Precision Irrigation Scheduling Using ECH2O Moisture ... - MDPI

water Article

Precision Irrigation Scheduling Using ECH2O Moisture Sensors for Lettuce Cultivated in a Soilless Substrate Culture Zhigang Liu 1, * and Qinchao Xu 2 1 2

*

School of the Environment and Safety Engineering, Institute of Environment Health and Ecological Security, Jiangsu University, Zhenjiang 212013, China College of Engineering, Huazhong Agriculture University, Wuhan 430070, China; [email protected] Correspondence: [email protected]; Tel.: +86-511-8879-0955





Received: 17 March 2018; Accepted: 21 April 2018; Published: 25 April 2018

Abstract: Soilless culture has become an effective technique to avoid continuous cropping obstacles in protected horticulture. The reliable measurement of substrate moisture and a rational irrigation are difficult tasks because of the low water-holding capacity of the substrate. Our objectives were to study the irrigation scheduling based on the ECH2 O moisture sensor(EC-5), using a matched model of wetting pattern and lettuce root zone in the substrate under drip irrigation. The EC-5 sensor was designed to connect to a controller, and a threshold value of 0.14 cm3 /cm3 was set for irrigation scheduling. The controller turned on the irrigation system via communication with a solenoid valve on the irrigation line and with the EC-5 sensor in response to a threshold value and stopped when the overlap area of the wetting pattern and crop root zone was more than 90%. The EC-5 sensors were installed at a horizontal distance from each plant and depth of 3 and 4 cm, respectively, under the substrate surface to the check substrate moisture for lettuce cultivation, and at (3,15) cm or (6,15) cm to monitor leakage. These parameters were determined by simultaneously considering the distance from the plants, the depths of effective root water extraction, and the region of substrate wetted volume under drip irrigation. Leakage occurred during each irrigation process, but the leakage ratewas15.7% lower than that of conventional irrigation, as a result of irrigation scheduling in the presence of the EC-5 sensors. Keywords: substrate; irrigation scheduling; lettuce; moisture measurement

1. Introduction In recent years, regulations on agriculture water use have become stricter because of increased urbanization and population growth [1] and are becoming worse because of the recent rapid expansion of land area occupied by greenhouses [2]. Several institutions have gained more utility to organize plant water use and to properly schedule agricultural, municipal, and residential irrigation by taking environmental measurements such as of evapotranspiration and of soil water content [3–5]. This method not only mitigates the severe water deficit, but also saves growers money, by ensuring that plants are not excessively irrigated [6]. However, most of these studies were conducted on soil. In fact, the soilless culture as a growing technique has been largely used in greenhouse because of its multiple advantages [7–9]. Greenhouse crops are commonly irrigated on the basis of the visual appearance of the substrate or plants or the use of irrigation timers, but these methods are not accurate and will not result in efficient irrigation [1]. To irrigate a crop grown in substrates with the right amount of water and achieve high yields, the water must be precisely applied to the crop because the water-holding capacity of the substrate is very low.

Water 2018, 10, 549; doi:10.3390/w10050549

www.mdpi.com/journal/water

Water 2018, 10, 549

2 of 10

When the volume of water absorbed exceeds the allowable water depletion of the substrate, the crop may suffer severe stress, and the yield is reduced [10]. For this reason, it is important to accurately measure the substrate water content to decide when and how much irrigation is required. The measurements and estimates of the water content for use in irrigation scheduling in the past are not practical. The main reasons are unreliable measurements, unsuitable size, and high costs. For example, the gravimetric water content measurement is destructive and requires a lot of time to dry the samples. Neutron probes require a large space for installation and are influenced by the soil properties [11], and time-domain reflectometry (TDR) probes are expensive [12–15]. Although a tensiometer can be easily placed in a substrate with high porosity, it causes erroneous and unreliable measurements. In recent years, electromagnetic techniques, such as ECH2 O probes and Theta probes have been used to estimate the substrate water content by indirectly measuring the dielectric permittivity of the substrate [16–18]. Estimates of water content based on electromagnetic techniques provide real-time and in situ measurements. Probes are available in convenient sizes for substrate cultivation troughs or growbags. The electromagnetic techniques have a good application potential in the measurement of substrate water content. The root system is the important organ to obtain water for a plant; the root growth and its distribution determine the ability of a plant to absorb the water from the wetting pattern, which then affects the crop biomass growth and grain yield [19]. The wetting pattern depends on the substrate properties and the irrigation application scheme [20]. The amount of irrigation water calculated on the basis of evapotranspiration is difficult to be administered, because the high porosity, the low bulk-density, and the initial water content of the substrate facilitate the formation of a finger stream, resulting in a reduced size of the wetting pattern and in increased losses of water, which then affects the crop root distribution and grain yield. A better understanding of the substrate wetting pattern and crop root zone are crucial for developing effective irrigation methods. However, to our knowledge, few studies have yet investigated the coupling of wetting pattern and crop root zone in substrates to design irrigation scheduling. Lettuce is a shallow rooted and moisture-sensitive crop, and moisture stress can cause negative impacts on its yield [21]. The aims of this paper were to study the irrigation scheduling of lettuce on the basis of (i) the nondestructive monitoring of the substrate water content; (ii) the matched model of wetting pattern and crop root zone for lettuce cultivated in a substrate under drip irrigation. 2. Materials and Methods 2.1. Substrates Preparation Aground vinegar residue (50%) and peat (50%) with a bulk density of 0.184 g/cm3 was used as growth media. The saturated hydraulic conductivity was found to be 0.051 cm/s by cutting ring. 2.2. Sensor Description and Operation ECH2 O(EC-5) is often used for sensor network applications because it is cheap, easy to use, and has low energy consumption. The EC-5 sensor is small (8.9 cm length, 1.8 cm width, 0.7 cm depth) and suitable for application in greenhouses. Since the EC-5 sensor is sensitive to the substrate temperature, the compensation method was used to obtain more reliable substrate water content estimates from the sensors. The compensation model considering temperature for EC-5 sensor estimates θ(θ sensor ) can be described as: θsensor = U − aT − b/c (1) where θ sensor is the substrate water content (%), U is the output voltage of EC-5 (mv), T is the substrate temperature (◦ C), a, b, c are constant and can be obtained by EC-5 sensor temperature calibration; the details of the process can be found in reference [16].

