Protein and Yield Response to Nitrogen Fertilizer and Variation of Plant Tissue Analysis in Wheat

Transcription

1 Protein and Yield Response to Nitrogen Fertilizer and Variation of Plant Tissue Analysis in Wheat Daniel Kaiser, Dept. of Soil, Water and Climate, U of M, St. Paul Research Questions The current N management guidelines recommend applying nitrogen based on the yield goal of wheat minus any N credits from a previous crop or from soil nitrate. This recommendation has served wheat growers for many years and is constant no matter where you are in the Red River Valley. For southern Minnesota table values are recommended for wheat because soil nitrate does not play as big of a role in recommendations. The idea of single sets of recommendations for large areas of the state are coming into question as farmers have adopted more precision agriculture technologies and have the ability to apply variable rates of crop inputs across fields and can collect yield data from individual field areas. In order to provide the information to make these recommendations data must be collected from multiple locations, especially if researchers want to better provide farmers with data that may provide differences between specific regions even within the larger regions within the state. Plant tissue analysis has become more increasingly offered to wheat and other crop growers as a way to assess the nutrient status within the plant at a particular time that may ultimately be related back to yield. Crop uptake generally follows a consistent pattern; however, nutrients vary in when they are taken up and where they may go in the plant. Therefore, if environmental conditions exist that may limit uptake certain elements may be deficient in some areas of the plant. For example, elements such as sulfur or zinc are taken up into the plant and are not remobilized into other plant tissues if the supply in the soil becomes low. Therefore, plant tissue analysis can be problematic and may not accurately reflect yield at the end of the season if the cause of the temporary deficiency is corrected. Also, varieties may differ in their growth patterns ultimately affecting crop uptake of nutrients and the concentrations in plant tissues. Therefore, particular normal values may be only normal for particular varieties in specific locations rather than a one size fits all number for a tissue concentration. No recent work has been published that shows variety to variety variation across locations. Therefore work should be undertaken to provide farmers with an idea of what their tissue analysis really means. The objectives of this study are to: 1) Study protein and yield response using nitrogen rate trials in Northern and Southern Minnesota 2) Evaluate economic optimum nitrogen rates for spring wheat Page 52 3) Evaluate the variability in plant tissue collected from the fully extended flag leaves 4) Compare variability in nutrient composition of plant tissue for spring wheat between varieties and locations in well fertilized trials. Results Nitrogen Rate Studies A summary of the locations studies are included in Table 1 and 2. Three locations were studies in northwestern Minnesota, one near, one at Crookson, and one near. Two locations were studied in southern Minnesota at and. Soil P and K levels were adequate for wheat (Table 1) for the Northern trials and were Low to Very Low in the southern studies. The two foot soil test N levels ranged from 34 to 88 lbs. of N per acre. A summary of grain yield data (adjusted to 13.5% moisture) is given in table 3. Grain yields were significantly increased by nitrogen application at all locations except for. Yield was maximized by 127 lbs of N at Crookston, 98 lbs at, 85 lbs at, and 15 lbs at. The actual yield increase was the greatest at the two northern locations averaging 2 bu/ac while the southern locations average only on average of 1. Table 4 summarizes yield data, estimated N need based on 2.5 x yield goal minus N credits, and an analysis of economic optimum nitrogen rates (EONR) that are needed to maximize grain yield only at individual locations. The predicted N need was greater than the actual need at Fergus Falls and, but was less than the actual predicted need at the other location. The predicted need and the actual need were close at most locations but differed greatly at and. The higher predicted need makes sense due to the fact that the amount of N needed to maximize yield is less than the amount of N needed to maximize yield and protein concentration. The predicted economic optimum N rate for yield varied by location. In most circumstances the ratio of the price per lb N to dollars per bu wheat should fall around the.15 price ratio or less. Since there was not a large response to N at most locations there was large difference in the amount of N recommended based on the economic model. Since the increase in yield per lb. of N was consistent across N rates (linear response) at, the agronomic optimum N rate will be the same at the EONR value as long as long as the change in yield per lb N is less than the price ratio used. For the location, the change was equal to.5 bu per lb of additional N

