A microplot design for 15N tracer research is described to accommodate multiple in-season plant and soil sampling events. Soil and plant sample collection and processing procedures, including grinding and weighing protocols, for 15N analysis are put forth.
Many nitrogen fertilizer studies evaluate the overall effect of a treatment on end-of-season measurements such as grain yield or cumulative N losses. A stable isotope approach is necessary to follow and quantify the fate of fertilizer derived N (FDN) through the soil-crop system. The purpose of this paper is to describe a small-plot research design utilizing non-confined 15N enriched microplots for multiple soil and plant sampling events over two growing seasons and provide sample collection, handling, and processing protocols for total 15N analysis. The methods were demonstrated using a replicated study from south-central Minnesota planted to corn (Zea mays L.). Each treatment consisted of six corn rows (76 cm row-spacing) 15.2 m long with a microplot (2.4 m by 3.8 m) embedded at one end. Fertilizer-grade urea was applied at 135 kg N∙ha-1 at planting, while the microplot received urea enriched to 5 atom % 15N. Soil and plant samples were taken several times throughout the growing season, taking care to minimize cross-contamination by using separate tools and physically separating unenriched and enriched samples during all procedures. Soil and plant samples were dried, ground to pass through a 2 mm screen, and then ground to a flour-like consistency using a roller jar mill. Tracer studies require additional planning, sample processing time and manual labor, and incur higher costs for 15N enriched materials and sample analysis than traditional N studies. However, using the mass balance approach, tracer studies with multiple in-season sampling events allow the researcher to estimate FDN distribution through the soil-crop system and estimate unaccounted-for FDN from the system.
Fertilizer nitrogen (N) use is essential in agriculture to meet the food, fiber, feed, and fuel demands of a growing global population, but N losses from agricultural fields can negatively impact environmental quality. Because N undergoes many transformations in the soil-crop system, a better understanding of N cycling, crop utilization, and the overall fate of fertilizer N are necessary to improve management practices that promote N use efficiency and minimize environmental losses. Traditional N fertilizer studies primarily focus on the effect of a treatment on end-of-season measurements such as crop yield, crop N uptake relative to the N rate applied (apparent fertilizer use efficiency), and residual soil N. While these studies quantify the overall system N inputs, outputs, and efficiencies, they cannot identify nor quantify N in the soil-crop system derived from fertilizer sources or the soil. A different approach using stable isotopes must be used to track and quantify the fate of fertilizer derived N (FDN) in the soil-crop system.
Nitrogen has two stable isotopes, 14N and 15N, that occur in nature at a relatively constant ratio of 272:1 for 14N/15N1 (concentration of 0.366 atom % 15N or 3600 ppm 15N2,3). The addition of 15N enriched fertilizer increases the total 15N content of the soil system. As 15N enriched fertilizer mixes with unenriched soil N, the measured change of 14N/15N ratio allows researchers to trace FDN in the soil profile and into the crop3,4. A mass balance can be calculated by measuring the total amount of 15N tracer in the system and each of its parts2. Because 15N enriched fertilizers are significantly more expensive than conventional fertilizers, 15N enriched microplots are often embedded within the treatment plots. The purpose of this methods paper is to describe a small-plot research design utilizing microplots for multiple in-season soil and plant sampling events for corn (Zea mays L.) and to present protocols for preparing plant and soil samples for total 15N analysis. These results can then be used to estimate N fertilizer use efficiency and create a partial N budget accounting for FDN in the bulk soil and the crop.
1. Field site description
NOTE: When performing 15N tracer field trials, selected sites should minimize variation due to soil, topography, and physical features5. Cross-contamination may occur following lateral soil movement due to slope, wind or water translocation, or tillage while the vertical distribution of soil N may be impacted by subsurface water flow and tile-drainage6.
