Primary Research Location:  Halfway/Helms

Project Title:  Equipment Development for Site-Specific Irrigation and Chemigation

Project Objectives:

1.      Modify a center pivot irrigation system to provide variable quantities of water in a pre-programmed manner as the system circles a field.         

2.  Begin development of delivery systems for site-specific application of agricultural chemicals along the length of a pivot.

Reporting Period:  January 1, 2001 – December 31, 2001

A.     Summary of Progress:  

The first objective, modification of a center pivot irrigation system to provide variable-rate (VR) irrigation, was completed.   Spans 6, 7, and 8 of a system at the Helms Research Farm were modified resulting in VR irrigation coverage of 73 acres.  The hydraulic and control components of the VR system were evaluated as cotton was being irrigated in July and August of this year.  A field evaluation comparing VR to uniform rate (UR) water application was conducted with the different water treatments beginning in August.  An additional VR section (span 5) has since been installed bringing the total VR area to 90 acres.

The second objective (site-specific chemigation) has been addressed by installing a 160-ft section of an Accu-Pulse™ (Valmont Manufacturing, Valley, NE) chemigation system on a one-tower center pivot at Halfway. The system has been divided into three sections similar in length to the manifold units of the VR irrigation system at Helms.  An attempt will be made to adapt the Accu-Pulse™ system to the site-specific control being used with the VR irrigation system.  Evaluations of this system’s utility for fertilizer and foliar chemicals will be made.  

Variable rate irrigation equipment construction and evaluation.

Equipment Construction.  The VR system routes water from the pivot lateral through pressure regulators and solenoid valves to each of three manifolds comprising the manifold unit.  There are three manifold units per 160-ft pivot span.  Hoses are used to direct water from the manifolds to modified LEPA irrigation applicators.  Nozzle sizes for each applicator provide flow rates of 1x, 2x, and 3x, which, when actuated in various combinations, provide 6 discrete irrigation amounts ranging from 25 to 150% of a base irrigation rate.

Prior to field evaluations, high winds toppled the Helms pivot carrying the VR irrigation system on April 6, 2001.  A new pivot was installed and crop irrigations began May 26.  Removal, reconstruction and installation of the VR irrigation components took place in May, June and July with VR equipment evaluation beginning in July.  Original LEPA application devices were extensively modified to accommodate high water volumes without causing runoff.  Acceptable water distribution was achieved by renozzling the VR system to match well output. 

A control system was installed that actuated solenoid valves of each manifold unit relative to field location, thereby controlling irrigation quantities at specific sites.  A SNAP-LCSX-PLUS industrial controller (Opto 22, Temecula, CA), two remote terminal units (SNAP-B3000), software, and related accessories were installed for this purpose.  The control system was programmed to provide four control signals to each manifold unit (3 signals for 3 water manifold solenoids and an additional signal for a future chemigation actuator).  Programming further allowed changes in solenoid status every 3⁰ around the 360⁰ perimeter of the pivot.  Therefore, the largest control area under this VR pivot was less than 0.1 acre (53’ manifold unit length x 71’ in a three degree arc) resulting in over 2000 potential water/chemical control areas under this 133-acre pivot.  A standard incremental encoder (DynaparÔ Series E15) was used to provide an input signal to the controller to determine pivot location.  Figure 1 shows the Helms pivot, the position sensor location, the valve controller, the remote terminal unit at tower 7, a manifold unit assembly, and field acquisition of flow rate data by technicians.

Equipment Evaluation.  The mechanical evaluation of the VR system included tests of the hydraulic and positioning systems.  Water flow rates were determined by volumetric catchment from individual LEPA applicators in each of the nine (9) manifold systems during irrigation events from July through August.  Water pressure taps were positioned at strategic locations throughout the manifold units to determine pressure losses and help diagnose the cause of poor water distribution.  Figure 2 displays hydraulic performance data of the VR system at initial startup (August 4) and, again, following several modifications (August 30).  These charts show comparisons of desired, achievable, and actual flow rates of applicators within each of the nine manifold units of spans 6, 7, and 8.  Flow rates of individual manifold systems were offset from adjacent manifolds due to programmed differences in flow rates relative to the location in the field.  Data from the initial date indicate actual applicator flow rates were somewhat higher and more scattered than the achievable flow rates.  System improvements were made by increasing and stabilizing inlet water pressure at the pivot, renozzling the VR applicators, modifying plumbing components to prevent flow restrictions, and eliminating low-pressure drain valves.  Actual flow rates were acceptable (within 5% of achievable flow rates) by the end of the growing season. 

