Cloud and Solar Power Prediction

The proper forecasting of clouds in the use case Cloud and Solar Power Prediction is important for the short-term predictions of photovoltaic power, photo-chemically impaired air quality, and precipitation. Cloud scenarios are best observed by satellites. The transfer of this space-borne information in prognostic models, expected to result in a demonstrated beneficial effect on cloud evolution and prediction capabilities, is an unresolved issue. Major Scientific Big Data Analytics (SBDA) methods to be applied are supervised learning as well as parallel and scalable classification algorithms.

This use case is contributed by the Institute of Energy and Climate Research - Troposphere (IEK-8).

Spaghetti plot of an ultra large meteorological ensemble with 1024 members. Geopotential height at 850 hPa (isohypses of 134 gpdam) at 12 UTC 9 August 2014 (initialized 00 UTC 8 August). Dispersion of isophypses indicates low predictability of the shape and core pressure of a shortwave trough which led to extreme errors in the day-ahead wind power forecast.