📄 plant biology

Yield losses associated with peanut smut incidence in Argentina: a quantitative synthesis across field studies

This study synthesizes data from 26 field trials in Argentina to quantify the linear relationship between peanut smut incidence and yield loss, revealing a consistent damage coefficient of 0.74–0.87% yield reduction per 1% increase in disease incidence regardless of environmental variability.

Original authors: Cazon, L. I., Gonzalez, N. R., Del Ponte, E. M., Costa de Carvalho, A. C., Asinari, F., Camiletti, B. X., Paredes, J. A.

Published 2026-08-11

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Cazon, L. I., Gonzalez, N. R., Del Ponte, E. M., Costa de Carvalho, A. C., Asinari, F., Camiletti, B. X., Paredes, J. A.

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

Imagine you are a farmer, and your field is a bustling city of peanut plants. In this city, there is a sneaky, underground troublemaker: a fungus called peanut smut. You can't see it at first because it hides inside the soil, waiting for the peanut plants to send their roots down. Once it gets in, it turns the tasty peanut kernels into useless, powdery spores. It's like a ghost that steals the food right out of the pantry. Farmers in Argentina have been fighting this ghost for decades, but they've always had a big question: "If I see 10% of my peanuts are infected, how much money am I actually losing?" Is it a small scratch on the car, or is the whole engine gone? This paper is like a giant detective story where scientists gather clues from hundreds of different fields to build a single, super-accurate map of exactly how much damage this fungus causes. They want to know if the damage happens in a straight line (like a slow leak) or if there's a sudden "tipping point" where everything goes wrong at once.

The researchers, led by Luis Cazón, decided to solve this mystery by looking at a massive pile of data: 922 snapshots from 26 different field studies in Argentina, collected between 2021 and 2025. Instead of just guessing, they used a statistical tool called a "meta-analysis," which is like taking 26 different maps and merging them into one giant, high-definition map to see the true picture. They wanted to find a simple rule that connects the percentage of infected peanuts (incidence) to the total weight of the harvest (yield).

Here is what they found: The damage is real, consistent, and surprisingly predictable. For every 1% increase in the number of infected peanut pods, the farmers lose about 24 to 29 kilograms of peanuts per hectare. To put that in perspective, if you have a field that could produce 3,000 kilograms of peanuts, a 1% infection rate knocks off about 0.8% of your total harvest. It's a steady, linear leak.

Now, here is the twist that the paper explicitly rules out. When the scientists first looked at the raw data, the graph looked a bit wobbly. It almost seemed like there was a "breakpoint" around 12% infection. They wondered: "Maybe the peanuts can handle a little smut, but once it hits 12%, the whole system collapses?" They tested this idea with a special mathematical model called a segmented regression. While the math said this "breakpoint" existed statistically, the deeper investigation proved it was a trick of the light. The paper argues that this apparent "tipping point" wasn't a biological rule of the fungus at all. Instead, it was caused by mixing together fields that had very different potential harvests. Some fields were naturally high-yield (like a rich, fertile valley), and others were low-yield (like a rocky hillside). When you mix them all together, the graph looks like it has a kink, but when you look at each field individually, the damage is actually a straight, steady line. The fungus doesn't get "angrier" or more destructive at 12%; it just steals the same amount of yield relative to the size of the harvest, no matter how big or small that harvest is.

The study confirms that the fungus is a consistent thief. Whether the field is producing 1,370 kg/ha or 5,409 kg/ha, the fungus steals roughly the same percentage of the crop for every percent of infection it causes. The authors suggest that this steady relationship is great news for farmers because it means they can predict their losses with confidence. If they know their infection rate is 20%, they can calculate their expected loss without worrying about a mysterious, sudden collapse of the crop. The paper concludes that the biggest difference between fields isn't how the fungus behaves, but how much the field could have produced if the fungus wasn't there. The fungus is a steady, predictable penalty, not a chaotic monster that changes its rules based on the weather or the soil.

Technical Summary: Yield Losses Associated with Peanut Smut Incidence in Argentina

Problem Statement
Peanut smut, caused by the soilborne fungus Thecaphora frezzii, is the most significant disease constraint to peanut (Arachis hypogaea L.) production in Argentina, particularly in the province of Córdoba. While the disease is known to cause substantial yield reductions (up to 35% in highly affected fields) and has achieved 100% field prevalence in core production areas, quantitative estimates of yield losses across diverse environments remain limited. Existing literature lacks a systematic, meta-analytic evaluation of the relationship between disease incidence and kernel yield, which is necessary for robust economic assessment and management decision-making.

