Multivariate ENSO Index (MEI)

Last update: 3 February 2010

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 in early 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 along the Equator, corresponding to westerly anomalies near the dateline. Negative loadings in the far western and eastern Pacific, as well as to the south of the positive loading center, show that easterly anomalies are almost equally pervasive in these regions during El Niño. The meridional wind field (V) features high negative loadings north of the Equator, denoting the southward shift of the ITCZ so common during El Niño conditions, juxtaposed with even stronger 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 the dateline, or warm anomalies during an El Niño event. They are flanked by negative loadings (cold anomalies) to its southwest and, to a lesser degree, to its northwest. At the same time, total cloudiness (C) tends to be increased over the central equatorial Pacific, sandwiched in between decreased cloudiness north of Australia and the eastern-most equatorial Pacific.

    Now just past its annual peak, the MEI explains 31.1% of the total variance of all six fields in the tropical Pacific from 30N to 30S. Although its temperature components dominate the MEI with over 40% of their possible variance, even P, V, and C join in with about a third, a quarter, and a fifth of their variance, respectively, at or close to their peak values during the year. The loading patterns shown here resemble the seasonal composite anomaly fields of Year 1 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 no lingering La Niña feature reaches the one sigma threshold. Significant positive anomalies (coinciding with high positive loadings) denote anomalously high sea level pressure (P) just east of Australia, strong westerly wind anomalies (U) near the Equator and dateline, very strong southerly wind anomalies (V) east of Australia, and positive sea surface and air temperature anomalies (S, A) in the central and eastern equatorial Pacific basin, as well as near Indonesia. It is noteworthy that the highest equatorial SST anomalies (almost two standard deviations) occur near the dateline where they have projected strongly on the Niño 3.4 SST index. Significant negative anomalies (coinciding with high negative loadings) denote strong easterly wind anomalies (U) north of Indonesia.

    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 transitions from La Niña to El Niño events

    In the context of recent positive MEI values, this section features a 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 from La Niña into El Niño).

    The most recent (December-January) MEI value has increased slightly to +1.16, its highest value since October-November 2006. The most recent MEI rank has risen from 50th (11th highest) out of 60 to 52nd (10th highest) out of 61, above the quintile (upper 20%) threshold for MEI rankings for this season. After its lapse from 10th to 15th highest rank in August-September 2009, this is the fourth month in a row that I would classify as moderate El Niño conditions for the MEI. If one compares the rankings of this event to other El Niño events of the last decade (2002-03, 2004-05, and 2006-07) from May-June of Year 0 to December-January of Year 1 (8 values for each event), the current event slightly outranks all three of these events to render 2009-10 as the strongest El Niño since 1997-98.

    Negative SST anomalies disappeared from the equatorial Pacific over the course of a single month (April), while positive anomalies have become more concentrated east of the dateline over the last few months, with peak values in excess of +2C just east of the dateline, 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 (7 January 2010), El Niño conditions are diagnosed, and are expected to continue at least into the Northern Hemisphere spring of 2010.

    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 have shown persistent anomalies above +0.5C since June 2009, with a recent pekk in monthly anomalies to reach +1.6C for Niño 3 and +1.8C for Niño 3.4 in December 2009, only to drop by 0.6C and 0.2C in January 2010, respectively. Nevertheless, the Niño 3.4 anomaly averaged +1.5C or higher for the last three months, hence it will probably be classified as a 'strong' event by CPC, higher than is justifable by any other ENSO index (Niño 3, SOI, MEI, ...). For extended Tahiti-Darwin SOI data back to 1876, and timely monthly updates, check the Australian Bureau of Meteorology website. This index has often been out of sync with other ENSO indices in the last few years, and did show near-normal conditions from June through September. In October 2009, it dropped by an unprecedented 1.86 standard deviations to reach -15 (-1.5 sigma), its lowest value since October 2006, and much better aligned with other ENSO indices than before. November and December showed a rebound to -7 (-0.7 sigma), while January 2010 saw a drop back down to -10 (-1 sigma). 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 November 2009). 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 March 6th) to see where the MEI will be heading next. During this time of year rapid changes of the MEI are less common, and the three most recent monthly increases by less than 1/10 each fit into that mold. At this point, it would be a surprise if this event were to grow any further than the current value of almost +1.2 standard deviations, in fact all 'analog' cases in the above figure declined after the December-January season. 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.