Understanding and Predicting El Niño and the Southern Oscillation Mike McPhaden NOAA/PMEL Seattle, Washington USA ECCO Summer School Friday Harbor Lab 21 May 2019 Topics § ENSO & Climate Impacts § Observing ENSO § 2015-16 El Niño Ø Description Ø Dynamics Ø Predictability & Prediction § ENSO in a changing climate El Niño and the Southern Oscillation (ENSO) El Niño—Warm phase of ENSO La Niña—Cold phase of ENSO Every few years, the trade winds weaken… H L Warm Water Slowly LeaksLatitudes to Higher ENSO: A Result of Coupled Ocean- Atmosphere Interactions Strong Trade Winds Weak Trade Winds Darwin Tahiti NINO-3.4 El Niño=NINO3.4 ≥ 0.5ºC for 5 months La Niña=NINO3.4 ≤ -0.5ºC for 5 months Atmospheric Teleconnections During El Niño PNA Pattern DJF Heavy Rain = Atmospheric Heating PSA Pattern JJA After Wallace & Gutlzer, 1981 Shifts in tropical rainfall patterns affect global atmospheric circulation via long range teleconnections Karoly, 1989 El Niño Alters Global Atmospheric Circulation DJF El Niño & La Niña Impacts on Global Patterns of Weather Variability El Niño El Niño and La Niña shift the probability for droughts, floods, heat waves, extreme weather events around the globe La Niña Impacts on Tropical Storms § El Niño tends to suppress the formation of Atlantic hurricanes § El Niño tends to enhance intensity and geographic range of Pacific hurricanes § Opposite tendencies occur during La Niña Observing ENSO Satellites Global Ocean Observing System for Climate In Situ Components 87% 100% 62% 81% 100% Elephant Seals 43% 24% 48% 79% Why we measure (the 5 D’s): Definition, Detection, Diagnostics, moDel Development, preDiction ATLAS Mooring: ü Ocean and atmosphere ü Rapid continuous sampling ü Real-time data TAO/TRITON Array built up over 1985-94 during international Tropical Ocean Global Atmosphere (TOGA) program 2015-16 El Niño Dec 2015 December 2015 SST & Precipitation Anomalies Precipitation Anomalies Dry Wet Atmospheric December 2015 Teleconnections Monsoon 2015 ends with 14% shortfall Flooding 'worst in 50 years', as Indonesian Forest Fires 150,000 flee in Paraguay, Out of Control Argentina, Brazil and Uruguay Week ending 4 Nov 2015 An ENSO Index Nino-3.4 El Niño=NINO3.4 ≥ 0.5°C for 5 months La Niña=NINO3.4 ≤ -0.5°C for 5 months Average of 15 El Niños from 1957-58 to 2015-16 December 2015 vs December 2016 MajorEl Niño El Niño WeakLa Niña 29ºC La Niña TRITON ATLAS December 2015 vs December 2016 20° C 20ºC Major El Niño 20ºC http://WWW.pmel.noaa.gov/gtmba/ Weak La Niña December 2015 Anomalies El Niño La Niña TRITON ATLAS ENSO Diversity 1997-98 Eastern Pacific (EP) El Niño 2009-10 Central Pacific (EP) El Niño 2015-16 Hybrid 2015-16 El Niño: a hybrid EP/CP event of extreme amplitude ENSO Dynamics Governed by Multi- Time Scale Processes § Deterministic seasonal time scale dynamics ü Coupled feedbacks between ocean and atmosphere ü Wind forced changes in ocean circulation that redistribute heat in the upper ocean § High frequency (days to weeks) weather noise forcing ü Westerly winds bursts most prominent ü Introduces irregularity (timing, duration, amplitude) Upper Ocean Heat Content as a Precondition (Recharge Oscillator Theory, Jin 1997) § Build-up of excess heat content along equator is a necessary