Asian soybean rust (ASR) can decimate soybean crops and lead to yield losses of up to 80%, as seen in Africa and South America. Gro’s spatial dynamics model determines the probability of ASR spreading within a region, given the infection data of neighboring regions. The existing model framework projects the movement of ASR within Brazil and can be applied to other tropical regions. The Brazil model can also be combined with Gro’s Soy Yield Forecast Model to assess the impact of ASR during the growing season.

Customers Use the Model to: 

  • Estimate the risk of ASR disease outbreak in a region or province 
  • Analyze how the disease is impacting yields  
  • Guide sales and marketing teams on where to provide crop protection products
  • Inform procurement and food safety teams on a crop’s health in key sourcing regions

Why It Matters

Gro first developed the Asian Soybean Rust Model for Brazil soybeans. Gro’s Rapid Response Data Science team organized disparate data and quickly applied it to make sense of a recurring issue which can devastate harvests. ASR occurs in Brazil every year in humid climates with temperatures that are between 14-28 degrees Celsius for at least 6 hours. ASR can be combated with fungicides and other crop protection inputs.

Methodology

By combining historical ASR presence data and the proximity to infection, Gro’s Asian Soybean Rust Model assesses the probability of ASR moving to neighboring regions. The predicted infection level of a region directly depends on other regions around it. By combining historical ASR presence data and spatial proximity between infection hotspots, the model is able categorize each region with one of four risk states, ranging from safe to critical, and thus estimate the likelihood of future outbreaks at the district level. The risk state of a region directly depends on other regions around it, as well as climate conditions. The model can be further tailored to a specific company's operations by training it with their proprietary ASR outbreak data.

Related Analytics: