Improving winter grain yields, grain quality, and nitrogen use efficiency in New England using adaptive management

As noted above, winterkill was a major issue preventing the completion of a number of the planned trials. As well, at some of our on-farm sites, data were high variability due to uneven plant stands and weed pressure, making it difficult to detect differences among topdress treatments. On-farm sites with highly variable results are indicated with an asterix (*) in Tables 1-4.

Given the reduction in data and increased interest in winter rye versus winter wheat among cooperating farmers, we chose to focus our efforts on evaluating the GS25 tiller count test. The discussion below summarizes results from both on-farm and research stations trials to answer the three research questions outlined above but phrased slightly differently.

1. Did N topdressing increase spike density and grain yield?
2. Can we use tiller counts at spring greenup to predict topdress N needs for winter grains?
3. Do we need to adjust recommendations for our region and non-wheat crops?

Did N topdressing increase spike numbers and grain yield?

The first step in our analysis was to determine if N topdressing is an effective strategy to increase tillering (evaluated here as an increase in spike density at maturity) and therefore yield. Prior research station trials in Maine and Vermont had shown this to be the case for winter wheat (Mallory and Darby, 2013) but we needed to verify in on-farm conditions and for other crops.

In the on-farm trials, topdressing significantly increased spike numbers relative to the no-topdress control treatment at four of five winter wheat sites, four of six rye sites, and neither spelt site (see Tables 1-3). At three of the winter wheat and one of the rye sites, the highest spike densities were achieved with the highest N rates, whereas at most of the rye sites and one wheat site, spike densities were not affected with N rates higher than 40 or 60 lb/acre. Note that for sites with relatively high initial tiller numbers where topdressing had an effect (sites 1, 8, and 9), final spike numbers were lower than initial tiller numbers. In these instances, topdressing did not actually increase tillering but rather increased the portion of tillers that developed into productive spikes, and therefore helped maintain yield potential.

Topdressing significantly increased grain yields at three of five wheat sites, three of six rye sites, and neither spelt site. Topdressing was a significant effect at one other wheat site (site 5) but in this case N applications actually reduced yields presumably due to more severe lodging (although 100% of plot areas lodged for all treatments). At the other sites, lodging increased with increasing N rates at seven of the sites indicating the importance of identifying the minimum N rate needed to improve yields. At the sites where topdressing increased yields, maximum yield levels were reached with 40 and 60 lb N/acre where tiller counts were low (sites 6 and 7) and at 20 lb N/acre where tiller density was high (sites 1 and 8). Also note that at sites with high initial tiller counts, topdressing at low rates caused more lodging than at sites with low tiller counts.

In the research station trials, topdressing clearly increased winter wheat spike density and grain yield, as shown by the N response curves in Figure 1. The figure isummarized all five planting treatments for 2015 and the cv Harvard at 300 with no preplant N treatment for 2014.

Can we use tiller counts at spring greenup to predict topdress N needs for winter grains?

Our on-farm results suggest that yes, spring tiller numbers can guide spring topdress decisions for winter rye but the data are inconclusive for wheat and spelt. Table 4 shows that at winter rye sites with tiller densities above 50 tillers/ft², yields maxed out with little or no topdress N (sites 3-6). In contrast, sites 6 and 7 had low initial tiller numbers and required substantially more topdress N to produce maximum yields. The winter wheat sites did not follow this same pattern. For wheat, response to topdress N appears unrelated to tiller density, however results from the one low tiller site are not reliable due to high variability in stands.  

Data from the research station trials and on-farm sites 3, 4, and 6-11 were used to evaluate the economic optimum N topdress rate for winter wheat and rye as a function of GS25 tiller density. The quadratic-plateau model fit the N response for wheat and some rye sites best (Fig. 2), whereas the quadratic model provided the best fit for the rye sites where higher topdress rates caused considerable lodging and declines in harvested yield (Fig. 3). The economic optimum N rate was identified as the lowest N rate at which ROVC reached maximum value.

For winter wheat, no relationship was found between the economic optimum N rate and tiller density, over the range of tiller densities that occurred in our trials (Fig. 4). Economic optimum N rates ranged from 54 to 77 kg/ha (48 to 69 lb/ac) regardless of tiller density. However, the range of tiller densities in these wheat trials was relatively narrow (63 to 97 tillers/ ft²) and all were close to or above the range in critical values that other states use for deciding whether to topdress at GS25 to stimulate tillering or wait until GS30 for soft winter wheat. This range is 50 to 70 tillers/ft² (Alley et al., 2009; Murdock et al., 2009; Weisz and Heininger, 2012) but these recommendations assume additional topdress will be applied at GS30. Our cooperating farmers indicated that they will likely topdress only once, either at GS25 or at GS30 but not both. For single topdress management, Virginia uses a slightly higher critical tiller density (70 to 90 tillers/ft²) and bases actual rates on soil nitrate levels. For our wheat trials, we found no significant relationship between optimum topdress N rate and soil nitrate (data not shown), but again tiller densities for our sites were relatively high and would suggest waiting to topdress at GS30.

