Tensiometer-based irrigation of cut-flower roses

1994 Project Report to the California Cut-flower Commission

Reduction of run-off in greenhouse cut flower crops through automated irrigation based on soil moisture tension.

Lorence R. Oki, J. Heinrich Lieth and Steve Tjosvold

Dept. of Environmental Horticulture, University of California, Davis, CA 95616-8587

and University of California Cooperative Extension, Watsonville, CA

P.I.: Dr. J. Heinrich Lieth; Cooperators: Loren Oki and Steve Tjosvold

Report date April 17, 1995

Bookmarks to move around quickly in this report:
Experimental setup Results and Discussion Figures:

1, 2, 3, 4, 5


1, 2, 3, 4

Introduction and Background

Agricultural use of water is an important issue in California with considerable attention being focussed on greenhouse growers. Water use in greenhouse production is frequently at excessive levels, causing high amounts of run-off water. While generation of some run-off is generally unavoidable and needed to prevent build-up of salts, excessive amounts of run-off represent an inefficient use of the resource and a significant source of ground water pollution since such run-off tends to be laden with fertilizer and pesticides. Growers have responded to this by attempting to make their irrigation systems more efficient. This may be accomplished by the placement of the applied water (drip irrigation), containment of effluent (capturing and reuse), or more precise methods of controlling irrigation (duration and frequency). The project discussed in this report focusses on developing new tools for growers attempting to improve irrigation efficiency and reduce run-off in in-ground cut-flower production.

A problem that is common with any irrigation system is determining when to irrigate and the amount of water to apply. Growers are generally "in tune" with their crops and may rely on their observations to decide when to water. Frequently, however, such visual observation corresponds to levels of water stress that are not optimal for crop production. Also, the moisture status may change rapidly so that watering may be required before a grower notices the need. This could be alleviated through an irrigation system which uses sensors to monitor the soil moisture status of the crop.

Also, when a decision is made to water, the amounts of water applied are generally excessive. In cut flower production, Tjosvold and Schulbach (1991), while examining alternative approaches to irrigation scheduling, found that growers commonly irrigate too infrequently and that too much water is applied when irrigations are made. In in-ground production it is very difficult to know when an adequate amount of water has been delivered. While the amount of excess water varies from grower to grower it is frequently substantial. A system that precisely monitors and controls the amount of applied water and the changing soil moisture status during an irrigation, may reduce the amount of generated run-off.

Sensors, such as tensiometers, may be used to measure soil moisture content and to make irrigation decisions. Tensiometers that have been commercially available were designed for use in field soils and their gauges have been optimized for these soils by having a wide tension range (0 to 100 kPa). However, previous research has shown that the optimum tension levels for producing plants in containers filled with highly-amended media are in the range of 1 to 5 kPa (Kiehl et al., 1992). Plaut et al. (1976) found that the moisture tension in cut-flower rose production should be below 6 kPa.

Figure 1: The design of the computer controlled irrigation system for monitoring the moisture content of the soil in the root zone with a tensiometer.

One problem with the use of field tensiometers in amended media, is their slow response. Burger and Paul (1987) developed a tensiometer for use in such soils. These tensiometers responded rapidly to changes in soil moisture tension (SMT) and were used in research to maintain a constant moisture level in potted plants (Lieth and Burger, 1989). This work demonstrated that the amounts of water used in greenhouse potted plant production is excessive and plants of commercial quality can be grown using substantially less water than previously thought necessary. Research supported by the California Association of Nurserymen has developed an adaptation of this system for commercial use and has shown that potted plants can be grown with significant reductions in water use (70-90%) without detrimentally affecting crop quality (Lieth, 1990; Lieth et al., 1990). The responsiveness of these "high flow" tensiometers is significantly better than the traditional field tensiometers. While tensiometers have long been used in making the decision "when" to apply water, these "high flow" tensiometers are able to also aid in the decision of "how much" to apply. When water is applied to the soil, these tensiometers are able to rapidly detect the addition of water and can stop the water application when a specified moisture level is reached. The objective of this project was to develop a complete system for use in cut-flower rose production, to test the system under commercial conditions and to compare its performance with standard irrigation procedures.

