Integration of Sea-Level Rise and Climate Change into Hurricane Flood Level Statistics

Size: px

Start display at page:

Download "Integration of Sea-Level Rise and Climate Change into Hurricane Flood Level Statistics"

Dylan Moore
5 years ago
Views:

edu Jen Irish and Don Resio November 1, 011 3 rd Coastal Hazards Symposium Sponsored by U.

1 Integration of Sea-Level Rise and Climate Change into Hurricane Flood Level Statistics Galveston Hurricane Ike Galveston 1900 Hurricane UPI Jen Irish and Don Resio November 1, rd Coastal Hazards Symposium Sponsored by U.S. Department of Energy U.S. Army Corps of Engineers and NOAA Sea Grant Limit to ~15 slides (0-minutes total) 1

2 Motivation A method for determining hurricane flood level extreme-value statistics which considers long-term environmental trends: Sea level rise Hurricane frequency Hurricane tracks Hurricane intensity Bender et al. 010 IPCC 007 Knutson & Tuleya 004

3 Conclusions: Climate, SLR, and Flood Level Statistics Conclusions & Methods Summary JPM-OS viable approach for future extreme-value statistics Mean flood statistics not too sensitive to climate scenario selected Error in projected future mean climate trends leads to error in mean flood statistics Differences in mean flood statistics on the order of model + sampling uncertainty Uncertainty significantly reduced by using adaptive management approach Methods summary: Use joint probability method with optimal sampling Consider an alongshore, uniform coast Consider three possible future climate scenarios: B1, A1B, A1FI Consider hurricane rate-of-occurrence and intensity, sea-level rise (SLR) 3

Outline Motivation Conclusions & methods summary Statistical framework Climate change scenarios Rate of occurrence Intensity Sea-level rise Application on an idealized coast Quantification of

4 Outline Motivation Conclusions & methods summary Statistical framework Climate change scenarios Rate of occurrence Intensity Sea-level rise Application on an idealized coast Quantification of uncertainty Results Mean flood statistics Errors due to projection means Adaptive management approach Conclusions (revisited) c p = 950 mb (historical) c p = 933 mb c p = 94 mb From Mousavi et al. 011 Climatic Change 4

5 Statistical Framework JPM with Optimal Sampling (JPM-OS) z max x (x, p o,r p,v f,, x o,msl) z 1 z tide where: surge simulation waves winds is a continuous flood response function, x is location of interest, x o is landfall location, R p is hurricane pressure radius near landfall (Thompson and Cardone 1996), θ is hurricane track angle with respect to the shoreline, v f is hurricane forward speed near landfall, MSL is mean sea level, and ε z is uncertainty in the flood response (Resio et al. 009). From Irish et al., 009 Natural Hazards and Irish et al. 011 GRL 5

6 Statistical Framework JPM with Optimal Sampling (JPM-OS) T R z max 1 f ( p o,r p,v f,, x o ) H z max (x, p o,r p,v f,,x o,msl) z p o R p v f x o dx d dv dr o f p dp o 1 Extreme-value distribution From Resio et al., 009 Natural Hazards 6

7 Statistical Framework JPM with Optimal Sampling (JPM-OS) T R z max 1 f ( p o,r p,v f,, x o ) H z max (x, p o,r p,v f,,x o,msl) z p o R p v f x o dx d dv dr o f p dp o 1 From Resio et al., 009 Natural Hazards 7

8 Statistical Framework JPM with Optimal Sampling (JPM-OS) T R z max 1 f ( p o, R p,v f,, x o ) H z max (x, p o, R p,v f,, x o,msl) z SST p o R p v f x o dx d dv dr o f p dp o dsst 1 f (SST, p o,r p,v f,, x o ) SST f ( p o x o ) a 1 (x o,t) exp p a (x,t) o o exp exp p a (x,t) o o (Gumbel Distribution) a 1 (x o,t) a 1 (x o,t) 5 g (t), x o SST? From Irish and Resio, in prep. ASCE Waterw. 8

9 Climate Scenarios SST and SLR IPCC 007 climate projection scenarios: B1 global carbon emissions minimized A1B global carbon emissions moderated A1FI global carbon emissions are high MAGICC/SCENGEN (Wigley 004): Sea-surface temperature (SST): Sea-level rise (SLR, eustatic only) Two time periods: 040s and 080s B1: This scenario assumes future use of clean-energy and energyefficient technologies such that future global carbon emissions are minimized. A1B: This scenario assumes a future balanced portfolio of energy sources such that future global carbon emissions are moderated. A1FI: This scenario assumes continued dominance of fossil energy sources such that future global carbon emissions are high. 9

