SGP/MWR/C1 - IRT not insulated

The downwelling IRT was insufficiently insulated to maintain an internal 
reference temperature above 0 degrees C (P971213.1). Measurements of 
sky temperature were over-estimated when instrument was below freezing.

• Sky Infra-Red Temperature(sky_ir_temp)

• Sky Infra-Red Temperature(sky_ir_temp)

SGP/MWR/C1 - Missing data

Data are missing and unrecoverable.

• 31.4 GHz sky signal(sky31)
• Water on Teflon window (1=WET, 0=DRY)(wet_window)
• Dummy altitude for Zeb(alt)
• (tknd)
• 31.4 GHz blac2body+noise injection signal(bbn31)
• Noise injection temp at 23.8 GHz adjusted to tkbb(tnd23)
• Noise injection temp at nominal temperature at 31.4 GHz(tnd_nom31)
• 23.8 GHz sky signal(sky23)
• 23.8 GHz blackbody+noise injection signal(bbn23)
• Sky brightness temperature at 31.4 GHz(tbsky31)
• lon(lon)
• Noise injection temp at 31.4 GHz adjusted to tkbb(tnd31)
• Averaged total liquid water along LOS path(liq)
• 31.4 GHz blackbody(bb31)
• Noise injection temp at nominal temperature at 23.8 GHz(tnd_nom23)
• Temperature correction coefficient at 31.4 GHz(tc31)
• Blackbody kinetic temperature(tkbb)
• Sky Infra-Red Temperature(sky_ir_temp)
• MWR column precipitable water vapor(vap)
• Ambient temperature(tkair)
• 23.8 GHz Blackbody signal(bb23)
• Sky brightness temperature at 23.8 GHz(tbsky23)
• Mixer kinetic (physical) temperature(tkxc)
• Temperature correction coefficient at 23.8 GHz(tc23)
• lat(lat)
• Time offset of tweaks from base_time(time_offset)
• base time(base_time)

SGP/MWR/C1 - Loss of thermal stabilization

Periodically during August and September 1995 all microwave radiometers
at the SGP CART generated error messages in the Site Operations Log
like:

  Time: Sat Aug 19 18:41:20 1995
  MWRLOS.C1, tkxc: Value above Maximum.

This indicates that the temperature of the microwave hardware
(specifically, the cross-coupler or "xc") exceeded its controlled
temperature (47-52 deg C) at which point it was no longer thermally
stabilized and the gain was uncontrolled.  During these periods which
typically last about 8 hours from about local noon until about sunset
the data behave anomalously and should be considered invalid.

Specifically the precipitable water vapor increases and the liquid
water path decreases (and becomes SIGNIFICANTLY NEGATIVE (-0.1 mm) on
clear sky days).  The RMS noise level in the data also increases
sharply.  The 'Tkxc' field appears ONLY in the a0-level data and does
NOT appear in either the a1 (mwrlos) or c1 (mwr5avg) files.  Therefore
THESE ANOMALOUS VALUES HAVE BEEN INCLUDED IN THE 5-MINUTE AVERAGES.

The microwave hardware is thermally stabilized to about +/- 0.1 deg C
by resistive heating.  When the internal temperature rises above the
set point the thermal stabilization of the instrument gain is lost.
>From an examination of the component temperature data it appears that
increasing the set point temperature to about 55 deg C (328 K) would
prevent a re-ocurrance of this problem at the SGP.  The manufacturer,
Radiometrics, concurs that raising the set point will fix this problem
and will not cause other problems.

I will have to carefully examine the MCTEX data to determine whether
this will be a problem for the TWP.  The manufacturer and I had
discussed this possibility prior to building the TWP MWRs (S/N 015,
016, and 017) and those instruments have set points above 50 deg C.
Note that MWR 018 has a set point near 52 deg C (like the TWP models)
but it still experienced a few loss-of-stabilization events.

Note that the instruments with the lowest set points had the most
loss-of-stabilization events.

