All farmers throughout the growing season ask this question. In the spring of 1988 the Canada Grains Council undertook a project entitled "Development of a Risk Management Guide for Wheat Producers". The objective of the project is to develop a risk management guide which will enable wheat producers to evaluate the risk and probability of success associated with the use of inputs or other management practices throughout the growing season.

The PMG project is the sequel to the previous investigation, "The Potential for Intensive Wheat Production in Canada". The intensive management project examined the potential to use 5 cropping factors which have been quoted as being largely responsible for the dramatic yield improvement obtained in Europe during the last 15 years. The factors examined included use of narrow row spacing and/or higher seeding rates, high yielding semi dwarf cultivars, split N applications, foliar fungicides and plant growth regulators. Depending on the year and location, net benefits were found for each of these factors, ranging from the narrow row spacing giving the most consistent benefits to plant growth regulators being the least consistent.

The RMG has been designed to help farmers respond to the yearly fluctuation experienced on farms in the Great Plains. The guide provides for initial yield targets based on environmental and economic factors. It provides farmers with the necessary information to update yield predictions at critical decision making stages of crop growth. The guide will provide benchmarks and threshold levels for farmers to assess the risk or livelihood of achieving net returns from each cropping input. The net effect of the guide will be to increase the probability of the farmer achieving the maximum economic yield.

The section of the RMG I would like to present today deals with our experience with Flexible Nitrogen Management and Yield Tracking. Prairie agriculture is by its nature highly variable. Boom to Bust is not uncommon. A particular farm with an average wheat yield of 40 bushels may well experience a yield range of 10 to 60 bushels/acre. The farmer's job is to obtain the maximum economic yield each year. This will involve minimizing the inputs and costs in the drought years and pushing the yields in the responsive years.

Flexible Nitrogen Management is aimed at assisting the producer to balance the nitrogen investment with the productivity of the year. While we are still considering different options, we have been examining the 'fit' of a system with the following features:

3. Yield tracking is used to help update the need for expanded N supply.

David R. S. Rourke

Ag-Quest, Inc.

Minto, MB

January, 1990

Flexible N studies have been conducted at Minto for 2 years using 3 different moisture regimes each year. In addition, a number of researchers and farmers from across Manitoba and Western Canada have participated in data collection for Target Yield and Yield Tracking analysis. At the present tire the Target Yield, Flexible N Management & Yield Tracking work is based on the following formula. Further refinements to these basic formulas are expected.

1. TARGET YIELD

(Plant available water at seeding + expected precipitation [inches]) - 5 x 4 -5 inches is the amount of moisture needed to get the first bushel of yield. 4 is the # of bushels of yield obtained from each available inch of moisture. 4 is the constant used for Katepwa spring wheat which is a traditional tall hard red spring wheat characterized by having high protein. Plant available moisture at seeding is determined with the aid of a Brown soil moisture probe.

2. NITROGEN REQUIREMENTS

2 lbs. of N required per bushel of yield (Source of information is the Phosphate and Potash Institute). This could be higher if high protein premiums are sought.

3. NITROGEN FERTILIZER REQUIRED

Target yield x 2 - soil N to 2 ft depth

4. YIELD TRACKING-YIELD ESTIMATE AT ZADOKS GROWTH STAGE

31 (5.5-6 leaf stage)(# of plants/m2 x tillers/plant x spikelets/head x .392)

67.25 = Bu/acre.

5. FLEXIBLE N MANAGEMENT

Additional nitrogen should be applied according to comparison of Target Yield Vs Yield Estimate at G.S. 31. Further refinements in terms of soil moisture and plant N content at G.S. 31 will be considered as data is collected.

 

EXPERIENCE WITH THE RISK MANAGEMENT STUDY

Harvested Yield Bu/acre

 

Ideal

60 -

50 -

40 -

30 -

20 -

10 -

‘ ‘ ‘

10 20 30 40 50 60

! ! ! ! ! !

10 20 30 40 50 60

Target yield Bu/acre based on predicted plant available water

FIGURE 1. CORRELATION OR RMG TARGET YIELD COMPARED TO HARVESTED YIELD

Katepwa Wheat, Minto 1988-89

 

Harvested Yield Bu/acre

Ideal

60 -

50 -

40 -

30 -

20 -

10 -

‘ ‘ ‘

10 20 30 40 50 60

! ! ! ! ! !

10 20 30 40 50 60

Target yield Bu/acre based on predicted plant available water

Figure 1B. CORRELTION OF TARGET YIELD COMPARED TO HARVESTED YIELD COOPERATOR DATA 1989

Figure 1 illustrates the relationship between Target Yield estimated at seeding compared to the actual yield. While the fit is not perfect, it appears to be within reasonable limits for most situations (r = .89). "R" is the correlation coefficient, which has a value between 1 and – 1. 1 or – 1 means data has a perfect fit. A value close to 0 would mean there is little or no correlation within the data set. Notable cases where the yield was underestimated were on which was a low area of the field (highly productive) and another, which was a trial, established on summerfallow. In these cases we have probably underestimated the water +/or nitrogen supply. Data from cooperator trials show the correlation coefficient to be lower (r = .69) than that found at Minto (Figure 1B). The lack of consistency with cooperators data may be attributed to the wide range of growing conditions experienced as well as lack of experience in handling the Brown soil probe.

