ArticlePublisher preview available
To read the full-text of this research, you can request a copy directly from the authors.

Abstract and Figures

El Niño events differ substantially in their spatial pattern and intensity. Canonical Eastern Pacific El Niño events have sea surface temperature anomalies that are strongest in the far eastern equatorial Pacific, whereas peak ocean warming occurs further west during Central Pacific El Niño events. The event types differ in their impacts on the location and intensity of temperature and precipitation anomalies globally. Evidence is emerging that Central Pacific El Niño events have become more common, a trend that is projected by some studies to continue with ongoing climate change. Here we identify spatial and temporal patterns in observed sea surface temperatures that distinguish the evolution of Eastern and Central Pacific El Niño events in the tropical Pacific. We show that these patterns are recorded by a network of 27 seasonally resolved coral records, which we then use to reconstruct Central and Eastern Pacific El Niño activity for the past four centuries. We find a simultaneous increase in Central Pacific events and a decrease in Eastern Pacific events since the late twentieth century that leads to a ratio of Central to Eastern Pacific events that is unusual in a multicentury context. Compared to the past four centuries, the most recent 30 year period includes fewer, but more intense, Eastern Pacific El Niño events.
This content is subject to copyright. Terms and conditions apply.
1School of Earth Sciences, University of Melbourne, Parkville, Victoria, Australia. 2ARC Centre of Excellence for Climate System Science, University of
Melbourne, Parkville, Victoria, Australia. 3Climate and Energy College, University of Melbourne, Parkville, Victoria, Australia. 4ARC Centre of Excellence
for Climate Extremes, University of Melbourne, Parkville, Victoria, Australia. 5School of Earth, Atmosphere and Environment, Monash University, Clayton,
Victoria, Australia. 6NESP Earth Systems and Climate Change Hub, CSIRO, Aspendale, Victoria, Australia. 7School of Earth, Atmospheric and Life Sciences,
University of Wollongong, Wollongong, New South Wales, Australia. 8Research School of Earth Sciences, Australian National University, Canberra,
Australian Capital Territory, Australia. 9ARC Centre of Excellence for Climate Extremes, Australian National University, Canberra, Australian Capital
Territory, Australia. 10ARC Centre of Excellence for Climate Extremes, Monash University, Clayton, Victoria, Australia. *e-mail:
Canonical Eastern Pacific (EP) El Niño events exhibit their
largest sea surface temperature anomalies (SSTA) in the far
eastern tropical Pacific near the Peruvian coast1. Over recent
decades, peak warming during several El Niño events has been
displaced by approximately 11,000 km, or 100° longitude, west-
wards into the central equatorial Pacific. These El Niño events are
described as Central Pacific (CP) El Niño, warm-pool El Niño2, El
Niño Modoki3 or Dateline El Niño4. The displacement of maximum
SSTA towards the central Pacific drives substantial shifts in atmo-
spheric convection and circulation57, which alter the location and
intensity of temperature and precipitation impacts associated with
El Niño around the globe3,4,811.
Evidence is emerging that changes in the El Niño Southern
Oscillation (ENSO) behaviour occurred during the instrumental
period1215. After the climate regime shift in 1976/1977, zonal SSTA
propagation during El Niño changed from westward to eastward16.
Coincident with the shift to a positive phase of the Interdecadal
Pacific Oscillation1618 in 1999/2000, Pacific trade winds strength-
ened19,20. Observations indicate an increasing El Niño event
amplitude21, decadal variations in event frequency22, changes in
maximum SSTA propagation direction12,23 and delays in the onset
of El Niño events24.
Since the late 1990s there has been a higher number of CP events
relative to EP events, unprecedented in instrumental records2,21,22.
It is unclear whether this recent increase is part of natural climate
variability25 or a consequence of anthropogenic climate change22. A
precise picture of El Niño diversity is a challenge due to model defi-
ciencies in simulating El Niño and the short and sparse coverage of
instrumental observations across the equatorial Pacific25,26. In this
study, we extend the record of El Niño diversity into the past using
a network of coral data that spans the tropical Indo-Pacific ocean.
Spatial and temporal patterns of El Niño types
We used a network of 27 seasonally resolved coral records to recon-
struct past EP and CP El Niño events (Methods and Supplementary
Information). The network includes four Sr/Ca records (a proxy
for sea surface temperature (SST)) and 23 oxygen isotope (δ18O)
records. The δ18O signal preserved in coral banding is determined
by the source isotopic composition of the surrounding seawater
(δ18OSW) and the equilibrium isotopic fractionation between the
seawater and carbonate, which is inversely related to temperature27.
The oxygen isotopes are fractionated throughout the annual cycle
(and thus preserve the SST variations across the calendar year) and
the δ18OSW is affected by the advection of water masses with dif-
ferent isotope signatures and precipitation–evaporation changes.
Precipitation–evaporation changes also affect the sea surface salin-
ity (SSS), so δ18O is often used as an SST, SSS or SST–SSS proxy. The
relationship between salinity and δ18OSW is complex and can depend
on local conditions28. In general, if the SSS is relatively constant,
the δ18O in corals is mainly determined by SST variability and vice
versa. High variability of SSS and SST can make the interpretation
of δ18O in corals more complex. We carried out extensive testing of
our methods and explored possible sources of error, which included
testing the δ18O signal in the individual coral records and our net-
work as a whole (Supplementary Information). We found that all of
the coral δ18O records in our network have a strong link to ENSO,
with correlations to SST and SSS, parameters which in turn vary
with the spatial and temporal patterns of CP and EP El Niño events.
