Bioengineering an Urban Solution

Sept. 1, 2003
There’s a reason they call the crescent-shaped swath of central Texas that curves northward from the Rio Grande River “flash-flood alley.” The potential for fairly frequent flooding of unusually high-magnitude has been proven to be higher in this region than anywhere else in the United States due to a combination of weather and topography. Central to the area’s semiarid landscape is the Balcones Escarpment, a crescent-shaped boundary more than 250 mi. long forming the “backbone” of central Texas’s flash-flood alley. This steeply sloped ledge runs just north of San Antonio and Austin, marking the transition from west Texas hill country to coastal plains. When a tropical storm drifting west from the Gulf of Mexico meets a polar air mass swirling down from the north along this geographic line of demarcation, sudden rains of incredible intensity can occur. That’s exactly what happened on the evening of May 24, 1981. Ominous thunderheads billowing as high as 45,000 ft. began to build late in the afternoon, and around 9:30 p.m. heavy rain started to fall. Lasting only four hours, the subsequent rainfall deposited more than 10 in. of water in parts of metropolitan Austin. The city’s Shoal Creek, which normally flows at a relatively placid 90 gal./min., reached a peak of 6 million gal./min. With potentially volatile weather conditions, generally hard-packed soil, and a landscape characterized by the near-ubiquitous presence of deeply cleft streambeds, central Texas is prone not only to frequently intense rainstorms but to almost instantaneous runoff and very limited infiltration. And in an urban environment, where development increases the amount of impervious cover and often presents a variety of obstacles to natural water flow, peak flood discharges can be magnified as much as 300%. This presented us with a challenge: Could we adequately address one of the most critical flood-control projects in the metropolitan area while still preserving the character and beauty of the riparian corridor? Identifying the Austin Metropolitan Area’s Most Troublesome LocationThe flash flooding that resulted from what has come to be known in Austin as the Memorial Day Flood of 1981 caused 13 deaths and more than $36 million in damage. The event also provided a galvanizing moment in the history of the city’s flood and erosion control efforts. The city had already begun to focus on these issues in the previous decade, passing the Waterway Development Ordinance (also known as the Creek Ordinance) in 1975, with regulations that required erosion controls from all new development, limited development within the 100-year floodplain, and emphasized the importance of retaining “the natural and traditional” character of existing waterways. The devastation of the Memorial Day Flood, however, led to a dramatically increased awareness of the need for more effective watershed protection and flood controls. In the years immediately following, voters approved more than $75 million in bond packages for capital improvement projects, and in the next decade, fee increases and more protective ordinances continued to promote Austin’s efforts to respond to its naturally flood-prone environment. In 1996, these efforts led to the creation of the city’s Watershed Protection Department to better manage the prioritization and funding of flood control, erosion control, and watershed protection.One of the new department’s first projects was a Watershed Protection Master Plan that identified 17 of the most critical watersheds throughout the metropolitan area, seven of them located south and 10 north of the Colorado River channel that winds through downtown Austin. An estimate was made of the number of structures affected in each watershed by two-year, 10-year, 25-year, and 100-year flood events. This survey located the greatest number of vulnerable structures in the Williamson Creek watershed, a 30-mi.2 system of tributaries and creeks in the far southeastern corner of the Austin metropolitan area. A Bridge Modification, Floodwall Construction, and Streambank Stabilization
Construction foundation (above) and geogrid for MSE bank(Below) First phase of bank stabilization completeWithin the Williamson Creek watershed, the highest concentration of vulnerable structures existed in the Creek Bend neighborhood, many originally built as armed-forces housing prior to Austin’s 1975 flood protection ordinances. Located on a horseshoe bend of the creek roughly 10 mi. from downtown Austin, 185 houses in the neighborhood were discovered to be vulnerable to 100-year flooding. Roughly 26 mi.2 of the Williamson Creek watershed drains to the Creek Bend neighborhood – flash flooding can increase the normal water flow of roughly 200 gal./min. to a raging menace carrying as much as 13 million gal./min. downstream and flooding straight across the neighborhood’s horseshoe bend. We began our search for solutions that have a minimal impact on the environment with a thorough investigation of alternatives. First, we conducted a detailed hydrologic and hydraulic study of Williamson Creek throughout the reach of the project area, determining possible contributions to increased flood