Identifying Stacked Conservation Practices that Optimize Water Use in Agriculture

Objective 1: Identify combinations of six water conservation practices with the greatest ability to optimize water and energy use.

In the early spring of 2019, the first of three water optimization trials were established at the Utah State University (USU) Wellsville Farm. The second trial was established at the USU Vernal farm in 2020. Establishment first included installing a lateral irrigation system to irrigate the ~15 acre test/demonstration area at each site.

Sixteen different irrigation strategies were tested and demonstrated in 2019 and 2020 at both sites. This included all combinations of:

1. four irrigation systems [mid-elevation (MESA), two low-elevation sprinklers (LEPA and LESA), and mobile drip irrigation (MDI)]
2. four irrigation rates (100%, 75%, 50% of ET, and 50% with partial-season irrigation)

In addition to these irrigation strategies, we also examined different crop management practices with potential to improve water use efficiency, including drought tolerant genetics, no-tillage, alternative crops, and cover crops. Corn and alfalfa were grown at each site each year. The alternative crops varied by site and year and were safflower, hemp, and teff. A total of 300-400 plots were established to test these at each site. 

Crop yield, profit, and irrigation productivity have been determined for each treatment. Crop water balances are also being determined in order to evaluate and demonstrate consumptive water use. We installed a total of 120-300 soil moisture sensors in all four irrigation systems at the 100% and the 50% irrigation rates at each site. We also installed sensors to test differences between “conventional corn practices” vs. “most advanced management” that included no-till, drought-tolerant genetics, and cover crops.

Objective 2: Expand variable-rate irrigation research to provide dynamic irrigation prescriptions based on soils and yield to improve water and energy use efficiencies.

The pilot VRI study that was initiated on an Idaho field in 2016 showed a 25% irrigation water savings, reduced variation but similar total yield compared to those expected under uniform irrigation. 

To expand the existing pilot study, three VRI research and demonstration sites are being established across Southern Idaho (Twin Falls, Pocatello, and Rexburg). The Pocatello, ID location will expand on the pilot study and will be led by our producer representative, who utilizes a VRI system on a field in a potato-wheat rotation. 

Here are reports for each of the three sites:

Rexburg: In 2019, a wheat field in Rexburg Idaho was chosen for VRI research. Soil samples were collected on April 29th just after planting, May 30th, June 25th, and August 29th, just after harvest. Sample locations were 66 points based two superimposed grids to have samples that were spaced at a range of sampling intervals (Figure 1). Field-moist soil samples were collected using augers driven into the soil with a jack hammer and stored in plastic bags. Samples were weighed, dried at 105^o for 24 hours and re-weighed. From these wet and dry weights volumetric water content (VWC) values were determined. These were interpolated (kriged) to a 5m grid then values for the four measurement times were averaged.  

       A k-means clustering algorithm was used in SPSS to group the average VWC data into 2-7 zones. The largest break of slope in the mean squared errors was for 3 zones so 3 zones was assumed to be optimal.  The 3-zone VWC map was used to compare with 3-zone maps produced using other ancillary variables that could be inexpensive and effective for delineating management zones.

       Ancillary variables included yield data from a yield monitor, normalized difference vegetation index (NDVI) from LandSat 8 imagery (June and July 2014-2019), elevation (collected from the GPS on the yield monitor) and derived topographic attributes such as slope, aspect and topographic wetness index (TWI).  These datasets were kriged to the same 15 ft grid that as the VWC data and were used individually and in different combinations to create more 3-zone maps using k-means clustering. The percentage agreement between the different 3-zone maps was determined.

In 2020 the field was again managed with uniform irrigation and soil moisture was measured on the same spatially dense grid on four dates. A full set of soil moisture sensors were installed in each zone. The set of sensors included water content sensors (Teros 12, Meter Group Inc) at 6, 18, and 30 inch depths) and water potential sensors (Teros 21, Meter Group Inc.) at 6 and 18 in depths.

Remote sensing objectives were accomplished using a DJI Phantom 4 outfitted with a Sentera single NIR sensor and RGB camera. The Phantom 4 Pro Camera has a 1 inch CMOS with 60 ft effective pixels. A 3 axis gimbal integrated with the inertial navigation system stabilized the camera during flight. The UAV followed automated flight paths created with DroneDeploy (DroneDeploy, 2020) that imaged the entirety of each field at 290 ft above ground level (AGL) and 308 ft above ground level (AGL) with at least 80% image side and end overlap in the flight paths. Images for Rexburg were collected May 30, 2019 and June 25, 2019 during optimum UAV operating times (11:00 am-2:00 pm). The data were processed in Web OpenDroneMap (ODM, 2020). All the individual photos of each field were imported into WebODM and Reviewed. WebODM then automated the stitching together of the separate images into a multiple orthomosaics of the imaged fields. These were exported as a GeoTIFF to be processed in ArcGIS Pro. The RGB images were acquired with a ground resolution of 0.02 ft at 290 ft height and were saved with GPS location information in GeoTIFF format.   

