1 Unique Perspectives on the Historic Atlanta Flood of 2009: How a Cut-Off Low and Urbanization May Have Contributed to Disaster Marshall Shepherd, Thomas Mote, Pam Knox, John Dowd, and Mike Roden Introduction The 2007 Intergovernmental Panel on Climate Change (IPCC) 4th Assessment Report (Trenberth et al. 2007) established that Earth s mean temperature has increased in recent decades primarily because of anthropogenic activities. A warmer climate, through the well-established Clausius Clapeyron relationship, suggests that the atmosphere s saturation point for water vapor should increase. It is hypothesized that this process will lead to higher evaporation and precipitation rates and an accelerated global water cycle. If IPCC (2007) projections are accurate, the frequency and severity of extreme hydroclimate events (e.g., droughts, floods) will likely increase in response to the acceleration in the water cycle. Groisman et al. (2004) argued that heavy and very heavy rainfall events have already increased in the 20th Century, and Brommer et al. (2007) found that long durations rainfall events are becoming wetter. Recent literature suggests that damage, loss of life and costs from flooding have risen in recent decades (Ashley and Ashley 2008, Brissette et al. 2003). Seager et al. (2009) recently noted that regions of the southeastern U.S. face increasing vulnerability to hydroclimate extremes because of population growth and increasing population density. In 2008, a majority of world s population lived in urban areas, and by 2030 this number is expected to reach 81 percent (UNFPA, 2007). It is more than plausible that a modified water cycle will affect urban infrastructure and societal activities, thereby requiring augmentation to design, management, and policy. Burian et al. (2004), Burian and Shepherd (2005), and Reynolds et al. (2008) have considered implications of urban precipitation on the design of drainage systems and surface hydrological processes. Table 1-8-day rainfall totals across the area from 14 September 2009 (8 am EDT) to 22 September 2009 (8 am EDT). The National Weather Service and COCORAHS Network provided the data. (Note: 1 inch = 25.4 mm). Station Rainfall Totals (mm) Athens (ASOS) Cartersville (ASOS) Columbus (ASOS) 81.7 Gainesville (ASOS) NE Atlanta (ASOS) Peachtree City (ASOS) Rome (ASOS) West Atlanta (ASOS) Canton (COOP) Alpharetta (COOP) Doraville (COOP) Dacula 2.1 SSW (COCORAHS, Lead Author s Home) Carrollton (COCORAHS) 331.7
2 In September 2009, the metropolitan area of Atlanta and surrounding areas in northern Georgia experienced historic urban flooding. The flooding closed major transportation arteries in the Atlanta areas, closed several major school systems, submerged the popular Six Flags theme park, and contributed to at least 10 deaths at the time of writing (figure 1). In many parts of the metropolitan Atlanta area, the rainfall totals exceeded the 100-year flood, an event that has a 1-in-100 chance of occurring in a given year. Preliminary assessments by the U.S. Geological Survey (USGS) have estimated that some parts of the region experienced a 500-year flood. The flooding was exacerbated by significant antecedent precipitation earlier in the week, which saturated the soils in advance of the historic rainfall. It was the worst flooding along the Chattahoochee River since December 1919, when the city of West Point in western Georgia was almost destroyed by another flood driven by significant upstream precipitation in advance of a broad area of low pressure. Table 1 illustrates 8-day rainfall totals across the area from 14 September 2009 (8 am EDT) to 22 September 2009 (8 am EDT). The National Weather Service Automated Surface Observing System values are as high as 335 mm. The cooperative observing station in Canton, Georgia (the lead author s home town) recorded a total of mm of rainfall over the period. The normal rainfall for September in Atlanta is mm, yet totals for September 2009 (227.0 mm) were 219% above normal at Hartsfield-Jackson International Airport. The storm totals are more impressive when placed in the context that extremely wet Septembers may often be associated with decaying tropical systems (Shepherd et al. 2007, Schumacher and Johnson 2006). The National Weather Service s (http://www.srh.noaa.gov/ffc/html/09wetsep.shtml) climatological assessment revealed that September 2009 was the fifth wettest in Atlanta s history and the fourth wettest for Athens, Georgia. There were other record totals across North Georgia as well. Figure 1-(top) Flooding on Interstate 285 loop around Atlanta (bottom) Six Flags theme park (Images courtesy of the Atlanta Journal Constitution, ajc.com).