• 31.4 GHz sky brightness temperature(31tbsky)
• Averaged total liquid water along LOS path(liq)
• MWR column precipitable water vapor(vap)
• 23.8 GHz sky brightness temperature(23tbsky)

• Sky brightness temperature at 31.4 GHz(tbsky31)
• Averaged total liquid water along LOS path(liq)
• MWR column precipitable water vapor(vap)
• Sky brightness temperature at 23.8 GHz(tbsky23)

• 23.8 GHz sky brightness temperature(23tbsky)
• Averaged total liquid water along LOS path(liq)
• MWR column precipitable water vapor(vap)
• 31.4 GHz sky brightness temperature(31tbsky)

• MWR column precipitable water vapor(vap)
• Sky brightness temperature at 23.8 GHz(tbsky23)
• Sky brightness temperature at 31.4 GHz(tbsky31)
• Averaged total liquid water along LOS path(liq)

SGP/MWR/C1 - Radio Frequency Interference during IOP

During the specified times a strong, continuous signal was 
measured by the 31.4 GHz of the MWR.  The signal was present
in all 31.4 GHz measurements including the sky measurement,
the internal reference target measurement, and the measurement
of the internal noise injection source from which the
instantaneous instrument gain is computed.

The source of the interference has not yet been identified.

Because the gain is computed using the difference of the
noise injection and target measurements, and because the
sky brightness temperature is computed relative to the
internal target temperature, the data appear anomalous
only for a period of an hour after the interference starts
and ends.  This is due to the low pass filter applied to
the instantaneous gain.  However the data should be
considered invalid or at least questionable during the
entire period for which the interference was present.

• Flag indicating where the initial surface water measurements are from: 0-> SMOS,
  1-> AERI(water_flag)
• Fraction of data in averaging interval with water on Teflon window(water_flag_fraction)
• 23.8 GHz sky brightness temperature(23tbsky)
• Standard deviation about the mean for the IR brightness temperature(ir_temp_sdev)
• Number of data points averaged out of 15(number_obs_averaged)
• base time(base_time)
• Standard deviation about the mean for the 31.4 GHz sky brightness temperature(tbsky31_sdev)
• Standard deviation about the mean for the total water vapor amount(vap_sdev)
• Averaged total liquid water along LOS path(liq)
• IR Brightness Temperature(ir_temp)
• Probability of level change in ratio of averaged brightness temps(prob_level_change)
• Standard deviation about the mean for the 23.8 GHz sky brightness temperature(tbsky23_sdev)
• lon(lon)
• Probability of slope change in ratio of averaged brightness temps(prob_slope_change)
• Number of data points averaged for 23tbsky, 31tbsky, vap & liq(num_obs)
• Probability of outlier in ratio of averaged brightness temps(prob_outlier)
• Sky brightness temperature at 23.8 GHz(tbsky23)
• Sky brightness temperature at 31.4 GHz(tbsky31)
• Number of points included in the ir_temp ensemble(num_obs_irt)
• 31.4 GHz sky brightness temperature(31tbsky)
• Standard deviation about the mean for the total liquid water amount(liq_sdev)
• Dummy altitude for Zeb(alt)
• MWR column precipitable water vapor(vap)
• Time offset of tweaks from base_time(time_offset)
• lat(lat)

• 31.4 GHz sky signal(sky31)
• Water on Teflon window (1=WET, 0=DRY)(wet_window)
• Dummy altitude for Zeb(alt)
• 31.4 GHz blac2body+noise injection signal(bbn31)
• Noise injection temp at nominal temperature at 31.4 GHz(tnd_nom31)
• 23.8 GHz sky signal(sky23)
• Sky brightness temperature at 31.4 GHz(tbsky31)
• lon(lon)
• Noise injection temp at 31.4 GHz adjusted to tkbb(tnd31)
• Averaged total liquid water along LOS path(liq)
• Noise injection temp at nominal temperature at 23.8 GHz(tnd_nom23)
• 31.4 GHz blackbody(bb31)
• Temperature correction coefficient at 31.4 GHz(tc31)
• Blackbody kinetic temperature(tkbb)
• Sky Infra-Red Temperature(sky_ir_temp)
• MWR column precipitable water vapor(vap)
• Ambient temperature(tkair)
• 23.8 GHz Blackbody signal(bb23)
• Sky brightness temperature at 23.8 GHz(tbsky23)
• Mixer kinetic (physical) temperature(tkxc)
• Time offset of tweaks from base_time(time_offset)
• base time(base_time)
• (tknd)
• Noise injection temp at 23.8 GHz adjusted to tkbb(tnd23)
• 23.8 GHz blackbody+noise injection signal(bbn23)
• Temperature correction coefficient at 23.8 GHz(tc23)
• lat(lat)

