The pressure imposed to the agricultural sector to increase the production of food, fuel and fiber to cover evolving demands, requires a rethinking of the broader agricultural process. The situation is exacerbated by climate change, as it leads to modifications on temperature, precipitation and the frequency of extreme weather events at a global scale.
Decision makers in the sector need tools that help predict the timing, intensity and coverage of such occurrences. The tools must provide concrete, actionable information towards increasing production, alleviating risks and mitigating the impact of destructive events and abrupt changes.
Predictive modeling can be used at various levels and stages of agricultural production, like land utilisation, variety selection and action scheduling. Different models must be invoked on each case, having important impact on the final outcome in terms of production, financial benefits and environmental impact.
Predictions of seasonal weather patterns like droughts or heat waves have an impact on the farmers’ decisions for which crops to grow and how much of each crop to grow, as well as, the expected benefits of crop insurance. This is particularly applicable to regions where there is traditionally a large variety of crops.
Moreover, climate prediction can provide valuable input for the optimisation of variety selection and crop input application (e.g. fertilisers, pesticides and fungicides). Information produced by such models can be used to determine the type, amount, timing and application method of the crop inputs on the proper crop variety for the given growing season conditions.
As an example, nitrogen fertilisation for maize cultivations can be more accurately designed if information about expected temperatures and precipitation during various timeframes of the growing season is available.
Furthermore, it is important to predict weather conditions and events of smaller timeframes that have nevertheless significant impact on crop yield. As most crops have developmental stages that are sensitive to weather conditions, e.g. the pollination stage for maize, climate prediction during these critical stages can improve the ability to determine crop yields and grain production.
Using statistical modeling, data mining and machine learning techniques, we offer predictive models for:
- local climate prediction
- optimum seed rate
- yield estimation
- seed-yield potential
- farm-level ROI estimates
- environmental impact