Journal list menu

Volume 6, Issue 1 e20029
CROP MANAGEMENT
Open Access

Developing sensor-based irrigation scheduling that maximizes soybean grain yield, irrigation water use efficiency, and returns above irrigation costs

C. W. Wood

C. W. Wood

Growers, Garner, NC, 27529 USA

Search for more papers by this author
L. J. Krutz

L. J. Krutz

Water Resource Research Institute, Mississippi State University, Starkville, MS, 39762 USA

Search for more papers by this author
W. B. Henry

W. B. Henry

The Graduate School, Mississippi State University, Starkville, MS, 39762 USA

Search for more papers by this author
T. Irby

T. Irby

Department of Plant and Soil Science, Mississippi State University, Starkville, MS, 39762 USA

Search for more papers by this author
J. M. Orlowski

J. M. Orlowski

Growers, Garner, NC, 27529 USA

Search for more papers by this author
C. J. Bryant

C. J. Bryant

Crop & Soil Sciences, University of Georgia, Tifton, GA, 31793 USA

Search for more papers by this author
R. L. Atwill

R. L. Atwill

Indigo Ag, Memphis, TN, 38103 USA

Search for more papers by this author
G. D. Spencer

Corresponding Author

G. D. Spencer

Delta Research and Extension Center, Mississippi State University, Stoneville, MS, 38776 USA

Correspondence

G. D. Spencer, Mississippi State University, Delta Research and Extension Center, Stoneville, MS.

Email: [email protected]

Search for more papers by this author
B. E. Mills

B. E. Mills

Delta Research and Extension Center, Mississippi State University, Stoneville, MS, 38776 USA

Search for more papers by this author
First published: 16 April 2020
Citations: 8

Abstract

Agricultural withdrawal from the Mississippi River Valley Alluvial aquifer (MRVAA) exceeds its recharge rate, causing groundwater declines and cones of depression across the midsouthern United States. This research was conducted to determine whether sensor-based irrigation scheduling of soybean [Glycine max (L.) Merr.] could improve yield and profitability while minimizing consumptive water use. The effects of season-long soil water potential (SWP) threshold (−50 and −85 centibar [cbar]) on consumptive water use and on agronomic and economic parameters were compared with scheduling irrigations as a function of growth stage (VN–R2, R3–R4, and R5–R6.5) and SWP threshold (−50, −85, and −125 cbar) on a Dundee silty clay loam (fine-silty, mixed, active, thermic Typic Endoaqualfs) near Stoneville, MS, from 2015 through 2017. Decreasing the season-long SWP threshold for soybean from −50 to −85 cbar reduced consumptive water use 49% while having no adverse effect on yield, returns above irrigation costs, or irrigation water use efficiency (IWUE). Adjusting the irrigation threshold as a function of growth stage did not improve yield or returns above irrigation costs. Our data indicate that maintaining a season-long irrigation threshold of −85 cbars optimizes yield, returns above irrigation costs, and IWUE efficiency while reducing withdrawal for soybean irrigation from the MRVAA up to 49% relative to the −50 cbar producer standard.

Abbreviations

  • IWUE
  • irrigation water use efficiency
  • MRVAA
  • Mississippi River Valley Alluvial aquifer
  • SWP
  • soil water potential
  • 1 INTRODUCTION

    The Mississippi River Valley Alluvial aquifer (MRVAA) is the primary water source for row-crop irrigators in the midsouthern United States, an area encompassing the bootheel of Missouri and the delta regions of Arkansas, Louisiana, Mississippi, and Tennessee. Agricultural withdrawal from the MRVAA exceeds its recharge rate, which has caused water levels in the aquifer to decline and cones of depression to appear in the delta regions of Arkansas, Missouri, and Mississippi (Karki, Tagert, Paz, & Pérez-Gutiérrez, 2015). Reducing row-crop withdrawal from the MRVAA without having an adverse effect on yield and net returns will improve the sustainability of the region's irrigated agriculture.

    Soybean occupies the majority of arable acres in the midsouthern United States and accounts for approximately 50% of the annual withdrawal from the MRVAA (Massey et al., 2017). Producers primarily seed maturity Group IV cultivars and deliver irrigation using a conventional, continuous-flow delivery system (Bryant et al., 2017). Depending on soil texture, producers irrigate on a 7- to 10-d calendar schedule, initiating during early reproductive growth and terminating near physiological maturity (T. Irby, personal communication, 2017). Soil moisture sensors may improve irrigation timing by measuring in situ soil moisture and thus offer an advantage over calendar-based scheduling.