2 which was below the ratios listed in the table. Therefore, the EONR value would likely be at this location. Grain protein was influenced by one or more nitrogen treatments at all locations except for. A linear model was fit to the data at, but the data was curvilinear at Crookston,, and and was maximized by 26, 225, and 88 lbs of N at the three locations, respectively. It took more N to maximize protein concentration than for yield at Crookston,, and while it took less at and. While the N rate never maximized protein at the amount needed to get to 14% was between 3 and 6 lbs of N. The amount needed to reach 14% was higher at Crookston, but the remaining sites averaged 14% protein concentration or higher with no or a low rate of N. Table 6 summarizes economic optimum nitrogen rates (EONR) for two price rations ($/lb. of N: $/bu. wheat price) adjusted by various protein discounts. This data combines data collected from 28 to 212 from a study previously funded by the Minnesota Fertilizer Research and Education Council and was collected from sites in northwest Minnesota. The variety at all locations was until 212 when was seeded at the northern locations. Figure 1 shows sample calculations for this data based on the.1 price ratio for northern locations while southern locations are summarized in Figure 2. Since protein discounts can have a significant impact on the net return per acre and the varieties used in the northern study were high yielding but moderate protein varieties, the calculated MRTN (maximum return to nitrogen: amount of N needed to return a dollar in crop value for a dollar spent on fertilizer N) value tended to increase. The biggest jumps in the MRTN value were when the discount levels were either or $.1 per fifth. In the table the low and high values represent plus or minus a dollar from the MRTN value and are considered a flexible range where a producer can select what to apply based many factors related to the field, environment, or perceived risk. The data shows that a high amount of N, 21 lbs., was needed to maximize yields when discounts were high, at the.1 price ratio, and were at 171 lbs of N for the.15 ratio. This is in sharp contrast to current recommendations. This is some caution in the data because this is a fairly limited data set in terms of the number of locations and years. Analysis for individual years or areas of the state might yield some differences in the calculated MRTN values. In addition all locations included spring applied urea that was incorporated soon after application which may affect the data compared with fall applied anhydrous ammonia. The data for southern Minnesota is given in Table 7. Discounts were not considered for the southern Minnesota data since protein concentrations were at or above 14% at nearly all locations when N was not applied. At the.1 price ratio rate of N that returned the maximum economic value (MRTN) was 143 lbs of N while 116 lbs was the MRTN value for the.15 price ratio. It should be noted that the 2 foot soil nitrate test was included in this analysis. Therefore the total amount of N determined with the test would need to be subtracted from this level to determine how much fertilizer N to apply. All samples were collected in the spring prior to seeding and appeared to give the best idea of total N requirements across the locations in both southern and northern Minnesota. Plant sampling study Plant sampling has increasingly been used in order to determine hidden nutrient deficiency symptoms in crops. In this study we sampled 15 varieties from the on-farm variety plots to compared the nutrient levels in flag leaf tissue. The variety trials were selected because they are managed with fertility as a non-limiting factor. For all sites a single sampling date was selected even though some varieties did vary in their maturity level at sampling. A single composite soil sample was collected from each location at the time of sampling. Soil data is given in Table 8 for macronutrients and Table 9 for micronutrients. A number of tests were run for both macro- and micro-nutrients. In most cases K levels were adequate for crop growth. The exceptions were 211 and and in 213, which tested Medium. Soil test levels at,, and in 211 and in 212 tested Low in P while all other sites were Medium or higher. However, fertilizer should have been applied to take care of any nutrient deficiencies. Soil test K was the lowest at however there should have been adequate K based on current recommendations. There are no suggested optimum soil test levels for calcium, magnesium, zinc, iron, manganese, and boron for Minnesota. Recommendations are made for sandy soils with low organic matter content for sulfate sulfur. However, no sites fit that description even though soils had a coarser texture at the and locations. Soil organic matter concentrations at these locations should have been high enough to mineralize enough plant available sulfur. In the case of copper a soil test between 2.5 and 5. ppm is considered marginal on organic soils. In this case we were not dealing with any organic soils so the same range would not hold and in fact all of the soil test values came in at or below this level. We do not expect a significant chance of a positive yield increase due to copper fertilization at any of the sites studied. It is not impossible to see a deficiency since small grains tend to be sensitive to copper deficiency. Table 1 summarizes the significance of each individual main treatment effects, site and variety, and their interaction. Part of the purpose of this study was to determine if 1) varieties significantly differed in their tissue concentration and 2) if the differences were similar across sites. We continued on page 54 Page 53