Water 2018, 10, 549 Water 2018, 10, x FOR PEER REVIEW

3 of 10 3 of 9

Thisstudy studywas wasconducted conductedinina agreenhouse greenhouse with a cultivation tank m length, 0.7width, m width, This with a cultivation tank (17(17 m length, 0.7 m 0.4 0.4 m depth).A filter collection device under the cultivation could be used to collect the leakage, m depth).A filter collection device under the cultivation could be used to collect the leakage, and a and a high-precision water on meter onbranch each branch wasto used to measure the applied Lettuces high-precision water meter each was used measure the applied water.water. Lettuces were were planted onspaced lines, spaced 0.2 mand apart, and the EC-5were sensors were atinstalled at fourdepths different planted on lines, 0.2 m apart, the EC-5 sensors installed four different (4, depths (4, 8, 15, and 20 cm) and at three different distances from the base of the plants (3, 6, and 9 cm) 8, 15, and 20 cm) and at three different distances from the base of the plants(3, 6, and 9 cm)(Figure 1). (Figure In order make the estimations EC-5 sensorand estimations and control irrigation, a solenoid valve In order 1). to make the to EC-5 sensor control irrigation, a solenoid valve and multiple and multiple mustto beaconnected a controller (WLT series products, brand) sensors (EC-5)sensors must be(EC-5) connected controller to (WLT series products, Weiqian brand)Weiqian for retrieving 3, calibration for retrieving the sensorsand estimations andirrigation. scheduling threshold value of 0.14 cm3 /cm3 , the sensors estimations scheduling A irrigation. threshold A value of 0.14 cm3/cm calibration and equations, and parameters of the EC-5were sensor wereinsaved in the controller programmatically. equations, parameters of the EC-5 sensor saved the controller programmatically. The The irrigation water was applied at a rate of 0.15 L/h, 0.35 L/h, and 0.5 L/h through an emitter irrigation water was applied at a rate of 0.15 L/h, 0.35 L/h, and 0.5 L/h through an emitter (SLD series (SLD series products, ShunLv brand) bythe regulating theThe pressure. Thethe rootwetted zone, the wetted and products, ShunLv brand) by regulating pressure. root zone, radius, andradius, the depth thedifferent depth atgrowth different growth stages of thewere lettuce were measured by method, the dig method, the water actual at stages of the lettuce measured by the dig and theand actual water content was measured the drying method. replications for experiment. each experiment. content was measured by theby drying method. Three Three replications were were used used for each

Figure 1. 1. Schematic Schematic diagram diagram of of the the experiment; experiment; (1) (1) Programmable Programmable controller; controller; (2) (2) Pressure-regulated Pressure-regulated Figure water source; (3) Solenoid valve; (4) Drip emitter; (5) EC-5 sensors. water source; (3) Solenoid valve; (4) Drip emitter; (5) EC-5 sensors.

2.3. Wetting Wetting Pattern Pattern Model Model 2.3. There are many substrate wetting pattern under dripdrip irrigation. Someof these There manyfactors factorsaffecting affectingthe the substrate wetting pattern under irrigation. Someof factorsfactors are: saturated hydraulic conductivity (Ks ), initial moisturemoisture content (θ), irrigation these are: saturated hydraulic conductivity (Ks),volumetric initial volumetric content (θ), amounts (V), irrigation flow (q), flow and emitter (z)depth [22]. (z) The[22]. wetted and(W) theand wetted irrigation amounts (V), irrigation (q), anddepth emitter The radius wetted(W) radius the depth (D) can be described as: wetted depth (D) can be described as: (n1 − 13 ) n1 K s θ ( n  1 ) (2) W = A1 V n1 ( K s) 1 3 W  A1V ( qz ) (2)

qz

1

K s θ ( n2 − 3 ) D = A 2 V n2 ( ) 1 Kqz ( n2  ) n

(3)

