SGP/EBBR/E12 - 3-month Check

During EBBR 3 month checks the Automatic Exchange Mechanism is disabled
(reflected by home signals near zero) so that the two aspirator units can be
placed at the same height.  The measurements during this period, at least,
do not produce correct outputs of bowen ratio (bowen), sensible heat flux (h),
or latent heat flux (e), nor do the air temperature and relative humidity
measurements indicate gradients.  Checks performed before or after the AEM
is disabled also lead to some incorrect data.

• temp_air_right
• temp_trh_left_rr
• rh_left_mv
• rh_right_mv
• temp_air_left
• temp_trh_right_rr

• temp_air_top
• vapor_pressure_bottom
• temp_trh_bottom
• temp_trh_top
• rh_top_fraction
• bowen_ratio
• Heat flux, latent, at 1.5-m height, 30-min intervals(latent_heat_flux)
• vapor_pressure_top
• rh_bottom_fraction
• temp_air_bottom
• Heat flux, sensible, at 1.5-m height, 30-min intervals(sensible_heat_flux)

• rh_top_fraction
• temp_trh_top
• rh_bottom_fraction
• vapor_pressure_bottom
• temp_air_bottom
• temp_air_top
• temp_trh_bottom
• vapor_pressure_top

SGP/EBBR/E12 - T/RH Check

During the month of November 1994, site operations personnel conducted
field comparisons of the temperature and relative humidity sensors in the
EBBRs and the HMI31 portable T/RH meter (with the removable probe).  As is
reflected in the home signal output, the automatic exchange mechanism was
disabled in order to conduct the checks.

Measurements were taken every 5 minutes, for a half hour each in the
aspirated radiation shields (the HMI31 was inserted into the EBBR aspirated
radiation shield to provide the same aspiration as the EBBR sensors
received).  Measurements of aspirator flow were also recorded.  This
information was provided to the mentor.  The mentor then compared the field
checks with the data recorded by the EBBRs.

The periods of data and data fields that are incorrect because of the checks
are listed below.  Times are those at which the data are reported.


Data Fields
5 minute:  tair_top, tair_bot, thum_top, thum_bot, hum_top, hum_bot,
           vp_top, vp_bot
15 minute: rr_thum_r, rr_thum_l, mv_hum_r, mv_hum_l, tair_r, tair_l
30 minute: tair_top, tair_bot, thum_top, thum_bot, hum_top, hum_bot,
           vp_top, vp_bot, bowen, e, h

5 minute:  1515, 1520, 1525, 1530, 1535, 1540, 1545, 1550, 1555,
           1600, 1605, 1610
15 minute: 1515, 1530, 1545, 1600, 1615
30 minute: 1530, 1600, 1630

(EDITOR'S NOTE:  The following analysis refers to all EBBRs installed in Nov 1994)

Results of the T/RH comparisons:
The combined accuracy of the HMI31 and EBBR temperature sensors
(thermocouples or PRTDs) is +/- 0.4 degrees C.  The combined accuracy of the
HMI31 and EBBR RH sensors is +/- 4% for RH less than 80%, and +/- 6% for RH
greater than 80%.  The combined accuracy for each pair of EBBR sensors of
the same type is essentially the same as the quantities stated above.

All of the pairs of EBBR sensors show differences that fall within the
stated combined accuracies.  Although this sounds wonderful, a few of the
differences are close to the combined accuracy, which, because of the small
differences in temperature being measured (normally smaller than the
magnitude of the combined accuracy) could lead to some incorrect estimation
of sensible and latent heat flux.  However, unless there would be substantial
differences of calibration slope of two sensors in a pair, the switching of
height of the sensors every 15 minutes by the automatic exchange mechanism
will remove the offset between the sensors.

The comparisons between the EBBR sensors and the HMI31 T/RH meter showed
much larger differences.  The range of differences were: -2.4 to 0.0 degrees
for the thermocouples, -2.1 to 0.1 degrees for the RH probe PRTDs, and -6.6 to
6.7 percent for the RH sensors.  Six Thermocouples indicated differences from
the HMI31 greater than the combined accuracy.  Seven PRTDs indicated
differences from the HMI31 greater than the combined accuracy.  Four RH
sensors indicated differences from the HMI31 greater than the combined
accuracy.  The temperature differences (for both thermocouples and PRTDs)
were nearly all of one sign; the HMI31 indicated the greater temperature,
which may suggest a bias in the HMI31 temperature measurement, even though
a dozen differences of +/- 0.2 degrees were measured.  Four differences of
zero and only one positive difference (out of 40) were measured.  Thirteen of
the 20 RH differences were positive (HMI31 indicating the lower
RH), with the differences near the level of the combined accuracies being
approximately 50% positive and 50% negative.  It does not appear that the
HMI31 has a RH bias.

