Demonstra+on  Storms  for   Iden+fying  Climate  Vulnerability     Ken  Po'er   Department  of  Civil  &  Environmental   Engineering   University  of  Wisconsin   Madison,  WI   Acknowledgements   •  David  Liebl  (Co-­‐PI),  Faculty  Associate,  Dept.  of   Engineering  Professional  Development,  UW  College   of  Engineering     •  Doug  Brugger  and  Daniel  Fletcher,  graduate   students  in  Civil  &  Environmental  Engineering,  UW-­‐ Madison   •  The  NOAA  Climate  and  Societal  InteracIons  –  Sector   ApplicaIons  Research  Program    (NOAA  CSI-­‐SARP   NA12OAR4310098)     Overview   •  General  CirculaIon  Models  (GCMs)  downscaled  for  Wisconsin   predict  significant  increases  in  the  magnitude  and  frequency   of  large  rainfall  events.   •  But  there  are  very  large  variaIons  across  models,  limiIng   their  direct  use  for  water  resource  design,  management,  and   planning.   •  We  believe  climate  adaptaIon  should  focus  on   –  IdenIfy  vulnerabiliIes;   –  AdapIng  design  and  management  methods  to  reduce  the  most  criIcal   vulnerabiliIes.   •  Storm  transposiIon  appears  to  be  useful  adaptaIon  tool   based  on  our  experience  in  Wisconsin.   Regional  Climate  Model  (RCM)  Projec+ons   for  Wisconsin   Comparisons  of  RCM   projecIons  of  storm  intensity   and  recurrence  interval  for  the   periods  1971-­‐2000  and   2041-­‐2070   Year   80   70   60   50   Historical   40   NARCCAP   30   20   10   0   Inch   1.0"   7   6   2.5"   3.0"   3.5"   4.0"   5.0"   Vavrus  and  Behnke   5   4   Historical   3   NARCCAP   2   1   0   1  year   5  years   10  years   25  years   50  years   100  years   Vavrus  and  Behnke   Both  intensity  and  RI  are   projected  to  increase.   Sta+s+cally  Downscaled  24-­‐Hour   Rainfalls  for  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  sta+s+cally  downscaled  data  developed  by  Kucharik,   Lorenz,  Notaro,  and  Vimont,  UW-­‐Madison.   Regional  Climate  Model  (RCM)   100-­‐Year,  24-­‐Hour  Rainfalls   8 9 7 8 7 Depth inches Depth inches 6 5 Observed 6 Observed 5 4 4 3 3 2 2 1970-1999 2040-2069 1970-1999 Red  bar  indicates  the  mean.   Green  Bay   Milwaukee   2040-2069   Quan+le-­‐Quan+le  Plots  for  Top  100  Daily   RCM  Rainfalls:  Madison     HRM3_hadcm3 6 5 5 4 Modeled Rainfall (inches) Modelled Rainfall (imches) CCRM_ccsm 4 3 3 2 1 2 0 1 1 2 3 4 Historical Rainfall (inches) 5 6 0 1 2 3 Historical Rainfall (inches) 4 5   Quan+le-­‐Quan+le  Plots  for  Top  100  Daily   Rainfalls:  Green  Bay     CRCM_ccsm HRM3_hadcm3 5 6 5 Modeled Rainfall (inches) Modeled Rainfall (inches) 4 3 2 1 0 4 3 2 0 1 2 3 Historical Rainfall (inches) 4 5 1 1 2 3 4 Historical (inches) 5 6   Quan+le-­‐Quan+le  Plots  for   Top  100  Daily  Rainfalls:  Milwaukee     CRCM_ccsm HRM3_hadcm3 5 7 6 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 0 1 2 3 4 5 Historical Rainfall (inches) 6 7 Given  GCM  uncertainIes,     how  shall  we  “adapt”  water   management  to  likely  increases  in  the   magnitude  and  frequency  of  extreme   precipitaIon  events?     Assess  Vulnerability   •     Floodplains   •     At-­‐risk  road-­‐crossings   •     Stormwater  BMPs   •     Sanitary  sewer  inflow  and  infiltraIon   •     Emergency  response  capacity   •     Wells  and  sepIc  systems   •     Hazardous  materials  storage   Build  Long-­‐Term  Resilience   •   Modify  structures   •   Change  design  standards   •   Strengthen  ordinances     How  can  we  best  iden+fy   vulnerabili+es?   We  can  begin  by  focusing  on  places   that  have  not  been  hit  by  extreme   events  in  recent  memory.   Large  Wisconsin  Rainfalls  Since  1950   Note  that  many  gages  have   not  measured  a  daily  event   larger  than  5  inches  since   1950,  while  others  have   measured  over  9  inches.   Infrastructure  expansion   in  such  areas  may  be   vulnerable.   11.75”   Liebl  and  Schuster   We  then  model  the  behavior  of  systems  when  stressed  by   an  extreme  storm  that  recently  occurred  somewhere   nearby.  