Higher frequency of Central Pacific El Niño events
in recent decades relative to past centuries
Mandy B. Freund 1,2,3*, Benjamin J. Henley 1,2,4,5, David J. Karoly1,2,6, Helen V. McGregor 7,
Nerilie J. Abram 8,9 and Dietmar Dommenget5,10
El Niño events differ substantially in their spatial pattern and intensity. Canonical Eastern Pacific El Niño events have sea sur-
face temperature anomalies that are strongest in the far eastern equatorial Pacific, whereas peak ocean warming occurs further
west during Central Pacific El Niño events. The event types differ in their impacts on the location and intensity of temperature
and precipitation anomalies globally. Evidence is emerging that Central Pacific El Niño events have become more common, a
trend that is projected by some studies to continue with ongoing climate change. Here we identify spatial and temporal patterns
in observed sea surface temperatures that distinguish the evolution of Eastern and Central Pacific El Niño events in the tropical
Pacific. We show that these patterns are recorded by a network of 27 seasonally resolved coral records, which we then use to
reconstruct Central and Eastern Pacific El Niño activity for the past four centuries. We find a simultaneous increase in Central
Pacific events and a decrease in Eastern Pacific events since the late twentieth century that leads to a ratio of Central to Eastern
Pacific events that is unusual in a multicentury context. Compared to the past four centuries, the most recent 30 year period
includes fewer, but more intense, Eastern Pacific El Niño events.
NATURE GEOSCIENCE | VOL 12 | JUNE 2019 | 450–455 |
The Nature trademark is a registered trademark of Springer Nature Limited.
... To investigate these matters, we considered a series of pertinent questions based on the impacts and processes observed in mangroves during mass dieback events and other abrupt declines in canopy condition across northern Australia [11,14,15] We also planned to evaluate these observations in the context of global climate change, especially where the frequency of extreme environmental drivers, like Taimasa and severe ENSO, were expected to increase [37]. But, while we consider suggestions for better management to minimise harm to mangroves, our primary goal with this article has been to raise awareness of the implications of the newly recognised impacts of Taimasa and other manifestations of extreme low sea level oscillations (coincident with severe El Niño) on shoreline mangroves. ...
... In this way, the unusually low 12-month SOI anomaly levels in 1982 ( Fig 4C) were both consistent with the earlier Taimasa, and definitive in identifying the specific months of occurrence of that dieback event in the absence of local sea level data. In summary, the two unusually extreme low sea level (Taimasa) periods in the western Pacific region in 1982/83 [16][17][18] and 2015/16 [9,27,37], were consistent with each event causing widespread mass dieback of mangroves [11]. ...
... In recent decades, these notably abrupt extreme lows in sea level have on two occasions, at least, resulted in severe desiccation and catastrophic mass dieback of shoreline mangroves in northern Australia [11]. The underlying drivers responsible have been the unusually severe El Niño events in 1982 and 2015, the two most severe such events in recent centuries, depicted in the 400 year coral record [37]. Of further concern, such occurrences have been predicted to re-occur as greenhouse warming escalates [58]. ...
Full-text available
Recent ENSO-related, extreme low oscillations in mean sea level, referred to as 'Taimasa' in Samoa, have destabilised shoreline mangroves of tropical northern Australia, and possibly elsewhere. In 1982 and 2015, two catastrophic Taimasa each resulted in widespread mass dieback of~76 km 2 of shoreline mangroves along 2,000 km of Australia's Gulf of Car-pentaria. For the 2015 event, we determined that a temporary drop in sea level of~0.4 metres for up to six months duration caused upper zone shoreline mangroves across the region to die from severe moisture deficit and desiccation. The two dramatic collapse events revealed a previously unrecognised vulnerability of semi-arid tidal wetland habitats to more extreme ENSO influences on sea level. In addition, we also observed a relationship between annual sea level oscillations and mangrove forest productivity where seasonal oscillations in mean sea level were co-incident with regular annual mangrove leaf growth during months of higher sea levels (March-May), and leaf shedding during lower sea levels (September-November). The combination of these periodic fluctuations in sea level defined a mangrove 'Goldilocks' zone of seasonal productivity during median-scale oscillations, bracketed by critical threshold events when sea levels became unusually low, or high. On the two occasions reported here when sea levels were extremely low, upper zone mangrove vegetation died en masse in synchrony across northern Australia. Such extreme pulse impacts combined with localised stressors profoundly threaten the longer-term survival of mangrove ecosystems and their benefits, like minimisation of shoreline erosion with rising sea levels. These new insights into such critical influences of climate and sea level on mangrove forests offer further affirmation of the urgency for implementing well-considered miti-gation efforts for the protection of shoreline mangroves at risk, especially given predictions of future re-occurrences of extreme events affecting sea levels, combined with ongoing pressure of rapidly rising sea levels.
... During EP "canonical" El Niño events, new production is depressed in the eastern equatorial Pacific and elevated west of 180°W when compared with neutral conditions (Strutton & Chavez, 2000;Turk et al., 2001Turk et al., , 2011. Central Pacific (CP) "Modoki" El Niño events have become more frequent relative to EP events since the 1990s (Ashok et al., 2007;Freund et al., 2019;Marathe et al., 2015;T. Lee & McPhaden, 2010), and have been shown to reduce new production and CO 2 flux in the western Pacific, but are similar to neutral conditions in the EP (Gierach et al., 2012;Liao et al., 2020;Messié & Chavez, 2013;Racault et al., 2017;Turk et al., 2011;Valsala et al., 2014). ...