In both years, wheat grain yield was measured using an Ag Leader Versa (Ames, IA, USA) with a TerraStar-C Pro GPS with a relative accuracy of +/- 5 cm. The yield monitor was calibrated using a mass flow sensor collecting yield data at a 32 ft² spatial density.  Erroneous data points were defined as outside the limits of ± 75% of the median. In addition, points where the combine did not harvest the full header width were removed.

 Twin Falls: The VRI project is located under a single-span VRI pivot at the University of Idaho Kimberly Research and Extension Center. The field site is on a hillside of eroded soil. The upslope half of the pivot is approximately 3% slope of eroded Portneuf silt loam soil. Low infiltration rates and excessive surface runoff due to surface seal formation are serious issues. The lower half of the pivot is approximately 1% slope Portneuf silt loam with relatively deep topsoil resulting from deposition of topsoil eroded upslope. This portion of the field has less tendency for surface seal formation and runoff issues. Crop productivity is also higher than for the more sloping portion of the pivot.

In 2019, plots were arranged so that equal numbers of plots of each crop were located in each quadrant of the pivot. In 2019, all plots received the same irrigation and fertility applications. Crop yield varied considerably, based on location on the hillside for each of the 3 crops. Part of the difference appears to be due to overall fertility / soil tilth conditions and part due to differences in water available for crop growth due to infiltration / runoff differences with slope position. End-of-season soil water content was measured for each of the 48 plots at 6-inch depth increments to 5 feet. Plots on the top (more sloping and eroded) half of the field had a noticeably lower post-harvest water content, more visual indications of crop water stress during the growing season, and lower crop yields relative to the lower slope, non-eroded portion of the field. Data from 2019 were used to develop VRI prescriptions for testing during the 2020 growing season for spring wheat. VRI was applied during the 2020 growing season. Soil moisture data, crop yield, and associated data are still being summarized.

Grace: The field site is near Grace, Idaho and winter wheat (Triticum aestivum L.) was growing in 2019. The field was equipped with a variable rate irrigation (VRI) system. The sampling scheme at Grace included 100 soil sample locations (Figure 1) with soil collected to compute volumetric water content (VWC) on April 23, May 30, June 25, and September 5 of 2019. Beginning sampling was at spring green-up, and the final sampling was shortly after harvest. In this study, the data from May 30th were focused on. Five irrigation zones were created and managed from soil sensors with sensor location installed in each zone. Soil sensors collected VWC and irrigation was applied to reach field capacity. A large portion of the west side of the field had ponding due to snow melt early in the season so wheat was replanted in spring, causing a late growth in this area compared to the rest of the field. In 2020 the field was managed using VRI based on measured soil water content in each zone, but no soil sampling was performed in 2020. The yield monitor and GPS in the combine collected spatially referenced yield data. SpaceStat and ArcPro were used for all spatial and statistical analysis.  Remote sensing was done as describe for the Rexburg location, with flights May 31, 2019, June 26, 2019 and July 8, 2020, during optimum UAV operating times (11:00 am-2:00 pm).

Figure 1. Map of research site near Grace, ID, USA (A), and research site in Rexburg, ID, USA (B) with VWC sample locations.

Objective 3: Determine producer’s attitudes, acceptance rates, and barriers to adoption of water conservation practices in order to better adapt and target Extension and outreach efforts.

Three nearly identical online 8-page surveys were developed in 2019 for corn, potatoes, and small grains. The questions were developed to ask Idaho and Utah growers how they manage irrigation and nitrogen needs. Questions were formulated to ask in an order of finding out the scope of the production, planting populations, fertilization plans, and irrigations plans. For the N fertilizer section, growers are asked questions relating to the time of year of application, rate of application, application practices, and application source. The irrigation questions are similar in nature, asking about irrigation type, scheduling, amounts, and frequencies. For both areas, growers are questioned about advanced practices such as variable rate applications. The last few questions of each crop survey for both nitrogen fertilizer and irrigation question growers on how they would handle a price increase of each of the inputs. This will allow us to assess both irrigation and nitrogen 4R adoptions and barriers. The survey was emailed to about 7,000 growers in Utah and Idaho during the early part of 2020. Preliminary results were presented in several local, regional, and national meetings. We are currently conducting the final analyses in preparation for publication.