MetraWeather UK lead forecaster Byron Drew, discusses how forecasting extreme weather events is the key for energy traders.
As anyone who’s tried to plan a family day out in Northern Europe knows, our weather is extremely unpredictable. Weather systems are immensely complex, and therefore forecasting them accurately requires vast amounts of data. Limited processing capability means the amount of data that can be included in models has to be restricted.
At the same time, many strategies deployed by energy traders rely on forecasting the weather correctly – particularly anticipating extreme weather events, which can have a huge impact on both energy supply and demand.
The good news for energy traders is that radical improvements in data processing have made it possible to include more data in models than ever before. Modern computing power means that processing that once took a week today takes minutes.
The models and data sets selected by forecasters depend on the type of forecast required to meet an energy trader’s particular strategies. Some methods are best for creating accurate short term forecasts (two to three days out), but cannot give an accurate picture over the longer term, while data and models designed to excel for long term forecasts (15 days and beyond) are less likely to give such a precise picture for the short term.
Energy traders’ strategies may be focused on long or short term events, but in each case, accurate forecasting is essential. In particular, accurate forecasting of extreme weather events can mean the difference between success and failure. Significant weather events often lead to increased demand for power (whether for air conditioning or heating) and yet may also create shortages in supply (particularly renewable energy, which in most forms is dependent on weather).
These types of events can be highly disruptive for energy traders. While the event may never happen, if the trader knows that there is a chance of an extreme condition that may disrupt supply or lead to increased (or reduced) demand, they can trade and hedge accordingly.
Probabilistic forecasting is ideal for anticipating extreme weather conditions as the data can provide not only the most likely events (as in deterministic forecasting), but also the probability of other, more extreme, events occurring. This also enables traders to calculate the risk in their decisions, so they can make informed decisions on how to hedge against them.
To be best placed in the energy market, traders need to be aware of these small likelihood forecasts in their strategies and hedging, which they cannot do with basic deterministic models. To get the most accurate probabilistic forecast possible, the forecaster needs to blend multiple sets of data and weather models to create a highly accurate probability distribution.
This means far fewer of the surprises traders dislike so intensely – and is therefore a more reliable basis on which to make trading decisions.
Contact us now to discuss weather solutions tailored to your particular needs.