2. Plot design
3. Soil and plant sample precautions
4. 15N enriched fertilizer preparation and application
5. Field sample processing: aboveground corn biomass
6. Field sample processing: soil
7. Lab sample processing: grind soil and plant samples
8. Weigh ground plant and soil samples for total N and 15N analysis
9. Calculations
The results presented in this paper come from a field site established in 2015 at the University of Minnesota Southern Outreach and Research Center located near Waseca, MN. The site was managed as a corn-soybean [Glycine max (L.) Merr] rotation prior to 2015 but was managed as a corn-corn rotation during the 2015 and 2016 growing seasons. The soil was a Nicollet clay loam (fine-loamy, mixed, superactive, mesic Aquic Hapludolls)-Webster clay loam (fine-loamy, mixed, superactive, mesic Typic Endoaquolls) complex. Soil fertility was managed according to university guidelines except for N18. Several N fertilizer treatments were arranged in a randomized complete block design with four replications but only the 135 kg N∙ha-1 rate applied as urea at planting is presented in this paper. Soil bulk density was measured at the center of 0- to 15-, 15- to 30-, 30- to 60-, 60- to 90-, and 60- to 120-cm depth layers from two 5-cm deep samples per replication using the intact core method19. Bulk density was averaged within depth across replications and assumed to be constant across the field. Plot setup and plant and soil samples were collected and processed as described in the protocol section.
Total (FDN + SDN) aboveground biomass N increased with each successive sampling event over the first growing season (Figure 4). Fertilizer derived N concentration was greatest earlier in the growing season accounting for 44 ± 4% (mean ± standard error) of the total aboveground biomass N at V8 and decreased with each successive sampling period (Figure 4A). However, SDN consistently was the greatest fraction of aboveground biomass N illustrating the importance of soil N supply for optimal corn growth. At physiological maturity in the first year, 27 ± 1% of aboveground biomass N was from FDN with similar proportions in grain, stover, and cob fractions (Figure 4B). At physiological maturity in the second year, only 2 ± 0.1% of first-year FDN was recovered in the aboveground biomass with 1.6 ± 0.2 kg of first-year FDN ha-1 exported in the grain (Figure 4A).
The soil-crop FDN budget is useful for quantifying FDN cycling within the system over time. Within 8 d of fertilizer application, the majority of FDN was in the top 15 cm of the soil profile, as expected (Figure 5). However, 22.2 ± 4.4 kg N ha-1 had already moved into the deeper depths while 4 ± 10% of the FDN was unaccounted for. Unaccounted-for FDN is likely primarily driven by N loss mechanisms including leaching, denitrification, and volatilization that either move FDN below the soil sampling depths or remove the FDN from the system entirely. At V8 and R1, unaccounted-for FDN increased to 60.4 ± 4.7 kg N ha-1 on average while soil N (0-15 cm) was 31.6 ± 6.8 kg N ha-1 on average. Corn's rapid growth and high N demand from V8 to R1 resulted in an increase of 19.0 ± 4.4 kg FDN ha-1 in aboveground plant biomass mirroring the 17.7 ± 5.2 kg FDN ha-1 reduction from the 15- to 60-cm soil depths. Soil temperature and moisture conditions between these corn development stages tend to favor microbial growth resulting in rapid turnover of organic residues and re-utilization of mineralized N. These results suggest that corn roots mined inorganic FDN from the 15- to 60-cm depths while FDN in the 0- to 15-cm depth was primarily cycled between soil organic matter and microbial fractions. Additional isotopic analysis of soil inorganic and organic N pools is necessary to validate this hypothesis and provide greater detail and insight into FDN cycling dynamics10. By post-harvest year 1, 59 ± 2% of the original FDN was unaccounted for while 18.1 ± 3.9 kg FDN ha-1 was in the top 30 cm of the soil (Figure 5) and 22.1 ± 2.3 kg FDN ha-1 was exported in the grain (Figure 4B). Fertilizer 15N use efficiency was 24% (Equation 7) and is at the low end of commonly reported F15NUE measures (25-45%) reported by other studies20. Although equipment was thoroughly cleaned between each sample, the lower F15NUE measures of the study could be an artifact of enriched sample dilution by processing enriched samples in order of lowest to highest expected enrichment. The amount of FDN in the top 30 cm doubled (36.0 ± 5.2 kg FDN ha-1) from post-harvest year 1 to pre-plant year 2 due to partial residue breakdown since the previous fall but by post-harvest year 2 only 17.3 ± 3.3 kg FDN ha-1 was still found within the soil-corn system (Figure 5). This study indicates that by the end of the first and second years, only 41 and 29%, respectively of first-year FDN was accounted for within the soil-corn system (including FDN exported in the grain) while the remainder was either lost to the environment or leached below the 90 cm soil sampling depth.