ExceeD ThresholD à El Niño precondition for El Niño to occur. § El Niño purges excess heat to higher latitudes, which terminates event. § The time between El Niños is determined by the time to recharge. Heat content (~average temperature anomaly in upper 300 m) From TAO/TRITON, Argo, XBT NINO-3.4 Upper Ocean Heat Content (“ENSO Fuel”) § Build-up of excess heat content along equator is a necessary precondition for El 0.25C Niño to occur. § El Niño purges excess heat to higher latitudes, which terminates event. § The time between El Niños is determined by the time to recharge. Heat content (~average temperature anomaly in upper 300 m) NINO-3.4 From TAO/TRITON, Argo, XBT Westerly Wind Bursts (“The Spark”) 2014, 2015, 2018 Fuel+Spark = El Niño Evolution: Jan 2015-Dec 2016 ”StateWesterly-dependent Wind Cooling by Warming by Warming by Air-Sea Zonal Downwelling Burstsnoise forcing”(WWB) Heat Flux Advection Kelvin Waves El Niño and La Niña are Predictable § First successful El Niño prediction in 1986 § ~20 organizations around the World noW regularly issue ENSO forecasts § Forecast models are reasonably accurate up to 6-9 months in advance § Predictability based variations in upper ocean heat content § Predictability limited by Weather noise, model bias, errors in initial conditions ENSO QUICK LOOK August 20, 2015 A monthly summary of the status of El Nino,˜ La Nina˜ and the Southern Oscillation, or “ENSO”, based on NINO3.4 index (120-170W, 5S-5N) During late July through early-August 2015 the SST was at a strong El Nino˜ level. All atmospheric variables support the El Nino˜ pattern, including weakened trade winds and excess rainfall in the east- central tropical Pacific. The consensus of ENSO prediction models indicate continuation of strong El Nino˜ conditions during the August-October 2015 season in progress. Some further strengthening into fall is likely, with the event lasting into spring 2016. Early-Aug CPC/IRI Consensus Forecast1 Mid-Aug IRI/CPC Plume-Based Forecast2 100 100 ENSO state based on NINO3.4 SST Anomaly ENSO state based on NINO3.4 SST Anomaly 90 o o 90 Neutral ENSO: −0.5 C to 0.5 C Neutral ENSO: −0.5oC to 0.5oC 80 80 El Nino Neutral 70 70 La Nina 60 60 50 50 Climatological Probability: 40 40 El Nino Probability (%) Neutral 30 30 La Nina 20 20 10 10 0 0 JAS ASO SON OND NDJ DJF JFM FMA MAM ASO SON OND NDJ DJF JFM FMA MAM AMJ 2015 Time Period 2016 2015 Time Period 2016 Niño-3.4 Predictions: July 2015 Initial Conditions Mid-Aug 2015 Plume of Model ENSO Predictions 3.0 Dynamical Model: El Niño NCEP CFSv2 2.5 IRI/CPC NASA GMAO “There is a greater Historical NINO3.4 SST Anomaly DYN AVG JMA STAT AVG SCRIPPS than 90% chance that 3 2.0 LDEO CPC CON AUS/POAMA El Niño will continue 2.5 1.5 ECMWF through Northern 2 UKMO C) C) o KMA SNU ← O Hemisphere winter 1.5 1.0 IOCAS ICM 1 COLA CCSM3 2015-16, and around 0.5 MetFRANCE an 85% chance it will 0.5 SINTEX-F CS-IRI-MM 0 last into early spring 0.0 GFDL CM2.1 −0.5 Neutral zone CMC CANSIP 2016.” GFDL FLOR −1 -0.5 Statistical Model: NOAA/NCEP NINO3.4 SST Anomaly ( NINO3.4 SST Anomaly ( −1.5 CPC MRKOV 13 August 2015 -1.0 CDC LIM −2 CPC CA −2.5 -1.5 Verification CPC CCA 1982 1986 1990 1994 1998 2002 2006 2010 2014 CSU CLIPR Time Period UBC NNET -2.0 