Transcript NOAA Atlas 14 - Wisconsin Initiative on Climate Change Impacts

2011 WAFSCM Annual Conference
November 3, 2011 - Pewaukee
Climate Change and Flooding in
Wisconsin
Ken Potter
Department of Civil & Environmental Engineering
University of Wisconsin
Madison, WI
Outline
• Global Circulation (Climate) Models (GCMs)
• WICCI findings based on statistical downscaling
of GCMs
• New findings based on Regional Circulation
(Climate) Models (RCMs)
• NOAA Atlas 14
• Storm transposition for vulnerability assessment
General Circulation Models
General Circulation Models (GCMs)
simulate the effects of incoming and
outgoing thermal radiation on global
circulation, and include:
Atmosphere
Clouds
Oceans
Topography
Rainfall
etc.
IPCC 2007
Dynamical Downscaling
(Regional Circulation Models, RCMs)
6 RCMs driven
by 4 GCMs
North American
domain
Modern: 1970-2000
Future: 2038-2070
3-hourly output,
Many variables
(T, P, H, SM, etc.)
50 km horizontal
resolution
Climate Model Projections
20 GCMs
Wisconsin
domain
Modern: 1950-2000
Future: 2001-2100
Statistical Downscaling
of GCM output
Daily output,
Fewer variables
(T, P, H?)
10 km horizontal
resolution
Wisconsin Initiative for Climate Change Impacts
• Used 14 General Circulation Models (GCM’s)
from IPCC 2007 assessment
• Debiased and downscaled using historical Wisconsin weather station data
• Result: a statistical range of potential climate change
GCM grid
Downscaled (8x8 km) grid
Downscaling:
Focus global
projections to a
scale relevant to
climate impacts
in Wisconsin
D. Vimont, UW-Madison
Temperature (°F)
Change in average annual temp
1980 to 2055
Significant warming is projected
Precipitation
Change in annual average
1980 to 2055 (inches)
Increase in 2” rainfalls
1980 to 2055 (days/decade)
1.25” to 2.25” and 2-3 days/decade = modest future increase
Winter Precipitation
1980 to 2055
Increasing (water inches)
Reduced Snowfall (%)
Notaro et al. 2010
Precipitation as snow reduced by 20% by mid-century
= 30% decrease in midwinter snow depth
Increased Winter Rainfall
Potter and Liebl, 2010
Increased Winter-Spring Flooding?
More rain in winter
+More intense rainfall
= More high water events?
= More groundwater recharge?
Gordy Stephenson
Cottle
Vulnerability is already high
during winter and spring
DNR
Quantile Projections: Madison
9
100-Year Recurrence Interval
24-hour Rainfall Depth (in)
8
7
6
5
4
2-Year Recurrence Interval
3
2
1961- 2000
2046 - 2065
2081 - 2100
Based on statistically downscaled data developed by
Kucharik, Lorenz, Notaro, and Vimont, UW Madison.
Large Uncertainty in Rainfall Projections
Projected size of 100-year 24-hour storm
for Madison, WI, based on 14 GCMs
Schuster, et al
Which 100-year event
do you prefer ?
6”or 9”
? Wetter or drier in summer ?
What about Regional Circulation
Models (RCMs)?