Plots of EBBR sensor/HMI31 differences as a function of time from the
installation of the individual sensors (a range of 10 to 30 months in the
field) shows no obvious pattern of sensor drift with time.  This is encouraging.

• temp_air_right
• temp_trh_left_rr
• rh_left_mv
• rh_right_mv
• temp_air_left
• temp_trh_right_rr

• temp_air_top
• vapor_pressure_bottom
• temp_trh_bottom
• temp_trh_top
• rh_top_fraction
• bowen_ratio
• Heat flux, latent, at 1.5-m height, 30-min intervals(latent_heat_flux)
• vapor_pressure_top
• rh_bottom_fraction
• temp_air_bottom
• Heat flux, sensible, at 1.5-m height, 30-min intervals(sensible_heat_flux)

• rh_top_fraction
• temp_trh_top
• rh_bottom_fraction
• vapor_pressure_bottom
• temp_air_bottom
• temp_air_top
• temp_trh_bottom
• vapor_pressure_top

SGP/EBBR/E12 - Invalid data due to blown AEM fuse at E12

On December 20, 1993, before 0730 GMT, the fuse in the AEM (Automatic
Exchange Mechanism) of Pawhuska, OK (E12) EBBR blew out.  The AEM remained
nonfunctional until the fuse was replaced just before 1500 GMT on
December 22, 1993.  This situation resulted in the following invalid data:

5 minute: tair_top, tair_bot, thum_top, thum_bot, hum_top, hum_bot,
          vp_top, vp_bot, home

15 minute: rr_thum_r, rr_thum_l, mv_hum_r, mv_hum_l, mv_home, tair_r,
           tair_l

30 minute: tair_top, tair_bot, thum_top, thum_bot, hum_top, hum_bot,
           vp_top, vp_bot, bowen, e, h, home_15, home_30

Bowen, h, and e can not be recalculated from the 15 minute and/or 5 minute data
since we cannot determine the position of the sensors on the AEM from the home
signal; home signal is near zero when there is no power to the AEM.

• temp_air_right
• home_signal
• temp_trh_left_rr
• rh_left_mv
• rh_right_mv
• temp_air_left
• temp_trh_right_rr

• temp_air_top
• vapor_pressure_bottom
• temp_trh_bottom
• temp_trh_top
• rh_top_fraction
• home_signal_15
• bowen_ratio
• vapor_pressure_top
• Heat flux, latent, at 1.5-m height, 30-min intervals(latent_heat_flux)
• home_signal_30
• rh_bottom_fraction
• temp_air_bottom
• Heat flux, sensible, at 1.5-m height, 30-min intervals(sensible_heat_flux)

• home_signal
• rh_top_fraction
• temp_trh_top
• rh_bottom_fraction
• vapor_pressure_bottom
• temp_air_bottom
• temp_air_top
• temp_trh_bottom
• vapor_pressure_top

SGP/EBBR/E12 - Wind Direction

For much of the period above, a malfunction in the E12 EBBR wind direction
sensor (insect larvae shorting out the circuit board) caused incorrect
values for 5 minute res_ws, wind_d, and sigma_wd, 15 minute mv_wind_d, and
30 minute res_ws, wind_d, and sigma_wd.  Wind direction normally indicated
north (near 0 or 360 degrees); there were some times during the period when
wind directions looked good, but the lack of variation in direction and sigma
implies that the data is suspect.  Therefore, it would be wise to flag all
of the values listed above during the entire time period as incorrect.
The wind direction sensor (S/N 3081) was replaced with S/N 3042 at 1515 GMT
on June 8, 1994.