Experimental setup

This work, funded by the California Cut Flower Commission, adapts the system developed for potted plant production for use in cut flowers. The system consists of commercially available soil tensiometers fitted with pressure transducers (to convert pressure into a voltage signal), signal amplifiers, an analog-digital signal converter, a computer, and the irrigation system with automatic valves (Fig. 1). The equipment is controlled by a custom-written program running on the computer. The program requires setting a high tension set-point (when irrigation is turned on) and a low tension set-point (when irrigation is turned off). Safety features include minimum and maximum on-times and a maximum elapsed time between irrigations to circumvent a failure of sensors.

The scenario for soil moisture tension-based irrigation is: the computer continuously monitors the tensiometers placed in the root zone of the crop. As water is extracted from the soil, the tension in the soil and the tensiometer increases. When the high tension set-point is reached, the automatic valve is opened. As water is applied, the soil moisture tension decreases and the valve is closed when the low set-point is attained.

The sensors used are tensiometers from Irrometer Corporation (Riverside, CA) manufactured with a "high flow" ceramic porous cup. We modified these by attaching a transducer to the tensiometer using brass fittings. This adaptation was simplified with the use of off-the-shelf parts and electronics purchased directly from vendors. The simplicity of the adaptation and the availability of parts is important since growers will be able to readily adopt the technology. Tests of the modified tensiometers has demonstrated that they respond significantly better than the standard field tensiometers and can be used in controlling automated irrigation systems.

The dynamics of moisture tension and moisture content were explored using 'Cara Mia' roses growing in ground beds at UC Davis. Irrigation was by spray emitters mounted on a perimeter system. Four different tensiometers were inserted in relatively dry soil: 6" and 18" Irrometer tensiometers with a high flow ceramic tip, a 12" Irrometer tensiometer with a standard tip, and a tensiometer of the type built by Lieth and Burger (1989) for use with potted plants (U.C. tensiometer). The Irrometer tensiometers were modified as described above. Before, during, and after irrigations (ranging in duration from 10 to 40 minutes) moisture tension were recorded with a data logger.

Figure 2. Comparison of responses of tensiometers with different ceramic porous cups at different depths to changes in soil moisture.

The result of this verified that the irrigation system was performing as designed and that the modification of the off-the-shelf components was working as planned. The response for the standard field tensiometer (Fig 2 C) was substantially slower than the "high-flow" versions (Fig 2 A, B, and D).

The sensors modified with the high flow tips responded to soil moisture changes rapidly. The rapid drop in tension registered by the high-flow tensiometer at 10 cm depth indicated that the system would be able to follow the moisture tension during an irrigation and be used to determine when adequate amounts of irrigation water had been delivered.

It was also possible to monitor the plants' extraction of water from the root zone soil using the tensiometers (Fig. 3). The slope of the tension curves are steeper when the light levels are higher (corresponding to more rapid removal of water from the soil by the plants).

Figure 3

Figure 3: Soil moisture tension levels monitored by four tensiometers over a three-day period.

Watsonville Nursery, a prominent rose grower (in Watsonville, CA), was selected as the site for testing the system. This nursery has a reputation for excellent rose quality and good nursery practices. They have, on staff, someone who can assist with problems related to soil, water and plant nutrition. They are attentive to irrigation practices and have, in the past, worked with field tensiometers.

One bay of one greenhouse, planted with 'Kardinal', was selected as the site for the experimentation. The plants on one side of the center aisle were irrigated by the grower; the plants on the other side were irrigated based on moisture tension using the system we developed and tested at UC Davis. Both sections were equipped each with three tensiometers. These were the high-flow type with porous cups at 5, 24, and 40 cm depth. Data were logged electronically at half-hour intervals.

