On‐farm corn phosphorus response reveals importance of soil testing

Concern for declining Great Lakes Basin water quality continues to emphasize improved phosphorus (P) management strategies. Nutrient (predominately P) and algae indicators used gauge were recently highlighted as fair unchanging (U.S. Environmental Protection Agency, 2019). The Tri-State fertilizer recommendations, covering Ohio, Indiana, Michigan, base P applications on the build maintain concept, which accounts soil test value yield potential while providing recommendations at or slightly above a critical (STP) concentration (Culman et al., 2020; Vitosh 1995; Warncke 2009). Critical is concept of an empirically derived point curve that relates level indicator (e.g., STP) with yield. probability corn (Zea mays L.) grain response in mineral soils increases when STP concentrations are deficient below 15 ppm Bray1-extractable (Bray-P1), 20 Mehlich-3 extractable (Mehlich 3-P). fertility framework recommends based crop nutrient removal (0.35 lb P2O5 bu−1) optimal (i.e., 15–30 Bray-P1 20–40 Mehlich Recent revisions consider 30 (40 3-P) sufficient no agronomic reason apply 2020). However, many growers continue excess increase profitability (Dodd & Mallarino, 2005). Some agricultural Michigan have two three times recommended current distribution estimates 19, 23, 58% testing deficient, optimal, sufficient, respectively (TFI, Continued plant uptake increased resulted legacy accumulation can impact years (Powers 2016). In 2018, western Lake Erie's shoreline, open waters, island shoreline declared impaired predominantly from non-point source loading (Ohio EPA, Corn production areas include regions within Saginaw Bay Erie Basins. Assessment stratification may occur due fertilization reduced tillage important where surface exports concerns determined multiple sampling depths (0–2 0–8 inch) NRCS, Trials across locations continued assessment trends affected by cropping system application Growers asked host on-farm trials demonstrate importance determining likelihood application. objective was determine evaluate relation guidelines. Twenty-four conducted 2016–2018 different sites previous history manure Plots between 300 600 ft length. Prior applications, sampled 0-to-2- 0-to-8-inch depths. Each sample composite cores collected randomly each replication. Soils air-dried, ground pass 2-mm sieve, analyzed pH (0–8 inch only) extractant (Table 1). Study represented diversity including concentrations. study arranged randomized complete block design four replications treatments, (+P) without (–P) fertilizer. rate product grower's estimated varied site-years. Field-specific guide practices included non-limiting rates nitrogen (N) balanced treatments if required, potassium (K) harvest data moisture. Grain adjusted 15.5% Analysis variance (ANOVA) using Proc Mixed procedures SAS 9.3 (SAS Institute, Cary, NC) site-year, utilized random location rep effects (p ≤ .10). UNIVARIATE procedure Levene's validate normality homoscedasticity residuals, .05). Relationships investigated Pearson product-moment correlations Relative yields (RGY) calculated per site-year mean zero-P plots divided fertilized plots. Cate–Nelson (Cate Nelson, 1971 ) divide into groups identity RGY maximized sums squares “rcompanion” package R (R Core Team, 2015; Mangiafico, package, studies, does not rely user-defined specified threshold but divides horizontally minimizes number observations occurring error quadrants (Fulford Culman, 2017). At depth, results 24 site-years indicated levels sites, eight 14 9 site-years, 3–18 bu acre−1 1, 4, 4 respectively. When maintenance range should expect A +P observed 50% (19–29 illustrating sustaining reduce risk loss. No be expected aboveground cool wet conditions limit diffusion early-season requirements (Havlin 2014). While one non-responsive despite non-significant trend treatment = .16). Mean site 133 grower >250 acre−1, other yield-limiting factors lack precipitation) contributed response. illustrate thresholds emphasizes rates. Surface-applied fertilizers residues combined limited mixing has stratification, non-uniform (Mallarino 1999). analyses identified similar values regardless depth. 0-to-2-inch (Figure analysis 11 16 (RGY 90%) same sum flexibility separating (11 presented; Figure 2). 95% sufficiency been previously suggested, 90% also minimized Blackmer, 1992). Shallower did improve relationship relative coefficients determination (r2 .34) low graphical relationships. study, Fulford Culman (2018) suggested more P-deficient validity analyses. We identify make statistical comparison value. Additional could help precision identifying new occurred. However widespread, often growth-limiting prevent mask Bray-P1, convolute results. grown (15 ppm, Bray-P1) provides validation Grower operations located critically watersheds still variable physical chemical property limitations. indicate depth require adjustment values. Validating foundation development effective strategies systems. Results illustrated recommendation guidelines loss insufficient supply. Moreover, they supported published recommendations. Current suggest managers conduct least every 3–4 part comprehensive program Our emphasized considered prior making decisions. authors thank USDA National Institute Food Agriculture, Marketing Program State University AgBioResearch, College Agriculture Natural Resources, Project GREEEN partial funding support this project. addition, John Mackson, Martin Nagelkirk, Bob Battel, Eric Anderson, George Silva, Marilyn Thelen, Phil Kaatz, Paul Gross their assistance collection. declare conflict interest.