3 The goal of this short communication is to present a brief overview of the hydrometeorological conditions leading to the historic Atlanta area flooding, present multiple observational perspectives on cumulative rainfall totals for the event, and consider a hypothesis for the preferred regions of enhanced rainfall totals around the Atlanta areas. The hydrometeorological scenario Through the period of interest, a low-pressure system was stalled (a cut-off low or COL) over the southern Plains and lower Mississippi Valley. Griffiths et al. (2009) described the structure and evolution of COLs. They noted that, in one case study, a COL generated very high rainfall resulting in flash flooding in the Adelaide Hills area of Australia. Llasat et al. (2007) used a mesoscale model to investigate the role of COLs in heavy rainfall episodes on the Iberian Peninsula. They concluded that role of COLs in heavy rainfall events was mainly dynamic. The cyclonic circulation draws in very warm, low-level moist potentially unstable air with the presence of cold air at mid-to-high levels. Stein et al. (2000) also discussed a flash flood case in Europe that was associated with a COL advecting moist Mediterranean air into the Alps. While cut-off lows do exist across the region, floods produced by these events are not well documented. In a review of flood events over Texas, Nielsen-Gammon et al. (2005) document the role of a COL in producing the July 2002 flood over south central Texas. Schumacher and Johnson (2006) found that tropical cyclones were the most common producers of heavy rainfall across the southeastern U.S and mesoscale convective systems were the second most common. (In all other regions of the contiguous U.S they found that MCSs were the most common producers of heavy rainfall.) Of the 24 heavy rainfall events in the southeastern U.S. identified by Schumacher and Johnson ( ), only 6 of the events were synoptic in origin (a category which could include COLs as well as other synoptic-scale systems.) This comparison is more complicated than it first appears, as mesoconvective systems may be embedded within a broad area of precipitation that is enhanced by a favorable synoptic-scale environment. It is not clear how the September 2009 event would fit into the Schumacher and Johnson classification. In their analysis of unseasonable floods in the southeastern U.S Gamble and Meentemeyer (1997) found the document mechanism for unseasonal heavy rain events in Northern Georgia to be frontal activity and/or upper air enhanced. From Monday (14 September) to Friday (19 September) rainfall amounts ranging from 1 to 10 inches were recorded over northern Georgia, including the Atlanta area (NWS, Peachtree City). Figure 2 is the surface analysis at 7 am (EST) on 20 September. Though this low-pressure system weakened and lifted further north, a persistent south-southwesterly flow associated with the cyclonic flow advected Gulf moisture into the region. A review of Hydrometeorological Prediction Center surface frontal maps and discussions for September 2009 shows that origins of the system can be traced to as early as 8-9 September over Texas. Anti-cyclonic flow around a surface high pressure along the East Coast also supplied Atlantic moisture to the region. This meteorological environment combined with a series of upper-level impulses propagating into the Southeast brought steady and, at times, heavy rainfall rates across northern Georgia. Hydrologically, the region s soil and surface storage (streams, lakes, and rivers) were near capacity. Antecedent soil moisture is an important hydrological variable in flood analysis because it influences the partitioning of rainfall between runoff and infiltration. This process affects flows at the outlets of catchments (Aubert et al. 2003). This scenario set the state for sustained, intense rainfall during September to exceed an apparent critical threshold for catastrophic flooding.