Since the automated irrigation system was to perform without any interaction, the grower imposed two requirements on us: 1. the system could only water at times when it would not interfere with other irrigation and never after 5 p.m. due to potential problems with high humidity in the late afternoon and early evening. Thus the software was modified to irrigate only in a specific time-frame. This "irrigation window" was from 6 a.m. to 7 a.m. each morning. Within this timeframe the system was set up to irrigate when a tension of 5 kPa was reached and to then irrigate until either a maximum on-time of 10 minutes had elapsed or until a tension of 1 kPa was reached. The tensiometer at 24 cm depth was used as the control point.

The particular greenhouse was selected since it allowed setting up two blocks of almost-identical rose plants. All plants had been planted several months earlier and were about 1.5 m tall at the start, and ready to go into full production. Although it was not apparent when the experiment was set up, the pinching pattern was slightly different between the two areas in the greenhouse with the flush of shoots on the growers plants being about 1 week earlier.

Water meters were installed at each section to monitor the actual amount of irrigation water applied to each section. The crop sections were measured to determine amount of area dedicated to each test section. At each harvest, all the harvested shoots were counted in each section and noted on a tally sheet by the harvesters.

On two occasions the quality to the rose production was tested by taking one bundle of harvested roses from each section and measuring the length of each stem.

Results and Discussion

The tensiometers at three depths provided a clear understanding of how tensions change in response to irrigations (Fig 4). Immediately following irrigation the tension dropped when the wetting front in the soil passed the tensiometer; with resulting tensions near 0 indicating that the soil was briefly saturated and then at field capacity. As the plants and surface evaporation remove the water the tension increases. The resulting pattern appears as a stair-step pattern with rapid increases in tension during the day and only minor changes at night (Fig 4 top curve).

Figure 4

Figure 4: Moisture changes at three soil depths during irrigation and subsequent drying

The moisture changes much more rapidly near the soil surface than at lower depths. Soil moisture tension at all depths decreases during the day, but changes only at the middle depth at night. Diurnal fluctuations at the 24 cm. level are the most pronounced probably because there are more roots present here than at the depths of the other sensors. At this depth, the most rapid increase in tension is during the period of highest water demand from the plant (afternoon) and decreases in tension occur during the night as water moves from deeper soils The fact that his system can track these changes represents a significant improvement over field tensiometers and is a vital aspect of the functionality of this system.

These sharp peaks became particularly prevalent in the tension-based irrigation treatment (Fig 5). This had a significant effect on the operation of the irrigation system since many of these peaks extend significantly higher than 5 kPa. When the tensions managed to return to a level below 5 kPa by 6 a.m. the next morning (e.g. on 10/12/94) there was no irrigation despite the high tensions during the previous day. Irrigations we done whenever the tensions did not go back below 5 kPa by 6 am (e.g. Fig 5, 10/13 and 10/17).

Figure 5

Figure 5: Changes in soil moisture during irrigation controlled by soil moisture tension measurements. The arrows indicate irrigations with the duration the valves were turned on

The applied water reaches the shallower depths but the amount is not great enough to affect the deeper soils. The tensions at 40 cm. fluctuate diurnally but do not decrease following irrigation as at the shallower soil depths. This is the desired performance of the irrigation control system, since the amount of water applied is not excessive and may not produce run-off.

Readings from the water meters show that there was less water applied, although at greater frequency, to the tension-based treatment than the grower beds (Table 1). The summary of data shows that an average of about 300 gallons of water were applied per irrigation per 1000 sq. ft. of greenhouse by the soil moisture tension-based system compared to about 850 gallons for the grower controlled system. On a per day basis, this translates to about 50 gallons of water used per day per 1000 sq. ft. for the soil moisture tension-based system versus about 80 gallons for the grower controlled system.

Table 1. Summary of irrigations during the test period
Total water applied Number of irrigations

Applied per irrigation

(gal/1000sq. ft.)

Average water use per day

(gal/1000 sq. ft.)

gallons gal/1000 sq. ft.
Grower 9042 12735 15 849 77.2
Test 4989 8179 28 292 49.6

Analysis of stems harvested from both the test and the grower beds show a dramatic effect on the production of flowers. The number of flowers harvested from the tensiometer-based treatment was much higher than those harvestd in the block being controlled by the grower (Tables 2). At the same time the quality was the same (Table 3). Although the pinching pattern was slightly different in the two blocks, it is not likely that this had a significant effect on the difference between the counts from the two blocks.