D  A2V ( ) (3) where W is the wetted radius (cm), D is the wetted qz depth (cm), V is the irrigation volume (L), Ks is the saturated hydraulic conductivity (cm/h), q is the emitter discharge (L/h), z is the depth of emitter where W is the wetted radius (cm), D is the wetted depth (cm), V is the irrigation volume (L), Ks is insert into the substrate (cm), A1 , A2 are constants, and n1 , n2 are exponents of equation. the saturated hydraulic conductivity (cm/h), q is the emitter discharge (L/h), z is the depth of emitter The models (2) and (3) are for predicting the wetted radius of the substrate surface and the wetted insert into the substrate (cm), A1, A2 are constants, and n1, n2 are exponents of equation. depth in the substrate. The wetting pattern under the substrate surface was estimated by an empirical The models (2) and (3) are for predicting the wetted radius of the substrate surface and the model, according to the experiment data and as demonstrated, as: wetted depth in the substrate. The wetting pattern under the substrate surface was estimated by an   1) empirical model, according to the experiment data and as demonstrated, as: ( n1 − (n2 − 13 ) (n2 − 13 ) 3 Y = [(W − D ) cos( a) + D ] sin( a) = ( A1 V n1 ( Ksθ − A2 V n2 ( Ksθ ) cos( a) + A2 V n2 ( Ksθ sin( a) (4) qz ) qz ) qz ) 2

s

3

1 1 1  Ks ( n1  3 ) Ks ( n2  3 ) Ks ( n2  3 )  Y  (W  D) cos(a)  Dsin(a)  ( A1V n1 ( )  A2V n2 ( ) ) cos(a)  A2V n2 ( )  sin(a) qz qz qz  

(4)

Water 2018, 10, 549

4 of 10

  (n1 − 13 ) (n2 − 13 ) (n2 − 13 ) X = [(W − D ) cos( a) + D ] cos( a) = ( A1 V n1 ( Ksθ − A2 V n2 ( Ksθ ) cos( a) + A2 V n2 ( Ksθ cos( a) qz ) qz ) qz )

(5)

where X and Y are the wetted radius and the wetted depth at a specific application time and at any position of the wetting front, and α is the angle between the wetted radius (W) and the wetted depth (D). 2.4. Root Zone Model Roots are the sole part of the crops that take up water and nutrients from the soil. Root growth and distribution are important plant traits, highly dependent on the environment where the crop is growing [23]. The root depth and root zone radius are used to describe the root system, are often found to increase with temperature [24], and can be described with a single function. The root depth (Z) and root zone radius (R) can be described as: dZ/dt

= bz ( Ta − Tb )

(6)

R = dZ3 + eZ2 + f Z + g

(7)

where dZ/dt is the growth rate of the root depth (cm/day−1 ), bz is a constant (cm◦ C−1 day−1 ), Ta is the daily mean air temperature (◦ C), Tb is the base temperature for root growth (◦ C), Z is the root depth (cm), R is the root zone radius (cm), d, e, f, g are coefficients that can be described as: C = hDD 3 + iDD 2 + jDD + k

(8)

where C is an undetermined coefficient d, e, f, g, h, i, j, k are the corresponding constants of d, e, f, g, respectively, according to Table 1. DD is the cumulative temperature (◦ C) and can be described as [19]:    0 DD = ∑ T = Ta − Tb   T −T m b

( Tb ≥ Ta ) ( Tb ≤ Ta ≤ Tm ) ( Ta ≥ Tm )

(9)

where Tm is the maximum temperature for root growth (◦ C). Table 1. The coefficients of the root zone radius model.

Coefficient d e f g

Root Zone Radius (R) h

i 10−10

5.04 × −1.15 × 10−8 4.69 × 10−8 −2.67 × 10−8

j 10−7

−5.61 × 1.48 × 10−5 −7.26 × 10−5 3.58 × 10−5

k 10−5

4.24 × −4.48 × 10−3 2.98 × 10−2 −9.95 × 10−3

0.08 −0.53 −0.34 −0.46

2.5. Matched Model of Wetting Pattern and Crop Root Zone A reasonable matching between substrate wetting pattern and crop root zone is one of the keys to irrigation scheduling using EC-5 sensors. In order to attain a reasonable match and irrigation, the following matching principles were proposed: (i) the water distribution uniformity and the overlap area of wetting pattern and crop root zone is more than 90%; (ii) the water in the crop root zone is enough for the growth of crops; (iii) the leakage from the crop root zone should be as little as possible. The matching results directly affect the irrigation scheduling and crops growth. The substrate must be fully wet before field transplanting. The substrate wetting pattern and crop root zone were regarded as relatively evenly distributed, and a proper overlap degree section area of substrate wetting

Water 2018, 10, 549

5 of 10

Water 2018, 10, x FOR PEER REVIEW

5 of 9

pattern and crop root zone (overlap(R,W)) was used to indicate the matching results and can be described as: area( R,W ) overlap R,W W)) =  area( R, W ) (10) overlap((R, (10) area area((RR) ) 2) and area(R,W)isisthe theoverlap overlap area of substrate wetting pattern and crop root (cm zone (cm2can ) and where area(R,W) area of substrate wetting pattern and crop root zone be can be described by the minimal area of the piecewise curve on the intersection points of described by the minimal area of the piecewise curve based on based the intersection points of Equations 2 ) and can be Equations and (7) (Figure 2), is and theroot areazone of crop (4), (5), and(4), (7)(5), (Figure 2), and area(R) thearea(R) area of is crop (cm2)root and zone can be(cm described as: described as: Z ZZ (11) area( R (R Rdz area ) )= Rdz (11)

0

0

Figure 2. A diagram of the overlap area of substrate wetting pattern and crop root zone. Figure 2. A diagram of the overlap area of substrate wetting pattern and crop root zone.

