During the late 1980s and early 1990s, residents of the small suburb of Sandy Springs, GA, north of Atlanta, began to notice an increase in siltation and flooding in their watershed. The affected neighborhood was built over a sustained period during the 1950s and 1960s, and most of the zoning in this portion of the watershed allowed one house per acre. Residential developments built in the 1980s and 1990s were zoned for higher density. As residential development expanded, large-scale commercial medical complexes were built at the headwaters of the watershed. To serve this new development, a connector road, the Glenridge Connector, was built from State Route 400 to the medical complexes, separating the residential neighborhoods from the upstream commercial development. It was after these zoning changes and subsequent construction that the residents began to notice a corresponding increase in flooding and siltation.As a result of these changes in the neighborhood, residents organized the Ridgeview Neighborhood Civic Association (RNCA) to address issues confronting their community and began to negotiate zoning resolutions in their area to ensure the compatibility of the new construction with existing zoning. Although they prevented commercial encroachment into their neighborhood, they were unable to achieve the erosion control and stormwater management ordinances they had sought or to influence zoning decisions.Site ConditionsThe watershed is drained by one perennial stream with a series of dry-weather drainage channels forming the typical dendritic morphology of a mature watershed. During residential construction in the 1960s, one developer-architect constructed a set of small lakes and ponds and diverted the stream into a wet-weather channel that circumvented the lakes. The diversion channel, or stream, was separated from the lakes and ponds by a 2- to 3-ft.-high berm. The base of the stream channel was roughly equal to the elevation of the top of the lake surface that it bypassed. Over time, two of the four constructed ponds silted in to form small upland-type riverine wetlands.As development progressed, the resident closest to the lower pond at the watershed study point recorded rainfall during storm events, photographed or videotaped the resultant stormwater runoff for more than one year, and established an empirical relationship between rainfall and stream flow. During moderate rain events (approximately 0.75 in./hr.), runoff in the diversion channel would overtop the berm and flow into the lower pond. If the rain event was more intense (yet still far below a 25-year event), flooding would be so severe that the surface elevation of the lower pond would equal that of the upper pond, a difference of 4.5 ft. over a 1.5-ac. area, creating one large pond out of the two. Homeowners testified that this phenomenon had not been seen until the commercial development and the Glenridge Connector were completed. The condition of the diversion channel supported the residents’ observations because the erosional features along the banks appeared to be recent. The stream channel architecture was widening and forming large gravel beds in response to increased flow events. It is important to note that during this time of increased flooding and streambank erosion, northern Georgia was experiencing a severe drought that has lowered the level of some large lakes in the region by nearly 14 ft. Measurements collected by the homeowners indicated that flow volumes were increasing while rainfall was not, indicating that it was a result of development and not a natural event.Water quality deteriorated as flooding intensified, increasing turbidity as a result of inadequate erosion control measures in place at the time of construction. Pollutants associated with urban stormwater runoff also contributed to the degradation of water quality in the community watershed. The water was a very dark brown-orange and acquired the consistency of a muddy milkshake. Receding water would leave brown-orange silt on the streambanks and in the yards, and the ponds would often remain turbid for weeks after it rained.
