Soil Moisture or ET-Based Smart Irrigation Scheduling: A Comparison for Sweet Corn with Sap Flow Measurements

Asiimwe, Gadson; Jaafar, Hadi; Haidar, Mustapha; Mourad, Roya

doi:10.1061/(ASCE)IR.1943-4774.0001668

Open access

Technical Papers

Mar 23, 2022

Soil Moisture or ET-Based Smart Irrigation Scheduling: A Comparison for Sweet Corn with Sap Flow Measurements

This article has a reply.

VIEW THE REPLY

This article has a reply.

VIEW THE REPLY

Authors: Gadson Asiimwe, Hadi Jaafar, A.M.ASCE https://orcid.org/0000-0003-2612-3191 [email protected], Mustapha Haidar, and Roya MouradAuthor Affiliations

Publication: Journal of Irrigation and Drainage Engineering

Volume 148, Issue 6

https://doi.org/10.1061/(ASCE)IR.1943-4774.0001668

PDF

Abstract

Smart irrigation is one application of digital agriculture that can be used to achieve improved crop yields while saving water and energy. Many variations of smart irrigation can be applied, but there is no consensus among the scientific community as to which method is the best practice. The objective of this research was to evaluate two smart irrigation methods: a soil moisture–based method using sensors and an evapotranspiration- (ET-) based technique. The selected crop was sweet corn. Growth variables (biomass, yield, harvest index, and water productivity) were compared for each of the two methods. Three irrigation regimes were applied for each method: 60%, 90%, and 120% of crop evapotranspiration (

{ET}_{c}

) for the ET-based method, and 25%, 30%, and 35% of the total soil moisture for the soil moisture–based method. Corn sap flow was measured in response to the three treatments in the soil moisture–based experiment to measure transpiration. The results showed that the ET-based method is easier to implement with less infrastructure, and it can result in similar yields compared to the soil moisture–based method. Although the fresh yield was 16% higher using the soil moisture–based method, grain yield can be sustained with the ET-based method using 8% less water. Selection of an appropriate irrigation scheduling method should be based on marketable yield: keeping the soil water content near field capacity will result in a higher fresh yield but will not translate into more dry matter. If grain yield is the target, the ET-based method is less expensive and more farmer friendly, but care should be taken to properly estimate irrigation losses to avoid underirrigation. An evaporative stress index (determined using sap flow measurements) of 70% will cause a 22% reduction in fresh sweet corn yield but only a 7% reduction in grain yield. Transpiration peaked when the soil moisture was above 85% of the available water (i.e., at 38% in the calcareous clay soil of the experiment).

Introduction

A large share of arid and semiarid regions’ freshwater resources is used in irrigated agriculture (Postel et al. 1996; Stephenson et al. 2004; Vaishali et al. 2017). The agriculture sector in these regions is facing severe water scarcities and challenges posed by increased water demands coupled with a growing population (Zwart and Bastiaanssen 2004), limited water resources (Dubois 2011; Margat et al. 2005), widely used traditional irrigation techniques (Cirelli et al. 2009; Finley 2016), and climate change–induced precipitation variability (Padakandla 2016). This situation has increased the need to shift toward real-time smart irrigation practices to optimize water use in agriculture (Adeyemi et al. 2018). Smart irrigation is a promising tool for enhancing crop yield and water productivity, improving irrigation scheduling, and decreasing farming costs while sustaining the environment (Dobbs et al. 2014; Paredes et al. 2018; Pereira et al. 2015; Rosa et al. 2012; Seidel et al. 2016; Zotarelli et al. 2011).

Through smart irrigation management, various irrigation scheduling methods have been adopted, such as evapotranspiration and water balance (ET-WB), soil moisture status, and plant water status approaches (Fereres and Soriano 2007; Jones 2004; Scherer et al. 1996). A combination of different irrigation scheduling methods has been practiced and endorsed as a means of improving the irrigation scheduling of crops and preventing water wastage (Deb et al. 2013; Navarro-Hellín et al. 2016; Torres Sánchez et al. 2016).

Several research works have reported the use of either evapotranspiration (ET), soil moisture sensors, or plant-based smart water irrigation technologies in scheduling irrigation events by providing site and crop-specific water requirements while considering weather factors, soil moisture conditions, or plant water status, respectively (Cardenas-Lailhacar and Dukes 2010; Dursun and Ozden 2011; McCready et al. 2009; Migliaccio et al. 2010; Sun et al. 2018). Many studies have been carried out on scheduling irrigation based on soil moisture status, such as in papayas (Migliaccio et al. 2010), tomatoes (Zotarelli et al. 2009), and chile peppers (Sharma et al. 2017), among many others. The advantage of soil moisture–based methods is the ease in practice and automation with some commercially available systems. Major drawbacks to the soil moisture sensor–based scheduling method are the spatial soil moisture heterogeneity, errors in sensor installation, the difficulty in the representation of the entire root zone, the need for sensor calibration, and the inaccuracy of measurements when gravel exists (Evett et al. 2011; Jones 2004). Examples of field studies demonstrating the use of the ET-based irrigation scheduling include those conducted by Jaafar et al. (2017) on biblical hyssop, Ertek and Kara (2013) and Garcia y Garcia et al. (2009) on sweet corn, and Di Paolo and Rinaldi (2008) and Irmak et al. (2016) on maize. However, ET-based irrigation methods strongly rely on the estimation of local climatic data. ET-based irrigation methods also depend on crop coefficients that are local-specific and may be inaccurate. Cumulative errors may occur with calculated ET-based scheduling approaches, and that is why field-based measurements are generally needed to correct or reset ET-based irrigation suggestions (Masmoudi et al. 2011; Pauwels and Samson 2006).

