Urea and Urea Additives as Fertilizer Sources for Corn Production in Minnesota
Study author(s): Fabián Fernández, Karina Fabrizzi, Jeffrey Vetsch, and Daniel Kaiser, University of Minnesota, Department of Soil, Water, and Climate
Years of study: 2017 – 2019
Location(s): Crookston MN, Morris MN, Lamberton MN, Waseca MN
Important: for the complete report, including all tables and figures, please download using the links to the right.
brief overview
Urea is increasingly an important nitrogen (N) source in Minnesota. Approximately 43% of our farmers use urea as their major N source. In the southwestern, south-central and west-central areas approximately 45% of the N is applied in the fall, 50% is applied in the spring, and 5% is applied at sidedress. While most of those that use urea as the major N source apply it in the spring, approximately 4% do the major application with urea in the fall and there are others that apply some of their N as urea in the fall as this application is part of the listed Best Management Practices. However, in recent years, due in part to wet spring conditions, fall urea applications have resulted in yield reduction.
The objectives of this study were to: 1) evaluate fall and spring applications of urea to determine their feasibility and calculate the economic optimum N rate for fall and spring applications when corn follows corn (CC), corn follows wheat (CWh) and corn follows soybean (CSb), 2) Investigate the role of placement (band vs. broadcast and incorporated) and of a nitrification inhibitor and polymer coating (ESN) to improve management of urea.
materials and methods
The study was conducted at the following Research and Outreach Centers during the 2019 growing season: Northwest (Crookston), West Central (Morris), Southwest (Lamberton), and Southern (Waseca). Following is the soil information for each site. At Crookston experimental sites were established under a corn-soybean (CSb) (Gunclub silty clay loam, 0 to 2 percent slopes) and a corn-corn (CC) cropping system (Colvin-Perella silty clay loams, 0 to 1 percent slopes). Lamberton experimental sites were established under a CSb (Canisteo clay loam, 0 to 2 percent slopes, Amiret loam, 2 to 6 percent slopes, Glencoe clay loam, 0 to 1 percent slopes) and a CC cropping system (Amiret loam, 2 to 6 percent slopes, small portion of Normania loam,1 to 3 percent slopes). Morris experimental sites were established under a CSb (Aazdahl-Formdale-Balaton clay loams, 0 to 4 percent slopes) and for CC cropping system (Aazdahl-Formdale-Balaton clay loams, 0 to 4 percent slopes; McIntosh silt loam, 1 to 3 percent slopes;Tara silt loam, 1 to 3 percent slopes). Waseca had an experimental site under a CSb cropping system (Nicollet clay loam, 1 to 3 percent slopes, Canisteo-Glencoe complex, 0 to 2 percent slopes, and a small portion of Webster clay loam, 0 to 2 percent slopes).
Treatments are presented in Table 1. The treatments included a full set of N rates for fall and spring pre-plant (PP) applications of broadcast and shallow incorporation of urea (BI) to allow us to determine N response and calculate the economic optimum N rate (EORN). Rates ranged from 0 to 240 lb N/acre in CC and since the yield response is expected to maximize at a lower rate for CSb and CWh the rate ranged from 0 to 200 lb N/ac. There is also a comparison of different sources, placements, and timings: the standard practice that consist of anhydrous ammonia (AA) with N-Serve in the fall and without N-Serve in the spring, ESN broadcast and incorporated by tillage (BI), and urea with or without the nitrification inhibitor Instinct HL (24 oz/ac) either as BI or banded application (SSB) all as spring and fall applications. The sub-surface banded fertilizer was applied below the crop row position (except for urea and urea+Instinct treatments at Waseca that were applied between the crop rows). For the comparison treatments, we used a sub-optimal rate (most responsive portion of the response curve) to be able to more easily detect differences due to treatment. Treatments 15 to 24 were applied at 120 lb N/ac in CC and at 80 lb N/ac in CSb except for AA that was always applied at 120 lb N/ac because a lower rate was not achievable with the available equipment. The treatments were organized in a randomized complete block design replicated four times.