    Scheduling irrigations based on in situ soil moisture measurements may decrease consumptive water use and improve or maintain irrigation water use efficiency (IWUE), yield, and profitability compared with calendar-based irrigation scheduling. Relative to the producer standard of −50 cbar, a soil water potential (SWP) threshold of −75 to −100 cbar decreased consumptive water use up to 40% and maintained or improved yield up to 3%, IWUE up to 260%, and net returns up to 16% in corn (Zea mays L.), cotton (Gossypium hirsutum L.), and soybean in the delta regions of Mississippi and Arkansas (Bryant et al., 2017; Plumblee et al., 2019; Spencer et al., 2019). There are concerns that the soil moisture potential threshold should vary according to plant growth stage rather than remaining constant throughout the growing season. The objective of this research was to determine the effects of scheduling irrigations as a function of growth stage (VN–R2, R3–R4, and R5–R6.5) and SWP threshold (−50, −85, and −125 cbar) on consumptive water use and on agronomic and economic parameters compared with season-long SWP thresholds of −50 and −85 cbar.

    2 SITE DESCRIPTION AND EXPERIMENTAL DESIGN

    Research was conducted near Stoneville, MS, from 2015 through 2017 on a Dundee silty clay loam (fine-silty, mixed, active, thermic Typic Endoaqualfs). Soybean variety AG4632 (Monsanto Co.) was seeded at a depth of 1.2 inches at 140,000 seeds acre−1 into 40-inch-wide raised seed beds with a John Deere MaxEmerge planter (John Deere Seeding Group). Experimental units were 100 ft long by four rows wide and in a split-plot arrangement within a randomized complete block design with four blocks. The whole-plot factor was soybean growth stage (VN–R2, R3–R4, R5–R6.5), and the subplot factor was irrigation threshold (−50, −85, and −125 cbar). Outside of a treatment's growth stage, the irrigation SWP threshold was −75 cbar. Three controls were included for comparison: rainfed, season-long −50 cbar, and season-long −85 cbar SWP thresholds.

    3 SENSOR-BASED SCHEDULING AND IRRIGATION DELIVERY

    Three Watermark Model 200SS soil moisture sensors (The Irrometer Co., Inc.) were installed in the third replication at depths of 6, 12, and 24 inches. Furrow irrigation was initiated once the weighted average of sensors over the 24-inch depth reached threshold (Table 1) (Plumblee et al., 2019). Irrigation was delivered to the experimental units through 12-inch-diameter by 9-mm-thick, lay-flat polyethylene tubing (Delta Plastics) placed perpendicular to the planted rows. Holes were punctured for every furrow, and hole size was selected according to the Pipe Hole and Universal Crown Tool (PHAUCET version 8.2.20, USDA-NNatural Resources Conservation Service) with input parameters as described by Bryant et al. (2017). Four acre-inches were applied at a flow rate of 3 gal−1 min−1 furrow−1 during each irrigation event to the middle three furrows of each plot. Flow rates and cumulative water applied were determined at the field inlet with a McCrometer flow tube with attached McPropeller bolt-on saddle flowmeter (McCrometer Inc.). Irrigation was terminated at the R6.5 growth stage on all treatments, and remaining agronomic practices were conducted according to Mississippi State University Extension Service recommendations (Bond et al., 2017; Catchot et al., 2017).