... The shift from cool to warm Pacific Decadal Variability around 2015 may have further influenced the CP like trends presented (Figure 3 right column and Figure 4, center column). Therefore, as CP and La Niña events are becoming more frequent Freund et al., 2019), they could be causing or amplifying the decadal trends in the carbon budget, as explained in detail below. ...
... CO 2 outgassing and new production are strongly influenced by ENSO conditions (Table 1; Figures 2, 4 and 5). ENSO is changing character over time, toward less frequent but more intense EP events and more common CP and La Niña events Freund et al., 2019;T. Lee & McPhaden, 2010;Yeh et al., 2009). ...
Full-text available
The equatorial Pacific is the largest oceanic source of carbon dioxide to the atmosphere. This outgassing varies depending on the El Niño‐Southern Oscillation (ENSO) and decadal climate variability. New production, the amount of phytoplankton net primary production driven by upwelled nitrate, plays a significant role in modulating air‐sea CO2 fluxes as the biological carbon pump removes carbon from the surface ocean. We aim to understand how the physical drivers of sea surface temperature and wind speed influence interannual and decadal variability of the equatorial Pacific carbon cycle. In the equatorial Pacific, there are three biogeochemical regimes: the upwelling cold tongue east of 140°W and south of the equator (3°N–15°S); the eastern Pacific warm pool north of the equator (3°–15°N); and the 28.5°C western Pacific warm pool, west of 140°W. We find that between 2000 and 2020, air‐sea CO2 flux and ΔpCO2 increased in the cold tongue (45 mmolC m⁻² yr⁻², 1.5 μatm yr⁻¹, respectively) but decreased elsewhere, while new production decreased everywhere. The western Pacific occasionally became a weak carbon sink, depending on ENSO and this sink was strongest at 165°E during central Pacific “Modoki” El Niño events. We find that changes in wind speed, temperature and ENSO frequency have altered the surface carbon budget. The mean basin‐wide (150°E−90°W and 15°N–15°S) new production for 2000–2020 was 1.2 ± 0.1 PgC yr⁻¹ and air‐sea CO2 flux was 0.5 ± 0.1 PgC yr⁻¹. New production decreased at −7.7 ± 1.6 TgC yr⁻², compared to the CO2 flux trend of −1.7 ± 1.4 TgC yr⁻².
... Relevant for society is the total damage resulting from consecutive or cascading hazards during several months. Modelling studies and measurements still do not provide a clear image about the effect of climate change on the El Niño Southern Oscillation, but the general expectation is that more extreme states will occur more frequently (Cai et al., 2015(Cai et al., , 2018Freund et al., 2019). For this reason, a stronger shift from the "emergency response"-mindset towards integrated disaster risk management would be advantageous for Latin America. ...
Der technologische Fortschritt erlaubt es, zunehmend komplexe Vorhersagemodelle auf Basis immer größerer Datensätze zu produzieren. Für das Risikomanagement von Naturgefahren sind eine Vielzahl von Modellen als Entscheidungsgrundlage notwendig, z.B. in der Auswertung von Beobachtungsdaten, für die Vorhersage von Gefahrenszenarien, oder zur statistischen Abschätzung der zu erwartenden Schäden. Es stellt sich also die Frage, inwiefern moderne Modellierungsansätze wie das maschinelle Lernen oder Data-Mining in diesem Themenbereich sinnvoll eingesetzt werden können. Zusätzlich ist im Hinblick auf die Datenverfügbarkeit und -zugänglichkeit ein Trend zur Öffnung (open data) zu beobachten. Thema dieser Arbeit ist daher, die Möglichkeiten und Grenzen des maschinellen Lernens und frei verfügbarer Geodaten auf dem Gebiet der Hochwasserrisikomodellierung im weiteren Sinne zu untersuchen. Da dieses übergeordnete Thema sehr breit ist, werden einzelne relevante Aspekte herausgearbeitet und detailliert betrachtet. Eine prominente Datenquelle im Bereich Hochwasser ist die satellitenbasierte Kartierung von Überflutungsflächen, die z.B. über den Copernicus Service der Europäischen Union frei zur Verfügung gestellt werden. Große Hoffnungen werden in der wissenschaftlichen Literatur in diese Produkte gesetzt, sowohl für die akute Unterstützung der Einsatzkräfte im Katastrophenfall, als auch in der Modellierung mittels hydrodynamischer Modelle oder zur Schadensabschätzung. Daher wurde ein Fokus in dieser Arbeit auf die Untersuchung dieser Flutmasken gelegt. Aus der Beobachtung, dass die Qualität dieser Produkte in bewaldeten und urbanen Gebieten unzureichend ist, wurde ein Verfahren zur nachträglichenVerbesserung mittels maschinellem Lernen entwickelt. Das Verfahren basiert auf einem Klassifikationsalgorithmus der nur Trainingsdaten von einer vorherzusagenden Klasse benötigt, im konkreten Fall also Daten von Überflutungsflächen, nicht jedoch von der negativen Klasse (trockene Gebiete). Die Anwendung für Hurricane Harvey in Houston zeigt großes Potenzial der Methode, abhängig von der Qualität der ursprünglichen Flutmaske. Anschließend wird anhand einer prozessbasierten Modellkette untersucht, welchen Einfluss implementierte physikalische Prozessdetails auf das vorhergesagte statistische Risiko haben. Es wird anschaulich gezeigt, was eine Risikostudie basierend auf etablierten Modellen leisten kann. Solche Modellketten sind allerdings bereits für Flusshochwasser sehr komplex, und für zusammengesetzte oder