Plant dry biomass and N uptake were measured at V6, V12 and R6 development stages. Canopy sensing was performed with the Crop Circle and normalized difference red-edge (NDRE) index were calculated for the V6 and V12 development stages. At harvest grain yield was calculated and grain N content measured. After harvest, soil samples from the 0-12, 12-24, and 24-36-inch depth increments were collected and analyzed for ammonium-N and nitrate-N and total inorganic N (TIN) was calculated. Statistical analysis was performed using the SAS software and program. Differences were established at P=0.05.
preliminary results and discussion
Waseca was 11 inches above normal for the Apr-Oct period with only June having below normal precipitation (Table 2). In Lamberton, precipitation was 9.4 inches above the normal average from Apr-Oct. (Table 2). Similarly, Morris was 11.4 inches above the normal. Crookston was at or below the normal during April through July, but was wetter than normal in August, September, and October, which resulted in 4.6 inches above the normal for the Apr-Oct period. The growing season across all sites in Minnesota was challenging. Generally, wet and cool conditions delayed planting and the crop growth rate was relatively slow.
Grain yield and grain N removal
During the 2019 growing season all sites and crop rotations, except Crookston-CC rotation in the fall and Morris CSb in the spring, had significant grain yield response to N application rate, but the response varied with sites and time of application (Table 3). Spring applications produced significantly greater yield than fall applications only at Waseca (Table 3, 9). Averaged across all sites and rotations the spring application produced 146 bu/ac compared to 141 for the fall application. As mentioned earlier, 2019 was a challenging and overall disappointing growing season. The highest yield we obtained in this study was only 205 bu/ac. It is clear that factors besides N rate or time of application had a substantial impact in the grain yield responses we obtained. Grain N removal increased with N rate at all sites (Table 4). Spring applications produced significantly greater N removal than fall applications in Waseca CSb and Crookston CSb. (Table 4).
In Waseca, the CSb crop had a linear response to N (yield was not maximized) for the fall application and a quadratic-plateau response for the spring application (Table 5). For the fall, the yield at the highest N rate (200 lb N/ac) was 189 bu/a. For the spring application, the EONR was 161 lb N/ac and the yield at the EONR was 203 bu/a. Across years for CC fall applications required 45 lb N/ac more than the spring application to reach the EONR and the yield at the EONR was 12 bu/ac lower than the spring application (Table 5). Similarly, for CSb fall applications required 26 lb N/ac more than the spring application to reach the EONR and the yield at the EONR was 11 bu/ac lower than the spring application. The data for all the years of the study clearly show that fall urea in south central Minnesota should not be used because the potential for N loss is too great. This follows current University of Minnesota BMPs.
In Lamberton, the CC crop for fall and spring application had quadratic-plateau response to N. For the fall application the EONR was 195 lb N/ac and the yield at the EONR was 159 bu/a, and for spring application the EONR was 190 lb N/ac and the yield at the EONR was 140 bu/a (Table 6). For the CSb crop, a quadratic response to N was also observed for both fall and spring. For the fall application the EONR was 131 lb N/ac and the yield at the EONR was 186 bu/a, and for spring application the EONR was 132 lb N/ac and the yield at the EONR was 178bu/a (Table 6). Across years for CC while the EONR for fall application was very similar to the spring application, the yield at the EONR was on average 23 bu/ac more for the spring application relative to the fall application.
Similarly, for CSb the difference in EONR was only 3 lb N/ac lower with spring than fall application, but the yield at the EONR was 6 bu/ac greater with spring than fall application. This indicates that overall the spring application results in better nitrogen use efficiency than the fall application.
In Morris, the CC crop had a linear response to N (yield was not maximized) for fall application and a quadratic- plateau response for spring applications (Table 7). Grain yield was 190 bu/a for fall at the highest N rate (240 lb N/ac). For the spring application the EONR was176 lb N/ac and the yield at the EONR was 175 bu/a. The CSb crop had a quadratic- plateau response to N for fall and spring applications (Table 7). For the fall application the EONR was165 lb N/ac and the yield at the EONR was 170 bu/a. Grain yield for the spring application was 118 bu/a and the yield at the EONR was 162 bu/ac. Across years for CC the EONR with spring applications was 37 lb N/ac lower than for fall, and the spring application yielded 11 bu/ac above the fall application. Similarly, for CSb, across years the EONR was 36 lb N/ac lower than for fall applications, tough the grain yield at the EONR was similar (3 bu/ac lower for spring than fall application). Again, as observed for Lamberton, while the results are not as consistent as observed in Waseca, the southwest and west-central portion of Minnesota showed that fall applications are not as efficient as spring applications. Overall, fall applications required additional N to reach the EONR and the yield at the EONR was typically lower, or at best very similar, to the yields obtained with the spring applications.
