Private Property, Information Disclosure and
the Roles of Insurance and Government
Otis Brown | Ken Kunkel | Jenny Dissen | Paula Hennon
CICS-NC / NC State University
NOAA National Centers for Environmental Information
March 19th – 20th, 2015
Discussion Topics
► Background
• CICS-NC: Who We Are and Our Research Focus
• Our Role in the Climate / Insurance Context
► Current State of the Climate
•
•
•
•
Updates from the National Climate Assessment
Projections and Downscaling
Where are We on Attribution
Extreme Events and Emerging Science Areas
► Where Do We Go From Here?
• What are the information needs?
CICS-NC
CICS-NC’s Role
CICS-NC is a catalyst –
Climate Information
• Supports NCEI, NOAA activities and
academic research in climate
• Brings innovative, state-of-the-art approaches
to solving science, stewardship and service
challenges
• Facilitates development of climate adaptation
decision support awareness and tools…and
Partnerships
• Strong economic development tie:
•
CICS-NC reports to the NC State University
Office of Research, Innovation and
Economic Development
Business/Policy Decisions
Examples of NOAA/NCEI Products
Local
Regional
National
& Global
Weekly
Monthly
Seasonal – Annual
Snowfall Impact
Index
–
Heating &
Cooling Degree
Days
–
Temperature &
Precipitation
Outlooks
–
Energy Sector
Agriculture
Emergency
Planners
Heat Wave
Prediction
–
Drought
Outlook
–
Climate
Normals
–
Public Health
Officials
Agriculture
Construction,
Infrastructure,
Agriculture
Drought
Monitor
Monthly State
of the Climate
Reports
–
Annual State of
the Climate
Reports
–
National
Climate
Assessment
Decision
Makers
Decision
Makers
FEMA, disaster
response
Hurricane Tracks
–
–
Agriculture
Decadal
Billion $ Disasters,
Climate Extremes
Index
–
Insurance
–
Decision
Makers
Climate / Insurance Context:
Our Understanding…
Data
OUR GOALS
• Expand the understanding of
the complex issues surrounding
the research and data on
climate change, the related
uncertainties and their impacts
on catastrophe risk
management
• Understand better the changes
to the frequency and severity of
weather events
• Understand better the impacts
of climate risks to the insurance
and reinsurance industry
Decisions
• Climate risk as a stand alone risk --- is this
likely?
• Complexities of the disaggregated nature
of the risk
• Singular events and its consequences
• Predictability vs. Acts of God?
• Changes in the frequency and
amplitude of extreme events
• Areas of current research
• Products and pricing? What is the level of
specificity…and where does climate data
fit?
• Timing and timeframe:
• Policies are annual
And how does it matter to each of you?
Climate Data,
Models,
Analyses,
Assessments,
Information
Exchange
Academics
Commissioners
/ Regulators
Associations
Markets,
Regulations,
Policies,
Standards
CLIMATE RISKS &
OPPORTUNITIES
Reinsurance
Companies
Actuaries,
Catastrophe
Modeling,
Data
Consumers /
Insured
Insurance
Companies
Types and Pricing
of Insurance:
-
Travel
Title
Property MortgageLife
-
Auto
Blanket
Casualty
Credit
Crop
-
Financial
Health
IP
General
Liability
Where Does Climate Matter?
Publications,
Research,
Workshops,
Education
Catastrophe Risk Models
Communication
s
Modeling
Data & Science
Reinsurance looking at mega
risks on climate
scales
Reinsurance
Companies
Insurers want more
downscaled information
to assess risk of insuring
property (coastal, floods,
wildfires)
INSURANCE
COMPANIES
- Products
- Policies
- Claims..
Insurance commissioners
POLICY HOLDER
DECISION
-
Land usage
Policy
Property
Adaptation
…
Climate - State of the Knowledge
What Do We Know Today
Where are the Gaps?