 

YIELD TRACKING VS ACTUAL YIELDS

During the past year, 13 trials have been monitored to examine the relationship between the estimated yield at growth stage 31 and the harvested yield. Figure 2 illustrates the results obtained.

 

Harvested Yield Bu/acre

Ideal

60 -

50 -

40 -

30 -

20 -

10 -

‘ ‘ ‘

10 20 30 40 50 60

! ! ! ! ! ! !

10 20 30 40 50 60 70

Target yield Bu/acre based on predicted plant available water

 

FIGURE 2. CORRELATION OF RMG G.S.31 PREDICTED YIELD

AND HARVESTED YIELD MINTO 1989

There has been a very strong correlation between the estimated yield at growth stage 31 and the actual yield (r=.91) from the data collected from the 1989 trials. There are two groups of data, which explain the major deviants from a perfect fit. One is where we have overestimated the actual yield; this tends to occur at yields less than 40 Bu/acre. Tiller mortality, small kernel size and lower kernel/head would be factors causing an overestimation of yield. The second group of data generally tends to underestimate yields and occurs with yields greater than 40 bushels/acre. Higher than expected tiller survival; large kernels and high number of kernels/head are the factors, which would lead to an underestimation of yield. The data collected from the 1988 and 1989 trials was also used to assess the accuracy of the RMG G.S. 31 yield prediction formula with the formula used by Manitoba Crop Insurance (Figure 3). The correlation coefficient was r=. 19 for the Crop Insurance formula compared to r=. 94 for the RMG formula. Yield predictions at this time of crop production can be useful to assist in determining the risk/benefit of applying a number of postemergent products such as top dressed nitrogen, fungicides and plant growth regulators. Figure 4 illustrates the results obtained when the yield tracking formula was used with data collected from cooperators. Essentially there was no fit (r=. 17). Reasons for the lack of correlation with cooperator data with Yield Tracking may be partly due to inexperience in measuring yield components at the early growth stages.

Figure 3. CORRELATION OF PREDICTED YIELDS AT G.S. 31 AND HARVESTED YIELDS

MINTO 1988 & 1989, KATEPWA WHEAT

Note: Figure 3 was included in the original proceeding.

Figure 4. CORRELATION OF RMG PREDICTED YIELDS AT G.S. 31 AND HARVESTED YIELDS

COOPERATOR DATA 1989

Note: Figure 4 graft was included in the original proceeding.

 

SPLIT N APPLICATIONS

A number of trials have been used to investigate the role which split N applications may have in crop management options. Trials have examined various factors such as type of nitrogen and method of nitrogen application, rate and timing for the topdressed nitrogen as well as the proportion of nitrogen applied proceeding relative to top dressing. The data presented in Figure 5 represents the combined data from 2 years and 2 moisture regimes, irrigated and natural. Our dry site typically had an excess of nitrogen relative to the requirements for the target yield, which eliminated the need for top dressed nitrogen.

FIGURE 5. RESPONSE OF KATEPWA TO SINGLE AND SPLIT N APPLICATION

MINTO 4 STATION YEARS

Note: Figure 5 graft was included in the original proceeding.

Figure 5 depicts a strong yield response to both total nitrogen as well as to topdressed nitrogen. Topdressed nitrogen was supplied at G.S. 31-32 as 34-0-0. The rate applied in each case was equivalent to 50 % of the N requirement to reach the target yield. For example if the target yield was 40 Bu/acre, total N required would equal 40 x 2=80 lbs./acre and 50 % of the target N requirements would be 80/2= 40 lbs. of N topdressed. Even applied as late as G.S. 31-32 the topdressed N was effective in recovering the yield obtained when the N was applied as a single application prior to seeding. By moving the yield curve for the topdressed N over to obtain the same total N (initial + topdressed = initial) one can find the yields were essentially the same.

The topdressing option has a number of advantages. The first is topdressing allows for corrections in the N supply to fi the requirements of the crop in that particular year. Secondly in wet years and with some wheat types, topdressing has been shown to result in yields higher than those obtained with single applications. In 60% of the trials conducted from 1985-87 at Minto, 5 Bu/acre increased HY320 yields when treated with an 80/40 lbs. N/acre split application compared to a 120-lbs. N/acre single application. In the other 40% of the trials, the split application treatment lowered yield by .8 Bu/acre compared to the single N application. Split application can be used to optimize the ratio between vegetative growth and grain production as well as in helping to control lodging in high rainfall years. There is of course some disadvantage to implementing the split N application technique. One is associated with the extra cost and time required for the application. A second may involve a price premium for ammonium nitrate fertilizer over NH3 or urea, which would have been used for the initial N application.