Plant-based irrigation methods can be based on either direct or indirect measurements of plant water status or plant physiological responses to drought. They can be based on measurements such as sap flow and stomatal conductance (Jones 2004; Padilla-Díaz et al. 2016). Sap flow measurements have been used in irrigation management in maize (Jiang et al. 2016), olives and apples (Fernández et al. 2008), and soybeans (Gerdes et al. 1994). The strength of these methods is their sensitivity to moisture deficit. However, plant-based methods require sophisticated instrumentation and expertise (Gu et al. 2020; Jones 2004).

In this paper, field experiments were conducted on sweet corn using two different irrigation scheduling methods: ET-based and soil moisture sensor–based methods, at three thresholds for each experiment. The objectives of this work were to compare ET-based and soil moisture–based irrigation scheduling methods and evaluate their impact on sweet corn morphometric parameters, water use, and water productivity under different treatments.

Materials and Methods

Study Site

The study site was located at the American University of Beirut’s Agricultural Research and Education Centre (AREC) in the Beqaa, Lebanon (33°55’83” N, 36°04’18” E; 990 AMSL) (Fig. 1). The sweet corn (Zea mays L. var. merkur) used in this field experiment is typically planted in the summer, requiring warm soil temperatures (20°C–30°C) for favorable growth. Sweet corn is generally cultivated over an extended period in order to provide a continuous supply of fresh corn. Varieties of sweet corn take 70 to 100 days to mature from the day after planting. A proper irrigation management would achieve maximum yield (beyond

20,000 kg / ha

) (Garcia y Garcia et al. 2009) in addition to enhanced water use efficiency with minimum water losses; conversely, poor irrigation practices with insufficient water provided to sweet corn would lead to low yield and, hence, economic loss (Archana et al. 2016; Dukes and Scholberg 2005; Mubarak 2020).

Fig. 1. Location of the study area in Beqaa, Lebanon, at AREC. (Sources: Esri, Airbus, USGS, NGA, NASA, CGIAR, N Robinson, NCEAS, NLS, OS, NMA, Geodatastyrelser, Rijkswaterstaat, GSA, Geoland, FEMA, Intermap and the GIS user community, Sources: Esri, HERE, Garmin, FAO, NOAA, USGS, © OpenStreetMap contributors, and the GIS user community.)

The field plots on which the experiment was carried out were flat and had no clear slope. A semiarid climate characterizes the experimental site, with dry hot summers from May to September and cold winters throughout the rest of the year. The average rainfall is around 500 mm per year. Daily crop evapotranspiration (

{ET}_{c}

) was calculated as a factor of grass-reference crop evapotranspiration (

{ET}_{ref}

) and crop coefficient (

K_{c}

) using the Ref-ET version 3.1 software (Allen 2009; Annandale et al. 2002) based on the ASCE standardized Penman-Monteith equation from weather data obtained from the AREC weather station (Campbell Scientific, Logan, Utah) (Fig. 2) within 200 m of the field. The soil of the experimental site was shallow, gravelly clay, with 19% sand, 36% silt, and 45% clay. The soil had a pH of 7.89, an electrical conductivity of

0.4 dS / m

,

{CaCO}_{3}

of 32.5%, and an organic matter content of 2.48% (Jaafar et al. 2017). The available nutrient supply of phosphorus, potassium, and nitrogen was 19.9, 530, and 30 ppm, respectively. Soil water characteristics were estimated using the model described by Saxton and Rawls (2006) (permanent wilting point was at approximately 22%, field capacity at 40%, and saturation at 49.7%). Bulk density was determined in this experiment by using a cylinder method and was found to be

1.33 g / {cm}^{3}

.

Fig. 2. Temperature (°C), and ${ET}_{ref}$ during the crop growing period at the local meteorological station, AREC, from June to September 2019.

Cultural Practices for the Sweet Corn Experiments

Sowing was carried out on June 6, 2018, with hybrid sweet corn seeds (Zea mays L. var. merkur) having a germination ratio of 90. Seeds were planted every 20 cm down the rows and with 75 cm between the rows. Each plot had a size of

13.5 m^{2}

with four rows. Seeds were sown at a depth of 5–6 cm. Nitrogen, phosphorus, and potassium (15:15:15 NPK ration) fertilizer was applied at a rate of

250 kg / ha

. To control weeds, periodic hand weeding and 2,4-D herbicide was applied 40 days after planting (DAP) at a concentration of 50 cm³ in 12 L of water (4.58 cc/L of water) for the whole field. All experimental plots were irrigated with the same amount of water until emergence to ensure uniform plant establishment. After the emergence of the sweet corn seedlings, irrigation was performed according to the prescribed irrigation treatments.

Experimental Design

The two experiments consisted of three replicates for each of the three treatments in the soil moisture sensor–based and evapotranspiration-based methods, making a total of 18 experimental plots of 6 m by 2.25 m and 1.5 m apart. Water was sourced from a well penetrating an underlying marl limestone aquifer with total dissolved solids of

320 mg / L

. A tapping was taken from one of the risers to feed the secondary network through a 32-mm pipe connecting to the six solenoid valves, each corresponding to a different irrigation treatment. The ET treatments (60%, 90%, and 120%) were operated using an automated control system. The total water flow for each treatment was measured with a flowmeter installed downstream of each solenoid valve. Irrigation was scheduled at an average interval of 2 days during initial crop growth stages and 3 days afterward. Drip emitters had an average discharge rate of

3.7 L / h

at a pressure of 100 kPa (1 bar) as an average of 20 samples. A layout of the experimental plots is shown in Fig. 3.

