Editors Note:   We are pleased to publish this excellent article by David Sedik describing his daughters high school senior project using the SwitchDoc Labs Smart Garden System V2.   This was a pleasure to read and understand their project.

Testing the Effect of Irrigation Scheduling on Green Onion Yields Using the Smart Garden System

Background

I worked for 20 years in the UN focusing on agricultural policy and food security. During my career, my family lived with me in various countries, first in Europe and then the Middle East. While the European countries where we lived were somewhat like the United States, the years we spent in Egypt were quite different. My daughter, now 18, spent 2 years with my wife and me in Egypt where she learned firsthand how water scarcity and water quality affect people’s lives. As an agricultural development economist, I often spoke to my daughter about the problems of water scarcity, irrigation and water waste, comparing the flood irrigation I saw in Egypt with the drip irrigation projects I had seen in Israel.

Perhaps those two years we spent in Egypt were a factor in my daughter’s decision to propose an experiment to test the effectiveness of irrigation methods on the yields of plants for her senior biology project. When I told her that perhaps we could conduct a test of irrigation methods using a raspberry pi, her eyes lit up. The chance to work with the pi and even program made this an interesting experiment for her.

This project required Minimal knowledge or electronics and no soldering.  Power source was 5V making it a great project for middle school and high school level students.

For another Smart Garden System example, check out the Hops Garden here. 

Irrigation methods

Most farms in temperate regions practice rain-fed farming to ensure crops get the right amount of moisture at the right time. This is difficult not only because it is difficult to predict the weather. Different crops require different amounts of water overall and plants require different amounts of water during their growth. In order to provide more control over watering, farmers practice irrigation all over the world, including Egypt. There are three main types of irrigation:

Surface irrigation. Water is applied and distributed over the soil surface by gravity. In Egypt, basin (or “flood”) irrigation has historically been used in small areas having level surfaces that are surrounded by earth banks. The water is applied rapidly to the entire basin and is allowed to infiltrate. Furrow irrigation is conducted by creating small parallel channels along the field length in the direction of predominant slope. Water is applied to the top end of each furrow and flows down the field under the influence of gravity.

Sprinkler irrigation. Sprinkler irrigation applies water to soil by sprinkling or spraying water droplets from fixed or moving systems.

Drip irrigation. This method applies frequent, small water applications by dripping, bubbling or spraying, and usually only wets a portion of the soil surface in the field. This method is the most water-conserving. Thus, for dry climates such as Israel or Egypt it is preferred.

Drip irrigation and the problem of scheduling

The Smart Garden System lends itself well to drip irrigation. However, even with drip irrigation there is still the issue of how much water to apply and how often. The Smart Garden System offers two ways of scheduling watering. The first is timed watering in which you choose the intervals and the time you want to water your plants, say once per day for 15 seconds. The second way to schedule watering is to install a capacitive moisture sensor to constantly monitor soil moisture and direct the system to keep soil moisture at a certain set level. All you need to do is to set the moisture level of the soil you want (say, 65%) and the system will maintain soil moisture at that level.

Neither I nor my daughter knew how much to water onion plants, nor did we know what moisture level the soil should be kept in order to water the plants properly. She proposed two management methods for water delivery: the first was to let the raspberry pi handle the watering schedule by setting the soil moisture level at 65% and leaving it at that. We called this the moisture sensor method. The second was to schedule plant watering the old-fashioned way, by “walking” from plant to plant a few times per week, feeling the moisture level with her thumb and guessing how much more or less the plants needed to be watered. This was called the scheduled method. This method called for much more direct management of the irrigation process, but it also gave my daughter more knowledge of what was happening with her crops.

 

The experiment

The purpose of the experiment was to test which of these two methods would deliver a better result measured by the average mass of the 36 plants under each method at the end of six weeks. For the experiment, my daughter seeded 72 starter pots with green onion seeds. For 36 of the plants, water was delivered according to the capacity moisture sensor method. For the other 36 plants, my daughter monitored the moisture of the soil with her thumb and reset the irrigation watering based on how dry or wet the soil was.

 

Below you can find a photo of the timed irrigation line with 36 drippers. In addition, there were 3 other lines independently controlled by three capacitive moisture sensors, each controlling the water to 12 plants. The other photo shows the capacitive moisture sensor that controlled watering to one row of 12 plants.