3. Results and Discussion 3. Results and Discussion 3.1. 3.1. Model Model Calibration Calibration The The core core of of irrigation irrigation scheduling scheduling is is the the accurate accurate precision precision of of the the wetting wetting pattern pattern model, model, root root zone zone model, and EC-5 sensor, which can be affected by substrate type, substrate temperature, crop model, and EC-5 sensor, which can be affected by substrate type, substrate temperature, crop species, species, weather weather conditions, conditions, and and so so on. on. Model Model calibration calibration becomes becomes an an integral integral part part of of the the modeling modeling exercise exercise to obtain the parameter values which will determine the quality of the model predictions. to obtain the parameter values which will determine the quality of the model predictions. All All the the simulated values were calculated experimentally through the formulas (1) to (9) with the estimated simulated values were calculated experimentally through the formulas (1) to (9) with the estimated calibration and 2); 2); the the models models could could correctly correctly forecast forecast the the substrate substrate wetting wetting calibration coefficients coefficients (Tables (Tables 11 and pattern, zone, and and EC-5 EC-5 sensor sensor detection detection value. value. The pattern, lettuce lettuce root root zone, The comparisons comparisons between between the the simulated simulated and the observed values of the substrate wetting pattern, lettuce root zone, and substrate moisture and the observed values of the substrate wetting pattern, lettuce root zone, and substrate moisture are are presented to show the results of our model calibration (because of the large volume of data, Table3 presented to show the results of our model calibration (because of the large volume of data, Table 3 shows shows only only part part of of them). them). The The simulated simulated and and observed observed values values were were very very close, close, and and the the RAE RAE (Relative (Relative Absolute Error) values were less than 12% for all calibration models. It can be concluded Absolute Error) values were less than 12% for all calibration models. It can be concluded that that the the calibrated wetting pattern model, root zone model, and EC-5 sensor calibration model can accurately calibrated wetting pattern model, root zone model, and EC-5 sensor calibration model can accurately simulate simulate the the substrate substrate wetting wetting pattern, pattern, lettuce lettuce root root zone, zone, and and substrate substrate water water content content under under drip drip irrigation condition, respectively. irrigation condition, respectively. Table root depth depth model, model, and and wetting wetting pattern pattern model. model. Table 2. 2. The The coefficients coefficients of of the the EC-5 EC-5 sensor, sensor, root

EC-5 Sensor EC-5 Sensor a b c a b c 1.83 522.24 597.41 1.83

522.24

597.41

Root Depth (Z) Root Depth (Z) bz bz 0.014 0.014

Wetted Radius Wetted Radius A1 n1 A1 n1 0.15 1.01 0.15

1.01

Wetted Depth Wetted Depth A2 n2 A2 n2 2.96 0.28 2.96

0.28

Water 2018, 2018, 10, 10, 549 x FOR PEER REVIEW Water

66ofof109

Table 3. Results of model calibration using experiment data obtained from lettuce cultivated in a Table 3. Results of model calibration using experiment data obtained from lettuce cultivated in a substrate under drip irrigation. substrate under drip irrigation.

Root Radius, R (cm) RAE Sim. Radius, Obs. R (cm) Root (%) 2.06 2.26 RAE 8.85 Sim. Obs. 3.58 3.47 (%) 3.17 4.58 4.38 4.57 2.06 2.26 8.85 5.33 5.61 4.99 3.58 3.47 3.17 5.26 4.38 5.85 4.57 10.09 4.58 6.74 5.61 7.52 4.99 10.37 5.33

Root Depth, Z (cm) RAE Sim. Depth, Obs. Z (cm) Root (%) 5.98 5.95 RAE 0.51 Sim. Obs. 7.74 8.50 (%) 8.94 9.10 9.43 3.50 5.98 5.95 0.51 11.07 12.20 9.26 7.74 8.50 8.94 13.67 9.43 13.15 3.50 3.95 9.10 15.24 12.20 16.12 9.26 5.46 11.07

Wetted Radius (cm) Wetted Depth (cm) RAE RAE Sim. Radius Obs. (cm) Sim. Depth Obs. (cm) Wetted Wetted (%) (%) 2.92 2.81 RAE 3.92 3.33 3.15 RAE 5.71 Sim. Obs. Sim. Obs. 3.85 4.20 (%) 8.33 4.43 4.60 (%) 3.70 4.43 4.61 3.91 5.62 5.23 7.46 2.92 2.81 3.92 3.33 3.15 5.71 5.84 5.24 11.45 5.69 6.33 10.11 3.85 4.20 8.33 4.43 4.60 3.70 6.71 4.61 7.34 3.91 8.58 5.62 6.49 5.23 7.03 7.46 7.68 4.43 7.02 5.24 7.84 11.45 10.46 5.69 7.75 6.33 7.04 10.11 10.09 5.84

Substrate Moisture (%) RAE Sim. Obs. Substrate Moisture(%) (%) 12.35 13.67 9.66 RAE Sim. Obs. 23.71 26.93 11.96 (%) 39.01 35.42 10.14 12.35 13.67 9.66 44.61 26.93 41.14 8.44 23.71 11.96 53.26 35.42 50.12 6.27 39.01 10.14 64.11 41.14 60.90 5.27 44.61 8.44

5.26RAE5.85 10.09 absolute 13.67 13.15 3.95 and 6.71 8.58 the6.49 7.03 and 7.68observed 53.26 values 50.12 of the 6.27 is relative error; Sim. Obs.7.34 represent simulated 15.24 16.12 5.46 7.02 7.84 10.46 7.75 7.04 10.09 64.11 60.90 5.27 6.74 7.52 10.37

relevant variables.

