SOIL MOISTURE MONITORING DEVICES

(First Year Results)

T.E. Harms

Please Note:  RF Moisture Sensor name has been changed to

AP Moisture Probe

            Irrigation management is about deciding when to irrigate and how much water to apply. When irrigations are managed properly, sufficient water is available for plant uptake for photosynthesis and transpiration during the growing season and excess water loss (through deep percolation and surface runoff) is minimized. Yields can suffer if irrigation is delayed and available moisture within the root zone of the crop is appreciably reduced. Similarly, irrigation efficiency is reduced if water is applied in excess of what the soil can infiltrate or retain. Knowing the soil moisture status within irrigated fields is critical information if irrigation efficiency is to be maximized and timely irrigation to obtain the highest yields and the best quality is to occur.

 

Gravity and wheel move irrigation systems continue to be replaced by center pivot irrigation systems. Labor savings and better control of application amounts with center pivot irrigation are some of the reasons for this conversion. Also, for some contracts to grow higher value crops, the processors insist that the field have center pivot irrigation. Soil moisture management under a center pivot differs from that for wheel move or gravity irrigation. Frequent applications, applying less water at a time is the operating recommendation for center pivot management and information regarding soil moisture status is important for timely irrigation scheduling.

 

            Realizing the benefit and need for frequent soil moisture monitoring for successful irrigation management, the question becomes what is available to the irrigator or irrigation management consultant to monitor the soil moisture levels in irrigated fields? The requirement for any tool or instrument is that it has to be relatively inexpensive, portable, accurate, easy to use, immediate display of results and have a visual display that is easily understood.

 

            A demonstration project commenced in spring of 2000 with the objectives to test, modify if necessary or as required, and demonstrate various commercially available soil moisture monitoring instruments that meet the criteria outlined above.

Method

 

            Five soil moisture-monitoring devices were tested at 10 sites within the Eastern Irrigation District. At nine of the sites at least one of the soil moisture instruments were evaluated and at one site all the instruments were installed for evaluation.

             

The soil moisture instruments were chosen to represent variation in methods of determining soil moisture and installation (Figure 1). The five instruments tested were the Hydrosense, Theta Probe, R.F. Soil Moisture Sensor (name has been changed to AP Moisture Probe), AM400 and Watermark.

 

Figure 1. Methods of installation for soil moisture determination

 

The Hydrosense probe manufactured by Campbell Scientific Inc. uses a soil property called dielectric permittivity to estimate volumetric moisture content. A high frequency electromagnetic wave pulse travels the length of a pair of rods (either 12 or 20 cm) inserted in the soil and returns to a sensor (Figure 2). The time it takes for the wave to complete the travel is an indication of the dielectric permittivity of the soil.

 

Figure 2. Hydrosense and hand held reader by Campbell Scientific

 

The readout of the Hydrosense can be either in volumetric moisture content percentage (VMC%) or relative water content when calibrated for field capacity and wilting point. The readout displays relative water content from 0 to 100% of available and also how much additional water (mm) is required to bring the depth of monitoring up to field capacity; sometimes referred to as deficit.

 

The Hydrosense was installed at 3 sites within the County. Soil texture varied from clay loam to silt loam. Access tubes were cut from 3 inch pvc pipe and installed to depths of 30, 60, and 90 cm and soil moisture readings from those depths were interpolated to represent the intervals 20-40, 40-60, 60-80 and 80-100 cm. Surface readings would represent the interval from 0-20 cm. The 12 cm length of probes were used for sampling with the Hydrosense.

 

The ThetaProbe manufactured by Delta-T uses a similar concept as the Hydrosense probe by sensing the apparent dielectric constant of the soil to estimate volumetric water content. The ThetaProbe has a configuration of 3 rods surrounding a center rod, all of which are inserted into the soil (Figure 3). The difference between voltage at a crystal oscillator (enclosed in the body of the probe) and that reflected by the rods is used to determine the dielectric constant of the soil. The readout from the Theta Probe is VMC%.

 

Figure 3. ThetaProbe and hand held reader from Delta-T Devices

 

The ThetaProbe was installed at 5 sites within the County. Soil texture varied from clay loam to sandy loam. Aluminum access tubes were cut and installed to depths of 30, 50, 70, and 90 cm and soil moisture readings from those depths would represent the intervals 20-40, 40-60, 60-80, 80 to 100 cm. Surface readings would represent the interval from 0-20 cm.

 

The R.F. soil moisture sensor (name has been changed to AP Moisture Probe) manufactured by AquaPro measures the dielectric coefficient of the soil using radio frequency waves. Soil moisture measurements can be taken at any number of locations to any depth.

 

This unit is a profiling probe meaning it is lowered into a polycarbonate tube (Figure 4) that has previously been inserted into the soil. The polycarbonate tubes come in 1 meter lengths but can be extended to 2 meter or greater lengths by connecting them together. The readout from the R.F. sensor (name has been changed to AP Moisture Probe) is percent available moisture. Since most irrigation decisions are based on how much available moisture is in the soil profile, this probe gives the irrigator an immediate readout of this value without the need for any conversions or detailed soil parameters.