kaskadierende Ereignisse mit Starkregen, Sturzfluten, und weiteren Prozessen, kaum vorhanden. Im vierten Kapitel dieser Arbeit wird daher getestet, ob maschinelles Lernen auf Basis von vollständigen Schadensdaten einen direkteren Weg zur Schadensmodellierung ermöglicht, der die explizite Konzeption einer solchen Modellkette umgeht. Dazu wird ein staatlich erhobener Datensatz der geschädigten Gebäude während des schweren El Niño Ereignisses 2017 in Peru verwendet. In diesem Kontext werden auch die Möglichkeiten des Data-Mining zur Extraktion von Prozessverständnis ausgelotet. Es kann gezeigt werden, dass diverse frei verfügbare Geodaten nützliche Informationen für die Gefahren- und Schadensmodellierung von komplexen Flutereignissen liefern, z.B. satellitenbasierte Regenmessungen, topographische und hydrographische Information, kartierte Siedlungsflächen, sowie Indikatoren aus Spektraldaten. Zudem zeigen sich Erkenntnisse zu den Schädigungsprozessen, die im Wesentlichen mit den vorherigen Erwartungen in Einklang stehen. Die maximale Regenintensität wirkt beispielsweise in Städten und steilen Schluchten stärker schädigend, während die Niederschlagssumme in tiefliegenden Flussgebieten und bewaldeten Regionen als aussagekräftiger befunden wurde. Ländliche Gebiete in Peru weisen in der präsentierten Studie eine höhere Vulnerabilität als die Stadtgebiete auf. Jedoch werden auch die grundsätzlichen Grenzen der Methodik und die Abhängigkeit von spezifischen Datensätzen and Algorithmen offenkundig. In der übergreifenden Diskussion werden schließlich die verschiedenen Methoden – prozessbasierte Modellierung, prädiktives maschinelles Lernen, und Data-Mining – mit Blick auf die Gesamtfragestellungen evaluiert. Im Bereich der Gefahrenbeobachtung scheint eine Fokussierung auf neue Algorithmen sinnvoll. Im Bereich der Gefahrenmodellierung, insbesondere für Flusshochwasser, wird eher die Verbesserung von physikalischen Modellen, oder die Integration von prozessbasierten und statistischen Verfahren angeraten. In der Schadensmodellierung fehlen nach wie vor die großen repräsentativen Datensätze, die für eine breite Anwendung von maschinellem Lernen Voraussetzung ist. Daher ist die Verbesserung der Datengrundlage im Bereich der Schäden derzeit als wichtiger einzustufen als die Auswahl der Algorithmen.
... Statistically significant relationships exist between the Laguna Pallcacocha flood history and records from both the central and western Pacific (Dang et al., 2020) and eastern Pacific (Koutavas and Joanides, 2012) (Fig. 6), indicates there is some degree of coherence between the many flavors of the ENSO over multi-millennial timescales. A newly developed network of coral records (Freund et al., 2019) shows that the ratio of CP to EP events has increased in recent decades, while EP events have become more intense. The development of a continuous Holocene chronology of EP El Niño events will be crucial to testing hypotheses regarding the relationship between different flavors of the ENSO under different background climate conditions. ...
Full-text available
The laminated sedimentary sequence of Ecuador's Laguna Pallcacocha is one of the most widely cited proxy records of Holocene El Niño Southern Oscillation (ENSO) variability. Previous efforts to reconstruct flood-driven laminae from Laguna Pallcacocha relied solely on sediment color, a useful but non-specific metric of flood events. We improved the chronology with ²¹⁰Pb and additional ¹⁴C dates over the past millennium, which allows for comparison of the sedimentary record with historically documented El Niño events. Additionally, we use elemental composition derived from X-ray Fluorescence (XRF) to reconstruct flood history at Pallcacocha. A principal component analysis (PCA) of the XRF dataset identifies minerogenic flood-driven clastic laminae. The first principal component (PC1) of the XRF data and red color intensity are positively correlated over the past 7.5 kyr, but the color record fails to capture high frequency variability that is preserved in the XRF dataset during the early Holocene (approximately 7.5-11 kyr BP). The new XRF dataset indicates moderate El Niño activity during the early Holocene, suppressed El Niño activity in the middle Holocene, and enhanced El Niño activity during the late Holocene. This pattern is relatively common among other ENSO records, and has been attributed to long-term changes in tropical insolation. Some intervals-most notably between 3-2 kyr BP and during the last millennium-deviate from expected trends if insolation was the sole forcing mechanism. Previously proposed mechanisms linking ENSO to latitudinal displacement of the ITCZ and ocean-atmospheric variabilities in other ocean basins appear to play an additional role in modulating Holocene ENSO development, as demonstrated by statistically significant correlations between the revised Laguna Pallcacocha flood history and proxy records from the Atlantic.
... On the other hand, it has been shown that the EP and CP events were alternatively prevalent every 10 to 20 years over the past century 61,62 . For example, the EP events were dominant in the 1980s while the CP El Niños were more frequently identified after 2000 63,64 . These findings imply that the decadal variability plays a crucial role in driving the transitions between the CP-and EP-dominant regimes. ...
