This project seeks to demonstrate improved agronomic nitrogen use efficiency (NUE) and sweet corn yield per unit N applied through a combination of deep fertilizer banding, strip tillage, and cover cropping. Improved NUE will increase sweet corn profitability and reduce N losses to the environment. Short-term outcomes of this project include an increased awareness of strip tillage and deep N banding as a viable option for sweet corn production in the North Central region; intermediate-term outcomes include increased adoption of reduced tillage, cover cropping, and deep banding of N by sweet corn growers region-wide; reduced N and herbicide use; reduced tractor use and fuel purchases; increased profitability; and reduced N leaching and nitrous oxide (N2O) release. We will compare sweet corn yield, quality, and profitability, as well as N loss to the environment from leaching and N2O flux in two treatments: one with broadcast fertilizer incorporated with conventional, full-width tillage and the other with deep-banded fertilizer and strip tillage. Cover crops are used in both treatments. We hypothesize that deep-banded fertilizer, compared to broadcast applications, will be more accessible to sweet corn roots and less accessible to weeds emerging near the soil surface, increasing crop growth and yield and improving weed management. Increased uptake by the sweet corn plant will result in higher yields and in less N lost through leaching and as N2O. Trials on collaborators’ farms will allow us to examine these practices in production systems and analyze profitability by creating partial enterprise budgets with farmers’ input costs and revenues for their standard practice and with deep-banded N fertilizer and strip tillage. Profitability analysis and on-farm demonstrations will increase grower awareness of these practices and hopefully lead to increased adoption.
Nitrogen (N) is an essential element for crop growth and N fertilizers are often over-applied to ensure optimal yields. When N is applied in excess to crop demand, however, it can be lost from the agroecosystem. Two common N loss pathways include nitrate (NO3-) leaching and nitrous oxide (N2O) flux as a result of partial denitrification. The latter typically represents only a small fraction of N loss, but is important because nitrous oxide is a potent greenhouse gas. In addition to causing environmental problems, N that is applied in fertilizers but is lost to either leaching or via nitrous oxide flux represents a direct cost to growers in wasted resources. Thus, agricultural practices that can mitigate nitrate leaching and nitrous oxide flux are beneficial to both growers and to society as well.
Reducing nitrogen application rates while maintaining yields and profits is a challenge that can be addressed by improving the agronomic nitrogen use efficiency (NUE) of a cropping system—getting more harvestable yield per unit nitrogen applied (Robertson and Vitousek 2009). NUE can be improved in part by placing nitrogen closer the developing plant roots, thereby increasing the likelihood that it will be taken up by the crop rather than lost to the environment. Strip tillage (ST) is a form of reduced tillage that offers a convenient way to band fertilizers at depths of 6” or more. In ST, strips are tilled only where the crop will be planted while the rest of the soil remains undisturbed. Yield of many crops, including sweet corn (Luna and Staben, 2002), is either similar or higher in ST compared to conventional, full-width tillage (FWT). ST offers other benefits as well, including lower cost because primary and secondary tillage is accomplished in one pass (Luna and Staben 2002). Improved NUE through banding and deep placement of N sources may also have important benefits for weed management—an ongoing challenge for sweet corn producers. Since many weeds are highly responsive to N fertility, deep-banding fertilizer may enhance weed suppression by placing resources beyond the reach of weeds emerging close to the soil surface.
This experiment compared sweet corn yield and profitability, weed management, and N loss to the environment from leaching and N2O flux in two tillage systems: one with broadcast fertilizer incorporated with FWT and the other with deep-banded fertilizer and ST. We hypothesized that deep-banded fertilizer, compared to broadcast applications, would be more accessible to sweet corn roots and less accessible to weeds emerging near the soil surface, thereby increasing crop growth and yield and improving weed management. Increased uptake by the sweet corn plant was expected to result in higher yields and in less N lost through leaching and as N2O.
The primary objectives of this experiment were to evaluate the impact of ST with deep N fertilizer banding on:
- sweet corn yield and profitability
- N losses via leaching and partial denitrification to nitrous oxide
- weed management efficacy with herbicides.
