Thirty-four communities in three Kentucky counties, spanning more than 245 mi.2 of hilly, flood-prone terrain, hug the Ohio River. When the decision was made to make Sanitation District No. 1 of northern Kentucky the stormwater management authority for the region encompassing the urbanized areas of Boone, Kenton, and Campbell counties, officials had a huge task ahead of them. Among the 34 communities, only a handful had stormwater maps. Virtually none had information on the location, connectivity, and condition of the various stormwater systems and their features. Northern Kentucky Sanitation District No.1
The district contracted with the Covington, KY, project office of Woolpert LLP, a Dayton, OH-based civil engineering firm, to perform an inventory and condition assessment of the region’s open and closed stormwater drainage systems. After the data were collected, they would be loaded into a geographic information system (GIS). The GIS would contain information to do the following:Assist in developing a Stormwater Management Plan (SWMP) for the district’s National Pollutant Discharge Elimination System (NPDES)/Kentucky Pollutant Discharge Elimination System (KPDES) Phase II permitHelp the district know the extent and condition of the stormwater drainage system in northern Kentucky as the district assumed management responsibilitiesHelp the district with future operations and maintenance and master planningIn all, 63,000 features and 2,600 mi. of open channels were surveyed between 1999 and 2003. Today, Sanitation District No. 1 has a continuous stormwater drainage system in an ArcInfo Workstation GIS. The database includes the attribution information collected in the field at each structure. The GIS is being linked to a computerized maintenance management system (CMMS). Here’s how this huge inventory and condition assessment project happened – on budget and on schedule.Growing Area Goes for RegionalizationUnderstanding northern Kentucky’s need for such information begins with understanding how the region – and Sanitation District No. 1 – evolved over the years. Northern Kentucky’s Boone, Campbell, and Kenton Counties are just across the river from Cincinnati, OH. As Cincinnati has grown, so too has northern Kentucky – in part because of Cincinnati but also in its own right. Boone County is home to Cincinnati/Northern Kentucky International Airport. The county is bisected by Interstate 75, a major transportation artery that runs from the top of Michigan to the southernmost tip of Florida. All three counties are bisected by Interstate 275. For many years, the sanitation district had served northern Kentucky solely as operator of the region’s wastewater treatment plant and pumping stations. Then around 1994 the State of Kentucky created legislation that allowed the district to assume responsibilities for operating the wastewater collection systems for the communities. Today, the sanitation district owns, operates, and maintains more than 1,200 mi. of combined and separated sewers, one major wastewater treatment plant, nine small treatment plants, 135 pumping stations, and 15 flood-pump stations. As the district assumed responsibility for sewage collection, it also brought sanitary-sewer GIS efforts “in-house,” building on a GIS that had been initiated by the Northern Kentucky Area Planning Commission (NKAPC). Officials from Boone, Campbell, and Kenton Counties determined that they needed a regional SWMP, and because regionalization of sanitary-sewer services had proved a success, they determined that the district would be the best entity to handle such a program. A regional SWMP would benefit the area as follows:By providing equitable, cost-effective management of stormwater problems spread among the 34 communitiesBy helping enhance water quality in local streamsBy helping ensure NPDES/KPDES complianceThus, in 1998, the State of Kentucky created legislation that allowed the district to assume from the communities responsibility for all stormwater systems and activities over the next several years. Base mapping of each community’s stormwater system, however, was even more limited than what had been available for each sanitary-sewer system. A few of the communities’ public works departments had paper maps. Some smoke and dye testing data and limited stormwater attribution information existed in a layer of the district’s sanitary-/combined-sewer GIS. In 1999, the district, NKAPC, and the Boone County Planning Commission (BCPC) used aerial photography to create planimetric, topographic, and digital orthophoto base-mapping datasets; this information was used to generate ArcInfo coverages of the stormwater drainage system. These photos and datasets became outdated, however, as the inventory project progressed, and in areas covered by trees or heavy brush, the aerial photos provided little support. As a result, project staff had to have a solid understanding of stormwater