    TABLE A. Useful conversions
    To convert Column 1 to Column 2, multiply by Column 1 suggested unit Column 2 SI unit
    0.304 foot, ft meter, m
    2.54 inch centimeter, cm
    0.405 acre hectare, ha
    3.78 gallon, gal liter, L
    67.19 60-lb bushel per acre, bu/acre kilogram per hectare, kg/ha
    1 centibar, cbar kilopascal, kPa
    10.26 acre-inch hectare-millimeter, ha-mm
    6.09 bushel per acre-inch, bu/acre-inch kilogram per hectare-millimeter, kg/ha-mm
    TABLE 1. Centibar thresholds used during a furrow-irrigated soybean study comparing season-long and growth stage–based irrigation threshold treatments near Stoneville, MS, from 2015 through 2017
    Growth stage
    Treatment VN R1 R2 R3 R4 R5 R6
    VN–R2, −50 −50 −50 −50 −75 −75 −75 −75
    VN–R2, −85 −85 −85 −85 −75 −75 −75 −75
    VN–R2, −125 −125 −125 −125 −75 −75 -75 −75
    R3–R4, −50 −75 −75 −75 −50 −50 −75 −75
    R3–R4, −85 −75 −75 −75 −85 −85 −75 −75
    R3–R4, −125 −75 −75 −75 −125 −125 −75 −75
    R5–R6.5, −50 −75 −75 −75 −75 −75 −50 −50
    R5–R6.5, −85 −75 −75 −75 −75 −75 −85 −85
    R5–R6.5, −125 −75 −75 −75 −75 −75 −125 −125
    Season-long, −50 −50 −50 −50 −50 −50 −50 −50
    Season-long, −85 −85 −85 −85 −85 −85 −85 −85
    Rainfed

    4 MEASURED PARAMETERS

    Soybean grain yield was determined by mechanically harvesting the center two rows of each plot at physiological maturity with a calibrated, onboard yield monitor (Ag Leader Technology) and adjusting yields to 13% moisture content. Irrigation water use efficiency was calculated as described by Vories, Tacker, and Hogan (2005):
    IWUE = Y IW A
    where IWUE is irrigation water use efficiency (bu acre-inch−1), Y is soybean grain yield (bu acre−1), and IWA is irrigation water applied (acre-inch).

    5 ECONOMIC ANALYSIS

    Returns above irrigation costs were determined using partial budget analysis techniques. Costs associated with irrigation setup and water lifting were obtained from Mississippi State University 2018 Delta planning budgets and are representative of a furrow-irrigated, 160-acre, precision-leveled field with lay-flat polyethylene pipe (Mississippi State University, 2017). Including all direct and fixed costs, irrigation setup costs were $74.30 acre−1 (Table 2). Water lifting costs were $2.01 acre-inch−1. Cost variation among irrigation treatments was due to differences in the volume of water applied to each treatment. Soybean price was held constant at $10.00 bu−1 for all 3 yr of the research.

    TABLE 2. Direct and fixed costs associated with irrigation setup on an acre basis for a furrow-irrigated soybean study conducted near Stoneville, MS from 2015 through 2017
    Operation/input Direct cost Fixed cost Total cost
    $ acre−1
    Land plane 1.56 1.58 3.14
    Set up engine 0.23 0.00 0.23
    Ditcher 0.35 0.31 0.66
    Polyethylene pipe 8.38 0.00 8.38
    Lay polyethylene pipe 2.54 0.66 3.20
    Pick up polyethylene pipe 1.02 0.98 2.00
    Land forming ($450) 0.00 31.93 31.93
    Well and pump, furrow 3.01 8.74 11.75
    Main line pipe 0.00 4.32 4.32
    Engine 0.00 8.43 8.43
    Soil moisture sensors 0.26 0.00 0.26
    Total 17.09 56.95 74.30

    6 STATISTICAL ANALYSIS

    All data were subjected to ANOVA with the GLIMMIX procedure of SAS (Release 9.4; SAS Institute Inc.). Season-long SWP thresholds were analyzed as a one-way treatment structure with year as a random effect. Growth stage–based SWP thresholds were analyzed as a two-way treatment structure with growth stage and threshold as fixed effects. Random effects for the growth stage × threshold analysis included year, replication within year, year × growth stage, replication × growth stage within year, year × cbar threshold within growth stage, and growth stage within year × cbar threshold. Degrees of freedom were estimated using the Kenward–Roger method (Kenward & Roger, 1997). All means were separated using the LSMEANS statement. Differences were considered significant at α = .05.

    7 SEASONAL RAINFALL

    Rainfall totals for the 2015 to 2017 growing seasons ranged from 5.6% below to 23.4% above the 10-yr average rainfall totals (Table 3). The 2015 growing season is considered a dry year with 5.6% less rainfall than the 10-yr average, whereas 2016 and 2017 are considered wet years with 23.4 and 9.9% more rainfall than the 10-yr average, respectively. Above-average rainfall in July and August of 2016 and 2017 coincided with peak soybean water demand, reducing irrigation requirements. These conditions resulted in three irrigation treatments (−85 cbar at VN–R2, −125 cbar at VN–R2, and −125 cbar at R3–R4) not applied throughout the duration of the study (Table 4). All other treatments were applied in at least one of the three years of the experiment.

