Weather Forecasting

Weather forecasting models are used to produce local weather and rainfall guidance for agricultural planning.

Business impact

• yield per hectare — Microsoft and ICRISAT report a 10-30% higher yield per hectare for farmers using the app; the pilot is also described as 30% higher.
• farmers reached — The pilot ran with 175 farmers and the expanded program reached more than 3,000 farmers.

Data requirements

• historical climate data (tabular) — The Microsoft case uses more than 30 years of climate history to calculate the crop-sowing period.
• daily rainfall measurements (tabular) — Real-time rainfall records are used to calculate current Moisture Adequacy Index values.
• weather forecasting models (tabular) — Forecast model output is used to estimate future moisture adequacy and weather conditions.
• satellite data (image) — Google says WeatherNext 3 includes real-time satellite data.

AI methods and techniques

• machine learning — Used to analyze historical climate data and support sowing-date recommendations.
• weather forecasting models — Used to estimate future weather conditions and feed the seven-day forecast and sowing guidance.
• predictive analytics — Used to calculate the crop-sowing period and the Moisture Adequacy Index from current and forecast data.

Industries

Real-world evidence

2 documented case studies on record.

Companies using this: Microsoft, Google.