The intelligence you need to anticipate weather-driven

fluctuations in energy supply and demand

Weather is a primary driver for commodity prices in energy, having an impact on both energy consumption (temperature) and production (wind and solar). CFAN’s OmniCast product provides weather and climate forecasts for energy traders, power providers and fund managers. CFAN’s research and product development team has created the most comprehensive and accurate platform available for making decisions related to energy supply and demand. 

CFAN’s OmniCast forecast products differentiated from competitor products:

  • Provision of both optimal and probabilistic forecasts

  • Proprietary algorithms minimize bias and distributional errors from the global model forecasts

  • Ensemble clustering to increase forecast skill at longer forecast horizons

  • Real-time forecast verification statistics and forecast confidence measures

  • Artificial Intelligence approach to longer range (seasonal) forecasts


OmniCast overview

OmniCast provides seamless access to CFAN's complete forecast products for surface weather for the main energy markets in the U.S. and Europe, on timescales from days to seasons. The OmniCast user interface is based on an interactive layered dashboard approach. Our solutions provide a combination of quick-response, decision-oriented outputs with more complex meteorological graphics.  Easy access is provided to download OmniCast forecast data streams.


OmniCast Forecast products:

  • Daily temperature forecasts (1-35 days) (U.S., Europe)

  • Seasonal forecasts of temperature and precipitation (1-6 months)  (U.S.)

  • Renewable energy forecasts (7-41 days)  (Western U.S.)

  • Forecasts of teleconnection regimes  (1-45 days; 1-6 months)

Daily temperature forecasts


OmniCast provides daily temperature forecasts out to 35 days.  Forecasts of daily maximum and minimum temperatures are provided twice daily out to 15 days, and twice weekly from days 16-35.  Forecasts are provided for over 300 U.S. cities and over 250 European cities.


For each city, the following temperature forecast information is provided:

  • Table showing daily raw model forecasts, CFAN calibrated forecasts, and CFAN consensus forecast, color coded to show the deviation from the 10 year or 30 year climatology for that date

  • Inter-percentile plume view to denote forecast uncertainty

  • Real-time and historical verification statistics for each forecast leadtime

  • Heat wave and cold event probabilities

  • Population-weighted heating/cooling degree days

  • Table showing 12 and 24 hour changes in the forecast


Insights into forecast uncertainty are provided by plume of the forecast ensemble. Real-time verification statistics are provided for each city, for each daily leadtime. Verification statistics are provided for the past 15 days, past 30 days, and past 90 days.


City temperature forecast products are illustrated for the record-breaking Dallas, TX heat wave in July 2018, with maximum temperature reaching 106oF on July 17 and 111oF on July 22.  A complete forecast summary and verification for this heat wave: [link].


In addition to the real time city verification tables, seasonal verification statistics are provided as a function of leadtime for each region.  Verification statistics for forecasted U.S. average temperatures for Jul•Aug•Sept 2018 compare CFAN's forecast with the ECMWF ensemble mean, GFSX, and climatology.


2019 verification report for CFAN's 15 day heat/cold wave forecasts [link]

Renewable energy forecasts

Wind Turbines
Solar Panel Installation

Matching wind and solar anomaly forecasts with temperature anomalies gives insight as to where and when power demand and supply issues might be most critical.  The need for wind and solar power forecast information is extending to longer projection windows with increasing penetration into the grid and also with diminishing reserve margins to meet peak loads during significant weather events. Maintenance planning and natural gas trading is being influenced increasingly by anticipation of wind and solar generation on timescales of weeks to months.


CFAN provides twice weekly forecasts of subseasonal forecasts (7-41 days) for 100 m wind speed and surface solar flux.  CFAN’s renewable energy forecast product for the U.S. includes:

  • Maps of rolling 5 day percentile anomalies

  • Clustered winds based upon weather regime cluster

  • Maps of rolling 5 day wind threshold probability exceedances:

         ≤ 4 m/s

          4.1 - 7 m/s

         7.1 - 10 m/s

         10.1 - 14 m/s

         14.1 - 25 m/s

         > 25 m/s


During the period May 20-28, 2019, wind power generation in ERCOT (Texas)

averaged 11-14 GW each day, around the clock.  CFAN identified this period of sustained winds beginning on May 6, with a predominance of wind speeds in the range 10-14 m/s. 


Seasonal forecasts


CFAN provides global model-based seasonal forecasts (1-6 months) for the U.S. and Europe:

  • Monthly forecasts from global models on the 5th day of the month of monthly temperature and precipitation anomalies

  • Maps of hemispheric height and temperature anomalies plumes of teleconnections indices

  • Probabilistic predictions of ENSO (Nino3.4, Nino1.2, Nino3, Nino4) and Modoki


Seasonal forecasts from the global models show little skill beyond two months unless there is a strong ENSO signal. To address this shortcoming, CFAN has developed a new data-based statistical/dynamical seasonal forecasting system using Artificial Intelligence techniques. Our statistical forecast system is based on climate dynamics research on global circulation patterns in the stratosphere, troposphere and ocean on regional, seasonal weather patterns.


CFAN’s statistical/dynamical temperature forecasts project 3-month temperature anomalies at lead times of 3 to 6 months, over 7 U.S. regions and for 1218 local sites.  These forecasts have been produced operationally since October 2017.  The forecasts are produced once per month and disseminated in the form of a report around the 14th of the month.  CFAN’s forecast report issued in May 2019 can be downloaded [link]


Teleconnections and ensemble clustering


Greater insights at extended forecast horizons can be obtained from our analyses of atmospheric circulation patterns.  OmniCast provides forecasts of weather regime (teleconnection) patterns for 1-15 day forecasts (twice daily), 16-45 day forecasts (twice weekly) and 1-6 months (monthly).  The following teleconnection patterns are predicted from global model forecast ensembles:

  • Arctic Oscillation (AO)

  • North Atlantic Oscillation (NAO)

  • Pacific North American pattern (PNA)

  • East Pacific Oscillation (EPO)

  • West Pacific Oscillation (WPO)

  • Antarctic Oscillation (AAO)

  • Madden Julian Oscillation (MJO)

  • Global Atmospheric Angular Momentum Index (GLAAM)


Probability plumes of the teleconnection predictions are provided for both the ECMWF and NOAA CFS forecast models.


At forecast time horizons beyond one week, clustering of the forecast ensemble members can help increase the sharpness of the distributions and in the assessment of uncertainty. Clusters are identified from analysis of pre-defined weather regimes based on historical analyses of geopotential patterns at 500 hPa. These clusters are then identified with dominant teleconnection patterns. Analysis of the clusters supports assessment of probabilities or likelihood of one weather regime versus another. The cluster probabilities help interpret the impact of major circulation transitions on the temperature anomaly patterns.


Pairing CFAN’s extended range temperature and wind forecasts with weather regime (teleconnection) information and clustering provides the basis for understanding of pronounced deviations from normal and interpreting the impacts of weather regime transitions. 

Trial subscription


CFAN offers trial subscriptions to OmniCast to qualified individuals and organizations. 

To inquire about a trial subscription [link]