    TABLE 3. Monthly rainfall totals received March through October, plus 10-yr average, during a furrow-irrigated soybean study comparing season-long and growth stage–based irrigation threshold treatments near Stoneville, MS, from 2015 through 2017
    Month 2015 2016 2017 10-yr average
    inches
    March 7.3 18.5 2.9 5.6
    April 6.3 4.3 6.6 5.6
    May 6.9 3.3 4.9 5.2
    June 2.6 5.1 7.6 2.9
    July 3.2 6.5 4.3 4.1
    August 0.8 5.5 10.7 2.9
    September 0.8 0.3 1.7 3.8
    October 5.5 0.2 0.2 5.3
    Total 33.4 43.7 38.9 35.4
    TABLE 4. Irrigation water applied at specific growth stages and season totals by year during a furrow-irrigated soybean study comparing season-long and growth stage–based irrigation threshold treatments near Stoneville, MS, from 2015 through 2017
    Growth stage
    Treatment VN R1 R2 R3 R4 R5 R6 Total
    acre-inch
    2015
    VN–R2, −50 4.0 8.0 8.0 20.0
    VN–R2, −85 4.0 4.0 8.0 16.0
    VN–R2, −125 4.0 8.0 12.0
    R3–R4, −50 4.0 4.0 4.0 12.0
    R3–R4, −85 8.0 8.0
    R3–R4, −125 4.0 4.0 8.0
    R5–R6.5, −50 16.0 12.0 28.0
    R5–R6.5, −85 4.0 12.0 12.0 28.0
    R5–R6.5, −125 4.0 8.0 12.0
    Season, −50 4.0 16.0 12.0 32.0
    Season, −85 4.0 8.0 12.0
    2016
    VN–R2, −50 4.0 4.0 8.0 4.0 20.0
    VN–R2, −85 4.0 4.0 8.0
    VN–R2, −125 4.0 4.0 4.0 12.0
    R3–R4, −50 4.0 8.0 12.0
    R3–R4, −85 4.0 4.0 8.0 4.0 20.0
    R3–R4, −125 4.0 8.0 4.0 16.0
    R5–R6.5, −50 4.0 4.0 8.0 4.0 20.0
    R5–R6.5, −85 4.0 4.0 8.0 16.0
    R5–R6.5, −125 4.0 4.0 4.0 12.0
    Season, −50 4.0 4.0 4.0 8.0 4.0 24.0
    Season, −85 4.0 4.0 8.0 4.0 20.0
    2017
    VN–R2, −50 4.0 4.0
    VN–R2, −85
    VN–R2, −125 4.0 4.0
    R3–R4, −50 4.0 4.0 8.0
    R3–R4, −85 4.0 4.0
    R3–R4, −125
    R5–R6.5, −50 12.0 4.0 16.0
    R5–R6.5, −85
    R5–R6.5, −125 4.0 4.0
    Season, −50 4.0 12.0 4.0 20.0
    Season, −85 4.0 4.0

    8 SEASON-LONG THRESHOLD ANALYSIS

    A central hypothesis of this research was that decreasing the season-long SWP irrigation threshold from −50 to −85 cbar will decrease total water applied with no adverse effect on yield, net returns, or IWUE. Experimental results supported this hypothesis. Relative to the −50 cbar threshold, a season-long threshold of −85 cbar decreased water use by 49% (Table 5). Yield, returns above irrigation costs, and IWUE were not different between season-long thresholds, with means of 61.1 bu acre−1, $485 acre−1, and 4.95 bu acre-inch−1, respectively. These data demonstrate that a −85 cbar threshold will reduce consumptive water use without sacrificing yield, profitability, or IWUE in midsouthern U.S. soybean production.

