NOAA: Major El Niño Still on Track

In its latest monthly outlook, issued on Thursday (see PDF), NOAA continues to project that the ongoing El Niño event, already close to record strength for August, will at least approach the highest overall strength observed at any time of year since 1950. As of last week (see PDF), sea-surface temperatures across a key part of the eastern tropical Pacific called Niño3.4 were running 1.9°C above the long-term average for this time of year. This month’s Niño3.4 values could end up warmer than those for any other August in the official NOAA database, which goes back to 1950. The most recent value of NOAA’s closely watched Oceanic Niño Index, which is based on three-month averages for Niño3.4, was +1.0°C for May-July 2015, which ranks behind only 1987 (+1.1°C) for May-July readings. The NOAA outlook released on Thursday notes that the atmosphere-ocean coupling remains strong across the tropical Pacific, with weaker-than-average trade winds. Also, showers and thunderstorms have shifted toward the central and eastern equatorial Pacific from the west. “Collectively, these atmospheric and oceanic features reflect a significant and strengthening El Niño,” noted the outlook.


Figure 1. Early-August status of the 1997 and 2015 El NIño events in terms of satellite-derived data showing departure from average sea-surface height for a given time of year, which is correlated with warmth in the upper ocean. This animation shows the side-by-side evolution of both events. Image credit: NASA/JPL.


More signs of a barn-burner El Niño can be gleaned from the international array of computer models scrutinized by the Australian Bureau of Meteorology. Each month BOM calculates a multimodel average of Niño3.4 values going out for several months. In July, BOM’s multimodel average indicated that the Niño3.4 anomaly (departure from the seasonal norm) will rise to around +2.6°C by October and +2.7°C by December. This would imply a three-month average (Oct-Dec) of between +2.6° and +2.7°C. Were this to materialize, it would be well above the previous record three-month average in the NOAA database of +2.3°C, observed in Sep-Nov and Oct-Dec 1997. Update (15 August]: The outlook from NOAA's Climate Forecast System, version 2 (CFSv2], which is regularly cited in the BOM outlook, has hovered close to a maximum Niño3.4 anomaly of +3.0°C in its forecasts over the last few weeks. Shown below is the "PDF-corrected" version of NOAA's CFSv2 outlook, which has been consistently in the neighborhood of a +2.0°C maximum. Studies indicate that the PDF correction can improve the anomaly predictions for the Niño3.4 region, although in this case the PDF-corrected outlook lies substantially below the multimodel average cited above. Thanks to WU reader Cheyne Mosher for calling attention to this.


Figure 2. NOAA’s Climate Forecast System model (CFSv2) continues to show El Niño intensifying into this autumn, then decreasing in early 2016, in a fairly typical pattern for a strong El Niño event. Niño3.4 sea-surface temperatures in this PDF-corrected version are projected by CFSv2 to rise to about 2.0°C above the seasonal norm. Without the PDF correction, CFSv2 predicts a maximum closer to 3.0°C. Both versions can be tracked here. The panels at right show SSTs for three-month windows from Aug-Oct 2015 (top left) to Feb-Apr 2016 (lower right). Image credit: NOAA.


There are well-known limits to how well models can simulate El Niño, and even a solid model can be temporarily “fooled” by short-term changes in the tropical Pacific. Nevertheless, the general consistency in El Niño outlooks across models and across time--and the steadily building warmth across the eastern tropical Pacific, both at and below the surface--suggests that an event as strong or stronger than any observed in modern times is still a real possibility. That said, NOAA forecasters stressed in a Thursday morning news conference that El Niño is not guaranteed to bring drought relief to California. In the crucial water-storage region of the central Sierra, for examples, the last four years brought only 56% of the cumulative average precipitation from October 2011 through July 2015, leaving a 71-inch deficit. To make this up, the region would need roughly 2.5 to 3 times its annual average precipitation over the coming year, said Kevin Werner, director of climate services for the National Weather Service’s Western Region. “We’d need something in excess of the wettest year on record to balance the four-year deficit,” Werner said. This message was reinforced by Mike Halpert, deputy director of NOAA’s Climate Prediction Center. “Just because something is favored, it doesn’t guarantee it will happen. One season of above-normal rain and snow is very unlikely to erase four years of drought,” said Halpert.

For more background on what impacts we might expect from El Niño over the next few months, see our North American roundup post from July 28 and our special post on potential Northeast U.S. impacts. Jeff Masters will take a look at global impacts of El Niño in a forthcoming post.

What happened to the El Niño of 2014-15?
If you thought we were already in an El Niño episode a few months ago, you might be puzzled to see that the official NOAA database no longer shows it. This change is due to a fairly minor update in the ocean temperature record that pulled one key period just below the required threshold.