10 Climate Scenarios Hurricane Rate of Occurrence Bender et al. 010 SST o SST SST 33%/1.7oC = 19%/1oC 10

11 Climate Scenarios Hurricane Intensity p SST p o 0.08 p SST SST p Knutson & Tuleya

12 Application Idealized Coast Alongshore-uniform: Slope = 1:1000 & 1: km (~ US GOM length) Surge response functions from idealized ADCIRC simulations Assume max MSL max ( x) max ( x) mx ( ) p p ( x xo ) x R p ( x) is dimensionless surge, max x is dimensionless alongshore distance from location of interest to peak surge, is specific weight of water, m(x) is a location-dependent constant, and δ is the average distance between landfall location (x o ) and peak alongshore surge normalized by R p. From Irish et al. 011 JGR 1

13 Application Idealized Coast Simplified JPM: f (SST, p o,r p,x o ) SST 1 5 Historical GOM meteorology: λ o = 0.36 Assume uniformly distributed: 5 (t) N x Assume uniformly weighted: SST constant From Irish et al. 011 JGR 13

14 Quantification of Uncertainty Sampling Bootstrapping z Climate Aleatory = 0.3 m, T R = 50 yr 0.55 m, T R = 500 yr SST SLR p... Epistemic = 0.70 m tide surge simulation waves winds... Aleatory + Epistemic = 0.74 m, T R = 50 yr 0.89 m, T R = 500 yr From Irish and Resio, in prep. ASCE Waterw. Resampling - accounts for not knowing population exactly Bootstrap if population known exactly 14

15 Results Mean Statistics and Uncertainty 040s 080s Parameter Mean Standard Deviation 040 / / 080 Rate of occurrence: 0.8 to 0.9 / 0.14 to to 0.03 / 0.04 to 0.07 Intensity (% change in deficit): 8.7 to 9.9 / 14.3 to to 3.5 / 4.7 to 8.6 SLR (m): 0.15 to 0.17 / 0.31 to to 0.04 / 0.07 to 0.09 From Irish and Resio, in prep. ASCE Waterw. Max delta =0.80 m for 080s B1 vs A1FI at Tr=50 yrs 15

16 Results Mean Statistics and Uncertainty Frequency Intensity SLR Parameter Mean Standard Deviation 040 / / 080 Rate of occurrence: 0.8 to 0.9 / 0.14 to to 0.03 / 0.04 to 0.07 Intensity (% change in deficit): 8.7 to 9.9 / 14.3 to to 3.5 / 4.7 to 8.6 SLR (m): 0.15 to 0.17 / 0.31 to to 0.04 / 0.07 to 0.09 From Irish and Resio, in prep. ASCE Waterw. 080s results in plots, only 16

17 Results Sensitivity to Error in Mean Climate Projections -19% per o C +3% per o C From Irish and Resio, in prep. ASCE Waterw only 17

18 Results Adaptive Management 040s 080s Parameter Standard Deviation St. Dev. with Adaption 040 / / 080 Rate of occurrence: 0.0 to 0.03 / 0.04 to to 0.0 / 0.01 to 0.0 Intensity (% change in deficit):.7 to 3.5 / 4.7 to to.0 / 1.0 to.5 SLR (m): 0.03 to 0.04 / 0.07 to / 0.0 to 0.03 From Irish and Resio, in prep. ASCE Waterw. St dev ranges for all three scenarios 18

19 Conclusions: Climate, SLR, and Flood Level Statistics Conclusions & Future Work JPM-OS viable approach for future extreme-value statistics Mean flood statistics not too sensitive to climate scenario selected Error in projected future mean climate trends leads to error in mean flood statistics Differences in mean flood statistics on the order of model + sampling uncertainty Uncertainty significantly reduced by using adaptive management approach Future work: Quantify climate uncertainty What about barrier islands, wetlands, human adaptation? 19

20 Questions? UPI ASCE Associated Press U.S. Army Corps of Engineers ASCE U.S. Army Corps of Engineers National Geographic 0

Future Tropical Cyclone Flood Hazard: Impacts of Vegetation Change and Sea Level Rise

Future Tropical Cyclone Flood Hazard: Impacts of Vegetation Change and Sea Level Rise : Impacts of Vegetation Change and Sea Level Rise Irish)et)al.)2014) 0.75 Es#mated)flooding)during)Hurricane)Katrina)2005)) Es#mated)flood9eleva#on)change)due)to)0.75)m) of)sea9level)rise)and)vegeta#on)loss)since)c.)1900)