. KNMI-WR 99-03.