Can soil health metrics improve standard soil fertility recommendations?

Authors: SVEDIN, JEFFREY - University Of Missouri; Kitchen, Newell; Ransom, Curtis; Veum, Kristen; ANDERSON, STEPHEN - University Of Missouri

Submitted to: North Central Extension Industry Soil Fertility Conference Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 11/1/2021
Publication Date: 11/17/2021
Citation: Svedin, J., Kitchen, N.R., Ransom, C.J., Veum, K.S., Anderson, S. 2021. Can soil health metrics improve standard soil fertility recommendations?. In: Proceedings North Central Extension Industry Soil Fertility Conference, November 17-18, 2021, Des Moins, Iowa. Available: https://www.greatplainssoilfertility.org/proceedings/?action=abstract&id=8591&title=Can+Soil+Health+Metrics+Improve+Standard+Soil+Fertility+Recommendations%3F+&search=types

Interpretive Summary: Corn requires phosphorus (P) and potassium (K) to complete its life cycle. To avoid nutrient deficiency, farmers rely on soil sampling and soil fertility laboratory tests to measure the concentration of nutrients left in the soil. These tests at times are not very accurate in helping farmers known how much fertilizer to apply. They may be inaccurate because they do not consider soil biological factors. Increased soil biological activity helps with nutrient cycling. Therefore, on fields that are actively improving soil health, and therefore increasing biological activity, less fertilizer may be required. The purpose of this study was to determine if soil health and soil fertility together could better explain when con grain yields respond to P and K fertilization. From 2018 to 2020, 532 fertilizer response plots on farmers’ fields in Northern Missouri were used to evaluate the grain yield response to P and K fertilizer. Using machine learning data analysis methods, we found that soil health factors were only minimally helpful in explaining when corn would respond to fertilization. Current soil testing methods alone were the most helpful. While improving soil health and biological activity may still provide additional agronomic and environmental benefits, this study shows that other factors may need to be considered (e.g., current management practices, cropping systems, and soil factors) to help farmers improve soil fertility recommendations.

Technical Abstract: It is speculated that integrating soil health (SH) testing with soil fertility (SF) testing would improve fertilizer recommendation decisions. However, quantified impacts of SH properties, specifically soil biological properties, on fertilizer demand have not been well established. The objective of this research was to explore corn (Zea mays L.) yield response to phosphorus (P) and potassium (K) fertilization as influenced by established SF analysis and common SH metrics. From 2018 to 2020, 532 fertilizer response plots (148/m) were implemented in 84 producer fields across central Missouri. Response plot treatments were 1) an unfertilized control, 2) 112 kg/ha of K2O, and 3) 112 kg/ha of P2O5. Each treatment received the same producer-specific nitrogen (N) rates, with an additional 45 kg N/ha applied near V6 corn growth stage to prevent N deficiencies. Random forest analysis was used to model yield response to P and K fertilization and to investigate the influence of SH and SF analysis on model performance. Two-thirds of established monitoring sites were below established P and K soil-test critical concentrations—with 32% and 36% of the low fertility plots responding to P and K fertilizer application. The most consistent yield responses occurred in established “Low” and “Very Low” fertility ratings, with up to 56% of these monitoring sites. However, integrating SH and SF for predicting yield response was only minimally helpful, resulting in r^2 values of 18% and 28% for the P and K treatments, respectively. The low r^2 values are likely due to the variability in P and K availability and crop demand introduced by the diversity of cropping systems, management practices, and soils in which the plots were deployed. Assessment of variable importance in the models indicated that the established University of Missouri recommended SF tests best predicted grain yield responsiveness to P and K fertilization. The addition of SH metrics provided minimal additional predictive power. Although improved SH may offer multiple environmental or agronomic benefits, this study indicates that across central and northern Missouri soils, established SF analysis remains the most effective tool to guide P and K fertilizer decisions in corn production.