Classification Documentation

Classification of El Niño and La Niña years in previous posters (1900 to 1997 and 1890 to 2004)

Please note classification of El Niño–Southern Oscillation years on previous posters differs from that described on the current poster (1890 to 2017)

A major component of the sustainable management of Australia's natural resources is management for climatic variability. Knowledge of historical climatic variability and likely causes is a key requirement. The major cause of year-to-year variability in eastern Australian rainfall is the El Niño-Southern Oscillation with 'El Niño' years having a greater chance of below median rainfall and 'La Niña' years having a greater chance of above median rainfall.

It has been necessary to classify years since 1890 into categories 'El Niño', and 'La Niña' for the purposes of publishing the poster of historical rainfall maps and the report describing degradation episodes. There is no single classification of historical years into 'El Niño' and 'La Niña'. For example, different historical time-series classification have been compiled by Quinn, Whetton and Allan et al. 1996. Currently on-line, there are several lists compiled by the NOAA Climate Prediction Centre (http://www.cpc.ncep.noaa.gov), Washington State University (http://www.atmos.washington.edu), Bureau of Meteorology (http://www.bom.gov.au). In the following description, we set out our procedure to classify years for the purpose of the poster and the publication of the report "Land and pasture degradation episodes in Australia's natural grazing lands".

When we first attempted this task in 1996 for the poster Twelve Month Australian Rainfall Relative to Historical Records (Peacock and Flood) we used an analysis done by Dr Rob Allan then of CSIRO and now UK Met Office. Allan, R.J., Beard, G.S., Close, A., Herczeg, A.L., Jones, P.D. and Simpson, H.J. (1996a). Average sea level pressure indices of the El Niño-Southern Oscillation: Relevance to stream discharge in south-eastern Australia. CSIRO Division of Water Resources, Divisional Report 96/1. Dr Allan is a recognised expert on El Niño-Southern Oscillation and had reconstructed the Southern Oscillation Index back to the 1870s for use in RAINMAN. In his report 'Mean Sea Level Pressure Indices of the El Niño–Southern Oscillation: Relevance to Stream Discharge in South-Eastern Australia' describing the reconstruction of the Southern Oscillation Index (SOI), he included tables classifying years into 'El Niño' and 'La Niña' years, based on the average value of SOI from June to November. The tables were labelled 'El Niño and 'La Niña' years and hence provided an independent reference by an acknowledged expert for the use of these terms in the poster. We now wish to update the poster and hence we need to be able to develop an algorithm that classifies past and future years into the same categories using readily available SOI time-series but maintain the integrity of the original poster.

Step 1 - We examined currently available historical records of monthly SOI to determine whether the classification based on the SOI would produce the same years that Dr Allan had tabulated in his 1996 report. Several SOI time-series are available and variation can occur because of differences in calculation procedure, the base period used to calculate anomaly pressures and how missing data have been reconstructed. We first used the SOI time-series from UK Met Office (www.metoffice.gov.uk). We found that applying the simple rule that Dr Allan had used with his original SOI time-series did not reproduce the identical classification in 6 years 1895, 1898, 1904, 1915, 1945, 1962. The rule that he had proposed was that 'El Niño' years were where the average SOI from June to November was less than or equal to -5, and La Niña years were when the average SOI was greater than or equal to +5. We found that by widening the SOI criteria to +/- 5.5, we were able to reproduce Dr Allan's original classification with only three outliers (1895, 1902, 1915).

We then used the more readily available SOI time-series from the Bureau of Meteorology (http://www.bom.gov.au) to examine whether similar criteria would reproduce those years. We found that the only differences from Dr Allan's original classification were in the years 1951 and 1957.

We then examined explicitly these individual years to assess whether they should be classified as 'El Niño' or 'La Niña'. We investigated two other time-series of reported El Niño/La Niña events. The Climate Prediction Centre (CPC) from NOAA National Weather Service, Camp Springs, Maryland, USA classified years as 'El Niño' and 'La Niña', commenting that the episodes tend to reach their peak during a northern hemisphere winter. Hence, their classified years span two different calendar years. They also indicated that Pacific warm (El Niño) or cold (La Niña) episodes can last for several years. Episodes were also classified in terms of their intensity as either 'weak', 'moderate' or 'strong'.

