Norway is known as the battery of Europe, so understanding the snow situation is paramount in price forecasting. Accurately predicting the price developments in the Nordic Power Market is difficult. There are a large number of influential factors as well as a lack of sufficient data on the amount of snow and water resources. This article aims to introduce snow cover data and to have you consider it as a valuable alternative data source.
Fractional Snow Cover is an important indicator for any snow season and corresponds to the fraction a specific area is covered in snow. The key challenge of traditional snow cover data is that the signal saturates at 100% snow cover and snow accumulation above 100% cannot be measured. EDInsights proprietary Fractional Snow Cover analytics excel from the competition in tracking accumulation and melting during the complete snow season providing a much better indication of the type and quality of snow season. These novel analytics describe and characterize the snow season in an unprecedented way and allow for improved water balance forecasts, calculation of the snow water equivalent during the melting phase and calibration of hydrological models. A plot of Snow-Covered Area for a region of interest can be seen below.
Multiple satellites are orbiting Earth which are suited to monitoring snow cover, satellites such as MODIS operated by NASA have been providing long time series of cross calibrated data for the last two decades. EDInsights’s spaceborne analytics tools convert these to over 3000 unique data points and indicators which enable accurate tracking of the snow season. In the next step, we translated these analytics into actionable information by cross-referencing 20 years of historical Nordic Fractional Snow Cover data together with Nordic Power Pricing data. The result is a list of indicators for Nordic System Price and an associated price model for the Summer period. More concrete, these indicators are combined with coal and EUA prices to predict the June Nordic System Future Price on NASDAQ.
Below is a plot of a single proprietary snow indicator and price development.
To validate the accuracy and robustness of the indicators and price prediction model it was back-tested on 9 years of historical NASDAQ Nordic Power System Price data with a simulated trade being placed on the June contract each year between 2010 and 2018. The direction of the trade would be placed depending on the relation between our predicted June Future Contract Price at May 31st and Current June Future Contract Price according to NASDAQ historical data; the contract would be traded long if we believed the price was to low and traded short if the price was too high. The results of the back-testing were positive with all simulated trades being on the correct side of the market. From this back-testing, we felt confident enough in our model to step up the validation and placed a live trade during 2019. We ran the price prediction model driven by EDInsights snow indicators on April 1st with a predicted June Nordic Future Price of €34.12. A live trade was placed on April 15th at a price of €41.9. We shorted the contract based on our price model output. The outcome of the trade was a success as the contract dropped to €34.5, meaning the contract moved by €7, a substantial figure given the median change during the same window between 2010-2018 was €2.8. Below is a summary of the Price Prediction Report on April 1st, 2019.
EDInsights Nordic Snow Cover Analytics proved to be invaluable in the price prediction as it was the key input into the price prediction model. The result of the above mentioned live trade performed beyond expectations with a price prediction error of 1%. This is an improvement on the backtested results between 2010 and 2018 with a median prediction error of 3%. This work was completed to demonstrate the value of EDInsights Snow Cover Analytics which was illustrated successfully via backtesting and live trading.
For more information contact sales.