DQRID : D970314.2
Start DateStart TimeEnd DateEnd Time Data Quality Metric
10/22/1996173002/05/19971400Suspect
more
Subject:
SGP/TWR/C1 - RH Probe affected by tower damage
DataStreams:sgp1440twr25mC1.a0, sgp30twr25mC1.a1, sgp30twr60mC1.a1, sgp1twr60mC1.a0,
sgp1440twr60mC1.a0, sgp1twr25mC1.a0
Description:
On Feb. 3 and 4 the SGP CART 60 m tower western elevator was repaired.
This included the replacement of cables on the 60 m carriage and connectors on
the 25 m carriage.  Although damaged, the 60 m carriage cable had been 
temporarily repaired by Site Operations several months ago, with apparent 
success.  However, evidence since then suggests that the repair was not 
totally successful.  Tests since then suggest that the low return for the 
RH sensor may have been floating, or at least the connection was poor.  At 
the same time, the 25 m carriage connectors and tower receptacles were bent
in planes perpendicular to each other, apparently resulting in the same 
condition as existed for the 60 m carriage.  The damage was partially the
result of improper routing of the cables on the 60 m carriage by the tower 
installer and partially the result of improper parking of the carriages 
(damaging the 25 m carriage connectors and receptacles).  The cables were 
reoriented and the connectors and receptacles replaced by the tower 
installer on 3-4 Feb 97.

Because of the damage to the 25 m carriage connectors, the 25 m carriage 
was left off of the tower and no correct data was collected from the 25 m 
level from 22 Oct 97 at 1730 GMT until 4 Feb 97 at 1400 GMT.  Only the 60 m 
level reported useful data during this period.  During this period a slope 
of 1.107 was applied in the datalogger programming before the 60 m RH value
was output to the logger memory.  Calibrations of the probe by Vaisala last 
month show that the logger slope was appropriate for the sensor behavior at RH 
less than 85%.  Calibration of sensors SN 109 and 231 by Vaisala showed the 
outputs to be several percent low near 0% RH and nearly correct at 100%.  The 
calibrations by the original manufacturer caused the outputs to be outside 
the specifications for the bottom 3/4 of the sensors' range.  Vaisala 
adjusted the slope and offset for both sensors.

In use at 60 m on the tower, the RH probe slope changed above 85% RH, becoming
larger with increasing RH.  At saturation the 60 m sensor RH value was being 
reported as 110.7% to 111.0% (in other words the slope in the logger of 
1.107 X 100%).  Apparently, at higher voltages the poor low side connection 
was less of a factor in the voltage output by the probe.  Comparison with 
sonde data for RH above 85% yields the correction in the next paragraph.

Therefore, my recommendation for the period listed above is to use the 60 m 
RH data as is for RH less than 85%.  Reported RH above 85% can be roughly 
corrected with the following: actual RH = 36 + 0.577 X (60 m reported RH) . 

Vapor pressure can then be recalculated by the user from the correct RH 
values and the temperature.

On Feb. 4 and 5, I conducted a number of tests on the carriages and RH 
probes (SN 231 at 25 m, SN 109 at 60 m), then replaced the RH probes with 
the probes recently (SN 234 at 25 m, SN 226 at 60 m) calibrated by Vaisala.  

One minute 60 m T and RH data is extremely consistent with the data 
from the sonde at 60 m height since 5 Feb 1997.  In fact, the 
comparisons have never been better, reflecting the problem with obtaining 
correct calibrations from the original manufacturer (even before the tower 
was damaged), and the problems created by the damage on the tower.

The problems resulting from damage to the tower carriage cables and 
connectors began around June 1995 (although the 60 m AC cable and 25 m 
receptacles had been damaged at least twice previously).  I made a number of
corrections to the RH data between June 1995 and when the tower was 
repaired this past February. 

25 and 60 m T and RH data since the tower repairs and the Feb. 5, 1997 
replacement of the RH probes are very consistent with the sonde data.  The 
only significant difference between the 60 m RH probe and the sonde RH probe 
occurs when the surface to 60 m RH gradient is large; this is expected since 
the RH probe on the sonde cannot respond to large changes in gradient quickly 
enough to produce the same measurement as the 60 m tower RH probe.

I will continue to compare the sonde, SMOS, and tower sensors for data quality 
purposes, as I have routinely for the past year.
Suggestions: 
Measurements:sgp1twr25mC1.a0:
  • vap_pres
  • rh
more
sgp30twr25mC1.a1:
  • rh
  • vap_pres
  • sd_vap_pres
  • sd_rh
more
sgp1twr60mC1.a0:
  • vap_pres
  • rh
more
sgp1440twr60mC1.a0:
  • min_rh
  • min_vap_pres
  • max_rh
  • max_vap_pres
more
sgp1440twr25mC1.a0:
  • min_vap_pres
  • min_rh
  • max_vap_pres
  • max_rh
more
sgp30twr60mC1.a1:
  • rh
  • sd_vap_pres
  • vap_pres
  • sd_rh
more

Close this window