Less Water

More Food?
Real-Time Three-Dimensional Imaging of
Soil Moisture Distribution for Intelligent
Irrigation

2017

 Rationale
Hunger, one of the most prominent problems in the world today, is primarily
caused by a shortage of fresh water. This problem is especially severe in third-world
countries and regions with low rainfall or prone to severe droughts. The most recent
California drought vividly demonstrated the importance of fresh water conservation. The
states large agricultural industry was most heavily impacted by the weather abnormality.
Many farmers were deprived of water for irrigation, which forced them to reconsider
their irrigation methods. For example, the use of traditional sprinkler systems is
accompanied by substantial evaporation and watering of unused soil regions. In response
to this, some have implemented subsurface drip irrigation systems, which can save up to
25-50 percent of water [1].
Significant optimization of water use can be achieved by knowing the time and
amount of watering, as well as the specific location of soil that requires irrigation.
Recently, various novel methods for soil moisture measurement were developed.
However, most existing techniques can only provide measurements at one point in the
soil, require frequent calibration, and remain very costly.
A novel, affordable, standalone device that provides real-time three-
dimensional imagining of soil moisture content can significantly improve the
efficiency of fresh water use, greatly increasing the availability of food and water
resources all around the world.

Figure 1: The recent California drought left an immense impact on the states agriculture
industry and posed concerns for the efficiency of present irrigation methods.

 Physical Properties of Soil for Moisture

Measurement
In order to measure the soil moisture content, variations in physical properties of soil may
be analyzed (dielectric constant, resistivity, thermal conductivity, neutron absorption, and
others). Dry soil has a dielectric constant of 2-5, while water is very different, with the dielectric
constant equal to approximately 81. Similarly, soils resistivity is 10 -1000 *m, while irrigation
water has a much lower value. Thermal conductivity also noticeably changes when soil contains
moisture. Neutrons are heavily scattered by hydrogen contained in water, but easily penetrate dry
soil. Various conventional sensors and devices have been used to measure these physical
properties of soil and obtain the moisture content. Among these known methods are:
1. Time Domain Reflectometry - ($2300)
2. Reflectometry (FDR) Technique - ($300)
3. Electrical Impedance Sensor - ($70)
4. Ground Penetrating Radar - ($20000)
5. Soil Resistivity Sensor Technique - ($190)
All traditional methods remain too expensive for wide use, most sensors do not store
historical data, and, most importantly, do not provide volumetric information. On the other hand,
my method is based on multiple measurements of electrical resistivity to build a real-time three-
dimensional image of soil moisture content.
Two arrays of copper electrodes (with least galvanic corrosion), are placed in soil around
plant roots. Electrical resistance is measured between all possible pairs of electrodes from the
two sets, with ~1 Hz frequency. The results are then used to calculate the resistance at
predetermined virtual cells located in between the two arrays of electrodes.

 Measurement of Soil Resistance

How the resistance is measured:
To measure the resistance of soil with the Arduino microcontroller, a voltage drop across a
test resistor is used. The voltage on the analog connection depends on the resistance of the soil,
according to this equation:

Controlling many electrodes with few Arduino outputs:

While my method requires simultaneous measurements of an array of electrodes, the Arduino
Nano has very few usable connections. However, I decided to use this specific microcontroller in
favor of its compact size and affordability. With the use of shift register and MUX integrated
circuits, 48 electrodes are controlled simultaneously with only three control signals. Control bits
are sent sequentially to a series of four 8-bit shift registers. They, in turn, the address of the MUX
connection is defined according to bytes received from the Arduino. Only two electrodes from
opposite electrode arrays are selected and measured at one moment, after which the next pair is
chosen and measured. This process allows to obtain resistance data from all pairs of electrodes
multiple times per second.

Optimization of Programmable Resistor:

The value of the test resistor needs to be optimized for specific soils and moisture conditions,
which significantly influence the resistivity values. To get relatively accurate measurements, the
test resistor must be manually replaced for different soils. To avoid inconvenient manual
calibration, I created an algorithm to automatically adjust the resistance of a nonvolatile digital

resistor. To digitize the array of resistances most precisely, the range of measured voltages should
be maximized:

Vmax Vmin = ;

The range of measured voltages will be maximum when == 0:

After the resistance measurements between all pairs of electrodes, Rmax and Rmin are
determined, and the optimal value of Rtest is calculated.
If the optimal resistance value is different from the current one, the Arduino sets the new
resistance with clock, enable, and direction pins.