Fig. 3. Irrigation system layout and experimental design for the ET-based and soil moisture–based experiments.

In the ET-based irrigation system,

{ET}_{ref}

is multiplied by the crop coefficients (

K_{c}

) of sweet corn to get crop evapotranspiration (

{ET}_{c}

). In the Mediterranean region,

K_{c}

is 0.3, 1.15, and 1.05 for

K_{c}

initial,

K_{c}

mid, and

K_{c}

end, respectively, with the initial period taking 20 days, the middle period taking 50 days, and the end period taking 20 days (Allen et al. 1998c). Irrigation run times varied according to

{ET}_{c}

, which is a factor of crop growth stage, and prevailing meteorological phenomena.

The soil moisture sensor–based irrigation had three irrigation treatments (25%, 30%, and 35%) representing the thresholds of soil volumetric water content at which irrigation was scheduled. Values of 0.3, 0.8, and

0.13 m / m

of available water, respectively, out of the full potential available water of

0.15 cm / cm

were maintained in each of those treatments. The valves were operated to start an irrigation event based on the thresholds in each treatment and were closed when the soil volumetric water content (VWC) exceeded field capacity at the sensor depth.

Field Measurements

Sap Flow Measurements

Sap flow was measured using a sap flow meter (SFM1) (Burgess and Downey 2014) with readings taken every 10 min from August 28 to September 4, 2018. The sap flow meter consisted of a set of three measurement probes and an integrated standalone data logger. The three probes were designed with two thermistors located 7.5 and 22.5 mm from the tip. The heater probe had a high resistance filament that produced a high and efficient amount of heat. The SFM1 sap flow meter measured the sap velocity and flow using the heat-ratio method (HRM). This method calculates the magnitude and direction of water flux by measuring the ratio of heat transported between two symmetrically spaced temperature sensors. The heat pulse velocity was calculated as per Eq. (1) (Barrett et al. 1995):

V h = \frac{k}{x} In (\frac{v 1}{v 2}) 3,600

(1)

where

V h

= heat pulse velocity (

cm / h

);

k

= thermal diffusivity of fresh plant

tissue = 2.5 \times 10^{- 3} {cm}^{2} / s

;

x

= distance (cm) between the heater and either temperature probe = 0.6 cm; and

v 1

and

v 2

= increase in temperature (from initial temperatures) at equidistant points downstream and upstream, respectively,

x

cm from the heater. Sap flow meters were installed on stems of random sweet corn in the three soil moisture irrigation regimes. The installation of an SFM1 was conducted via the following steps. First, measurements of the corn stem circumference (ranging from 6 to 6.5 cm) and diameter (average of 2 cm) were conducted. Based on these measurements, three parallel probes were inserted.

Measurements from the sap flow meter include corrected sap velocity, flow, and needle temperature (°C). Because the data generated from the sap flow meters were in hourly intervals, we used the Sap Flow Tool software version 1.4.1 to convert the corrected sap flow to daily accumulated sap volume (

{cm}^{3}

) and sap flow rates (

{cm}^{3} / h

) by inputting the correction factors for the stem and sensor properties (Steppe et al. 2009) as in Table 1.

Table 1. Correction factors and the values made in the sap flow tool settings

Stem and sensor properties	Corrections
Stem circumference	6.9 cm
Stem diameter	2 cm
Bark thickness	0.004 cm
Xylem radius	1 cm
Sapwood depth	0.45 cm
Thermal diffusivity	$0.0025 {cm}^{2} / s$
Sapwood fresh weight	1 g
Sapwood dry weight	0.63 g
Probe spacing	0.5 cm
First thermistor depth	1.5 cm
Wound diameter	17 m

Soil Moisture Measurements and Sensor Calibration

CS655 soil moisture sensors (Campbell Scientific) were calibrated and installed in each of the experimental plots, ensuring good sensor–soil contact. Soil moisture data (1-min increments) were collected using a CR850 Campbell Scientific data logger. The soil volumetric water content was computed based on the Topp equation [Eq. (2)], representing only

C_{0}

and

C_{1}

(Campbell Scientific 2017):

Q_{v} (K_{a}) = C_{0} + C_{1} K_{a} \times C_{2} K_{a}^{2} + \dots + C_{n} K_{a}^{n}

(2)

where

Q_{v}

= volumetric water content (% or

m^{3} / m^{3}

);

K_{a}

= bulk dielectric permittivity (unit-less) of the soil; and

C_{n}

= calibration coefficients.

Fig. 4 shows the equation that was derived from calibration. Soil water at a depth of 30 cm was measured throughout the growing season in every treatment using 16 soil moisture sensors. One soil moisture sensor was installed in two of the three replicates of the ET-based treatments, and two sensors were installed in irrigation treatment ET60%. Average sensor data from the subplots representing the same treatment were assumed to be representative of the soil moisture content in the entire treatment. For soil moisture irrigation regimes, the soil moisture readings from the three replicates were averaged to schedule the irrigation events (Jones 2004) considering the variability of soil moisture in the different plots of the same treatment (Dabach et al. 2015). The sensors were installed at a depth of 30 cm to represent the active root zone of corn, which has been observed to reach a depth of 30–40 cm (Wiesler and Horst 1994).