Methodology
The study synthesized data from 26 independent field studies conducted between 2021 and 2025 in the core peanut-producing regions of Córdoba (General Deheza and General Cabrera). The dataset comprised 922 observations, where each observation represented a 1 m² sampling unit evaluated for disease incidence (percentage of infected pods) and kernel yield (kg ha⁻¹, standardized to 9% moisture).

The analysis employed a multi-faceted statistical approach to quantify the incidence–yield relationship:

Linear Regression: Separate models were fitted for each study to estimate study-specific intercepts (attainable yield) and slopes (damage coefficients).
Exploratory Segmented Analysis: A LOESS smoothing function and segmented regression were used to test for non-linearity or breakpoints in the relationship.
Meta-Analysis: Random-effects meta-analytic models were used to synthesize study-specific regression coefficients, quantifying between-study heterogeneity using Cochran's Q, I², and $\tau^2$ .
Moderator Analysis: Studies were stratified by disease incidence (using an exploratory 12% threshold) and yield levels to test if damage coefficients varied across strata.
Mixed-Effects Modeling: Linear mixed-effects models with random intercepts and slopes were fitted to account for the hierarchical structure of the data (observations nested within studies), simultaneously estimating population-average effects and study-specific variability.

Key Results

Global Relationship: A consistent negative association was found between peanut smut incidence and kernel yield. The global linear regression estimated an average yield reduction of 26.9 kg ha⁻¹ for every 1% increase in disease incidence.
Damage Coefficients: Across all analytical approaches, the damage coefficient (slope) ranged from 24.2 to 28.7 kg ha⁻¹ per 1% incidence increase. When expressed relative to attainable yield, the relative damage coefficient ranged from 0.74% to 0.87%. This indicates that each 1% increase in incidence reduces yield by approximately 0.8% of the attainable yield.
Attainable Yield Variability: In contrast to the stability of the damage coefficient, attainable yield (intercept) varied markedly among studies, ranging from 1,370 to 5,409 kg ha⁻¹. Meta-analysis estimated a population-average attainable yield of 3,281 kg ha⁻¹ under negligible disease pressure, with extremely high heterogeneity ( $I^2 = 97.98\%$ ).
Linearity and Breakpoints: While exploratory LOESS and segmented regression suggested a potential breakpoint near 12% incidence, subsequent moderator analyses, study-specific regressions, and normalized response curves provided no evidence of a biologically meaningful change in the damage coefficient. The apparent non-linearity in pooled data was attributed to variability in attainable yield among studies rather than a threshold effect in the disease damage function.
Consistency: Mixed-effects modeling confirmed that variability in slopes (damage) was comparatively small (SD = 6.2) relative to variability in intercepts (SD = 874), reinforcing that the disease exerts a proportionally similar penalty across environments.

Significance and Claims
The paper claims to provide the first quantitative synthesis of peanut smut yield losses across field environments in Argentina. Its primary contribution is the establishment of a stable, population-level damage function that can be used for yield-loss assessment.

The authors assert that:

Stability of Impact: Peanut smut reduces yield at a relatively constant rate (approx. 0.8% relative loss per 1% incidence) regardless of the environmental context or baseline productivity.
Operational Utility: The derived damage function (approx. 240–290 kg ha⁻¹ loss per 10% incidence increase) offers a biologically interpretable metric for estimating economic losses and evaluating management strategies.
Incidence as a Predictor: Due to the high proportion of severely damaged pods in infected plants, disease incidence serves as a robust and operationally efficient indicator for predicting yield penalties, potentially obviating the need for more complex severity scales in large-scale monitoring.
Methodological Value: The study demonstrates the utility of meta-analytic and mixed-effects approaches in disentangling true disease effects from environmental noise in highly variable agricultural systems.

The authors conclude that these findings support the development of evidence-based management strategies and provide a quantitative basis for economic assessments, while acknowledging limitations regarding the geographic concentration of the data and the observational nature of the source studies.

More like this