UV-B Monitoring and Research Program at Colorado State University

Project Staff | Research Publications | Contact Us

UV-B Monitoring and Research Program

Quality Control Procedures

Every day, the instruments voltage in the dataloggers are downloaded to the UVMRP database servers at Colorado State University. This process typically takes four hours. Any sites which experience connectivity issues are addressed by UVMRP staff on the next business day. After the download processes are completed, numerous automated scripts begin the tasks of applying quality control and calibration factors to ensure that the data is of the highest standard.

Voltage Range Checks

As the raw voltage is being loaded into our system, the below range checks are applied:

Visible MFRSR

The wavelength data has a value over 4090 millivolts the saturation code is set for that wavelength(s#;) where # is the wavelength.
The head temperature is below 2100 millivolts, the code htl; is set.
The head temperature is above 3100 millivolts, the code hth; is set.

UV MFRSR

The wavelength data has a value over 4090 millivolts the saturation code is set for that wavelength(s#;) where # is the wavelength.
The head temperature is below 2500 millivolts, the code htl; is set.
The head temperature is above 3100 millivolts, the code hth; is set.
The bias wavelength value is below -100 millivolts. This indicates a failure for that wavelength and a f#, where # is the wavelength is set.

UVB-1

The broadband dome temp is less than 700 millivolts or greater than 875 millivolts, a bt; code is set.

PAR

The solar elevation is less than -10, and the par sensor is greater 27 or the par sensor is less than -10 millivolts, the code pd for a par problem
The solar elevation is greater than 25 and the sensor value value is 0 or less than 0, the par sensor has failed and an pf; code for a par failure is set.

Barometer

The barometer is less than 400 millivolts or over 2400 millivolts, a barometer failure has occured(bf;)

UVA

If the uva dome temperature is less than 900 millivolts or greater than 1100 millivolts, it is out of specification and the uva data may be compromised. The at; code is set.

Humidity and Air Temperature

If the air temperature is less than 0 or greather than 10000 millivolts, there is an issue and a temperature failure flag(tf;) is set.
If the humidity is less than 0 then there is a humidity failut and hf; is set.
If the humidity is greater than 1100, a humidity high flag is set(hh;).
If the humidity is between 1030 and 1050 millivolts a hc; flag is set for a warning.

The Quality Control (QC) procedures for the UVMRP database consist of a combination of daily automated detection and manual visual inspection of the raw data files. The visual inspection of segments of the raw and calibrated data is based upon known problems with specific instruments, error codes found during the capturing of the data, and errors due to weather-related problems, which can cause aberrant positive or negative spikes in the data. The QC codes for this type of data are either manually entered for short duration problems or via automatic batch programming for longer duration problems. QC codes are divided into three levels which indicate how serious the problem is, in order to provide automated caution statements for file transfers and web users.