4 Figure 2-Surface analysis from 20 September 2009 at 7 am (EST). Figure is courtesy of the National Weather Service. Multiple Observation Perspectives on the Historic Flooding: Traditional and Non-traditional Multiple observing systems were available to quantify the magnitude of the storm. These resources include satellite-based estimates, Doppler radar-based estimates, traditional NWS gauge networks (e.g., Table 1), and a relatively new community volunteer network. It is interesting and informative to note the particular merits of each of these resources for this unprecedented event. The Community Collaborative Rain, Hail, and Snow Network (CoCoRaHS, is an informal precipitation observing network that was formed in the aftermath of a flash flood event in Ft. Collins, Colorado, that produced mm of rainfall in July 1997 (Cifelli et al. 2005). Volunteer observers enter daily rainfall totals to a centralized database, and it is immediately available for use by stakeholders and the public. The CoCoRAHS data was particularly valuable for quantifying the 24-hour rainfall totals during the period of heavy rainfall that tipped the area over the aforementioned flood threshold. Figure 3 is the 24-hour rainfall totals for the period 20 September (7 am) to 21 September (7 am) over north Georgia. The heaviest cumulative rainfall totals are clustered around the western Atlanta suburban counties of Douglas and Carroll counties and the northeastern suburb of Gwinnett county, two regions that were hardest hit by flood damage. A CoCoRaHS observer just south of Douglasville in Douglas County reported a rainfall in excess of mm overnight, but was unable to empty the rain gauge before it overflowed and may have missed several inches of precipitation accumulation during that time. It is not uncommon to have persistent rainfall over particular regions due to a phenomenon called training, however, this process is most commonly observed in banded precipitation often associated with tropical systems or orographic features, which is a common factor in many of the aforementioned literature documented COL-flooding. Later in this manuscript, we speculate on another explanation for the distribution related to urban land cover.
5 Figure 3-24-hour rainfall totals (in) for the period 20 September (7 am) to 21 September (7 am) over north Georgia (courtesy of CoCoRaHS website). (Note: 1 inch=25.4 mm). Space-base observations also capture this extreme rainfall event. Figure 4 is the cumulative daily rainfall (12-23 September) as measured by the Tropical Rainfall Measuring Mission (TRMM) Multisatellite Precipitation Analysis (TMPA). TMPA is a 0.25-degree product based on the multi-satellite precipitation analysis (TMPA) described by Huffman et al. (2007). The daily product is composed of available microwave (e.g., TRMM microwave imager, Special Sensor Microwave Imager (SSM/I), Advanced Microwave Scanning Radiometer (AMSR) and Advanced Microwave Sounding Unit (AMSU) and calibrated infrared (IR) estimates. TMPA is available as at 3-hourly to monthly resolution. TMPA has been recently validated against ground-based gauges (Ebert et al. 2007, Su et al. 2008, Hand and Shepherd 2009) and has proven to be quite viable at the daily to monthly scale. It is clear that the heaviest rainfall totals (> 350 mm) over the period fell over the metropolitan Atlanta areas. There is an area of rainfall > 400 mm in the western suburbs of the Atlanta area. Figure 5 is a time-latitude Hovmoller diagram for the period September. This figure was constructed by averaging the longitude range of 85W-83W and then plotting the daily rainfall as a function of latitude and time. The figure seems to suggest the significant rainfall was initializing on the southern fringe of the metropolitan Atlanta area and intensifying over the urbanized metropolitan region of Atlanta.
6 Figure 4- Cumulative daily rainfall (mm) during September as measured by the Tropical Rainfall Measuring Mission (TRMM) Multi-satellite Precipitation Analysis (TMPA).
