RDG-ENSO-Final (1)
1 like311 views
ENSO events like El Niño and La Niña can impact global climate patterns and increase risks of natural catastrophes. While ENSO predictions have improved, uncertainties remain due to complex factors. Insurers can leverage knowledge of ENSO to support risk analysis and strategies like adjusting reinsurance terms for expected active or quiet hurricane seasons. Considering multi-year ENSO patterns may also help mitigate risks and reduce earnings volatility.
RDG-ENSO-Final (1)
- 1. WILLIS RE CATASTROPHE MANAGEMENT SERVICES Return of the Damned Child: Uncertainty in ENSO Prediction, Climatic Impacts and Insurance Implications By: Alexander Pui
- 2. Outline How predictable is ENSO? What are its climatic impacts? What are its economic and insurance implications? How to develop an ‘ENSO-readiness’ plan?
- 3. 3 HOW PREDICTABLE IS ENSO?
- 4. Section Outline What is ENSO? What are the key ‘Ingredients’ for an El Niño? How do we know an El Niño has occurred? What are the complexities of ENSO? How accurate are ENSO predictions? 4
- 5. What is ENSO? El Niño Southern Oscillation A global climate phenomenon that describes changes in both the ocean and atmosphere. Has 3 states: – El Niño – Neutral – La Niña 5
- 6. ENSO state (Neutral) Schematic of ENSO during neutral state (neither El Niño or La Niña) Source: BOM 2011 West (Warm) East (Cool) Walker Circulation SSTA in Nino 3.4 region Nino 3.4 SSTAs in Nino 3.4 region
- 7. Cool ENSO state (La Niña) Schematic of ENSO during La Niña state : Note warmer SST and increased convection in the West Pacific Ocean. Source: BOM 2011 SSTA in Nino 3.4 region West (Warmer) East (Cooler) Nino 3.4
- 8. Warm ENSO state (El Niño) Schematic of ENSO during El Niño state : Note warmer SST and increased convection in the West Pacific.Source: BOM 2011 SSTA in Nino 3.4 region East (Warmer) West (Cooler) Nino 3.4
- 9. Key Ingredients for an El Niño Westerly Wind Bursts Spread of Sea Surface Ocean Warmth Spread of Sub Surface Ocean WarmthSource: NOAA
- 10. Evolution of current El Niño Note spread of Pacific Ocean warmth towards the West Coast of American Continent Source: NOAA
- 11. Different ‘Flavors’ of El Niño 11 East Pacific/Classical El Niño Hybrid El Nino Source: NOAA Central Pacific / Modoki (もどき) El Niño 2. Find Covariance Matrix of Y 1. Let Y = SSTA Field in Space & Time 3. Solve Eigen Equation for C …AND HOW ARE THE PATTERNS & INDICES OF THESE ‘ FLAVORS’ DERIVED? Where, e1 =1st EOF / Classical Pattern e2 =2nd EOF / Modoki Pattern (~52%Variance Explained) . (~11%Variance Explained) .
- 12. Has an El Niñodeveloped? Is the monthly Nino 3.4 SSTA >= 0.50 C? NO Not El Niño Conditions YES Will SSTAs stay >= 0.50 C for next few seasons? Insufficiently weak Walker Circulation YES What does atmosphere look like? El Niño Conditions! Indication of weak Walker Circulation (i.e. atmospheric response!) NO Source: NOAA
- 13. 13 IPO (inter-decadal variations) NOISE Internal Variations (i.e. Modoki) Annual Cycle (i.e. Seasons) ENSO Adapted from: Axel Timmerman,2015 COMPLEXITIES OF ENSO
- 14. 14 IPO (inter-decadal variations) Annual Cycle (i.e.Seasons) NOISE Internal Variations (i.e. Modoki) ENSO COMPLEXITIES OF ENSO (ICE AGE)
- 15. 15 IPO (inter-decadal variations) Annual Cycle (i.e. Monsoon) NOISE Internal Variations (i.e. Modoki) ENSO COMPLEXITIES OF ENSO (GLOBAL WARMING)
- 16. 16 COMPLEXITIES OF ENSO (ORBITAL CYCLES)
- 17. How accurate are climate crystal balls? 2010-11 La Niña 2014-15 El Niño Model ‘Hit’? Model ‘Miss’? Source: IRI (Columbia University)
- 18. Who Killed the 2014 ‘Super’ Niño? 18 SSTA at Time of Prediction Outcome April 1997 April 2014 September 1997 September 2014 Adapted from: M.McPhaden,2015