Full-text available
El Ni\~no-Southern Oscillation (ENSO) is the most prominent interannual climate variability in the tropics and exhibits diverse features in spatiotemporal patterns. In this paper, a simple multiscale intermediate coupled stochastic model is developed to capture the ENSO diversity and complexity. The model starts with a deterministic and linear coupled interannual atmosphere, ocean and sea surface temperature (SST) system. It can generate two distinct dominant linear solutions that represent the eastern Pacific (EP) and the central Pacific (CP) El Ni\~nos, respectively. In addition to adopting a stochastic model for characterizing the intraseasonal wind bursts, another simple stochastic process is developed to describe the decadal variation of the background Walker circulation. The latter links the two dominant modes in a simple nonlinear fashion and advances the modulation of the strength and occurrence frequency of the EP and the CP events. Finally, a cubic nonlinear damping is adopted to parameterize the relationship between subsurface temperature and thermocline depth. The model succeeds in reproducing the spatiotemporal dynamical evolution of different types of the ENSO events. It also accurately recovers the strongly non-Gaussian probability density function, the seasonal phase locking, the power spectrum and the temporal autocorrelation function of the SST anomalies in all the three Ni\~no regions (3, 3.4 and 4) across the equatorial Pacific. Furthermore, both the composites of the SST anomalies for various ENSO events and the strength-location bivariate distribution of equatorial Pacific SST maxima for the El Ni\~no events from the model simulation highly resemble those from the observations.
... The canonical El Niño is a pattern of SST anomalies observed in central and western tropical Pacific areas, while El Niño Modoki is characterized by positive SST anomalies in the central equatorial Pacific and negative anomalies in eastern and western zones (Capotondi et al. 2015). EP-El Niño events have become less frequent in recent decades (Lee and McPhaden, 2010), contrasting with CP-El Niño events, which show a decadal periodicity (Ashok et al. 2007;Di Lorenzo et al. 2010;Couto et al. 2013;Freund et al. 2019). Phytoplankton functional groups are particularly useful for identifying the effects of El Niño events on oceanic ecosystems. ...
A long-term information baseline is necessary to identify the seasonal and interannual variability of diatom production driven by environmental dynamics. This motivated our compiling hydrographic data and concurrent diatom fluxes at the Alfonso Basin sub-zone. The diatom data time series spanned from February 2008 to September 2012. These data were analyzed along with primary productivity and chlorophyll-a concentration estimates derived from satellite imagery from 2002–2012. Planktonic diatom flux (range: 1.62 × 10⁶–3.55 × 10⁷ valves·m⁻²d⁻¹) was significantly correlated (r² = 0.37) with chlorophyll-a concentration. The bimodal production season (December–July) exhibited a winter mixing period influenced by water from the Gulf of California. The diatom flux was dominated by fast-growing surface species such as Chaetoceros spp. and Thalassiosira spp. The second peak in May occurred under the influence of southwest winds concurrent with the entry of tropical surface water and a cyclonic eddy that injected nutrients into the base of the euphotic zone, eliciting a bloom of Pseudo-nitzschia spp. The summer stratification resulted in the deepening of Subtropical Subsurface Water, and the diatom flux included large species of oceanic affinity such as Chaetoceros (Phaeoceros), Rhizosolenia acuminata, and Rh. imbricata. The environment at Alfonso Basin promotes diatom sinking; thus, material from sediment traps accurately reflects the primary production cycle associated with hydrographic processes. Interannual variability was characterized by anomalously low diatom fluxes between August 2009 and July 2010. A significant negative association between diatom flux anomalies and El Niño Modoki Index values suggests that oceanographic changes are related to this event. Additional reductions in diatom fluxes were coeval with SST positive anomalies recorded from July to February 2008–2009 and 2011–2012, which showed a significant relationship with the tropical Pacific Meridional Mode. The temporal trends like the additive signals of interannual (CP-El Niño) and decadal (Pacific Meridional Mode) time scales indicate that climatic disturbances affected the local ecosystem in the basin, reducing the vertical flux of diatoms initially dominated by Pseudo-nitzschia spp., and boosting the presence of oceanic warm-water species and small highly silicified species such as Thalassionema nitzschioides var. parva.
... We distinguish regional WRs from global oscillations that occur at larger spatial scales [e.g., El Niño-Southern Oscillation (ENSO), Pacific Decadal Oscillation (PDO), Interdecadal Pacific Oscillation (IPO), Atlantic Multidecadal Oscillation (AMO), Madden-Julian Oscillation (MJO)]. A significant body of literature has focused on reconstructing the latter using environmental proxies (Hernández et al. 2020), including tree-ring chronologies (D'Arrigo et al. 2001;MacDonald and Case 2005;Cook et al. 2008;D'Arrigo et al. 2015;Buckley et al. 2019), ice cores (Chylek et al. 2012;Vance et al. 2015), speleothems (Chen et al. 2016), corals (Urban et al. 2000Crueger et al. 2009;Cobb et al. 2013;Freund et al. 2019) and sediments (Dean and Kemp 2004;White et al. 2018). Reconstructions vary in length across the Common Era and Holocene depending on the chosen proxy, and they can often conflict due to nonstationarity in the teleconnections (Schmutz et al. 2000;An and Jin 2004;Raible et al. 2014;Newman et al. 2016). ...
