Student or educator? Sign up with your school email and get 6 months of Pro free.

The DOI matches its Crossref record; 47 of the 49 references it cites were found in a registry or an index; 18 of the 50 sources its claims cite were checked for support.

A Bayesian framework to model variance of grain yield response to plant density

Open access

Abstract

BACKGROUND: The expected grain yield response to plant density in winter wheat (Triticum aestivum L.) follows a diminishing returns function. To our knowledge, all previous studies dealing with plant density have assumed constant variance. The gap relies on quantifying the optimum plant density that optimizes grain yield at the lowest risk. Here, we propose a Bayesian hierarchical framework to model the variance of grain yield response to plant density. We demonstrate our framework by identifying the plant density in each seed size, seed treatment and environment combination that maximizes the expected yield and minimizes yield variance. RESULTS: To fit the model, we used data from field experiments conducted in the Canadian Prairies to identify informative priors and Kansas experiments to demonstrate and validate our framework. Kansas experiments were conducted in 25 environments and consisted of a complete factorial combination of three seed cleaning methods leading to three different seed sizes (light, moderate, heavy), two or three seeding rates, and two seed chemical treatments (insecticide + fungicide vs. none). We described both expected yield and variance of yield in response to plant density. The proposed model allowed us to quantify the minimum risk plant density (minRPD), which represents the minimum plant density at which grain yield variance becomes constant. Plant density at the minRPD was always greater than the agronomic optimum plant density (AOPD, i.e.: the plant density that maximizes expected yield); thus, minRPD could be used to estimate the minimum plant density that maximizes expected yield and minimizes yield variance. When compared at the AOPD, four seed cleaning × chemical treatments combinations resulted in similar yield advantages over the control under high and low yielding environments. However, in low-yielding environments, only two cleaning × chemical treatments combinations resulted in smaller variance when compared at the minRPD against the control. All seed cleaning × chemical treatments combinations resulted in similar AOPD. However, two cleaning × chemical treatments combinations had greater minRPD in low-yield environments compared to the control. CONCLUSION: Modeling grain yield response to plant density with the proposed framework is suitable for heteroscedastic data scenarios. Future research may focus on exploring how genotypes, environments and their interaction modulate the difference between AOPD and minRPD and, extend the framework to a variety of processes involving crop management decisions.

Overview

CoChat Summary

TL;DR

The paper proposes a Bayesian hierarchical framework that models both expected grain yield and yield variance as functions of plant density, introducing the minimum risk plant density (minRPD). In Kansas winter wheat trials, minRPD was consistently greater than the agronomic optimum plant density (AOPD), and following minRPD increased net income on average by 37 $ ha−1 in low-yielding and 19 $ ha−1 in high-yielding environments.

Key findings

  • The model separates expected yield and yield variance; minRPD is the minimum plant density at which yield variance becomes constant. In the Canadian Prairies data, minRPD was consistently higher than AOPD (posterior probability = 0.0001), ranging from 138 to 199 plants m−2 versus AOPD from 79 to 128 plants m−2.
  • Across Kansas environments, grain yield averaged 4.10 Mg ha−1 (q.025 = 1.87; q.975 = 7.25) and plant density ranged from 8 to 275 plants m−2. Low-yielding environments averaged 3.37 Mg ha−1, while high-yielding environments averaged 5.59 Mg ha−1.
  • Heavy and treated seeds conferred greater yields at the AOPD compared with untreated light seed, averaging +0.27 Mg ha−1 (q.025 = 0.06; q.975 = 0.47). In low-yielding environments, heavy or moderate seed and treated seed reduced yield variance compared with the control, reducing variance on average by −0.13 Mg ha−1 (q.025 = 0.02; q.975 = −0.31).
  • Net income following minRPD versus AOPD was on average 37 $ ha−1 higher (q.025 = 6; q.975 = 77) in low-yielding environments and 19 $ ha−1 higher (q.025 = −8; q.975 = 54) in high-yielding environments.

Methods

  • A hierarchical Bayesian model was fitted with plot-level grain yield following a normal distribution; expected yield was modeled with a quadratic-plateau function of plant density, and yield variance was modeled as an exponential-quadratic function increasing to a maximum and then remaining constant at minRPD. Informative priors were developed from Canadian Prairies data and previous literature.
  • Data included Canadian Prairies field experiments in 26 environments across three cropping seasons (2010–2013), with two seeding rates (200 and 400 seeds m−2), three seed sizes, two seed treatments, and cultivar CDC Buteo. Kansas experiments covered 25 environments (2018–2021) in a split-split plot design with 18 treatments: three seeding rates (148, 222, 297 seeds m−2), three seed cleaning levels, two seed treatments, and cultivar SY Monument.
  • Environments were classified as high or low yielding using k-means clustering. The model was fitted using a divide-and-conquer approach to 12 subsets corresponding to each seed size × seed treatment × environment combination; posterior medians and 95% credible intervals were calculated, and economic comparisons used a wheat price of 220 $ Mg−1 and seed cost assumptions.

