Dr. Josh McGrath - Precision applications of nutrients
0 likes53 views
The document discusses the precision management of phosphorus (P) and potassium (K) in agriculture, highlighting the four key activities involved in soil testing: sampling, analysis, interpretation, and recommendations. It emphasizes the importance of high-resolution nutrient management and the need for precise mapping of nutrient needs to enhance fertilizer application strategies. Additionally, it critiques common soil sampling methods, suggesting that more frequent and strategically varied sampling can lead to better fertility recommendations and outcomes.
Dr. Josh McGrath - Precision applications of nutrients
- 1. UNIVERSITY OF KENTUCKY Precision management of P and K: What do we need and what do we have? Josh McGrath Follow @NPK_Professor
- 2. Soil Testing •Four separate activities: 1. Soil sampling 2. Soil analysis 3. Interpretation of results 4. Recommendations 2
- 3. Soil Testing •Four separate activities: 1. Soil sampling 2. Soil analysis 3. Interpretation of results 4. Recommendations 3 • Traditionally viewed as limiting function when managing field average • When planning VR we often grid sample and interpolate between points
- 4. Soil Testing •Four separate activities: 1. Soil sampling 2. Soil analysis 3. Interpretation of results 4. Recommendations 4 • Traditionally viewed as limiting function when managing field average • When planning VR we often grid sample and interpolate between points • We’re pretty good at soil analysis
- 5. Soil Testing •Four separate activities: 1. Soil sampling 2. Soil analysis 3. Interpretation of results 4. Recommendations 5 • Traditionally viewed as limiting function when managing field average • When planning VR we often grid sample and interpolate between points • We’re pretty good at soil analysis • We have a long way to go here
- 6. What do we need and what do we have for VR? •~40 - 70% report VR nutrient management– what’s the agronomic basis? •High resolution characterization of spatially variable nutrient need • Spatial distribution of nutrient availability (soil testing) •Interpretation of soil test results with matching precision •Recommendations developed for VR application •We can vary fertilizer at a pretty fine resolution •We can’t (precisely) map need at the same resolution •We don’t develop recommendations at that resolution 6
- 7. Soil Testing •Four separate activities: 1. Soil sampling 2. Soil analysis 3. Interpretation of results 4. Recommendations 7
- 8. Soil Sampling for Precision Ag 10 13 11 22 7 9 15 3512 45 13 18 ? 10 13 11 22 • Precision ag breaks a field into smaller management zones • Grid Sampling (grid point and grid cell): • Typically involves interpolation of grid sample data • Previous management has altered soil nutrient levels • Combined small fields into large field • Directed sampling: • Break field into zones and collect “average” sample for each zone • Requires other data: yield maps, remotely sensed images, or other sources of spatial data • Requires experience with field
- 9. Most VRP likely based on interpolated grid soil data •Interpolation estimates unknown value between two sample points •Samples must be collected close enough that they are correlated •r > 0.3 •Requires samples on ¼ acre grid or less 9 Lauzon, J.D., I.P. O’Halloran, D.J. Fallow, A.P. von Bertoldi, and D. Aspinall. 2005. Spatial Variability of Soil Test Phosphorus, Potassium, and pH of Ontario Soils. Agronomy Journal 97(2): 524–532.
- 10. How does interpolation perform? •Often field average is closer to true value than coarse sampling (>1/4 acre) •At small scale soil properties tend to be stochastic •Random such that they can be predicted accurately, but not necessarily precisely •On average the estimated values are right, at each spot they are off further than the field average 10 ¼ acre grid 1 acre grid Courtesy John Spargo, Penn State Soil Testing Lab
- 11. High Precision Low Accuracy Low Precision High Accuracy Reminder: Accuracy v. Precision http://en.wikipedia.org/wiki/File:High_precision_Low_accuracy.svg 11
- 12. Soil Testing •Four separate activities: 1. Soil sampling 2. Soil analysis 3. Interpretation of results 4. Recommendations 12
- 13. Soil analysis •Soil testing should provide •Nutrient in soil solution (intensity) and in stored pools (quantity) •Buffer capacity (Q/I). 13
- 14. Soil analysis •Soil testing should provide •Nutrient in soil solution (intensity) and in stored pools (quantity) •Buffer capacity (Q/I). 14 •Multiple factors affect Q/I •Soil testing can only provide an index of nutrient supplying capacity of a soil.