- 19. Predictions not perfect, but still useful 19 Model View Odds of La Niña following strong El Niño? Source: NOAA
- 20. 20 REGIONAL CLIMATIC IMPACTS (ASIA PACIFIC)
- 21. Rainfall (El Niño) 21 Historical Rainfall Anomalies (Sep-Nov, 1901: 2013) Less Rain More Rain Dry Conditions in Indonesia and East Australia Dry Conditions in Northern India Source: KNMI Rainfall Anomalies (Jul – Sep 2015)
- 22. Temperature (El Niño) 22 Seasonal Temperature Anomalies (Dec-Feb) Warmer Cooler Warmer in Thailand , Malaysia and East Australia Warmer in India Source: KNMI
- 23. Forest Fires and Drought (El Niño) 23 2015 Indonesian Haze Event 24th Sep 2015 NASA terra satellite image of Peat Forest Burning Source: NASA Reduced Visibility Levels rivalling that of 1997 record haze year
- 24. Bushfires in Australia (El Niño) 24 Canberra 2003 Fires co-occuring with the 2002-2003 El Nino Modoki. Decreasing Trend in Soil Moisture to exacerbate Bushfires? Bushfire Risk heightened during El Nino, +ve IOD (or both co-occuring) Source: Cai et al., Positive Indian Ocean Dipole events precondition South East Australia Bushfires, GRL, 2009
- 25. Typhoons (All Years) 25 Largest ACE since 1976 The Pacific is primed for a very active Typhoon season. (give e.g. of elapsed Typhoons thus far) 6 tropical cyclones at once! . Source: Phil Klotzbach (Colorado State via RMS)
- 26. Cyclones (East Australia) 26 Source: KNMI Feature Frequency Genesis Location Track Bias Intensity Cyclone Lifespan Unknown? Lower Unknown? Coral Sea (SWPacific) Cyclone Characteristics during El Nino Response Lower North EastShift
- 27. Source: Wang, G., and H.H. Hendon, 2007: Sensitivity of Australian rainfall to Inter-El Nino variations. J. Climate, 20, 4211-4226. Impacts on Rainfall are different Max. SSTAs at different locations Differences in rainfall response according to position of maximum SSTAs. Uncertainty in ENSO impacts (Rainfall)
- 28. Source: Risbey et al. On the Remote Drivers of Rainfall Variability in Australia, Monthly Weather Review, 2009. Max. SSTAs at different locations Uncertainty in ENSO impacts (Rainfall) Strength of correlation between rainfall and SOI index (proxy for ENSO activity). Note differences in extent and strength of correlations at different periods of time.
- 29. Reinforcing Climate Drivers (Australia) 29 ‘3 Headed Dog’ : ENSO IOD SAM Joint Impact of the ‘3 Headed Dog’ during 2011 Brisbane Floods: ENSO (La Niña) IOD (-ve phase) SAM (-ve phase) Source: Risbey et al. On the Remote Drivers of Rainfall Variability in Australia, Monthly Weather Review, 2009.
- 30. Climate Change Factor Increased frequency of extreme ENSO events? Extreme = 101 Extreme = 212 Increased risk of coastal flooding from storms/ cyclones due to sea level rise? Source: Cai et al. 2014,Increasing frequency of extreme El Nino events due to greenhouse warming, Nature Pole-wards migration of cyclones/hurricanes?
- 31. 31 ECONOMIC IMPACTS & RISK MANAGEMENT STRATEGIES
- 32. Section Outline 32 Is there evidence of ENSO signature in property, BI and crop economic costs? How to incorporate ENSO into risk management and RI decisions? Other Applications
- 33. Economic costs of ENSO (Property) 33 Event El Niño Neutral La Niña Total Bushfire 2 1 1 4 Cyclone 3 1 6 10 Hail 4 2 4 10 Flood 0 0 4 4 Storm 3 2 2 7 Non-Bushfire 10 5 16 31 Total 12 6 17 35 Count of Seasonal Natural Catastrophe Events Exceeding $500M AUD in 2015 Values - ByENSO Phase Source: Prepared by Ned Scandrett, using ICA data (1967 – current) Higher Losses in La Niña and El Niño years compared to Neutral. Differences especially stark after considering ‘residence’ time spent in each ENSO phase.
