A Statistical Approach to Forecasting of Seasonal Anomalies in Wind for Renewable Power Production

With the increasing proliferation of renewable wind power projects across North America, South America and the Caribbean Sea, there has been a growing need from both operators and the investment community for long-range forecasting of seasonal wind anomalies, both for operational and financial considerations and for planning and maintenance purposes. These forecasts require knowledge regarding expected seasonal variations on a quarterly basis, and should be updated at least once a month in order to look for any changes in the long-range signals.

Using a proprietary process focused on long-range teleconnections, our method been able to demonstrate skill at providing this type of information to the at-large renewable and wind power community across the affected areas. This process was used experimentally in 2015, and has become increasingly more reliable as we have been able to use ground-truth verification and improve our skill as a result. The forecasting product is constructed using historical anomalous wind patterns for specific locations with extrapolated water temperature anomalies associated with these wind patterns, combined with identification of analog seasons based on current water temperature anomalies in key areas of the Pacific and Atlantic oceans. Using the outcome of this method, operating, logistics and financial components of the industry are better able to plan and be prepared for upcoming seasonal variations in wind production.