7 Figure 5- Time-latitude Hovmoller diagram of daily rainfall (mm) for September. Possible Urban Interactions Shem and Shepherd (2008) and Niyogi et al. (2006) recently used modeling and radar-based observational studies to illustrate how urban land cover might enhance pre-existing convective storms. Shepherd et al. (2009) and Shem and Shepherd (2008) discuss the mechanisms for this urban enhancement (enhanced convergence, destabilization due to urban heat island warming, and increased sensible heat flux) for the city of Atlanta. An examination of the Doppler rainfall-totals for September (figure 6) further suggests that there were preferred regions of enhanced precipitation. The radar-derived totals clearly indicate an arc of enhanced rainfall totals (light magenta and white) from the northwest suburbs (Douglas, western Cobb Counties) over to northeast suburbs (Gwinnett County). The Peachtree City soundings over the period resolve a prevailing southeasterly to southwesterly flow at low levels during much of the period. Shepherd et al. (2002) and Hand and Shepherd (2009), building on work of Changnon et al. (1981), developed a framework for describing where rainfall might experience urban enhancement under specific prevailing wind flows. Several studies (see Hand and Shepherd 2009 or Mote et al. 2007) have indicated that the preferred region for urban-enhanced precipitation is anywhere from 25 to 100 km downwind of the central business district. Under the scenario for the Atlanta flood event, the aforementioned northern arc of counties would be located in the downwind region part of the Atlanta land cover boundary (figure 6). It is equally apparent that parts of northeastern Georgia and South Carolina experienced elevated pockets of heavy rainfall (light magenta) that were likely orographically enhanced. Our preliminary results suggest that mesoscale forcing from urbanization (and orography) might have affected the distribution of heaviest rainfall within broader rainfall event forced by synopticscale forcing.
8 Concluding Statements Herein, we have described, what appears to be a rare Southeast flood episode associated with a cutoff-cyclone or low. The combination of this meteorological set-up coupled with an overburdened hydrological system led to a historic hydrometeorological event in the Atlanta area. Additionally, qualitative analysis indicates that it is plausible that some urban land cover interactions, as reported in the literature, could have influenced the distribution of the heaviest rainfall around the city. There is no conclusive evidence supporting this hypothesis at this time, but we are conducting a series of coupled atmosphere-land surface model runs to re-engineer the storm to ascertain how urban land cover and soilmoisture capacity likely affected the hydroclimate and land surface hydrology for this event. Figure 6-Doppler-radar rainfall totals (in) for (September 2009) with the urban rainfall enhancement framework (Shepherd et al. 2002) superimposed on the image. Black arrows represent an approximation of the mean low-level wind flow during the period. The radar image was provided by the National Weather Service, Peachtree City. (Note: 1 inch=25.4 mm).
9 References Ashley ST, Ashley WS (2008a), Flood fatalities in the United States. J Appl Meteorol Climatol. 47, Aubert, D., Loumagne, C., Oudin, L. (2003), Sequential assimilation of soil moisture and streamflow data in a conceptual rainfall-runoff model. J. Hydrol. 280, Brissette, F.P., Leconte, R., Marche, C., Rousselle, J. (2003), Historical evolution of the flooding risk in a USAQuebec river basin, J. Am. Water Resources Ass. Vol 39(6), Brommer, D. M., R. S. Cerveny, and R. C. Balling Jr. (2007), Characteristics of long-duration precipitation events across the United States, Geophys. Res. Lett., 34, L22712, Burian, S, Stetson W, Han WS, Ching JKS, Byun DW (2004), High-resolution dataset of urban canopy parameters for Houston, Texas. Proceedings of the AMS Fifth Conference on the Urban Environment, Vancouver: 9 pp. Burian SJ, Shepherd JM (2005), Effect of urbanization on the diurnal rainfall pattern in Houston. Hydrological Processes, 19(5), Changnon, SA, Semonin RG, Auer AH, Braham RR, and Hales J (1981), METROMEX: A review