Floods. Brian McCallum, USGS GA Water Science Center Georgia Association Floodplain Managers Annual Conference March 24, 2010

Transcription

1 USGS Monitoring i of Historic i September 2009 Atlanta Floods Brian McCallum, USGS GA Water Science Center Georgia Association Floodplain Managers Annual Conference March 24, 2010 U.S. Department of the Interior U.S. Geological Survey

2 Overview USGS in Georgia The Making of a Disaster Storm Operations Photo Gallery P t St A l Post-Storm Analyses Lessons Learned Summary

3 USGS in Georgia Collected data since the 1890 s Over 50 different cooperators Very strong programs with the State of Georgia, Georgia Power, and the USACE-Mobile District

4 The USGS in Georgia District Office Atlanta About 80 employees Four field offices Albany Norcross Savannah Tifton SE Region Norcross, GA HQ-Reston, VA

5 The USGS in Georgia Priority Water Science Issues Water Availability Hydrologic Hazards Impacts from Non-point Sources Water Resources Development p in Coastal Georgia

6 USGS in Georgia A National Hazards Risk and Resilience Assessment Program Enhance our ability to collect and deliver RT info from earth-observation networks With our partners, assess the vulnerability of cities and ecosystems Ensure science is effectively applied to reduce losses Facing Tomorrow s Challenges, USGS Science in the Decade , Circular 1309

7 USGS in Georgia 313 surface-water stations 251 have raingages 50 have continuous water-quality 100% transmit hourly All have emergency random transmissions

8 USGS in Georgia Housing with raingage Pressure orifice line buried Outside gage (staff)

9 USGS in Georgia Why is a streamgage important? Flood warning/forecasting Flood control Hurricane surge Drought monitoring State Water Planning Water wars Water supplies for continued growth Water effluent discharges Hydro-electric power generation Navigation Safe bridge and roadway design Recreation Tourism Long-term climate analyses Modeling Allatoona Dam

10 The Making of a Disaster Average annual rainfall = 50.2 inches (Wikipedia)

11 The Making of a Disaster 72-hour radar rainfall accumulation 9/22/09 Average annual rainfall = 50.2 inches (Wikipedia)

12 The Making of a Disaster Average annual rainfall = 50.2 inches (Wikipedia)

13 The Making of a Disaster Courtesy: GEMA Average annual rainfall = 50.2 inches (Wikipedia)

14 The Making of a Disaster Pumps were primed September was a wet month due to tropical moisture Weekend of September 18 th had persistent rainfall Evening of Sunday, g y, September 20 th was when the sky began falling By Monday, September 21 st, 7 am, USGS crews were all in

15 The Making of a Disaster Streams respond Weekend rains put streams above normal By Tuesday 22 nd By Tuesday 22 nd most streams in North Georgia were setting new daily highs

16 Storm Operations By Monday afternoon, almost 50 streamgages g were above their ratings By Monday evening, more than 15 streamgages were not reporting because they were under water By Tuesday, the Chattahoochee at GA 280 and Fairburn gages were lost to floodwaters

17 Storm Operations In a five day period starting Monday, September 21 st, USGS crews: Made nearly 75 discrete flood measurements Flagged HWMs at about two dozen locations Repaired or installed temporary real-time gages at all 20 destroyed d gages Deployed hurricane storm surge sensors to attempt to get recession limb data and prep for next storm

18 Storm Operations In a five day period starting Monday, September 21 st, USGS crews: Extended ratings at more than 49 stations Computed flood frequencies on-thefly Collected water-quality samples throughout flooded region Routine briefings to USGS HQ Brought in two crews and spare equipment from NC & SC Released two press releases Performed at least 5 media interviews Captured event in digital photos

19 Photo Gallery Monday, Sept. 21 st USGS crews measuring Crooked Creek at Norcross

20 Photo Gallery Tuesday, Sept. 22 st USGS crews measuring Sweetwater Creek near Austell in i a boat on I-20!

21 Photo Gallery Tuesday, Sept. 22 st USGS hydrographer measuring flood flows with Acoustic Doppler Current Profiler (ADCP) on Sweetwater Creek

22 Photo Gallery Monday, Sept. 21 st Chattahoochee at Morgan Falls Dam Tuesday, Sept. 22 st USGS hydrographer h preparing to measure by boat

23 Photo Gallery Monday, Sept. 21 st USGS gage Bridge and nearby area flooded at Powder Springs Creek at Powder Springs

24 Photo Gallery Monday, Sept. 21 st USGS Hydrographer inspects flooded car for emergency officials at Peachtree Creek at Atlanta (Northside Drive) (courtesy AJC)

25 Photo Gallery Tuesday, Sept. 22 st Chattahoochee River at Atlanta (Vinings)

26 Photo Gallery Dog River at GA 5 near Fairplay Water was 12 feet above the gage house

27 Post-Storm Analyses Indirect measurements were surveyed and computed at a dozen locations High-water marks were used at about 6 more locations to extend ratings

28 Post-Storm Analyses Rural flood frequencies determined by Log Pearson Type III regression with a generalizedskew coefficient using Bayesian general-least squares regression model. (Gotvald, others, 2009) USGS has effort underway to update the same report for urbanized areas. Due in FY Flood frequencies from this event were calculated using provisional equations from unpublished urban effort.