Fig. 4. Calibration graph of calculated volumetric water content (fraction) against dielectric permittivity.

Measured Parameters and Statistical Analysis

The measured parameters from the six regimes’ harvests were as follows: shoot height and ear length, aboveground fresh and dry biomass, cob fresh and dry weight, root biomass, and grain yield. Shoot height and cob length were directly measured on the day of harvest, 90 DAP for 10 corn plants from the middle two rows within each treatment. The fresh aboveground weight was immediately weighed, oven-dried at 65°C to a constant weight and weighed. The fresh cob weight was measured right after harvest and oven-dried under the same conditions of the aboveground biomass. Root biomass was measured by uprooting the sweet corn using a shovel, and washing the roots of soil under running water before taking their weights. Grain yield of the sweet corn was measured after drying the cobs in the oven. Subsequently, the grains were removed from the cob and weighed. The calculated variables were as follows: harvest index (HI), grains-to-cob ratio, and water productivity (WP). Grains-to-cob ratio represents the weight of grains divided by the total weight of the cob. HI was calculated as the ratio of fresh weight of cobs to fresh aboveground biomass. WP is a measure of the biophysical gain in terms of consumed water. WP is calculated as the ratio of fresh and dry biomass of cob yield to cumulative

{ET}_{c}

or total water use (WU) (Fereres and Soriano 2006). The impact of the studied irrigation regimes was analyzed using ANOVA implemented with the SPSS statistical package version 24. Significant differences between group means were calculated using least significant difference (LSD) at significance level of

p < 0.05

(Steel and Torrie 1980).

Results and Discussion

Soil Moisture Analysis in ET-Based Scheduling

Soil moisture monitoring in the ET-based scheduling experiment was crucial for interpreting the yield and biomass response to the various treatments. The cumulative irrigation water applied in the three ET treatments as measured by flow meters was 566, 770, and 971 mm for ET60%, ET90%, and ET120%, respectively. In the ET60% irrigation treatment, the applied water did not recurrently reach field capacity (FC), and soil moisture was depleted to levels close to the permanent wilting point (PWP) before the next irrigation event. Life-saving irrigations were carried out at different times to rejuvenate the corn under these deficit conditions. Although wilting point was reached in this treatment, most plants in the three replicates of the ET60% treatment were able to recover. In irrigation treatment ET90%, irrigation was carried out at slightly above the permanent wilting point. Most of the irrigations took place at around 27% to 30% VWC. No water losses through runoff or deep percolation were observed. Soil moisture measurements in irrigation treatment ET120% indicated that upon an irrigation event, the VWC often exceeded field capacity, which led to irrigation water percolating below 30 cm. Irrigation was mostly carried out at an average of 32% VWC until August 18, 2018, when it was 35% of VWC. This represents approximately a manageable allowable depletion (MAD) of 23%.

Soil Moisture Analysis of Soil Moisture–Based Scheduling

In the three soil moisture (SM) treatments, the cumulative irrigation water applied was as follows: 717, 969, and 1,050 mm for irrigation treatments SM25%, SM30%, and SM35%, respectively. The irrigation intervals in irrigation treatment SM25% were the longest in comparison with the others, at an average of 3 days. Soil moisture measurements in irrigation treatment SM25% showed that irrigation was very close to a wilting threshold point of 22.2%, equivalent to 18.4% of MAD. Although the average VWC was 31%, water stress on the sweet corn in this treatment still occurred, because the soil water content dropped below the recommended 50% of the manageable allowed depletion (Allen et al. 1998b). In irrigation treatment SM30%, soil moisture stayed within allowable limits, with an average soil moisture of 34%. The irrigation interval in this treatment averaged 2 days before the next irrigation event. Irrigation at 30% VWC represented 50% MAD before the next irrigation event was scheduled. Soil moisture in treatment SM35% had the shortest irrigation interval of 1 day on average. This treatment also recorded the highest in terms of average moisture percentage at 38%, which was slightly above the field capacity of 38%. Irrigation at 35% VWC meant that only 23% of available soil moisture had been depleted (MAD), which is considerably higher than the maximum recommended 50% MAD (Fig. 5).

Comparison between Irrigation Regimes

Soil Moisture

Here we compare the average soil moisture data taken every 10 days in the irrigation treatments (Table 2) for 40 days. The amount of water applied during this period was linearly related to soil moisture content, as shown in Fig. 6. Overall, average soil moisture content was highest (38%) in irrigation treatments ET120% and SM35%. Treatments ET90% and SM25% also had a similar average soil moisture content of 31%. Soil VWC was lowest in the ET60% treatment at 25%, significantly different from all the other treatments.

Table 2. Decadal average soil moisture in each of the irrigation treatments

Irrigation scheduling type	Irrigation treatments	Applied water (mm)	Average decadal soil moisture (fraction)
Irrigation scheduling type	Irrigation treatments	Applied water (mm)	59 DAP	69 DAP	79 DAP	89 DAP	Overall
ET (%ET)	60	322	0.24^a	0.24^a	0.27^a	0.26^a	0.25^a
	90	460	0.30^bc	0.31^b	0.33^bc	0.31^b	0.31^b
	120	614	0.36^d	0.37^bc	0.40^d	0.39^c	0.38^c
Soil moisture (%SM)	25	462	0.30^c	0.32^b	0.30^b	0.32^b	0.31^b
	30	522	0.33^bc	0.35^bc	0.35^c	0.36^c	0.34^c
	35	568	0.36^d	0.39^c	0.39^d	0.39^c	0.38^d

Note: Statistical comparison between the two different scheduling types. Within the same columns, means with the same letter are insignificantly different at $p < 0.05$ .

