The trade winds are one of the most persistent climatological features of the earth. These easterly winds, which exist over much of the global tropics, are also the driving force for equatorial ocean currents. Although the phenomenon of the trade winds has been known for hundreds of years, the relationship between them and ocean has become known only during the past century and may still not be completely understood. The trades converge near the equator at the Intertropical Convergence Zone (ITCZ) where warm moist air rises (Philander 1990). In the Pacific, convection occurs primarily in the west over the warm pool as shown in Fig. 1.1.
Fig. 1.1 Average sea surface temperatures in the Equatorial Pacific.
The term "trade" originally meant "track" or "path," referring to the constant path of the trade winds. The intensity, however, of the wind along this track is not constant. When the trade winds over the Pacific are strong enough, the Western Pacific stays warmer than the Eastern Pacific. In the absence of strong winds the current weakens. The existence of a temperature gradient in the Equatorial Pacific in a weaker-than-normal trade-wind regime allows the equatorial counter current to dominate, deepening the thermocline in the east, and submerging cold waters below the level of upwelling. In turn, surface temperature the Eastern Pacific rise-the phenomenon known as El Niño. The reverse, cooling in the Eastern Pacific, is known as La Niña. Coincident with this pattern of warming and cooling (El Niño and La Niña) is the pressure variation between Darwin, Australia, and Tahiti in the South Pacific Ocean that is known as the Southern Oscillation. Together the two phenomena are called El Niño-Southern Oscillation (ENSO).
Methods of quantifying ENSO phase and strength have expanded as our understanding and the observational network have grown. Two of the earliest metrics used to quantify ENSO are the sea surface temperature (SST) anomalies off the coast of Peru (Niño Index) and the difference in sea level pressure between Darwin, Australia, and Tahiti (Southern Oscillation Index, or SOI). Initially, it was not known that warming off the coast of Peru extended thousands of kilometers west. Until the 1960s El Niño events were recorded only in the waters near Peru (Philander 1990). As late as the 1980s many ENSO definitions still focused primarily on the SSTs along the coast of South America (Trenberth 1997). This area is now known as the Niño 1+2 region (see Fig. 1.2). Several problems exist with the Niño 1+2 time series. First, this region has a weak response to cold events. Additionally, warm phases occasionally arise without significant change of SST in the coastal region-a phenomenon that seems to be occurring more often in recent years (Yu and Kao 2007).
As understanding of ENSO increased, so did the methods of quantifying it. Today, the larger Niño regions 3, 4, and 3.4, which tend to have less noise than Niño 1+2, are more commonly used in defining ENSO events. The Niño 4 has a weak response to warm events and the Niño 3 has the strongest correlation with SOI. Additionally, the Japan Meteorological Agency (JMA) uses a region similar to Niño 3, but confined to 4 degrees from the equator to define the JMA index. Newer methods include combining ENSO metrics [e.g. the Multivariate ENSO Index (MEI) and the Trans-Niño Index (TNI)] or using statistical methods. Weare (1986) defines an index using the first Principal Component (PC) of the Equatorial Pacific SST Empirical Orthogonal Function (EOF). The Niño 3, 3.4, and the SOI are the most widely used indices in operational meteorology in the United States. The National Weather Service (NWS) retrospectively defines ENSO events by the three-month running average SST in Niño 3.4 region. The NWS determines warm (cold) events as any period with of at least five consecutive months with temperatures 0.5°C above (below) normal. For the SOI, negative (positive) values indicate warm (cold) phase. Each measurement of ENSO has limitations, but each demarcates warm and cold events of ENSO nearly unanimously; only the weakest events are sometimes contested. Furthermore, most indices are well correlated with each other (above .9 in some instances), so the choice of index usually depends on the preference of the author. The reader is referred to Hanley et al. (2003) for a more detailed discussion on ENSO indices.
Fig 1.2 Regions used to determine Niño Indices.
These methods of quantifying ENSO do not sufficiently describe how meridional or zonal the area of anomalous SSTs is and whether the event is located in the Central or Eastern Pacific. Since 2007, research into variations involving eastern Pacific warm (EPW) events and central Pacific warm (CPW) events has become an active research topic. Variance of the EPW, the more familiar type of warm event, peeks in the Niño 3 region and will be explained in the following sections. The CPW, often dubbed "Modoki"' ENSO [Modoki is a Japanese word meaning "a similar but different thing" (Ashok et al. 2007)] peeks in the Niño 3.4 and Niño 4 regions (Kao and Yu 2009). Generally, an event with a larger maximum anomaly will also have a larger spatial structure; however, the area of warming is not stationary. Some indices may be better suited for phase onsets but miss the maximum anomaly. For this reason, this paper will examine the typical structure associated with the ENSO event to best analyze the performance of the models.