29 Post-Storm Analyses For large magnitude floods, the rural and urban curves begin to converge* At these magnitudes, alterations to watershed are drowned out Still large effect of urbanization for smaller floods % impervious = 30.6 (>10% is considered urban) *Results are provisional and subject to change upon further review

30 Post-Storm Analyses Outliers: Overlapping exceedance curves* Almost identical between rural and urban % impervious = 22.1 (>10% is considered urban) *Results are provisional and subject to change upon further review

31 Post-Storm Analyses Outliers: Non-intersecting ti exceedance curves* Curves do not cross, but confidence limit envelopes still do converge % impervious = 19.5 (>10% is considered urban) *Results are provisional and subject to change upon further review

32 Post-Storm Analyses Sweetwater Creek near Austell, GA Peak stage was 30.8 feet Peak flow was 31,500 cfs Annual probability of occurrence was 0.2 percent (500-year) flood Previous peak was July 2005 at a stage of feet with 1 percent (100-year) probability

33 Post-Storm Analyses Dog River at GA 5, near Fairplay, GA Peak stage was 33.8 feet on 9/21/09 Peak flow was 59,900 cfs Annual probability of occurrence is 0.2 percent (500- year) flood 1 percent (100-year) flood discharge is 9,390 cfs

34 Post-Storm Analyses Yellow River at GA 124, near Lithonia, GA Peak stage was feet on 9/22/09 Peak flow was 17, cfs Annual probability of occurrence was between 1.0 to 0.5 percent (100- to 200- year) flood Previous peak was feet on July 7 Previous peak was feet on July 7, 2005 with flow of 8,430 cfs

35 Post-Storm Analyses Chattahoochee River at Atlanta (Vinings) Peak stage = feet on 9/22/09 Peak flow = 40,900 cfs Annual probability of occurrence was between 1.0 to 0.5 percent (100- to 200-year) flood Highest level gage ever recorded in 81 years Highest discrete flow measurement ever

36 Post-Storm Analyses Snake Creek near Whitesburg Peak stage = 17.3 feet on 9/21/09 Peak flow = 10,900 cfs Annual probability of occurrence was between 0.5 to 0.2 percent (200- to 500-year) flood NOTE: Chattahoochee at Whitesburg turned out to be between 50- to 100-year event

37 Post-Storm Analyses Chattooga River near Summerville Peak stage = feet on 9/22/09 Peak flow = 23,500 cfs Annual probability of occurrence was a 2 percent (50-year) flood Potential dam failure

38 Post-Storm Analyses Lake Lanier Was just over 1064 feet Rose to over 1068 feet Full pool is 1071 feet Lake Allatoona Was at 837 feet before Rose to feet Full pool is feet USACE estimated that flood on Chattahoochee would have been 4 feet higher

39 Post-Storm Analyses Other general items of interest: The 2009 Atlanta floods, when comparing drainage areas yields of similar creeks, the magnitude was worse than the 1977 Kansas City floods, considered d one of the worst in the past century, where 25 persons lost their lives NWS estimates that in areas that had 10 inches of rain in 24 hours, the rainfall probability bilit chance would be 1 in 10, Closest thing in Georgia was Tropical Storm Alberto in 1994

40 Post-Storm Analyses Other general items of interest (Nov. 2009): 16,981 dwellings impacted (FEMA) 3,482 businesses impacted (SBA) 10 lives lost (Georgia Medical Examiner) $193,403,000 total estimated damages $50,253,000 individual assistance $143,150,000 public assistance projects Hundreds of roads & bridges damaged Courtesy AJC

41 Lessons Learned The gages did their job! Even though 20 were lost, they showed early on that the magnitude and extent of event was beyond major Peak discharges compared to long-term records allowed for probability calculations Raingage data was invaluable to NWS

42 Lessons Learned Investments in Streamgaging Network paid off! Most gages were upgraded to HDR Acoustic technologies allowed for faster, safer measurement of flood Network is almost 3X bigger than 10 years ago better density of gages Rapid Deployment Gages Not so rapid!

43 Lessons Learned Never bet against Mother Nature! Event was a surprise Magnitude and extent was amazing No chance for preparation Flood chasing in deadlock traffic was extremely difficult

44 Lessons Learned Still a ways to go to better communicate flood threat Worked well with NWS Communicated frequently with county officials Sheer magnitude of flood fight precluded communications Still some misunderstanding of how real-time data is presented RT-HNS coming very soon!

45 Lessons Learned Need a better way to explain post-storm t estimates t of probability Limits of 0.2 percent chance of exceedence (500-year) flood don t allow for proper presentation of size of event Have to resort to If a flood were a cup of coffee Courtesy AJC

46 Lessons Learned Media A blessing and a curse Take time to do the interviews i Chance to provide clearer story Most don t have a technical background (which is actually good) Courtesy AJC

47 Summary The September 2009 Atlanta floods added a new chapter or two to the record books The role of the USGS in events like this sometimes goes unnoticed, but is a vital one Not much could be done to prevent a disaster like this from happening except provide real-time warnings The losses in life and property are very significant, but could have been much worse

48 Questions? Brian McCallum (770)