systems.Basically the district had to start from scratch to obtain the location, attribution, connectivity, and condition information on many features:ManholesInlets and catch basinsPipes (12 in. or greater in diameter)Pipe inletsPipe outletsCulvertsRetention/detention pondsDitches and open channelsThe district again contracted with Woolpert to survey the region’s stormwater channels and structures and from this information create a comprehensive GIS.Determining Necessary Information, Establishing ProtocolsWoolpert started the process by meeting with district personnel – specifically engineering, maintenance, and field crews – to learn what kinds of attribution information would be used on a regular basis and needed for future uses. The project team discussed how attribution information would aid in preventive maintenance, modeling, a CMMS, illicit discharge determination and elimination, and asset transfer and management. Only truly necessary attribution would be included in the inventory project.Ultimately district staff determined that the following stormwater attributes needed to be collected:Location (horizontal coordinates at meter-level accuracy)ConnectivityMaterialDepth (e.g., pipe inverts, structures)Size (e.g., pipes, culverts)Flow characterization (turbidity)Digital photographs In addition, the condition of structures would be assessed for debris and cleaning and for rehabilitation and repair. Woolpert and the district worked to develop the initial field procedures for collecting objective attribution (i.e., pipe size, depth, and material) and subjective attribution (i.e., physical structure conditions and flow characterization). The inventory of the stormwater drainage system would be conducted using a ruggedized, pen-based field computer system with the real-time differential (RTD) global positioning system (GPS) survey method. RTD was chosen instead of real-time kinematic (RTK) GPS because RTD would provide the x and y (horizontal) coordinates at meter-level accuracy, which was precise enough for the district’s purposes and more cost-effective than the x, y, and z (vertical) sub-meter-level accuracy of RTK. The region’s hilly topology had a hand in this decision: Because of the slope in the area, adequate flow is not a problem, and therefore the district did not plan to do high-end hydraulic modeling, which would have required sub-meter-level accuracy. It was decided that top-of-structure elevations would be derived through an interpolation process using an existing digital terrain model provided by NKAPC and BCPC. Next came the logistics of surveying such a huge land area. Woolpert recommended breaking the 245-mi.2 service area into manageable basins initially based on United States Geological Survey 14-digit Hydrologic Unit Code (HUC) watersheds; these areas were further broken into sub-basins. As crews got into the process, they sometimes found a storm sewer flowing into another watershed, so sub-basins were modified as needed to reflect drainage patterns of the field. In all, 256 urban and rural sub-basins, each approximately 1 mi.2 in size, would be surveyed. Each sub-basin was assigned a field manager; two field staff members, who comprised a crew; and a GIS analyst. If any problems arose and the sub-basin needed to be revisited – whether in person or “virtually” via the GIS – the original personnel assigned to the sub-basin would handle the problem. This method made the entire project flow smoothly, from personnel management to data collection and processing to deliverables.Additionally, separate protocols for surveying sub-basins in urban and rural areas were established:For urban areas,crews would walk the entire open-channel system (including tributaries) until it intersected with a closed storm system or transitioned from overland flow in the upstream regions. Crews then would inventory the closed system.For rural areas, field managers and crews would first examine aerial photography to get a visual on whether stormwater structures even existed in the sub-basin and then choose which areas to walk. But, in some cases, entire subdivisions in rural areas (and urban ones) were not on the aerial photography because it had been flown in 1999 and new developments had been built since then. So crews also would drive the public roads of some rural areas for evidence of stormwater structures. Road culverts and open channels within the road right of way would be inventoried, as would closed sewer systems that existed within the road right of way, which could be seen from the road. If a closed system crossed the road right of way but was not contained completely within it, field staff would track and collect attribution data until the termination points of the system in each direction were reached. Fieldwork BeginsWoolpert began with a 6.1-mi.2 pilot area in Boone and Kenton Counties in 1999, which helped determine the appropriate methods and level of