3 would expect site averages to differ due to differences in environmental conditions, soil types, fertilizer rates, or management practices. For all nutrients the effect of site was significant, except for tissue Iron concentration. The effect of variety was highly significant for all nutrients except for zinc and iron. The effect of zinc was marginally significant but still failed to reach significance at the accepted level (P<.5). A summary of tissue concentration for each variety is given in Figures 3 and 4 in the Appendix. A review of the literature found very few published sufficiency values for the time of sampling that was done in this work. A summary of values given by the University of Kentucky extension is given in Table 11 and follows somewhat closely to values given by Agvise labs to their customers. Most data found were for analysis of whole plant samples not individual flag leaves. Using this data the concentration data can be compared to determine if any of the median values for the varieties were sufficient. Median values were sufficient for all varieties for P, Ca, Mg, S, Mn, Cu, Fe, and B concentration. Potassium was consistently low across hybrids with the Median value testing lower than everything else for. In the case of N, Zn, and Cl, tissue concentrations were at the low end of the sufficient range and into the low range. Plant nutrient concentration summaries by location are given in Figures 5 through 8. Calcium, Mg, Fe, and B levels were all in the sufficient range across locations. Flag leaf N concentration was in the low range at 9 of the locations. For the remaining 8 locations median values were sufficient but on the low end of the range. Phosphorus was considered low only at in 212 while K was low at all locations except for in 212. Sulfur concentration was mostly sufficient except for Fergus Fall and in 211 and in 212 which were sufficient to low. Zinc concentration was borderline low at all locations except for 212. Manganese was sufficient except for Humboldt in 211 and during both years. Copper was low at 211 and sufficient to border line low at 211. Chlorine was sufficient to low at Fergus Falls 211 and 212. Median Cl levels were considered low at and during both years of the study and and in 211. When considering yield there was little correlation across locations and hybrids with any of the tissue concentrations (Table 12). The greatest correlation was between Cl concentration and yield. However, when regressed there was very little predictability of chlorine concentration on final yield. Grain protein was not significantly related to any of the plant tissue concentrations except for Cl which was negatively correlated to yield. Individual correlations within location are given in Table 13 for grain yield and Table 14 for grain protein. The best correlation was with most of the macronutrients (N, P, Ca, Mg, and S) at in 211. While many factors can be considered significant there level of significance was not high for most concentrations at most locations. There were several significant negative correlations within locations. For instance, K concentration at Humboldt in 211 was negatively correlated with yield. In terms of protein we were interested whether there was any relationship with N at any of the locations. Only one location showed a highly significant correlation between N and protein, in 211 while two locations, 211 and 212 and 212, showed moderate significance. Another element that has been questioned to effect protein is S. However, there was no evidence of a positive effect of S on grain protein concentration. Overall, there is no evidence of a clear single predictor of yield at any location. The average values for each nutrient concentration were compared to soil test values taken at the time of sample and there was no clear relationship between soil test values and tissue nutrient concentration (Table 15). The best relationship was between tissue S concentration and soil test sulfate. However, there was no indication that S was low at any location. Significant main treatment interactions for most nutrients may indicate that the lines do not always vary the same way at all locations or that the effect of line may only be significant at some of the locations. There was only a significant interaction for all the macronutrients and half of the micronutrients. There was no evidence of a location and variety interaction for Zn, Fe, and Cu. Since there was a lot of variation in the effects between locations an analysis was conducted on ranks to simplify the data. Rank data across locations for macronutrients, grain yield, and grain protein concentration are given in Table 16. There was no variety that consistently ranked high or low for all macronutrients. For instance, generally ranked among the highest varieties for N and Mg concentration but among the lowest for K. While was among the lowest for N and P, but in the middle for K, Ca, Mg, and S. In the case of micronutrient concentration, ranked among the highest for Zn and Mn, but low for Fe, Cu, B, and Cl. The lack of consistency across all nutrients indicate that the effect of variety is somewhat random across locations and which indicate that separate sufficiency levels would not be required for different varieties. If some varieties were consistently low or high then separate levels may be warranted. In addition, the somewhat random differences between locations indicate that factors other than the measured soil test values, which include environmental stresses, are likely more important in determining why differences may occur between locations. If sufficiency rates need to be determined separate fertility trials with multiple nutrient rates are needed. We were curious if some varieties would show larger differences in certain nutrient concentration. Other than the K difference seen with, there was no evidence that variety would play a significant enough difference to affect results. It should be cautioned that since all Page 54