In Crookston, the CC crop showed no significant response to N application both for fall and spring applications (Table 8). The CSb crop had quadratic response to N for spring application and no response to N for fall application (Table 8). For the spring application, the EONR was 189 lb N/ac and the yield at the EONR was 198 bu/ac. While there were several non-N response curves at Crookston (seven out of 12), the years where there was a response to N showed that fall applications required more N to achieve the EONR compared to the spring application to achieve a lower, or at best similar, yield at the EONR.
While studies were conducted in south-central Minnesota to provide a comparison, the major objective of this study was to evaluate whether fall urea application could still be considered an acceptable practice for wester Minnesota. The results from southcentral Minnesota consistently confirmed that fall application of urea is not acceptable for this region, which had been known and used as the guide for many years as a Nitrogen Best Management Practices (N-BMP) (Table 5, 9). For western Minnesota (Lamberton representing the southwest region, Morris representing the west-central region, and Crookston representing the northwest region) we observed less consistent results compared to southcentral Minnesota (Table 6-9). However, having 40 response curves over four distinct growing seasons across the western portion of the state allow us to describe on average what is most likely to happen when comparing fall to spring applications of urea. Averaged across all these 40 site-years and crop rotations (Table 6-8) the spring urea application needed 27 lb N/ac less nitrogen than the fall application to achieve the EONR and produced 9 bu/ac more yield than the fall application at the EONR. Based on these results and the need to improve nitrogen use efficiency for both economic and environmental reasons, fall urea application should not be considered a N-BMP for Minnesota. Though we feel confident on the results, at this point these conclusions can be considered preliminary. We will continue to examine the data in rigorous scientific ways, publish results in peer-reviewed journals, and adjust N-BMPs guidelines as needed.
The source of N showed difference in yield response only at Waseca for fall application, and Lamberton-CC and Crookston-CSb for spring applications (Table 10). Only these three situations will be discussed. At Waseca ESN and the treatments with sub-surface band (SSB) applications produced better yields than the urea broadcast and incorporated (BI) treatments, but there was no difference due to using the nitrification inhibitor (+I). At Lamberton, anhydrous ammonia (AA) outperformed both urea with inhibitor sub-surface band (U+I/SSB) and urea broadcast and incorporated (U/BI). Also, ESN was better than U+I/SSB. In Crookston ESN was better than broadcast and incorporated urea regardless of nitrification inhibitor usage (Table 10). In Morris, fall AA+I produced greater yield than urea BI (Table 11). In terms of Grain N removal, there were only a few situations where N source/placement made a difference (Table 12-13); only the significant differences will be discussed. Morris CC had greater yield with ESN than U+I/BI, U/BI, and U/SSB, also U+I/BI produced lower yields than the AA treatments (Table 12). In Crookston the CSb rotation had greater N removal with ESN than urea when both were BI.
The effect of urea placement across N inhibitors (with and without Instinct) for all the evaluated years shows that SSB of urea has the greatest potential to increase yields for fall applications (Table 14 and 16). That said, the benefit only occurred about half of the time (52% of the time), so the additional difficulty (time and expense) of applying urea as a SSB might not be justified. The data also shows that SSB applications cannot be used as a way to improve urea performance for fall applications. For those cases where SSB produced a benefit, the yield increase was 33 bu/ac compared to the BI application. For spring applications, the placement does not have a clear pattern.
The use of Instinct as a nitrification inhibitor regardless of placement (BI or SSB) is not justified for fall or spring applications (Table 15, 16). Using the inhibitor as a way to justify fall urea applications is not valid as it only improved yields 7% of the time.