Quality Control Codes

QC Code	Description
QC Code	Description
ad	UVA sensor problem, UVA data should not be used.
adc	A/D converter failure, shadowband data and auxliary data should not be used.
at	UVA dome temp is out of specifications, manufacture calibration and uva data may be compromised.
ba	band is mis aligned and is not adequately shadowing the diffuser, calibrated shadowband data should not be used, the raw total voltage is available by request
bb	suspect bias applied to diffuse channel, total and diffuse shadowband data may be compromised.
bc1	no des_factor in broadbands_calibrations for this serial_num, uvb broadband not calibrated.
bc2	no erythemal yearly annual factor found for this serial_num, uvb broadband not calibrated.
bd	Possible UVB-1 sensor problem, erythemal data may be questionable
bf	barometer failure, barometer data should not be used.
bg	UVB-1 sensor problem, significant electronic noise, erythemal data may be questionable
bm	poor band control, indicates motor problems and band is likely not shadowing the diffuser, only the total shadowband data may be used, not the diffuse or direct normal shadowband data
bt	UVB-1 sensor dome temp is out of range, effect is minimal on the erythemal data
bx	UVB-1 sensor problem, erythemal data should not be used
bz	UVB-1 sensor dome is damaged or obscured resulting in low transmission, erythemal data should not be used
cd	different shadowband calibration facility's files were used as the opening and closing calibration data.
ci	only the initial/opening shadowband calibration file used in calibrating the uv-mfrsr instrument at this site.
cn[#]	shadowband calibration file not found for this site, or this wavelength, this data is not available.
co	shadowband calibration file for the uv-mfrsr instrument at this site is over 18 months old.
coy	yankee angular correction file is being used on the uv-mfrsr instrument at this site. Only applies to shadowband data after year 2000.
cu[1..7]	the shadowband calibration file had 0 in this channel's head or board gain. This shadowband channel was not calibrated.
cy	yankee calibration file is being used on the the uv-mfrsr instrument at this site. Only applies to shadowband data after year 2000.
db[1..7]	channel failure for listed channels caused by board damage, shadowband channel data should not be used.
dc	no diffuse angular correction file found for this site, diffuse and total values do not have the diffuse correction applied but are angularly corrected.
dd	diffuser is damaged, shadowband channel voltages and lamp calibrated irradiances may be compromised.
dh[1..7]	channel problem for listed channels caused by loss of sensitivity of photodiode, lamp calibrated should not be used.
dr	diffuser is dirty, shadowband channel voltages and mlo/lamp calibrated irradiances may be compromised, please use in situ calibrated irradiances only
eca	The broadband sensor is erythemally corrected by using the constants provided by yankee. The constants are 0.141 if Solar zenith angle is <= 65 and 0.154 if > 65
ed	epply sensor problem, epply data should not be used.
ef	epply failure, epply data should not be used.
fp	power failure
f[1..7]	shadowband channel failure for listed channels, shadowband channel data should not be used for listed channels.
hc	The humidity is between 103 and 105 degrees.
hf	humidity failure, humidity data should not be used.
hh	humidity is too high. Humidity data should not be used.
hth	Head temperature is too high. Shadowband channel data may be compromised.
htl	Head temperature is too low. Shadowband channel data may be compromised.
i[1..7]	signal interference for listed channels. Shadowband channel data for listed channels should not be used.
lc[1..7]	The lamp calibrated factor is applied instead of the langley factor, due to problems with the langley factor
ld	licor problem, licor data should not be used.
lf	licor sensor failure, licor data should not be used.
lv[1..7]	low voltage, due to aging of instrument and/or location and/or time of year
nb	no bias was available to apply to diffuse channel data.
nd[1..7]	Voltage problem for listed wavelengths(1-7) due to noise. Irradiances may be compromised.
ng	Diffuse and/or direct values are negative. Please see Data Processing Procedures for possible reasons. Data could possibly be reprocessed.
nid	No head or board id found on instrument. Data cannot be corrected.
no	No nighttime bias was found for the UVB-1 sensor, effect is minimal on the erythemal data
np	no par correction data found for this site, par data is not corrected.
nsb	A bias was created and used for the nasa instrument, by putting a dome over the instrument. The nighttime values were NOT used.
pd	A par sensor problem, par sensor data should not be used.
pf	A par sensor failure, par sensor data should not be used.
pr	A par sensor problem, the par sensor was obsured. The par data should not be used.
sl[1..n]	problem with listed solar light instruments, solar light data should not be used.
sn	no angular correction file found for this site, data not angularly corrected.
so	cosine correction file for uv-mfrsr instrument is over a 18 months old.
sp[1..7]	sensitivity problem for listed channels, irradiances should not be used.
s[1..7]	signal saturated for listed channels. Shadowband channel data for listed channels may not be used.
tf	air temperature failure, temperature data should not be used.
tx	Board not synchronizing with the computer's time, timestamp and air mass may be off. If timing bias can be determined, the total horizontal and ancillary data may be usable.
ubf	UVB-1 sensor problem, erythemal data should not be used

Intercomparison platform at Mauna Loa Observatory (MLO)

Another critical issue for continued operation of the monitoring program is the accurate and frequent characterization of the radiometric, spectral, and angular responses of the instruments in a manner that is traceable to the National Institute of Standards and Technology (NIST). The Central UV Calibration Facility (CUCF), established in 1994 at the NOAA Earth System Research Laboratory in Boulder, CO, is recognized as the premier UV calibration lab by the World Meteorological Organization. To maintain scientifically-credible calibration of the instruments, UVMRP primarily uses in-situ Langley calibration techniques with an improved cloud screening module and a stable total optical depth selection module for the daily data, augmented by cycling the instruments through an intercomparison platform at the world-renown Mauna Loa Observatory (MLO), and supplemented on an as-needed basis with laboratory NIST-traceable lamp calibrations from CUCF. MLO, HI and CUCF, Boulder, CO are used by several other atmospheric monitoring organizations for similar calibration purposes. The original project instrumentation deployed at MLO in November 1997 as part of the climatological network has now become part of a suite of reference instruments. In December 2009 mounting pads and associated dataloggers were installed for additional reference instruments and for field instruments that are cycling through for periodic re-calibration.

An enhanced quality control procedure will be implemented which combines the new cloud screening algorithm and radiative transfer modeling. Basically, the new day’s data from each site will be scanned for cloud free periods and the MODTRAN5 radiative transfer model will be used to calculate clear sky, filter function weighted irradiances. The process will be seasonally and geographically adjusted to take into consideration the site elevation, solar geometry and climatological aerosol optical loading.

An analysis program has been implemented for UV-MFRSR and UVB-1 broadband instruments that have recently been calibrated at MLO. The purpose is to confirm the quality of the Langley calibration process which is used to calibrate UV-MFRSR heads at MLO, and the calibration program that is used to calibrate UVB broadband instruments deployed concurrently with the MFRSRs. The technique uses the synthetic spectrum algorithm applied to the UV MFRSR measurements to obtain erythemal irradiances averaged over the calibration deployment period (usually about a month) and compares these averages with the UVB broadband instruments’ measurements averaged over the same period. In an ideal world all the mean values should agree. The extent to which the means show disagreement serves as an additional quality control measure for our data collection network and avoids complications that arise in the rare case of a nonconforming instrument being deployed to a field site.