RAE is relative absolute error; Sim. and Obs. represent the simulated and observed values of the relevant variables.

3.2. Sensor Number and Placement 3.2. Sensor Number and Placement The use of sensors for substrate water monitoring should take under consideration factors such The use of of sensors forfor substrate water shouldThe take undernumber consideration factors such as the number stations readings andmonitoring their placement. sensor and placement are as the number of stations and placement. The sensor numberThe andmoisture placement are important factors affectingfor thereadings efficiency of their drip irrigation scheduling systems. sensor important factors affecting the efficiency of drip irrigation scheduling systems. The moisture sensor is highly priced, and their numbers increase the final cost. In addition, their poor placement that is is priced, and their numbers increaseconditions the final cost. addition, their poor placement is nothighly representative of the substrate moisture in theInroot zone can result either in cropthat water not representative of the substrate conditions in placement the root zone can result stress or in over-irrigation. The moisture sensor number and depend noteither only inoncrop thewater root stress or in over-irrigation. The sensor number and placement depend not only on the root distribution, distribution, but also on the water state in the root zone. but also the water state in root the root zone. Theon concept of effective depth and root distance taking into account 80% of total root length The concept of effective root depth and root system distance(CID taking into account 80% Camas, of total root measured by the plant root growth monitoring Bio-Science/CI-600, WA,length USA) measured by the plant root growth monitoring system (CID Bio-Science/CI-600, Camas, WA,length USA) was used to express the main root distribution. It was found that 86.37 to 100% of total root was to express It was thatless 86.37 to 100% totalthe rootplant length was used at depths of 0 the to 9main cm, root and distribution. 80.25 to 100% at a found distance than 6 cmoffrom inwas the at depths of 0 to 9 cm, and 80.25 to 100% at a distance less than 6 cm from the plant in the longitudinal longitudinal and orthogonal profiles to plant row at different times after lettuce planting (Figure and orthogonal profilesfor tosensor plant row at different times after lettuce from planting (Figureof3A,B). The6best 3A,B). The best region placement should be at a distance the plants less than cm region sensor placement bethe at abasis distance from the plants less than 6 cmdistance. and at a maximum and at afor maximum depth ofshould 9 cm, on of the effective root of depth and root However, depth 9 cm, on basis the effective root root distance. However, it is not to it is notofenough to the define theofpositions where to depth install and the sensors, the adequate position forenough installing define the positions where to install sensors, the adequate position for content installing thehighly sensors should the sensors should also consider thethe region in which variations in water can influence also consider the region in which variations in water content can highly influence root water extraction. root water extraction.

(A)

(B)

Figure 3. 3. Percentage distribution at at different different times times after after lettuce lettuce planting. planting. (A) (A) Horizontal Horizontal Figure Percentage of of root root distribution distanced from from plant, plant, cm; cm; (B) (B) Substrate Substrate depth, depth, cm. cm. distanced

The main root distribution is not enough to define the position for installing sensors by itself, The main root distribution is not enough to define the position for installing sensors by itself, because the water state at different depths in the root zone, which reflects the ability of water because the water state at different depths in the root zone, which reflects the ability of water extraction extraction by the root, must be considered. It can be expressed by the variability and cluster analysis, by the root, must be considered. It can be expressed by the variability and cluster analysis, and allows and allows to further optimize the sensor number and placement. The parameters of average value, to further optimize the sensor number and placement. The parameters of average value, standard standard deviation, and coefficient of variation in Table 4 were the average values of three EC-5 deviation, and coefficient of variation in Table 4 were the average values of three EC-5 sensors, 23 h sensors, 23 h after stopping the irrigation. The coefficient of variation decreased with increasing substrate depth, the values at (3,4) cm and (3,8) cm under the substrate surface were very obvious

Water 2018, 10, 549

7 of 10

after stopping the irrigation. The coefficient of variation decreased with increasing substrate depth, Water 2018, 10, FOR PEER REVIEW 7 of 9 the values atx(3,4) cm and (3,8) cm under the substrate surface were very obvious and close, and show that the region in the range of 0–8 cm is the main region for root water extraction. This is consistent and close, and show that the region in distance the rangeand of 0–8 cm is the region for root extraction. with the effective root depth and root suggests thatmain the sensors will be water more accurate at This is consistent with the root depth and depth root distance distances of 3 cm from the effective plants and a maximum of 8 cm.and suggests that the sensors will be more accurate at distances of 3 cm from the plants and a maximum depth of 8 cm. Table 4. Substrate water content analysis at different depth. Table 4. Substrate water content analysis at different depth. Sensor Position