4 samples were taken at the same time the difference could be due to differences in maturity of the crop. Application/Use The nitrogen rate data when combined with previous work shows that in many cases N rates applied based on historical yield goals likely will not limit yields when yield potential is high but will limit protein. Economic data shows that limiting protein can have a significant impact on economics within a field, especially with a variety that does not have a high genetic protein potential. In order to maximize profit more N likely needs to be applied. However, using the current recommendations if the yield goal can be accurately predicted within a given year the current recommendation of 2.5 x yield goal minus N credits may be used. The biggest challenge to wheat growers is accurately predicting yield goals. In addition, there are justifiable concerns for yield loss due to lodging from the over-application of N. The risk of over- or under application of N needs to be weighed before making the decision on how much to apply, which could hinge on what variety is planted within a field. The plant analysis data can be important in that it can tell us how much variability in tissue concentration we can expect based on sites and varieties. Since we do not have specific fertilizer rates we cannot fully determine optimum concentrations with this data. Yields can be used in order to determine if there is a link between yield and concentration, but since yield potentials varied significantly by location it can be difficult to determine relationships across locations. Since we are dealing with varieties, there is also some yield differences inherent between varieties which can make it difficult to determine if the difference in achieved yield is related to the tissue concentration or is just inherent in the particular variety. Some cautions should be taken when looking at the plant tissue data. In particular, the data presents a particular concentration at a particular time in development. We know that tissue concentration can change significantly during a growing season therefore if a sample is taken it should be known when it was taken in order to compare back to a specific nutrient concentration at a specific time in development of the plant. As plants mature mobile nutrients, such as phosphorus, can be re-mobilized out of the flag leaf therefore tissue levels may decline over time. If samples are collected too late they may return erroneous information. That is why suggestions are made for a specific plant at a specific stage in development. Materials and Methods Nitrogen (N) rate studies: Small plot trials will be conducted at two to three locations in conjunction with the onfarm variety trials. Small plot studies replicated four times will include eight nitrogen rates starting at and ranging in 3 lb increments up to 21 lbs of N per acre. All nitrogen will be applied as broadcast urea incorporated prior to planting. Additional P, K, and S will be applied in order to make sure these nutrients will not limit yield or protein response. For trials in Northern Minnesota a single variety will be planted across locations. Yield will be determined after harvesting the entire plot with a small plot combine and protein will be determined with a NIR. Soil samples will be taken from -6 for P, K, ph and nitrate and nitrate only will be determined on 6-24 samples taken in northern Minnesota and 6-12 and samples collected from southern Minnesota locations. Yield data will be analyzed to determine the amount of nitrogen required to maximize agronomic yield as well as determine economic optimum nitrogen rates based on several combinations of the ratio of the price of N ($ per lb.) to the value of wheat ($ per bu.) and protein discount levels. Plant tissue survey: ed wheat varieties will be sampled at selected variety trial locations. A total of 17 locations were sampled over 211 and 212 with 8 locations sampled in 211 and 9 sampled in 212. The preferred sampling will be conducted at Feekes 9. (prior to boot) in which about 2-3 flag leaves will be sampled out of each plot. However, differences in maturities between varieties may give us a range in maturities at the time of sampling. This sampling is different from the recommended sampling of the whole plant but represents a common analysis being used by farmers. Samples will be dried at 65 o c, ground to pass through a 2mm sieve, and sent in a single batch and analyzed for nutrient concentration. Analysis was conducted for nitrogen, phosphorus, potassium, calcium, magnesium, sulfur, zinc, iron, manganese, copper, boron, and chlorine concentration. At the time of sampling a single composite soil sample was collected from each site (-6 depth) to characterize soil test levels at the sampling time for each location. This would reflect any fertilizer applied in the spring before wheat was seeded at each location. Statistical Analysis: All analysis was completed using SAS. For both studies and analysis of variance was conducted to determine if treatment effects were significant for each study. For the N rate trials, if the analysis indicated that the effect of nitrogen was significant that the yield or protein concentration was regressed with the N rate applied to fit a model to predict protein or yield respose to N at each location. To study effects across locations, the amount of nitrate nitrogen in the two foot soil test was included with the N rate applied and regressed against the yield for all N rate trials conducted from 28 to 212 for both northern and southern Minnesota locations. This data was used to determine the maximum return to nitrogen for several combinations of N price ratio ($ per lb N:$ per bu wheat) and continued on page 56 Page 55