Full-text available
Extreme weather variability has long posed difficulties for food, energy, and water systems in the Western US and is only projected to become more variable in a changing climate. The key to understanding weather in the West lies in understanding variability of large-scale weather regimes that dictate regional weather patterns. In this study, we propose a novel, multi-objective optimization and regression-based framework that reconstructs the annual frequency of regional weather regimes based on a gridded, tree-ring based reconstruction of cold season precipitation. The approach optimally smooths and selects the tree-ring based information used in the weather regime reconstruction to enhance out of sample performance and minimize model complexity. Multiple objectives are considered within the optimization to balance the preservation of low and high frequency modes of variability in the multivariate weather regime dynamics. We reconstruct weather regime frequencies back to 1400 CE and show that the reconstructed weather regimes are consistent with previously identified megadroughts and pluvials. Further, the reconstructed weather regimes exhibit significant variability in the 3–15-year frequency band and extend far beyond the bounds of the instrumental period. Overall, the weather regime reconstructions developed here provide important insight into the extent of natural atmospheric variability that can influence Western US weather.
... It oscillates irregularly every 2-8 years and significantly disrupts global rainfall and SST anomalies patterns (McPhaden et al., 2006;Ropelewski and Halpert, 1987). The teleconnection regulations of ENSO in the global ocean climate system show evident spatiotemporal heterogeneity in terms of intensity, spatial scope, onset, duration, and cessation (Freund et al., 2019). The SCS is located on the northern edge of the Western Pacific Warm Pool (WPWP), and its interannual climate change is primarily regulated by ENSO and associated climate systems (Xie et al., 2016;Zhou and Chan, 2010;Wang et al., 2000). ...
Full-text available
Massive Porites corals are widely used for paleoclimate reconstruction, however the potential of Porites coral microatoll δ¹⁸O as high-resolution archive of paleoclimate has only been recognized recently, and the systematic chemistry-climate relationship has yet to be well characterized in the western Pacific. In this study, we examined the reproducibility of microatoll δ¹⁸O records against two adjacent Porites corals from the Xisha Islands in the northern South China Sea (SCS) and evaluated the reliability of microatoll δ¹⁸O as a proxy for reconstructing regional climate and El Niño-Southern Oscillation (ENSO) variability. The seasonal to interannual variability in microatoll δ¹⁸O was primarily controlled by local sea surface temperature (SST), while the δ¹⁸O signal might be suppressed by sea surface salinity (SSS) variations on interannual timescale. Despite the overlapping coral δ¹⁸O records exhibited similar patterns of variability, the mean values were consistently offset by ~0.2‰ and the sensitivity of δ¹⁸O proxy to climate also varied across different coral colonies. The microatoll δ¹⁸O exhibited relatively high proxy-SST sensitivities and amplitude of the seasonal variabilities. These results suggested that intercolony δ¹⁸O variability was a significant source of uncertainty in coral-based paleoclimate reconstructions. Microatoll δ¹⁸O anomaly appeared to serve as a sensitive and relatively reliable proxy for ENSO variability, although the imprints of weak-to-moderate ENSO events could not be fully captured due to the complex relationship between the East Asian Monsoon and ENSO, as well as the local seawater salinity changes. This study further strengthened the evidence for microatoll as an alternate climate archive in the SCS and highlighted its potential in helping resolve poorly understood paleoclimate before instrumental observations.
... However, no observational studies have examined the relationship between Arabian Sea aerosols and Indian monsoon rainfall and its sensitivity to ENSO. Also, recent studies have indicated that El-Nino-like conditions will increase in the future due to anthropogenic climate change 32,33 . Therefore, it is essential to understand how ENSO may modulate the relationship between AS dust aerosol and monsoon rainfall. ...
Full-text available
The intensity of Indian summer monsoon rainfall (ISMR) over Central India (CI) is known to be positively correlated with the dust aerosol loading over the Arabian Sea (AS) on short time scales of about a week. However, global oscillations such as the El-Nino Southern Oscillation (ENSO) modulate both the rainfall over India and aerosol loading over the AS. This study uses long-term satellite-based aerosol and gridded rainfall datasets to explore the correlation between AS aerosol and CI rainfall and their relationship to ENSO. It is found that the highest correlation is during El-Nino (0.53), followed by Normal (0.44) and La-Nina (0.34) years, closely following the overall dust aerosol loading over the AS. Spatially, irrespective of the phase of ENSO, the high aerosol loading conditions are associated with increased winds over the AS, shifting eastward towards the Indian mainland and enhancing rainfall over CI and elsewhere across the Indian landmass. In contrast, the low aerosol loading conditions over the AS are associated with reduced winds, shifting westward away from the Indian mainland, suppressing rainfall over CI. In response to anthropogenic climate change, the El-Nino-like conditions are likely to increase in the future, making the dust aerosol-induced monsoon rainfall enhancement/modulation significant.