• Wind speed vector mean(wspd_vec_mean)
• Wind direction vector mean(wdir_vec_mean)
• wdir_vec_std

• wdir_vec_mean_mv

• wdir_vec_std
• Wind speed vector mean(wspd_vec_mean)
• Wind direction vector mean(wdir_vec_mean)

SGP/EBBR/E12 - Aspirator Failure

On June 18, 1994 the aspirator fan in the EBBR at E12, Pawhuska stopped
working.  This situation was not discovered until a Preventative
Maintenance visit by site operations personnel on June 22, 1994.  I
determined the time of the beginning of the malfunction from close
inspection of the sensible and latent heat flux data.  There is no aspiration
status output for the EBBRs, nor is one presently possible.  Wind speeds during
most of the period listed were sufficient to provide some (although
inadequate) aspiration to the temperature and relative humidity probes, making
it difficult for most users of the data to be able to tell (except in a few
circumstances) that the sensible and latent heat flux values produced were
incorrect.  All of the data fields listed below are incorrect for the period
listed above.  Latent and Sensible heat fluxes cannot be recomputed, as the
measurements that they are based on are what was corrupted.

30 minute: tair_top, tair_bot, thum_top, thum_bot, hum_top, hum_bot,
           vp_top, vp_bot, bowen, e, h

15 minute: rr_thum_r, rr_thum_l, mv_hum_r, mv_hum_l, tair_r, tair_l

5 minute:  tair_top, tair_bot, thum_top, thum_bot, hum_top, hum_bot,
           vp_top, vp_bot

• temp_air_right
• temp_trh_left_rr
• rh_left_mv
• rh_right_mv
• temp_air_left
• temp_trh_right_rr

• temp_air_top
• vapor_pressure_bottom
• temp_trh_bottom
• temp_trh_top
• rh_top_fraction
• bowen_ratio
• Heat flux, latent, at 1.5-m height, 30-min intervals(latent_heat_flux)
• vapor_pressure_top
• rh_bottom_fraction
• temp_air_bottom
• Heat flux, sensible, at 1.5-m height, 30-min intervals(sensible_heat_flux)

• rh_top_fraction
• temp_trh_top
• rh_bottom_fraction
• vapor_pressure_bottom
• temp_air_bottom
• temp_air_top
• temp_trh_bottom
• vapor_pressure_top

SGP/EBBR/E12 - Stopped Ventilation Fan

On August 4, 1994, a site operator discovered the ventilation fan of the
EBBR to be stopped; the fan guard was clogged with animal remains.
Inspection of the EBBR data from Pawhuska, OK, E12 indicates that this
condition occurred from July 22 through August 4.  Lack of ventilation of
the temperature and relative humidity sensors attached to the AEM results
in questionable measurements of air temperature, relative humidity, and all
quantities calculated from them, including sensible and latent heat fluxes.
The quantities that are questionable for the period indicated are:

5 minute:  tair_top, tair_bot, thum_top, thum_bot, hum_top, hum_bot,
           vp_top, and vp_bot

15 minute: rr_thum_r, rr_thum_l, mv_hum_r, mv_hum_l, tair_r, tair_l

30 minute: tair_top, tair_bot, thum_top, thum_bot, hum_top, hum_bot,
           vp_top, vp_bot, bowen, e, h

• temp_air_right
• temp_trh_left_rr
• rh_left_mv
• rh_right_mv
• temp_air_left
• temp_trh_right_rr

• temp_air_top
• vapor_pressure_bottom
• temp_trh_bottom
• temp_trh_top
• rh_top_fraction
• bowen_ratio
• Heat flux, latent, at 1.5-m height, 30-min intervals(latent_heat_flux)
• vapor_pressure_top
• rh_bottom_fraction
• temp_air_bottom
• Heat flux, sensible, at 1.5-m height, 30-min intervals(sensible_heat_flux)

• rh_top_fraction
• temp_trh_top
• rh_bottom_fraction
• vapor_pressure_bottom
• temp_air_bottom
• temp_air_top
• temp_trh_bottom
• vapor_pressure_top

SGP/EBBR/E12 - Miscellaneous Problems Affecting Data

Three hardware problems occurred at E12, Pawhuska, OK in late August 1994.

On August 12 the wind direction sensor malfunctioned.  Therefore, all wind
direction, sigma, and resultant wind speed values for the period above are
incorrect.  The wind direction sensor was replaced on August 31.  Affected
data are:

        5 minute:  wind_d, sigma_wd, res_ws
        15 minute: mv_wind_d
        30 minute: wind_d, sigma_wd, res_ws.

Water accumulated in the AEM motor box and apparently caused a reduction in
the voltage level of the home signal (battery condition was very good and so
was not the cause); by late August 27 the hom30 level had fallen below the
minimum qc check level.  The AEM continued to exchange properly.  The minimum
qc level check is intended to help detect either a non-operating AEM or low
battery condition.  The Campbell CR10 software operates properly unless the
hom30 signal is thirty-five or greater.  So the low hom30 values do not
result in any incorrect data!!!.  I point this out because this situation
demonstrates a condition flagged by the qc checks, even though no effect on
the data has occurred.  The AEM box was dried out and new dessicant placed
in it on August 31.