Spurious results may be obtained when samples are cross-contaminated affecting calculations of Nf, FDN, and SDN. For example, suppose a 15N enriched plant sample with an actual enrichment of 3.000 atom % 15N is contaminated with unenriched material diluting the 15N concentration to 2.500 atom % 15N. Further, assume Total NPlant is 100 kg N ha-1, the atom % 15N enrichment of the fertilizer was 5.000, and the atom % 15N enrichment of the unenriched plant sample was 0.366. The 15N enriched plant sample Nf would be reduced from 0.568 (actual) to 0.461 (contaminated sample) underestimating the true FDN by 10.7 kg N ha-1. Overestimations of FDN may occur when samples with low 15N enrichment are contaminated with additional 15N. Thus, extreme care should be taken in all steps of sample collection and processing to minimize sample contamination, but most especially when sample masses are reduced (e.g., grinding and weighing procedures).
Figure 1: Plot design for the treatment plot and microplot. The figure illustrates the dimensions and relative placements of the border areas, unenriched sampling area, harvest area, and microplot area within the treatment plot. Please click here to view a larger version of this figure.
Figure 2: Microplot plant and soil sampling diagram. The figure illustrates the relative plant and soil sampling positions at each sampling stage that avoids altering corn N uptake patterns of later sampled corn plants. Sampling events occurred 8 days following the 15N enriched fertilizer application, at the V8 and R1 corn physiological development stages, at physiological maturity in the year of 15N enriched fertilizer application (PMY1) and the following year (PMY2), and prior to planting the second year (PPY2). Please click here to view a larger version of this figure.
Figure 3: Chronological depiction of microplot management. (A) Dissolve 15N enriched urea into 2 L of deionized water and spray onto the microplot at planting. (B) Collect and chop a six-aboveground corn plant composite sample from within the sampling area (15N unenriched) and a six-aboveground corn plant composite sample from the 15N enriched microplot at the pre-determined sampling times. (C) Following sample collection at physiological maturity, remove all remaining aboveground biomass from within the microplot. (D) Post-harvest, rake unenriched aboveground corn biomass from the microplot area. Chip and reapply the microplot corn aboveground biomass to the microplot area. Please click here to view a larger version of this figure.
Figure 4: Example of aboveground biomass N partitioned into fertilizer derived N (FDN) and soil derived N (SDN) fractions. Total aboveground biomass N was separated into its individual sources of FDN (solid color) and SDN (hashed color) in (A) and (B). Error bars represent the standard error of the mean. (A) Aboveground biomass N was measured at the V8 and R1 corn physiological development stages and at physiological maturity in the year of 15N fertilizer application (PMY1) and the year following 15N fertilizer application (PMY2). The value above each column represents the percentage of the total N that was FDN. (B) Aboveground biomass N measured at PMY1 and PMY2 is shown in its individual parts of cob (only Year 1), stover (stalk and leaves; includes cob for PMY2), and grain for FDN and SDN. Please click here to view a larger version of this figure.
Figure 5: Example of the soil-corn fertilizer derived N (FDN) budget. The mass of FDN recovered in aboveground (Abvgd) corn biomass and at various soil sampling depths is reported for six sampling events over two growing seasons. Sampling events occurred 8 days following the 15N enriched fertilizer application (PA), at the V8 and R1 corn physiological development stages, at physiological maturity in the year of 15N enriched fertilizer application (PMY1) and the following year (PMY2), and prior to planting the second year (PPY2). The difference between the applied fertilizer rate (135 kg N ha-1) and the mass of FDN recovered in the soil-corn portions is the unaccounted for FDN fraction. The total mass of FDN for PPY2 and PMY2 was 113 kg FDN ha-1 because 22 kg FDN ha-1 was exported out of the soil-corn system as first-year grain. Error bars represent the standard error of the mean. Please click here to view a larger version of this figure.