Limitations

  • The paper states that k-means clustering does not account for uncertainty in environment classification and requires post hoc yield data, limiting real-world applicability. It also states that only one cultivar was evaluated, ignoring cultivar × plant density interactions.
  • The authors acknowledge that the Canadian Prairies and Kansas are widely differing environments, but used Canadian-derived priors for Kansas with assumed wide support. The text also notes that environment-specific practices such as sowing date and tillage were not explored for interactions.
  • Evident from the text: the framework relies on specified process forms for expected yield and variance, and no empirical model comparison is presented; the authors argue model selection should be based on biological basis rather than fit criteria alone.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1186/s13007-025-01355-y

Cites

47 of 47 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 Posterior distributions of yield response to plant density model parameters (Eqs. 2 and 3) pooling all 12 combinations of seed treatment, seed size and environment. Derived quantities from these distributions (e.g., the expected value, the standard deviation, and quantiles) were used to build informative priors described in Eqs. 6–10. Purple bars refer to model parameters related to the expected value (μij) and yellow bars to model parameters related to the variance ((σ)ij²) of grain yield. Abbreviations: sample size (n), linear term of expected yield response to plant density (β1), agronomic optimum plant density (AOPD), maximum variance plant density (maxVPD), minimum risk plant density (minRPD), linear term of variance of yield response to plant density (α1)
Fig. 1 Posterior distributions of yield response to plant density model parameters (Eqs. 2 and 3) pooling all 12 combinations of seed treatment, seed size and environment. Derived quantities from these distributions (e.g., the expected value, the standard deviation, and quantiles) were used to build informative priors described in Eqs. 6–10. Purple bars refer to model parameters related to the expected value (μij) and yellow bars to model parameters related to the variance ((σ)ij²) of grain yield. Abbreviations: sample size (n), linear term of expected yield response to plant density (β1), agronomic optimum plant density (AOPD), maximum variance plant density (maxVPD), minimum risk plant density (minRPD), linear term of variance of yield response to plant density (α1)CC BY-NC-ND · PubMed Central
Fig. 2 Panel A shows expected grain yield (μij) in response to plant density (sij, purple); panel B shows standard deviation of yield (derived from yield variance,σij²) in response to plant density (yellow); panel C shows the posterior predictive distribution of grain yield (yij) response to plant density (brown) for the jth treatment × environment combination (in this example untreated, light seeds in low yield environment cluster). Solid black lines represent the median and ribbons the 95% credible interval. Purple refers to parameters or quantities related to modelling the expected value (μij) and yellow to parameters or quantities related to modelling the variance ((σ)ij²) of grain yield
Fig. 2 Panel A shows expected grain yield (μij) in response to plant density (sij, purple); panel B shows standard deviation of yield (derived from yield variance,σij²) in response to plant density (yellow); panel C shows the posterior predictive distribution of grain yield (yij) response to plant density (brown) for the jth treatment × environment combination (in this example untreated, light seeds in low yield environment cluster). Solid black lines represent the median and ribbons the 95% credible interval. Purple refers to parameters or quantities related to modelling the expected value (μij) and yellow to parameters or quantities related to modelling the variance ((σ)ij²) of grain yieldCC BY-NC-ND · PubMed Central
Fig. 3 A Posterior distributions of expected grain yield response to seed management evaluated at the AOPD (Eq. 11); B standard deviation of yield response to seed management evaluated at the minRPD (Eq. 11) in high (circles, opaque) and low (diamond, transparent) yielding environments. Contrasts represent heavy, treated seeds against light seed, untreated control treatment (Eq. 11). Points represent the median and whiskers are the 95% credible interval; p represents the posterior probability of Δμj, Δσj being greater than zero. Purple refers to derived quantities related to modelling the expected value (μij) and yellow to derived quantities related to modelling the variance ((σ)ij²) of grain yield
Fig. 3 A Posterior distributions of expected grain yield response to seed management evaluated at the AOPD (Eq. 11); B standard deviation of yield response to seed management evaluated at the minRPD (Eq. 11) in high (circles, opaque) and low (diamond, transparent) yielding environments. Contrasts represent heavy, treated seeds against light seed, untreated control treatment (Eq. 11). Points represent the median and whiskers are the 95% credible interval; p represents the posterior probability of Δμj, Δσj being greater than zero. Purple refers to derived quantities related to modelling the expected value (μij) and yellow to derived quantities related to modelling the variance ((σ)ij²) of grain yieldCC BY-NC-ND · PubMed Central
Fig. 4 Posterior distributions of AOPD (A) and minRPD (B) for contrasting seed management treatments (no treatment and light seeds, “None_Light”; treated and heavy seeds, “Treated_Heavy”) under high (circles, opaque) and low (diamond, transparent) yielding environments. Points represent the median and whiskers are the 95% credible interval. Purple refers to model parameters related to the expected value (μij) and yellow to model parameters related to the variance ((σ)ij²) of grain yield
Fig. 4 Posterior distributions of AOPD (A) and minRPD (B) for contrasting seed management treatments (no treatment and light seeds, “None_Light”; treated and heavy seeds, “Treated_Heavy”) under high (circles, opaque) and low (diamond, transparent) yielding environments. Points represent the median and whiskers are the 95% credible interval. Purple refers to model parameters related to the expected value (μij) and yellow to model parameters related to the variance ((σ)ij²) of grain yieldCC BY-NC-ND · PubMed Central
Fig. 5 Posterior distributions of AOPD (A) and minRPD (B) as affected by seed management treatments under high (circles, opaque) and low (diamond, transparent) yielding environments. Contrasts represent heavy, treated seeds against light, untreated (control) seeds. Points represent the median and whiskers are the 95% credible interval. Equation 11 describes the calculations used for contrasts between treatments. Purple refers to derived quantities related to modelling the expected value (μij) and yellow to derived quantities related to modelling the variance ((σ)ij²) of grain yield. Abbreviations: agronomic optimum plant density (AOPD), minimum risk plant density (minRPD)
Fig. 5 Posterior distributions of AOPD (A) and minRPD (B) as affected by seed management treatments under high (circles, opaque) and low (diamond, transparent) yielding environments. Contrasts represent heavy, treated seeds against light, untreated (control) seeds. Points represent the median and whiskers are the 95% credible interval. Equation 11 describes the calculations used for contrasts between treatments. Purple refers to derived quantities related to modelling the expected value (μij) and yellow to derived quantities related to modelling the variance ((σ)ij²) of grain yield. Abbreviations: agronomic optimum plant density (AOPD), minimum risk plant density (minRPD)CC BY-NC-ND · PubMed Central
Fig. 6 Posterior distribution of the difference in net income from plant density reccomendations following minRPD compared to AOPD, under (A) low and (B) high yielding environments and varying seed chemical treatment and cleaning methods. Positive ΔNet Income values represent greater net income by following minRPD compared to AOPD (Eq. 12, 13). Points represent the mean and whiskers the 95% credible interval
Fig. 6 Posterior distribution of the difference in net income from plant density reccomendations following minRPD compared to AOPD, under (A) low and (B) high yielding environments and varying seed chemical treatment and cleaning methods. Positive ΔNet Income values represent greater net income by following minRPD compared to AOPD (Eq. 12, 13). Points represent the mean and whiskers the 95% credible intervalCC BY-NC-ND · PubMed Central

