Notes on Quality Control


The pressure data was mostly correct throughout the duration of the project aside from a number of spikes that had to be removed between February 25th and March 4th, for the Irgason sensors. In all instances the data were obvious outliers.



The Qsoil signal at the EHS location experienced a signal degredation which worsened as the project progressed <>. The corrupted signal was unrelated to moisture intrusion and suspected to be a byproduct of the wireless communication; therefore the sensor was rewired on March 16th to bypass the wireless link and establish a hard connection to the serial mote <>. There were some additional instances where the Qsoil measurements exhibited a degredation in signal and the data had to be removed. One such example was between the period of March 23rd 19:00 until March 25th 09:05 for the 10 and 20 cm sensors. These instances were obvious from an observational standpoint because of the trasition in data resolution. There was also an significant degredation in data quality between March 30th and and April 14th where "Qsoil.ehs" returned a highly digitized version of the signal. The data throughout this time was too corrupted to salvage and so it was all removed.

In addition to the edits listed above, there were several instances where spikes were removed from the data set.



The humidity (as well as tempreature) data was initially QC'd using the fan current of the TRH sensor as a cue for the data integrity. If the value of "Ifan" was at or near zero, the data was removed, as well as if there were abnormalities in the signal such as spikes deviating beyond the normal operating current of 20 to 60 mA. Up until March 5th at 10:30, "Ifan.50m" averaged too low of a current, and so the data was removed. Between March 18th 13:45 and 19th 09:50, "Ifan.2m.bao" had a low fan current and eventually needed to be replaced, the data was removed <>. Between April 23rd 13:15 and April 28th 11:50, the current for "Ifan.300m" read too low, and so that data was removed <> (see also <>). On May 5th between 09:40 and 15:00 there was another instance of erratic Ifan values for the 100 meter data, and that was removed. On May 23rd at 13:20 there was a significant shift in the current of "Ifan.200m" and "Ifan.300m," but aside from removing some spikes, most of the data was maintained <>.

There was an instance where the "RH.50m" sensors froze up and could not be restarted through software and so it had to be replaced; only data between 14:15 and 21:20 on April 24th was removed <>.

For the 200 and 300 meter sensors, there were a sizable number of spikes/outliers leading up to as well as following the BAO total power outage event <>. The 26th, 27th, and 28th had the bulk of these outliers (as well as several on June 4th and 5th for the 300 meter sensor).



The temperature data QC process included all the events from the RH QC processes. In addition, on April 1st between 15:25 and 18:15, data for "T.300m" was removed <>. 


...more notes are still being added