On August 28, 1994 the left AEM housing ventilation fan ceased to operate
because of clogging of the ventilation system with debris.  Fifteen minute
data seems to indicate that the fan stopped operating just before 1245 GMT.
This is at least the third such incidence of clogging this summer at different
EBBRs.  The clog was removed and ventilation restored on August 31.  The lack
of ventilation caused improper measurements of temperature and
relative humidity in the left housing (these are quite obvious in the 15
minute data but difficult to detect in the 30 minute data) and therefore
incorrect values of sensible and latent heat.  Affected data are:

        5 minute:  tair_top, tair_bot, thum_top, thum_bot, hum_top,
                   hum_bot, vp_top, vp_bot
        15 minute: rr_thum_l, mv_hum_l, tair_l
        30 minute: tair_top, tair_bot, thum_top, thum_bot, hum_top,
                   hum_bot, vp_top, vp_bot, bowen, e, h.

• temp_air_top
• vapor_pressure_bottom
• temp_trh_bottom
• temp_trh_top
• Wind speed vector mean(wspd_vec_mean)
• rh_top_fraction
• bowen_ratio
• Wind direction vector mean(wdir_vec_mean)
• vapor_pressure_top
• Heat flux, latent, at 1.5-m height, 30-min intervals(latent_heat_flux)
• home_signal_30
• rh_bottom_fraction
• wdir_vec_std
• temp_air_bottom
• Heat flux, sensible, at 1.5-m height, 30-min intervals(sensible_heat_flux)

• wdir_vec_mean_mv
• home_signal
• temp_trh_left_rr
• rh_left_mv
• temp_air_left

• rh_top_fraction
• wdir_vec_std
• temp_trh_top
• rh_bottom_fraction
• vapor_pressure_bottom
• temp_air_bottom
• temp_air_top
• temp_trh_bottom
• Wind speed vector mean(wspd_vec_mean)
• vapor_pressure_top
• Wind direction vector mean(wdir_vec_mean)

SGP/EBBR/E12 - Net Radiometer Hemisphere Changes

The upper hemisphere of the net radiometer at the site was punctured
allowing water to collect in the lower (and sometimes upper) section of the sensor. The 
punctures are most likely the result of migratory birds (a problem that we have experienced 
almost every year here at ANL also). The punctures (and water in some net radiometers) 
did not result in obvious invalidation of the data.

Whether there is water in the lower hemisphere or not, a small puncture of the upper 
hemisphere would not have a major effect on the net radiation measurement.  Significant 
amounts of water in the lower
hemisphere should have a major effect on nighttime measurements particularly, although 
that effect is hard to detect in the data. 
The date when the puncture occurred is approximate - but is certainly 
no more than two weeks before the hemisphere was replaced.

• Heat flux, latent, at 1.5-m height, 30-min intervals(latent_heat_flux)
• Heat flux, sensible, at 1.5-m height, 30-min intervals(sensible_heat_flux)
• Radiation, net(net_radiation)

• net_radiation_mv

• Radiation, net(net_radiation)

SGP/EBBR/E12 - Low home_15 Values

The home signal values for the EBBR system installed at E12, Pawhuska, OK have
generally been lower than for the other EBBR systems.  However, during the
period above the home_15 value was often lower than 35, which is then
improperly interpreted by the CR10 software.  A home_15 value less than 35 is
interpreted as meaning that the AEM aspirated housings (which hold the
temperature and relative humidity sensors) are in the opposite orientation to
what they actually are.  The AEM continues to exchange properly, but the
resulting calculations of Bowen ratio, latent heat flux, and sensible heat
flux are incorrect.  On September 15, 1994 Site Operations personnel
refurbished the wiring for the left housing and the home_15 signal returned
to proper levels.

When 30 minute home_15 values are below 35, the data fields listed below are
incorrect.

5 minute:  home, tair_top, tair_bot, thum_top, thum_bot, hum_top, hum_bot,
           vp_top, vp_bot

15 minute: mv_home

30 minute: home_15, tair_top, tair_bot, thum_top, thum_bot, hum_top,
           hum_bot, vp_top, vp_bot, Bowen, e, h

It is possible to recalculate the Bowen ratio, e, and h by using the 15
minute data to compute the correct gradients of temperature and relative
humidity (two 15 minute periods are averaged to give a half hour value); the
orientation of the AEM is obvious in the 15 minute data, based on the home
signal (which is largest for the first 15 minutes of the half hour).  More
information on performing this recalculation can be obtained from the
mentor.