    TABLE 5. Soybean grain yield, water use, irrigation water use efficiency (IWUE), and returns above irrigation costs for season-long threshold treatments in a furrow-irrigated soybean study comparing season-long and growth stage–based irrigation threshold treatments near Stoneville, MS, from 2015 through 2017
    Centibar threshold Yield Water use IWUE a Net returns
    bu acre−1 acre-inch bu acre-inch−1 $ acre−1
    −50 62.4 (2.8)a 23.8 (3.0)a 2.6 (2.4) 487 (17.6)
    −85 59.7 (2.8)a 12.2 (3.0)b 7.3 (2.4) 483 (17.6)
    Rainfed 53.8 (2.7)b 513 (17.2)
    • Note. Values within parentheses are the SEM. Values within a column followed by the same letter are not different from one another at the .05 level of significance.
    • aIrrigation water use efficiency defined as irrigation water applied divided by yield.

    Optimizing the timing of irrigations for soybean and other row crops in the midsouthern United States with soil moisture sensors may reduce withdrawal from the MRVAA up to 39.5% relative to producer standards. At the field scale, optimizing delivery of water in furrow-irrigated environments and scheduling irrigations for soybean and corn based on an SWP threshold of −75 to −100 cbar decreased total water applied up to 39.5% relative to the regional standard (Bryant et al., 2017; Spencer et al., 2019). Similarly, at the plot scale, irrigating cotton at an SWP threshold of −90 cbar rather than −50 cbar decreased the amount of water applied up to 63% in two of three years (Plumblee et al., 2019). Sensor-based scheduling of irrigations for rice (Oryza sativa L.) managed under alternate wetting and drying reduced water applied up to 40% relative to rice managed under a conventional 2- to 4-inch flood (Gholson, 2020). These data indicate that sensor-based scheduling reduced the amount of water applied to row crops, thereby increasing sustainability of the MRVAA if producers adopt these irrigation water management strategies. However, producers are reluctant to adopt best management practices if they perceive the new technology will have an adverse effect on yield or profitability (Asci, Borisova, & VanSickle, 2015).

    Sensor-based irrigation scheduling of soybean and other midsouthern United States crops either maintains or improves yield relative to the producer standard. Decreasing the irrigation threshold from −50 to approximately −87.5 ± 12.5 cbar increased corn grain yield 3% at 18 paired locations in the prairie region of Arkansas and the delta region of Arkansas and Mississippi (Spencer et al., 2019). Alternating wetting and drying cycles based on soil moisture thresholds rather than maintaining the standard 2- to 4-inch flood improved rice grain yield 2.5% (Gholson, 2020). Scheduling the irrigation of cotton at −90 cbar rather than the regional standard maintained or improved lint yield up to 33% (Plumblee et al., 2019). Similar to the results of this study, decreasing the irrigation threshold from approximately −50 to −92.5 ± 7.5 cbar had no adverse effect on soybean grain yield for soil textures ranging from very fine sandy loam to clay (Bryant et al., 2017). These data indicate that sensor-based scheduling will either have no adverse effect or improve profitability if the cost of soil moisture sensors is offset by an increase in yield or a reduction in fuel costs associated with water lifting.

    Relative to the regional standard, scheduling irrigations based on soil moisture sensors either maintains or slightly improves net returns. Producer-generated, field-scale research for corn, soybean, and rice indicate that if the irrigation threshold is based on Mississippi State University (MSU) Extension Service recommendations rather than the regional standard, net returns were either maintained or increased up to $80 acre−1 for pumping depths ranging from 18 to 400 ft and diesel prices ranging from $1.59 to $3.70 gal−1 (Bryant et al., 2017; Gholson, 2020; Spencer et al., 2019). Similarly, decreasing the irrigation threshold for cotton from −50 to −90 cbar increases net returns 16% (Plumblee et al., 2019). Transitioning from traditional irrigation thresholds to those recommended by MSU Extension decreases diesel and maintenance costs, which, at a minimum, offsets the additional costs of purchasing and maintaining sensor scheduling tools. Because sensor-based SWP thresholds maintain or improve profitability, it is probable that producers will adopt this irrigation water management practice (Asci et al., 2015). The widespread adoption of sensor-based scheduling should have a positive effect on the amount of yield derived per acre-inch of water applied.