Fig. 6. Soil volumetric water content against applied irrigation water (between 60 and 90 days after planting). The Treatments SM25% and ET90% had a similar average soil water content and applied water, while treatment ET120% had 8% more applied water than the SM35% treatment but a similar soil moisture content, indicating some deep percolation losses below the sensor depth.

Applied Water in the Various Treatments

The depths of applied water (mm) in the soil moisture and ET-based treatments were compared across the growing season at intervals of 10 days (the first row in Table 3 shows 30 mm of irrigation that was applied right after planting to bring the soil to field capacity). Treatment SM35% had the highest amount of water applied, followed by SM30% and ET120% (which had lower applied water until midseason). Applied water in ET90% and SM25% was nearly the same, and ET60% had the lowest amount (Table 3). Fig. 7 shows the cumulative readings from the flow meters on the different irrigation treatments throughout the growing season. In the ET-based treatments, the amount of irrigation water applied at the start of the growing season exponentially increased with time as

{ET}_{ref}

increased (Allen et al. 1998a). In the soil moisture sensor–based irrigation treatments, there was a general gradual increase across the growing season.

Table 3. Comparison between the irrigation regimes in terms of applied water (mm) across the growing season (days)

Days after planting	ET-based (%ET)			SM-based (%SM)
Days after planting	60	90	120	25	30	35
0	30.0	30.0	30.0	30.0	30.0	30.0
10	52.2	44.2	41.5	12.7	111.2	114.1
20	41.1	37.1	35.8	21.4	70.6	72.1
30	10.6	31.4	61.1	96.8	76.0	102.9
40	43.8	49.0	46.1	38.1	56.1	53.0
50	66.7	118.1	142.8	57.0	103.7	110.9
60	63	108.3	126.4	127.7	83.1	133.7
70	70.0	97.7	133.4	131.6	178.4	141.1
80	97.7	136.8	185.8	87.0	92.2	116.2
91	91.1	117.5	168.3	115.1	168.2	176.9
Total	566.4	770.2	971.1	717.5	969.5	1,051.0

Fig. 7. Flowmeter readings from the SM sensor–based and ET-based irrigation experiments during the various growth stages: (A) emergence; (B) 0 to 8 leaves; (C) 8 to 16 leaves; (D) tasseling and silking; and (E) maturity.

Irrigation Regime Impact on Crop Growth

Impact on Shoot Height

The impact of the applied water and the average soil moisture within the season was noticeable in the variations of the crop growth parameters. Mean shoot height (SH) was significantly affected by the ET treatments, increasing concurrently with higher water applications. The mean SH in ET120% was 21% and 3% higher than it was in the ET60% and ET90% treatments, respectively (Table 4).

Table 4. Effects of irrigation on the morphometric characteristics (shoot height and cob length) of field-grown sweet corn at harvest in different irrigation treatments

Statistical comparison	Irrigation scheduling type	Irrigation treatment	Shoot height (cm)	Cob length (cm)
Within each irrigation scheduling type	ET-based (% ET)	60	151.67^a	17.20^a
		90	195.20^b	20.97^b
		120	203.17^c	21.37^b
	Soil moisture–based (% SM)	25	185.83^a	19.867^a
		30	208.60^b	21.783^b
		35	213.00^b	21.417^b
Between the two different irrigation scheduling types	ET-based (% ET)	60	151.67^a	17.20^a
		90	195.20^b	20.97^b
		120	203.17^c	21.37^bc
	Soil moisture–based (% SM)	25	185.83^d	19.87^d
		30	208.60^ce	21.78^bce
		35	213.00^e	21.42^bce

Note: A statistical comparison within the same and between the irrigation scheduling types is shown. Values within the same columns with different letters are significantly different at $p < 0.05$ ( $n = 30$ from three replicates) based on LSD.

For the SM-based treatments, sweet corn in the SM25% treatment had the shortest SH, significantly lower than the other treatments. There were no significant differences between SH in the SM30% and SM35% treatments, although the latter was higher. Among both irrigation regimes, SM35% recorded the highest shoot height, with a SH 29%, 8%, 5%, 12%, and 2% greater than SH in ET60%, ET90%, ET120%, SM25%, and SM35%, respectively (Table 4).

Cob Length

The analysis showed that cob length increased with increasing %ET and %SM. In the ET-based irrigation treatments, the highest and lowest recorded length means were observed in ET120% and ET 60%, respectively. The average cob lengths showed no significant differences between ET120% and ET90%. However, cob lengths in the ET60% treatment significantly differed from the other treatments (

F = 15.57

,

p < 0.05

). As for the SM-based irrigation treatments, cob length in SM35% was significantly higher (9%) than in SM25%. Comparison between both irrigation scheduling regimes showed no impact on cob length except that cob length in the ET60% treatment was significantly lower than in all the other irrigation treatments (Table 4).