• 31.4 GHz blac2body+noise injection signal(bbn31)
• Noise injection temp at 31.4 GHz adjusted to tkbb(tnd31)
• Sky brightness temperature at 31.4 GHz(tbsky31)
• 23.8 GHz sky signal(sky23)
• Temperature correction coefficient at 23.8 GHz(tc23)
• MWR column precipitable water vapor(vap)
• (tknd)
• Sky brightness temperature at 23.8 GHz(tbsky23)
• Ambient temperature(tkair)
• Noise injection temp at nominal temperature at 23.8 GHz(tnd_nom23)
• 23.8 GHz Blackbody signal(bb23)
• Mixer kinetic (physical) temperature(tkxc)
• Water on Teflon window (1=WET, 0=DRY)(wet_window)
• 31.4 GHz sky signal(sky31)
• Dummy altitude for Zeb(alt)
• lon(lon)
• Temperature correction coefficient at 31.4 GHz(tc31)
• 31.4 GHz blackbody(bb31)
• Blackbody kinetic temperature(tkbb)
• lat(lat)
• Sky Infra-Red Temperature(sky_ir_temp)
• Noise injection temp at 23.8 GHz adjusted to tkbb(tnd23)
• Noise injection temp at nominal temperature at 31.4 GHz(tnd_nom31)
• Averaged total liquid water along LOS path(liq)
• 23.8 GHz blackbody+noise injection signal(bbn23)
• base time(base_time)
• Time offset of tweaks from base_time(time_offset)

SGP/MWR/C1 - IRT lens replaced

The lens of the uplooking "IR thermometer" (a 10 micrometer pyrometer)
was replaced on 28 October 1996 at 16:18 GMT.  At the time the sky was
heavily overcast (0.1-0.5 mm liquid water path, IR temperature =
281-285 K) with about 4 cm integrated water vapor.  A comparison was
carried out against the Atmospheric Emitted Radiance Interferometer
(AERI) both prior to and subsequent to the lens change.  The statistics
of the differences (IRT-AERI) are as follows:

         Mean (K)  Std Dev (K)   No.
Before:    0.64      0.24       122
After:     0.35      0.20        89

In addition, the downward step change in the time series plot of the
temperature difference (IRT-AERI) at the time of the lens replacement
is obvious.

This suggests, but is not conclusive evidence, that the primary cause
of the differences between the IRT and AERI reported previously
(PIF no. P960809.2) may have been the weather-worn lens.  However,
as this instrument has not been calibrated since it was deployed in
December 1993, there may also be calibration drift to contend with.

Arrangements are being pursued with NREL to check the calibration of
this instrument.

It is preferable to carry out the IRT vs AERI comparison for clear sky
conditions to be certain that both instruments are observing the same
scene at the moment when they make sky measurements.  However, it is
not possible at this time of the year to carry out a clear sky
comparison because the integrated water vapor is generally less than
2.0 cm which means that the IR temperature is less than 223 K, the
lower limit of the D/A converter on the IRT.

On the positive side, the AERI and the IRT are within 10 m of each
other and have similar fields of view (about 2 degrees) so their
scenes should be comparable for clear or cloudy skies.

• Sky Infra-Red Temperature(sky_ir_temp)

• Sky Infra-Red Temperature(sky_ir_temp)

SGP/MWR/B1/B4/B5/B6/C1 - Thermal Stabilization Adjustment

In order to correct a thermal stabilization problem identified earlier
I adjusted the thermal set point of the microwave radiometers at the
SGP upward from 48-50 deg C to 55 deg C in early August 1996 according
to the schedule given below.

B6    5 August 1996
C1    6 August 1996
B1    7 August 1996
B5    8 August 1996

Subsequent to making this adjustment the MWRs were put in TIP mode to
check on whether the change in set point temperature affected their
calibration.  Because clear sky conditions were quite intermittent, it
is difficult to determine whether the substantial variability in the
tip data were attributable to the change in thermal set point.  The
instrument calibration was not altered in August.

