The North American ‘Wind Drought’: Is it the new normal?

It seems a long time ago now, but it was only last March that my hometown of Boston was digging out from the last of a series of storms that dumped almost 3 meters of snow on the city. It was the heaviest accumulation since Boston started keeping records in 1872.

That was only one of a number of strange weather events in North America in 2015. Last spring, much of Texas and Oklahoma suffered weeks of torrential rains that destroyed crops, drowned cattle, and flooded houses. Even worse, California was and remains in the throes of one of the most severe droughts on record. Though it started in late 2011, the drought has only intensified this year.

Winds have also behaved strangely, albeit in a less dramatic way. Wind resources have been unquestionably low in much of North America – so low that the possibility of a prolonged ‘wind drought’ has been on the minds of many in the wind industry.

When I talk to our clients about this, they want to know three things:

  • How severe was the wind drought, really?
  • Why have winds been so low this year, and does it have anything to do with El Niño or climate change?
  • Most important: What is the outlook for 2016?

I’ll do my best in this blog to answer these questions (though the last one is tough).

How low were winds in North America in 2015?

The simple answer is, in the first half of the year, very low indeed. The three maps below show wind anomalies for the first 3 quarters of 2015. Blue colors represent wind speeds below the 1988-2014 average for the same quarter. Yellow and orange represent above-normal speeds. (The maps were generated by AWS Truepower using a custom blend of 3 reanalysis data sets. For more maps and other information see our Wind Trends Bulletin.)

In the first 2 quarters, average wind speeds were at least 9% below average, and in some places more than 18% below average, throughout most of the West, the Great Plains, and Mexico. These deficits affected some of the continent’s largest concentrations of wind plants, and consequently had a huge impact on the wind industry’s production and revenues in this period. (Typically, every percent decrease in average wind speed produces roughly a 1.5%-2% decrease in average output.)

Wind Anomoly NA Q1-Q2-Q3-wind drought2

Maps of wind speed anomalies in North America for (left) Q1 (January, February, March), (middle) Q2 (April, May, June), and (right) Q3 (July, August, September). The colors represent the percent departure in the average speed for that quarter from the 1988-2014 average for the same calendar quarter. Data are from a custom combination of 3 reanalysis datasets.

The news is not altogether grim, though. In the third quarter of 2015 wind patterns returned more or less to normal. Furthermore, unlike the California water drought, the low winds in Q1 and Q2 were not the continuation of an earlier pattern. On the contrary, 2014 was pretty normal, even slightly above normal, in most of the United States and Canada; only Mexico suffered from broadly below-normal winds in that year. (See our 2014 Annual Wind Trends roundup.)

A time series of wind speed anomalies for a point in West Texas based on MERRA and ERAI. The blue line is a rolling 2-year average.

A time series of wind speed anomalies for a point in West Texas based on MERRA and ERAI. The blue line is a rolling 2-year average.

This low-wind period, in other words, was more of a dry spell than a real drought.  What’s more, though a rare event, it was not entirely unprecedented. As an example, the chart to the right shows the quarterly wind anomalies for a location in West Texas. You can see that the large wind deficit in early 2015, which was around 15% below normal, was matched by a similar deficit in early 1987 and one nearly as large in 1992. In 1980, there was an equally large positive departure, when winds were about 15% above normal.

Another encouraging observation is that despite rising global temperatures and other signs of climate change, wind resources have actually been quite stable for the past 30 years (at least, in West Texas – but we’ve seen no sign of trends elsewhere, either). There have been ups and downs, to be sure. The rolling-average line on the chart suggests they typically last 3-9 years. Since about 2011, however, the region has actually been in one of the good periods. If the historical pattern is any guide, that might be about to change – but I’m getting ahead of myself.

In short, while early 2015 was tough for the wind industry in North America, it does not seem to be part of a trend. That argues against climate change being a leading cause – which is good news, of course, though it doesn’t rule out climate-change effects in the future.

What caused the wind deficits in 2015?

