Presentation on theme: "Preliminary Results from ATDD’s Soil Moisture/Temperature Testbed Soil Moisture and Soil Temperature Observations and Applications: A Joint U.S. Climate."— Presentation transcript:
Preliminary Results from ATDD’s Soil Moisture/Temperature Testbed Soil Moisture and Soil Temperature Observations and Applications: A Joint U.S. Climate Reference Network (USCRN) – National Integrated Drought Information System (NIDIS) Workshop, Oak Ridge, TN, March 3-5, 2009 William Collins USCRN
Characteristics of Oak Ridge Data From 2000 day 172 of 2007 through 0900 day 100 of 2008 Instrument Deployment— 4 holes with temperature and moisture sensors: 3 at 5 cm depth 3 at 10 cm depth 2 at 20 cm depth 1 at 50 cm depth 1 at 100 cm depth The moisture sensors are Vitel moisture instruments, measuring the dielectric constant to get water content. The water content is given in m 3 water/m 3.
Study Objective This study will exclusively examine the characteristics of the soil moisture and temperature data at 5 and 10 cm since only those 2 depths have 3 sensors. The use of 3 sensors at the same location has proven invaluable in the quality control of temperature and precipitation (from the 3-wire Geonor gauges) for the U. S. Climate Reference Network. This study will take a first look at that utility here.
Statistics for Full Time Period by Individual Hole and Depth: Mean Temperature by Sensor Inter-sensor Temperature Differences Distributions (0.2 C bins) Means Standard Deviations (Only a few typical examples are shown.)
Comparison of Moisture for 3 Co-located Sensors The moisture during the period of record begins dry and ends wet. The data were divided as follows: Dry period: 2000 day to 2400 day Wet period: 0100 day to 0900 day Transition period: data in between The following plots show a comparison of the moisture at 5 and 10 cm depth, individually for each sensor and hole for a selected time period spanning dry to wet.
Note! Same vertical scale used for all these plots.
Comparison of Average Moisture At 4 Holes Individually at 5 and 10 cm Depth Selected Time Period (Dry to Wet) Note! Without quality control, the best estimate of the moisture should be the average of the 3 sensor values. The following plots have the 3 sensor values averaged, and in addition, they are smoothed in time with values (0.25, 0.50, 0.25).
Examples of Large Inter-Sensor Differences Lower Limits for Example Choices (8 : Moisture (m 3 water/m 3 soil): 5 cm – cm – Temperature (C): 5 cm – cm – 5.44
yeardaytimehole depth (cm)T1T2T Large Temperature Difference (Dry)
yeardaytimehole depth (cm)T1T2T Large Temperature Difference (Transition)
yeardaytimehole depth (cm)T1T2T Large Temperature Difference (Wet)
yeardaytimesensor depth (cm)m1m2m Large Moisture Difference (Dry)
yeardaytimesensor depth (cm)m1m2m Large Moisture Difference (Transition)
yeardaytimesensor depth (cm)m1m2m Large Moisture Difference (Wet)
Comparison of Mean Moisture for Dry and Wet Periods For the Same Sensor
Line has slope 1:1 Best response would be alone line of slope 1:1, with vertical displacement and sensors from same hole grouped together.
Comparison of 1-Hour Changes in Moisture Between Pairs of Sensors at Hole 1 Full Time Record Possible use in quality control…
Line is not linear fit; has 1:1 slope. Ideal fit would be along this line.
Line is not linear fit; has 1:1 slope.
Comparison of Moisture Standard Deviations— Inter-sensor differences versus Inter-sensor 1-hour change differences Individually for Dry and Wet Periods Implications for Quality Control
Since the inter-sensor standard deviation for 1-hour change becomes smaller as the inter-sensor standard deviation becomes smaller, the inter-sensor 1-hour change provides independent information for the quality control of soil moisture.
Steeper slope of line for wet period indicates even greater value of 1-hour change for quality control than for dry period.
Summary Preliminary comparison was made between the 3 co-located soil temperature and moisture sensors from nearly 300 days of data from four holes at Oak Ridge, TN. The inter-sensor differences show wide differences in the performance of individual sensors. Also, there are significant differences between the mean values of moisture at the 4 holes at 5 and 10 cm depths. The results have implications for quality control for both temperature and moisture— The use of 3 sensors is invaluable for error and value determination. Sensor value and change in time are both useful for qc. Individual sensor characteristics, if developed, might be valuable. Question for research: Can differences in individual sensor characteristics be linked to calibration, installation, or other factors?