Full-text available
Multiple publications over the past decades argue that the Gulf of California is a region with high biological diversity and productivity. However, ecosystem resilience to climate disturbances and anthropogenic stressors in the Gulf of California remains poorly explored. One method to assess ecosystem resilience based on ecological indicators is the analysis of continuous records of critical environmental variables. Here we analyze satellite time-series of sea surface chlorophyll-a concentration (Chl-a) over the past two decades (1997−2020) and hydrographic data obtained from the central Gulf of California (2005−2019) to detect abrupt transitions (tipping points) and shifts in the temporal trends and their association with the most prominent modes of climate variability in the northeastern Pacific. In addition, based on the critical “slow-down theory”, we estimated the autocorrelation time (AcT) and standard deviation (SD) of satellite sea surface Chl-a, as a resilience indicator (RI), to monitor whether early-warning signals anticipate any impacts of climate change. We observed a clear negative trend in SST in the pre-2012 period, related to decadal and multiannual modes of variability of the Pacific decadal Oscillation (PDO), El Niño Modoki (EMI), and Pacific Meridional Mode (PMMSST). In contrast, a positive SST trend in the period post-2012 to 2017 was associated with the multiannual warming event in the northeastern Pacific that peaked during the intense 2015−2016 El Niño. These trends differentially regulate the Chl-a response during the cold (November to April) and warm (June to October) seasons, in line with the shift of regime in 2012. The critical transition early-warning signal depicted better consistency in the use of increasing SD in Chl-a time series, but still, AcT provides an effective predictor of a slowdown in most cases. GAM results showed that the main mode of climate variability that affects Chl-a was PMMSST. EMI, NPGO, and PDO modes had a less significant influence on Chl-a than PMMSST. The monitoring of high-frequency satellite records in the Gulf of California central region provided insight into temporal trends and their association with modes of climate variability. It represents an indicator of the effectiveness of the application of RIs for resilience monitoring that can be used to inform resource management decisions.
Full-text available
El Nino events, characterized by anomalous warming in the eastern equatorial Pacific Ocean, have global climatic teleconnections and are the most dominant feature of cyclic climate variability on subdecadal timescales. Understanding changes in the frequency or characteristics of El Nino events in a changing climate is therefore of broad scientific and socioeconomic interest. Recent studies(1-5) show that the canonical El Nino has become less frequent and that a different kind of El Nino has become more common during the late twentieth century, in which warm sea surface temperatures (SSTs) in the central Pacific are flanked on the east and west by cooler SSTs. This type of El Nino, termed the central Pacific El Nino (CP-El Nino; also termed the dateline El Nino(2), El Nino Modoki(3) or warm pool El Nino(5)), differs from the canonical eastern Pacific El Nino (EP-El Nino) in both the location of maximum SST anomalies and tropical-midlatitude teleconnections. Here we show changes in the ratio of CP-El Nino to EP-El Nino under projected global warming scenarios from the Coupled Model Intercomparison Project phase 3 multi-model data set(6). Using calculations based on historical El Nino indices, we find that projections of anthropogenic climate change are associated with an increased frequency of the CP-El Nino compared to the EP-El Nino. When restricted to the six climate models with the best representation of the twentieth-century ratio of CP-El Nino to EP-El Nino, the occurrence ratio of CP-El Nino/EP-El Nino is projected to increase as much as five times under global warming. The change is related to a flattening of the thermocline in the equatorial Pacific.
Full-text available
The far-reaching impacts of central Pacific El Niño events on global climate differ appreciably from those associated with eastern Pacific El Niño events. Central Pacific El Niño events may become more frequent in coming decades as atmospheric greenhouse gas concentrations rise, but the instrumental record of central Pacific sea-surface temperatures is too short to detect potential trends. Here we present an annually resolved reconstruction of NIÑO4 sea-surface temperature, located in the central equatorial Pacific, based on oxygen isotopic time series from Taiwan tree cellulose that span from 1190 AD to 2007 AD. Our reconstruction indicates that relatively warm Niño4 sea-surface temperature values over the late twentieth century are accompanied by higher levels of interannual variability than observed in other intervals of the 818-year-long reconstruction. Our results imply that anthropogenic greenhouse forcing may be driving an increase in central Pacific El Niño-Southern Oscillation variability and/or its hydrological impacts, consistent with recent modelling studies.
Full-text available
Focusing on ENSO seasonal phase locking, diversity in peak location, and propagation direction, as well as the El Niño–La Niña asymmetry in amplitude, duration, and transition, a set of empirical probabilistic diagnostics (EPD) is introduced to investigate how the ENSO behaviors reflected in SST may change in a warming climate. EPD is first applied to estimate the natural variation of ENSO behaviors. In the observations El Niños and La Niñas mainly propagate westward and peak in boreal winter. El Niños occur more at the eastern Pacific whereas La Niñas prefer the central Pacific. In a preindustrial control simulation of the GFDL CM2.1 model, the El Niño–La Niña asymmetry is substantial. La Niña characteristics generally agree with observations but El Niño’s do not, typically propagating eastward and showing no obvious seasonal phase locking. So an alternative approach is using a stochastically forced simulation of a nonlinear data-driven model, which exhibits reasonably realistic ENSO behaviors and natural variation ranges. EPD is then applied to assess the potential changes of ENSO behaviors in the twenty-first century using CMIP5 models. Other than the increasing SST climatology, projected changes in many aspects of ENSO reflected in SST anomalies are heavily model dependent and generally within the range of natural variation. Shifts favoring eastward-propagating El Niño and La Niña are the most robust. Given various model biases for the twentieth century and lack of sufficient model agreements for the twenty-first-century projection, whether the projected changes for ENSO behaviors would actually take place remains largely uncertain.
