Multivariate ENSO Index (MEI)

Last update: 4 September 2008

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 CDC. 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 at 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 comparison of recent MEI conditions against a large subset of these La Niña events.

  • 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. I have discontinued its comparison plot against other weaker events in favor of a new comparison with other La Niña events below (click on "Discussion" to jump to the bottom of the page).

    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: low pressure anomalies in the west and high pressure anomalies in the east correspond to negative MEI values, or La Niña-like conditions. Consistent with P, U shows positive loadings mostly west of the dateline, corresponding to easterly anomalies near the dateline. The meridional wind field (V) features its biggest negative loadings north of the Equator across the eastern Pacific basin, denoting the northward shift of the ITCZ so common during La Niña 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 the dateline, or cold anomalies during a La Niña event. During La Niña conditions, total cloudiness (C) tends to be decreased over the central and western equatorial Pacific, and increased close to equatorial South America.

    The MEI now stands for 23.4% of all six fields in the tropical Pacific from 30N to 30S, having regained almost 5% since May/June. The loading patterns shown here resemble the seasonal composite anomaly fields of Year 0 in Rasmusson and Carpenter (1982).

    Consistent with near-neutral ENSO conditions, there are only a few observed key anomalies in the MEI component fields that exceed or equal one standard deviation, or one sigma (compare to loadings figure), with some of them indicating lingering La Niña features, while others would be consistent with emerging El Niño features. Significant negative anomalies (coinciding with high positive loadings) denote strong easterly anomalies (U) along the Equator and west of the dateline, consistent with La Niña. Significant positive anomalies (coinciding with high negative loadings) flag anomalously high sea level presssure over the eastern equatorial Pacific (P), strong westerly anomalies (U) off the Central American coast, and positive sea surface and air temperature anomalies (S and A) northeast of Australia. All of these anomalies are also consistent with La Niña. On the other hand, large negative anomalies (coincident with high negative loadings) denote strong northerly anomalies (V) south of Hawaii, and reduced cloudiness (C) near Galapagos. Both of these anomaly features are more typical for El Niño.

    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 La Niña conditions, this section displays a comparison figure of historic La Niña events since 1949 vs. current conditions. The most recent (July-August) MEI value has dropped below "0" for the first time in three months (now at -0.21). While running higher than in the 1974 'analog year', it is noteworthy that the 1973-74 period remains the most similar stretch in the MEI record. In 1974, this was followed by a return to La Niña conditions well int o 1976. The most recent MEI rank has dropped further, from 29th to 23rd, still close to the middle of the MEI rankings for this season.

    Weak negative SST anomalies cover the equatorial Pacific from just west of the dateline to about 160W, while positive anomalies have shrunk to cover most of the equatorial Pacific east of about 140W. East Pacific El Niño conditions appear to coexist with Central Pacific La Niña conditions, at least in terms of the latest weekly SST map.

    For an alternate interpretation of the current situation, I 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 August 2008), ENSO-neutral conditions were diagnosed and expected to continue through the northern hemispheric fall season.

    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, but not 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 (region 3.4) through February, but have warmed up considerably since then. Weekly SST data shows that the warming may have peaked in August, with a cooling trend found towards the end of the month. For extended Tahiti-Darwin SOI data back to 1876, and timely monthly updates, check the Australian Bureau of Meteorology website. This index caught on to the current La Niña event last November, reaching +12 to +14 (1.4 sigma) in December, January and March, and higher yet (+21) in February. The average for February-March (+16.8) was ranked fourth for that season (since 1950, but this ranking is true for the full record as well). It has bounced near 0 from April through July, while the August SOI rebounded right up to +9, close to one standard deviation in La Niña territory. 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 more than one year behind). Extended SST-based ENSO data can be found at the University of Washington-JISAO website, currently updated through April 2007.

    Stay tuned for the next update (by October 4th, if there is no 'October surprise') to see where the MEI will be heading next. We have now reached the time of year when persistence will be hard to beat again, although I would not be surprised if the current near-neutral ENSO classification regains a La Niña flavor as we get closer to winter. This slight tilt in the odds is based on MEI behavior during the last decade of fall seasons. Stay tuned for next month's update, I am still not ready to declare the 2007-08 La Niña event over. 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/N MC/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.