Sensor Position

Parameter

Parameter

Average %% Averagevalue value Standard deviation % % Standard deviation Coefficient of variation %

Coefficient of variation %

(3,4) (3,4)cmcm (3,8) (3,8)cm cm 16.3 19.8 16.3 19.8 2.5 2.8 2.5 2.8 15.3 14.1

15.3

14.1

(3,15) cm (3,15) cm 21.4 21.4 2.2 2.2 10.2 10.2

(6,4) (6,4)cm cm 8.6 8.6 1.1 1.1 12.8

12.8

(6,8) cm (6,8) cm 12.1 12.1 1.5 1.5 12.4 12.4

(6,15) cm cm (6,15) 20.3 20.3 2.3 2.3 11.3 11.3

The cluster the substrate water variables at (3,4) cm cm andand (3,8)(3,8) cm, clusteranalysis analysisininFigure Figure4 4indicated indicatedthat that the substrate water variables at (3,4) and (3,15)(3,15) cm and cm first together, respectively. This means that the substrate water cm, and cm (6,15) and (6,15) cmmerged first merged together, respectively. This means that the substrate variables were similar at (3,4)atcm and (3,15) cm and cm, so ansoEC-5 sensor can water variables were similar (3,4) cm(3,8) andcm, (3,8)and cm,atand at (3,15) cm(6,15) and (6,15) cm, an EC-5 sensor be installed at (3,4) cm or (3,8) cm, and at (3,15) cm or (6,15) cm. When combining the region of can be installed at (3,4) cm or (3,8) cm, and at (3,15) cm or (6,15) cm. When combining the regionthe of main rootroot distribution and and root root water extraction withwith the EC-5 sensor detection area (cylinder with the main distribution water extraction the EC-5 sensor detection area (cylinder awith column diameter of 4 cm), the adequate position for installing an EC-5 sensor waswas found to be at a column diameter of 4 cm), the adequate position for installing an EC-5 sensor found to be (3,4) cm under the substrate surface to check substrate moisture for lettuce cultivation, and at (3,15) cm at (3,4) cm under the substrate surface to check substrate moisture for lettuce cultivation, and at (3,15) or cm to the bottom layer leakage. cm(6,15) or (6,15) cmmonitor to monitor the bottom layer leakage.

Figure 4. Substrate under different different depths. depths. Figure 4. Substrate water water content content cluster cluster diagram diagram under

3.3. Moisture Dynamic in Irrigation 3.3. Moisture Dynamic in Irrigation The irrigation begun when the substrate moisture content detected by the sensors in the crop The irrigation begun when the substrate moisture content detected by the sensors in the crop root root was lower than the threshold value of 0.14 cm3/cm3, and the irrigation stopped when the overlap was lower than the threshold value of 0.14 cm3 /cm3 , and the irrigation stopped when the overlap area of wetting pattern and crop root zone (overlap(R,W) was more than 90%. The substrate moisture area of wetting pattern and crop root zone (overlap(R,W) was more than 90%. The substrate moisture dynamic at the depths of 4 cm and 15 cm under the substrate surface was monitored by theEC-5 dynamic at the depths of 4 cm and 15 cm under the substrate surface was monitored by theEC-5 sensors (Figure 5). It was observed that the water content at the depths of 4 cm and 15 cm changed in sensors (Figure 5). It was observed that the water content at the depths of 4 cm and 15 cm changed a similar manner and was always higher than the threshold value of 0.14 cm3cm−3, but the water in a similar manner and was always higher than the threshold value of 0.14 cm3 cm−3 , but the water content at 15 cm lagged behind compared with that at 4 cm. This shows that the EC-5 sensor’s ability content at 15 cm lagged behind compared with that at 4 cm. This shows that the EC-5 sensor’s ability to to maintain a reasonable water content in the main root zone, the wetting pattern, and the main root maintain a reasonable water content in the main root zone, the wetting pattern, and the main root zone zone could fulfill the real-time requirements, and when irrigation was stopped, the bottom layer could fulfill the real-time requirements, and when irrigation was stopped, the bottom layer leakage leakage was already occurring. Irrigation was performed only three times within more than one was already occurring. Irrigation was performed only three times within more than one month, and month, and the maximum leakage occurred during the second irrigation, mainly because of the large the maximum leakage occurred during the second irrigation, mainly because of the large emitter emitter discharge of the second irrigation and of the substrate low water-holding capacity caused by large particles of the ground vinegar residue. Although leakage occurred, the leakage ratewas15.71% lower than that occurring during conventional irrigation, as a result of precise irrigation achieved with the EC-5 sensors (Table 5). Because of this, the lettuce did not show signs of water stress and was green and healthy. In addition, compared with conventional ways of providing sufficient irrigation, this irrigation scheduling, saves

Water 2018, 10, 549 Water 2018, 10, x FOR PEER REVIEW

8 of 10 8 of 9

discharge of the second the irrigation and of of thethe substrate lowwater, water-holding capacity caused by large irrigation water, increases productivity irrigation and makes good use of limited particles of the ground vinegar residue. water resources.

Figure 5. Substrate water content measured by EC-5 sensors at two depths of 4 cm and 15 cm. Figure 5. Substrate water content measured by EC-5 sensors at two depths of 4 cm and 15 cm. Table 5. Total irrigation, leakage, and leakage rate.