5 discount levels ($ per fifth). The actual yield values for each location were used for the MRTN analysis but the protein data was estimated using the total lbs. of protein produced per acre. Initial analysis was conducted using average protein concentrations across locations but this tended to overestimate protein at low N rates. The current database includes 15 locations from northern Minnesota and 8 locations from southern Minnesota. For the plant analysis study an analysis of variance procedure was used to study the effects of location and variety and their interaction (whether the effects of variety differed by location of vice versa). If the analysis indicates a significant interaction between variety and location normally the effects of variety within individual locations would be further studied. Since our main question was whether nutrient concentrations were similar for varieties across locations we simplified our analysis to look at the relative ranking of varieties within location if there was evidence of a significant effect of variety from the initial analysis of variance. The varieties were ranked from 1 to 14 with 1 being the variety with the highest level and 14 being low. Only 14 varieties were used since the variety was not used both years. In 212 was sampled in place of but both could not be considered across all locations. In addition 1 location,, in 212 was not used since not all varieties were sampled due to some not being mature enough at the time. For the analysis variety means for the nutrient concentration as well as yield and protein concentration were ranked and these rankings were then analyzed using analysis of variance with the particular location being considered a single replication. If the effect of variety were significant it indicates that there are consistent differences across locations. An LSD procedure was used to separate out varieties into general groupings at the.3 probability level. Box and whisker plots are given to summarize concentrations by variety and location. The data within the box represents the median concentration value and the upper and lower lines represent the 25 th and 75 th quartile of the data. The whiskers represent the 1 th and 9 th quartile and any points outside represent high or low outlier values within the plots. The data was presented this way in order to show where the majority of the values occurred and give an idea of the range in values across all locations. Economic Benefit to a Typical 5 Acre Wheat Enterprise The direct benefit from the plant analysis study to a 5 acre operation is less clear that the direct benefit of the nitrogen study. However, if plant analysis is being used there likely will be some recommendation for a foliar application since it is likely that one or more nutrients will be considered deficient. With the variability we have been finding it is likely that one or more nutrients may be low due to environmental or other stresses. The lack of correlation with the soil test data indicates this since the values were generally sufficient. Soil testing should be the preferred option in order to make nutrient application decisions prior to seeding to ensure yield will likely not be limiting. In the nitrogen study, not hitting the correct rate can have significant impacts on profitability. For instance, if discount levels are high ($.2 per fifth), under-application of fertilizer by 25 lbs would result in an average loss of $25 per acre and $5 per acre for 5 lbs under-application. This would be a $12,5 loss for a 5 acre operation and $25, for the 5 lb under-application rate. Over application can also cause significant loss in profit, but the losses would be different since there likely would be less of an issue with decreased grain protein. The major challenge to growers is predicting the appropriate discount levels. An average level for both the price ratio and discount levels could be used but some caution should be taken. The current recommendation, 2.5 x yield goal n credits, can be used but may over-estimate the nitrogen need. However, the appropriate yield goal needs to be selected in order for the calculation to work. Related Research This study is a continuation of nitrogen rate work funded previously by the Minnesota Fertilizer Research and Education Council which ended in 21. The overall goal is to incorporate data from the current study with previous work to develop a N rate response database for spring wheat. In addition to the plant sampling work in this study, a separate grant was awarded to look at nutrient accumulation over the growing season to study how nutrient levels change over time in specific plant parts. In addition, a related project on ear leaf sampling in the corn hybrid trials is underway to study differences between hybrids and locations across the state of Minnesota. Recommended Future Research The current research does show the large potential economic benefit to N application for a high yielding moderate protein producing spring wheat variety. However, future research on varieties with varying protein potentials is key in order to determine if the MRTN approach can be adequately used for spring wheat. For example, the question could be raise if a variety with a higher protein potential may need as much N since it may be easier to achieve 14% protein. Further research would be helpful to determine if MRTN values should be not only adjusted based on the ratio of the price on nitrogen to the value of wheat but also to the protein potential for a variety. Page 56