- 34. Economic costs of ENSO (BI) 2011 Thai Floods (La Niña) 4th costliest disaster (approx. USD 45-50B economic loss) USD 15-20B of which is insured, and; 80% Supply Chain Related 34 Ayutthaya Aerial view of Honda vehicles at the flooded Honda factory in the Rojuna Industrial district, Nov 2011. Source: Bloomberg
- 35. Neutral years tend to experience best yield outcome (with the exception of Soybean). Economic costs of ENSO (Crop) 35 Source: Izumi et al., Impacts of ENSO on the global yield for crops, Nature 2014 El Nino La Nina Neutral ENSO is predictable - impacting climate and economic costs! How can we leverage this knowledge to hedge financial risk?
- 36. ENSO to support technical risk view 36 ‘Best’ view of risk • Exposure Analysis (e.g. RMS Cat Modelling) • Experience Analysis • Climate Variability Technical Technical/Commercial Risk Appetite • How much risk to take? • Shareholder expectations Capital Requirements • How much RI purchased? • Regulatory Requirements Commercial RI Quotes • Soft/Hard Market? • Relationship with Reinsurers ENSO INPUT
- 37. ENSO to support technical risk view 37 1 in 200 year event (w/o ENSO considerations) 1 in 100 year event (with ENSO considerations) +50%
- 38. Example of Short Term RI Strategy 38 Hypothetical Property Biz • Fixed Capital of $250m • XOL reinsurance • Constant P(ruin) = 0.1% • Ruin if losses > 30% net worth Idealized Hurricane scenarios • Quiet vs Active Seasons • 1000 realizations, each 100yrs • Wind Damage only Source: Emanuel et al. 2012: Potential Economic Value of Seasonal Hurricane Forecasts Weather, Climate and Society, Vol 4, 110 -117
- 39. ‘Baseline’ vs ‘Strategic’ scenarios 39 RI Program Options Baseline (Control) *Retains $30m every year Strategic Perfect *Forecast 100% correct *Retain $18m ($119m) if active(quiet) years Imperfect * Forecast 60% correct *Retain $28m ($50m) if active (quiet) years Source: Emanuel et al. 2012: Potential Economic Value of Seasonal Hurricane Forecasts Weather, Climate and Society, Vol 4, 110 -117
- 40. ‘Baseline’ vs ‘Strategic’ revenue 40 Distribution of accumulated revenue after 50 years for control, flawed and perfect forecasts. Baseline $39m Strategic (Perfect) $136m Strategic (Imperfect) $63m Source: Emanuel et al. 2012: Potential Economic Value of Seasonal Hurricane Forecasts Weather, Climate and Society, Vol 4, 110 -117
- 41. Reinsurance Treaty Inception Dates 41 Jan (Y2)Jan(Y1) Jan (Y3)Predictability Barrier Early Year Inception Mid Year Inception Peak Bushfire/Flood Season Window of (arbitrage) Opportunity? Time window to consider: Increase Limits? Lower Retentions? Additional Sideways Protection? Reinstatements?
- 42. Longer Term Strategies Multi-year Reinsurance Contracts Hedge against price hikes 42 State IPO + All IPO - P(Niño|Niño) 17% 10% 0% P(Niña|Niño) 42% 33% 22% P(Niño|Niña) 31% 29% 25% P(Niña|Niña) 15% 25% 38% How likely are back-to-back adverse ENSO events? Source: Pui et al. 2015, Representing low frequency variability in continuous rainfall simulations, Journal of Hydrology (submitted)
- 43. Other Applications Reduce volatility in earnings through Natural Perils ‘allowance’ Underwriting (Embargoes, pricing adjustments) Claims Handling (Event Response Capability) 43
- 44. Conclusion ENSO is predictable up to 9 months in advance* ENSO has quantifiable impacts on climate and ensuing economic costs. Insurers should incorporate ENSO in risk management decisions and when formulating business strategies. 44* In a probabilistic sense
- 45. Subtitle 1 Subtitle 2 Subtitle 3 WHAT IS YOUR 2015-16 ENSO STRATEGY GOING TO BE?