We also used this experiment to address a secondary objective: to evaluate the impact of relative strip placement on yield. Consultation with growers led to the development of this secondary objective as they were interested in knowing how strip placement from year to year would affect crop yield.
Detailed methods are available in Appendix 1. Briefly, these experiments were conducted from 2011-2013. We conducted a detailed experiment in which we measured sweet corn yield, N loss, and final weed biomass at the Kellogg Biological Station in Hickory Corners, Michigan. We also conducted three trials on farmer cooperator fields to compare ST with the farmers’ standard FWT practice. The timeline of field operations for these experiments is given in Table 1.
Research station trials The research station trials had three tillage treatments as the main plot factors: FWT, ST with the strips located in the same position from year to year (ST same), and ST with the strips located between the previous year’s strips (ST offset). Cover crops were used over the whole experiment—either spring-planted oats (in 2011 and 2012) or fall-planted winter rye (in 2013). In 2013, we also included two weed management intensities as subplot treatments—the high intensity weed management included both pre-emergence herbicides and a single hand-weeding pass while low intensity management included just the herbicides. In FWT, P, K, and some N (45 kg N/acre) was broadcast immediately prior to tillage while this fertilizer was banded 15 cm deep in the tilled in-row (IR) zone in the two ST treatments (Table 2). Forty-five kg of N/acre was also applied at sweet corn planting and at side-dress in each treatment. At harvest, all ears were collected and sorted into two categories—marketable (>4 cm diameter and silks dried down) and unmarketable (would not be marketable for > 3 days after harvest). The number and weight of each category were recorded. In 2013, we added an additional category to reflect the large number of immature ears that were present at both sites. This category was called immature (marketable size, but silks not completely dried down; these would be marketable within a day or two of harvest). For the on-farm trials, just the number of ears is presented as this is how these growers measure yield. Biomass is also presented for the research station trials.
Soil samples (0-20 cm) were collected from the IR and between-row (BR) zones at least bi-weekly and analyzed for nitrate and ammonium. In 2011 and 2013, we also measured nitrous oxide flux (NOF) throughout the growing season using the static chamber method described by Kahmark and Millar (2008); chambers were located in each zone. To determine the amount of soil nitrate remaining after sweet corn harvest, we collected deep soil cores (1 meter deep) following harvest in each year. These were separated into 20 cm sections and nitrate and ammonium were measured in each section. Since the BR zone is approximately twice the width of the IR zone, a weighted average of both NOF and the residual soil nitrate were calculated to estimate plot-wide N loss.
On-farm trials Treatments for the on-farm trials in 2012 included both ST and the farmer’s standard FWT practice—either a chisel plow followed by two passes of a rototurner (for Mr. Van Houtte) or a chisel plow followed by two passes of a field cultivator (for the Zilkes). Each tillage type had plots with and without a spring-planted oat cover crop. Fertilization was similar to that used on the research station trials, with some being deep-banded in the IR zone in ST and some being broadcast over both zones in FWT. In 2013, we tilled (ST and FWT; FWT=moldboard plow followed by two passes of a field cultivator) directly into a wheat cover crop that had been established in fall 2012. Fertilization and planting were identical to the other on-farm trials. Wheat in the BR zone in ST was terminated with glyphosate that was tank-mixed with our pre-emergence herbicides applied at planting. Yield was assessed in these trials, as was final weed density and biomass.
Profitability Tillage costs were estimated using Machdata.xlsm (Lazarus 2014), which uses an economic engineering approach. We estimated the operating and ownership costs of three different ST options, as well as the cost of common FWT equipment. Please refer to Appendix 1 for the assumptions used in these calculations. Since these cost estimates are highly dependent on the assumptions used, we conducted a sensitivity analysis varying three of these assumptions—farm size, (acres or hours that the implement is used annually), how long the equipment is kept (hours to trade in), and the inflation rate.