Anhydrous Ammonia applied in the fall produced the greatest and most consistent grain yield benefit (Table 16). Compared to urea BI there was a 60% chance of improving yield. In those cases, the yield increase was 49 bu/ac on average. Compared to urea SSB in the fall the benefit of AA was still observed, but it was only for 30% of the time. This is likely, because as explained earlier, a SSB application in the fall tends to improve urea performance compared to urea BI. Compared to ESN, AA was better only 25% of the time during fall applications. For spring applications AA performed better than urea (regardless of urea application method) approximately 30% of the time. These results illustrate that urea in the fall and in wet springs has greater potential for N loss than anhydrous ammonia. Using ESN in the fall as a way to reduce concerns regarding N loss with urea is not justified. In the spring, the results were similar to those in the fall. However, from previous studies we have observed that in wet springs the use of ESN can be beneficial compared to urea. Additional analyses of the data in this study in conjunction with weather and soil information are needed and will be conducted in the future.
Summary
- Fall urea is problematic due to wet springs and warmer falls/winters. While this is a preliminary analysis, most likely current BMP’s for wester Minnesota will need to be adjusted to reflect changes in weather and cropping conditions as reflected in this study.
- All variables equal, spring applications produced more grain than fall applications.
- Anhydrous ammonia is superior to urea, especially in the fall.
- Banding urea (SSB) or using ESN in the fall had limited advantage.
- The use of the nitrification inhibitor Instinct with urea did not increase yield.
This last section of the report provides additional information collected during the study. At this stage in the data analysis, these values constitute ancillary information that will be included along with previous years of data to evaluate in greater depth the results presented in the earlier sections. At this point the data has been evaluated statistically, but little attempt has been made to relate these results to results presented in the earlier sections.
Canopy sensing and Plant N uptake
Canopy sensing measurements both at V6 and V12 development stage were linearly correlated to grain yield, but the relationship was low for V6 (R2=0.29) and V12 development stage (R2=0.26) (Fig. 1). These results are in contrast to previous years where we normally observe that in general, the relationship improves, as the plant becomes a better integrator of growing season conditions up to the time of sensing. Regardless, the data highlight the fact that it may be difficult to use sensing technologies alone to improve N management. The data also highlights the fact that the 2019 season was very challenging. It is possible that by V12 deficiencies due to N supply had not fully developed. Also, as mentioned earlier, during 2019 other stressors might have had a greater role than N treatment on the crop responses we measured.
In general, plant N uptake increased with N rate at all locations except Lamberton CSb where there were no differences (Appendix-1, 2). At V6, generally N uptake increased with N rate to 80 or 120 lb N/ac rate, but for total N at R6 uptake continued to increase with additional fertilizer N rates.
Plant N uptake was greater under spring application than fall application at all stages in Waseca CSb; at V6 at Lamberton CSb, at R6 in Crookston CC, V12 and R6 at Crookston CSb, and at V6 in Morris CC (Appendix-1, 2). No differences were observed between fall and spring applications across N rates in Lamberton CC and Morris CSb (Appendix-1, 2).
The N source and placement treatments produced small and inconsistent differences in plant N uptake across the various locations and crop rotations and for the different development stages (Appendix-3, 4, 5). At Lamberton CC and Crookston CC had no difference in plant N uptake at any of the development stages (Appendix-3). However, in Morris CC total N uptake at R6 was greatest in the AA/I, followed by ESN/BI, and all the urea treatments had the lowest uptake (Appendix-3). Plant N uptake was significantly greater with spring application than fall application only at V12 in Morris CC (Appendix-3). Lamberton CSb and Morris CSb had no difference in plant N uptake at any of the development stages (Appendix-4). Waseca CSb had only differences in V12 (Appendix-4). Crookston CSb had greater plant N uptake at R6 in the ESN than Urea treatments (Appendix-4). For the CSb rotations only at V6 in Morris, Urea had significantly greater yield than AA/I (Appendix-5).
Soil measurements
At Waseca in CSb, amounts of soil NH4+-N for all the depths were similar for different N rate at V4 and Post-harvest (Appendix-6). Soil NO3–N increased with increasing N rates at V4 development stage at all depths (Appendix-6). The fact that there was substantially greater NO3–N in the top two feet of the soil correspond with the greater yields we observed for this location with spring compared to fall applications. Also, the fact that there were no differences in soil N at post-harvest despite mediocre grain yields; indicate that there was substantial potential for N loss. There were not significant differences on the soil NNH4+-N and NO3–N content at any depth among source/placement treatment at V4 and post-harvest (Appendix-7) ,except that U/BI had significantly greater N-NO3- content at12-24 and 24-36” for V4 development stage (Appendix-8).