If climate change was not responsible for the wind shortfall in early 2015, what was?

Explaining wind patterns is always a challenge, since so many factors can come into play. The earth’s weather is chaotic, with storms and other systems forming and dispersing in quasi-random ways. That’s one reason why weather forecasting is so difficult. And it means that no explanation can ever be regarded as complete or definitive.

Illustration of the “Ridiculously Resilient Ridge” and its influence on storm tracks in winter and spring 2015. Source: KQED.

Illustration of the “Ridiculously Resilient Ridge” and its influence on storm tracks in winter and spring 2015. Source: KQED.

Still, certain patterns are clear. The most direct reason for the low winds across much of North America in early 2015 was a large high-pressure system, known as a ridge, that formed over the eastern North Pacific (right). This system caused winter and spring storms to take a big detour into Canada before dipping back into the United States. Among the consequences of that shift were drought in the West, huge rainstorms in Texas and Oklahoma – and, yes, lots of snow in Boston.

The larger question, though, is where did this system come from? What has become affectionately known in weather circles as the “ridiculously resilient ridge” (RRR) actually first appeared in 2013. It is closely linked to the formation of a large region of warm water in the eastern North Pacific Ocean, referred to as the “blob”, which was also first detected in 2013.

Scientists don’t entirely understand the reasons why the RRR and the blob have occurred or lasted so long. Some think it is tied to a cyclical change in temperature and wind patterns across the Pacific Ocean called – if you think I’m about to say El Niño, think again! – the Pacific Decadal Oscillation (PDO). The PDO changes phase every decade or two (hence its name). During its “cold” phase, the western part of the North Pacific tends to be warmer than usual, and the eastern part tends to be cooler. During its “warm” phase, the pattern reverses. Many scientists believe the PDO definitively entered a warm phase around 2014.

That is not to say El Niño plays no role. In fact, El Niño – which is more properly called the El Niño-Southern Oscillation, or ENSO – and the PDO often move in tandem. El Niño’s signature is a weakening of trade winds and the accumulation of unusually warm surface water in the eastern equatorial Pacific. (The opposite pattern is La Niña.) There were signs in the spring of 2014 of a developing El Niño, but by late 2014 it had fizzled. The present El Niño cycle started in April or May 2015, and is much stronger.

Normally, however, El Niño is associated with heavy rainfall in California as the winter and spring storm track is not diverted to the north but on the contrary intensifies and heads straight across the southern United States and northern Mexico. This time, despite being in a positive overall El Niño phase, there has been a deepening drought in the West. What happened? It could be that El Niño’s effects were delayed or offset by the RRR, and that now that a really strong El Niño is under way, its influence will win out. For California’s sake, let’s hope so.

What is the outlook for 2016?

Where does all this leave us? Unfortunately, there are few tools for forecasting weather conditions more than a few weeks into the future, and all carry a lot of uncertainty.

One approach we have been looking at is to examine so-called teleconnections relating events such as El Niño to weather conditions in areas of interest. They can be useful for long-range forecasts because many organizations forecast El Niño – and also because El Niño is linked to sea-surface temperatures, which are relatively slow to change and hence more easily forecast. If we can understand the relationship between El Niño (and other climate cycles) and wind resources, then perhaps we can say something meaningful about wind conditions in the future.

That there is a relationship between El Niño and wind resources is clear enough. The global maps below show the correlation between the average wind speed anomaly and a leading ENSO index (the multivariate ENSO index, or MEI) for the months of January-February, April-May, July-August, and October-November, for 1988-2014. Depending on how El Niño is defined, this period spans about 7 or 8 El Niño events.

The red colors indicate areas where winds tend to be lower when the ENSO index is positive – meaning El Niño. This is the case over most of North America in winter and spring. The reason has to do with the path of storms, which tends to be farther south when El Niño is present, bringing rainfall to the southern United States and Mexico. During La Niña, storms follow a more northerly track, bringing stronger winds to much of the rest of the continent. El Niño has noticeably less influence in summer and fall.