 

Figure 4. R.F. Sensor (name has been changed to AP Moisture Probe) and hand held reader from AquaPro

 

The R.F. Sensor (name has been changed to AP Moisture Probe) was installed at 6 sites within the County. Soil texture varied from clay loam to loamy sand and at one site there was a textural discontinuity at 60 cm with the sandy loam overlying clay loam to clay-textured soil.

 

The WatermarkÔ sensor manufactured by Irrometer works on the principle of electrical conductivity of moist gypsum, which is strongly dependent on the water tension. The sensor consists of a matrix of granular material and two electrodes embedded in gypsum (Figure 5). As water is "pulled" from the matrix, the electrical resistance between the two electrodes increases. The probes are buried and two leads from the electrodes are connected to a handheld meter during readout. The readout is in centibars (a unit of soil tension) and to properly convert or interpret this value as VMC% or percent available, a soil water characteristic curve must be constructed for the specific soil.

 

            Figure 5. Watermark sensor and hand held reader from Irrometer.

 

Watermark sensors were installed at 3 sites within the County. Soil texture varied from clay loam to fine sandy loam. At two sites, the sensors were installed to depths of 30, 60, and 90 cm and soil moisture readings from those depths were interpolated to represent soil moisture status for the intervals 0-30, 30-60 and 60-100 cm. At one site the sensors were installed at depths of 10, 30, 60, 90 and soil moisture readings from those depths were interpolated to represent soil moisture status for intervals 0-20, 20-40, 40-60, 60-80, 80-100. A soil characteristic curve was constructed from comparisons of the Watermark readings in centibars and the VMC% of the soil to determine percent available soil moisture.

 

The AM400 is not a soil probe but a datalogger that uses the WatermarkÔ sensor as the soil probe component. The WatermarkÔ sensors are buried into the soil and the leads are connected to the AM400. The logger records soil moisture readings from the WatermarkÔ sensors (up to 6 individual sensors can be connected to the logger) every eight hours and graphically displays the readings from the sensors showing five weeks of soil moisture readings. The logger displays soil tension in centibars and similar to the WatermarkÔ, a soil water characteristic curve is required to convert soil tension to volumetric moisture content (Figure 6).

 

(a)                                                                                                             (b)

Figure 6. AM400 logger (a) and close-up of digital display (b) from M.K. Hansen.

 

The AM400 was only tested at one location with a clay loam textured soil. The Watermark sensors were installed at depths of 10, 30, 50, 70 and 90 cm and the readings were interpolated to represent soil moisture in zones 0-20, 20-40, 40-60, 60-80 and 80-100 cm. A soil characteristic curve was constructed from the readings of the sensors (centibars) compared to VMC% measured with a neutron probe.

 

Accuracy of the probes was initially determined by using regression to compare available soil moisture percentage calculated from the readings of the various sensors to the value calculated from the readings of a CPN 503 soil moisture probe. There are a few limitations for this method of comparisons. A single reading soil moisture reading at a specific depth does not necessarily accurately represent variations in soil moisture for a 20 to 30 cm depth interval. Additionally, the soil moisture monitoring probes were not placed in exactly the same location as the neutron probe at a sample site. Soil moisture can vary over fairly short distances, especially within an irrigated field.

At a few sites, available moisture calculated from the neutron probe readings were consistently higher or lower than that calculated from the readings of the various soil moisture sensors. Upon auguring and assessing soil moisture by hand feel method, it was determined that the soil moisture was in fact wetter/dryer between the neutron probe access tube location and that of the other sensor. Weekly available soil moisture percentages were converted to natural log form and the slope of the regression line between the neutron probe and soil moisture sensor readings were calculated and compared (Figure 7).

Figure 7. Comparison of slope of regression line between neutron and theta soil moisture probe at one location.

 

RESULTS

            The R.F. soil moisture probe (name has been changed to AP Moisture Probe) scored the highest for the criteria assessed in Table 1. Readout from the hand-held meter is continuous (from 0 to 100% of available soil moisture) as the probe is lowered into the access tube. The only difficulty with the probe came during installation. The soil auger provided by the supplier for installation of the polycarbonate tubes was inappropriate for clay-textured soils. It would be more effective if the auger came equipped with a t-handle so more force could be exerted when auguring in dry clay or clay loam textured soils. (AquaPro~Sensors agrees with this assessment of our free soil auger.  Our free soil auger now has a durable t-handle).

 

            The Hydrosense required a 7.8 cm diameter access hole and did not come with any extension devices. An extension device was made locally so depth readings could be obtained. A special auger was purchased for installation of the larger access tubes and auguring the larger diameter hole was difficult in clay or clay loam textured soils. The 12 cm metal probes, rather than the 20 cm probes were used with the Hydrosense. Even with the 12 cm length, the probes would often bend when inserting them into dry soil. Inaccurate readings result if the probes are not inserted parallel to each other in the soil. Values for VMC% at field capacity and wilting point are required to convert VMC% reading to percent available moisture.