6 Appendix Table 1. Trial location and planting information for spring wheat N rate studies. Location County Soil Type Soil Texture Variety Crookston Polk Wheatville Sandy Loam Otter Tail Formdale-Buse Clay Loam Kittson Northcote Clay Redwood Normania Loam Nicollet Loam RB7 Table 2. Spring soil test averages across replications for Spring wheat N trials. Soil Test (-6") Location P K SOM ph Nitrate N ppm %--- --lb/ac-- Crookston P, Olsen phosphorus; K, ammonium acetate potassium; SOM, soil organic matter; ph, soil ph. to 2 foot soil nitrate level. Table Yield averages for nitrogen rate treatments for the 212 spring wheat N rate study. Nitrogen Applied (lbs N/acre) Statistics Location avg. Sig. Model MaxN bushels/acre P>F- lbn/ac Crookston <.1 QP QP <.1 QP LP 15 Sig., significant for main treatment effects (nitrogen rate) according to ANOVA; Model, regression model that best fits the data (lin, linear; Quad, quadratic; QP, quadratic plateau; LP, linear plateau; NM, no model); MaxN, N rate where response was maximized (a rate of 18 indicates no maximum was achieved with N rates applied). Page 57

7 Table 4. Summary of estimated nitrogen needs based on maximum yields at each location and economic optimum N rates. Estimated N Need EONR Location Yield 2.5x STN Calculated AONR bu/ac lbs N/ac lb/ac lbs N/ac Crookston Yield, maximum agronomic yield; 2.5x, nitrogen rate at 2.5 x AONR; STN, avearage 2 soil nitrate; calculated, calculated N rate (YGx2.5-Ncredits.). Economic optimum nitrogen rate at specific fertilizer:wheat price ratios based on yield response data. Maximum agronomic nitrogen rate based on yield response. Table grain protein averages for nitrogen rate treatments from the 212 spring wheat N study. Nitrogen Applied (lbs N/acre) Statistics Location avg. Sig. Model MaxN % P>F-- lbn/ac Crookston <.1 QP 26 Fergus Falls <.1 Lin QP <.1 QP 88 Sig., significant for main treatment effects (nitrogen rate) according to ANOVA; Model, regression model that best fits the data (lin, linear; Quad, quadratic; QP, quadratic plateau; LP, linear plateau; NM, no model); MaxN, N rate where response was maximized (a rate of 18 indicates no maximum was achieved with N rates applied). Table 6. Summary of maximum return to N (MRTN) data for yield data summarized over five years in Northern Minnesota. Economic optimum N rate includes N applied in fertilizer + 2 soil test. EONR (.1) EONR (.15) Discount $/fifth low MRTN high low MRTN high lbs total N/ac lbs total N/ac MRTN, Maximum return to N for specified nitrogen price ($/lb):crop price ($/bu); EONR, Economic optimum nitrogen rate. Table 7. Summary of maximum return to N (MRTN) data for yield data summarized for wheat grown in Southern Minnesota. Economic optimum N rate includes N applied in fertilizer + 2 soil test. Page 58 EONR (.1) EONR (.15) Discount $/fifth low MRTN high low MRTN high lbs total N/ac lbs total N/ac not applicable MRTN, Maximum return to N for specified nitrogen price ($/lb):crop price ($/bu); EONR, Economic optimum nitrogen rate.

