Tornado casualties and climate: past, present, and future

Paper presented at the Climate Science Lab meeting
College Station, TX

Transcript

1. Tornado casualties and climate: past, present, and future Tyler Fricker Department of Geography, Texas A&M University February 28, 2020
2. Who survives? Credit: Dayton Daily News Tornado casualties
3. Goal and objectives In this talk, I will provide an overview of our current understanding of tornado casualties and the potential connection between tornado casualties, climate, and society. More specifically: Describe the rate of tornado casualties as a function of tornado strength and population Investigate future rates of tornado casualties in conjunction with projected changes in tornado behavior and community structure Tornado casualties
4. What is a tornado? Tornado casualties
5. An understanding of past and present tornado casualties Tornado casualties
6. Tornado casualties A tornado casualty is any injury or death directly attributable to the tornado event itself: A direct fatality is any death attributed to the tornado A direct injury is any injury that requires treatment by a first-responder or subsequent treatment at a medical facility Tornado casualties
7. Tornado casualty trends 0 2000 4000 6000 1955 1965 1975 1985 1995 2005 2015 Year Number of Tornado Casualties A 40 80 120 160 1955 1965 1975 1985 1995 2005 2015 Year Number of Casualty−Producing Tornadoes B 0 5000 10000 15000 20000 25000 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Number of Tornado Casualties C 1 10 100 1000 1 10 100 Number of Casualties Number of Tornadoes D Tornado casualties
8. Tornado fatality locations Tornado casualties
9. Tornado injury locations Tornado casualties
10. Predicting tornado casualties Tornado casualties
11. Prediction approaches There are two main approaches to make future predictions of tornado casualties: 1. Statistical models 2. Empirical models Tornado casualties
12. Statistical models Statistical models can estimate how sensitive casualties are to changes in population and how sensitive casualties are to changes in tornado strength. These statistical models are evaluated through: An additive model An interactive model Tornado casualties
13. Tornado strength Surface energy dissipation (E) is given by E = Apρ J j=0 wj v3 j , where Ap is the area of the approximate path (width times length), ρ is the air density (assumed to be 1 kg m−3 at the surface), vj is the midpoint wind speed for each damage rating j, and wj is the corresponding fraction of path area. Tornado casualties
14. Additive model The additive model is given by C ∼ NegBin(µ, n) ln(µ) = ln(β0) + βP ln(P) + βE ln(E) where NegBin(µ, n) indicates that the conditional casualty counts are described by negative binomial distributions with mean (rate) µ and size n. The coefficient βP is the population term, and the coefficient βE is the energy term. Tornado casualties
15. Additive model results A doubling of the population under the path of a tornado leads to a 21% increase in the casualty rate A doubling of the energy dissipated by the tornado leads to a 33% increase in the casualty rate Tornado casualties
16. Interactive model The interactive model is given by C ∼ NegBin(ˆ µ, n) ln(µ) = ln(β0) + βP ln(P) + βE ln(E) + βP·E [ln(P) · ln(E)] where [ln(P) · ln(E)] is the interactive term, the coefficient βP is the population term, and the coefficient βE is the energy term. Tornado casualties
17. Interactive model results The percentage increase in casualties with increasing energy dissipation increases with population density The percentage increase in casualties with increasing population density increases with energy dissipation Tornado casualties
18. Which model is better? 1 2 5 10 20 50 .01 .1 10 1,000 .01 .1 10 1,000 Population Density [people per sq. km] Energy Dissipation [GW] A Tornado casualties
19. Which model is better? 1 2 5 10 20 50 .01 .1 10 1,000 .01 .1 10 1,000 Population Density [people per sq. km] Energy Dissipation [GW] B Tornado casualties
20. Empirical models Empirical models can estimate casualty rates at a per-capita or per-housing unit level using the number of people or the number of housing units within the tornado damage path. Tornado casualties
21. The dasymetric method Population < 10 10 − 20 21 − 30 31 − 40 > 40 Tornado casualties
22. Empirical casualty rates Per-capita casualty rate = number of casualties number of people Per-housing unit casualty rate = number of casualties number of housing units Tornado casualties
23. Empirical model results Over the period 1995–2016: The 22-year per-capita tornado casualty rate is 2.3% with a per-capita tornado fatality rate of .15% and a per-capita tornado injury rate of 2.1% The 22-year per-housing unit tornado casualty rate is 5.4% with a per-housing unit tornado fatality rate of .36% and a per-housing unit tornado injury rate of 5.1% Tornado casualties
24. Future tornado casualties Tornado casualties
25. Casualty factor trends 0 20000 40000 1980 1990 2000 2010 Year Exposure [Number of People] A 0 50 100 150 1980 1990 2000 2010 Year Tornado Power [TW] B Tornado casualties
26. Changes in population Atlanta Indianapolis Cincinnati Oklahoma City Tulsa Memphis Nashville Austin Corpus Christi Dallas Houston San Antonio A Number of People [millions] .00 to .25 .25 to .50 .50 to 1.0 1.0 to 1.5 1.5 to 2.0 > 2.0 Atlanta Indianapolis Cincinnati Cleveland Toledo Oklahoma City Tulsa Memphis Nashville Dallas B Number of People [millions] .00 to .25 .25 to .50 .50 to 1.0 1.0 to 1.5 1.5 to 2.0 > 2.0 Tornado casualties
27. Changes in environmental variables Tornado casualties
28. Statistical model prediction (population only) 500 510 520 530 540 2020 2040 2060 2080 2100 Year Casualties [Number of People] Tornado casualties
29. Statistical model prediction (population and tornado power) 500 550 600 650 700 2020 2040 2060 2080 2100 Year Casualties [Number of People] Tornado casualties
30. Empirical model prediction (per-capita) 700 800 900 1000 1100 1200 2020 2040 2060 2080 2100 Year Casualties [Number of People] Tornado casualties
31. Predictive model results The statistical model predicts higher-than-baseline casualty rates in the future under both increasing population and increasing population/tornado power scenarios The empirical model predicts higher-than-average casualty rates in the future as population increases Tornado casualties
32. Predictive model results For the year 2100: The statistical model predicts a percent change of 6% above present values under an increasing population scenario The statistical model predicts a percent change of 30% above present values under an increasing population/tornado power scenario The empirical model predicts a percent change of 30% above present average values as population increases Tornado casualties