Biomass and Yield Response

Aboveground Biomass

The analysis showed that both dry and fresh aboveground biomass increased significantly with increasing irrigation applications in the ET scheduling regimes. The highest recorded aboveground fresh and dry biomasses were in ET120%; they were 46% and 39% higher than in ET60%, respectively, and 18% and 15% higher than in ET90%, respectively. These results were consistent with previous studies showing similar effects of irrigation applications on aboveground biomass (Ertek and Kara 2013; Kresović et al. 2016). The effects of the soil moisture–based irrigation strategies on the aboveground fresh and dry biomass are shown in Table 5. SM35% produced the highest fresh and dry aboveground weight, 21% and 14% higher than SM25%, respectively, and 11% and 17% higher than SM30%, respectively. Among the ET and SM irrigation regimes, SM35% accumulated the highest aboveground fresh and dry biomass, even higher than the highest ET irrigation level (ET120%) by 13% and 12%, respectively (Table 5).

Table 5. Effects of irrigation on the morphometric characteristics of field-grown sweet corn at harvest

Statistical comparison	Irrigation scheduling type	Irrigation treatment	AG biomass ( $t / ha$ )		Cob weight ( $t / ha$ )		Root biomass ( $t / ha$ )	Grain yield ( $t / ha$ )
Statistical comparison	Irrigation scheduling type	Irrigation treatment	Fresh	Dry	Fresh	Dry	Root biomass ( $t / ha$ )	Grain yield ( $t / ha$ )
Within each irrigation scheduling type	ET-based (% ET)	60	26.54^a	8.66^a	11.92^a	5.53^a	7.17^a	2.54^a
		90	39.87^b	12.03^b	21.81^b	6.53^b	13.00^b	4.95^b
		120	48.75^c	14.17^c	26.42^c	7.41^c	15.39^b	6.41^c
	Soil moisture–based (% SM)	25	44.33^a	13.75^a	25.43^a	7.58^a	10.50^a	6.08^a
		30	50.41^a	13.44^a	23.42^a	7.62^a	11.27^a	6.20^a
		35	56.29^b	16.14^b	33.70^b	9.12^b	13.89^a	6.47^a
Between the two different irrigation scheduling types	ET-based (%ET)	60	26.54^a	8.66^a	11.92^a	5.53^a	7.17^ac	2.54^a
		90	39.87^b	12.03^b	21.81^b	6.53^b	13.00^bc	4.95^b
		120	48.75^c	14.17^c	26.42^c	7.41^ce	15.39^b	6.41^c
	Soil moisture–based (%SM)	25	44.33^bc	13.75^c	25.43^d	7.58^cd	10.50^bc	6.08^c
		30	50.41^cd	13.44b^c	23.42^d	7.62^cd	11.27^abc	6.20^c
		35	56.29^d	16.14^d	33.70^e	9.12^ce	13.89^bc	6.47^c

Note: A statistical comparison within each and between the irrigation scheduling types is shown. Values within the same columns with different letters are significantly different at $p < 0.05$ ( $n = 120$ except for root biomass, where $n = 12$ ). AG = above-ground.

Root Biomass

The root biomass increased with increased irrigation levels in the ET-based regimes, with the highest value recorded in ET120%. However, the root biomass means in ET120% did not significantly differ from those in ET90%; however, the values for ET60% significantly differed from both ET90% and ET60%.

Moreover, the various SM treatments did not significantly affect root biomass, despite a 24% difference between the highest (SM35%) and lowest (SM25%) root biomass means. We can attribute this to the low number of samples in the root biomass measurements. Comparison between the highest irrigation levels also showed no significance difference between the root biomass means in SM35% and ET120% and all other treatments except for ET60%, which had means significantly lower than those of SM35% (Table 5).

Cob Weight, Grain Yield, and Grain to Cob Ratio

An increasing trend in cob fresh weight (FW) and dry weight (DW) was observed as the level of ET increased, with the highest noted in ET120%, 55% and 17.5% higher than the cob weight means in ET60% and ET90%, respectively. These outcomes were in agreement with those of other studies (Ertek and Kara 2013; Kresović et al. 2018; Vial et al. 2015).

In the SM-based treatments, the mean fresh and dry weights of the sweet corn cobs were highest in SM35%, and they were significantly different from the values for SM25% and SM30%. Mean fresh and dry weights of cobs in SM25% and SM30% revealed no significant differences. The low cob weight in SM25% was due to the deficit form of irrigation being carried out below the recommended 50% management allowable depletion. The irrigation treatment SM35% at a 50% MAD yielded high cob dry and fresh weights. Grain yield was significantly impacted by the ET irrigation levels (

p < 0.001

,

F = 82.04

), where ET120% produced the highest grain yield. In addition, the grain per cob ratios increased with increasing %ET, with values of 0.72, 0.76, and 0.87 for ET60%, ET90%, and ET120%, respectively (Table 5, Fig. 8).

Fig. 8. Range and median of collected data on aboveground fresh and dry weight ( $t / ha$ ), root biomass ( $t / ha$ ), grain yield ( $t / ha$ ), and grain to cob ratio of sweet corn in the ET- and SM sensor–based irrigation treatments.

The SM irrigation levels did not significantly affect the grain yield in the SM-based irrigation treatments; the highest grain yield mean was realized in irrigation treatment SM35% but was not more than 6% greater than the yields in SM25% and SM30%. The greater amount of applied water and average soil moisture translated into more fresh and dry cob weight, with SM35% having the dominant response with the highest fresh and dry cob weight and crop yield (Fig. 8). These findings were similar to other related works showing an increase in cob weight with the amount of applied water (Kiziloglu et al. 2009; Liu et al. 2017; Medrano et al. 2015).