Most of the newer developments had stormwater detention ponds that had been designed and constructed in accordance with standard practices required by Fulton County. There were no detention facilities or stormwater best management practices (BMPs) to control runoff from roadways that were built or widened to meet the needs of an expanding commercial development. Because of these changing conditions in the watershed, the RNCA board decided to focus on a study point at the lower pond, where the impacts were most obvious.MethodsTo determine the cause of flooding, a reconnaissance of the watershed was conducted by walking the streams and dry-weather channels, reviewing United States Geological Survey 7.5-quadrangle maps, reviewing development site plans, and observing rain events in the field. The initial watershed assessment did not identify any single specific condition, facility, or structure upstream that contributed to the flooding in the lower pond. Because no “culprit” structure was found, the problem was identifying the source of flooding and recommending the appropriate action to resolve the problem. Because detention facilities had been built in accordance with ordinances requiring a reduction of runoff to predeveloped rates and were certified by a number of engineers, the problem of initiating a solution to the flooding was difficult for RNCA. Clearly the runoff rates experienced in the neighborhood exceeded the predeveloped rate of stormwater runoff, because significant floods were occurring from rain events producing far less than five-year and 25-year rainfall amounts. RNCA had to answer one particular question: How could this extensive and frequent flooding occur when all new development had facilities to reduce the runoff to predeveloped (natural) rates? To unravel the problem, RNCA decided on a systematic, step-by-step approach to incrementally reduce the flooding and improve water quality. RNCA knew that this approach would take a great deal of commitment and unity among its members because first attempts might not produce noticeable results. Fortunately, this was not the case.The problem of siltation and turbidity was addressed first because it was the most dramatic and obvious. Turbidity was the best representation of the degradation of water quality in the basin and was the most easily resolved by the use of BMPs. Because there was only one new construction project within the boundaries of the watershed at the time of this investigation, the source of the siltation was easily identified. RNCA persistently observed, photographed, and/or videotaped the construction site and notified the Fulton County Department of Environment during every rain event to ensure that compliance was enforced. Resident “watchdogs” inspected the site from an adjacent parking structure and from public streets. Photos were sent to the county to show failed BMPs and the extent of sedimentation leaving the site. Finally, a toll-free hotline was used repeatedly to report these conditions. As a result, Fulton County required the developer and contractor to comply with erosion control standards with aggressive use of BMPs at all times. All exposed soil was covered with a thick layer of mulch, additional silt fences were installed correctly and maintained regularly. Additional BMPs, such as rock check dams, were installed in dry-weather drainage channels, and gravel exit pads were placed at the entrance to the site. Some stormwater was diverted to the other side of the site into a large detention pond that discharged into a neighboring watershed. All this activity greatly reduced the turbidity of runoff flowing to the study point.
Turbidity of stormwater during July 6, 1999, storm.
Same shot of lake during July 3, 2000, storm. Color of lake is significantly clearer in the 2000 storm due to improved EC methods at a construction site upstream.The photos demonstrate the effectiveness of this approach. The top photo was taken less than a year before the bottom photo. At the time the first photo was taken, construction had just begun at the commercial site and land clearing and grading were nearly completed. A 1.5-in. rain event generated the volume and turbidity seen in the photo. The bottom photo was taken during a similar rain event immediately after the site been brought into compliance by the rigorous program the homeowners had pursued. Notice the substantial improvement in turbidity. The first photo demonstrates the “red clay milkshake” consistency of the floodwaters, as described by the residents and county inspectors who witnessed the events.At the time that the regional pond was under consideration, an incident occurred with a homeowner adjacent to a detention pond in an upstream subdivision (Woodmore detention pond). This pond was constructed as a large concrete retaining wall across a small dry-weather “draw,” similar to a dam across a ravine. During one rain event, the retaining wall was nearly overtopped by stormwater. From an investigation of the county’s records, it appeared that the developer had not complied with a request by the county to restore the detention capacity of the pond to the original volume recommended in the Land Disturbance Permit. This pond had been used as a sediment retention pond during the construction of the subdivision. It had filled in as planned but was not excavated to the original capacity in order to properly serve as a stormwater detention pond after the construction phase. After excavation of the Woodmore pond and the modification to the OCS of the Falcon Chase Pond, detention capacity in two ponds was increased at the same time.
Click here for larger viewThe Woodmore pond was not considered by any of the parties to contribute a significant portion of runoff in the watershed because of its limited capacity and comparatively small drainage area. In addition, there was another pond between it and the watershed study point, so peak runoff from the Woodmore pond was delayed in the intermediate pond, thus mitigating the full effects of runoff during storms from that subdivision. However, the impact of runoff through the Woodmore pond produced an interesting effect on the diversion channel around the lower pond. Stormwater discharged from this pond into a sharp bend in the diversion channel, which impeded flow in the channel much as a weir or an obstruction in the channel would affect flow. The runoff contributed to water backup in the diversion channel and caused the flow to overtop the banks sooner than it would have if there had been no discharge from the Woodmore pond. Therefore, increasing the storage capacity of the Woodmore pond nearly eliminated the “tailwater” effect and prevented further overtopping of the diversion channel during moderate rain events. As a result of these modest changes to only two detention ponds in the watershed, flooding events in the RNCA neighborhood diminished in both frequency and intensity. Success was occurring slowly and incrementally as planned. Although improvements were noticeable, the goal of achieving predeveloped flow rates during storm events at less than the 25-year rate had not yet been realized. These results encouraged RNCA to continue looking for problem areas upstream. Several ponds and an underground detention vault were identified as potential sources of flooding. As before, RNCA adopted a program of surveillance to monitor their outflow.
