Identifying Stacked Conservation Practices that Optimize Water Use in Agriculture

Objective 1:

Irrigation technologies, irrigation rates, and drought-tolerant genetics had only minor impacts on yield in 2019 at the Logan site (Data not shown) due to an abnormally wet spring. In contrast, most factors influenced silage corn yield in 2020 in Logan, and at the new established site in Vernal. The interaction of irrigation technology and irrigation rate was significant at both sites, with results suggesting that the higher efficiency sprinklers (LEPA, and LESA) can result in higher yields when irrigation rates are reduced. Treatment type was heavily influenced by the irrigation technology as well at both sites. The addition of the soil wetting agent did not improve yield at reduced rates and thus would not be warranted. At reduced irrigation rates, drought-tolerant genetics rarely improved yield. No-till only sometimes decreased yield and reductions were more pronounced when water stress was greater. Targeted 50% irrigation rate reductions to critical crop growth stages was often worse and never better than a straight 50% reduction in rate all season. The corn study will be repeated at the Vernal site with the additional factor of cover crops, and the third site at Cedar City, Utah will have a corn study with the similar factors. In conjunction with corn at the sites, alfalfa, small grains, and alternative crops will be examined at the three sites. This study aims to determine the long-term effect on crops when shorting irrigation consistently.

Vernal Corn 2020 Results. Bars with the same letter had the statistically equivalent yield. NT = no-till, DT = drought tolerant genetics, and CONV = conventional (tillage and no DT).

Corn soil moisture results for plots with tillage and with full irrigation (100% rate). This is only for the top 1 foot of soil (sensor placed at 6-inch depth). The second and third foot were also measured, but this was shown for simplicity. In tilled plots, the mobile drip irrigation (MDI) had much lower soil moisture all season long. LEPA had the greatest soil moisture for much of the growing season. This indicates that LEPA’s application efficiency may have been greater than other systems.

Corn soil moisture results for plots with no-till and drought tolerant corn genetics with full irrigation (100% rate). This is only for the top 1 ft of soil. In no-till plots, LEPA again had the greatest soil moisture for most of the growing season and MDI generally had the least.

A three-way mix of small grains was planted in April and harvested in July at the Vernal site in 2020. The mix included wheat, barley, and oat. It was planted on April 23 at 100 lbs/acre and harvested in mid-July. Small grain forage yield in 2020 range from 1.5 to over 3.5 tons per acre. The LESA system at full irrigation produced the greatest yield, while the MDI produced the least. All systems except LESA were able to maintain yield with 25% less water. MESA was the only system where the 50% reduction also maintained yield compared to full irrigation. This indicates that small grain forage has a great ability to withstand water stress, and that traditional MESA systems may have some of the greatest potential.

Small grain forage quality was only impacted by irrigation rates. Forage quality increased as water stress increased. This is likely because less growth in water-stressed areas can reduce plant stem, which can improve forage quality. Quality and yield data will be combined to examine the economics of these treatments. Soil moisture data in the top 1 foot of soil showed that LEPA had the greatest moisture in the first half of the season, but had the least during the second half. The other three systems had very similar trends throughout the season. 

Teff was the alternative crop in Vernal. Two teff grass trials were established. The first examined how the 16 irrigation treatments impacted yield and quality. At full irrigation, all irrigation systems besides MDI (poor yield in all treatments) performed equally well. None of the three advanced irrigation systems could maintain yield with 25% less irrigation. However, for MESA and LESA, the 50% reduction yielded the same as the 25% reduction in irrigation rate.

The second project was to examine how three different teff varieties handled water stress. Results showed that there was a different. Variety “A” and “B” could maintain yield with 25% less water, but variety “C” could not.

Teff forage quality was also measured in LESA irrigation systems only. Variety only impacted two quality parameters – digestible fiber and a forage quality metric called RFV (relative feed value). The partial irrigation treatment that had 50% irrigation with targeted applications had the best overall forage quality of the three varieties. Soil moisture was measured in one teff variety in two irrigation rates of each of the four irrigation systems. LESA had the lowest soil moisture for much of the growing season. This is likely because it had the best yield and the teff was likely using the most water. Likewise, MDI had the greatest moisture because teff growth was poor in these plots.

The studies were repeated at Logan and Vernal in 2021, and the third site in Cedar City was established. This resulted in a total of six site-years of data (3 from Logan, 2 from Vernal, and 1 from Cedar City). The results from all six site-years showed that drought-tolerant genetics for corn and alfalfa rarely impacted crop yield or forage quality. Likewise, we saw few to no interactions of crop genetics and irrigation rate. This indicated that drought tolerant genetics were not providing enhance resilience to water stress and may not be an effective option for optimizing water use. 

