In early August, we begin to compile our ideas on winter weather and how we feel it would be prudent to position oneself in the energy markets into the winter given weather-related risks. Last year, around this time we began to have high confidence in the winter forecast, as it became clear that a strong El Nino was forming. That El Nino overwhelmed global weather patterns and took charge, and accordingly through the winter we saw what was generally a Nino-like pattern. 

This winter there are no such guarantees. In the natural gas market, we are currently seeing the winter strip trade at around a 40-cent premium to the current prompt month September contract, as demand is of course higher in the winter months and there are more risks for freeze-ins disrupting production. 

Similarly, it is rare for the market to significantly price in weather expectations more than 3-4 weeks in advance (at most), as the winter strip was similarly elevated compared to the prompt month last fall. Many traders remain skeptical of longer-range forecasts, as they should at times given the amount of noise that can cloud them. The result was that last winter, despite very high confidence in a warm December from El Nino, prices waited until late November and early December to reach their lows for the year, not rallying until the first significant cold snap came in early January. Even after that, cold was short-lived, so prices retreated to new lows in early March as it became clear no significant, sustained cold was arriving. 

This year, the confidence in winter forecasts has been significantly decreased. As we have been discussing, a La Nina looks likely for winter, though the latest update from the Climate Prediction Center only gives it around a 60% chance of forming, with chances falling off later in the winter. 

Additionally, we have seen a wide spread in climate models, from no ENSO influence whatsoever to a moderate La Nina forming. 

The consensus, however, seems to be for a weak La Nina, that will exert some influence on the pattern but potentially not all that much. Meanwhile, the Quasi-Biennial Oscillation (QBO) measuring winds at the stratosphere has broken down in cycle. As can be seen below, with it we typically oscillate between negative (easterly winds) and positive (westerly winds) phases. The easterly phases typically last about twice as long and are stronger. However, at the 30-hpa level, this oscillation has broken down, as we expected to trend back easterly but instead rapidly reversed. 

This level of a breakdown has not been observed since at least 1960, and is throwing a wrench in a number of forecasts as many scramble to determine where we go from here. The July print due out in the next couple of day should provide some guidance, and the very strong El Nino of last year may have played some role, but what is normally a stalwart index for winter forecasts (being predictable on timescales of up to 6-12 months) has now become highly variable. Add in the fact that a recent positive phase of the Pacific Decadal Oscillation (PDO) may be breaking down (associated with but not directly caused by La Nina formation) and uncertainty is clearly clouding winter forecasts. 

Most traders are unlikely to be looking at winter forecasts at this stage regardless, figuring them to be clouded by noise, and at least thus far such an assumption would appear valid. The result of these lower-confidence winter developments could mean additional short-term weather-driven volatility as we near the winter months. More traders may just have to keep their eyes on the noisiest things we have: weather computer models.