It’s long been clear that the tropical Pacific Ocean influences seasonal climate in many parts of the globe. Here in the Northwest, we often have somewhat warmer, drier winters during a tropical El Niño event. So it’s no surprise that dozens of researchers have looked at details of sea surface temperature (SST) patterns to understand specific climate anomalies.
That’s what Richard Seager and Naomi Henderson did in a recent effort to understand California’s drought. Their paper, published in The Journal of Climate, focuses on precipitation during the dry winter of 2013-2014.
In their work, the researchers used an atmosphere-only model (the vintage Community Climate Model, CCM3, of the National Center for Atmospheric Research), changing SST in different parts of the Pacific in order to see which parts were most likely to lead to the atmospheric ridge of higher-than-normal pressure that effectively parked itself off the West Coast and produced the drought.
Seager and Henderson ran the model 100 times for some experiments, providing adequate characterization of the uncertainties. (If you’re new to climate modeling, you can learn more on our website.) The researchers’ results suggest that the observed drought-making ridge was in fact only partly caused by SST.
The researchers came to this conclusion by comparing the atmospheric responses to SSTs actually observed with responses to SSTs that were different only in one patch or another of the ocean. But here the plot thickens in an unexpected way: We can’t even be sure what “observed” really means, even in this era of satellite observations and high-quality thermometers found on almost every major ship traversing the world’s oceans. The datasets available are painstakingly developed but end up with slight—and for these purposes—important differences.
The researchers employed three different gridded datasets of SSTs that differed by as much as one degree Fahrenheit in some places. Now, this may not sound like much, but this little degree of difference led to notable differences in the atmospheric simulations and, most important, in the precipitation anomalies simulated on the US West Coast. Because these are exactly the sorts of meteorological phenomena that many suspect led to that drought-creating ridge, having such a disagreement in SSTs really matters if you’re trying to determine the role SSTs played in the recent drought.
The experiments Seager and Henderson ran comparing “observed” SSTs to their computer simulations in which they manipulated SSTs looked something like this: Using their atmosphere model, the researchers changed SSTs relative to the long-term average in five latitude-longitude boxes centered on the equator. The experiments were run one box at a time. The five boxes ranged from a box in the Indian Ocean to a box in the eastern equatorial Pacific. Seager and Henderson found that by carefully choosing anomalies in two of the boxes they could match the pattern, but not the amplitude, of the observed ridge.
Much like safecrackers listening intently with a stethoscope for the tell-tale clicks of a safe’s dial, the researchers found the right setting of oceanic knobs to make a California drought.
Seager and Henderson also took advantage of having a large ensemble—remember, they ran the model 100 times for some SST configurations—to see whether any individual simulations had a ridge with the strength of the observed ridge, and indeed there were. Moreover, the distribution of “ridgeiness” in their forced runs was significantly closer to the observed pattern than an ensemble with regular SSTs.
In short, the ridge that led to the drought was a product both of the specific SST pattern that year and of internal (unforced) atmospheric variability. The role of SSTs can help predict future droughts.
STUDY: The Journal of Climate
Citation: Seager, Richard, and Naomi Henderson. “On the Role of Tropical Ocean Forcing of the Persistent North American West Coast Ridge of Winter 2013/14 a,” Journal of Climate 29, no. 22 (2016): 8027-8049.
Citation Link: http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-16-0145.1