Employing weather derivatives to assess the economic value of high-impact weather forecasts out to ten days - indicating a commercial application

A "real-time" trial of a methodology utilised to generate Day-1 to Day-7 forecasts, by mechanically integrating judgmental (human) and automated predictions, has been ongoing since 20 August 2005 (Stern, 2007; refer also to http://www.weather-climate.com/10dayforecasts.html). After 500 days, to 1 January 2007, the trial revealed that, overall, the various components (rainfall amount, sensible weather, minimum temperature, and maximum temperature) of Melbourne forecasts so generated explained 42.3% variance of the weather, 7.1% more variance than the 35.2% variance explained by the human (official) forecasts alone.

The trial continues.

Shapiro and Thorpe (2004) note that "THORPEX addresses the influence of sub-seasonal time-scales on high-impact forecasts out to two weeks, and thereby aspires to bridge the 'middle ground' between medium range weather forecasting and climate prediction". Stern (2005) identified a modest level of forecast skill out to ten days.

Since 20 August 2006, forecasts have also been generated for beyond Day-7 (Day-8 to Day-10). After 135 days, to 1 January 2007, Day-8 forecasts explained 13.1% of the variance, Day-9 forecasts explained 7.9% of the variance, and Day-10 forecasts explained 3.9% of the variance - for these longer range forecasts, the variance explained was mainly for the temperature components.

Although it might have been expected that mechanically integrating judgmental (human) and automated predictions via some kind of an averaging procedure would have resulted in a significantly inferior set of forecasts for extreme (high-impact) events, this proved not to be the case.

To illustrate, the RMS Error of the combined forecasts leading to the ten coldest nights was 2.47 deg C, whilst in comparison, the RMS Error of official forecasts leading to the ten coldest nights was 2.51 deg C. Furthermore, the RMS Error of the sequence of combined forecasts leading to the ten hottest days was 5.54 deg C, whilst in comparison, the RMS Error of official forecasts leading to the ten hottest days was 5.81 deg C.

What is particularly interesting about the verification data is that the combined forecasts are more consistent than the official forecasts. For example:

- The consistency, that is, the RMS inter-diurnal change in the sequences of combined forecasts of minimum temperature (7 days in advance, 6 days in advance, 5 days in advance, 4 days in advance, 3 days in advance, 2 days in advance, 1 day in advance) is 1.17ºC (this RMS inter-diurnal change being well below the 1.36ºC associated with the official forecasts); and,

- The consistency, that is, the RMS inter-diurnal change in the sequences of combined forecasts of maximum temperature is 1.36ºC (this RMS inter-diurnal change being well below the 1.86ºC associated with the official forecasts).

In a 1992 paper presented to the 5th International Meeting on Statistical Climatology, the author introduced a methodology for calculating the cost of protecting against the onset of global warming. The paper, “The likelihood of climate change: A methodology to assess the risk and the appropriate defence”, was presented to the meeting held in Toronto, Canada, under the auspices of the American Meteorology Society (AMS). In this first application of what later was to become known as 'weather derivatives', the methodology used options pricing theory from the financial markets to evaluate hedging and speculative instruments that may be applied to climate fluctuations. What now follows is an application of options pricing theory where one uses the theory in a weather risk management context.

The theory shows that the more consistent forecasts are from one day to the next, between Day-7 (when they are first issued) and Day-1 (the final issue), the cheaper are the prices of option contracts that one may wish to purchase to protect against the eventuality that the forecasts might be incorrect. The implication from this is that, the more consistent forecasts are from one day to the next, the more valuable are the forecasts.

The American Marketing Association (2006) notes that "a “competitive advantage” exists when there is a match between the distinctive competences of a firm and the factors critical for success within the industry that permits the firm to outperform its competitors. Advantages can be gained by having the lowest delivered costs and/or differentiation in terms of providing superior or unique performance on attributes that are important to customers."

From the foregoing, it may be said that the value of a series of weather forecasts with a low volatility, that is, a series of forecasts that display a high level of consistency from one day to the next, is greater than the value of a series of forecasts with a high volatility. This is because the cost of protecting against the possibility of such weather forecasts being incorrect by adopting a strategy of purchasing weather derivatives is lower.

This means that sellers of weather derivatives, who utilise low volatility forecasts to price their call and put options, are provided with a competitive advantage over sellers of weather derivatives who utilise high volatility forecasts. This arises because sellers of weather derivatives who utilise low volatility forecasts being able to charge lower, and, therefore, more competitive, prices to purchasers of weather derivatives who wish to use those weather derivatives to protect against the possibility of the weather forecasts being incorrect.

Hence, the verification data, in showing the combined forecasts are more consistent than the official forecasts, are also showing that the combined forecasts are more valuable than the official forecasts.

Furthermore, that the combined forecasts are also more accurate than individual currently available predictions taken separately, also provides the small to medium sized companies involved in weather risk management and weather broadcasting with a potential competitive advantage over their peers should they choose to adopt a strategy of mechanically combining existing predictions.

And, there is a multiplicity of existing predictions to choose from.