Figure 3. Departures from seasonal average for sea-surface temperatures (SSTs) across the Niño3.4 region over three-month intervals from 1997 to 2015. The shaded box (Jan-Mar 2015), originally 0.5°C, was “demoted” to 0.4°C with a July upgrade to SST analysis techniques. Image credit: NOAA Climate Prediction Center.


To qualify as a full-fledged El Niño episode, the Niño3.4 departure must be sustained at +0.5°C or greater for at least five overlapping three-month-long periods. When NOAA analysts are tracking El Niño in real time, they rely on a series of daily and weekly analyses of sea-surface temperature called OISST (optimum interpolation SST). OISST incorporates data from a variety of sources, including satellite-based measurements that are useful for short-term needs but which can introduce biases if they’re folded into a longer-term dataset that predates a particular satellite. For those longer-term purposes, the NOAA National Centers for Environmental Information (NCEI, formerly NCDC) produces the Extended Reconstructed Sea Surface Temperature (ERSST), a monthly dataset that goes back to 1854 and that uses statistical techniques to fill in data gaps.

Every few years, the techniques used in ERSST are updated and a new version of the full dataset is released. This occurred with the advent of ERSSTv4 this past July. As it happens, ERSSTv4 brought down the Niño3.4 anomaly for January through March 2015 from +0.5°C to +0.4°C (see Figure 3). Even though every other three-month window since Oct-Dec 2014 is still at +0.5°C or greater, this “demotion” of Jan-Mar 2015 means that we have yet to see five consecutive three-month periods of El Niño. Without those five periods, we haven’t yet met the formal definition for an El Niño episode (shown as red intervals in the NOAA historical database and in Figure 3 above). By next month, though, we’ll have five consecutive periods, and the El Niño episode will again become official, extending back to Feb-Apr 2015. The ERSSTv4 introduction also downgraded a few other El Niño and La Niña events from the last 65 years. These are identified by Jan Null (Golden Gate Weather Services) on a website that classifies El Niño and La Niña events by strength. One more wrinkle: it’s possible that Jan-Mar 2015 could be “undemoted” next year, when the climatology that underlies the above- and below-normal categorizations goes through a scheduled five-year update that takes long-term warming into account. In any case, the atmospheric response to the warm Niño3.4 readings in the winter of 2014-15 fell short of the usual El Niño standards, according to Michelle L’Heureux (NOAA Climate Prediction Center), so it’s best not to place too much emphasis on this borderline event.

The Case of the Not-Quite El Niño Episode reminds us of an important point stressed by L’Heureux: “Our SST observations are estimates. This is why we always encourage looking at multiple indices and datasets when trying to assess the state of ENSO. No one dataset or index will ever be perfect.” The further back in time we go, the more piecemeal is our knowledge of the SSTs that prevailed at the time. This is why NOAA’s most commonly used database of ENSO episodes only extends back to 1950. It’s quite possible to use ERSSTv4 or analyses from other research centers to calculate Niño3.4 values prior to 1950. However, these must be used with caution, as the data become increasingly scant going back in time (see Figure 4 below). To help flesh out the picture, scientists look to independent measurements related to El Niño, such as the Southern Oscillation Index, based on the observed difference in barometeric pressure between Darwin, Australia, and Tahiti. The greatest confidence in pre-1950 ENSO history is for the very strongest events, which are typically reflected in a wide range of land-based repercussions consistent with El Niño and La Niña behavior. For example, the El Niño of 1877-78 appears to have been at least as strong as the “super” 1982-83 and 1997-98 events. Calculations based on ERSSTv4 by wunderground member Eric Webb (@webberweather) suggest that the Niño3.4 value topped 2.5°C for several months. Drought associated with the 1877-78 El Niño may have contributed to horrific multiyear famines that took an estimated 5 million lives in India and 9-13 million lives in China.

For an update on the latest tropical cyclone action, including a fizzling Tropical Storm Hilda and twin typhoons predicted to develop next week, see this morning's post by Jeff Masters.

Bob Henson


Figure 4. Distribution of sea surface temperature observations from the International Comprehensive Ocean Atmosphere Data Set for each 20-year period from 1860 to 1979. This dataset underpins the NOAA ERRST long-term reanalysis discussed above. Color shading indicates the percentage of months that have at least one measurement within a 2°-latitude by 2°-longitude grid box (roughly 140 by 140 miles near the equator). Image credit: Used with permission from “Sea Surface Temperature Variability: Patterns and Mechanisms,” Annual Review of Marine Science 2009, doi: 10.1146/annurev-marine-120408-151453, courtesy Clara Deser, National Center for Atmospheric Research.