• Recently developed by North American Regional
Climate Change Assessment Program
(NARCCAP)
• 6 RCMs driven by 4 GCMs
• North American domain
• 1970-99; 2040-69
• 3-hourly output
• 50 km resolution
RCM 100-Year, 24-Hour
Rainfalls for Madison
9
8
Depth
inches
7
6
5
Observ ed
4
3
2
1970-1999
2040-2069
RCM 100-Year, 24-Hour
Rainfalls for Green Bay
8
7
Depth
inches
6
5
Observed
4
3
2
1970-1999
2040-2069
RCM 100-Year, 24-Hour
Rainfalls for Milwaukee
9
8
Depth
inches
7
6
Observed
5
4
3
2
1970-1999
2040-2069
Quantile-Quantile Plots for Top 100
Daily RCM Rainfalls: Madison
CCRM_ccsm
HRM3_hadcm3
6
5
4
Modeled Rainfall
(inches)
Modelled Rainfall
(imches)
5
4
3
2
3
2
1
1
0
1
2
3
4
Historical Rainfall
(inches)
5
6
0
1
2
3
His torical Rainfall
(inche s )
4
5
Quantile-Quantile Plots forTop 100
Daily Rainfalls: Green Bay
HRM3_hadcm3
CRCM_ccsm
6
5
5
Modeled Rainfall
(inches)
Modeled Rainfall
(inches)
4
3
2
4
3
2
1
1
0
0
1
2
3
Historical Rainfall
(inches)
4
5
1
2
3
4
Historical
(inches)
5
6
Quantile-Quantile Plots for
Top 100 Daily Rainfalls: Milwaukee
CRCM_ccsm
HRM3_hadcm3
7
5
6
Modeled Rainfall
(inches)
Modeled Rainfall
(inches)
4
3
2
5
4
3
2
1
1
0
0
1
2
3
Historical Rainfall
(inches)
4
5
0
1
2
3
4
5
Historical Rainfall
(inches)
6
7
Quantile-Quantile Plot for
Top 100 Daily Rainfalls: Milwaukee
RCM3_gfdl
8
Modeled Rainfall
(inches)
7
6
5
4
3
2
1
1
2
3
4
5
6
Historical Rainfall
(inches)
7
8
What do others think about
precipitation modeling?
Science Magazine, October 2011:
When the Seattle Public Utility asked
University of Washington climate scientist
Clifford Mass how big they should build the
pipes in a $750M storm drainage system, he
“couldn’t give them an answer.”
And a University of Wisconsin
professor…
In the same Science article, University of
Wisconsin Professor Greg Tripoli points
out that global circulation models can’t
“create the medium-size weather
systems that they should be sending
into any embedded regional model.”
So what should we do about
engineering design?
Hydrologic design is based on experience.
(i.e. history)
The record we use may actually reflect
a drier period (TP40, 1938-1958).
Largest Daily Rainfall
Madison, WI
Mississippi River at Clinton
1 105
4
8 104
Discharge
(cfs)
5
3
Rainfall
(inches)
2
4 104
2 104
1
0
1860
6 104
0
1880
1900
1920
1940
Year
1960
1980
2000
2020
1880
1900
1920
1940
Year
Are we designing for historical climate?
1960
1980
2000
NOAA Atlas 14
Updating TP-40
Provides data for locations
TP-40 vs. NOAA Atlas 14
Record
Hourly Stations
Daily Stations
n
Years (avg)
200 vs. 994
14 vs. 40
1350 vs. 2846
16 vs. 63
South Beloit. IL
More stations
Longer period of record
Revised statistical method
vs. TP-40 isohyetal maps
NOAA Atlas 14 – Updated Design Storms
Davis Todd, C.E., J.M Harbor, B. Tyner, Increasing Magnitudes and Frequencies of Extreme
Precipitation Events Used for Hydraulic Analysis in the Midwest,
Journal of Soil and Water Conservation, (61)4:179-184, 2006
Davis Todd, et al 2006
NOAA Atlas 14 - a work in progress
Midwest States - due late 2012
http://www.nws.noaa.gov/oh/hdsc/current_projects.html
What else can we do, given GCM uncertainties?
Storm
transposition can
be used to asses
vulnerability?
The 2008 storm in
the Midwest could
be used this way.
Vulnerability assessment
“Build upon the experiences of communities
that have experienced recent extreme
rainfalls to guide a state-wide evaluation of
vulnerabilities...” - WICCI Stormwater Working Group
Assess:
• Floodplains and surface flooding
• At-risk road-crossings
• Stormwater BMPs
• Sanitary sewer inflow and infiltration
• Emergency response capacity
• Wells and septic systems
• Hazardous materials storage
Building Long Term Resilience
• Planning for impacts 25 or 50 years out is challenging
• Adaptation to low-risk high-cost events requires
political support
• Can use simulations to understand high water impacts
Conclusions
• Rain event intensities will likely increase in WI
due to climate change, but the modeling results
do not provide a sufficient basis for engineering
design.
• Winter-spring precipitation changes appear to be
better supported. Runoff implications need
further study.
Conclusions
• NOAA Atlas 14 should be adopted statewide
when available.
• Storm transposition should be explored as a
method of evaluating vulnerabilities and
increasing resilience.
Questions?