    Similar to the results from this study, sensor-based irrigation scheduling either maintained or improved IWUE for the majority of row crops seeded in the midsouthern United States. Scheduling with sensors rather than the producer standard increased IWUE on production-scale irrigation sets 36, 51, and 70% for soybean, corn, and rice, respectively (Bryant et al., 2017; Gholson, 2020; Spencer et al., 2019). Transitioning from the producer standard to a −90 cbar threshold either had no effect or increased IWUE of cotton up to 260% (Plumblee et al., 2019). The in situ quantitative nature of sensor-based scheduling enhances the ability to predict soil moisture levels as they approach the maximum allowable deficit. More accurate forecasting with sensor-based scheduling increases rainfall capture, decreases water applied, and has no adverse effect on yield, thereby increasing IWUE relative to regional standards.

    9 ADJUSTING IRRIGATION THRESHOLDS AS A FUNCTION OF GROWTH STAGE

    A secondary objective of this study was to determine if adjusting the irrigation threshold as a function of growth stage could reduce the amount of water applied and improve IWUE with no adverse effect on yield or net returns. Growth stage and SWP threshold interacted to effect water use (p = .0002), soybean grain yield (p = .0012), and returns above irrigation costs (p = .0439) (Table 6). There was no effect of growth stage or irrigation threshold, and there was no interaction between these parameters on IWUE (p ≥ .1996). Adjusting the irrigation threshold above or below −85 cbar within a growth stage never decreased the amount of water applied or improved IWUE without having an adverse effect on yield and net returns. For example, changing the threshold from −85 to −125 cbar during the R5 to R6.5 growth stages decreased the amount of water applied by 46% and had no effect on IWUE or net returns but decreased yield by 6.7%. Altering the irrigation threshold from −85 to −50 cbar during the R5 to R6.5 growth stages had no effect on yield, net returns, or IWUE but increased consumptive water use by 54%. These data indicate that adjusting the irrigation threshold as a function of growth stage cannot reduce consumptive water use or improve IWUE without having an adverse effect on yield.

    TABLE 6. Soybean grain yield, water use, irrigation water use efficiency (IWUE), and returns above irrigation costs for growth stage–based threshold treatments in a furrow-irrigated soybean study comparing season-long and growth stage–based irrigation threshold treatments near Stoneville, MS, from 2015 through 2017
    Growth stage Centibar threshold Yield Water use IWUE a Net returns
    bu acre−1 acre-inch bu acre-inch−1 $ acre−1
    VN–R2 −50 59.6 (2.5)ab 13.3 (1.7)bc 7.0 (2.6) 480 (27.9)a
    VN–R2 −85 60.4 (2.5)ab 8.0 (1.7)d 8.8 (2.8) 499 (27.9)a
    VN–R2 −125 60.1 (2.5)ab 9.3 (1.7)cd 8.2 (2.6) 493 (28.1)a
    R3–R4 −50 58.4 (2.5)bc 9.3 (1.7)cd 6.2 (2.6) 476 (27.9)ab
    R3–R4 −85 58.8 (2.5)abc 10.7 (1.7)bcd 8.0 (2.6) 478 (27.9)ab
    R3–R4 −125 60.1 (2.5)ab 9.3 (1.7)cd 8.4 (2.6) 493 (27.9)a
    R5–R6.5 −50 61.9 (2.5)a 22.7 (1.7)a 2.9 (2.6) 483 (27.9)a
    R5–R6.5 −85 59.7 (2.5)ab 14.7 (1.7)b 4.0 (2.9) 477 (28.1)ab
    R5–R6.5 −125 55.7 (2.5)c 8.0 (1.7)d 5.6 (2.9) 452 (28.1)b
    • Note. Values within parentheses are the SEM. Values within a column followed by the same letter are not different from one another at the .05 level of significance.
    • aIrrigation water use efficiency defined as irrigation water applied divided by yield.

    10 CONCLUSION

    This research was conducted to determine whether sensor-based irrigation scheduling of soybean can improve yield and profitability while minimizing consumptive water use from the MRVAA. Maintaining a season-long irrigation threshold at −85 cbar rather than the regional standard of −50 cbar reduced consumptive water use without sacrificing on-farm yield and profitability. Adjusting the irrigation threshold as a function of growth stage only decreased consumptive water use or improved IWUE with a corresponding decrease in productivity. Our data indicate that maintaining a season-long threshold of −85 cbar maximizes yield and net returns while reducing withdrawal from the MRVAA for the irrigation of soybean by 49%.

    CONFLICT OF INTEREST

    The authors report no conflicts of interest.