Validation of Soil-Test-Based Phosphorus and Potassium Fertilizer Recommendations for Rice and Soybean.

摘要

The science of soil-testing for nutrient management and fertilizer recommendations is widely accepted among scientists and agronomists. Although this science is unsurpassed in predicting soil nutrient availability, soil-test interpretations are seldom validated. Major research objectives for irrigated soybean [Glycine max (L.) Merr.] and direct-seeded, delayed-flood rice (Oryza sativa L.) were to: i) validate the accuracy of Mehlich-3 soil-test P (STP) and K (STK) interpretations and ii) published critical tissue-P and -K interpretations in predicting the yield response to fertilizer at different significance levels (p≤0.05 to 0.25), iii) examine how seed nutrient concentrations are influenced by fertilization and crop response to fertilizer, and iv) evaluate how soil samples from two different depths (0-10 and 0-30 or -45 cm) influence crop response to fertilization. Seventeen (soybean) and 16 (rice) total field experiments were conducted in 2013 and 2014. Soil-P interpretations showed an overall accuracy of 40 (p≤0.05) to 55% (p≤0.25) for soybean and 35% (p≤0.05 and 0.25) for rice in predicting the yield response to fertilizer. Overall accuracy of STK interpretations regarding yield response predictions was 71 (p≤0.05) to 84% (p≤0.25) for soybean and 14 (p≤0.25) to 20% (p≤0.05) for rice. Complete evaluation of critical tissue-P (soybean) and -K (rice) concentrations were unfeasible due to the lack of site-years in deficient categories. Rice tissue-P concentrations were 50 (p≤0.05) and 39% (p≤0.25) accurate in predicting the fertilizer effect on yield. Yield response to fertilizer was accurately predicted at 48 (p≤0.05) to 62% (p≤0.25) of the sites by tissue-K concentrations of soybean. Seed nutrient concentrations of soybean were influenced by fertilizer more frequently and to a greater measure compared with rice. Regression analysis of soil nutrient indices of the shallow and deep soil depths for both crops were relatively strong with a slope close to 1.0 suggesting nutrients at deeper depths had minimal effects on the yield response of either. With exception to STP in rice, nutrient concentration interpretations of soil and tissue for both crops and nutrients were affected by the significance level used to evaluate the data. Tissue concentrations were an inferior (soybean) and superior (rice) predictor of crop response to fertilizer.