 

Formally, the experiment was a test of which of the two methods would deliver the greater yield, defined as the mean weight of the plants under the two methods of water management. If the average of the 36 moisture-sensor plants was statistically greater than that for the scheduled plants, then we concluded that the moisture-sensor was superior.

 

Assembly of the experiment

Here is a photo of the fully assembled experiment in action. The top shelf of the rack contains (left corner) the remote boards which connect to the raspberry pi via Wi-Fi and control the lights and pumps. You can see the USB lines on the left and right sides that go into the 5-gallon pails where the pumps were located.

Below you can see a close-up of the housing for the remote boards (left).

 

On the second shelf from the top is the raspberry pi (right side). You can also see a closeup of the pi in the below photo. The ribbon cable emerging from the top of the case is the camera which was positioned on top of the plants to photograph them every few hours. The grow lights were also attached to the second shelf and were lit 18 hours per day controlled by the remote board according to instructions received from the raspberry pi. The pi also received information from a number of sensors. Since we live in an apartment, we set up the outdoor SkyWeather unit inside our apartment to monitor temperature and humidity levels. You can see it in the photo of the experiment on the right side. However, we also used the indoor unit to monitor temperature and humidity levels as well. Both communicate with the pi via radio antennae. You can see the radio receiver on the pi in its case to the right of the Pi.

What kind of information did the Smart Garden System 2 give us?

We monitored the information sent to the pi from The SkyWeather unit and from the Pi itself. Below, we can see the data received from the moisture sensors displayed on the dashboard which my daughter was able to monitor on a daily basis. As I indicated above, the moisture level was set to 65% and you can see that the system controlled the level of moisture quite well.

 

Indoor humidity and temperature were monitored as well using the SkyWeather2 system.

 

The Smart Garden System dashboard also allowed me to monitor when the pumps were turned on an off to keep the moisture level of the soil at 65%.

 

Results of the experiment

After 46 days the plants were harvested. The below photograph shows the plants ready to be harvested and weighed.

The final figure below shows the results of the experiment, the mean weight of the 36 plants raised using the timed irrigation scheduling method (left bar chart) and the mean weight of the 36 plants raised using the sensor-based irrigation scheduling method (right bar chart). The blue bars represent the means of the two samples. The black whiskers show the 95% confidence interval around the point estimates, the margin of error. Because the two margins of error do not overlap, we can say that the two means are different with a high degree of confidence.

 

 

Interpretation

What does this all mean? Here I go beyond what my daughter handed in as her experiment. Recall before that I stated that neither my daughter nor I knew how much to water onion plants, nor did we know what moisture level the soil should be kept in order to water the plants properly. We basically guessed at the moisture level we set for the moisture-controlled plants based on what we read in the SGS2 manual. For the timed irrigation plants, she actually “walked the crop” learning about the moisture level of the soil by pressing her thumb into it and sensing the degree of soil moisture. We also watched the timed irrigation plants carefully to check for signs of water stress, such as sagging stems and slow growth. Though we did not measure it, my guesstimate is that the timed irrigation plants received more water than the moisture-controlled plants, because I also occasionally checked the moisture level of the moisture-controlled plants as well. It just goes to show that a good farmer doesn’t leave watering entirely to technology. There is no substitute for walking the crop and constantly updating crop care based on observation.

 

Technical requirements to replicate this experiment

• Raspberry Pi with Smart Garden System 2 software loaded on a 16 GB SD card
• The Smart Garden System (V2)
• SkyWeather system to monitor indoor humidity and temperature.
• Seeding trays with seedling soil
• Rubber boot tray to put under the seeding trays for drainage
• 5-gallon plastic buckets for irrigation watering
• Green onion seeds
• Grow lights (purple spectrum filter)
• Gram scale to weigh the green onions
• PC, laptop running windows 10 or mac
• 1/8”x1/4” clear PVC tubing
• USB water pumps (4)
• 1/2 GPH Pressure Compensating Drippers
• Gram Scale 600g x 0.01g
• ¼” barbed end plug (6), ¼” barbed 3-way T connector (72) and ¼” barbed 4-way connector (1) fittings
• Irrigation Support Stakes for 1/4-Inch Drip Tubing