1. NCA Findings
2. Projections and Downscaling
3. Where Are We on Attribution?
4. Extreme Events and Emerging
Science Areas
Observed U.S. Temperature Change
Observed U.S. Trends in Heavy Precipitation
Trends in Flood Magnitude
Projected Temperature Change
Global Sea Level Rise
The 2000 Assessment
The 2014 Assessment
Vulnerability to Sea Level Rise
Projected Change in Precipitation Intensity
Projected Changes in Water Withdrawals
Attribution – State of Knowledge
Extreme Events
Let’s discuss trends, projections and
confidence:
 Heavy precipitation
 Severe thunderstorms
 Hurricanes
 Winter storms
 Heat and cold waves (maybe)
 Drought (maybe)
Upward U.S. Trends in Extreme Precipitation
1 in 3.7 yrs
1 in 4 yrs
1 in 5 yrs
(Normal)
1 in 6 yrs
Rank of 2001-2011 “Decade”
Global Extreme Precipitation Trends
Projected Extreme Precipitation Episodes
Areas of Research:
Global Warming->Saturation Water Vapor Increases
Global Temperature
Water Vapor Content
+7%/°C
RCP 8.5
Increases
in
Rainfall
Rate
(Capacity)
Observed
Changes in Meteorological Systems (Opportunity)
Observed Precipitation vs. Temperature
blue: cool climates
orange: moderate climates
red: warm climates
Warmer Locations->more heavy events
Percent of Annual Precipitation
(Karl and Trenberth, 2003, Science)
Confidence in Projected Extreme Precipitation
• High
• Why? Because there is a direct link between
air temperature and the amount of water
vapor in the atmosphere over the oceans. As
temperature increases, the capacity of the
atmosphere to hold water vapor increases.
• Climate model simulations show increases in
the most extreme precipitation amounts as a
direct consequence of increased water vapor
US Annual Tornadoes
Severe Thunderstorms
• Data generally lacking in quality to assess
trends
• Small spatial scale environment critical to
these phenomena, particularly tornadoes
Projections of Severe Thunderstorms
• Small spatial scale of important
processes is a challenge for climate
models
• Competing effects:
– thermodynamic instability, which promotes
thunderstorm development; likely to
increase due to warmer surface
temperatures and higher water vapor content
– vertical wind shear, which is necessary for
thunderstorm rotation; likely to decrease
Confidence in Projected Severe Thunderstorms
• Low
• Why? Because the two key atmospheric
conditions necessary for severe
thunderstorms appear likely to change in
opposite directions in response to global
warming
• Recent research results point to increases
in frequency of favorable conditions
Hurricanes
0.7
North Atlantic
0.5
0.4
0.3
0.2
0.1
0
1970 1975 1980 1985 1990 1995 2000 2005 2010
Year
Best Track
Satellite-Based Reanalysis
2
Western North Pacific
1.8
1.4
3
-2
m s )
1.6
1.2
12
PDI (10
• Observational
changes over
time are a
challenge to
century-scale
trends analysis
PDI (1012 m3 s-2)
0.6
1
0.8
0.6
0.4
0.2
1970 1975 1980 1985 1990 1995 2000 2005 2010
Year
Projections of Hurricanes
Confidence in Projected Hurricane Intensity
• Medium
• Why? Because hurricane development requires
several conditions including (among others)
– high sea surface temperatures (relative?)
– Low vertical wind shear
– Low pressure disturbances
• There is high confidence sea surface
temperatures will increase in future but not so for
relative changes
• There is lower confidence about changes in
vertical wind shear and the number of low
pressure disturbances
35
250
30
200
25
20
150
15
100
10
Number of Grid Cells
40
300
(a)
50
5
0
0
1900
1920
1940
1960
1980
2000
300
(b)
35
250
30
200
25
20
150
15
100
10
50
5
0
1900
Number of Grid Cells
Percent Area Coverage > 90th Percentile
• Percent area
with seasonal
snowfall
above 90th
percentile
and below
10th
percentile
40
Percent Area Coverage < 10th Percentile
Winter Storms
0
1920
1940
1960
1980
2000
NCDC/NESDIS/NOAA
Projections of Winter Storms
• Regionally variable
– Higher moisture content would favor larger
snowstorms in areas where it is cold enough
to snow
– Decreases on southern margins of
climatologically snowy areas – becomes too
warm to snow
Confidence in Projected Winter Storm Frequency
and Intensity
• Medium
• Why? Because there is high confidence
about increases in temperature and
associated changes in where it can
snow and in the amount of water vapor
in the atmosphere but lower confidence
in whether the number and intensity of
low pressure storms will change over
the U.S.