Stable isotope research is a useful tool for tracking and quantifying FDN through the soil-crop system. However, there are three main assumptions associated with N tracer studies that if violated may invalidate conclusions drawn from using this methodology. They are 1) the tracer is uniformly distributed throughout the system, 2) processes under the study occur at the same rates, and 3) N leaving the 15N enriched pool does not return3. Because this study is interested in the distribution of total FDN throughout the soil-crop system, assumptions 2 and 3 are of minimal concern21.
The high cost of 15N enriched material generally limits the size of 15N tracer studies. Therefore, prior to initiating a N tracer study, the researcher should carefully plan the research project's objectives considering: the number of sampling events, the length of the study (days to years), the N fertilizer application rate, and the 15N enrichment concentration required to measure differences from natural abundance (0.366 atom %) following 15N enriched fertilizer dilution by bulk soil2. Commonly used 15N enrichment levels and application rates are reported for different types of agronomic research in Ref. 2. After determining the study objectives, the microplot must be sufficiently large to accommodate soil and plant sampling and avoid edge effects. The plot design described in this protocol uses a non-confined plot requiring that non-sampled border areas be employed6. The 15N concentration in border areas is diluted by mass flow across the microplot boundary and N uptake from outside the microplot by lateral corn roots growing in rows 1 and 6. Confined plots, where physical barriers are driven into the soil, do not require border areas but do require additional work during microplot establishment and may limit routine field operations6. References 3, 6, 22-25 provide additional guidance on selecting microplot sizes, border widths, and when confined or non-confined plots may be most appropriate.
The plant and soil sampling scheme of this study is designed to allow for multiple sampling events over two consecutive growing seasons. Early season plant and soil samples are taken near the outside edges of the microplot. Each successive sampling event moves closer to the center of the microplot to avoid sampling previously sampled areas. At least two corn plants separate each sampled plant to minimize changes in corn physiological development. One challenge with this study's soil sampling technique is that the soil core sampling method may not accurately intercept the heterogeneous distribution of 15N in the soil profile3. Spatial variability of soil total N is high with an estimated coefficient of variation of 15%3. Complete microplot excavation would improve 15N quantification accuracy but requires processing significant volumes of soil and limits sampling to a single event3 that is not in line with the objectives of this study. Subdividing the microplot into smaller sampling units allows for multiple excavation events but may increase the required microplot size to ensure non-sampled units are unaffected by modifications to the crop canopy and soil water dynamics. Despite the potential reduction in accuracy, many studies use the soil core technique for microplots ≥1 m2 9,22,26,27,28. Sample precision may be increased by increasing the number of soil cores collected and composited per microplot using the following formula13:
n = (Z2)(CV2)/(d2)
where n is the number of soil cores, Z is the standardized normal variate for the corresponding alpha level (1.96 for 0.05 and 1.65 for 0.10), CV is the coefficient of variation, and d is the margin of error in the plot mean (as a decimal). Based on this formula, the authors expect that 15 cores per microplot would estimate total N to ±7.6% on 95% of the plots (n = 15; Z = 1.96; CV = 15%; d = 0.076). Reference 25 used a similar number of cores but subdivided the microplot into 32 sampling units collecting plant and soil samples from four units at each sampling event.
Others have shown that the microplot data can be extrapolated to the entire plot29. However, for this assumption to be valid, the treatment plot and microplot must be similarly managed. If possible, fertilizer N should be applied in the same chemical and physical forms (e.g., urea dissolved in water) as these properties impact fertilizer-soil dynamics including N loss mechanisms, immobilization, and availability to soil microbes and plants3.