We  call  this  storm  transposiIon.   Past  Uses  of  Storm  Transposi+on   •  Design  of  the  flood  control  pracIces  associated  with   the  Miami  Conservancy  District  (Arthur  Morgan)   •  Development  of  probable  maximum  precipitaIon   •  Floodplain  mapping  based  on  conInuous  rainfall-­‐ runoff  modeling  (Bradley  and  Po2er)   •  EvaluaIon  of  vulnerabiliIes  to  flooding,  landslides,   and  coastal  erosion  and  inundaIon  in  California   (USGS)   Basic  Steps   •  Reconstruct  spaIal  and  temporal  pa'ern  of  one  or  more   major  historical  storms  based  on  NEXRAD  and  rainage  data.   –  Requires  specialized  sonware.   •  Use  the  transposed  storm  as  input  to  a  hydrologic  model  of   “system”  under  invesIgaIon.   –  Requires  specialized  sonware.   –  To  make  fullest  use  of  rainfall  data,  model  should  be  able  to  predict  at   a  sub-­‐daily  Ime  step.   •  Use  the  storm  and  model  to  explore  management   quesIons.   Example:  Yahara  Lakes  Watershed   This  is  an  urbanizing  watershed   that  has  experience  some   flooding  in  the  past,  but  was   south  of  the  most  extreme  part   of  the  2008  storm.   –  What  would  have  happened  had   the  2008  storm  been  centered   on  the  lakes?   –  Are  the  Dane  County  stormwater   ordinances  sufficiently  restricIve   with  respect  to  volume  control?   –  What  about  closed  watersheds?   Stakeholder  Par+cipa+on      All  steps  are  being  conducted  in  consultaIon  with   a  stakeholder  group  that  includes  representaIves   of  the     •  •  •  •      City  of  Madison;   Dane  County  Land  &  Water  Resources  Dept;   Madison  Metropolitan  Sewer  District;   Wisconsin  Department  of  Natural  Resources.   Rainfall  Data   NOAA-­‐NCDC  NEXRAD  Data   •  ReflecIvity  of  radar  signal  -­‐>  precipitaIon  depth   –  High  SpaIal  ResoluIon   •  Processed  by  Prof.  Jim  Smith  and  Mary  Lynn  Baeck  at   Princeton  University   –  Mean  field  bias  correcIon  using  USGS  and  CoCaRaHS   gages  (~130)   •  15-­‐minute  cumulaIve  precipitaIon  depths   •  SpaIal  coverage  by  three  radar  staIons   –  LaCrosse,  Milwaukee,  and  Davenport   Sobware  Development       TranStorm   –  Shins  locaIon  of  storm  rainfall  to  either   •  OpImize  total  rainfall  over  a  designated  area,  or   •  To  respond  to  user’s  specificaIons   –  Calculates  total  rainfall  at  centroid  of  model   subareas.   Preliminary  Results   What  would  have  happened  if  the   2008  rainfall  had  been  centered  over   the  Yahara  Lakes?   Lake  Mendota   Rainfall   Rise   Stage   Ouclow     2008   storm   10.7″₺   2.3  ʹ′   852  ʹ′   632  cfs   Transposed   storm   13.6″₺   2.9  ʹ′   853  ʹ′   752  cfs   100-­‐year     flood   852  ʹ′   -­‐   Note  that  the  853’  stage  is  almost  a  foot  higher  than  the  outlet  structure!   Other  Applica+ons   •  Black  Earth  Creek   –  Largely  rural  watershed  that  will  likely  have  significant   populaIon  increases.   –  Should  the  county  increase  detenIon  standard  from  a  10-­‐ year  event  to  a  100-­‐year  event?   •  Starkweather  Creek   –  Watershed  above  the  Dane  County  Airport  is   undeveloped.   –  Is  the  conveyance  through  the  AP  sufficient?   Conclusions   •  For  much  of  the  US,  the  magnitude  and  frequency  of  large   rainfall  events  is  likely  to  increase  under  climate  change.   •  But  there  are  very  large  variaIons  across  climate  models,   limiIng  their  direct  use  for  water  resource  design,   management,  and  planning.   •  Instead,  we  believe  climate  adaptaIon  should  focus  on   –  IdenIfy  vulnerabiliIes;   –  AdapIng  design  and  management  methods  to  reduce  the  most  criIcal   vulnerabiliIes.   •  Storm  transposiIon  appears  to  be  a  useful  adaptaIon  tool   based  on  our  experience  in  Wisconsin.