• home_signal

• home_signal_15
• bowen_ratio
• temp_air_top
• vapor_pressure_top
• Heat flux, latent, at 1.5-m height, 30-min intervals(latent_heat_flux)
• vapor_pressure_bottom
• temp_trh_bottom
• temp_trh_top
• rh_bottom_fraction
• rh_top_fraction
• temp_air_bottom
• Heat flux, sensible, at 1.5-m height, 30-min intervals(sensible_heat_flux)

• home_signal
• rh_top_fraction
• temp_air_top
• temp_trh_top
• rh_bottom_fraction
• temp_trh_bottom
• vapor_pressure_bottom
• vapor_pressure_top
• temp_air_bottom

SGP/EBBR/E12 - Wind Direction Malfunction

For most of the period above, a malfunction in the E12 EBBR wind direction
sensor caused the following data values to be incorrect:

5 minute: res_ws, wind_d, and sigma_wd
15 minute: mv_wind_d
30 minute: res_ws, wind_d, and sigma_wd.

Wind direction normally indicated near north or 135 degrees (the latter
indicates a malfunction).

The wind direction sensor (S/N 3036) was replaced with S/N 3034 at 1620 GMT
on April 26, 1995.

• Wind speed vector mean(wspd_vec_mean)
• Wind direction vector mean(wdir_vec_mean)
• wdir_vec_std

• wdir_vec_mean_mv

• wdir_vec_std
• Wind speed vector mean(wspd_vec_mean)
• Wind direction vector mean(wdir_vec_mean)

SGP/EBBR/E12 - Wind Direction Malfunction

For the period indicated, a malfunction in the E12 EBBR wind direction
sensor caused the following data values to be incorrect:

5 minute: res_ws, wind_d, and sigma_wd
15 minute: mv_wind_d
30 minute: res_ws, wind_d, and sigma_wd.

Wind direction normally indicated smaller or larger than it should have.
The affected time period was determined from comparison with data from E7,
Elk Falls EBBR, which is located north of E12.

The wind direction sensor (S/N 3034) was replaced with S/N 3039 at 1822 GMT
on June 7, 1995.

• Wind speed vector mean(wspd_vec_mean)
• Wind direction vector mean(wdir_vec_mean)
• wdir_vec_std

• wdir_vec_mean_mv

• wdir_vec_std
• Wind speed vector mean(wspd_vec_mean)
• Wind direction vector mean(wdir_vec_mean)

SGP/EBBR/E12 - Wind Direction Malfunction

For the period indicated, a malfunction in the E12 EBBR wind direction
sensor caused the following data values to be incorrect:

5 minute: res_ws, wind_d, and sigma_wd
15 minute: mv_wind_d
30 minute: res_ws, wind_d, and sigma_wd.

Wind direction normally indicated small values (15 to 25 degrees) from 3 July
1995 at 0600 GMT through 6 July 1995 at 1500 GMT (at which time the sensor
was removed).  No sensor was in place between 6 July 1995 at 1500 GMT and 7
July 1995 at 1605 GMT (at which time a new sensor was installed).
 Old sensor S/N K3039, new S/N K3033.

• Wind speed vector mean(wspd_vec_mean)
• Wind direction vector mean(wdir_vec_mean)
• wdir_vec_std

• wdir_vec_mean_mv

• wdir_vec_std
• Wind speed vector mean(wspd_vec_mean)
• Wind direction vector mean(wdir_vec_mean)

SGP/EBBR/E12 - Wind Direction Malfunction

For the period above, a malfunction in the E12 EBBR wind direction
sensor caused the wind direction to output a constant 212 degrees.  Site
operations replaced the sensor (S/N K3033) with K3042.  The following data
values are incorrect for the period listed:

5 minute: res_ws, wind_d, and sigma_wd
15 minute: mv_wind_d
30 minute: res_ws, wind_d, and sigma_wd.

• Wind speed vector mean(wspd_vec_mean)
• Wind direction vector mean(wdir_vec_mean)
• wdir_vec_std

• wdir_vec_mean_mv

• wdir_vec_std
• Wind speed vector mean(wspd_vec_mean)
• Wind direction vector mean(wdir_vec_mean)