 Algorithm of Three-Dimensional Soil

Moisture Imaging

The goal of the main algorithm is to create a real-time three-dimensional representation

of soil moisture content according to an array of electrical resistance measurements. All data (the
resistance between all possible pairs of electrodes) is sent from the Arduino microcontroller to a
computer or mobile device, where all results are processed and then saved to a file, which is used
for moisture content visualization.
The volume between the two sets of electrodes is divided into virtual cells. The
resistance inside each virtual region needs to be calculated from raw resistance data (e.g. 256
values for a 16x16 array of electrodes). The algorithm assumes that current flows between each
pair of electrodes in a straight line, crossing multiple cells. The resistance is equally divided
between all cells which are crossed by the current. By the end of analysis, each virtual cell
contains an averaged resistance value representing a soil region.

 Measurements and Results

Watering #1 (at 7:29 GMT, the left side of the soil volume was irrigated)

Figure 2: The resistance averaged over all pairs of electrodes.

Figure 3: Raw data obtained from pairs of opposite electrodes. Readout noise is visible
for fast single measurements.

Figure 4: The calculated ratio of present soil resistance to dry soil resistance values. Two horizontal and
two vertical sets of electrodes are shown (as indicated by the legend).

Watering #2 (at 3:16 GMT, the right side of the soil volume was irrigated)

Figure 5: The calculated ratio of present soil resistance to dry soil resistance values. Two horizontal
and two vertical sets of electrodes are shown (as indicated by the legend). Note the different in initial
resistance values among different electrodes.

 Discussion of Results

1. After the application of water, there was a substantial change in soil resistivity across all
measurements. This justifies that resistivity can be used for a qualitative assessment of water
content (Figure 1).

2. A continuous measurement session over several days showed a very slow dynamic of water
content depletion (Figure 2).

3. A distinct difference in resistance near the top and bottom of the soil volume was observed.
The dynamics of water penetration downward could be clearly observed (Figure 4.d).

4. When the soil was watered near electrode #9, there was a difference in the moisture content
along the horizontal plane (e.g. electrodes 9, 10, 11, 12) (Figures 3.a and 4.a).

5. For some time, the water accumulated near electrode #9 and did not seep through the soil.
This resulted in a rapid dip in resistance, shortly followed by a rapid increase. Based on this,
irrigation abnormalities such as water pockets can be spatially resolved (Figure 3.a, Figure
4.a).

6. Even before the application of water, various electrodes exhibited different resistance
measurements, constant with time (Figure 3). In order to account for these irregularities, a
ratio of present resistance values to the resistance of dry soil is computed. This process
separates the variation of resistance due to moisture from the variation of resistance due to
instrumental factors.

7. By assessing the differences in resistance measurements, quantitative soil moisture content

can be spatially evaluated by an equation: W = A*e-b*R , according to reference [2]. However,
this requires an accurate calibration of constants A and b.

 Conclusions

The results of soil resistance measurements indicate a possibility for an evaluation of Soil
Moisture distribution in three dimensions, enabling intelligent irrigation management.

A novel system capable of real time assessment of soil moisture content was built. The
system consists of a self-sustained data acquisition device, three-dimensional data analysis
algorithm, and real time visualization.

The system is very low cost compared to analogous instruments (can be under 15 dollars), is
capable of wireless data transmission, solar-powered, features one-button calibration, provides
accurate time and location from GPS signal, and can store historical data over long periods of
time.

The system is flexible and can be effectively used with data acquisition devices of various
geometries, including a laboratory research Smart Pot (created by me).

By significantly increasing the efficiency of Agricultural irrigation, this unique system

can help to preserve precious fresh water resources and potentially alleviate the problem
of hunger all around the world.

 Three-dimensional Visualization of Soil
Moisture

 Circuit Diagram

 References

B. Hanumantha Rao, D. N. Singh. "Moisture Content Determination by TDR and Capacitance

Techniques: A Comparative Study." International Journal of Earth Sciences and

Engineering. October 2011. pp. 132-137. Print.