Maps showing the relationship between one leading ENSO index (MEI) and winds around the world. Colors indicate the (signed) r2 correlation coefficients for (top left) January-February, (top right) Apr-May, (bottom left) July-August, and (bottom right) October-November. Reds imply lower winds during El Niño, blues imply higher winds. The correlations are calculated with respect to the two-month trailing MEI for the first month of each month pair. The data span 1988-2014.

Maps showing the relationship between one leading ENSO index (MEI) and winds around the world. Colors indicate the r correlation coefficients for (top left) January-February, (top right) Apr-May, (bottom left) July-August, and (bottom right) October-November. Reds imply lower winds during El Niño, blues imply higher winds. The correlations are calculated with respect to the two-month trailing MEI for the first month of each month pair. The data span 1988-2014.

Forecasts of sea-surface temperature anomalies for equatorial eastern Pacific. (Source: NOAA)

Forecasts of sea-surface temperature anomalies for equatorial eastern Pacific. (Source: NOAA)

Now we can combine this information with the ENSO forecasts. The chart at left shows an ensemble of forecasts of the equatorial sea-surface temperature anomaly in the eastern Pacific – a key signature of El Niño. The periods are 3-month rolling averages; NDJ is November-December-January, for example. Almost all models suggest that the present El Niño will persist through late spring and early summer of 2016.

We took this information and asked a simple question: For periods when the MEI has been above 0.75, what was the average wind speed anomaly? The maps below show the answer for North America.

Mean wind speed anomalies over North America, in percent, for (left) December-January-February (DJF) and (right) March-April-May (MAM), during  El Niño events. For this purpose an El Niño event is defined as an average 2-month trailing MEI greater than 0.75 over the corresponding 3 month period. The historical period covers 1988-2014 and includes 8 El Niño events in winter (DJF) and 6 in spring (MAM).

Mean wind speed anomalies over North America, in percent, for (left) December-January-February (DJF) and (right) March-April-May (MAM), during El Niño events. For this purpose an El Niño event is defined as an average 2-month trailing MEI greater than 0.75 over the corresponding 3 month period. The historical period covers 1988-2014 and includes 8 El Niño events in winter (DJF) and 6 in spring (MAM).

The result, unfortunately, is a little discouraging. If the average historical pattern holds, average wind speeds will be around 3% to 6% below average in the United States this coming winter and spring.

Of course, the average historical pattern may not hold: that’s the problem with seasonal weather forecasts! There have been times when winter and spring winds were near or above normal during El Niño episodes. Moreover, El Niño may fade more rapidly than forecast, making the spring outlook especially uncertain.

Still, based on what we know right now, I wouldn’t bet on winds being above normal in the first few months of next year.

(Many thanks to Katherine Rojowsky and Llorenç Lledó for doing the analysis behind this report.)

Dr. Michael Brower can be reached at michael.brower@awstruepower.com.

 

 

 

 

 

Back to #smarterpower (blog)

4 Responses to The North American ‘Wind Drought’: Is it the new normal?

  1. […] showing after a challenging first half of 2015 when much of the country experienced a severe “wind drought.” The India wind industry, at 4.9% below normal, was not so fortunate. China (-0.8%) […]

  2. […] showing after a challenging first half of 2015 when much of the country experienced a severe “wind drought.” The India wind industry, at 4.9% below normal, was not so fortunate. China (-0.8%) […]

  3. […] showing after a challenging first half of 2015 when much of the country experienced a severe “wind drought.” The India wind industry, at 4.9% below normal, was not so fortunate. China (-0.8%) […]

  4. […] Still, based on what we know right now, I wouldn't bet on winds being above normal in the first few months of 2016.Michael Brower is president and chief technical officer of Albany, N.Y.-based AWS Truepower. He can be reached at michael.brower@awstruepower.com. This article was excerpted with permission from AWS Truepower. To read the full blog post, "The North American "Wind Drought': Is it the new normal?" click here. […]

Leave a Reply

Your email address will not be published. Required fields are marked *


*