Verification

Claims the paper makes

13 Valid · 3 Not valid · 9 Not checked · 37 Own results and interpretations

34 more claims · 2 with their cited sources checked

Create a free account to see all 62 claims and every source verdict.

See all claims — free account
  • Probably not a substantial claim. 17
  • It may state two separate claims as one. 13

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

1 of 1

Citations

Cites

Show all 49 references

Metadata

Authors
Giordano, N., Hayes, D., Hefley, T. J., Lacasa, J., Beres, B. L., Haag, L. A., & Lollato, R. P.
Affiliations
  • Kansas State University (US)
  • Agriculture and Agri-Food Canada (CA)
Corresponding author
  • Giordano, Nicolas
  • Lollato, Romulo P.
Venue
Plant Methods
Published
2025-03-17
Volume
21
Issue
1
Pages
38-38
ISSN
1746-4811 (ISSN-L)
OpenAlex
W4408498771
Cited by
3
References
49
Citations by year
2026: 3
FWCI
3.29
Type
journal article
Language
en
Open access
Licence
Locations
  • Plant Methods (publishedVersion, cc-by-nc-nd)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • Kansas Crop Improvement Association
  • Kansas Wheat Commission
  • Kansas Wheat Cooperative Extension Service
Topics
  • Life Sciences › Agricultural and Biological Sciences › Agronomy and Crop Science › Crop Yield and Soil Fertility
  • Life Sciences › Agricultural and Biological Sciences › Plant Science › Genetics and Plant Breeding
  • Life Sciences › Agricultural and Biological Sciences › Agronomy and Crop Science › Agriculture, Soil, Plant Science
Keywords
winter wheat, plant density, grain yield, yield variance, Bayesian hierarchical model, heteroscedasticity, seed cleaning, seeding rate, yield risk

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-02, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-02 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: A Bayesian framework to model variance of grain yield response to plant density. https://app.cochat.ai/papers/10.1186/s13007-025-01355-y
The paper
Giordano, N., Hayes, D., Hefley, T. J., Lacasa, J., Beres, B. L., Haag, L. A., & Lollato, R. P. (2025). A Bayesian framework to model variance of grain yield response to plant density. Plant Methods. https://doi.org/10.1186/s13007-025-01355-y

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.