Nitrogen management poses myriad challenges to the sustainable production of corn. In typical corn production systems, N fertilizer is the most expensive input, the largest contributor to greenhouse gas emissions, the nutrient most limiting to crop growth, and the nutrient with the most pathways to become a pollutant in the environment (4). Research and extension efforts in recent decades have focused on various methods to improve N fertilizer management, especially the 4 R’s- the right rate, in the right form, in the right place, at the right time. However, between 2003 and 2011, there was an overall decline in the acreage receiving appropriate N fertilizer management in the Chesapeake Bay watershed and a 9% increase in the commercial N fertilizer application rate (5).
Concurrent with research on improved N fertilizer management, precision agriculture technology rapidly evolved. Precision agriculture uses the ability for global position satellite (GPS) receivers to interface with machinery controls and crop and soil sensors to record data and vary management practices, such as fertilizer application rates, within a field. While improved N fertilizer management was an early promise of precision agriculture, widespread success and implementation of precision N fertilizer management in corn has been elusive (6). One reason for the lack of success and adoption of these fertilizer N management practices is that N fertility recommendations that take into account the ecological factors controlling N cycling are not well developed.
Despite current limitations in N fertility recommendations, the advent of the soil health movement in mainstream agriculture has led to increasing farmer interest in using cover crops to recycle N and build soil organic matter (7). Along with this interest has come the need to credit the N supply potential from cover crops and soil organic matter when calculating fertilizer N application rates. Recently, methods to predict N supply from cover crop residues and soil organic matter have been developed from research experiments across Pennsylvania (2, 3). There is potential for these methods to be paired with precision agriculture technologies to develop field maps of biological N supply that could inform variable rate N fertilizer applications.
Our goal is to decrease N pollution in the environment and increase farmer profitability by improving NUE in corn production. Our research question is whether the spatially explicit crediting of biological N supply from cover crops and soil organic matter when calculating a variable rate N fertilizer prescription can improve fertilizer NUE.
The first objective is to use precision agriculture technologies to map cover crop biomass N content and delineate soil sampling zones for the measurement of soil organic matter and soil texture. The second objective is to use the maps of cover crop biomass N, soil organic matter, and soil texture to calculate spatially explicit biological N supply credits. The third objective is to use the map of biological N supply credits to calculate variable rate N fertilizer prescriptions and measure whether NUE improves and overall corn yields can be maintained by using a variable rate application compared to the standard practice of a fixed N rate.
Predicting N mineralization from soil organic matter and cover crop residues has been a long standing topic in soil science research and is a crucial component to improving fertilizer NUE (8). There are many soil tests that have been proposed as indices of the N mineralization potential of soil organic matter and our understanding of the controls on N mineralization from cover crop residues is strong (9). However, our ability to predict relevant agronomic outcomes, such as corn yield response, based on N mineralization from soil organic matter and cover crop residues is either weak or has not been tested under real environmental conditions. To overcome the gap between laboratory studies and field studies, we recently used a large dataset of corn yield responses from research station and on-farm experiments across Pennsylvania to develop and calibrate equations to predict N supply from soil organic matter and cover crops (2, 3).
These equations are designed to predict the unfertilized yield of a corn crop based on easily measurable variables, specifically total soil carbon (C) concentration (a measure of organic matter), soil texture, cover crop biomass N content and cover crop biomass C:N ratio (see supplemental attachment). The equations are formulated on well understood processes of N mineralization and the most recent advances in scientific understanding of soil C dynamics, with coefficients calibrated to capture regionally-specific environmental controls on decomposition and N losses. The equation to predict N supply from soil organic matter reflects the relatively new understanding that soil organic matter is stabilized not by its chemical recalcitrance, but by sorption to the surfaces of mineral particles (10). Thus, soil organic matter decomposition and N mineralization is adjusted in the equation based on soil texture, with silt and clay reducing N supply because of their increased surface area relative to sand.
Because the variables needed to predict N supply from cover crops and soil organic matter in these equations are easily measured, but could vary across the scale of a field, there is an opportunity to use existing precision agriculture technologies to create field maps of biological N supply potential. Soil ECa sensors (e.g., Veris model 3100) are currently used by fertilizer suppliers in Pennsylvania to map fields and delineate soil sampling zones that correspond to differences in soil chemical and physical characteristics, including soil texture. Crop canopy sensors measuring normalized difference vegetation index (NDVI) and logging to a GPS receiver are commercially available, and we have previously calibrated these sensors to predict cover crop biomass N content (11). Some fertilizer companies in Pennsylvania already have the equipment to offer variable rate fertilizer application services and could use N fertilizer prescriptions that include spatially explicit N credits from soil organic matter and cover crops.
