Enhancing sunflower head rot assessments with a purpose-built visual tool

Article published on AgriOrbit on 13 August 2025

Sunflower head rot (SHR), caused by Sclerotinia sclerotiorum, reduces yield and seed quality. In South Africa, SHR affects key sunflower regions such as North West, the Free State, and Mpumalanga, especially under cool, wet conditions during flowering and grain filling. Understanding its impact remains difficult without a consistent method to score head rot intensity, making accurate and standardised measurement essential for guiding research and management efforts.

Accurate and precise measurements of disease help estimate yield losses, evaluate control strategies, support predictive modelling, inform management decisions, and justify funding by demonstrating economic impact.

Researchers often focus on incidence, which is a count of how many plants are diseased out of the total number of plants evaluated. However, this only provides a partial view of the disease’s impact. For example, high incidence does not always translate to high infection and yield loss if the affected plants have low severity.

From field trials we conducted in 2023/24, there was a moderate negative correlation between the incidence of sunflower head rot and the yield of sunflowers (Figure 1). Measuring severity, which is how much tissue of the sunflower head is infected, may give us a clearer understanding of the actual impact compared to incidence.

Accurate and consistent data 

This leads to the practical question: How do we turn visual perceptions of disease into usable data? The process of visual assessment depends on an individual’s judgement to estimate disease severity based on what they see, drawing from their perception experience and sometimes requiring a rough guess of how much tissue is infected. Often, this is the quickest method for performing disease severity assessments during experimental evaluations.

While this is an advantage of visual estimates, they are prone to errors and compromise reliability, which is the extent to which repeated estimates under different conditions produce similar results. Differences can occur between raters (known as inter-rater reliability) when estimating disease severity on the same affected plant, reducing consistency between raters. Additionally, a single rater might show greater variation in their ratings of the same plant organ over time (known as intra-rater reliability), which impacts their individual consistency.

This inconsistency can stem from the subjective interpretation of symptoms, such as lesions’ size, shape, or colour, which can vary even on the same plant organ. The experience and training levels of raters in identifying symptoms and giving a disease estimation can influence the accuracy of their ratings. Other factors, such as lighting conditions at assessment times or fatigue during long hours of disease rating, can reduce the consistency of estimates.

These challenges highlight the need for an improved method to ensure disease severity is more accurately and consistently estimated, especially when symptoms are as variable as those in SHR. To meet this need, standard area diagrams (SAD) offer clear, standardised reference points of severity visualised as illustrations or photographs to support consistency across different individuals.

SADs minimise subjectivity in visual assessments, making plant disease evaluations more reliable and comparable across raters and studies. Improved consistency helps increase accuracy, which is how close estimates are to actual severity. It also enhances precision by bringing estimates closer to the actual value and reducing bias, such as over- or underestimation.

Utilising standard area diagrams 

Developing a SAD must follow a purposeful and evidence-based process grounded in real disease expression to ensure these benefits in practice. Plant organs displaying a range of disease severity levels are photographed and then measured for actual disease severity using computer-based image analysis tools, such as ImageJ (Figure 2). Based on the frequency and distribution of observed severity, selected images are arranged to represent incremental severity values likely to be encountered in the field.

We designed the SAD for SHR using 212 photographs, each analysed for actual tissue damage using ImageJ. The selected images were then arranged to represent a scale of disease severity as shown in Figure 3. To illustrate the effectiveness of the SAD, we validated it by asking 16 individuals to rate the severity of head rot, first without a visual aid and then using the SAD.

Once a SAD has been validated, the focus shifts to ensuring it is correctly utilised in practice. Start by visually inspecting the diseased area on the sunflower head you are assessing. Then, compare it to the reference images in the SAD, each representing a specific, predefined severity level. Choose the image with the severity value that most closely matches the severity observed in your sample. (Although SADs are valuable tools for improving disease assessments, they should be accompanied by proper training on recognising the specific disease symptoms and understanding how to apply the scale.)

Conclusion 

The designed SAD will be utilised in future field evaluations, supporting fungicide evaluation efforts, and improving our understanding of how severity levels relate to yield outcomes. This data will be incorporated into a decision-support tool to help producers evaluate the return on investments when choosing the timing of intervention. Beyond these applications, the SAD holds potential value for plant breeders, researchers, and extension personnel seeking a method to evaluate disease more consistently across environments and with many raters. – Kwanele Sabela and Lisa Rothmann, University of the Free State