Forecasts on Demand for Renewable Energy and Agriculture at The Weather Company, an IBM Business

A new global 1-15 day weather forecast capability to serve renewable energy and agriculture industry needs has been developed at The Weather Company, an IBM Business. The system is an extension of The Weather Company's Forecast on Demand infrastructure (Neilley et al., 2015), using spatial and temporal downscaling and a multi-model weighted consensus methodology to provide forecasts at any specified global location. Forecasts are created using the freshest NWP model data available that is downscaled and blended at the time of the request. Forecast timesteps are hourly for all forecast variables, with sub-hourly solar irradiance forecasts provided for the first 7 days. The forecasts are delivered via a robust, cloud-based API infrastructure comprising a collection of public-facing web services designed for extremely low latency and high availability. A proprietary and fast just-in-time backend science computation engine enables high throughput performance.

The new agriculture and energy forecasts include:

• Solar Energy: Global 15-day hourly, and 7-day 15 minute-resolution forecasts of global horizontal irradiance (GHI) and direct normal irradiance (DNI).
• Wind Energy: Global 15-day hourly forecasts of wind speed, wind direction, and air density at a user-specified altitude between 10 and 260 m above ground.
• Agriculture: Global 15-day hourly forecasts of soil moisture and soil temperature at any depth from the surface to 200 cm; reference evapotranspiration based on the Penman-Monteith formulation for a short grass crop; crop-specific evapotranspiration for over 100 crop types based on crop maturity; and model evapotranspiration based on latent heat flux.

The presentation will review the new forecast system, its numerical design philosophy, scientific challenges encountered and pragmatic solutions implemented. It will also discuss applications of these specialized forecasts to weather-driven decision systems, contributing to balanced integration of intermittent renewable energy in existing electricity grid infrastructure and to sustainable management of agricultural production in a resource-constrained world.

Reference

Neilley P.P. et al., 2015: Overview of The Weather Company's Principal Forecasting Methodologies. 27th Conference On Weather Analysis And Forecasting / 23rd Conference On Numerical Weather Prediction, Chicago, IL, American Meteorological Society, 3A.7. [Available online at https://ams.confex.com/ams/27WAF23NWP/webprogram/Paper273524.html]