Multivariate ENSO Index (MEI)

Last update: 5 August 2009

The views expressed are those of the author and do not necessarily represent those of NOAA.

El Niño/Southern Oscillation (ENSO) is the most important coupled ocean-atmosphere phenomenon to cause global climate variability on interannual time scales. Here we attempt to monitor ENSO by basing the Multivariate ENSO Index (MEI) on the six main observed variables over the tropical Pacific. These six variables are: sea-level pressure (P), zonal (U) and meridional (V) components of the surface wind, sea surface temperature (S), surface air temperature (A), and total cloudiness fraction of the sky (C). These observations have been collected and published in COADS for many years. The MEI is computed separately for each of twelve sliding bi-monthly seasons (Dec/Jan, Jan/Feb,..., Nov/Dec). After spatially filtering the individual fields into clusters (Wolter, 1987), the MEI is calculated as the first unrotated Principal Component (PC) of all six observed fields combined. This is accomplished by normalizing the total variance of each field first, and then performing the extraction of the first PC on the co-variance matrix of the combined fields (Wolter and Timlin, 1993). In order to keep the MEI comparable, all seasonal values are standardized with respect to each season and to the 1950-93 reference period. The MEI is extended during the first week of the following month based on near-real time marine ship and buoy observations (courtesy of Diane Stokes at NCEP) summarized into COADS-compatible 2-degree monthly statistics at NOAA-ESRL PSD. Caution should be exercised when interpreting the MEI on a month-to-month basis, since the input data for updates are not as reliable as COADS, and the MEI has been developed mainly for research purposes. Negative values of the MEI represent the cold ENSO phase, a.k.a.La Niña, while positive MEI values represent the warm ENSO phase (El Niño).

You can find the numerical values of the MEI timeseries under this link, and historic ranks under this related link. You are welcome to use any of the figures or data from the MEI websites, but proper acknowledgment would be appreciated. Please refer to the (Wolter and Timlin, 1993, 1998) papers (NOW available online as pdf files!), and/or this webpage.

If you have trouble getting the data, please contact me under (Klaus.Wolter@noaa.gov)

How does the 1998-2000 La Niña event compare against the seven previous biggest La Niña events since 1949? Only strong events (with a peak value of at least -1.2 sigma) are included in this figure. Note that some events last through the full three years shown here (for instance, 54-56), while others revert to "normal" or El Niño conditions by the second or third year (especially in 64-66). The 1998-2000 La Niña does not resemble any previous event in this comparison figure. It started late (about three months later than the previous latest case), and it featured a superimposed annual cycle (peaking around May and troughing around November) that does not match the other events displayed in this figure. However, the weak La Niña period after the 1982-83 El Niño had similar characteristics. Click on the "Discussion" button below to find a new comparison of recent MEI conditions against several La Niña events that transitioned into El Niño in the same calendar year.

  • Discussion and comparison of recent conditions with historic El Niño events

    • How does the 2002-04 El Niño event compare against the seven previous biggest El Niño events since 1949? Aside from 2002-04, only strong events (with a peak value of at least +1.4 sigma) are included in this figure. The 2002-03 El Niño event peaked below that threshold, with just over +1.2 sigma in early 2003. Overall, I would rank it just barely in the top 10 El Niño events of the last half century. In its evolution, it bears some resemblance to the 1965-67 event (highest temporal correlation), but shared with 1991-93 its reluctance to drop below the zero line once it had run its course. The most recent El Niño event of 2006-07 reached a similar peak as the 2002-03 event, but lacked 'staying power', and collapsed rather early in 2007. Click on the "Discussion" button below to find a new comparison of recent MEI conditions against several La Niña events that transitioned into El Niño in the same calendar year.

    The six loading fields show the correlations between the local anomalies and the MEI time series. Land areas are flagged in green, and typically noisy regions with no coherent structures and/or lack of data are shown in grey. Each field is denoted by a single capitalized letter and the explained variance for the same field in the Australian corner.

    The sea level pressure (P) loadings show the familiar signature of the Southern Oscillation: high pressure anomalies in the west and low pressure anomalies in the east correspond to positive MEI values, or El Niño-like conditions. Consistent with P, U shows positive loadings centered along the Equator, corresponding to westerly anomalies near the dateline during El Niño-like conditions. The meridional wind field (V) features negative loadings north of the Equator across the eastern Pacific basin, denoting the southward shift of the ITCZ so common during El Niño conditions, juxtaposed with large positive loadings northeast of Australia.

    Both sea (S) and air (A) surface temperature fields exhibit the typical ENSO signature of a wedge of positive loadings stretching from the Central and South American coast to just east of the dateline, or warm anomalies during an El Niño event. At the same time, total cloudiness (C) tends to be increased over the central and western equatorial Pacific (east of Indonesia).

    The MEI now stands for 20.7 % of all six fields in the tropical Pacific from 30N to 30S, one month after its annual minimum of importance. For comparison, this value is 2.3% lower than the one registered 10 years ago, attesting to an overall weakening of ENSO variability in the last decade. The loading patterns shown here resemble the seasonal composite anomaly fields of Year 0 in Rasmusson and Carpenter (1982).