To compare profitability of ST to FWT for sweet corn production, we used a partial budget approach in which only revenues and costs differing between the two tillage types are considered. Changes in revenue were determined by considering yield following the two tillage types without cover crops—data from the on-farm trials in 2012 were used, as were yield data from previous research station trials (see Appendix 1 for details of these trials). Changes in cost resulted from operations that differ between the two tillage types; these included a broadcast fertilizer pass since fertilizer could be deep-banded with the strip tiller, all tillage passes, and cultivation for weed management. Weed management costs may differ for ST compared to FWT, so we also created a partial budget using different weed management tactics.
Environment. Temperature and precipitation for the Kellogg Biological Station for May through September in 2011-2013 are shown in Table 3. Environmental conditions varied widely during this period. 2011 was a relatively warm and wet year. 2012, in contrast, was hot and dry—precipitation in 2012 was much lower than the ten year average, with only 227 mm of precipitation during the growing season. 2013 was a relatively cool year, with slightly below-average temperatures in June, July, and August but, until September, had relatively normal precipitation.
Sweet corn yield. On the research station trials, averaged over all years, sweet corn yield measured by weight was 18% higher in ST offset with the deep fertilizer banding than in CT with broadcast fertilizer (Figure 1; contrast p=0.010). Yield in the ST same treatment was similar to both CT and ST offset. Sweet corn marketable yield was also affected by year, with 2013 having the highest yields, 2012 the lowest, and 2011 having intermediate yield (Figure 1). We anticipated that biomass production and marketable yield of both crops would be higher in ST compared to CT because of the potential direct benefits of ST on soil moisture (Wilhoit et al., 1990; Haramoto and Brainard, 2012), N retention (Sainju and Singh, 2008), and temperature moderation (Mochizuki et al., 2007), as well as indirect benefits of deep-banded N that is facilitated by ST (Maddux et al., 1991; Malhi et al., 2001).
For the on-farm trials in 2012, neither tillage nor cover crop treatment affected the number of marketable sweet corn ears produced (Figure 2). At the Zilke farm in 2013, the number of marketable plus immature ears was similar between tillage types (Figure 3). However, when only marketable ears were considered, yield was marginally higher in ST than in FWT (p=0.104). This suggests that ears matured slower in FWT than in ST. There was a lot of flooding in this wet year, and we observed more saturated soils in FWT than in ST; this could have led to delayed corn plant maturation in FWT.
Residual soil nitrate after harvest. Tillage effects on deep soil nitrate (20-100 cm) remaining after harvest were influenced by year, so years are presented separately. Following sweet corn harvest in 2011, ST offset had 17 kg NO3–N/ha less in the deep soil fraction than in CT (Table 4). In 2012, FWT had 26 kg NO3–N/ha more deep soil nitrate than ST same. In 2013, deep soil nitrate was similar between the tillage treatments. By the spring of 2012, deep soil nitrate was low in all tillage types—in most cases less than 5 kg NO3–N/ha—which suggests that the nitrate present in the previous fall was leached beyond the sampling depth over the winter. In spring 2013, following the drought year when soil N remained high in the fall after harvest, FWT had more deep nitrate than the two ST treatments in sweet corn (Table 4).
Whole-plot estimates of IN remaining in the top 100 cm of soil after harvest (Table 5) can be considered a “worst case scenario” of N loss in the case of poor cover crop establishment or growth. Tillage influenced total residual nitrate in 2 of 3 years. In 2012, ST reduced the total potential N loss compared to FWT by NO3–N/ha. In 2013, the year with lowest residual soil nitrate throughout the soil profile, ST-same increased total potential N loss compared to FWT by NO3–N/ha.
There was much more soil N left in 2012 than in 2013 in both the deep and surface sections (Figure 4). Higher amounts of residual soil nitrate in 2012 were likely due to hot, dry conditions in this year (Table 3) that limited plant biomass production. Yields were lowest in 2012 (Figure 1), as was total plant biomass production (not shown). There was also more soil nitrate remaining after harvest in the surface layer (0-20 cm) than in the deep layer in 2012 (Figure 4). It is likely that this is also a result of dry conditions in 2012—downward movement of nitrate was limited by dry soil conditions.
