Brazil is the largest soybean producer in the world, surpassing the United States in 2018. In the past 40 years, Brazilian soy yield has increased more than twentyfold.
Gro’s Brazil Soy Yield Model is updated daily, providing in-season yields at the district, province, and national levels throughout the growing season.
The model’s yield predictions can be paired with Gro’s forecasts for soybean harvested area or official CONAB area data to estimate Brazil’s final soybean production numbers before state agencies collect farmer surveys and planting data.
Gro’s Brazil Soy Yield Model, combined with our United States and Argentina Soy Yield Models, enables Gro to forecast more than 80% of global soybean production in a given year, providing users with a highly accurate, real-time window into global soybean supply.
How the Model Was Built
The fundamental concept behind the Brazil model is similar to our Argentina and United States Soy Yield Models — it produces district-level estimates and aggregates them to the national and provincial levels. Brazil’s soil and weather characteristics are also similar to Argentina, allowing us to leverage our prior work.
Significant expansion in Brazilian soybean acreage in recent years meant we had to model both acreage and yield to properly estimate the country’s total soy production. There are also significant differences in climate, soil condition, and technology adoption throughout Brazil, resulting in high spatial variation of crop yields.
Gro’s Brazil Soy Yield Model uses the following variables:
- Yield, area, and production history in Brazil
- 8-day normalized difference vegetation index (NDVI)
- Evapotranspiration difference from 10-yr median (2003-2013)
- Land surface temperature
- Weather forecasts of air temperature and precipitation
- Soil characteristics
- Latitude and longitude
- 30-meter resolution cropland masks