The controller, remote terminal units, and solenoid valves worked flawlessly (with the exception of one lightning strike), however, the positioning system used to actuate valves at appropriate locations in the field failed to perform as precisely as desired.  An evaluation was conducted which compared actual pivot location to the pivot location sensor outputs of both the VR positioning sensor and the pivot manufactures sensor.  Output data was systematically recorded as the pivot rotated around the field in both the clockwise and counter-clockwise directions.  A comparison of pivot and VR sensor response to actual position is shown if Figure 3.  The pivot and VR sensors showed deviations of up to 6⁰ from the actual at the 0/360⁰ field position (true north).  This represents a positioning error at the outer edge of the pivot of approximately 140 feet.  As the pivot rotated through the 120 to 200⁰ arc, the output signals of both sensors were consistently within a few degrees of the actual pivot position.  Position data were generally similar in both pivot directions at multiple revolutions.  The systematic difference between pivot and VR outputs indicates possible mechanical problems with the rack portion of the rack and pinion sensor mechanisms.  This error may be reduced by replacing pivot parts or by reprogramming the count sequence within the VR controller.  Error of up to 2⁰ may be acceptable for most irrigation or chemical applications.

Crop response to VR irrigation.

Methodology.  A field experiment was conducted to in an attempt to document advantages of variable-rate irrigation to standard uniform application rates in cotton production on the High Plains.   The experiment was conducted in a 12.2-acre area covered by the VR irrigation system.  This portion of the field contained the greatest elevation changes and the most notable differences in surface soil texture.  The 60⁰ arc was divided into 9 strips with each strip either 20 or 22 rows wide and falling beneath one of the 9 VR manifold units.  The experimental protocol called for alternating strips to be irrigated by either variable-rate (VR) or the uniform rate (UR) irrigation.  Comparisons of crop responses from these areas would be used to evaluate the VR system.  Figure 4 shows the position of the 12.2-area as it relates to the Helms pivot and the locations of the nine VR and UR treatment areas.

Past work at Halfway had shown variability in cotton lint yield most strongly correlated with factors associated with crop water use such as slope, elevation, soil texture, and seasonal irrigation (Bordovsky, et al., 1999 Annual Report, Precision Agriculture Initiative for Texas High Plains). Profile elevations and soil texture at 64 sites within the area were used as a basis to determine transitions in irrigation quantities in the VR strips.  Elevation data and row direction were used to determine the “slope down the furrow” at each of 64 referenced sites (Figure 5).  Soil texture below 40-cm had not been determined prior to initial VR irrigation on August 2, therefore, the only textural data used in the initial decision on water placement in VR strips was clay content in the top 40-cm (Figure 6).  A decision was made to divide the field into three zones for VR application.  The low-rate zone was irrigated at a rate equal to 75% of the uniform rate (UR) in the area where “slope down the furrow” was 0% and clay content in the top 40-cm was greater than 40%.  This area contained soils with high water holding capacity and limited risk of rainfall runoff.  The medium-rate zone was irrigated at 100% of the UR and included the area of “slope” between 0.0 and 0.5% and clay content of less than 40%.  The high-rate zone was irrigated at 125% of the UR in the area where slope down the furrow was greater than 0.5%.  The high-rate zone had the highest risk of rainfall losses.  Previously defined sampling sites also affected decisions on irrigation boundary positions since yield analysis required representative numbers of sites per zone.

 A Microsoft Excel program was written to create coded map files from the desired irrigation application map.  The application sequence was then loaded into the VR controller with a laptop computer.  The boundaries between zones of different irrigation levels are shown in Figures 4, 5, and 6.

Cotton was planted in the test area on May 24 and the crop maintained using normal cultural practices.  Nutrients were applied based on aggregate soil sampling and pests were treated at recommended thresholds.  Irrigation was initiated on May 26 and continued through August 30.  Due to the dry growing season and limited pumping capacity, irrigations in UR treatments were considerably less than 80% of estimated evapotranspiration.  Irrigation amounts of 5.59 inches were uniformly applied over the entire 12.2-acre field from May 26 to July 27.  From August 2 through August 30, irrigations totaled 3.87, 5.15, and 6.44 inches in the VR strips of the low-, medium-, and high-irrigated zones, respectively.  Therefore, the difference in total irrigation quantity between the low and high irrigation zones within the VR treatments was 2.54 inches.

Cotton yield response.  Yield response was determined by hand sampling at the 64 geo-referenced sites, by harvesting the entire area using a cotton stripper equipped with yield monitor, and by using stripper harvest, boll buggy weights from each of the 9 treatment areas (strips) under the pivot.  No differences in total yield or total irrigation water use efficiency were evident between VR and UR treatments.  Table 1 includes weighted irrigation amounts; yield based on burr cotton weights (boll buggy), average hand harvested lint weights within each strip, and integrated hand harvest lint weights; and total irrigation water use efficiency for VR and UR irrigation treatments.  Yield based on boll buggy weights were 719 verses 713 lb/ac for VR verses UR treatments.  Yield based on average hand samples were 966 lb/ac (981 lb/ac, integrated) from the VR irrigation treatment compared to 1004 lb/ac (1015 lb/ac, integrated) from the UR treatment.  asEstimates of water use efficiency were similar for the two treatments.

Although average yields were the same, spatial distribution of yields were quite different within the VR and UR irrigation treatment areas.  Figure 7 represents the integration of hand harvest data obtained at the 32 sites in the UR treatments.  This represents the yield response from uniformly irrigating the entire 12-acre area.  This map shows two general areas of lower yields, an area with no slope and high clay content (East side) and a sloping area (> 0.5%) with low clay content (West).  For comparison, the VR map shown in Figure 8 used only the yield data from the 32 VR sites.  This map indicates that shifting water from the west side of the field to the east side reduced lint yield in the low water zone and increased yield in the high water zone.    High yields seen on the west side of the VR map may be due to irrigation from the adjacent field (VR controller not actuating valves at the precise location).  

The small total yield difference between the two irrigation treatments was not unexpected.   Irrigations treatments were started late in the growing season, initial irrigations were being made with VR equipment that had not been fully optimized, data used to base VR irrigation transition zones were limited, and the strategy for creating the zones was based on more normal rainfall.  Timelier field experiments will be conducted in 2002.

B.          Education/technology transfer:          Elements related to site-specific irrigation were contained in the following presentations:

Bordovsky, J.P.  2000.  Irrigation research update.  Irrigation Update for the Texas High Plains.  Texas Agricultural Irrigation Association, TAES, TAEX.  Nov. 27, 2000, Amarillo, TX.

Bordovsky, J.P.  2001.  Workshop – water management problem solving.  Beltwide Cotton Conference.  Jan. 9-13, 2001.  Anaheim. CA.

Bordovsky, J.P.  2001.  Irrigation efficiency options.  Hale County Cotton and Soil Fertility Conference.  January 26, 2001.  Plainview, TX.

Bordovsky, J.P.  2001.  Irrigation research update.  Lubbock County Water Management Seminar.  Jan. 30, 2001.  Lubbock, TX.

Bordovsky, J.P.  2001.  Irrigation research in the Southern High Plains.  LEPA Irrigation System  Seminar.  Central Curry Soil and Water Conservation District.  March 8, 2001.  Clovis, NM.

Bordovsky, J.P.  2001.  Irrigation research.  Texas Cooperative Extension Crop and Drip Irrigation Tour.  Halfway, TX. September 11, 2001.

C.          Milestones achieved: 

August 2001.  The first variable-rate pivot using LEPA technology in the United States was used to irrigate cotton.

D.          Publications:

Machado, S., E.D. Bynum, Jr., T.L. Archer, R.J. Lascano, L.T. Wilson, J.P. Bordovsky, E. Segarra, K. Bronson, D.M. Nesmith, and W. Xu.  2000.  Spatial and temporal variability of corn grain yield: site-specific relationships of biotic and abiotic factors.  Precision Agriculture, (2) 359-376.

E.            Precision agriculture proposals:

None.

F.           Precision Agriculture meetings attended/papers (posters) presented:

None

G.         Other developments: (Anything that impacted research progress, positive or negative)

Negative

High wind toppled the VR irrigation system on April 6, 2001.  This interruption, along with very limited rainfall (requiring constant irrigation), greatly reduced the evaluation time planned for the VR system and limited the time and money available for site-specific chemigation development. 

Positive 

The VR system was quickly rebuilt due to the intense coordinated efforts of the staffs at Halfway and Lubbock.

A protocol for comparing VR to traditional irrigation methods was developed and was used in a field experiment during 2001.  The protocol continues to be evaluated.