Weaker La Nina Trends May Portend Colder Winter: Could Energy Bulls Benefit?

With the new month, more climate models are beginning to print out their late fall/winter forecasts. The climate model consensus has generally been one that is quite warm, though there is a decent range. The NMME climate model, which just had its July run come out very recently, is forecasting widespread warmth across the country for December/January/February, as seen below. This also fits with recent runs of the other American CFSv2 climate model.

There are reasons to be skeptical. The model does show a weak/moderate La Nina through the winter, as seen below.

In such a scenario, we could see widespread warmth across the country. However, it is far from a guarantee. In fact, recently we have seen confidence decreasing on just how much of a La Nina we will have come winter, and some aspects of the upper level pattern on this model do not fit what we would expect with a weaker La Nina. The Australian Bureau of Meteorology shows either neutral or very weak La Nina conditions through the winter, but recently the ensemble climate members have been skewed higher, with some trending neutral through the winter.

The American CFSv2 has a similarly wide spread. In blue we see the 8 most recent runs of the model, and it shows everything from a moderate La Nina to almost entirely neutral ENSO conditions for the winter. The ensemble mean now favors a very weak La Nina, potentially not even reaching technical La Nina classification if it unable to persist for the 5-month minimum.

Both dynamic and statistical models were favoring a stronger La Nina a couple months ago, but this more recent chart from the CPC has not been updated since June 16th; we are expecting the new one in the next week or so and expect some of these models will back off their stronger La Nina forecasts even more as we have not seen quite as rapid a transition to La Nina conditions as expected.

In fact, just today the Southern Oscillation Index (SOI) flipped back weakly negative, indicating no significant short-term transition to a more La Nina-dominated atmosphere. As can be seen, April had a strongly negative SOI and since then we had been trending positive, though not in any very significant or convincing way.

This is as the daily Nino 3.4 Index has finally stabilized following a very volatile period in mid-Juny, and we see neutral conditions lingering with no clear cooling since late May. This index had previously been clouded with noise thanks to Pacific buoy placement, but the day-to-day stability we see now is giving us higher confidence in using the daily index as a proxy before Monday's next update from the CPC on La Nina development.

All of these trends are raising questions about the pending La Nina development, and these trends could have drastic impacts on the winter forecasts. Winter forecasts created within a La Nina base state may need to be modified, though there is still time for development to pick back up again as our modeling guidance is notoriously poor at nailing longer-term ENSO trends. Similarly, we see a wide spread that shows a wide range of possibilities for the coming winter, which is in almost direct contrast to last winter when we had very strong confidence in a strong El Nino. This opens up the potential for more variability in winter forecasts through the coming months.

Looking through historic ENSO data, we see that most strong El Ninos tend to be followed by La Ninas, but not always.

In 1998 we had an extremely strong El Nino rapidly transition into a strong La Nina that winter. This is looking increasingly unlikely. Rather, the year 1983 has our attention, where we had a strong El Nino but more gradually transitioned into weakly negative conditions late that year that were not classified as La Nina. By January and February negative SST anomalies were rising gradually, much like most climate models (including the European ECMWF ensembles below) have been showing.

As the image above shows, individual ensemble members are still printing out a massive spread with the ENSO for winter months, meaning that longer-term forecasts and expectations could be widely volatile. Should the stronger La Nina come, we could see the 1998 analog return as something to benchmark our conditions against, and that winter was not particularly cold. However, should we see the La Nina not develop as strongly, which is looking increasingly possible as doubt is injected into the forecast, then we could see 1983 dominate more. For comparison, November and December in 1983 were very cold, as seen below.

January and February in 1984, meanwhile, were not quite as cold, with warmer weather returning across western Canada into portions of the Great Plains and Midwest.

Still, a closer analysis shows that January was colder than average across the country, and it was not until February that warmer weather began to dominate the pattern. Cold then came flooding right back in for March as well for a winter with significantly above average heating demand. If a similar pattern sets up, we could be looking at a boost for natural gas prices and potentially heating oil prices/spreads.

What our preliminary analysis has continued to indicate is that on a fundamental level natural gas bulls should be rooting for less La Nina development, which has been the recent trend across climate guidance. Of course, as winter nears, there will be any number of variables to take into account, so the ENSO base state will not be the sole determiner of what the winter pattern will look like, especially as the positive PDO has weakened significantly over the week and the QBO is acting uncharacteristically. Still, as a strong La Nina is looking increasingly unlikely, this would seem to open up some colder risk in the winter forecast, since ENSO may not be as major a driver of the winter pattern, and that may be something that some of our climate models are just not picking up on yet.