Observed Extreme
Temperature
Episodes
#Records per station per year
7
Daily Records
6
5
4
3
2
1
0
1910s 1920s 1930s 1940s 1950s 1960s 1970s 1980s 1990s 2000s
– Daily records
– Monthly
temperature
records
Record High
0.3
#Records per station per year
• Record hot and
cold
temperatures
Decade
Record Low
Monthly Records
0.2
0.1
0.0
1910s 1920s 1930s 1940s 1950s 1960s 1970s 1980s 1990s 2000s
Decade
Projected Extreme Temperature Episodes
Confidence in Projected Extreme Temperatures
• High
• Why? Because there is high confidence
about increases in mean temperature
and to a first approximation climate
models simulate similar changes in
extreme temperatures
Drought
Trends (% per century)
•
•
•
1900 to 2011: -0.1%
1930 to 2011: -10.0%
1971 to 2011: +31.6%
• Widespread persistent drought
– 1930s (Central and Northern Great Plains, Northwest, Great Lakes)
– 1950s (Southern Plains, Southwest), 1980s (West, Southeast)
– First decade of the 21st century (West, Southeast)
Peterson, T. C. et al., 2013.
Drought
Drought
• Increased
temperature
leads to
increased
evaporation
rates and
tendency
toward
decreased
soil moisture
Confidence in Projected Extreme Drought
• Medium
• Why? Because there is high confidence
about increases in mean temperature
and this will lead to increases in
evaporation and decreases in soil
moisture. But there is low confidence in
changes in precipitation over much of U.S.
What Do We Know – Key Messages
• Observed changes in extremes
–
–
–
–
–
Increase in extreme rainfall events
Data insufficient to determine trends in severe thunderstorms
Increase in number of intense hurricanes
No trends in winter storm severity
Increase in unusual heat and decrease in unusual cold
• Future projections
– Continuation of recent trends in extreme temperature and
precipitation extremes
– Uncertainty about trends in certain types of extremes, most
notably severe thunderstorms and tornadoes
– Possible increase in intensity of most extreme hurricanes
– Regional changes in winter storms
Confidence in Projections
• Highest confidence is for those extreme
types that are directly correlated with
temperature
– Heat waves and cold waves (follows mean
temperature)
– Extreme Precipitation (higher water vapor content)
– Drought (higher evaporation)
– Hurricanes (higher sea surface temperatures)
• Lower confidence where the linkage to
temperature is indirect or complex
– Severe thunderstorms (competing effects)
– Winter storms (complex linkages)
Catastrophic Events-Challenges
• Examples: Sandy, Katrina, 1930s Dust
Bowl, 1993 Mississippi River flood, etc.
• Physical Understanding
– We can describe how they develop, but the
“why”, the larger causes, is another matter.
Their very rarity is the challenge
– This means we have great uncertainty about
projecting into the future
Catastrophic Events-Challenges
• Statistical Modeling
– Provides a mechanism for projecting the future, even
without full physical understanding
– Extreme Value Theory is the state-of-the-art framework for
statistical modeling of climate extremes
– However, this does not have the maturity of standard
statistical modeling based on the “central limit theorem”
(normal distribution)
– The small number of “events” at the extreme tail of the
distribution and the lack of a mature, powerful foundation
(like the central limit theorem) inserts considerable
uncertainty into projections based on such statistical
modeling (does EVT actually represent nature)
Catastrophic Events-Challenges
• Fundamental limitations for future
projections
– Physical understanding of changes in very rare
circumstances (e.g. the Sandy “left hook”).
Occurrences are too rare in climate models to
confidently assess future changes
• Statistical model limitations
– The small number of “events” at the extreme tail of
the distribution and the lack of a mature, powerful
foundation (like the central limit theorem) inserts
considerable uncertainty into projections based on
such statistical modeling (does EVT actually
represent nature)
Availability of Climate Projections
• Wide variation in level of data
– Summary graphics and data
(small file sizes)
– Raw climate model output
(huge file sizes)
• Type of data
– Direct output from global
climate models
– Downscaled data sets
Links:
National Climate Assessment 2014
NCA3 Data page
Downscaling
• Global climate model data is at a
coarse resolution, typical grid box size
being 50-100 miles
• This is too coarse for many impacts
studies
• “Downscaling” uses various scientific
methods to increase the spatial
resolution
Downscaling
• Statistical downscaling is the most widely used
because it is relatively inexpensive
• Data sets being produced at resolutions of 110 miles
• Dynamical downscaling has the potential to
produce more accurate projections, but is
much more expensive
Sources of detailed data
• Earth System Grid for global climate
model data
– http://cmippcmdi.llnl.gov/cmip5data_portal.htm
• Statistically-downscaled data
– http://gdodcp.ucllnl.org/downscaled_cmip_projections
/dcpInterface.html
– https://cds.nccs.nasa.gov/nex/
So - Where Does Climate Matter?
