Novel Application of UV Disinfection Targeting Urban Area TMDLs

Weracoba Creek in Columbus, GA is 303(d) listed as impaired with a TMDL Implementation Plan targeting an 85% reduction in fecal coliform bacteria. The creek is also listed as impaired with respect to macroinvertebrate biology. Construction structural BMP in the watershed was completed in October of 2007. The BMP, a unique flow control and offline treatment system, functions passively to:Disinfect all or a portion of dry weather flow through a 2 MGD medium pressure UV unit, reducing the concentrations of viable bacteria downstream in a highly used City park.Allow unimpeded movement of aquatic biology across the flow control during dry weather conditions (for operation when less than 100% dry weather disinfection).Attenuate streambed hydrograph velocities during wet-weather events and divert small events and extended hydrograph tail flows to offline high rate 10 MGD filtration and 2 MGD UV disinfection,Trap floating debris and sediments in an access area for periodic removal.The BMP operated hydraulically to achieve the functions as designed. The flow control was operated initially to divert all dry weather flow through the UV disinfection unit easily meeting the criteria for fecal coliform bacteria during steady state conditions. The flow control was changed for the later testing events to allow aquatic biology to pass back and forth unimpeded across the flow control. Although good for aquatic biology, bacteria standards could not be reached for partial dry weather disinfection due to the higg dry weather concentrations.Passive wet weather runoff attenuation and diversion to offline treatment worked extremely well. Approximately 7,000 pounds of organic matter and course and fine sediments were removed for every inch of rain. The most problematic element of the passive facility was the course screen that clogged during heavy leaf loadings. A diversion boom improved these conditions. Compressed media filtration removed the lighter particles to allow effective UV disinfection. Filter effluent TSS levels were generally between 10 mg/l and 15 mg/l. The BMP was able to completely treat the small more frequent runoff events from rainfalls of 0.2 inches or smaller. The offline BMP also treated the much longer lower flow tail end of the runoff event.Implementation capital and actual operation and maintenance costs were used with measured removals to compute cost-benefits. The BMP serves a 1350-acre watershed having a capital cost of $648 per acre. Annualized life cycle capital was computed as $57 per acre with an additional annual O&M cost of 47 per acre. The UV component of the O;amp;M cost was 13 per acre. The cost per pound of TSS and sediment removed was computed as 0.34. The cost per 10 million fecal coliform colonies inactivated is 0.03.Calibrating the EPA BASINS model using the dry and wet weather data was used to assess compliance with bacteria standards of magnitude, duration and frequency. Continuous daily output from the model over a five-year period in a rolling 30-day geometric mean format was compared to the winter and summer criterion of 1,000 cfu's and 200 cfu's respectively. Disinfecting all of the dry weather and a portion of wet weather flows reduced geometric mean excursions from 67% to just under 10%, meeting the threshold criteria for impaired waters in Georgia. Steady-state bacteria monitoring upstream of the BMP and downstream throughout the park and beyond showed that the BMP was effective at protecting the immediate park area and children play areas, but were increasing in numbers at the lower reaches of the park and beyond. E. coli concentrations were found to be well within EPA criteria throughout the watershed, indicating the sources were non-anthropogenic. It was also found that improper land operations, infrastructure maintenance and poor habits of individuals contributed to the high steady state fecal coliform levels.Aquatic biology surveys were conducted before and after the BMP was constructed. Weracoba Creek is highly armored with little sinuosity. The greatest natural stream conditions exist through the park having frequent passive and active recreational uses including children playgrounds adjacent and accessible to the creek waters. Macroinvertebrate communities in the Park improved significantly as a result of the BMP. Their recovery, represented as a one-category improvement, is expected to reverse the current impaired listing and meet any TMDL allocation anticipated in the future. Their quick recovery was attributed to their “fly in” capabilities and significantly improved substrate conditions due to reduced habitat-scouring and organic sediments. Fish communities showed little improvement but are expected to re-colonize due to improved substrate and food sources but will take more time as they must swim from downstream sources or be artificially stocked.