The roller jar grinding method described in this protocol is capable of pulverizing large volumes of plant and soil samples, ideal for ensuring a representative, homogenized sample. However, the technique requires significant manual labor and time to load, unload, roll, and clean the roller jars. Sample processing is limited by the available number of roller jars, the capacity of the conveyor belt unit, and the size of the acid bath. Commercial grinding vials may be an alternative to roller jars but may limit the volume of plant and soil samples processed. Lab-made, single-use grinding vials may be constructed that potentially serve as both the grinding and sample storage vessel. The main consideration of any of these grinding methods is to minimize cross-contamination between samples.
Finally, because 15N enriched fertilizer material is expensive, 15N enriched aboveground biomass and soil samples may be retained and homogenized for use in future studies. These products may be especially useful when investigating residue decomposition, mineralization potential, or other nutrient cycling processes21.
The authors have nothing to disclose.
The authors acknowledge the support of the Minnesota Corn Research & Promotion Council, the Hueg-Harrison Fellowship, and the Minnesota's Discovery, Research and InnoVation Economy (MnDRIVE) Fellowship.
20 mL scintillation vial | ANY; Fisher Scientific is one example | 0334172C | |
250 mL borosilicate glass bottle | QORPAK | 264047 | |
48-well plate | EA Consumables | E2063 | |
96-well plate | EA Consumables | E2079 | |
Cloth parts bag (30×50 cm) | ANY | NA | For corn ears |
CO2 Backpack Sprayer | ANY; Bellspray Inc is one example | Model T | |
Coin envelop (6.4×10.8 cm) | ANY; ULINE is one example | S-6285 | For 2-mm ground plant samples |
Corn chipper | ANY; DR Chipper Shredder is one example | SKU:CS23030BMN0 | For chipping corn biomass |
Corn seed | ANY | NA | Hybrid appropriate to the region |
Disposable shoe cover | ANY; Boardwalk is one example | BWK00031L | |
Ethanol 200 Proof | ANY; Decon Laboratories Inc. is one example | 2701TP | |
Fabric bags with drawstring (90×60 cm) | ANY | NA | For plant sample collection |
Fertilizer Urea (46-0-0) | ANY | NA | ~0.366 atom % 15N |
Hand rake | ANY; Fastenal Company is one example | 5098-63-107 | |
Hand sickle | ANY; Home Depot is one example | NJP150 | For plant sample collection |
Hand-held soil probe | ANY; AMS is one example | 401.01 | |
Hydraulic soil probe | ANY; Giddings is one example | GSPS | |
Hydrochloric acid, 12N | Ricca Chemical | R37800001A | |
Jar mill | ANY; Cole-Parmer is one example | SI-04172-50 | |
Laboratory Mill | Perten | 3610 | For grinding grain |
Microbalance accurate to four decimal places | ANY; Mettler Toledo is one example | XPR2 | |
N95 Particulate Filtering Facepiece Respirator | ANY, ULINE is one example | S-9632 | |
Neoprene or butyl rubber gloves | ANY | NA | For working in HCl acid bath |
Paper hardware bags (13.3×8.7×27.8 cm) | ANY; ULINE is one example | S-8530 | For soil samples and corn grain |
Plant grinder | ANY; Thomas Wiley Model 4 Mill is one example | 1188Y47-TS | For grinding chipped corn biomass to 2-mm particles |
Plastic tags | ULINE | S-5544Y-PW | For labeling fabric bags and microplot stalk bundles |
Sodium hydroxide pellets, ACS | Spectrum Chemical | SPCM-S1295-07 | |
Soil grinder | ANY; AGVISE stainless steel grinder with motor is one example | NA | For grinding soil to pass through a 2-mm sieve |
Tin capsule 5×9 mm | Costech Analytical Technologies Inc. | 041061 | |
Tin capsule 9×10 mm | Costech Analytical Technologies Inc. | 041073 | |
Urea (46-0-0) | MilliporeSigma | 490970 | 10 atom % 15N |