Climate Change and Flooding in Wisconsin Ken Potter Department of Civil & Environmental Engineering University of Wisconsin Madison, WI 2011 WAFSCM Annual Conference November 3, Pewaukee

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 IPCC 2007 General Circulation Models (GCMs) simulate the effects of incoming and outgoing thermal radiation on global circulation, and include: Atmosphere Clouds Oceans Topography Rainfall etc.

Dynamical Downscaling (Regional Circulation Models, RCMs) Statistical Downscaling of GCM output 6 RCMs driven by 4 GCMs North American domain Modern: Future: hourly output, Many variables (T, P, H, SM, etc.) 50 km horizontal resolution 20 GCMs Wisconsin domain Modern: Future: Daily output, Fewer variables (T, P, H?) 10 km horizontal resolution Climate Model Projections

Downscaling: Focus global projections to a scale relevant to climate impacts in Wisconsin 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 D. Vimont, UW-Madison

Temperature (°F) Significant warming is projected Change in average annual temp 1980 to 2055

Precipitation Change in annual average 1980 to 2055 (inches) 1.25” to 2.25” and 2-3 days/decade = modest future increase Increase in 2” rainfalls 1980 to 2055 (days/decade)

Winter Precipitation 1980 to 2055 Precipitation as snow reduced by 20% by mid-century = 30% decrease in midwinter snow depth Notaro et al Reduced Snowfall (%) Increasing (water inches)

Increased Winter Rainfall Potter and Liebl, 2010

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 Increased Winter-Spring Flooding?

Based on statistically downscaled data developed by Kucharik, Lorenz, Notaro, and Vimont, UW Madison. Quantile Projections: Madison

Large Uncertainty in Rainfall Projections Which 100-year event do you prefer ? 6”or 9” Projected size of 100-year 24-hour storm for Madison, WI, based on 14 GCMs Schuster, et al ? 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 ; hourly output 50 km resolution

RCM 100-Year, 24-Hour Rainfalls for Madison

RCM 100-Year, 24-Hour Rainfalls for Green Bay

RCM 100-Year, 24-Hour Rainfalls for Milwaukee

Quantile-Quantile Plots for Top 100 Daily RCM Rainfalls: Madison

Quantile-Quantile Plots forTop 100 Daily Rainfalls: Green Bay

Quantile-Quantile Plots for Top 100 Daily Rainfalls: Milwaukee

Quantile-Quantile Plot for Top 100 Daily Rainfalls: Milwaukee

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?

The record we use may actually reflect a drier period (TP40, ). Are we designing for historical climate? Hydrologic design is based on experience. (i.e. history)

NOAA Atlas 14 Updating TP-40 Provides data for locations vs. TP-40 isohyetal maps South Beloit. IL More stations TP-40 vs. NOAA Atlas 14 Record n Years (avg) Hourly Stations 200 vs vs. 40 Daily Stations 1350 vs vs. 63 Longer period of record Revised statistical method

Davis Todd, et al 2006 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: , 2006

NOAA Atlas 14 - a work in progress Midwest States - due late

Storm transposition can be used to asses vulnerability? The 2008 storm in the Midwest could be used this way. What else can we do, given GCM uncertainties?

“Build upon the experiences of communities that have experienced recent extreme rainfalls to guide a state-wide evaluation of vulnerabilities...” - WICCI Stormwater Working Group Vulnerability assessment 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?