Tip data were again collected with these instruments in September prior
to the beginning and at the close of the Water Vapor IOP.  For example,
the calibration of the instrument at the central facility (S/N 10)
derived from the September data was essentially the same as that
derived from calibration data acquired in February 1996.  Although this
would lead one to believe that altering the thermal set point did not
affect the instrument calibration, it may be that some transient effect
was induced.

In comparing soundings launched from the central facility with the
microwave radiometer there, I noticed that those sondes calibrated in
June 1996 consistently reported lower integrated water vapor than the
radiometer in July and September (during the IOP) but were in better
agreement with the radiometer for the two weeks period immediately
after the set point was adjusted.  I suspect that adjusting the thermal
set point may have temporarily increased the radiometer gain
(kelvins/volt) thereby lowering the measured brightness temperature and
the retrieved integrated water vapor.

It is not clear why a temporary change in gain should occur or even
whether it did.  But users of the data should be aware that the data
from the microwave radiometers at the SGP may be anomalous during
August 1996.

• MWR column precipitable water vapor(vap)
• 31.4 GHz sky brightness temperature(31tbsky)
• Averaged total liquid water along LOS path(liq)
• 23.8 GHz sky brightness temperature(23tbsky)

• 23.8 GHz sky brightness temperature(23tbsky)
• Averaged total liquid water along LOS path(liq)
• 31.4 GHz sky brightness temperature(31tbsky)
• MWR column precipitable water vapor(vap)

• 23.8 GHz sky brightness temperature(23tbsky)
• 31.4 GHz sky brightness temperature(31tbsky)
• Averaged total liquid water along LOS path(liq)
• MWR column precipitable water vapor(vap)

• 23.8 GHz sky brightness temperature(23tbsky)
• Averaged total liquid water along LOS path(liq)
• MWR column precipitable water vapor(vap)
• 31.4 GHz sky brightness temperature(31tbsky)

• Averaged total liquid water along LOS path(liq)
• 23.8 GHz sky brightness temperature(23tbsky)
• MWR column precipitable water vapor(vap)
• 31.4 GHz sky brightness temperature(31tbsky)

• Sky brightness temperature at 31.4 GHz(tbsky31)
• Averaged total liquid water along LOS path(liq)
• MWR column precipitable water vapor(vap)
• Sky brightness temperature at 23.8 GHz(tbsky23)

• Averaged total liquid water along LOS path(liq)
• 23.8 GHz sky brightness temperature(23tbsky)
• MWR column precipitable water vapor(vap)
• 31.4 GHz sky brightness temperature(31tbsky)

• 23.8 GHz sky brightness temperature(23tbsky)
• MWR column precipitable water vapor(vap)
• 31.4 GHz sky brightness temperature(31tbsky)
• Averaged total liquid water along LOS path(liq)

• Sky brightness temperature at 31.4 GHz(tbsky31)
• MWR column precipitable water vapor(vap)
• Sky brightness temperature at 23.8 GHz(tbsky23)
• Averaged total liquid water along LOS path(liq)

• Averaged total liquid water along LOS path(liq)
• 31.4 GHz sky brightness temperature(31tbsky)
• 23.8 GHz sky brightness temperature(23tbsky)
• MWR column precipitable water vapor(vap)

• MWR column precipitable water vapor(vap)
• 31.4 GHz sky brightness temperature(31tbsky)
• 23.8 GHz sky brightness temperature(23tbsky)
• Averaged total liquid water along LOS path(liq)

• MWR column precipitable water vapor(vap)
• 31.4 GHz sky brightness temperature(31tbsky)
• 23.8 GHz sky brightness temperature(23tbsky)
• Averaged total liquid water along LOS path(liq)

• Averaged total liquid water along LOS path(liq)
• 31.4 GHz sky brightness temperature(31tbsky)
• 23.8 GHz sky brightness temperature(23tbsky)
• MWR column precipitable water vapor(vap)

• 23.8 GHz sky brightness temperature(23tbsky)
• Averaged total liquid water along LOS path(liq)
• 31.4 GHz sky brightness temperature(31tbsky)
• MWR column precipitable water vapor(vap)

• MWR column precipitable water vapor(vap)
• Averaged total liquid water along LOS path(liq)
• 23.8 GHz sky brightness temperature(23tbsky)
• 31.4 GHz sky brightness temperature(31tbsky)

• Averaged total liquid water along LOS path(liq)
• MWR column precipitable water vapor(vap)
• 31.4 GHz sky brightness temperature(31tbsky)
• 23.8 GHz sky brightness temperature(23tbsky)

SGP/MWR/C1 - IRT Calibration check

IRT was out of service for calibration check.  MWR was powered
down on 96/12/03 1937-1949 to remove IRT and on 96/12/12 1848-1920
to re-install IRT.

