Gro’s East Africa Cereal Yield Forecast Models, like our other yield models, provide in-season forecasts of crop yields. The 10 East Africa models provide country-level forecasts for five cereals — wheat, corn, sorghum, barley, and millet. The models take into account the unique crop cycles and the areas where they are grown, along with current weather and environmental information.
Customers Use the Model to:
- Predict future trends such as crop yields and production, agricultural input demand, and significant weather patterns
- Gauge crop availability and crop prices
- Understand how weather impacts yields in microclimates
- Inform other models focused on damage caused by pests and diseases
- Monitor crop production, the most significant part of a balance sheet, by tracking its most variable component — in-season yield
- Monitor crop health across the East African region, keeping sight of the implications of the locust outbreak
Why It Matters
Gro developed the East Africa Yield Forecast Models along with the Locust Vegetation-Impact Model in response to the 2020 locust invasion, which started in the Horn of Africa and spread as far as India and Pakistan. By May, the FAO estimated that locusts threatened the food security of more than 42 million people across 10 countries. Having a reliable in-season yield model (along with a separate model for the area impacted by locusts) provides information that can help prioritize food distribution and aid efforts in situations like the 2020 locust infestation. More generally, whether it’s for a locust infestation, a drought, or planning for future growth, these models provide automated real-time predictions on a daily basis, providing critical information and enabling decision-making in regard to the supply side of these major crops in this region.
To forecast yields, we use machine-learning models driven by inputs reflecting long-term trends as well as in-season changes. High resolution satellite data is transformed into robust district-by-district signals uniquely adapted to the situation using domain expertise. The models run daily throughout the growing seasons. The yield model for each crop/country pair is trained using approximately 4 million data points from the Gro Platform, which together represent 40 million points of raw data from nine different sources. These consist of:
- Historical annual yields for each crop from up to three different sources
- Land cover area for each crop at the district level
- Eight-day NDVI
- Daily land surface temperature
- Daily rainfall anomaly
- Monthly evapotranspiration anomalies
- Crop calendars
During model training, relevant time series consisting of millions of data points are selected. Selected series are combined to form input signals: For example, daily rainfall is combined with corn land cover area at the district level to form a corn-area-weighted rainfall input signal.