- K. G. Cassman, A. Dobermann, D. T. Walters, Agroecosystems, nitrogen-use efficiency, and nitrogen management. Ambio. 31, 132–140 (2002).
- C. M. White, D. M. Finney, A. R. Kemanian, J. P. Kaye, A Model–Data Fusion Approach for Predicting Cover Crop Nitrogen Supply to Corn. Agron. J. 0, 0 (2016).
- C. White, A. Kemanian, J. Kaye, in Annual Meeting, Soil Science Society of America: Phoenix, AZ. 9 November 2016. (2016).
- G. P. Robertson, P. M. Vitousek, Nitrogen in Agriculture: Balancing the Cost of an Essential Resource. Annu. Rev. Environ. Resour. 34, 97–125 (2009).
- NRCS., Impacts of Conservation Adoption on Cultivated Acres of Cropland in the Chesapeake Bay Region, 2003-06 to 2011. United States Dep. Agric. Nat. Resour. Conserv. Serv. Washington, DC (2013).
- T. A. Doerge, Variable-Rate Nitrogen Management for Corn Production- Success Proves Elusive. Int. Plant Nutr. Inst. Site Specif. Manag. Guidel. SSMG–36.
- S. Groff, Mixtures and cocktails: Soil is meant to be covered. J. Soil Water Conserv. 63, 110A–111A (2008).
- A. J. Franzluebbers, Should Soil Testing Services Measure Soil Biological Activity? Agric. Environ. Lett. 1, 0 (2016).
- H. H. Schomberg et al., Assessing Indices for Predicting Potential Nitrogen Mineralization in Soils under Different Management Systems. Soil Sci. Soc. Am. J. 73, 1575 (2009).
- M. W. I. Schmidt et al., Persistence of soil organic matter as an ecosystem property. Nature. 478, 49–56 (2011).
- C. M. White, Using Cover Crop Mixtures to Reduce Nitrate Leaching and Supply Nitrogen to Corn: Models to Inform Adaptive Management (2015).
- C. M. White, A. R. Kemanian, J. P. Kaye, Implications of carbon saturation model structures for simulated nitrogen mineralization dynamics. Biogeosciences. 11, 6725–6738 (2014).
Two variable rate nitrogen experiments were established, one in a Penn State Farm Services field in State College, PA (hereafter referred to as the State College location) and another in a field on Ed Quigley’s dairy farm in Spruce Creek, PA (hereafter referred to as the Spruce Creek location). At the State College location, a triticale cover crop was planted after corn silage harvest. Corn silage was harvested on 26 September 2017, a burndown herbicide was applied to control weeds on 1 October 2017, and the Hy Octane triticale variety was drilled on 7.5 inch row spacing with a seeding rate of 90 lbs/acre on 5 October 2017. At the Spruce Creek location, cereal rye was planted after corn silage harvest in September 2017.
Both fields were mapped with an electric conductivity (EC) sensor (Veris 3100) using a 50′ pass width and a driving speed of ~ 10 miles/hr. The sensor logged georeferenced EC measurements for the 0-1′ and 0-3′ depth ranges at a frequency of 1/sec. An EC map of each field was created by interpolated EC measurements using kriging in SAS 9.4 (Figures 1 and 2).
Soil samples were collected from 24 points in the State College field on 17 November 2017 and from 24 points in the Spruce Creek field on 26 June 2018. A stratified random sampling approach was used to determine sampling points that were representative of the range of mapped EC values. Each soil sampling point was georeferenced (< 2 meter GPS accuracy) and a composite sample of 8 cores (3/4″ diameter, 8 inch deep) was collected from random spots within a 5′ radius of the georeferenced point. Soil samples were air dried, ground in a flail mill, sieved to < 2 mm, and sent to the Penn State Agricultural Analytical Services Lab to measure nutrients, soil organic matter and soil texture.
In the spring, cover crops were terminated with herbicides on 8 May 2018 at the State College location and 3 May 2018 at the Spruce Creek location. During the cover crop burndown, two Greenseeker sensors (RT200 sensor package, Trimble, Inc., Sunnyvale, CA) were mounted on the spray boom to map cover crop NDVI throughout the field. Measurements were mapped using the Farmworks Mobile software installed on a tablet PC. Measurements were logged with georeferenced coordinates every 10 feet of forward motion of the sprayer. Sprayer passes were made on 90 foot widths at the State College location and 60 foot widths at the Spruce Creek location. Prior to cover crop burndown, cover crop biomass samples were collected from eight 0.25 square meter quadrats evenly distributed across each field. Cover crop biomass was weighed, ground, and analyzed for carbon and nitrogen contents.