- 15. Soil Chemical Analysis •There are multiple soil tests that use different procedures and chemicals 15
- 16. Soil Chemical Analysis •There are multiple soil tests that use different procedures and chemicals •Each extracts different nutrient amounts and forms from soil pools •Require local correlation and calibration to be useful 16
- 17. Soil Testing •Four separate activities: 1. Soil sampling 2. Soil analysis 3. Interpretation of results 4. Recommendations 17
- 18. Soil test interpretation & Fertilizer recommendations •Correlation •Relative yield versus soil test value •Plant response to application of element • Requires check plot and sufficient plot •Conduct experiment at multiple sites (multiple soil concentrations) 18 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0 50 100 150 200 RelativeYield Soil Test Result Soil critical concentration Relative Yield = 0.95
- 19. Soil test interpretation & Fertilizer recommendations •Calibration •Amount of applied nutrient versus soil test value •Multiple fertilizer rates at one site (one soil test value) •Conduct at multiple sites •Build fertilizer recommendations for different soil test values 19 0 20 40 60 80 100 120 140 160 180 200 0 50 100 150 Yield(bu/acre) Fertilizer Rate (lbs/acre) Yield response to fertilizer rate at different sites Soil Test = 30 Soil Test = 20 Soil Test = 10
- 20. Variability in correlation and calibration results? 20
- 21. Soil Testing •Four separate activities: 1. Soil sampling 2. Soil analysis 3. Interpretation of results 4. Recommendations 21
- 22. Soil Test Calibration: How much fertilizer? 1. Sufficiency approach • When soil test level is below optimum, apply only enough nutrients to meet crop needs 2. Buildup and maintenance approach • Rapidly build low soil test concentrations to optimum level • Replace nutrients removed by crop at higher soil test levels where response is not expected 3. Hybrid Approach 22What about BCSR? -100 0 100 200 300 400 500 600 700 0 10 20 30 40 50 60 70 FertilizerPrate(lb-P2O5/a) Initial soil test P (lb/a) Fertilizer P rate to move STP10 units CornP2O5 Rec
- 23. Calibration: Fertilizer Recommendation Systems 23 0 50 100 150 200 250 0 5 10 15 20 25 30 Fertilizerrate(lb-P2O5/a) Mehlich 3 Soil P (mg/kg) Sufficiency only recommendations Sufficiency 0 50 100 150 200 250 0 5 10 15 20 25 30 Fertilizerrate(lb-P2O5/a) Mehlich 3 Soil P (mg/kg) Example of Build and Maintain Recommendations Buildup Maintenance ActualCriticalLevel ActualCriticalLevel
- 24. How should we make precise recommendations? •Build and maintain ignores soil buffer capacity •To move STP 1 lb/acre •Low STP required 10 – 25 lb- P2O5/acre •Optimum STP or above required ~5 lb-P2O5/acre •Initial STP <6 •600 – 200 lb/a P2O5 to move soil test +10 lb/a •Soils don’t pay interest (Thom and Dollarhide, 2002) 0 5 10 15 20 25 0 50 100 150 200 250 300 Lb-P2O5added/lb-STPchange Initial Soil Test P (lb/acre)
- 25. Soil nutrient storage – buffer capacity 25 This Photo by Unknown Author is licensed under CC BY-NC-ND
- 26. How should we make precise recommendations? •Precision ag -- frequent soil testing and sufficiency rates •We need to know the yield maximizing (sufficiency) rate •Sufficiency rate < build & maintain •Sufficiency probably < crop removal •buffer capacity makes up difference (Thom and Dollarhide, 2002) 0 50 100 150 200 250 0 5 10 15 20 25 30 Fertilizerrate(lb-P2O5/a) Mehlich 3 Soil P (mg/kg) Example of Build and Maintain Recommendations Buildup Maintenance Sufficiency
- 27. Designing soil testing for precision •Four separate activities: 1. Soil sampling 2. Soil analysis 3. Interpretation of results 4. Recommendations •How do we spatially characterize response potential? •Do we need additional points of information besides current soil test? •What would this data look like? •Philosophical approach to precision ag? 27
- 28. How do we move forward? •I would argue we need new research design and analysis methods to get precise •Five P rates, randomized in five blocks, at four locations, for three years… 28 We don’t need more data; we need different data!
- 29. Spatial variability in soil test correlation •Reduce plot size to limit variability •Two treatments: sufficient or none •One phosphorus application rate using APP in 2x2 •29 kg ha-1 P (60 lb/acre P2O5) •56 kg ha-1 N (50 lb/acre) balanced using UAN 29
- 30. Early growth response: All sites • 𝑅𝑒𝑙𝑎𝑡𝑖𝑣𝑒 = )*+,-../ )*+,-..0 ×100% •Biomass in kg/ha •Red line indicates University critical level •Green line indicates 95% Relative Biomass •Majority of the time starter P increased biomass V4-V6 30
- 31. Delta Yield Paired T-Test: Princeton •Delta Yield = YP-Y0 •Found to be highly significant •Mean difference 9.4 bu/a 31
- 32. On average soil test correctly predicted yield response 32 Regardless of soil test we had a yield response in about 50% of the plots 0 20 40 60 80 100 120 140 0 5 10 15 20 25 30 35 RelativeYield Mehlich 3 Phosphorus (mg kg-1)
- 33. Variance decreases as soil P increases 33
- 34. Soil testing for SSM: New challenges 34 0% 20% 40% 60% 80% 100% 120% 0 10 20 30 40 50 60 RelativeYield Soil test phosphorus Where do we place the critical level? •We have focused on mapping soil P status spatially •Correlation and calibration were designed to make accurate recommendations •What if in addition to soil P concentration varying, the critical level varies? •spatially and temporally?
- 35. Grid sampling •Interpolated soil sample maps (>1/4 acre grid) are unreliable AT BEST. •There is nothing wrong with the grid approach, the problem is interpolation of the data •More frequent sampling is better use of money •If you insist on grid sampling, then shift grid over time to get denser sample map •Look at soil test range, median, average, and deviation – just don’t interpolate 35
- 36. Zone sampling •Intensively sampled zones might work •Use topography, soil texture, or even historic grid samples • Look at summary statistics from grid data by zone •Use yield to test zones – but not necessarily to create zones •We’re working on guidance for zone development using free software (e.g. QGIS, MZA) 36
- 37. Future opportunities •Even with a decent soil test map (grid or zone) our recommendations are very coarse and were intended to be an average • Use your technology to insert check strips (High, Low, None) within your prescription and evaluate recommendations yourself •To truly practice precision ag you need to be closer to sufficiency rates 37 0 50 100 150 200 250 0 5 10 15 20 25 30 Fertilizerrate(lb-P2O5/a) Mehlich 3 Soil P (mg/kg) Example of Build and Maintain Recommendations Buildup Maintenance Sufficiency