Publications,
Research,
Workshops,
Education
Catastrophe Risk Models
Communication
s
Modeling
Data & Science
Reinsurance looking at mega
risks on climate
scales; aiming
to diversify
Reinsurance
Companies
Insurers want more
downscaled information to
assess risk of insuring
property (coastal, floods,
wildfires, wind); also
diversifying risks
INSURANCE
COMPANIES
- Products
- Policies
- Claims...
Insurance commissioners
POLICY HOLDER
DECISION
-
Land usage
Policy
Property
Adaptation
…
Need for improved risk analysis & catastrophe modeling
 Hurricanes – tends to be the most popular model (i.e. “The gold standard”)
 Models focus on wind damage, but starting to consider storm surge and rain
induced flooding.
 We understand the complexity: difficult to develop stable statics with only a
few instances per year vs. millions of collisions per year for auto policies
 Severe convective storms – Focusing on straight line wind, hail, and tornadoes.
 Winter storm – Especially snowfall and freezing risk.
 Flooding – A few modellers are working on riverine flood, inland flooding, and
storm surge (sometimes as part of the Hurricane model).
 Wildfire –models tend to focus on wind and fuel sources.
 Earthquake / tsunami
 Space weather
 Any peril
Value of Uncertainty
“There’s uncertainty around everything, but in insurance
it has a cost associated with it.”
• We are in a dynamic climatic state; things are not going
back to baseline based on perturbation
– Impact of singular events: predictable or Acts of God?
– Frequency and amplitude of events…and how do we handle
small occurrences (e.g. the tails) --- active area of research
• What can the climate research community do to
facilitate the information content?
• How can we integrate the value of climate risk across
the board? Is disaggregation helpful…?
Where Do We Go from Here?
We see a need for…continued discussions
• Awareness
• Case studies or examples of adaptation successes
• Collaborative partnerships
• More than a SEC Guideline
• Decision support tools and capabilities
But…
• There is a need for a venue or a mechanism for science leaders and insurance
leaders to meet on equal turns
• Decision support approaches and/or tools which facilitate integration of
climate adaptation into the insurance industry – harder than it appears
• Need Collaborative Partners
Partnership for Resilience
Connect. Communicate. Collaborate.
Public, Private, Non-profit and Academic Partners
Working together to
enhance resilience by
developing
innovative
approaches based on
the latest science,
private-sector
expertise and best
practices.
Partner needs drive the
development, direction
and outcomes of working
groups
Contacts and Information
For Additional Information:
• Otis Brown
Director, CICS-NC
Otis_Brown@ncsu.edu
• Ken Kunkel
Lead Scientist NCA, Research Professor CICS-NC
Kekunkel@ncsu.edu
• Paula Hennon
Partnership for Resilience
Pahennon@ncsu.edu
• Jenny Dissen
Corporate Relations and Strategic Engagement
JennyDissen@cicsnc.org
Additional References:
Websites:
•
Cooperative Institute for Climate and Satellites – North Carolina: www.cicsnc.org
Appendix
Meteorology Causing Extreme Precipitation
(contiguous U.S., daily, 5-yr events)
Dominated by Large Systems
*
* Research area…
*
*
*
Extreme Precipitation (5-yr Storms)
CMIP5 Precipitation
Crop Yields Decline…Fisheries Shifting
Observed Increases in Frost-Free Season
Surface Temperature and Sun’s Energy
Carbon Emissions in the Industrial Age
NCEI Data Assets
LAND-BASED
STATION DATA
SATELLITE
RADAR
MODE
L
WEATHER
BALLOON
PALEOCLIMA
TOLOGY
SEVERE
WEATHER
MARINE /
OCEAN
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
U.S. Local Climatological Data
Global Historical Climatology Network-Daily (GHCN-D)
Integrated Surface Data, Hourly, Global
U.S. Climate Normals Products
Storm Data Publication/Database
Climatological Data Publication
Hourly Precipitation Data Publication/Database
U.S. Annual Climatological Summary
Weather Maps/Charts
Comparative Climatic Data (CCD)
Climatic Wind Data Publication
Climate Maps of the United States
CD-ROMs/DVDs
Climate Indices
Global Summary of the Day (GSOD)
U.S. Historical Climatology Network (USHCN)
Climates of the World/Historical Climate Summary
Historical Significant Events Imagery
Regional Climate Centers/National Weather Service
Products
Various Webpage Resources
Analyses/Climate Monitoring
Model Data
Radar Data
Temperature Change by Decade
Summer Arctic Sea Ice Now Projected to
Disappear by 2050