For winter rye, there was a larger range of tiller densities and clearer differences among sites in economic optimum N rate than for wheat. The two rye sites with tiller densities around 400 tillers/m² (37 tilllers/ft²) had an optimum N rates of 94 and 101 kg/ha (84 and 90 lb/ac), whereas optimal N rates at the sites with higher tiller densities were 0 or 36 kg/ha (32 lb/ac).

Do we need to adjust recommendations for our region and non-wheat crops?

As noted before, the tiller density test is commonly used to decide whether to apply topdress N at GS25 to stimulate tillering or to wait until GS30. To evaluate the test for this decision for winter wheat, we calculated the relative yields (or ratio) of plots topdressed at GS25 versus an additional set of plots that were topdressed at GS30. Relative yields above 1 indicate higher yields when topdressing was applied earlier than later.

The relative yield (GS25:GS30) of winter wheat was negatively related to tiller density over the range of tiller densities that occurred (Fig. 5). At tiller densities below 954 tillers/m² (87 tillers/ft²), yields were higher when topdress N was applied at GS25 as compared with GS30. This relationship did not appear to be affected by preplant manure. A similar approach was used to evaluate the critical tiller density in terms of difference in profits when topdress is applied at GS25 versus GS30 (Fig. 6). In this case, it was more profitable to topdress earlier than later at tiller densities below 906 tillers/m² (84 tillers/ft²) given the assumed prices for N fertilizer and wheat grain. These values are within the 70 to 90 tillers/ft² critical range identified in the recommendation from Virginia for single topdress application management (Alley et al., 2009). At tiller densities above 90 it is advised to postpone application until GS30 in order to avoid forming too many tillers and losing N to leaching. While more testing is needed to verify the critical level for topdress timing for wheat in the Northeast, the fact that our results agree with existing recommendations from another region suggests that we may not need to adjust those recommendations very much.

For winter rye, we were not able to conduct a similar analysis of the effect of delaying topdress application because topdress was only applied at GS25 in the on-farm rye trials. However, it appears from the rye data presented in Fig. 4 that winter rye has a lower critical tiller density than wheat when price assumptions are equal, and it would be even lower if rye grain were assigned a lower value depending on the market.

References:

Alley, M.M., P. Scharf, D.E. Brown, W.E. Baethgen, and J.L. Hammons. 2009. Nitrogen management for winter wheat: Principles and recommendations. Virginia Cooperative Extension, Blacksburg, VI.

Mallory, E.B., and H. Darby. 2013. In-season nitrogen effects on organic hard red winter wheat yield and quality. Agron. J. 105:1167-1175.

Murdock, L., J. Grove, and G. Schwab. 2009. Fertilizer management. In: A comprehensive guide to wheat management in Kentucky. The University of Kentucky college of Agriculture, Lexington, KY.

Weisz, R., and R. Heiniger. 2012. Nitrogen management for small grains, p. 27-33, In R. Weisz, ed. Small grain production guide 2011-2012. North Carolina State University, Raleigh.

Table 1. Effects of spring N topdressing rates on winter wheat spike density, grain yield, and lodging in on-farm trials. Site 2 was managed organically.

Averages within a site that share the same letter are not significantly different from one another (using Tukey’s Least-Significant Difference Tests). No letters means the treatment effect was not significant.

Table 2. Effects of spring N topdressing rates on winter rye spike density, grain yield, and lodging in on-farm trials at six site-years. All sites were managed conventionally.

Averages within a site that share the same letter are not significantly different from one another (using Tukey’s Least-Significant Difference Tests). No letters means the treatment effect was not significant.

Table 3. Effects of spring N topdressing rates on winter spelt spike density, grain yield, and lodging in on-farm trials at two site-years. Site 12 was managed organically.

Averages within a site that share the same letter are not significantly different from one another (using Tukey’s Least-Significant Difference Tests). No letters means the treatment effect was not significant.

Table 4: Average tiller counts and the spring greenup (GS25) topdress N rate that produced maximum yield for the on-farm trials with rye, wheat, and spelt.

* Uneven crop plant stands and weed pressure caused high variability in yields.

** Only one GS25 N rate tested at these sites. Other treatments included GS30 topdressing.

Figure 1. Spike density and grain yield as affected by rate of N topdress applied at GS25 in research station trials conducted in Old Town, Maine in 2014 and 2015. Data points are the average of 5 replicate plots for a variety of seeding rate and preplant nitrogen treatments.

Figure 2. Example of how the economic optimum GS25 N rate was determined by planting treatment or on-farm site using a quadratic-plateau model. Data points are from individual plots for the planting treatment cv Harvard seeded at 200 seeds/m² with no preplant nitrogen.

Figure 3. Example of how the economic optimal GS25 N rate was determined by planting treatment or on-farm site using a quadratic model. Data points are from individual plots from site 8.

Figure 4. Economic optimum N rate as affected by GS25 tiller density for winter wheat (blue) and rye (green). For number/ft², divide number/m² by 10.8.

Figure 5. Relative yield of topdressing at GS25 versus at GS30 as influenced by tiller density for winter wheat that received no preplant N (dark blue circles) or preplant manure N (light blue diamonds). Data are from the Maine research station trials.

Figure 6. Difference in profits for topdressing at GS25 versus at GS30 as influenced by tiller density for winter wheat that received no preplant N (dark blue circles) or preplant manure N (light blue diamonds). Data are from the Maine research station trials.