**2. The ENSO Phenomenon**

The ENSO phenomenon refers to the variations in atmospheric and ocean conditions, or in the climate conditions, arising from variations in sea surface temperatures and atmospheric pressure across the tropical Pacific Ocean. ENSO is comprised of two major components that reflect its complex coupled nature, the El Niño or ocean, and the Southern Oscillation or atmospheric component. Human society has chronicled the impacts of El Niño for far longer than its atmospheric counterpart. Peruvian fishermen in the 1500s understood well the impact on fisheries of unusually warm waters that occasionally occurred off the coast of Peru around Christmas time. Because El Niño events are associated with anomalously high sea surface temperatures, they are also referred to as "warm events". The "cool event" counterpart carries the name of La Niña.

H. Hidebrandsson, working in the late 1890s, is often credited with unearthing the rudiments of what we now know as the Southern Oscillation or the atmospheric pressure "seesaw" (barometric seesaw) between the eastern and western Pacific [10]. Building on this work and that of Norman and Lockyer in 1902, and extensive research of his own and that of his collaborator E W Bliss, Gilbert Walker in 1928 named and presented the first coherent ideas about the Southern Oscillation (SO) and extensively described the implications of the SO for inter-annual climate variability across the tropics, including how a Southern Oscillation Index (SOI) could be applied to climate forecasting a season

ahead [10]. While Troup [11] reconfirmed and refined many of the earlier long distance associations (now referred to as teleconnections) discovered by Walker, Bjerknes [12] conceptualised the association between El Niño and the Southern Oscillation as an outcome of air-sea or ocean-atmosphere interaction that led to the coining of the term ENSO.

El Niño and La Niña are part of the ENSO cycle that lasts from 12–18 months, over periods of 2–7 years, associated with alterations of the SO; although some El Niño and La Niña events can last beyond 24 months. The ENSO cycle refers to the alteration of climate fields associated with the development, peak, and decay of sea surface temperature anomalies in the eastern and central Pacific along with alterations to the atmospheric circulation and weather patterns across vast areas. El Niño (La Niña) or warm (cool) event conditions first begin to manifest as positive (negative) sea surface temperature (SST) anomalies in the central and eastern Pacific around July. These continue to develop as the ENSO cycle progresses, reaching a peak in the Northern Hemisphere around January to February of the following year, trailed by a decay or lessening of SST anomalies in the subsequent months of March to July/August, and cessation of the El Niño (La Niña) event by the end of summer. The swing between El Niño and La Niña phases is not immediate and successive. Rather, El Niño and La Niña events can be punctuated by "neutral" conditions when SST conditions in the eastern and central Pacific are in and around "normal".

ENSO events can be viewed as a self-sustained and naturally oscillatory mode of the coupled ocean-atmosphere system, or a stable mode triggered by stochastic forcing [13], and positive ocean-atmosphere feedback processes, with negative feedbacks required to terminate events [12]. No two ENSO events are alike. From exhaustive analyses of ocean and atmosphere climate fields, two broad types of ENSO have emerged: Eastern Pacific (EP) and Central Pacific (CP). The two types were identified in relation to where maximum SST anomalies tend to occur, with the CP type also referred to by a variety of other names particularly the widely used "El Niño Modoki" [13]. While in the context of ENSO and health the different types or flavours of ENSO events might appear inconsequential, the nuanced differences in their climate impacts may hold implications for the temporal and spatial dynamics of ENSO-related health responses.

A further characteristic of ENSO that holds possible implications for climate and health associations is the multi-decadal changes observed for ENSO's amplitude, period, propagation characteristics, asymmetry, onset, and predictability [13,14]. For example, the variance of the 2–7 year periodicity of ENSO was relatively high during the periods 1875–1920 and 1960–1990, but relatively low from 1920 to 1960 [14]. A clear shift in the amplitude of SST anomalies in the EP occurred in and around the mid-1970s. Such a shift and the variation in the variance of the 2–7 year periodicity appears to be related to the background climate state of the Pacific Ocean, or the phase of the Pacific Decadal Oscillation (PDO). As the Pacific Ocean transitions from a cool (warm) phase with lower (higher) than normal SSTs to a warm (cool) phase with positive (negative) SST anomalies over a period of 3–4 decades, ENSO characteristics and their link to climate impacts are affected [9,15–19]. The implication is that ENSO-related health impacts may be non-stationary at the multi-decadal scale. That there are non-symmetric relationships between ENSO and the Indian Ocean Dipole (IOD), another form of ocean climate variability [20], also raises the question as to whether the strength and direction of ENSO health links in the broad region of IOD influence might be IOD phase dependent.