The results of a second series of Split Nitrogen trials are shown in Tables 1and 2. Nitrogen was applied at 2 basic rates, target yield N and target yield N x 1.5. Various split application timings and rates was examined as well as comparing granular and liquid fertilizer for topdressing. The second of the two trials was conducted using irrigation both before and after seeding.

TABLE 1. SPLIT NITROGEN MANAGEMENT TRIAL – NATURAL RAINFALL MINTO 1989

SEEDING	TREATMENT % N at G.S. 31	G.S. 49	TYPE OF N USED FOR TOPDRESSING	TARGET N YIELD TREATMENT* BU/ACRE	TARGET N x 1.5 YIELD TREATMENT BU/ACRE
100			--	32	31
50	50		34-0-0	40	31
50	50		20-0-0(3)	29	32
75	25		34-0-0	38	30
75	25		20-0-0(3)	30	29
50	25	25	34-0-0	35	30
50	25	25	20-0-0(3)	28	28
75	12	12	34-0-0	30	26
75	12	12	20-0-0(3)	27	33
75				28	28
50				36	28

* Available soil water at seeding = 6.6"

Expected May 1 to July 31 rainfall = 7.5"

Target yield = 36 Bu/acre

May 1 - July 31 rainfall 5.2"

 

TABLE 2. SPLIT NITROGEN MANAGEMENT TRIAL – IRRIGATED

MINTO 1989

SEEDING	TREATMENT % N at G.S. 31	G.S. 49	TYPE OF N USED FOR TOPDRESSING	TARGET N YIELD TREATMENT* BU/ACRE	TARGET N x 1.5 YIELD TREATMENT BU/ACRE
100			--	64	50
50	50		34-0-0	65	60
50	50		20-0-0(3)	57	52
75	25		34-0-0	62	55
75	25		20-0-0(3)	56	51
50	25	25	34-0-0	58	53
50	25	25	20-0-0(3)	53	52
75	12	12	34-0-0	57	51
75	12	12	20-0-0(3)	48	51
75				47	51
50				57	34

* Available soil water at seeding = 8.8"

Expected May 1 to July 31 rainfall = 11.5"

Target yield = 61 Bu/acre

May 1 - July 31 rainfall 9.2"

 

In both trials the 50/50 split using 34-0-0 as the topdressing source resulted in the highest yield. The benefit to the split application was the most dramatic in the natural rainfall trial. The actual yield obtained with the best treatments was very close to the target yields predicted before seeding for both trials. Increasing the N rate to 1.5 times the target rate resulted in lower yields, clearly indicating excess nitrogen can be as detrimental as too little.

A third series of topdressing trials involved examining five different sources and/or methods of nitrogen for topdressing. Table 3 summarizes the results obtained from 2 trials conducted in 1989. One trial was irrigated. Both trials were conducted as a split block design with one half of each replicate covered with a rain shelter for 10 days compared to the other half receiving either natural rainfall (.75") or natural rainfall plus supplemental water (.75 + 2.0= 2.75"). The split N treatments represented adding 50% of target yield goal to plot previously fertilized to 100% of target yield. A more dramatic trial could have been achieved had we only initially fertilized to 50% of target yield.

 

TREATMENT	COVERED	YIELD BU/ACRE UNCOVERED	AVERAGE
34-0-0 broadcast	48	52	50
28-0-0 strip band	45	46	45.5
28-0-0 sprayed	42	46	44
28-0-0 injected	42	49	45.5
20-0-0 (3) sprayed	40	50	45
Check	41	44	42.5

The trial results indicate that 34-0-0 gave the highest and most consistent results compared to all other sources +/or methods of topdressing nitrogen. The results were somewhat surprising, especially on the covered plots. It was expected the 28-0-0 injected would have been less dependent on rainfall than the other treatments. The 20-0-0(3) was the least suitable product under the stress conditions (covered) but its performance improved dramatically in the uncovered or low stress conditions.

 

CONCLUSIONS

1. Target yield calculations prior to seeding (Minto 1988 + 1989) provided a very good estimate of harvested yield. Calculations using cooperator data were less accurate.

2. Yield tracking conditions at G.S.31 (Minto 1988 + 1989) provide a very good estimate of harvested yield and was much more accurate than the formula currently used by Manitoba Crop Insurance. Unfortunately the use of the RMG yield tracking has had poor initial results when used by cooperators from across Western Canada.

3. Split N applications have been shown to be effective in supplementing initial nitrogen fertilizer applications. In some cases the split application out yield the single applications. The most consistent product for topdressing wheat has been 34-0-0 (ammonium nitrate). The positive responses obtained from split N applications make Flexible N management a viable option.

5. The role of yield tracking to predict the need for topdressing is still under investigation. It is felt that by using a combination of factors including yield tracking, soil moisture at S.31 and soil and plant N at G.S.31, accurate prediction of the need for supplemental nitrogen can be made, allowing farmers to save more money by not applying in unfavorable years and increasing revenue applying in responsive years.