Statistical Models for Plant Disease Epidemiology

Statistical models for plant disease epidemiology encompass a spectrum of quantitative approaches designed to characterise, predict and manage outbreaks of pathogens in cropping systems. These models range from mechanistic compartmental frameworks that divide host populations into healthy, latent, infectious and removed classes, to data-driven techniques such as geostatistical interpolation and machine learning algorithms. Spatial models exploit patterns of disease incidence across landscapes, employing tools like kriging or Ripley’s K function to locate hotspots and guide targeted interventions. Temporal or spatio-temporal formulations integrate weather variables, host phenology and pathogen life-history traits to forecast epidemic development and assess the impact of control measures. Recent advances have leveraged high-resolution climate forecasts, field observations and remote sensing to improve predictive accuracy and inform real-time decision support systems. By coupling rigorous statistical inference with practical applications—from early warning of rice blast to optimised fungicide schedules in orchards—these models play a central role in safeguarding global food security and enhancing the sustainability of agricultural production.

Research from Nature Portfolio

Recent studies have applied advanced geostatistical methods to map the spatial distribution of rice blast severity across diverse ecosystems. By analysing multi-year field surveys with spatial autocorrelation metrics and semivariogram-based interpolation, researchers generated predictive maps of disease risk in unvisited locations. Hierarchical clustering and local indicators of spatial association revealed ecosystem-specific vulnerability, highlighting regions where targeted management strategies could most effectively reduce pathogen spread. This work represents the first intensive application of such approaches in the region and demonstrates how integrated spatial models can inform tailored disease control programmes.

Statistical Models for Plant Disease Epidemiology publication trend

The graph below shows the total number of articles in statistical models for plant disease epidemiology across all publications each year (not limited to Nature Index journals).

Technical terms

Geostatistical interpolation: A set of spatial statistics methods, such as kriging, used to estimate values at unsampled locations based on the spatial dependence of observed data.

Semivariogram: A function that quantifies how the variance between paired observations changes with the distance separating them, used to characterise spatial autocorrelation.

Machine learning: Computational techniques in which algorithms learn predictive patterns from data, often without explicit mechanistic assumptions, to forecast disease occurrence.

HLIR epidemic model: A compartmental framework partitioning a host population into Healthy, Latent, Infectious and Removed classes to simulate disease progression over time.

Multiple regression model: A statistical method that quantifies the relationship between a response variable and multiple explanatory variables, producing a predictive equation.

References

1. Rice Blast (Magnaporthe oryzae) Occurrence Prediction and the Key Factor Sensitivity Analysis by Machine Learning. Agronomy (2021).
2. Spatial distribution and identification of potential risk regions to rice blast disease in different rice ecosystems of Karnataka. Scientific Reports (2022).
3. Mathematical Model for Rice Blast Disease Caused by Spore Dispersion Affected from Climate Factors. Symmetry (2022).
4. A disease predictive model based on epidemiological factors for the management of bacterial leaf blight of rice. Brazilian Journal of Biology (2024).