UPDATE: I have a Davis Weather Station with a radiation shield that is gathering dust in my workshop. I'm going to rebuild the housing for the BME280 on my instrument using parts from the Davis station and share details shortly. I'll also include data from before an after the radiation shield upgrade to see how much bias is introduced by a poorly designed shield. I have some ideas for the exposed DS1520 thermometers as well, so looks like I have my next project lined up.

This video summarizes the first suite of data collected by a new longterm soil monitoring installation in my back yard using temperature (DS18B20) and capacitive soil moisture sensors. The design of the instrument is summarized in this video:
Long Term Soil Monitoring Installation:    • LongTerm Soil Monitoring Installation   .

The full playlist of soil monitoring experiments is available here:    • Soil Monitoring Experiments  

Prior experiments:

Capacitive soil moisture sensors that meet quality assurance criteria realize a signal response of 280 in water to 785 in air.

Most makers will use the signal for wet and dry endpoints to develop a linear curve for approximating soil moisture as a function of signal response. Based on soil column experiments, I’ve determined these sensors do NOT respond linearly to water losses. This can realize an over estimation of soil moisture which is something I confirmed in my latest experiments.

New experiments:

(1) Baseline data collection

The following table summarizes collected temperature data collected at various heights/depths over 16 days with no moisture encountered or introduced at the site.

DateTime, AirTemp, T0, T3_3, T9_5, T15_5
min, 54.97, 59.00, 62.60, 81.50, 85.10
max, 119.71, 133.70, 152.60, 102.20, 91.40
average, 85.07, 90.43, 96.28, 91.68, 88.37
stdDev, 16.64, 19.43, 24.53, 5.00, 1.44

In summary, temperatures at ground maxed out at 153 degrees Fahrenheit when ambient temperatures in the shade were about 100 degrees Fahrenheit. Temperatures 28 inches above the basin maxed out at 120 degree Fahrenheit. At a depth of 1foot, maximum temperatures dropped to just over 90 degrees Fahrenheit with little standard deviation.

Baseline data demonstrates that capacitive soil moisture sensors are influenced by diurnal temperature patterns. In Southern Arizona, sensors buried at a depth of 3 inches in a silty clay soil can yield a +/25 unit signal variation (on average) during the summer at a constant (dry) soil moisture content.

(2) Simulated Storms

Water was introduced at the site at volumes approximating two consecutive 2yr 15minute storms.

Capacitive soil moisture sensors placed in a silty clay at a depth of 3 inches yielded a drop of 140 to 160 units in response to the simulated storms. This suggests these sensors may be useful for monitoring infiltration at various depths in response to a rainfall event. However, their insensitivity to moisture loss limits their use in how a soil column dries over time.

Temperature sensors placed at ground level and 3 inches above ground will report up to a 20 F degree difference in dry conditions during the summer. That difference is halved to 10 F in response to a simulated 2yr 15minute storm. The change results from air temperatures at 3 inches above ground decreasing even as general ambient air temperatures are increasing.

Sensor disturbance can impact capacitance and temperature readings. If your sensors are attached to a staff or rod in the ground, be careful not to move your sensors when servicing or downloading your installation.

(3) Future Experiments

Further explore the deltas between ground and air temperatures in dry vs wet conditions using R for a more quantitative analysis. How does mulch impact the holding time for respective changes after a rainstorm? What impact do different mulches have on soil temperature over time?

Data published here: https://docs.google.com/spreadsheets/...

Subscribe for Updates!

Closing music sampled from Frederick Noad Solo Guitar Playing 1 Part B: Allegro (Haydn)