8 Table 8. Summary of extractable macronutrients and soil organic matter for studies conducted in 211 and 212 Ammonium Acetate Extractable Year Location Nitrate-N P K Ca Mg SO 4- S O.M ppm % Humboldt na na na na na na na na na Table 9. Summary of extractable micronutrients and soil ph. DTPA Extractable Year Location Zn Fe Cu Mn B ph ppm Humboldt Page 59

9 Table 1. Summary of statistical analysis for main treatment effects and their interaction. Main Effect Significance Nutrient Site Line Site x Line P>F Nitrogen <.1 <.1 <.1 Phosphorus <.1 <.1 <.1 Potassium <.1 <.1 <.1 Calcium <.1 <.1 <.1 Magnesium <.1 <.1 <.1 Sulfur <.1 <.1.1 Zinc < Iron Manganese <.1 <.1 <.1 Copper <.1 <.1.68 Boron <.1 <.1.4 Chlorine <.1 <.1 <.1 Effects are considered significant at P<.5 Table 11. Summary of nutrient sufficiency values published in AGR-92 by the University of Kentucky publication for small grains. N P K Ca Mg S Zn Fe Mn Cu B Cl % % % % % % Ppm ppm ppm ppm ppm % to to To to to to To to to to to to Table 12. Pearson correlation coefficients for data collected across all locations and varieties. P K S Ca Mg Zn Fe Mn Cu B Cl YLD Twgt Protein N P K S Ca Mg Zn Fe Mn Cu B Cl Yield Twgt -.24 Page 6

10 Table 13. Pearson correlation coefficients for nutrient concentration in flag leaves and wheat grain yield at all locations. Correlations are not significant (P<.5) when r is between.3 and -.3 (n=45) Prob > r N P K Ca Mg S Zn Fe Mn Cu B Cl locations: 1, ; 2, Humboldt; 3, ; 4, ; 5, ; 6, ; 7, ; 8,. 212 locations: 9, ; 1, ; 11, ; 12, ; 13, ; 14, ; 15, ; 16; ; 17,. Table 14. Pearson correlation coefficients for nutrient concentration in flag leaves and wheat grain protein concentration at all locations. Correlations are not significant (P<.5) when r is between.3 and -.3 (n=45) Prob > r N P K Ca Mg S Zn Fe Mn Cu B Cl locations: 1, ; 2, Humboldt; 3, ; 4, ; 5, ; 6, ; 7, ; 8,. 212 locations: 9, ; 1, ; 11, ; 12, ; 13, ; 14, ; 15, ; 16; ; 17,. Table 15. Pearson correlation coefficients between soil test values and individual nutrient concentration averages for each location. Soil Test SOM ph N na P K Ca Mg S Zn Fe Mn Cu B Page 61

11 Table 16. Summary of average macronutrient, yield, and protein ranks for individual varieties across locations. Low values indicate a variety ranked high across many locations. Variety N P K Ca Mg S Yield Protein RB Table 17. Summary of average micronutrient, yield, and protein ranks for individual varieties across locations. Low values indicate a variety ranked high across many locations. Variety Zn Fe Mn Cu B Cl RB Page 62

12 Dollars per Acre Net Return Crop Value Fertilizer Cost Estimated Discounts Dollars per Acre Net Return Crop Value Fertilizer Cost Estimated Discounts Nitrogen Rate (lbs per acre) Nitrogen Rate (lbs per acre) Figure 1. Summary of MRTN model showing crop value, fertilizer cost, estimated discounts, and net return for the.1 price ratio with no discounts (left graph) and $.2 per fifth of a percent (right graph) in northern Minnesota Dollars per Acre Net Return Crop Value Fertilizer Cost Estimated Discounts Dollars per Acre Net Return Crop Value Fertilizer Cost Estimated Discounts Nitrogen Rate (lbs per acre) -5 Nitrogen Rate (lbs per acre) Figure 2. Summary of MRTN model showing crop value, fertilizer cost, estimated discounts, and net return for the.1 price ratio with no discounts (left graph) and $.2 per fifth of a percent (right graph) in southern Minnesota. Page 63

