With a total length of 58 miles, Clear Creek is divided into two (2) designated segments for the Texas Natural Resource Conservation Commission’s (TNRCC) surface water quality monitoring program. The tidal Segment 1101 begins at the confluence with Clear Lake in Galveston/Harris County and ends at a point 100 meters (110 yards) upstream of FM 528 in Galveston/Harris County. The non-tidal Segment 1102 begins 100 meters (110 yards) upstream of FM 528 in Galveston/Harris County and ends at Rouen Road in Fort Bend County. Both segments have been designated contact recreation waters in the State of Texas Water Inventory (305(b) Report). Both segments are nonsupportive of the contact recreations use because domestic point sources and urban runoff contribute to elevated bacteria levels. Conversely, both segments are on the 303(d) list of “impaired waters” in regards to “bacterial levels sometime exceeding the criterion established to assure the safety of contact recreation.” Likewise, according to the monthly fecal coliform data collected by the Galveston County Health District (GCHD) Pollution Control Division, both segments have repeatedly failed to meet the surface water quality standard (SWQS) for contact recreation. GCHD’s data indicates that Clear Creek fecal coliform levels are historically high during both wet and dry weather periods. Therefore, the overall goals of the Clean Rivers Illicit Connections Program is to reduce the fecal coliform bacteria concentrations in both the tidal and non-tidal segments of Clear Creek, and collect baseline monitoring data for portions of Clear Creek outside of Galveston County.

Introduction
With a total length of 58 miles, Clear Creek is divided into two (2) designated segments for the Texas Natural Resource Conservation Commission’s (TNRCC) surface water quality monitoring program. The tidal Segment 1101 begins at the confluence with Clear Lake in Galveston/Harris County and ends at a point 100 meters (110 yards) upstream of FM 528 in Galveston/Harris County. The non-tidal Segment 1102 begins 100 meters (110 yards) upstream of FM 528 in Galveston/Harris County and ends at Rouen Road in Fort Bend County. Both segments have been designated contact recreation waters in the State of Texas Water Inventory (305(b) Report). Both segments are nonsupportive of the contact recreations use because domestic point sources and urban runoff contribute to elevated bacteria levels. Conversely, both segments are on the 303(d) list of “impaired waters” in regards to “bacterial levels sometime exceeding the criterion established to assure the safety of contact recreation.” Likewise, according to the monthly fecal coliform data collected by the Galveston County Health District (GCHD) Pollution Control Division, both segments have repeatedly failed to meet the surface water quality standard (SWQS) for contact recreation. GCHD’s data indicates that Clear Creek fecal coliform levels are historically high during both wet and dry weather periods. Therefore, the overall goals of the Clean Rivers Illicit Connections Program is to reduce the fecal coliform bacteria concentrations in both the tidal and non-tidal segments of Clear Creek, and collect baseline monitoring data for portions of Clear Creek outside of Galveston County.

Methods
GCHD employed a three-phase approach to identify and eliminate sources of fecal coliform contamination throughout the Clear Creek watershed. In Phase I, area maps were developed, existing data was reviewed, and baseline data collection began for areas outside of Galveston County. GCHD learned that many cities and utility districts did not have complete maps, and therefore, were unable to share information about the various drainage systems located within their jurisdictions. Investigators finally went into the field with street maps and developed drainage maps of their own. These crude, but effective maps, identified all ditches and end-of-pipe outfalls within those ditches.

Besides using the data from 22 sample sites within the Galveston County portion of the Clear Creek watershed, data was acquired from TNRCC, the Harris County Flood Control District, and the City of Houston’s Environmental Health Division. The data was assimilated to ascertain two (2) pieces of information. One, there was a large portion of the Clear Creek watershed that was not being monitored. And two, when fecal coliform averages were calculated for each sample site, only a few locations were found to have averages below the pre-determined screening level of 1,000 MPN/100 ml. Review of this historical data also confirmed that GCHD had the only tributary data available and high fecal coliform numbers were common throughout the watershed. Therefore, the Phase II intensive surveys could begin almost anywhere.

While maps were being completed, bridge to bridge sampling was conducted to begin collecting baseline data for areas outside of Galveston County and to focus in on “hot spots.” This method of sampling proved to be good for ambient sampling but was too generalized for detecting “hot spots.” The illicit connections detection method was changed to sub-watershed investigations where samples were collected from every outfall having a flow, no matter how small. These Phase II investigations involved sampling each discharge twice to identify which outfalls had continuing problems. Those outfalls repeatedly exceeding the screening level were placed on a list for further study.