FSU REGR OBS FORECAST La Niña UCLA-TCD -2.5 UNB/CWC MJJ Jul JAS ASO SON OND NDJ DJF JFM FMA MAM AMJ 2015 2016 Compiled by the International Research Institute for Climate and Society (IRI) Historically Speaking El Nino˜ and La Nina˜ events tend to develop during the period Apr-Jun and they: - Tend to reach their maximum strength during Dec-Feb - Typically persist for 9-12 months, though occasionally persisting for up to 2 years - Typically recur every 2 to 7 years 1Based on a consensus of CPC and IRI forecasters, in association with the official CPC/IRI ENSO Diagnostic Discussion. 2Purely objective, based on regression, using equally weighted model predictions from the plume. DJF 2016 Precipitation Forecast vs Historical El Niño Pattern Forecast Precipitation Forecast vs Verification DJF 2016 ý þ þ þ þ þ ý þ Verification Forecast ENSO QUICK LOOK August 20, 2015 A monthly summary of the status of El Nino,˜ La Nina˜ and the Southern Oscillation, or “ENSO”, based on NINO3.4 index (120-170W, 5S-5N) During late July through early-August 2015 the SST was at a strong El Nino˜ level. All atmospheric variables support the El Nino˜ pattern, including weakened trade winds and excess rainfall in the east- central tropical Pacific. The consensus of ENSO prediction models indicate continuation of strong El Nino˜ conditions during the August-October 2015 season in progress. Some further strengthening into fall is likely, with the event lasting into spring 2016. Early-Aug CPC/IRI Consensus Forecast1 Mid-Aug IRI/CPC Plume-Based Forecast2 100 100 ENSO state based on NINO3.4 SST Anomaly ENSO state based on NINO3.4 SST Anomaly 90 o o 90 Neutral ENSO: −0.5 C to 0.5 C Neutral ENSO: −0.5oC to 0.5oC 80 80 El Nino Neutral 70 70 La Nina 60 60 50 50 Climatological Probability: 40 40 El Nino Probability (%) Neutral 30 30 La Nina 20 20 10 10 0 0 JAS ASO SON OND NDJ DJF JFM FMA MAM ASO SON OND NDJ DJF JFM FMA MAM AMJ 2015 Time Period 2016 Comparison2015 Time Period of 20142016 and 2015 Nino3.4 Forecasts Mid-Aug 2015 Plume of Model ENSO Predictions 3.0 Dynamical Model: NCEP CFSv2 El Niño 2.5 IRI/CPC El Niño NASA GMAO Historical NINO3.4 SST Anomaly DYN AVG JMA STAT AVG SCRIPPS 3 2.0 LDEO CPC CON AUS/POAMA 2.5 1.5 ECMWF 2 UKMO C) C) o KMA SNU ← O 1.5 1.0 IOCAS ICM 1 COLA CCSM3 0.5 MetFRANCE 0.5 SINTEX-F CS-IRI-MM 0 0.0 GFDL CM2.1 −0.5 CMC CANSIP Neutral zone GFDL FLOR −1 -0.5 Statistical Model: NINO3.4 SST Anomaly ( NINO3.4 SST Anomaly ( −1.5 CPC MRKOV Verification -1.0 CDC LIM −2 Verification CPC CA −2.5 -1.5 CPC CCA 1982 1986 1990 1994 1998 2002 2006 2010 2014 CSU CLIPR La Niña Time Period UBC NNET (ºC)Anomaly SST Nino3.4 -2.0 La Niña FSU REGR OBS FORECAST UCLA-TCD -2.5 UNB/CWC MJJ Jul JAS ASO SON OND NDJ DJF JFM FMA MAM AMJ 2015 2016 2014 | 2015 Historically Speaking Good! Bad! El Nino˜ and La Nina˜ events tend to develop during the period Apr-Jun and they: CompileD by the International Research Institute for Climate anD Society (IRI) - Tend to reach their maximum strength during Dec-Feb - Typically persist for 9-12 months, though occasionally persisting for up to 2 years - Typically recur every 2 to 7 years 1Based on a consensus of CPC and IRI forecasters, in association with the official CPC/IRI ENSO Diagnostic Discussion.