stages and potential treatments. More than seven sets of alternatives were considered to evaluate a variety of flood and erosion control components against a matrix of benefits-versus-impacts. These included traditional methods of channel enlargement, the possibility of building a bypass channel, and even the possibility of building a bypass tunnel to nearby Onion Creek. With each alternative providing equal safety against flooding, overall project costs, constructability, and environmental benefits proved to be the deciding factors. Conceived in two phases to lessen the total funding impact, solutions to flooding of the Creek Bend neighborhood include modifying a downstream bridge originally built within the floodplain, acquiring flood-prone property within the neighborhood, constructing an extended floodwall and a related expansion of a floodplain terrace above the natural channel, stabilizing a badly eroding streambank, and improving the neighborhood’s internal drainage. The design of the creek-related improvements was guided by two principles: maintaining natural flow to the greatest extent possible, and controlling the impact of floods with minimal environmental impact and maximum environmental benefit.Creating a Greenbelt Segment Using Neighborhood InputThe recent history of erosion control and channel stabilization projects relying on biotechnical solutions has developed, to a great extent, on the basis of well-established empirical and biological data. Nevertheless, we felt we had to come up with some project-specific design solutions based on traditional engineering principles and standards, primarily because of the risk of continued flooding and the sheer magnitude of potential water flow being considered in the Creek Bend project. This resulted in what we consider to be an unprecedented combination of engineering analysis and traditional design methodology combined with innovative design solutions, especially in the application of biotechnical methods, evident throughout the Creek Bend project. Key considerations in the first phase of this unique mix of flood control techniques focused on the selection and design of the concrete floodwall and a related floodplain bench as well as the specifications and construction details of the reinforced streambank. Producing a considerably smaller footprint than the option of an earthen levee, the choice of a concrete floodwall and floodplain bench ultimately resulted in the maintenance of Williamson Creek’s natural channel capacity while requiring considerably less channel excavation to provide for adequate floodplain conveyance within the natural stream corridor. Bordering the Creek Bend neighborhood and designed to comply with Federal Emergency Management Agency requirements, the floodwall also will guard against the safety hazard of a high-velocity surge of floodwater coursing through neighborhood streets. Choosing the option of a concrete floodwall provided many additional benefits as well. Along with the acquisition of 22 houses, the construction of the floodwall provided the opportunity to effectively create a new greenbelt segment within the Creek Bend neighborhood, expanding open park land, saving many large trees, and extending adjacent backyard areas. This also conformed with the City of Austin’s plans to link the neighborhood with nearby park areas by means of a future hiking and biking trail. And since the floodwall would result in a significant new feature defining the neighborhood landscape, obtaining support from the neighborhood’s local residents became an important consideration. This led a coordinated program of resident input, with neighborhood meetings held to discuss preliminary and final designs. At these meetings, Creek Bend residents were able to voice their own preferences for floodwall surface treatments, vegetative covering, and landscaping features. Old Landscaping Techniques With New Materials TechnologyTo offset conveyance losses created by the floodwall, the Creek Bend project design also called for more than 1,500 ft. of related channel improvements. These improvements included broad, overbank terraces designed specifically to maintain the natural flow of the active channel while addressing the full range of flood control and erosion concerns. As opposed to the more conventional solution of wide, trapezoidal, gabion- or concrete-sided channels used in recent decades, vegetated floodplain terraces have been created where required using a combination of mechanically stabilized (reinforced) earth (MSE) planted with native-adapted vegetation and gabion toe walls to both reinforce the low-flow channel banks and maintain the existing stream morphology. A goal of the floodplain terrace was to allow for extreme event flows while maintaining the natural sediment-transporting characteristics of Williamson Creek during lower-flow, more frequent storm events. Concrete floodwall and open green spaceParticularly challenging was the treatment of high-bank erosion within the project reach, especially in those areas related to the sharp bend of Williamson Creek downstream from the floodwall that gives the Creek Bend neighborhood its name. Flood events had badly eroded the bend’s steep