Maps of cover crop NDVI were interpolated with kriging in SAS 9.4 and NDVI values were converted to cover crop biomass N content using a previously established calibration equation (Figures 5, 6, 7).
At each site, maps of cover crop biomass N content and soil EC were overlaid and clusters with similar biomass N content and soil EC were identified to be used for calculation of variable rate N fertilizer prescriptions that credited N availability from cover crops and soil organic matter. A Cover crop and soil organic matter N credits were calculated within each cluster using measurements of soil texture and soil organic matter from the 24 soil sampling points, cover crop biomass N content from the Greenseeker NDVI maps, and cover crop C:N ratio from the 8 cover crop biomass sampling points. The N credits were calculated using the equations presented in Figure 8. Equation 1 first calculates a microbial carbon use efficiency based on the soil texture. As clay content increases or sand content decreases, microbial carbon use efficiency, or the fraction of decomposed carbon that is assimilated into microbial biomass, increases. An increasing carbon use efficiency will ultimately translate into less N mineralized from the decomposition of cover crop residues and soil organic matter. Equations 2 and 3 calculates the credits from soil organic matter and cover crop residue decomposition. Note that the carbon use efficiency calculated in Eq. 1 is used in both Eq. 2 and Eq. 3. The N credits calculated in Eq. 2 and Eq. 3 are expressed as levels of corn yield that can be achieved without supplemental N fertilizer applied. In Eq. 2, as soil organic matter increases, the corn yield increases. In Eq. 3, as cover crop biomass N content increases and as C:N ratio decreases, the corn yield increases. Note that if the C:N ratio increases above a critical threshold (defined as 10/ε), the N credit from cover crops will become negative due to immobilization of N. In Eq. 4, the N credits from soil organic matter and cover crops are added together and then adjusted to account for a quadratic yield response to N availability. The result of Eq. 4 predicts the corn yield that could be supported by N mineralization of cover crop residues and soil organic matter without the addition of any supplemental N fertilizer. Equation 5 determines the required amount of supplemental N fertilizer by taking the difference between a fully fertilizer yield goal for the site and the yield that can be achieved without N fertilizer, multiplying that difference by a factor of 1.2 lbs N required per bushel of yield, and dividing by a fertilizer efficiency (efficiency value between 0 and 1). The factor 1.2 lbs N/bushel is the quantity of N physiologically required by the corn plant to produce the biomass needed for a bushel of grain. It is known that N fertilizer is not always recovered by a crop with 100% efficiency, therefore the physiological N requirement is modified by an expected fertilizer efficiency value to determine the rate of N fertilizer to apply to meet the physiological N requirement. In our experiment, we tested N fertilizer rates developed assuming 100% efficiency, 75% efficiency, and 50% efficiency of N fertilizer recovery to determine what efficiency level should be used when crediting N availability from soil organic matter and cover crops as we do in this process. In the experiments, we also included a standard N fertilizer recommendation based on the Penn State Agronomy Guide, which is 1 lb N fertilizer per bushel of expected yield, minus any N credits from residual manure history or previous legume crops. Neither location had an N credit from a previous legume crop, but the State College site had a 20 lbs N/ac credit from a frequent manure history (2-3 out of 5 years) and the Spruce Creek location had a 35 lbs N/ac credit from a continuous manure history (4-5 out of 5 years). The target corn yield used in the development of N fertilizer rates was 165 bu/ac at the State College site and 250 bu/ac at the Spruce Creek site.
Variable rate N fertilizer prescriptions were developed for each zone as described above. At each site, a fixed amount of N fertilizer was applied prior to corn planting based on existing farm management practices and the remaining N fertilizer needed according to the prescription was applied as a variable rate sidedress application of 30% urea ammonium nitrate (UAN) solution. At the State College location, 65 lbs N/ac was applied as UAN solution used for the herbicide carrier in the burndown application (19.5 gal/ac UAN solution). At the Spruce Creek location, 50 lbs N/ac as urea was spun onto the cover crop on 3 April 2018 and 65 lbs N/ac was applied as UAN solution used for the herbicide carrier in the burndown application (19.5 gal/ac UAN solution). Because of relatively sparse cover crop growth at the Spruce Creek site and the limited amount of N uptake in the biomass, we credited the 50 lbs N/ac of urea as fertilizer applied to the corn rather than as N recycled through cover crop residues and credited in Eq. 3. The sidedress application used to apply the remaining N requirement was applied with a 30 foot wide boom sprayer equipped with TeeJet flexible orifice nozzles and drop tubes spaced on 60″ centers, dribbling a band of UAN solution between every other corn row. In both locations, fertilizer treatment plots (consisting of the 100%, 75%, and 50% fertilizer efficiency comparisons and an Agronomy Guide rate) were applied in a randomized block design with treatment strips running the length of the field. In each treatment strip, the variable rate prescription for the appropriate treatment (100%, 75%, or 50% efficiency level) was loaded into Farmworks Mobile. Based on the location of the sprayer in the field, Farmworks Mobile sent the N fertilizer rate for that prescription zone to a Raven 450 sprayer control console, which adjusted the control valve on the sprayer to apply the specific rate of UAN solution. For the fixed rates of N used in the Agronomy Guide treatment, the sprayer controller was manually set to the necessary rate.
To measure N status in the corn during the growing season, earleaf samples were collected between tasseling and silking growth stages from 25 plants in each treatment plot on 31 July 2018 at State College and 1 August 2018 at Spruce Creek and measured for N concentration. The late season corn stalk nitrate test was used at the end of the growing season to measure N status of the corn. At approximately the 1/2 milkline growth stage (7 September 2018 and 11 September 2018 at State College and Spruce Cree locations respectively), 15 plants in each treatment strip were sampled by cutting 8″ sections of stalk from 6″ to 14″ above the ground. Samples were further cut into 2″ segments, dried, ground and measured for nitrate concentration. To determine whether N fertilizer rates that credited soil organic matter and cover crops could achieve similar yields as the Agronomy Guide recommendation, corn yields were measured at the end of the season. At the State College site, corn grain was harvested with a 12 row combine and at Spruce Creek corn silage was harvested with an 8 row silage chopper. The mass of either grain or silage harvested from each treatment plot was measured on trucks driven over pad scales. Moisture content was determined on subsamples of grain and silage and reported yields are adjusted to 15.5% moisture for grain and 65% moisture for silage.
Education & Outreach Activities and Participation Summary
This methods used in this project were the topic of a 1 hr field day presentation at the Penn State Extension Agronomic Diagnostic Clinic (July 18 and 19, 2018), attended by 75 ag service providers, including extension educators, NRCS, and private consultants.
Results from the project were presented at the Pennsylvania 4R Alliance Fall Field Day (October 4, 2018) in Hershey, PA, attended by 60 ag service providers and 20 farmers.
Results from the project were presented as a Water Insights Seminar on the campus of Penn State (October 9, 2018), organized by the Environmental and Natural Resources Institute at Penn State, attended by 15 faculty members, graduate students, and professional staff.
Results from the project were presented at the Northeast Soil Testing Workgroup Meeting in Milford, PA (October 11, 2018), attended by 15 researchers and extension specialists from Land Grant Universities in the Northeast.
Results from the project were presented at the New York Certified Crop Advisors Conference in Syracuse, NY (November 27, 2018) attended by 70 ag service providers.
Results from the project were disseminated via a poster presentation at the Agronomy Society Annual Meeting in Baltimore, MD and the Northeast Cover Crop Council Annual Meeting in State College, PA (both November 2018)
One additional outcome of the grant was that a graduate student was able to be recruited with funds from the Penn State Plant Science department to conduct the on-farm work of the project. In addition to the on-farm work, he was able to replicate the study at the research station. The Greenseeker equipment purchased with Northeast SARE funds was used for mapping the cover crop NDVI at the research station but other expenses (time and analytical costs) were paid for separately.
Results from this project were also used to develop a multi-state multidisciplinary grant proposal to the USDA Sustainable Agricultural Systems initiative ($10 million proposal) focused on nutrient use efficiency in the Chesapeake Bay. Collaborations in Maryland and Delaware have also been fostered based on the concepts of this NE-SARE project, for which a research proposal to the 4R Research Fund has been developed (PI, Charlie White, Co-PIs Gurpall Toor and Amy Shober). Finally, a consultant in New York who learned about the project at the NY CCA Conference is planning to coordinate with several farmers in New York to implement on-farm experiments based on the design of this project to test the practices in the New York climate and soils.