Alex Mak, P. Eng. " Corrosion of Steel, Aluminum and Copper in Electrical Applications."

General Cable. Print.

Ceocor, Biarritz. " Corrosion Resistance of Stainless Steels in Soils and in Concrete." October

2001. Print.

Rachel Cardell, Mark Kranz. "A Reactive Soil Moisture Sensor Network: Design and Field

Evaluation." International Journal of Distributed Sensor Networks. 2005, pp. 149

162. Print.

Matthew D. Steinberga, Biserka Tkalcec, Ivana Murkovic Steinberg. "Towards a passive

contactless sensor for monitoring resistivity in porous materials." Sensors and

Actuators. Volume B. May 2016. Print.

Benjamin F. Schwartz, Mazdeline E. Schreiber. "Quantifying field-scale soil moisture using

electrical resistivity imaging." Journal of Hydrology. 2008. pp. 234 246. Print.

Susha Lekshmi , Maryam Shojaei Baghini. " A critical review of soil moisture measurement."

Elsevier Series Volume. April 2014. Print.

Ralf C. Dietrich, Anthony G. Bengough, Hamlyn G. Jones. "A new physical Interpretation

interpretation of plant root capacitance" Journal of Experimental Botany. August

2012. Print.

H.R. Bogena, J.A. Huisman, H. Vereecken. " Evaluation of a low-cost soil water content sensor

for wireless network applications." Journal of Hydrology. 344 . June 2007 32 42.

Print.

J. L. Davis, A. P. Annan. " Electromagnetic Determination of Soil Water Content: Measurements

in Coaxial Transmission Lines." Water Resources Research, vol. 16, no. 3, pp. 574-

582, June 1980. Print.

Ferhat Ozcep, Eray Yldrm, Okan Tezel. "Correlation between electrical resistivity and soil-

water content based artificial intelligent techniques." International Journal of

Physical Sciences. Vol. 5, pp. 47-56. January 2010. Print.

Sudhir Bhatt, Maulana Azad. " Correlation between electrical resistivity and water content of

sand a statistical approach." American International Journal of Research in Science,

Technology, Engineering & Mathematics. pp. 14-342. Print.

Hal Werner. " Measuring Soil Moisture for Irrigation Water Management." SDSU College of

Agriculture & Biological Sciences. Print.

Dr. Freddie R. Lamm. "Advantages and disadvantages of subsurface drip irrigation." Kansas

State University. Print.

Jos O. Payero, C. Dean Yonts, Suat Irmak, David Tarkalson. "Advantages and Disadvantages

of Subsurface Drip Irrigation." The University of NebraskaLincoln. Print.

Jeffrey P. Walker, Garry R. Willgoose, Jetse D. Kalma. "In situ measurement of soil moisture: a

comparison of techniques." Journal of Hydrology. (293), 2004. 8599, Print.

I. Gravalos, D. Moshou , S. Loutridis, Th. Gialamas, D. Kateris, E. Bompolas, Z. Tsiropoulos ,

P. Xyradakis, S. Fountas. " 2D and 3D soil moisture imaging using a sensor-based

 platform moving inside a subsurface network of pipes." Journal of Hydrology. 2013.

146153. Print.

I. Gravalos, A. Georgiadis, D. Kateris, O. Haralampous, T. Gialamas, P. Xyradakis, Z.

Tsiropoulos, E. Bompolas, E. Manolakoudis. " 3D soil moisture sensing and

imaging." Precision agriculture. 2015. Print.

"Water Use in California." Just the Facts. Print.

Alex Park, Julia Lurie.. "It Takes How Much Water to Grow an Almond" Mother Jones. February

2014. Print.

Erin Brodwin, Samantha Lee. "Chart shows how some of your favorite foods could be making

Californias drought worse." Business Insider April 2015. Print.

"Subsurface Drip Irrigation Demonstration and Research Project." The University of Arizona.

2001. Print.

T. L. Thompson, Pang Huan-cheng, Li Yu-yi. "The Potential Contribution of Subsurface Drip

Irrigation to Water-Saving Agriculture in the Western USA." Agricultural Sciences in

China. July 2009. pp. 850-854. Print.