To further analyze cob and grain weight response to irrigation levels, the relative yield to cob weight was calculated for each treatment as a ratio of the respective irrigation treatments and the yield per cob of the treatment with the highest applied water (SM35%) (Fig. 9). ET60% had the least relative cob fresh and dry weight and grain yield, while ET120% was the second highest-ranking treatment in terms cob fresh and dry weight and grain yield, with no significant difference between the relative grain yields in ET120% and SM35%. The relative yield stabilized with increasing amount of water and soil moisture, as shown in Fig. 9.

Fig. 9. Relative cob fresh and dry weight and grain yield in relation to applied water (mm) and average soil moisture (%). Treatments ET120% and SM35% had similar soil moisture and dry grain yield, but the fresh cob weight was lower in the ET120% treatment.

Harvest Index Response

The HI was significantly affected by the irrigation treatments, increasing simultaneously with increased %ET and %SM. In the ET-based irrigation regimes, ET60% yielded the lowest mean HI (0.449), significantly different from the others. Although HI was highest in irrigation treatment ET120%, it was not significantly different from ET90% (0.551). Regarding the SM-based irrigation regimes, there was a significant difference in mean HI between sweet corn in irrigation treatments SM35% and SM25%, which represented the highest and lowest values at 0.6 and 0.47, respectively (Fig. 10). Among both irrigation scheduling regimes, SM35% had the highest HI.

Fig. 10. Range, mean, and standard deviation values of the collected data for HI of sweet corn in the different irrigation treatments.

Water Productivity Response

The current analysis showed that both cob fresh and dry WP decreased with increased irrigation applications, as reported in related studies showing the increase in sweet corn WP as the amount of irrigation decreased (Cid et al. 2018; Di Paolo and Rinaldi 2008; Viswanatha et al. 2002). This was true in all treatments except for the ET60% treatment, which had a low WP due to a low yield (Table 6). Increased WP for both fresh and dry weight in SM25% indicates that the crop was utilizing the water more efficiently although it had a lower yield than the crops under ET120%, SM30%, and SM35%.

Table 6. Mean comparisons for water use and water productivity of field-grown sweet corn at harvest

Statistical comparison	Irrigation scheduling type	Irrigation treatment	Applied water (mm)	Cob WP ( $kg / m^{3}$ )		Grain WP ( $kg / m^{3}$ )	AGB ( $kg / m^{3}$ )
Statistical comparison	Irrigation scheduling type	Irrigation treatment	Applied water (mm)	Fresh	Dry	Grain WP ( $kg / m^{3}$ )	Fresh	Dry
Within each irrigation scheduling type	ET-based (%ET)	60	566	2.80^a	0.83^a	0.6^a	6.24^a	2.03^a
		90	770	3.77^b	1.14^b	0.86^b	6.9^a	2.08^a
		120	971	3.62^b	1.02^b	0.88^b	6.7^a	1.94^a
	Soil moisture–based (%SM)	25	717	4.73^a	2.56^a	1.41^a	8.23^a	2.55^a
		30	969	4.28^a	2.05^b	1.15^b	7.14^a	2.05^b
		35	1,050	3.22^b	1.85^b	1.04^b	6.93^a	1.85^b
Between the two different irrigation scheduling types	ET-based (%ET)	60	566	2.80^a	0.83^a	0.6^a	6.24^a	2.03^a
		90	770	3.77^bc	1.14^a	0.86^b	6.9^ab	2.08^a
		120	971	3.62^bc	1.02^a	0.88^b	6.7^a	1.94^a
	Soil moisture–based (%SM)	25	717	4.73^d	2.56^b	1.41^c	8.23^b	2.55^b
		30	969	4.28^cd	2.05^c	1.15^d	7.14^ab	2.05^a
		35	1,050	3.22^ab	1.85^c	1.04^bd	6.93^ab	1.85^a

Note: Statistical comparison within the same and between irrigation scheduling types. Values within the same columns with different letters are significantly different at $p < 0.05$ from three replicates ( $n = 30$ ) based on LSD. WP = water productivity ( $kg / m^{3}$ ); and AGB = aboveground biomass.

Sap Flow

Sap flow has been estimated to be proportional to transpiration (Fernández et al. 2006), because almost 99% of daily water uptake (sap flow) is lost through transpiration (Bethenod et al. 2000). In addition, sap flow transpiration previously investigated was within 88%–95% of actual evapotranspiration (Bethenod et al. 2000; Miner et al. 2017; Uddin et al. 2014). Sap flow measurements were taken at full cover for 7 days (from 84 to 90 DAP) from the SM25%, SM30%, and SM35% crops (Fig. 11). During the period of measurement, these treatments had applied water of 7.18, 8.78, and 10.90 mm, respectively. The diurnal changes of sap flow velocity showed that the highest observed velocity was recorded during the daytime between 11:00 and 2:00 local time when solar radiation was maximum. As light decreased, sap flow decreased to near-zero sap flow at night, correlating to no photosynthesis (Gerdes et al. 1994; Jiang et al. 2016; Uddin et al. 2014). The highest peaks of sap flow were observed in irrigation treatment SM35%, averaging

110 {cm}^{3} / h

, followed by SM30% at

85 {cm}^{3} / h

; the lowest peaks were in SM25% at

65 {cm}^{3} / h

. This indicates that corn responds very well to soil water stress by a reduction in transpiration.

