The reason for weather risk management, he said, is “to transfer the risk of normal weather variability from one party to another” through a financial instrument. Weather-indexed financial contracts, for example, include such things as temperature and degree-day derivatives.

For the end user, this allows you “to unload weather risk and make costs and revenues more stable.” For the risk takers/investors, they earn a premium for taking the weather risk.

The average conditions for coverage are over a season, such as over a heating or cooling season. Dischel noted that weather risk is not about extreme events like floods. It’s about a departure from normal, e.g., a cooler than normal summer.

Weather financial contracts include derivatives or insurance products. Payment is not necessarily based on a potential loss. So you don’t have to have a loss to get paid. Payment is keyed to an objective weather index.

Hedging Your Bets

“When you hedge weather exposure well, you take weather risk off the table and you can focus on business competencies,” Dischel said.

Aaron Studwell, director of research for CMS Energy, Market-ing, Services and Trading, then presented “Weather Derivatives 101.”

A weather derivative is a financial instrument introduced in 1996 that is traded in the over-the-counter market. It is “a risk management tool as well as a speculative tool,” he explained.

A natural gas company, for example, may have a mild winter season with less gas usage, reducing its revenue. If it purchased a weather derivative, the company can collect a preset amount per heating degree day below a predetermined level.

The option structure of a derivative is to measure by degree days or by the number of days with the temperature above or below a preset level. The rate is in dollars per degree day or dollars per number of days above/below a particular temperature. Limits are placed on the amount of the payout.

“Weather derivatives provide a variety of methods to reduce the weather-related exposure of operating expenses and revenue,” said Studwell.

“Tailor-Made Weather Solu-tions” was the topic of Frank Caifa, associate director of Swiss Re New Markets. In assessing weather risks, he said, you determine how weather affects your profits. Is it temperature-related events? Is it precipitation-related events? And, can you quantify the effects in dollar terms?

In developing solutions, you can do a simple risk transfer. Or, Caifa remarked, you can do a tailored combination of risk financing and risk transfer. This provides a layer of contingent capital and smoothes out cash flow and the revenue stream.

A tailored combination is “a customized solution that closely correlates to the specific exposure,” he said. It could be based on average temperature each month during the summer for a particular area.

The duration can be from months to years. The index can be temperature-based, precipitation-based, or multi-triggered. The structure can be simple or complex. The market players can include insurance companies, banks, and commodity markets.

How’s the Weather?

Shifting the focus to weather prediction, Jan Dutton, president of Weather Ventures Ltd., talked about “Short-Term and Long-Term Weather Forecasting: Differences and Similarities.”

“To predict the weather for the next few days, you look upstream,” he declared. “To predict climate conditions for the next month, you look at global surface conditions.”

Weather risk occurs at all time scales, he said. For the short term, you have high forecast skill; you have less skill for the longer term.

Virtually all forecasting is based on numerical model information. To start a forecast, you need to know such conditions as winds, temperature, humidity, pressure, etc. This data is collected by satellites, radiosonde balloons, ground stations, and aircraft.

To prepare a model forecast, the future weather is simulated from these initial conditions. For the short term, the initial state of the atmosphere is most important. For the long term, global conditions are more important.

Tropical Pacific temperatures are most significant for long-term forecasts, Dutton indicated. This includes extreme tropical sea surface temperature events such as El Niño and La Niña.

Long-range forecasts, he said, are probabilistic, i.e., telling you the probability of being above or below normal temperature.

John Snook, research scientist with Colorado Research Assoc-iates, discussed “Utilization of Advanced Numerical Weather Prediction Techniques.”

The first numerical weather forecast was done in 1950, he said, and it has become a primary basis for modern weather prediction. Numerical weather prediction is “the forecasting of atmospheric behavior by the numerical solution of the governing fundamental hydrodynamic equations subject to observed initial conditions.”

The National Weather Service (NWS) provides the majority of the freely available weather products, but these are general forecasts. Private companies step in to provide specific local forecast products and forecast interpretation, related Snook. They organize, simplify, and repackage the information to meet end-user requirements.

For medium-range prediction, ensemble forecasting is more accurate than any individual forecast, he said. Ensemble forecasting takes the initial conditions and produces forecasts for slight variations of those conditions.

The other co-chairman, John A. Dutton, dean of the College of Earth and Mineral Sciences and professor of meteorology at Pennsylvania State University, covered “Advancing Capabilities in Atmospheric Science.”

According to his numbers, more than $2 trillion worth of business is affected by weather risk. There is increased interest in long-term time scales as well as short term.

The traditional method of incorporating weather information in decision making involves NWS computer forecasts being analyzed by a human forecaster; then the forecast and advice are given to the decision-maker.

The new era system, stated Dutton, includes taking NWS computer forecasts and applying private numerical analysis using impact variables and decision aids specific to an industry, and then integrating that data with risk models before presenting it to the decision-maker.

An important weather issue, he said, is developing reliable data sets. Data from New Jersey, for example, may not match New York data right next door.

Roland Madden, senior scientist, Climate and Global Dynamics Division, National Center for Atmospheric Research, then explained “The Basis for Long-Range Weather Forecasting.”

“Weather forecasts are made by solving dynamical equations that describe wind motions and energy inputs to the atmospheric system,” he said.

Long-term climate forecasts “predict time-averaged weather for intervals of a month or a season,” Madden observed. They predict changes in the probability or the likely occurrence of certain weather. They are most often based on slowly changing boundary conditions like sea surface temperatures.

The American Meteorological Society’s position is that scientifically based, skillful, seasonal forecasts are possible one to two seasons in advance for some parts of the world and for some seasons.

Risks for The Little Guy

Looking again at weather risk, Agbeli Ameko, chief executive officer of Foresight Weather, LLC, presented “Weather Tools for the Little Guy in the Energy Industry.”

Weather derivatives are a useful tool in managing weather risk, he commented. “But you must understand your exposure first.”

Exposure equals your potential dollar losses. Many businesses view their exposure in “a limited manner,” said Ameko, and they may view much of it as unmanageable. In fact, tools exist to understand and manage risk on all time scales, from short to medium to long term, he maintained. For the medium range, ensemble forecasting provides quantitative and qualitative results. For the long range, seasonal forecasts and new probabilistic methods can be used.

Finally, Simon Atkins, chief executive officer and atmospheric scientist, Advanced Forecasting Corp., discussed “Case Studies Using Meteorological Intelligence to Gain a Competitive Edge.”

Two goals to always have in mind with risk management, he stated, are to stabilize your earnings volatility caused by weather and protect sales and profits from unseasonable variations. Important concepts to minimize weather-related losses, said Atkins, include: 1. Do not accept weather risk as standard protocol; 2. Identify cost and revenue streams that are exposed to weather risk; 3. Quantify corporate risk; 4. Customize a set of unique solutions; and 5. Optimize the risk/reward profile.

For more information, contact Jan Dutton of Weather Ventures at 804-295-9989; jfd@weatherventures.com (e-mail), or contact John A. Dutton of Pennsylvania State University at 814-865-6546; dutton@ems.psu.edu (e-mail)

Publication date: 07/16/2001