 

Reliability was an issue with the Hydrosense soil moisture probe. Shortly after installation, the hand held reader failed to work and had to be sent in for repair. The hand held reader again failed near the end of the season when the mode was set to VMC% but the reading was in relative water content. The display on the hand-held reader would often “freeze-up” and the batteries would need to be removed and replaced to reset the display. The auto shutoff would often not work and the batteries would quickly run down. A few modifications to the hand-held reader that would improve the operation would be:

*   a manual reset switch on the body of the hand held reader.

*   visual display on only when the button for readout is depressed, otherwise off.

*   a manual switch to choose between VMC% and relative water content.

 

The Theta Probe has a smaller, cylindrical design and required a 5 cm diameter access hole in the soil for obtaining soil moisture readings at depth. Extension tubes were provided by the manufacturer to extend the length of the probe for depth readings. The shorter (6 cm) and thinner probes on the ThetaProbe are more suitable for inserting into soil and no bending problems were encountered. The readout from the Theta Probe is VMC% and must be converted to percent available by inputting the soils VMC% at field capacity and wilting point for each depth increment. We had no problems with failure of the probe or the hand held reader but felt the connection from the probe to the hand held reader could be made more robust to withstand constant connecting and disconnecting.
           
The Watermark sensors worked well as long as the soil was moist. In one field of clay loam soil, the upper zone of the soil dried to near wilting point prior to irrigation. After irrigation, the sensors did not detect the increase in soil moisture. The manufacturer reports that the sensors may have to be reinstalled after a clay textured soil has been allowed to dry. Aggregates of clays high in montmorillinite have the property of expanding when wet and shrinking when dry. When the clay textured soils are allowed to dry, the clays shrink and a good soil/probe contact is not maintained.

 

The conversion of the readings from centibars to available moisture content percentage is quite difficult. A soil characteristic curve has to be constructed to relate soil tension readings to the corresponding VMC%. It may be more practical for the irrigator to decide when to irrigate by relating the centibar reading from the sensor to simultaneous hand auguring and hand feel. Once the relationship is established, the irrigator can use the readout from the Watermark sensors to schedule subsequent irrigations.
 
   

Instrument

Portable

Ease of Installation

Scale (1-5)

1 easy.... 5 difficult

Ease of Use

Scale (1-5)

1 easy.... 5 difficult

Interpretation of Readout

Scale (1-5)

1 easy.... 5 difficult

Hydrosense

Yes

4

3

3

Theta Probe

Yes

3

1

3

R.F. Sensor

(AP Moisture Probe)

Yes

2

1

1

AM400

No

4

1

5

Watermark

No

3

1

5

Table 1. Subjective ratings for each of the soil moisture sensors.

             

ACCURACY

             

            Comparisons were made between the weekly neutron probe readings and weekly soil moisture sensor readings at the various locations (Table 2). The R.F. Sensor (name has been changed to AP Moisture Probe), ThetaProbe and Watermark were all comparable in their accuracy. Limited testing of the Hydrosense was a result of the problems encountered with reliability of the hand held meter that came with the probe.


 

Sensor

Number of sites 

R2

Hydrosense

3

0.57

ThetaProbe

5

0.80

R.F. Sensor

(AP Moisture Probe)

7

0.83

Watermark

4

0.76

Table 2. Regression analysis of the various probes compared to the CPN soil moisture probe.

             

            Average difference in slope of the least squares regression line of the natural logarithm of weekly available soil moisture percentages between the neutron probe and the various instruments is shown in Table 3. The R.F. Sensor (name has been changed to AP Moisture Probe) again scored the highest when compared to the neutron probe readings.


 

Sensor

Average difference in slope compared to neutron probe.

Hydrosense

0.11

ThetaProbe

0.07

R.F. Sensor

(AP Moisture Probe)

0.01

Watermark

0.12

Table 3. Average difference in slope for least squares regression line for the log transformed available soil moisture percentage between the various probes and the neutron probe.                  

             

            The ThetaProbe consistently gave higher readings for VMC% than the neutron probe at all sites. The readings varied with the level of force that was used when inserting the metal electrodes in the soil; when more force was applied, VMC% increased. Robinson et al., 1999 speculated that the cylindrical arrangement of the electrodes compacted the soil and would confine the reading of VMC% to the zone close to the middle electrode.

 

CONCLUSION

 

            This is the first year of a 3-year study for testing soil moisture monitoring equipment. Continued and expanded testing of select soil moisture monitoring equipment will commence in spring of 2001. Plans are underway to test the R.F. Sensor (name has been changed to AP Moisture Probe) and ThetaProbe at all irrigation district office locations in southern Alberta and to use them in conjunction with the neutron probe at the demonstration farm in Lethbridge. Suggestions and recommendations for improvement will be forwarded to the manufacturers of the instruments.