Although leakage occurred, the leakage ratewas15.71% lower than that occurring during Irrigation Method Irrigation Amount Leakage Amount (kg) Leakage Rate (Table (%) 5). conventional irrigation, as a result of precise(kg) irrigation achieved with the EC-5 sensors Conventional Irrigation 10.48 26.03 Because of this, the lettuce40.26 did not show signs of water stress and was green and healthy. In addition, Precise Irrigation 32.45 3.35 irrigation, this irrigation 10.32 scheduling, saves compared with conventional ways of providing sufficient irrigation water, increases the productivity of the irrigation water, and makes good use of limited 4. Conclusions water resources.

In this study, a controller Table with 5. EC-5 sensors and a solenoid valve was used to control irrigation Total irrigation, leakage, and leakage rate. in a greenhouse. The EC-5 sensor compensation model, wetting pattern model, and root zone model were used to calibrate sensor, simulate pattern and the zoneRate of lettuce, Irrigation Methodthe EC-5Irrigation Amountthe (kg)wetting Leakage Amount (kg)main root Leakage (%) 3 3 determine the relationship between them, and calculate a threshold value 0.14 cm /cm , which were Conventional Irrigation 40.26 10.48 26.03 used to establish an irrigation scheduling that was saved within the controller programmatically. In Precise Irrigation 32.45 3.35 10.32 this setting, the controller turns on the irrigation system via communication with the solenoid valve on 4.the irrigation line and the EC-5 sensor when a threshold value is reached and stops when the Conclusions overlap area of wetting pattern and crop root zone is more than 90%. this study, a controller EC-5 sensors and a solenoid valve was usedand to control irrigation TheIn distance from lettuce, with the depths of effective root water extraction, the region of in a greenhouse. The EC-5 sensor compensation model, wetting pattern model, and root zone model substrate wetted volume under drip irrigation should be considered for the definition of the EC-5 were used to calibrate EC-5 sensor, simulate the wetting and the root of lettuce, sensor’s placement. The the adequate position for installing an pattern EC-5 sensor is main at (3,4) cmzone under the 3 /cm3 , which were determine the relationship between them, and calculate a threshold value 0.14 cm substrate surface to check substrate moisture for lettuce cultivation, and at (3,15) cm or (6,15) cm to used to an irrigation monitor theestablish bottom layer leakage. scheduling that was saved within the controller programmatically. In Even this setting, the controller oneach the irrigation via communication the than solenoid if leakage occurred turns during irrigation,system the leakage ratewas15.7%with lower that valve of on the irrigation line and the EC-5 sensor when a threshold value is reached and stops when conventional irrigation, when using irrigation scheduling with the EC-5 sensors, and the lettuce grewthe overlap area of wetting pattern and crop root zone is more than 90%. green and healthy. The distance from lettuce, the depths of effective root water extraction, and the region of substrate Author Contributions: Zhigang designedshould and carried out the simulations wrote theofpaper; Qinchao Xu wetted volume under dripLiu irrigation be considered for theand definition the EC-5 sensor’s analyzed the data and revised the manuscript. placement. The adequate position for installing an EC-5 sensor is at (3,4) cm under the substrate surfaceThis to check substrate moisture for lettuce cultivation, and atFoundation (3,15) cm orof(6,15) cmgrant to monitor Funding: research was funded by [the National Natural Science China] numberthe bottom layer leakage. [51509098]. Even if leakage occurred during each irrigation, the leakage ratewas15.7% lower than that of Conflicts of Interest: The authors declare no conflict of interest. conventional irrigation, when using irrigation scheduling with the EC-5 sensors, and the lettuce grew green and healthy. References 1. 2.

Nemali, K.S.; Montesano, F.; Dove, S.K.; Van Iersel, M.W. Calibration and performance of moisture sensors in soilless substrates: ECH2O and Theta probes. Sci. Hortic. 2007, 112, 227–234. Sánchez-Molina, J.A.; Rodríguez, F.; Guzmán, J.L.; Ramírez-Arias, J.A. Water content virtual sensor for tomatoes in coconut coir substrate for irrigation control design. Agric. Water Manag. 2015, 151, 114–125.

Water 2018, 10, 549

9 of 10

Author Contributions: Zhigang Liu designed and carried out the simulations and wrote the paper; Qinchao Xu analyzed the data and revised the manuscript. Funding: This research was funded by [the National Natural Science Foundation of China] grant number [51509098]. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2.

3. 4. 5. 6. 7. 8.

9. 10. 11. 12. 13. 14.

15. 16. 17. 18. 19. 20. 21.