Results from the four irrigation systems over the six site-years showed that irrigation system performance varied by site and by year at sites with multiple years (Logan and Vernal). The LESA, LEPA, and LENA systems were often (50-100% of cases for alfalfa and corn) able to maintain or increase alfalfa and corn yield with less water. However, the MESA systems sometimes also had equivalent yield with 25% less water than the full irrigation rate. The MDI system rarely could maintain yield with full or reduced irrigation rates and is likely not a viable option unless filtration issued can be adequately addressed. These results indicated that low-elevation systems can be an effective tool for reducing water diversions, but that their benefits are highly site-specific and in some cases crop-specific. We are in the process of calculating and estimating evapotranspiration from soil moisture sensors to compare water consumption differences among all treatments. This will be completed in the summer of 2023. 

Early data on no-tillage and cover crops across all site-years indicates that they often reduced or maintained corn silage yield but were not able to increase yield, especially with water stress treatments. This data is preliminary as these treatments had only been implemented for 1-3 years. The second phase of this long-term study will provide robust data on how cover crops and no-till individually and combined will impact water use and silage corn yields over time. 

Objective 2:

An objective at the VRI field sites was to characterize the spatial variability of soil moisture and compare it to variability of several ancillary variables. Under uniform irrigation management in both Rexburg (2019 and 2020) and Grace (2016), there was significant variation in measured soil volumetric water content an every sampling date. Variation was large enough  to justify application of VRI.  At the Rexburg site, the volumetric water content, averaged over all sampling dates, best clustered into three unique zones (Figure 2A). When those zones were compared to three-zone maps of the field generated from ancillary variables (slope, elevation, yield, soil apparent electrical conductivity(EC)), all had between 40% and 50% agreement, with EC and elevation having the highest agreement. At the Grace location, the volumetric water content, averaged over all sampling dates, best clustered into five unique zones (Figure 2A). At this location elevation had the best agreement among the ancillary variables evaluated. Based on the results from both locations, we conclude one wants to delineate VRI zones based on an individual, simple ancillary variable, then elevation is likely to be a good choice. 

Figure 2A-2E. Three zone maps for the Rexburg site: average 2019 volumetric water content (VWC) (2A), slope (2B), Elevation (2C), Yield (2D), shallow EC (2E).

Figure 3A-E. Three zone maps for the Grace site: average 2016 VWC (3A), slope (3B), Elevation (3C), Yield (3D), shallow EC (3E).

Another objective of the VRI study was to determine the spatial variation of crop growth within the field and whether that should be considered in making VRI recommendations. We evaluated visible vegetation indexes (VVIs) based on UAV images for their ability to predict leaf area index (LAI) and describe the spatial variation of LAI within irrigated fields.  The relationship between wheat LAI and 12 unique VVIs were then compared to measured LAI values using simple linear regression. The results showed that three VVIs had near identical outputs. Visible atmospherically resistant index (VARI), normalized green-red difference index (NGRDI), and modified green-red vegetation index (MGRVI) all estimated LAI equally well (R²= 0.66, 0.66,0.66; RMSE=0.74,0.73,0.73; MAE=0.57,0.56,0.56). Those VVIs estimated most accurately when resampled to 3 m spatial resolution and resampled through the ladder resampling method. This study introduces a promising solution to remotely estimating LAI using only RGB cameras and consumer grade drones, potentially leading to other applications such as remotely estimating evapotranspiration of irrigated fields and delineation of zones in variable rate irrigation systems.  We then used the VARI derived equation to characterize the spatial variation of LAI for each field and each sampling date. Figure 4 shows an example of the variation of LAI at the Rexburg location and how the LAI data could be clustered into zones.  We have also documented that there was significant spatial and temporal variation of mean LAI across the pre-defined VRI zones at both field locations.

Figure 4 a: Raster image representing estimated LAI values for Rexburg May 2019 based on best LAI estimation methods and b: K means clustered image of Rexburg May 2019 with 4 equal classes of leaf area index values

The Grace location was managed using VRI in 2019 and 2020. For each irrigation event, soil moisture sensors in each zone were evaluated to determine the zone specific irrigation rate. Cumulative applied irrigation by zone is shown in figure 5 and is compared to the irrigation that would have been applied using the growers standard practice (GSP). There was meaningful variation in applied irrigation over the zones, with a difference of 60 mm of irrigation between the least and greatest zones. On average, the VRI approach reduced applied water by 40 mm. Although there was not a direct way to compare grain yield in the VRI system to what it would have been under the GSP, we did conduct a relative yield analysis between zones under the VRI to the same analysis from a previous year under uniform irrigation. This analysis suggests that the yield under the VRI system is equivalent to that of the GSP but used less water.    