The following is from the NREL Metrology Laboratory test report of 12/10/97:

Temperature     Nominal Value     Measured Value
-----------     -------------     --------------
  0 C             273.2 K           274.1 K
 10               283.2             283.6
 20               293.2             293.6
 30               303.2             303.2
 40               313.2             313.2

In the range tested, the temperature difference is within the
resolution of the instrument, so the IRT was not adjusted and
the calibration factor was not changed.

• Sky Infra-Red Temperature(sky_ir_temp)

• Sky Infra-Red Temperature(sky_ir_temp)

SGP/MWR/B1/B4/B6/C1 - software change

The MWR operating software was changed on 12 October 1998 to provide additional 
functionality as described below. This change affects the format of the raw and ingested data. 
   
NEW FEATURES
1. Faster sampling rate
   
Standard line-of-sight (LOS) observations can now be acquired at 15-second intervals vs. 
20-second intervals previously. (The standard LOS cycle is comprised of one sky sample per 
blackbody sample and gain update.)
   
2. More flexible sampling strategy
   
Multiple sky observations can be acquired during a LOS cycle, up to 1024 per gain update. 
This permits sky samples to be acquired at intervals of 2.67 seconds for improved 
temporal resolution of cloud liquid water variations and better coordination with the millimeter 
cloud radar during IOPs.
   
3. Separation of zenith LOS observations from TIP data
   
When the radiometer is in TIP mode, the zenith LOS observations are now extracted, the PWV 
and LWP computed and reported separately in the output file. This eliminates the periods 
of missing LOS data during calibration checks/updates.
   
4. Automatic self-calibration
   
The software now permits the calibration to be updated at specified intervals or 
continuously. In the first case, LOS mode is automatically changed to TIP mode at user-specified 
intervals or whenever clear sky conditions occur, the tip data reduced, the calibration 
updated ,and the radiometer returned to LOS mode without operator intervention. In the 
second case, the radiometer is continuously is TIP mode until changed by the operator.
   
5. Graphical user display
   
The graphical display is comprised of a status display, a message display, a temperature 
plot, a plot of the retrieved PWV and LWP, and (in TIP mode) a plot of the latest tip 
curves.

Editor's Note: The SGP.C1 data were reprocessed in 2004 and enhancement #3 described above 
was applied to the data prior to Oct 1998.  The SGP.BF data are queued for reprocessing 
as well.

• Noise injection temp at 31.4 GHz derived from this tip(tnd31I)
• Noise injection temp at 23.8 GHz derived from this tip(tnd23I)

• Noise injection temp at 23.8 GHz adjusted to tkbb(tnd23)
• Ambient temperature(tkair)
• Noise injection temp at 31.4 GHz adjusted to tkbb(tnd31)
• (tknd)

• Noise injection temp at 31.4 GHz adjusted to tkbb(tnd31)
• Noise injection temp at 23.8 GHz adjusted to tkbb(tnd23)
• (tknd)
• Ambient temperature(tkair)

• (tknd)
• Ambient temperature(tkair)
• Noise injection temp at 31.4 GHz adjusted to tkbb(tnd31)
• Noise injection temp at 23.8 GHz adjusted to tkbb(tnd23)

• (tknd)
• Noise injection temp at 23.8 GHz adjusted to tkbb(tnd23)
• Noise injection temp at 31.4 GHz adjusted to tkbb(tnd31)
• Ambient temperature(tkair)

• Noise injection temp at 31.4 GHz derived from this tip(tnd31I)
• Noise injection temp at 23.8 GHz derived from this tip(tnd23I)

• (tknd)
• Noise injection temp at 23.8 GHz adjusted to tkbb(tnd23)
• Noise injection temp at 31.4 GHz adjusted to tkbb(tnd31)
• Ambient temperature(tkair)

• Noise injection temp at 31.4 GHz derived from this tip(tnd31I)
• Noise injection temp at 23.8 GHz derived from this tip(tnd23I)

