Hourly VWC and temperature Data from five Decagon 5TM probes (Decagon Devices, Inc.) are logged in the field by Decagon Em50 logger. Each sensor uses a 5-meter long cable with a 3.5mm stereo connector to the Em50 logger. Probes and cables are buried horizontally at 10cm of depth, except on the UMBC plots where they are buried in 10 cm increments to 50 cm. The Em50 logger is mounted above-ground on a stake. Data are collected at 30-60 day intervals using a laptop and the ECH20 Utility program. Battery levels are checked and replaced upon reaching 1/2- 1/3 capacity.
Volumetric Water Content:
The 5TM sensor uses an electromagnetic field to measure the dielectric permittivity of the soil surrounding the probe. This permittivity is proportional to the volumetric water content. Factory calibrations were used. These calibrations use the Topp equation (Topp et al., 1980) and have an accuracy of +/- 0.03 m3/m3 for a mineral soil.
Temperature:
The 5TM uses a surface-mounted thermistor to take temperature readings. It is located underneath the sensor overmold, next to one of the prongs, and will read temperature in degrees C. The equation used offers 0.5 degrees C accuracy in the range of -40C-+60C.
Data notes:
• Not all of the data are collected at 60-minute intervals. There are segments of the dataset that were collected at 30-, 20-, or 10-minute intervals instead of 60-minute intervals.
• Throughout the years of data collection, some sensors needed to be removed and replaced. It should not be assumed that each data stream (data from a given port at a specific site) was collected from the same sensor buried in exactly the same location throughout the length of data collection.
• There are a handful of rows with identical dates and timestamps. See explanation of 777 flag that can be used to filter these rows out when analyzing the data.
Data file formatting notes:
Each sensor is plugged into a port on the data logger and identified by the port number, resulting in data columns with names like “Port_1_Temperature”, “Port_1_VWC”, “Port2_Temperature”, and on through “Port_5_VWC”. Each of those temperature and VWC columns has a corresponding flag column that contains a value reflecting how a given data value was flagged. Flag values include:
• 0: Data value was not flagged.
• 1: Data value was missing (NA).
• 2: Data value was outside of an acceptable range.
◦ For temperature: outside of -10 to 38 degrees Celsius.
◦ For VWC: outside of 0 to 1.
• 3: For temperature data. Data value is more than 3 degrees Celsius away from the data value immediately before or immediately after it.
• 4: Data point has no neighbors within one hour before or after it (there is a time gap on both sides of the data point, so was unable to check if the value should be flagged 3).
• 5: For VWC data. VWC value corresponds to a temperature value that was flagged for any reason. If a temperature data value was flagged for any reason, the corresponding VWC value was flagged with a 5.
• 999: Data value was manually flagged.
Additional flags:
• 777: There are some parts of the dataset that had two rows with the same date and timestamp. If the two rows with the same date and time were not identical, all flag values in those rows were set to 777. (If the two rows held identical values for all measurements, one row had all flags set to 999 while the other was left unchanged.)
• 888: There are some parts of the dataset that appeared to have reasonable VWC data while the temperature value is marked as missing. To recover these data, VWC data that were previously flagged as 5 because their corresponding temperature value was flagged as 1 were re-flagged as 888.
After flags 0 through 5 were initially set, the data were viewed and manual flags were set, including changing some 4 flags to 999 flags if a data value looked out of line with its surroundings or was so isolated that it was not possible to tell if it was in line. For analysis, data to include have been those flagged as 0, 4, and 888, with data flagged as 1, 2, 3, 5, 777, and 999 being excluded. For more recent data, data values were left in the dataset for data flagged with a 1. For older data, out of range values were changed to NA.
Site-specific metadata:
For some sites, site-specific metadata were recorded in the earlier years of data collection. Below are notes that may help to explain some of the gaps in the data and/or explain some jumps in the data at specific sites.
• ORLR:
◦ 2/9/2012: Logger stopped logging
◦ 7/3/2012: Probes 1 and 3 are in fully saturated soil
◦ 7/27/2012: All probes (except probe 2) were knocked out/disabled by trees that fell on the plot during a large storm
◦ 9/21/2012: One cord was severed and all probes were dug up
◦ 10/15/2013: Logger removed from field to install waterproof casing
◦ 5/6/2014: Reinstalled logger in field with waterproof casing
• ORU1:
◦ 9/21/2012: Probe 4 was replaced
• ORUR:
◦ 2/9/2012: Logger stopped recording and interior of case was visibly wet with some corrosion on the copper plates of the logger
◦ 9/21/2012: Probe 3 was not working and was removed
◦ Overall: Probe 3 produces strange temperature readings while producing normal-looking VWC readings
◦ Overall: Probe 4 is supposedly on a hummock but gives consistently high VWC values