13 Flag Leaf N Concentration (%) Nitrogen Concentration Flag Leaf P Concentration (%) Phosphorus Concentration 2.5 RB7. RB7 3. Potassium Concentration 1.2 Calcium Concentration Flag Leaf K Concentration (%) Flag Leaf Ca Concentration (%) RB7. RB7 Flag Leaf Mg Concentration (%) Magnesium Concentration Flag Leaf S Concentration (%) Sulfur Concentration. RB7 RB7 Figure 3. Average macronutrient concentrations for varieties averaged across locations. Box plots represent the median and 25 th and 75 th quartile for the data. Whiskers are set at the 1 th and 9 th percentile. Points outside the whiskers represent extreme values across locations.. Page 64

14 Flag Leaf Zn Concentration (ppm) Zinc Concentration Flag Leaf Mn Concentration (ppm) Manganese Concentration 5 RB7 RB7 Flag Leaf Cu Concentration (ppm) Copper Concentration Flag Leaf Fe Concentration (ppm) Iron Concentration RB7 2 RB7 Flag Leaf B Concentration (ppm) 14 Boron Concentration Chlorine Concentration Flag Leaf Cl Concentration (%) RB7 RB7 Figure 4. Average micronutrient concentrations for varieties averaged across locations. Box plots represent the median and 25 th and 75 th quartile for the data. Whiskers are set at the 1 th and 9 th percentile. Points outside the whiskers represent extreme values across locations. Page 65

15 Flag Leaf N Concentration (%) Flag Leaf P Concentration (%) Flag Leaf K Concentration (%) Locations 211 Locations 211 Locations 212 Locations Figure 5. Average N, P and K concentrations for individual locations in 211 and 212 averaged across varieties. Box plots represent the median and 25 th and 75 th quartile for the data. Whiskers are set at the 1 th and 9 th percentile. Points outside the whiskers represent extreme values across locations. Flag Leaf N Concentration (%) Flag Leaf P Concentration (%) Flag Leaf K Concentration (%) Locations 212 Locations Page 66

16 Flag Leaf Ca Concentration (%) Flag Leaf Mg Concentration (%) Flag Leaf S Concentration (%) Locations 211 Locations 211 Locations 212 Locations Figure 6. Average Ca, Mg, and S concentrations for individual locations in 211 and 212 averaged across varieties. Box plots represent the median and 25 th and 75 th quartile for the data. Whiskers are set at the 1 th and 9 th percentile. Points outside the whiskers represent extreme values across locations. Flag Leaf Ca Concentration (%) Flag Leaf Mg Concentration (%) Flag Leaf S Concentration (%) Locations 212 Locations Page 67

17 Flag Leaf Zn Concentration (ppm) Flag Leaf Fe Concentration (ppm) Flag Leaf Mn Concentration (ppm) Locations 211 Locations 211 Locations 212 Locations Figure 7. Average Zn, Fe and Mn concentrations for individual locations in 211 and 212 averaged across varieties. Box plots represent the median and 25 th and 75 th quartile for the data. Whiskers are set at the 1 th and 9 th percentile. Points outside the whiskers represent extreme values across locations. Flag Leaf Zn Concentration (ppm) Flag Leaf Fe Concentration (ppm) Flag Leaf Mn Concentration (ppm) Locations 212 Locations Page 68

18 Flag Leaf Cu Concentration (ppm) Flag Leaf B Concentration (ppm) Flag Leaf Cl Concentration (%) Locations 211 Locations 211 Locations 212 Locations Figure 8. Average Mn, B, and Cl concentrations for individual locations in 211 and 212 averaged across varieties. Box plots represent the median and 25 th and 75 th quartile for the data. Whiskers are set at the 1 th and 9 th percentile. Points outside the whiskers represent extreme values across locations. Flag Leaf Cu Concentration (ppm) Flag Leaf B Concentration (ppm) Flag Leaf Cl Concentration (%) Locations 212 Locations Page 69