slope, exposing a nonvegetated, earthen creek bank that was gradually eating into the backyards of residences above the stream. In these areas, a reinforced earth embankment roughly 20 ft. high was designed using a combination of geosynthetic and vegetative materials layered on a foundation of gabion anchors to prevent scour and undercutting of the embankment system. Polypropylene geogrids used in conjunction with flexible channel liners stabilize the design approach, creating a series of small, terraced embankments covered by soil and planted with native grasses and shrubs. The plantings, beyond their obvious aesthetic appeal, serve to enhance the retention of embankment soils while protecting the synthetic materials from ultraviolet radiation. As the plastics ultimately degrade, the long-term stability of the this embankment system combining old landscaping techniques and new materials technology will become increasingly dependent on the imbedded root systems of the planted vegetation.The selection and construction of a floodwall as a central flood control solution for the Creek Bend neighborhood led to several other improvements as well, both within the neighborhood itself and within the stream corridor. Since flood stages associated with Williamson Creek prevent local runoff from draining into the creek, an internal ponding facility was designed to contain 100-year flows behind the floodwall. Other drainage improvements within the neighborhood included two small bridge replacements, the addition of inlets and larger culverts to capture flows and deliver them to the local drainage system, and the replacement of a badly eroding drainage ditch with a 500-ft.-long, double box culvert designed in conjunction with the floodwall. The buried box culvert will convey local flood flows underground while allowing residents in proximity to the floodwall to reclaim backyards previously affected by channel erosion.Managing Flood Conditions and Preserving the EnvironmentTwo major flood events, one in October 1998 and another in November 2001, occurring during the project period provided us with the opportunity to field-test our designs for Williamson Creek and, in some cases, even make adjustments due to the presence of flood conditions. In fact, the design strategy for the first phase of the Creek Bend flood control project proved successful enough to significantly alter plans for phase two. Those efforts, which are currently approaching completion, include a major extension of the neighborhood’s floodwall and the related internal ponding facility designed to contain neighborhood runoff generated by 100-year flood events and prevented by flood stages from draining into Williamson Creek. Phase two activities also include lengthening the first roadway bridge downstream of the Creek Bend neighborhood and implementing additional channel improvements – both designed to reduce the bridge’s encroachment in the 100-year floodplain. The anticipated mitigating effects of the extended floodwall and the ponding facility allowed project designers to scale back the planned length of the bridge extension, from three spans to two spans, and reduce the required amount of channel improvements. When completed, the Creek Bend project will have removed all affected houses from the 100-year floodplain at a total cost of $10 million, a relatively small portion of the approximately $150 million total spent by the City of Austin on flood mitigation projects since 1981.The Creek Bend project has proven the viability of adopting flood control strategies that fit into the riparian corridor. The bank stabilization measures have proven capable of maintaining Williamson Creek’s shape and flood-carrying capacity while also preserving its rugged character. More importantly, the Creek Bend project demonstrates the existence of an expanding toolbox of flood control strategies that can replace traditional (bigger, wider) solutions while significantly expanding the range of options and benefits for both project designers and local communities. Foremost among these ought to be honoring the needs of the floodplain, re-creating greenspace within both urban and suburban environments, combining flood and erosion control objectives while reducing unnecessary construction, and restoring natural landscapes. The benefits of environmentally friendly flood and erosion control efforts also include an additional community and human-resources component, encouraging resident involvement, requiring cross-departmental cooperation, and building expertise in disciplines applicable in future projects. This certainly has been the case with Creek Bend, which now is serving as a model for the next generation of Austin flood control projects. Resting on a foundation of traditional engineering, bioengineering expertise and traditional landscaping practice, the success of the Creek Bend flood control project also shows that the management of major flood conditions and the preservation of the natural landscape can be mutually achieved, even in an urban environment, and that visually “transparent” and functionally effective flood control solutions exist that can both protect local residents and respect the natural environment.