Nemali, K.S.; Montesano, F.; Dove, S.K.; Van Iersel, M.W. Calibration and performance of moisture sensors in soilless substrates: ECH2 O and Theta probes. Sci. Hortic. 2007, 112, 227–234. [CrossRef] Sánchez-Molina, J.A.; Rodríguez, F.; Guzmán, J.L.; Ramírez-Arias, J.A. Water content virtual sensor for tomatoes in coconut coir substrate for irrigation control design. Agric. Water Manag. 2015, 151, 114–125. [CrossRef] Cardenas-Lailhacar, B.; Dukes, M.D. Precision of soil moisture sensor irrigation controllers under field conditions. Agric. Water Manag. 2010, 97, 666–672. [CrossRef] Dobbs, N.A.; Migliaccio, K.W.; Dukes, M.D. Interactive Irrigation Tool for Simulating Smart Irrigation Technologies in Lawn Turf. J. Irrig. Drain. Eng. 2013, 139, 745–854. [CrossRef] Kenneth, F.; Heaney, J.P.; Miguel, M. Predicting and managing residential potable irrigation using parcel-level databases. J. Am. Water Works Assoc. 2013, 105, 45–46. Blonquist, J.M., Jr.; Jones, S.B.; Robinson, D.A. Precise irrigation scheduling for turfgrass using a subsurface electromagnetic soil moisture sensor. Agric. Water Manag. 2006, 84, 153–165. [CrossRef] Raviv, M.; Lieth, J.H. Soilless Culture Theory and Practice; Elsevier: New York, NY, USA, 2008; pp. 6–7. Naddaf, O.A.; Livieratos, I.; Stamatakis, A.; Tsirogiannis, I.; Gizas, G.; Savvas, D. Hydraulic characteristics of composted pig manure, perlite, and mixtures of them, and their impact on cucumber grown on bags. Sci. Hortic. 2011, 129, 135–141. [CrossRef] Liu, Z.G.; Li, P.P.; Hu, Y.G.; Wang, J. Modeling the wetting patterns in cultivation substrates under drip irrigation. J. Coast. Res. 2015, SI73, 173–176. [CrossRef] Rodriguez, D.; Reca, J.; Martínez, J.; Urrestarazu, M. New Adaptive Hybrid–Automatic Irrigation Control System for Soilless Culture. J. Irrig. Drain. Eng. 2015, 141, 04014083. [CrossRef] Li, J.; Smith, D.W.; Fityus, S.G.; Sheng, D. Numerical analysis of neutron moisture probe measurements. Int. J. Geomech. 2003, 3, 11–20. [CrossRef] Pelletier, M.G.; Karthikeyan, S.; Green, T.R.; Schwartz, R.C.; Wanjura, J.D.; Holt, G.A. Soil Moisture Sensing via Swept Frequency Based Microwave Sensors. Sensors 2012, 12, 753–767. [CrossRef] [PubMed] Park, J.S.; Son, J.E. Analysis of Moisture Characteristics in Rockwool Slabs using Time Domain Reflectometry (TDR) Sensors and Their Applications to Paprika Cultivation. Prot. Hortic. Plant Fact. 2009, 18, 238–243. Seo, T.C.; Kim, Y.C.; Kim, K.Y.; Yun, H.K. Optimal Supply Amount and Strength of Nutrient Solution for Ripe-harvesting Tamatioes Grown under Perlite Culture System of Semi-forcing Cropping. Korean J. Hortic. Sci. Technol. 2003, 21, 79–85. Morel, P.; Michel, J.C. Control of the moisture content of growing media by time domain reflectometry (TDR). Agronomie 2004, 24, 275–279. [CrossRef] Liu, Z.G.; Li, P.P.; Hu, Y.G.; Zhang, X. Calibration method for dectection of matrix water content with dielectric-type moisture sensor. Trans. CSAE 2011, 27, 109–202. Burnett, S.E.; Van Iersel, M.W. Morphology and irrigation efficiency of Gauralindheimeri grown with capacitance sensor-controlled irrigation. Hortscience 2008, 43, 1555–1560. Mendes, L.B.; Li, H.; Xin, H.; Barbosa do Nascimento, J.W. Evaluation of EC-5 soil moisture sensors for real-time determination of poultry manure or litter moisture content. Appl. Eng. Agric. 2014, 30, 277–284. Liu, Z.G.; Wang, J.Z.; Xu, Y.F.; Li, P.P. Root Growth Model of Lettuce under Different Cultivation Substrates and Irrigation Modes. Trans. Chin. Soc. Agric. Mach. 2014, 45, 284–288. Elnesr, M.N.; Alazba, A.A. The effects of three techniques that change the wetting patterns over subsurface drip-irrigated potatoes. Span. J. Agric. Res. 2015, 13, e1204. [CrossRef] Satchithanantham, S.; Krahn, V.; Sager, S. Shallow groundwater uptake and irrigation water distribution within the potato root zone. Agric. Water Manag. 2014, 132, 101–110. [CrossRef]

Water 2018, 10, 549

22. 23.

24.

10 of 10

Liu, Z.G.; Li, P.P.; Hu, Y.G.; Wang, J. Wetting patterns and water distributions in cultivation media under drip irrigation. Comput. Electron. Agric. 2015, 112, 200–208. [CrossRef] Ahmadi, S.H.; Sepaskhah, A.R.; Andersen, M.N.; Plauborg, F.; Jensen, C.R.; Hansen, S. Modeling root length density of field grown potatoes under different irrigation strategies and soil textures using artificial neural networks. Field Crops Res. 2014, 162, 99–107. [CrossRef] Kage, H.; Kochler, M.; StÜtzel, H. Root growth of cauliflower (Brassica oleracea L. botrytis) under unstressed condition: Measurement and modeling. Plant Soil 2000, 223, 131–145. [CrossRef] © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).