Fig. 11. Graphical representation of sap flow ( ${cm}^{3} / h$ ) against time in irrigation treatments SM25%, SM30%, and SM35%.

The generated hourly sap flow was converted to transpiration (T) (mm/day) by dividing the daily flow rate (

q

) in

{cm}^{3} / day

by the crop spacing (

0.75 \times 0.2 m

) assuming 100% canopy coverage. Transpiration (

mm / day

) was compared to

{ET}_{ref}

(

mm / day

) as displayed in Table 7. Analysis of transpiration rates showed that corn in irrigation treatment SM35% had 21% higher transpiration. In general, average transpiration rates in the SM35% treatment were 3% higher than the average daily

{ET}_{ref}

for the period of measurement, indicating that the corn was not enduring any water stress in this treatment. Transpiration rate of corn plants in Treatment SM35% was 3% higher than ETref. Transpiration rates of corn in SM30% and SM25% where 9% and 18% lower than ETref, respectively (Table 7). The ratio of transpiration to reference evapotranspiration (evaporative stress indicator) decreased by 26% in the SM25% treatment.

Table 7. Daily transpiration rates (T) of corn plants in the soil-moisture based treatments and ratio of transpiration to reference Evapotranspiration (T/

{ET}_{re f}

)

DAP	T SM25% ( $mm / day$ )	T SM30% ( $mm / day$ )	T SM35% ( $mm / day$ )	% $T / {ET}_{ref}$ SM25%	% $T / {ET}_{ref}$ SM30%	% $T / {ET}_{ref}$ in SM35%	${ET}_{ref}$ ( $mm / day$ )
84	4.24	4.11	6.45	74%	72%	112%	5.74
85	4.75	4.97	6.09	83%	87%	106%	5.73
86	4.79	5.03	6.11	82%	86%	105%	5.82
87	5.07	6.13	6.43	86%	104%	109%	5.9
88	5.29	6.81	6.02	82%	106%	94%	6.43
89	4.58	5.26	6.04	74%	84%	97%	6.23
90	5.13	5.8	5.8	88%	99%	99%	5.86
Total	33.85	38.11	42.94	81%	91%	103%	41.71

Conclusion

In this research, we compared and evaluated sweet corn yield, biomass, water productivity, and other morphometric characteristics based on irrigation scheduling using the irrigation amounts estimated from ET (60%, 90%, and 120% of

{ET}_{c}

) and SM irrigation regimes (25%, 30%, and 35% of soil moisture) on sweet corn. Water stress in both scheduling methods had a significant effect on plant height, aboveground biomass, and yield. With increased application of irrigation water, cob yield increased linearly at a rate of

32 kg / ha

per mm of added irrigation water. However, deficit irrigation in both the ET-based and soil moisture sensor-based irrigation treatments showed an increase in water productivity, except severe water stress in irrigation treatment ET60% showed lower water productivity and yield.

We found that average soil moisture in treatments SM35% and ET120% was identical (38%) for the period of the experiment, and that irrigations can be reduced by 8% while sustaining grain yield. The highest irrigation level resulted in 27% higher fresh cob weight. Sap flow measurements indicated that transpiration peaked when the soil moisture was above 85% of the available water (i.e., at 38% in the calcareous clay soil of the experiment). The highest soil moisture level resulted in a larger fresh cob weight and grain dry weight than was obtained for the other soil moisture treatments. Irrigation water can be saved by using smart irrigation with appropriate soil moisture sensors or ET-based schedulers. ET-based systems can be less costly for large fields in which soils tend to be less homogenous. Installation of soil moisture sensors within the root zone and operational logistics could be challenging, especially in large fields. Overall, both methods are comparable in scheduling irrigations.

Data Availability Statement

All data, models, and code generated or used during the study appear in the published article.

Acknowledgments

This work was made possible via a grant from the MasterCard Foundation Scholars Program at American University of Beirut (AUB) (Award No. 103148).

References

Adeyemi, O., I. Grove, S. Peets, Y. Domun, and T. Norton. 2018. “Dynamic neural network modelling of soil moisture content for predictive irrigation scheduling.” Sensors 18 (10): 3408. https://doi.org/10.3390/s18103408.

Abstract

Introduction

Materials and Methods

Study Site

Cultural Practices for the Sweet Corn Experiments

Experimental Design

Field Measurements

Sap Flow Measurements

Soil Moisture Measurements and Sensor Calibration

Measured Parameters and Statistical Analysis

Results and Discussion

Soil Moisture Analysis in ET-Based Scheduling

Soil Moisture Analysis of Soil Moisture–Based Scheduling

Comparison between Irrigation Regimes

Soil Moisture

Applied Water in the Various Treatments

Irrigation Regime Impact on Crop Growth

Impact on Shoot Height

Cob Length

Biomass and Yield Response

Aboveground Biomass

Root Biomass

Cob Weight, Grain Yield, and Grain to Cob Ratio

Harvest Index Response

Water Productivity Response

Sap Flow

Conclusion

Data Availability Statement

Acknowledgments

References

Information

Published In

Copyright

History

Authors

Affiliations

Metrics

Citations

Download citation

Figures

Other

Share

Copy the content Link

Share with email

Share

Request Username

Create a new account

Change Password

Password Changed Successfully

Verify Phone

Congrats!