1 2 This is what we see in the field. Where we apply no fertilizer N - PDF document

1

2

This is what we see in the field. Where we apply no fertilizer N we see significant differences in rice grain yield, but we have no way of predicting when and where this will happen. This is why we need to develop a N soil test. 3

4

We know that rice roots can go much deeper than most people think. Unlike the non ‐ mobile nutrients P and K, where a 4 ‐ 6 inch depth sample is sufficient with a mobile nutrient such as N, we have to soil sample to the rice plants effective rooting depth. Which is the depth from which the rice plant is able to take up N. 5

6

Here is why we have to sample to an 18 in soil depth. In this calibration curve we only sampled the soil to a 6 inch depth. The line that is developed suing the 6 inch sample depth can only account for 36% (R 2 =0.36) of the variability in the data, which means we are about 36% accurate in our recommendation. 7

Notice when we sample to 12 inches our R 2 increases to 0.61 (about 61% accuracy). Also notice how the points start to spreading out or getting closer to the calibration line. 8

Here we have sampled down to 18 inches look how the points have spread out and gotten even closer to the calibration line and our R 2 has increased to 0.89, which means that we are about 89% accurate in our N rate recommendation. When we sample to 24 inches the R 2 or accuracy of the calibration curve decreased which indicates that 0 ‐ 18 inches is the correct soil sampling depth for silt loam soils. 9

Here is the completed calibration curve, with all of the site ‐ years of data included. 10

These graphs indicate the benefits of using a 18 inch sample depth and the influence on field variability. When we sample to a 6 inch depth we can see roughly 3 management zones with a range of 115 ‐ 145 units of N per acre. Notice when we sample the 0 ‐ 18 inch depth our units of N per acre only varies from 100 ‐ 115 indicating that the field is not as variable as the 6 inch sample would suggest and would result in an erroneous N rate recommendation. 11

We have now completed the N ‐ STaR calibration curve for rice produced on clayey soils and found a similar accuracy of the method and that a 12 inch sample depth gave us the highest predictive ability. 12

Farmers have different N management philosophies. Some use the rates that we currently recommend, while others apply more and some apply less. For those who are concerned about lodging and disease or the price of N fertilizer they would probably choose the 90% Relative grain yield (RGY) recommendation. Whereas those farmers who want to ensure they produce maximum yield would most likely choose the 95% RGY curve, which is what we recommend. Those that are not concerned with fertilizer cost, lodging or disease and want to ensure that N will not limit rice yields will probably want to choose the 100% RGY curve curve. 13

Once we had completed the calibration curves we felt that it was important to validate the calibration curves since this is the first soil test of its kind and there is no precedent for this type of soil testing procedure. Validation studies are designed to test the ability of the calibration curves to predict the correct N rate over a wide range of soils with varying levels of native soil N. To do this we compared the standard recommendation for silt loam soils of 150 units of N per acre to the N rates recommended using the three N ‐ STaR calibration curves. 14

This table highlights the differences in native soil N and the resulting N ‐ STaR rate recommendations that were observed in the validation trials. Under the 90% calibration curve recommended N rates you will notice a range from 20 to 160 units of N per acre indicating the variability that exists within our silt loam soils and the need for a soil ‐ based N test to predict field ‐ specific N rates. Also notice the differences in the three rate recommendations (90, 95 and 100% RGY) within a location. There is a 25 unit of N per acre difference between the 90 and 95% RGY recommendations and there is a 35 unit difference between the 95 and 100% rate recommendations This indicates that the last 5% and 10% between the 95 and 100% rate recommendations. This indicates that the last 5% and 10% of yield is the most expensive to gain in regards to N fertilization. 15

This table indicates the rice yields obtained when N ‐ STaR rates were compared to the standard recommendation. Prod ‐ H is the prime example of the accuracy of the N ‐ STaR rate recommendation. The yield obtained using the 100% RGY curve equaled the standard recommendation. Also notice that when the 90 and 95% RGY recommended N rates were applied they resulted in grain yields equal to 90 and 95% of the maximum for this location. The LHRF had a substantial level of native soil N and all three calibration curves recommended lower N rates than the standard recommendation. Notice that grain yields were maximized when the N ‐ STaR 95% N rate recommendation was used Lets look at this were maximized when the N STaR 95% N rate recommendation was used. Lets look at this location a little closer and see what the rice plots actually looked like. 16

Here is an example of what the rice plots looked like at midseason when the N ‐ STaR and standard N rate were applied. Notice that we obtain the same grain yield with 95 units of N as we did with 150 units of N, but the 150 unit plots had a much denser canopy that could result in mutual shading and an increase in fungal disease. Whereas the 95 units of N per acre plots had more open canopy to allow air flow and sunlight penetration and reduce the potential for fungal disease and the need for fungicides. Consequently, rice does not have to be tall and dark green in order to maximize rice yield. This demonstrates the need to educate the rice industry on what rice needs to look like in order to maximize yields educate the rice industry on what rice needs to look like in order to maximize yields. 17

18

N ‐ STaR recommended 30 units of N less than the producer had intended to apply and resulted in similar grain yields. The reduced N rate resulted in a higher profitability for the producer based on the N rate savings. 19

Similarly but more dramatic this producer was able to cut his N rate by 105 units of N per acre utilizing N ‐ STaR and realize an increase in yield. The extremely high native N at this location resulted in lodging and excessive disease pressure when the standard N rate was applied and therefore lowered his yield when he used the standard recommendation. The increased profitability at this location is a combination f the N rate savings as well as an increase in yield due to the application of the correct N rate to maximize yield using N ‐ STaR. 20

We have mentioned the ability of N ‐ STaR and the potential for N rate savings, however there are fields with low levels of native soil N that could benefit from increased N rate applications and N ‐ STaR has the ability to identify these fields as indicated in the table. This producer was able to increase his grain yield by 18 bushels per acre by increasing his N rate from 150 to 165 units of N per acre and resulted in an increased profitability of $116 per acre. 21

These are the counties where N ‐ STaR generally recommended a lower N rate than the standard recommendation. These are the Grand Prairie soils that are naturally more productive and have more native N than the forest soils in north central Arkansas. 22

These soils north of 1 ‐ 40 and west of Crowleys ridge generally come out of forest and have lower native N than our prairie soils. Resulting in N rate recommendations equal to or more than the standard recommendation. 23

24

This slide is typically what we measure in our soils where the surface soils have the highest levels of native N and native soil N decreases as depth increases. Because the native soil N varies with depth it is important that the proper depth is sampled. Too shallow sampling depths will result underestimation of N fertilizer needs. Sampling depths that are too deep can result in N rate recommendations that are too high. 25

26

These are the requirements that producers need to keep in mind when they use the N ‐ STaR program. In order to use this technology we have to ensure that we are maximizing our N use efficiency and consistency. 27

28

29

30

31

N ‐ STaR sample bucket. 32

33

34

35