Phase III intensive investigations were performed on the underground municipal separate storm sewer systems (MS4's) having the highest contamination or the most visible contamination. Illicit connections were confirmed using any number of the following tools: laboratory tests for phosphate, fecal, enterococcus and E.coli bacteria, field ammonia and chlorine tests, and/or dye. Once identified, the responsible party was served a notice to make repairs or the problem was referred to another agency to facilitate a resolution. GCHD would always attempt to resample the outfall after repairs were made, but many times the outfall no longer had a flow.

Baseline monitoring has been conducted throughout the project period and will be continued. Thirteen (13) ambient sites were chosen to represent Brazoria, Ft. Bend and Harris Counties in addition to the sites currently monitored in Galveston County. Basic water quality parameters are being collected at each location on a monthly basis (See Appendix A&B) and data is being submitted to the Houston-Galveston Area Council (HGAC) for inclusion in the regional data base.

Discussion
Many obstacles arose which required restructuring of the program in order to meet grant objectives and work more efficiently . The first obstacle encountered was the lack of city MS4 maps. Having these MS4 maps would have allowed GCHD to begin pinpointing illicit connections immediately for better lab coordination. GCHD’s method for mapping the sub-watershed was the only alternative for determining the number and locations of every outfall. The second obstacle was identifying sources of contamination. Baseline sampling proved ineffective for locating specific high fecal coliform sources. Only two cross-connections were found during a three month period and no obvious “hot spots” were identified. Subsequently, the method was changed to sampling each outfall having a discharge. This technique is time consuming and labor intensive but very effective in locating contaminated discharges. The third obstacle was baseline data collection. Routine baseline data collection began mid-year rather than monthly from the beginning. Miscommunication between the project manger and field personnel plus a change in personnel caused the delay in regular ambient data collection. Ambient data will be collected on a monthly basis for the next two (2) years hoping to show an overall improvement in water quality.

Minor setbacks to the program also included malfunctioning meters and inclement weather. The YSI-85 and field ammonia meters were necessary when sampling and tracking, but malfunctions and needed repairs created down time. The weather also posed problems because the project involves dry weather sampling only. Sampling could not be conducted for 3-5 days after a significant rainfall event. There were at least two significant rainfall events within the months of September, October, and November of 1998 and also May and June of 1999. Plus, December 1998, January, and July 1999 had at least one significant rainfall event (See Table 17).

Despite the difficulties encountered, several lessons were learned involving sampling methods and investigation techniques. First, during the initial sampling event each outfall with discharge should be sampled for E.coli and enterococci along with fecal coliform which are better indicators of cross connections. Second, when resampling outfalls within a sub-watershed it is best to sample every discharge regardless of whether the fecal coliform count was above or below the screening level. To date, only outfalls with fecal coliform numbers greater than the screening level were resampled. Third, ammonia and chlorine field meters should not be used exclusively to investigate possible illicit connections. They were not always effective in locating the source of the contamination. Rather, dye should be used in conjunction with the meters to pinpoint the source.  

As of October 1, 2000, ninety-seven (97) sites with high fecal numbers have been identified. Sixty-five (65) of these sites have been investigated with the intent of locating the source of the fecal contamination. Investigations have pinpointed fourteen (14) sources and twelve (12) illicit connections have been eliminated. Nine (9) sources were identified as being related to intermittent sanitary sewer overflows, and six (6) problems were referred to the another agency for resolution. Approximately thirty (30) of these investigations remain unresolved and require further attention. In addition to the 97 problem sites, eleven (11) investigations were conducted to locate the sources of high chlorine residuals and soapy water. Nine (9) of these investigations identified broken drinking water lines, while two (2) located car washes discharging directly into the storm drainage system.  

Conclusion
Sixteen (16) tributaries or sub-watersheds were identified or designated within the Clear Creek watershed. A total of 1,140 storm water outfalls were identified and plotted. The outfalls exhibiting flow were sampled for fecal coliform, E. coli and enterococcus bacteria. When fecal coliform concentrations exceeded the pre-determined screening level of 1,000 MPN/100 ml, they were resampled. Outfalls with continuing high fecal numbers were considered “problem sites” and were placed on a list for further investigation. A total of 52 problems were identified and resolved. Until the project ends in August of 2001, the remaining problem sites will be resampled and investigated further so that the illicit connections can be pinpointed and repaired.   Each time an illicit connection is identified and eliminated both environmental and economic benefits are realized. One environmental benefit includes a reduction in total fecal coliform within the watershed. By reducing bacterial contamination, the quality of life in the watershed ecosystem improves and the threat of disease transmission from contact recreation activities decreases.

Cities, too, can realize an economic benefit. Broken sewer lines and illicit connections cause inflow and infiltration (I&I) problems. Storm water entering the sanitary sewer system generates hydraulic overloading of both area collection systems and wastewater treatment plants. The outcome is increased costs caused by plant washouts, elevated chemical and electrical usage, and potential permit violations.