Figure 5. Cumulative irrigation totals for a wheat crop in 2019 at the Grace, ID location under variable rate irrigation (VRI) with five irrigation zones compared to the estimated cumulative grower standard practice (GSP) total.

Another aspect of this work in 2020-2021 was to examine variable-rate irrigation approaches for alfalfa. This study was provoked by the potential of using leaf area index (LAI) to inform VRI recommendations within irrigated alfalfa. LAI has a direct connection with crop coefficient (Kc) used to estimate crop evapotranspiration (ETc). Irrigation zone specific Kc could improve ETc. Measuring LAI manually can be destructive. One approach to estimating LAI is the correlation to visible vegetation indices (VVIs) from an unmanned aerial vehicle (UAV). There are three objectives; (1) manually measure LAI and canopy height, (2) identify VVIs with canopy height that are correlated to LAI, and (3) translate UAV imagery into maps of LAI. The study was conducted in 2021-2022 in Rexburg, Idaho. Measured LAI varied from 0.23-11.28 and canopy height varied from 6 cm to 65 cm. There is strong spatial clustering in measured LAI but the spatial patterns are dynamic between dates. The VVI model with the highest coefficient of determination for estimating LAI was Modified Green Red Vegetation Index (MGRVI) and was improved when field average canopy height was included in the model R2=0.93. The MGRVI and canopy height model was applied to all the UAV imagery and there was statistically but not practical difference between zones for predicted LAI. These results have indicated that UAV imagery coupled with field average crop height can be a useful tool to predict LAI in alfalfa. This could help guide future precision irrigation in alfalfa and other forage crops. 

 Objective 3:

Surveys of Utah and Idaho growers were conducted to determine their current irrigation and nitrogen practices in corn, potatoes, and small grains. A total of 58 corn responses, 137 small grain responses, and 25 potato responses were collected from February to May 2020. Removing and cleaning the responses lead to much fewer usable responses of 41 corn, 90 small grain, and 15 potato, mutually exclusive of one another (n = 146). The corn responses included 16 grain corn and 25 silage corn and the small grain responses included 30 small grains forage and 60 small grains harvested for grain. The corn responses represented roughly 6,900 ha of production, small grain at 14,500 ha, and potato responses representing about 4,000 ha. This sample supplied three percent of responses from the contacted target population, which was calculated based on the number of usable responses compared to the number of participants emailed. Online only surveys typically have lower response rates compared to paper and phone counterparts. All the potato responses were from Idaho as Utah has minimal potato production. Most responses came from southeastern Idaho (8%) and northern Utah (32%).

The independent variables of acreage managed, percentage of acreage owned, manager years, and test plot occurrence played the biggest roles 4R nitrogen and irrigation practices for corn, potatoes, and small grains. There could be other factors at play such as income from off farm jobs and other aspects to the farm such as livestock components. Nitrogen management had the largest number of full responses. This is probably due to the questions being towards the beginning of the survey before respondents began to drop-off as the survey continued. The nature of the sensitivity to change questions or environmental concern questions may have also turned off participants from completing the survey.

It should be noted that the survey was conducted post-2019 season, and answers were based on the 2019 growing season – a record wet year for Idaho and Utah after several drought years. The survey occurred at the same time as the start of COVID19 pandemic shutdown. Responses could have been low as many of the growers are typically informed in person about Extension projects, or they have not provided up-to-date email addresses to the listserv. The combination of these factors could have led to the low response rate and the type of responses received as all online, opt-in type surveys get low response rates. Further digital surveys often have lower response rates than main-in or phone surveys (Dillman et al., 2021). Repeating this survey post- 2021, an extreme drought year, would lead to more realistic or change viewpoints in sensitivities to change. Though growers should have the ability to look forward towards change, many of the responses suggest this was not the case or answers had an elevated positive outlook of the future. In 2021 and going into the 2022 growing seasons, growers were faced with doubling costs of fertilizer; a follow-up survey would be interesting to see if growers did have to cut fertilizer rates, or they offset the costs in other ways.

The opportunities to expand the use of conservation practices are evident from the responses received, especially in the areas of precision application adoption with nitrogen and irrigation. Simple changes such as changing nitrogen or irrigation rates, especially in a precision technology application, could lead to higher investment returns and reduced resource losses. The sample population gave insight to a group that appears more willing to change – the less experienced smaller operations. The less experienced with smaller operations might be more willing to try new technologies, products, and methods as there could be a lower implementation cost or they are less set in their current ways. There are plenty of areas that need improvement as evident from the survey results, though the sample was small and not as ideally representative of all of Idaho and Utah crop growers. The survey results have been drafted into a journal article that will be submitted for publication in the spring of 2023.