Because of the relatively large amount of soil nitrate left after harvest in 2012 (86-148 kg NO3–N /ha), the effect of ST in this year is important. There was also a non-significant trend towards ST reducing total residual soil nitrate relative to FWT following sweet corn in 2011 (Table 5). These two crop*year combinations have the highest levels of residual soil nitrate. However, in any year, only one of the ST treatments was effective in reducing deep soil nitrate relative to FWT—ST offset in 2011 and ST same in 2012 (Table 4). The effect of each ST treatment was inconsistent between years. This inconsistency was also observed by Al-Kaisi and Licht (2004)—lower deep soil nitrate (15-120 cm) was observed with ST compared to FWT, but only after two years in ST at one site and not at the other.
Nitrous oxide flux (NOF). Cumulative NOF in 2011 is shown in Figure 5; data from 2013 are still being analyzed. In 2011, cumulative NOF over the sweet corn growing season was greater IR than BR (p<0.0001) but was not affected by tillage treatment. Cumulative IR NOF averaged 745 g N2O-N/ha over 105 days, while BR NOF averaged 232 g N2O-N /ha over this period. The observed spatial heterogeneity in NOF was likely due primarily to differences in the location of IN fertilizer application—most of the N fertilizer was applied to the IR zone (Table 2). Generally, cumulative NOF was higher with higher levels of soil nitrate, expressed as the season-long average (Figure 6); NOF is often correlated with increased soil nitrate content (McSwiney and Robertson, 2005).
Area-corrected NOF measured in sweet corn (350-430 g N2O -N/ha/105 days) was lower than that typically measured out of ST field corn. With ST and deep urea banding, similar to our trial, reported NOF from field corn ranges from approximately 1.7 kg N2O-N/ha in irrigated production (Halvorson et al., 2013) to over 5 kg N2O -N/ha in heavier, poorly drained soils (Nash et al., 2012). In irrigated sweet corn, NOF ranged from 555-668 g N2O/ha over a 12 week period, about 80% of our season (Haile-Mariam et al., 2008). Our cumulative NOF was lower than these, though we also fertilized with a much lower N rate (135 kg N/ha compared to 225 kg N/ha). Overall, NOF represented only a small fraction of the nitrate remaining in the soil profile—between 0.2-1.4% over the different years and treatments (not shown). While NOF is important because of its impact as a greenhouse gas, it does not contribute much in the way of nitrogen loss (Halvorson et al., 2013, Haile-Merriam et al., 2013).
Weed biomass. Weed biomass data are still being analyzed.
Profitability of ST compared to FWT. Using Machdata.xlsm (Lazarus 2014), we determined that one pass from an 8-year-old chisel plow cost a total of $11.40/acre, while one pass from a similar age field cultivator cost $10.60. Our high cost ST scenario cost $25.60/acre, the middle cost scenario was $18.00/acre, while the low cost scenario was $17.30/acre (Table 6). The high cost option was $7.60 and $8.30 more per acre than the medium and low cost options, respectively. Implement ownership costs for the STH option are more than double those for the STM option—a reflection of the higher purchase price for the STH option. Since fuel, lubrication, and labor costs were the same for all ST options, differences in operating costs are tied to repairs and maintenance. These are linked to equipment age so are higher for the STL option that purchases used equipment than for the STM option that purchases new equipment; they are also linked to the purchase price so are higher for the more expensive STH option and this is mostly responsible for cost differences. The difference of $0.30/acre in repair and maintenance costs between the STL and STM options translates to $68 in annual costs assuming the equipment is used on 200 acres. A sensitivity analysis showing how changing the operation size, the number of years the ST equipment is kept, and the inflation rate influence the total cost per acre of the medium cost ST option is shown in Appendix 2.