• Noise injection temp at 31.4 GHz adjusted to tkbb(tnd31)
• Noise injection temp at 23.8 GHz adjusted to tkbb(tnd23)
• (tknd)
• Ambient temperature(tkair)

• Noise injection temp at 31.4 GHz adjusted to tkbb(tnd31)
• Ambient temperature(tkair)
• Noise injection temp at 23.8 GHz adjusted to tkbb(tnd23)
• (tknd)

• Ambient temperature(tkair)
• Noise injection temp at 23.8 GHz adjusted to tkbb(tnd23)
• (tknd)
• Noise injection temp at 31.4 GHz adjusted to tkbb(tnd31)

SGP/MWR/B1/B4/B5/B6/C1 - software upgrade (version 3.27)

At 00:00 GMT on 7 January version 3.27 of the MWR operating program was installed and made 
operational at the SGP central facility (C1).  No problems were noted over the next few 
days and the boundary facility MWRs (B1, B4, B5, B6) were upgraded at 20:00 GMT on 11 
January.  This version includes a beam width correction I developed as well as providing the 
capability to automatically level the elevation mirror (that is, to automatically detect 
and correct offsets in the elevation angle stepper motor position.)

On 12 January I discovered that the '486-based MWR computers at B1, B4 and B6 were not 
executing the system command to move and rename the data files so that the ARM data system 
could retrieve them.  Reducing the length of the storage arrays in the auto-leveling 
feature from 1000 to 250 resolved the problem.  This results in the auto-leveling being based 
on only 4 hours of clear sky data rather than 16 hours at B5 and C1.  This version of the 
program is 3.28. Version 3.27 (running at B5 and C1) can be installed if and when these 
computers are upgraded to Pentium-class machines.

The improvement in the quality of the tip curves resulting from the auto-leveling has been 
dramatic: differences in the brightness temperatures at 3 airmasses (19.5 and 160.5 
degrees) have been reduced from +/- 5 K to +/- 0.5 K.  The goodness-of-fit coefficient for 
the tip curves has improved from about 0.995 to over 0.998.  In order to take full 
advantage of this improvement to detect and reject cloudy tip curves, the minimum value of the 
goodness-of-fit coefficient for a valid tip curve has been increased from 0.995 to 0.998.

Editor's Note: The SGP.C1 data were reprocessed in 2004 to produce a common DOD for all 
time.  The 1996-1998 data reprocessing included beam width and mirror-leveling corrections, 
but the data prior to that range did not have these corrections applied.

• 31.4 GHz sky signal(31tipsky)
• 23.8 GHz sky signal(23tipsky)

• 31.4 GHz sky brightness temperature(31tbsky)
• MWR column precipitable water vapor(vap)
• Averaged total liquid water along LOS path(liq)
• 23.8 GHz sky brightness temperature(23tbsky)

• 31.4 GHz sky signal(31tipsky)
• 23.8 GHz sky signal(23tipsky)

• 23.8 GHz sky brightness temperature(23tbsky)
• MWR column precipitable water vapor(vap)
• Averaged total liquid water along LOS path(liq)
• 31.4 GHz sky brightness temperature(31tbsky)

• Sky brightness temperature at 31.4 GHz(tbsky31)
• Averaged total liquid water along LOS path(liq)
• MWR column precipitable water vapor(vap)
• Sky brightness temperature at 23.8 GHz(tbsky23)

• 23.8 GHz sky signal(23tipsky)
• 31.4 GHz sky signal(31tipsky)

• 31.4 GHz sky signal(31tipsky)
• 23.8 GHz sky signal(23tipsky)

• Sky brightness temperature at 31.4 GHz(tbsky31)
• MWR column precipitable water vapor(vap)
• Sky brightness temperature at 23.8 GHz(tbsky23)
• Averaged total liquid water along LOS path(liq)

• 31.4 GHz sky brightness temperature(31tbsky)
• Averaged total liquid water along LOS path(liq)
• 23.8 GHz sky brightness temperature(23tbsky)
• MWR column precipitable water vapor(vap)

• Averaged total liquid water along LOS path(liq)
• 31.4 GHz sky brightness temperature(31tbsky)
• 23.8 GHz sky brightness temperature(23tbsky)
• MWR column precipitable water vapor(vap)