A second set of classifications is available from Washington State University (WSU), web address http://www.atmos.washington.edu. Years were classified based on northern hemisphere winters December to February and whether "NINO3.4 exceeds 0.5 standard deviations or about 0.47 degrees C". The indicated year is the year at the end of the northern hemisphere winter (February), i.e. one year later than Dr Allan's indicated year. Using this information 1915 was classified as a 'neutral' year, although the reconstructed UK Met Office SOI values classified it as a La Niña; 1895 was a 'neutral' year since it does not appear in WSU or CPC list; 1951 was an 'El Niño' year as it appears in both the WSU and CPC lists; 1957 was an 'El Niño' year since it occurs in both WSE and CPC lists. Hence, the attached list was generated using the BoM SOI, criteria of average SOI (June-November) + 5.5, and editing 1951 as an 'El Niño' year. The only variation to the existing poster is the reclassification of 1957 as an El Niño year.

La Niña period 1998/99 to 2000/01
With regard to recent years, the CPC list 1998 to 2001 as a 'cold' episode with the comment 'strong during northern winter seasons 1998-99, 1999-00, moderate during 2000-01'. The classification using BoM SOI indicates 1998/99 and 2000/01 as La Niña years but not 1999/00. In the attempt to maintain rigour we have left the 1999/00 as a neutral year even though it rained like a La Niña in western NSW and western Queensland.

La Niña period 2007/08
This La Niña period was not a typical episode. Although it was anticipated as early as April 2007 (e.g. La Niña – Detailed Australian Analysis) the SOI did not go positive (>+5) until October 2007.

El Niño period 2009/10 
This El Niño period was not a typical episode with the Southern Oscillation Index not going negative (<-5) until October 2009. However the World Meteorological Organisation (WMO) state that the El Niño event began in June 2009 based on warm sea surface temperature anomalies in the equatorial Pacific (WMO http://www.wmo.int/pages/prog/wcp/wcasp/documents/El_Nino_Dec09_Eng.pdf (PDF))

CVAP report classifications
Two further classification systems were used in the CVAP report QNR14 Can seasonal forecasting prevent land and pasture degradation in Australia's grazing lands. The SOI was used in combination with the Inter-decadal Pacific Oscillation (IPO) to classify years. The SOI was calculated from June-November following Dr Allan's previous analysis. The criteria were widened to be +/- 4 to increase the number of years in the group of IPO-positive - SOI-positive years to 11 allowing more accurate estimate of median rainfall under these conditions. As a result, several years fell in the zone when the absolute value of the June-November SOI was between 4 or 5 units these being 1962, 1939, 1922, 1942, 1895, 1976. Analysis of the WSU and CPC lists supported the inclusion of these years as either 'El Niño' (1939, 1976) and 'La Niña' (1922, 1942, 1962) type even though the absolute SOI value was no greater than 5.5. The year 1895 was included as an 'El Niño' as the SOI values from Rob Allan's SOI time series was -6.8.

A second classification was made based on June-October SOI using the same absolute criteria of 4 SOI units. This resulted in four new years being classified as 'El Niño' or 'La Niña', e.g. 1899, 1920, 1931, 1989. Whilst this classification was necessary for use of evaluating forecasts for November-March period, the classification is limited in that it does not include the subsequent development of ENSO after October. In these cases, further development of warm or cool episodes did not occur based on CPC lists. As the Washington State and CPC analysis indicates the December-February period is important in making some judgement of whether a year should be classified as El Niño or La Niña.

Citation
This classification has been described in the following report:
McKeon, G.M., Hall, W.B., Henry, B.K., Stone, G.S. and Watson, I.W. (2004), Pasture degradation and recovery in Australia's rangelands: Learning from History. Queensland Department of Natural Resources, Mines and Energy, pp.256.

Last updated: 8 November 2018