    Consistent with moderate El Niño conditions, there are several of the observed key anomalies in the MEI component fields that exceed or equal one standard deviation, or one sigma (compare to loadings figure) that also flag typical El Niño features, while a few La Niña features are still lingering as well. Significant positive anomalies (coinciding with high positive loadings) denote positive sea surface and air temperature anomalies (S, A) west of South America. Significant negative anomalies (coincident with high negative loadings) denote easterly wind anomalies (U) off the Pacific coast of Mexico and negative SST anomalies (S) east of Australia, also consistent with El Niño. On the other hand, negative sea level pressure anomalies (P) east of Papua New Guinea and positive SST anomalies (S) in the same region are more typical during La Niña conditions.

    Go to the discussion below for more information on the current situation.

    If you prefer to look at anomaly maps without the clustering filter, check out the climate products map room.

    Discussion and comparison of recent conditions with historic El Niño events

    In the context of recent positive MEI values, this section features a new comparison figure of weak-to-moderate El Niño events that emerge from La Niña conditions in the same calendar year. Note that there was a 30-year hiatus with no such transitions between 1976 and 2006, attesting mostly to the lack of La Niña conditions in this period (as well as the lack of fast transitions into El Niño from La Niña).

    The most recent (June-July) MEI value has risen a little higher to reach +0.90, its highest value since January 2007. The most recent MEI rank has now reached 50th out of 60, right above the quintile (upper 20%) threshold for MEI rankings for this season. I continue to classify this as a border-line moderate El Niño condition for the MEI.

    Negative SST anomalies have disappeared from the equatorial Pacific over the course of a single month (April), while positive anomalies of close to +1C continue along the equatorial cold tongue from the coast of South America westward to about 150W, at least as reflected in the latest weekly SST map.

    For an alternate interpretation of the current situation, I highly recommend reading the latest NOAA ENSO Advisory which represents the official and most recent Climate Prediction Center opinion on this subject. In its latest update (9 July 2009), El Niño conditions are diagnosed, and are expected to last through the Northern Hemisphere winter of 2009-10.

    There are several other ENSO indices that are kept up-to-date on the web. Several of these are tracked at the NCEP website that is usually updated around the same time as the MEI, just in time for this go-around. Niño regions 3 and 3.4 showed persistent temperature anomalies of -1.3C or lower from September (region 3) and October 2007 (region 3.4) through February 2008, but hovered much closer to 0C anomalies between about June and November 2008. After another La Niña interlude from December 2008 through March 2009, both regions have seen a rise to above 0C anomalies in May 2009. In fact, the most recent monthly anomaly of +0.9C for Niño 3.4 was last exceeded in December 2006. For extended Tahiti-Darwin SOI data back to 1876, and timely monthly updates, check the Australian Bureau of Meteorology website. This index belatedly caught on to the La Niña event in November 2007, reaching +12 to +14 (1.4 sigma) in December, January and March, and higher yet (+21) in February 2008. The average for February-March 2008 (+16.8) was ranked fourth highest for that season (since 1950, but this ranking is true for the full record as well). It then hovered near 0 from April through July 2008, while the August SOI rebounded right up to +9, followed by values at or above that value right through February 2009. While oscillating around 0 since then, it went mildly negative in May and stayed below 0 in June (-2), confirming the demise of the most recent La Niña episode in all ENSO indices. However, its July value of +2 runs counter to all other ENSO indices for this month, virtually the same as last July's average. It remains noteworthy that the SOI was the ENSO index with the strongest La Niña footprint in 2008-09, and that it currently shows the weakest amplitude in terms of El Niño-like conditions, in fact, it is the only ENSO index to not show El Niño conditions at all in July. An ever longer Tahiti-Darwin SOI (back to 1866) is maintained at the Climate Research Unit of the University of East Anglia website, however with less frequent updates (currently through December 2008). Extended SST-based ENSO data can be found at the University of Washington-JISAO website, currently updated through June 2009.

    Stay tuned for the next update (by September 5th) to see where the MEI will be heading next. We have now entered the time of year when rapid changes of the MEI become less common, and the recent switch to El Niño is now 'locked in' for at least several months. This is a major change compared to my assessment a few months ago, but consistent with last month's discussion that such a big rise in the MEI and such high MEI values as in the last two months have always been followed by continued El Niño conditions through the remainder of the calendar year, at least in the post-1950 MEI record. This does not exclude an early weakening of this event either, in fact, the best 'analog' year of 1951 ended up with very weak El Niño conditions by the end of that year. Note that I have discontinued my monthly e-mail announcements in favor of keeping the discussion right here on this webpage.

    REFERENCES

    • Rasmusson, E.G., and T.H. Carpenter, 1982: Variations in tropical sea surface temperature and surface wind fields associated with the Southern Oscillation/El Niño. Mon. Wea. Rev., 110, 354-384. Available from the AMS.
    • Wolter, K., 1987: The Southern Oscillation in surface circulation and climate over the tropical Atlantic, Eastern Pacific, and Indian Oceans as captured by cluster analysis. J. Climate Appl. Meteor., 26, 540-558. Available from the AMS.
    • Wolter, K., and M.S. Timlin, 1993: Monitoring ENSO in COADS with a seasonally adjusted principal component index. Proc. of the 17th Climate Diagnostics Workshop, Norman, OK, NOAA/NMC/CAC, NSSL, Oklahoma Clim. Survey, CIMMS and the School of Meteor., Univ. of Oklahoma, 52-57. Download PDF.
    • Wolter, K., and M. S. Timlin, 1998: Measuring the strength of ENSO events - how does 1997/98 rank? Weather, 53, 315-324. Download PDF.

    Questions about the MEI and its interpretation should be addressed to:
    (Klaus.Wolter@noaa.gov), (303) 497-6340.