Table 2. Quantity of stems harvested


stems harvested by grower stems harvested in test plot
total number number/1000 sq. ft. total number number/1000 sq. ft.
September 1171 1649 2500 4098
October 1175 1655 1783 2923
November 72 101 912 1495
Total 2418 3406 5195 8516

Table 3. Lengths (cm) of harvested stems. (Sample size = 50 stems.)


Grower Test
mean stem length std dev mean stem length std dev
Feb 6 62.88 7.19 61.28 7.36
Feb 13 63.56 7.45 60.16 7.52

When combining the harvest and water use data (Table 4) a very peculiar result emerges: the number of stems harvested per gallon of water used is very different in the two areas. The crop irrigated based on moisture tension produced 1.2 stems per gallon of water used compared to 0.4 stems per gallon resulting from the grower-controlled system. This is a tremendous difference and we must run additional tests before we will allow ourselves to believe this result. It is difficult to accept that the tension-based irrigation system would be 3 times more efficient. Another way to look at this is that the soil moisture tension-based system produced flowers using one third of the amount of water and fertilizer.

It was hoped that the amounts of subsurface run-off water could be estimated by using data from the meters and the deepest (40 cm.) tensiometers. However, as can be seen in figure 3, the soil at this depth is almost always saturated (Fig 4 and 5), making it impossible to estimate run-off water as expected.

Table 4. Analysis of harvest and water usage data
Grower Test
Period stems

/1000 sq. ft.

gal. used

/1000 sq. ft.

no. stems/gal no. stems

/1000 sq. ft.

gal. used /1000 sq. ft. no. stems/gal
9/6- 9/30 1649 3554 0.46 4098 2018 2.03
10/1-10/31 1655 4338 0.38 2923 3836 0.76
11/1-11/16 101 724 0.14 1495 1230 1.22
Total 3406 8615 0.4 8516 7084 1.2

The soil moisture tension-based system is efficient in controlling irrigation for the production of cut-flower roses. The tensiometers modified with transducers and high flow ceramic tips are able to effectively measure soil moisture tensions in the root zone of plants grown in beds and are capable of responding to rapid changes in soil moisture. The system, using these modified tensiometers, is able to continuously monitor the moisture level of the soil and will not only demand irrigation when the soil requires water, but turn off the water when an adequate amount has been applied.

The environmental benefit of this system is a more efficient use of water compared to a grower controlled system which may result in the reduction in run-off water. The reduction in water use also has economic benefits to the grower in reducing the cost to produce and use water (pumping, etc.). Additional costs savings are realized if a liquid feed program is used since the amount of fertilizer applied is reduced.

Literature Cited:

Burger, D.W. and J.L. Paul. 1987. Soil moisture measurements in containers with solid-state, electronic tensiometers. Hortscience. 22(2):309-310.

Kiehl, P.A., J.H. Lieth, and D.W. Burger. 1992. Growth response of chrysanthemum to various container medium moisture levels. J. Amer. Soc. Hort. Sci. 114(2):224-229.

Lieth, J.H. 1990. Irrigation Technology: Reducing run-off in greenhouse and nursery production. Proceedings of the Society of American Florists.

Lieth, J.H. and D.W. Burger. 1989. Growth of chryssanthemum using an irrigation system controlled by soil moisture tension. J. Amer. Soc. Hort. Sci. 114(2):387-392.

Lieth, J., D. Burger, P. Kiehl, S. Tjosvold, G. Vogel. 1990. Reduce run-off from your potted crops by watering based on soil moisture. Grower Talks. September 1990, p.24-32.

Plaut, Z., N. Zieslin, and N. Levev. 1976. Effect of different soil moisture regimes and canopy wetting on 'Baccara' roses. Scientia Hort. 5: 277-285.

Tjosvold, S.A. and K.F. Schulbach. 1991. How to reduce water use and maximize yields in greenhouse roses. Califonia Agriculture. May-June 1991, p. 31-32.

For more information or feedback contact jhlieth@ucdavis.edu