Full-text available
Inferring climate from paleodata frequently assumes a direct, linear relationship between the two, which is seldom met in practice. Here we simulate an idealized proxy characterized by a nonlinear, thresholded relationship with surface temperature, and demonstrate the pitfalls of ignoring nonlinearities in the proxy–climate relationship. We explore three approaches to using this idealized proxy to infer past climate: (i) methods commonly used in the paleoclimate literature, without consideration of nonlinearities, (ii) the same methods, after empirically transforming the data to normality to account for nonlinearities, (iii) using a Bayesian model to invert the mechanistic relationship between the climate and the proxy. We find that neglecting nonlinearity often exaggerates changes in climate variability between different time intervals, and leads to reconstructions with poorly quantified uncertainties. In contrast, explicit recognition of the nonlinear relationship, using either a mechanistic model or an empirical transform, yields significantly better estimates of past climate variations, with more accurate uncertainty quantification. We apply these insights to two paleoclimate settings. Accounting for nonlinearities in the classical sedimentary record from Laguna Pallcacocha leads to quantitative departures from the results of the original study, and markedly affects the detection of variance changes over time. A comparison with the Lake Challa record, also a nonlinear proxy for El Niño–Southern Oscillation, illustrates how inter-proxy comparisons may be altered when accounting for nonlinearity. The results hold implications for how nonlinear recorders of normally distributed climate variables are interpreted, compared to other proxy records, and incorporated into multiproxy reconstructions.
Full-text available
Despite the tremendous progress in the theory, observation and prediction of El Niño over the past three decades, the classification of El Niño diversity and the genesis of such diversity are still debated. This uncertainty renders El Niño prediction a continuously challenging task, as manifested by the absence of the large warm event in 2014 that was expected by many. We propose a unified perspective on El Niño diversity as well as its causes, and support our view with a fuzzy clustering analysis and model experiments. Specifically, the interannual variability of sea surface temperatures in the tropical Pacific Ocean can generally be classified into three warm patterns and one cold pattern, which together constitute a canonical cycle of El Niño/La Niña and its different flavours. Although the genesis of the canonical cycle can be readily explained by classic theories, we suggest that the asymmetry, irregularity and extremes of El Niño result from westerly wind bursts, a type of state-dependent atmospheric perturbation in the equatorial Pacific. Westerly wind bursts strongly affect El Niño but not La Niña because of their unidirectional nature. We conclude that properly accounting for the interplay between the canonical cycle and westerly wind bursts may improve El Niño prediction.
Full-text available
A new index is developed for the Interdecadal Pacific Oscillation, termed the IPO Tripole Index (TPI). The IPO is associated with a distinct ‘tripole’ pattern of sea surface temperature anomalies (SSTA), with three large centres of action and variations on decadal timescales, evident in the second principal component (PC) of low-pass filtered global SST. The new index is based on the difference between the SSTA averaged over the central equatorial Pacific and the average of the SSTA in the Northwest and Southwest Pacific. The TPI is an easily calculated, non-PC-based index for tracking decadal SST variability associated with the IPO. The TPI time series bears a close resemblance to previously published PC-based indices and has the advantages of being simpler to compute and more consistent with indices used to track the El Niño–Southern Oscillation (ENSO), such as Niño 3.4. The TPI also provides a simple metric in physical units of °C for evaluating decadal and interdecadal variability of SST fields in a straightforward manner, and can be used to evaluate the skill of dynamical decadal prediction systems. Composites of SST and mean sea level pressure anomalies reveal that the IPO has maintained a broadly stable structure across the seven most recent positive and negative epochs that occurred during 1870–2013. The TPI is shown to be a robust and stable representation of the IPO phenomenon in instrumental records, with relatively more variance in decadal than shorter timescales compared to Niño 3.4, due to the explicit inclusion of off-equatorial SST variability associated with the IPO.
Decision trees are a simple but powerful prediction method.
One of the core problems of modern statistics is to approximate difficult-to-compute probability distributions. This problem is especially important in Bayesian statistics, which frames all inference about unknown quantities as a calculation about the posterior. In this paper, we review variational inference (VI), a method from machine learning that approximates probability distributions through optimization. VI has been used in myriad applications and tends to be faster than classical methods, such as Markov chain Monte Carlo sampling. The idea behind VI is to first posit a family of distributions and then to find the member of that family which is close to the target. Closeness is measured by Kullback-Leibler divergence. We review the ideas behind mean-field variational inference, discuss the special case of VI applied to exponential family models, present a full example with a Bayesian mixture of Gaussians, and derive a variant that uses stochastic optimization to scale up to massive data. We discuss modern research in VI and highlight important open problems. VI is powerful, but it is not yet well understood. Our hope in writing this paper is to catalyze statistical research on this widely-used class of algorithms.
A much-anticipated 'monster' El Nino failed to materialize in 2014, whereas an unforeseen strong El Nino is developing in 2015. El Nino continues to surprise us, despite decades of research into its causes. Natural variations most probably account for recent events, but climate change may also have played a role.
El Nino-Southern Oscillation (ENSO) is the leading mode of interannual global climate variability, which in its essence is often described by the equatorial dynamics of the recharge oscillator with a fixed pattern. Here we explore the idea that ENSO can be simulated in a model with a fixed pattern of sea surface temperature variability following the recharge oscillator mechanism, which interacts with the thermodynamic red noise of a slab ocean. This model is capable of simulating the leading modes of sea surface temperature variability in the tropical Pacific in good agreement with the observations and most coupled general circulation models. ENSO dynamics, amplitude, seasonality, the structure of the leading patterns, its meridional extension, its variations in an eastern and central Pacific pattern and associated positive feedbacks are all influenced and simulated well by including the interaction of recharge oscillator and the thermodynamic coupling to the slab ocean model. We further point out that much of the ENSO diversity in the spatial structure is a reflection of this interaction. However, it also has to be noted that some equatorial dynamics are missing in this model and in coupled general circulation models that are important for the ENSO diversity.