and summary. Meteor. Monogr., No. 40., Amer. Meteor. Soc. 181 pp. Cifelli, R., N. N. Doesken, et al. (2005), The Community Collaborative Rain, Hail, and Snow Network: Informal education for scientists and citizens. Bulletin of the American Meteorological Society, Volume 86, Issue 8 (August 2005), Ebert, E.E., J.E. Janowiak, and C. Kidd (2007), Comparison of near-real-time precipitation estimates from satellite observations and numerical models, Bull. Amer. Meteor. Soc., 88, Gamble DW, Meentemeyer VG. (1997) A synoptic climatology of extreme unseasonable floods in the southeastern United States, , Phys. Geog., 18, Griffiths, M., M.J. Reeder, D.J. Low, R.A. Vincent (2009), Observations of a cut-off low over southern Australia, Quarterly Journal of the Royal Meteorological Society, 124, Groisman, P.Y., R.W. Knight, T.R. Karl, D.R. Easterling, B. Sun, and J.H. Lawrimore (2004), Contemporary Changes of the Hydrological Cycle over the Contiguous United States: Trends Derived from In Situ Observations, J. Hydrometeor., 5, Hand, L.M., and J.M. Shepherd (2009), An Investigation of Warm-Season Spatial Rainfall Variability in Oklahoma City: Possible Linkages to Urbanization and Prevailing Wind, J. Appl. Meteor. Climatol., 48, Huffman, G.J., R.F. Adler, D.T. Bolvin, G. Gu, E.J. Nelkin, K.P. Bowman, Y. Hong, E.F. Stocker, and D.B. Wolff (2007), The TRMM Multisatellite Precipitation Analysis (TMPA): Quasi Global, Multiyear, Combined-Sensor Precipitation Estimates at Fine Scales, J. Hydrometeor., 8,
10 Llasat, MC, Martin, F, Barrera, A (2007), From the concept of Kaltlufttropfen (cold air pool) to the cutoff low. The case of September 1971 in Spain as an example of their role in heavy rainfalls, Meteorol Atmos Phys., 96, Mote, T.L., M.C. Lacke, and J.M. Shepherd (2007), Radar signatures of the urban effect on precipitation distribution: A case study for Atlanta, Georgia, Geophys. Res. Lett., 34, L20710, doi: /2007gl Nielsen-Gammon, J. W., F. Zhang, A. M. Odins, and B. Myoung (2005), Extreme rainfall in Texas: Patterns and predictability, Phys. Geog., 26, Niyogi, D., T. Holt, S. Zhong, P. C. Pyle, and J. Basara (2006), Urban and land surface effects on the 30 July 2003 mesoscale convective system event observed in the southern Great Plains, J. Geophys. Res., 111, D19107, doi: /2005jd Reynolds, S., Burian, S., Shepherd, J.M., and Manyin, M. (2008), Chapter 7: Urban induced rainfall modifications on urban hydrologic response, In: Reliable Modeling of Urban Water Systems. Edited by W. James et al. Computational Hydraulics International. Guelph, Ontario, CA. pp Schumacher, R. S., and R. H. Johnson (2006), Characteristics of United States extreme rain events during , Wea. Forecasting, 21, Seager, R., A. Tzanova, and J. Nakamura (2009), Drought in the Southeastern United States: Causes, Variability over the Last Millennium, and the Potential for Future Hydroclimate Change, J. Climate, 22, Shem, W, Shepherd JM (2008), On the impact of urbanization on summertime thunderstorms in Atlanta: Two numerical model case studies, Atmospheric Research, 92, Shepherd JM, Pierce H, Negri AJ (2002), Rainfall modification by major urban areas: Observations from spaceborne rain radar on the TRMM satellite, J. of Appl. Meteor., 41(7), Shepherd, J. M., A. Grundstein, and T. L. Mote (2007), Quantifying the contribution of tropical cyclones to extreme rainfall along the coastal southeastern United States, Geophys. Res. Lett. 34, L23810, Stein J., E. Richard, J. P. Lafore, J. P. Pinty, N. Asencio, and S. Cosma (2000), High-resolution nonhydrostatic simulations of flash-flood episodes with grid-nesting and ice-phase parameterization, Meteor. Atmos. Phys., 72, Su, F., H. Yang, and D.P. Lettenmaier (2008), Evaluation of TRMM Multi-satellite Precipitation Analysis (TMPA) and its utility in hydrologic prediction in La Plata Basin, J. Hydrometeor., 9(4), Trenberth, K.E., P.D. Jones, P. Ambenje, R. Bojariu, D. Easterling, A. Klein Tank, D. Parker, F. Rahimzadeh, J.A. Renwick, M. Rusticucci, B. Soden and P. Zhai (2007), Observations: Surface and Atmospheric Climate Change. In: Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. UNFPA (2007), The State of World Population 2007: United Nations Population Fund, United Nations Publications, 1 pp.