Click here for larger viewBefore action was taken on any of the monitored detention facilities, homeowners who lived adjacent to a tributary dry-weather channel (Lot 5, Figure 2) complained about flooding in their backyards. Site inspections revealed recent erosional features that had undercut trees along segments of the banks in the channel. These features indicated that a much higher flow rate was discharged from the culvert near their property line than had previously flowed through the channel. Discharge through the culvert was the result of drainage from a 2.5-ac. sub-basin that included the Meridian Mark Medical Building. Runoff from this small sub-basin was routed through a large underground detention vault south of the medical building and then discharged through the culvert under the Glenridge Connecter near Lot 5. The vault was constructed like a concrete tunnel more than 120 ft. in length and about 10 ft. in diameter beneath the building. It was intended to contribute a major portion of the detention capacity to the watershed but did not appear to be performing as designed.After several meetings with the Fulton County Department of Environment, a dye test was conducted to inspect the movement of stormwater through the storm sewer system. The test was performed after a rainstorm had left measurable precipitation, but no water was flowing through the system at that time, so water from a fire truck flushed the dye into the detention vault. As water began to flow into the vault (observed at a maintenance manhole), it immediately flowed out over the weir at the same rate it had entered. This clearly demonstrated that the vault was not detaining stormwater during rain events. RNCA and Fulton County reviewed design drawings and inspection records for the vault, but nothing could be found to indicate the reason for a malfunction. The county sent inspectors into the vault at the request of RNCA to sketch and photograph the weir structure.
Click here for larger viewComparison of the as-built drawing and the inspection photos indicated that a 2-in.-diameter weephole had not been drilled near the bottom of the OCS weir during construction (Figure 3). The weep hole was intended to allow slow, controlled drainage of the vault after rain events. The problem was aggravated by the accumulation of sediment and debris at the weir wall that clogged the 12-in.-diameter hole in the middle of the OCS. The combination of these two factors prevented the vault from detaining any stormwater because water was permanently stored behind the weir. As a result, all the property owners along the dry-weather channel to the study point received full peak flow from this drainage basin during rain events. Several months after the weephole was installed and an additional 12-in.-diameter hole was built into the weir wall of the OCS, a 2.25-in. rain event occurred in the area over a two-hour period without flooding. This was the largest rain event since efforts to control flooding began in July 1999 and demonstrated a remarkable improvement over conditions a year earlier. This amount of rainfall one year earlier would have caused severe overtopping of the streambanks and overflow from the upper pond to the lower pond. ResultsAlthough flooding has been reduced with the modification of the three ponds, this neighborhood has not experienced a five-year to 25-year rain event since completion of large-scale commercial projects upstream. Further modifications might be needed after the watershed has received precipitation from a storm of that magnitude. The upstream area is still the object of intense speculation for further commercial development because a 13.5-ac. tract remains undeveloped, so additional capacity or additional modifications of existing ponds might still be necessary in the future.Results of these actions by RNCA significantly reduced the volume and frequency of flooding and improved water quality in its small watershed. Because such dramatic improvements were made to the flow regime and water quality, wildlife that had been absent has returned. The ponds in the area are no longer choked with sediment and resultant algal blooms. Habitat areas are no longer frequently flooded. Properties along the streambanks are now protected from the ravages of flooding and siltation, and homeowners no longer panic during rain events.ConclusionsThe actions taken by RNCA can serve as a model for other watersheds with similar problems. Adherence to strict land-use rules is not sufficient to maintain water-quality standards within suburban watersheds. Land use must include the aggressive use of a variety of erosion control and water-quality BMPs and a more conservative approach to estimating the predeveloped flow rates in the stormwater design and permitting process. Runoff coefficients used to estimate predeveloped flow rates are too high and result in inadequate storage capacity of detention ponds. In addition, improved stormwater management facilities are not sufficient to control runoff if they are not effectively maintained. By implementing this program, RNCA was able to improve water quality and reduce flooding at minimal cost and with no disruption to the natural watershed. The unintended benefit was a restored wildlife habitat, which is an improvement to the quality of life affecting every homeowner association.