accuracy necessary for development of the districtwide stormwater GIS. Actual surveying began in 2000, with crews starting in Kenton County (the central county) and working west toward Boone County and east toward Campbell County.Two-person survey crews would start at the low end of each sub-basin and work their way up. As they came across channels and structures, they would “GPS” the locations and collect the applicable connectivity, attribution, and condition assessment information on pen-based computers loaded with Woolpert’s SmartSurveyor software (see sidebar). Each structure would receive a unique ID number based on the HUC basin, the sub-basin, and a feature identifier.In addition to surveying channels and structures, if crews saw something of interest, they would note it on their field computer. For example, crews noted illegal dumping in and along creeks, septic systems discharging into stormwater channels, broken laterals, and broken sanitary-sewer pipes. Basically anything that impacted stormwater quality was noted in preparation for NPDES/KPDES Phase II permitting requirements. Also, applicable communities’ public works departments were notified of any deficiencies found in their systems because transfer of responsibilities to the district still was to occur.If a public health or safety problem was found, such as sewage in catch basins and open channels or a missing manhole cover, crews immediately reported the problem to district dispatch. If it was a sanitary-sewer problem, the district would send personnel; if the problem was related to stormwater, because the district had yet to assume management responsibilities, it would notify the applicable community, which would respond with personnel. Woolpert crews did not leave the site until community or district personnel arrived. During the inventory and condition assessment portion of the project, crews encountered a number of challenges:Lack of access to buried structures. Crews used a probing rod or a metal detector to verify a buried structure’s existence and used a TV camera if an internal view of a pipe was needed. On occasion, crews exposed structures, but for the most part the ground was left undisturbed. Working in SmartSurveyor, crews coded the structure to note that it was buried. This notation also appeared in the GIS. Blocked GPS signals. For structures located in heavily wooded areas or urbanized areas where the GPS signal was blocked, rendering the RTD GPS inaccurate, locations were made by digitizing the obscured structure into the pen-based field computer using the NKAPC and BCPC planimetric mapping as a reference.City of Florence, KY. The City of Florence has its own stormwater system and program and, as such, was not a part of the district’s stormwater project. Portions of Florence’s stormwater system, however, flow into the district’s system, and vice versa. Crews worked around the city (which is located in Boone County near the Kenton County border) to obtain survey information and then, in SmartSurveyor, used inferred lines or segments to create map connectivity. Inferred lines were noted as such and coded in a separate color in the GIS. The result is a seamless and continuous system.Cincinnati/Northern Kentucky International Airport. Similarly, crews worked up to the boundaries of the 7,000-ac. airport, which maintains its own private stormwater system, and used inferred segments to create a seamless system.Other sites with private stormwater systems. If a public sewer collected runoff from a private site, crews would collect information until the structure was on private property. If a public sewer (main trunk lines only) ran through a private site, crews would inventory the structure through the private site.Highways. Woolpert worked with the Kentucky Department of Highways to obtain records drawings. Crews collected information up to the right of way along an interstate but, for safety reasons, did not enter the roadway. They used the records drawings to know where to start picking up information on the other side of the road. Inferred segments were used to create a seamless system.Other gaps in the stormwater drainage system. Woolpert and district staff attempted to resolve other connectivity issues in the field, such as with dye testing or the use of TV cameras in pipes, to discern why a gap existed. Crews also tried to obtain information from applicable communities’ public works departments. Sometimes the conclusion was to defer further investigation until the district assumed management of the entire system. If connectivity was not achievable, inferred segments were used to create a seamless system.Unique structures and situations. Unique structures and situations were flagged and photos were taken as necessary for review by district staff. Photos and reports were submitted to the district upon completion of each sub-basin so district personnel could conduct further investigation.