At Lamberton in CC, soil NH4+-N was similar among rates, however soil NO3–N increases with N rate at all depths (Appendix-9) at V4. At V4, fall applications had greater soil NO3–N than spring applications at the 12-24 inches depth, but soil NO3–N was similar for fall than spring application timing for the other depths (Appendix-9). At post-harvest sampling, no differences were detected among N rates or time of application at any depth (Appendix-9). There were not significant differences on the soil NH4+-N and NO3–N content at any depth among source/placement treatment at V4 and post-harvest, except NH4+-N content at 0-12” that was greater in the U+I/SSB and lower in the U/BI (Appendix-10).
At Lamberton in CSb, similar NH4+-N concentrations were detected among N rates at V4 and Post-harvest sampling at all depths (Appendix-11). Soil NO3–N increased with increasing N rates both at V4 development stage at 0-12 and 12-24” and at all depths for post-harvest sampling (Appendix-11). When N source and placement treatments were evaluated, similar amount of soil NH4+-N and NO3–N were observed among treatments (Appendix-12, 13).
At Morris in CC, concentration of soil NO3–N increased with N rate at V4 development stage at 0-12 and 12-24” soil depths (Appendix-14). At post-harvest, only the 24-36” depth showed differences among N rates but in general low concentrations were observed at the end of the growing season (Appendix-14). At V4, soil NO3–N concentration at 0-12” was greater under the AA/I and ESN/I treatment than the other sources (Appendix-15), and at 12-24” soil NO3–N were similar among ESN and SSB treatments but greater than BI treatments (Appendix-15). No differences among sources/placement were detected on soil NH4+-N and NO3–N concentrations at post-harvest (Appendix-15).
In Morris CSb, the concentration of soil NO3–N increase when N rates increased at all depths at V4, and 12-24 and 24-36” at post-harvest (Appendix-16). Soil NH4+-N and NO3–N content was not affected by N source and placement treatments at both sampling times (Appendix-17).
At Crookston CWt, soil NO3–N content increased significantly with N rates at 0-12” and 12-24” at V4 (Appendix-19), and at all depths at post-harvest sampling (Appendix-19). No differences were found among N source and placement treatments at V4 development stage and Post-harvest (Appendix -20).
At Crookston CSb, at V4 soil NO3–N content was significantly affected by N rates at 0-12” soil depth, and at 12-24” at post-harvest sampling (Appendix-21). Soil NH4+-N and NO3–N content was not affected by N source and placement treatments at both sampling times (Appendix-22).
Total inorganic N (TIN) (NH4+-N plus NO3–N) in the top 36 inches of the soil were highest at V4 at Waseca 228 lbs N/ac for the spring application. Soil TIN was also high at V4 at Crookston for CC for the spring application. For these two locations at V4 fall TIN was also high with 161 lb N/ac at Waseca and 169 lb N /ac at Crookston. These two sites also had the highest TIN at post-harvest for both fall and spring applications, ranging between 102 and 106 lb N/ac. These results might help explain why there was no grain yield response to N at Crookston for CC. The fact that Waseca had a grain yield response that was linear (for fall) and a relatively high EONR (164 lb N/ac) for spring indicate that N was either not the limiting factor or that substantial N loss occurred before the crop was able to utilize substantial N. The fact that substantial N was present at post-harvest may be a reflection of either unused N (other factors were more limiting to the crop) or substantial mineralization after the crop reached physiological maturity. It was also interesting to observe that in Crookston CSb the EONR was high, but the TIN at V4 and post-harvest was relatively high and did not change substantially between the two sampling times for either fall or spring (range 124 to 102 lb N/ac). Overall, there were very small differences between CC and CSb rotations at V4 or post-harvest. Averaged across N rate at V4 for fall applications TIN was 132 lb N/ac for CSb and 138 lb N/ac for CC while for spring applications TIN was 146 lb N/ac for CSb and 145 lb N/ac for CC. Averaged across N rate at post-harvest for fall applications TIN was 92 lb N/ac for CSb and 87 lb N/ac for CC while for spring applications TIN was 94 lb N/ac for CSb and 95 lb N/ac for CC. These results need to be further explored in relation to crop grain yield and N use. Similarly, additional analysis will be conducted with the data collected across all four years of the study.