A data-driven blog that uses chart, maps, cartograms, and other visualizations to understand important societal issues.

Climate, Environment

The World is Getting [en]So Hot: El Niño, La Niña, and Rising Global Temperature

The world has been rapidly warming, at a rate of 1.4°C per 100 years, since 1950 according to NOAA data. The chart illustrates this warming for El Niño years (red), La Niña years (blue), and ENSO-neutral years (grey) (ENSO is the El Niño–Southern Oscillation). This warming trend (dotted lines) is significant regardless of the ENSO signal. Further, there is very little overlap between the ENSO years and trends. For example, since 1950, no La Niña years (blue circles) were warmer than the El Niño trend (red dotted line), and only one El Niño year (red circles; 1976–1977) was cooler than the La Niña trend (blue dotted line).
The data in the chart is ENSO years, not calendar years, running from July one year to June the following year. The ENSO cycle typically does not follow a calendar year, with El Niño and La Niña cycles often starting around late spring or early summer in the northern hemisphere. The traditional description of an ENSO cycle uses the Oceanic Niño Index (ONI) for identifying warmer- and cooler-than normal temperatures in the tropical Pacific. The ONI is the running 3-month mean SST anomaly, and events are defined as five consecutive overlapping 3-month periods at or above the +0.5°C anomaly for warm (El Niño) events and at or below the -0.5°C anomaly for cold (La Niña) events.1 This approach can mischaracterize the impact on global temperatures if looking at calendar years. For example, the calendar year 1998 was the hottest year on record at the time (it’s now only 9th hottest; http://chartedterritory.us/2018/01/23/its-getting-hotter-monthly-global-temperatures-1880-2017/), largely due to the 1997–1998 El Niño being one of the three strongest on record. However, the second half of calendar year 1998 was actually characterized by strong La Niña conditions and 1998–1999 was one of the strongest La Niña cycles on record (as was 1999–2000).
Here we use the concept of ENSO years for both the global temperatures and for defining the ENSO signal. We calculate global temperatures for July-through-June (southern hemisphere years?!) years rather than January-through-December. For example, the latest “year” in the chart is for the ENSO year 2016.5 (or 2016–2017), running from July 2016 through to June 2017 for both global temperatures and the ONI. Further, instead of identifying five consecutive ±0.5°C normal SST months, we want to relate global temperatures to the ENSO behavior through the entire ENSO year. Consequently, we define an ENSO year as La Niña if the SSTs are ⅓°C less than normal, El Niño if the SSTs are ⅓°C greater than normal, and neutral if neither. This approach works very well for ENSO years, but fails if we only used calendar years. For example, 1998 would have a whole-year ONI average of +0.07°C, only the 38th highest value in the 68-yer record (this “mixed- signal” problem does not occur when using ENSO years).
This new approach helps reveal trends that might be obfuscated using calendar years. It more clearly separates the warming trend into El Niño, La Niña, and ENSO-neutral years. And more clearly illustrates the connection between neighboring years. For instance, warmer-than-average El Niña years tend to follow El Niño years (e.g., 1998–1999), and cooler-than-average El Niño year tend to follow La Niña years (e.g., 1976–1977 and 1986–1987). Similarly, like 2017 was the hottest ENSO neutral calendar year on record (http://chartedterritory.us/2018/01/23/its-getting-hotter-monthly-global-temperatures-1880-2017/), the ENSO year 2016–2017 was by far the warmest neutral year on record having followed two consecutive El Niño years (2014–2015 and 2015–2016). [EDIT: the year 2016–2017 was, using the traditional ONI metric, a weak La Niña year.]
Data and software: Global temperatures and El Niño–Southern Oscillation (ENSO) data are sourced from NOAA2,3. Monthly ENSO data are publicly available from NOAA since 1950 (annual data, on a calendar year basis, are available back to 1900). The data was compiled and visualized using Microsoft Excel4.

  1. http://ggweather.com/enso/oni.htm
  2. http://origin.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/ONI_v5.php
  3. https://www.ncdc.noaa.gov/cag/time-series/global/globe/land_ocean/p12/12/1880-2017.csv
  4. https://products.office.com/en-us/excel



  1. philip

    Keeping to calendar months and years is convenient, but its not the way weather and climate works. When having record cold/hot/dry/wet month is not always a good indication because one could have very wet period but spread over Christmas and New Year the very wet period does not appear too wet as it could be spread over 2 separate years.

    • Comment by post author

      Well, that’s not really true here. First, we’re looking at a (strong) trend of year-to-year temperature changes over several decades, and so 12 months is a pretty good temporal unit (less than or more than 12 months makes little sense here). Second, it doesn’t matter much which month we start with; in this case, the trend varies between 1.344°C/100 years (starting with April) and 1.364°C/100 years (starting with August). And third, using “southern hemisphere” years works pretty well when categorizing annual observations by the ENSO signal (and my new ENSO categorization approach seems to work better than the traditional method, which doesn’t necessarily capture the year-long impact of ENSO).

Leave a Reply