The partial budget for changing tillage from FWT to ST, with the three different ST cost scenarios, is presented in Table 7. Given the amount of tillage done by the Zilkes and Mr. Van Houtte (one primary tillage pass followed by two passes for secondary tillage) and cost estimates per acre (Table 6) we estimate total tillage costs to be $32.55/acre in FWT (Table 7). In all ST scenarios, $18.84/acre was saved by eliminating an additional broadcast fertilizer application and the cultivation pass. Assuming weed management costs stay the same, this represents a total savings of $25.82, $33.40, and $34.08 for the high, medium, and low cost ST options, respectively, relative to FWT. Yields were similar between FWT and ST at the on-farm trials, so there is no change in revenue and these savings represent increased profit. The partial budget for two alternative weed management scenarios is included in Appendix 2
Educational & Outreach Activities
Results of these trials were presented to an audience of approximately 100 sweet corn growers, dealers, and extension personnel at the Great Lakes Fruit, Vegetable, and Farm Market Expo in Grand Rapids, Michigan, in December 2012. Poster presentations were also given at the Tri Societies meeting (Agronomy Society, Crop Science Society, and Soil Science Society of America) in October 2012 and the Ecological Society of American meeting in August 2013. Trace gas results were also presented to the North Central Region SARE Carbon, Energy and Climate Conference in September 2012.
Written results were published as part of my doctoral dissertation. Additional publications in peer-refereed journals and as an extension bulletin are in preparation.
Averaged over three years of trials at our research station site, ST offset combined with deep fertilizer banding resulted in higher sweet corn yields than FWT with broadcast fertilizer. This is an important finding for growers who are interested in pursuing ST for growing sweet corn.
Our results also suggest that ST with deep N fertilizer banding can influence soil N dynamics, but that these effects are also highly dependent on climatological conditions during the growing season as weather also played a role in determining the amount of soil nitrate remaining after harvest—both in surface and in deep samples. The impacts of weather were sometimes direct and sometimes indirect. One direct effect of weather was observed in sweet corn in 2012—reduced rainfall lowered downward movement of soil nitrate, resulting in a large difference between surface and deep residual nitrate after sweet corn harvest. An indirect effect of weather, however, contributed to large amounts of residual N over the whole plot in this year as crop growth and biomass accumulation was limited by lack of available moisture.
Our study demonstrated several important instances in which strip tillage reduced potential N losses relative to FWT. Strip tillage with deep N banding was successful in lowering potentially leachable nitrate at the 20-100 cm depth in one out of three years—this is important in systems with over-wintering cover crops to take up residual N in surface soils. ST also reduced total residual soil nitrate that could be leached in systems (0-100 cm) without these cover crops in this year. Our results suggest that the potential benefits of reduced tillage systems are likely to be greatest in years when environmental stresses limit plant uptake. This may be of increasing importance as climate change scenarios forecast increasing drought conditions in some areas.
Farmers were very interested in our on-farm trials and often stopped by to talk about the research. Interviews with sweet corn growers throughout the state showed interest in ST for soil conservation. Some important constraints to ST adoption, however, exist. Despite our finding that ST with fertilizer banding can save money (as well as time) compared to FWT, purchasing ST equipment to replace existing equipment may be cost-prohibitive for smaller growers, and for those growing a variety of crops for which ST is not as good an option (e.g. those on flexible row spacing like carrots or requiring black plastic mulch like peppers and tomatoes).
Areas needing additional study
Future research can be focused on two areas:
- In terms of reducing potentially leachable nitrate, what are the relative contributions of ST and of the deep N banding? How did relative strip placement influence residual soil nitrate and why was this effect variable?
- Why was sweet corn yield higher in ST offset compared to FWT? In ST offset, one year’s strips are tilled into soil that was undisturbed the previous year. Does this one undisturbed year give sufficient time for pools of labile soil organic matter to form that are then oxidized when tillage does occur? Also, within this context, what soil quality benefits might be conferred in the ST same system? If yields are maintained in ST same, relative to FWT, how can these soil quality benefits be incentivized to convince farmers to adopt this system?