Left and right: A crew chief measures the dimension of an outfall pipe. Quality Assurance and ControlQuality assurance (QA) and quality control (QC) were integral to all portions of the project – from fieldwork to GIS creation to final deliverables. QA was used to reduce the potential for error during the production phase of the project (i.e., data collection and development of ArcInfo coverages). QC was performed at the end of processes.The complexity and size of the district’s project demanded accurate, frequent, and coordinated communication – verbal and written – as the foundation for all QA/QC processes. Project team members used the following communication methods to ensure project success:A comprehensive reference manual. At the beginning of the project, Woolpert created a reference manual – a “living document” – which spelled out all work processes and procedures. The manual ensured that everyone involved in the project followed the same work methods. This communication tool enhanced productivity, consistency, and accuracy. Also, if a process or a procedure had to change – whether for safety, technical, or other reasons – it was tracked and documented in the reference manual. Changes to the manual became known as “one-pagers” that Woolpert and district team members would review and agree on. Now that the project is completed, the manual is an asset that has been turned over to district staff. It will be used for years to come and modified, as needed, to include future updates.Regular meetings. Woolpert project team leaders regularly met with district and Woolpert staff. By design, the district was actively involved in the project from inception: During the first phase of the project, Woolpert project leaders and district representatives met every week; as the project “matured” and processes were refined, meetings were reduced to twice and then once a month.Communication protocol. With the large number of Woolpert staff involved (30-plus people at one point), a communication protocol was developed to ensure that messages delivered to different staff members reached a primary focal point within the company and that any responses to those messages were delivered with one “voice” back to the district. Additionally, learning to speak the language of the various disciplines involved was important for accurate communication and overall project success.QA/QC also was task-specific. Fieldwork QA/QC included the following:Reviewing field procedures with field crews and appropriate office staff before beginningDemonstrating examples of objective and subjective attribution collection to field crewsSupplying field crews with field packets containing planimetric/topographic mapping, any existing sanitary-sewer maps, and such information as smoke/dye testing data, which would aid in data-collection efforts Checking in at a control network of known National Geodetic Survey (NGS) monuments at the beginning and end of the workday to compare the coordinates generated by the RTD GPS survey to the published First-Order NGS horizontal surveys.Using SmartSurveyor, which contains built-in QA/QC “prompts,” including notification when the pipe material or diameter noted at each end of a pipe span does not match and when data are missing from a pipe or structure tableWoolpert developed a data management process for handling the data, from the daily downloading of field information to the development of the final GIS database. The goal of the management process was to minimize the number of times the data was handled, thereby helping reduce potential mistakes that could develop as a result of information being transferred between different software and staff.Field managers used several methods to perform QA/QC on collected data:Downloading daily data collection efforts from the field computers onto the Covington, KY, Woolpert office network in date-specific temporary project folders to protect the data from being corrupted or destroyedImporting the raw data into ArcView 3.2 and reviewing it against any provided hard-copy (paper) informationRunning the data through an automated review program (SQL and Avenue script formats for ArcView 3.2, which primarily checked for missing and conflicting attribution). QA/QC at the GIS Level
An analyst performs quality control by checking plots for accuracy.At approximately the same time that the field manager was performing the SQL/Avenue review, a copy of the raw data was forwarded to Woolpert’s GIS department in Dayton, OH, to perform a combination of manual (visual) and programmatic reviews of the field structure locations. Early in the project, GIS staff had to convert the data from shapefile format to coverage format; however, Woolpert developed a program that permitted GIS staff to maintain the ArcView format without requiring conversion. The field structure reviews were performed by comparing the structure placement against the 1999 orthophotos and digital structure data to help to evaluate whether the GPS or digitized (manually placed) field locations were within the permitted tolerance ranges. After GIS staff completed their reviews, reports were forwarded to the field manager. GIS deleted its copy of the sub-basin structure data to avoid the possibility of conflicting data sets in the future. The field managers took the GIS-reviewed reports and performed the following procedures:If GIS staff noted that a GPS point was more than 4.5 ft. away (for 100-scale mapping) or 6 ft. away (for 200-scale mapping) from the location shown on the orthophotos, the field manager would send a field crew back to the structure to take a revised shot.If GIS staff noted that there was not a GPS field location structure within a 4.5-ft. radius (100-scale mapping) or 6-ft. radius (200-scale mapping) of the existing structure data or there was not a digitized field location within 10 ft. of the structure, the field manager would compare the NKAPC/BCPC information to the orthophotos. If it appeared that the shot was not an actual structure, the field manager would dismiss the error. Otherwise, the field manager would send a field crew back to the location to review and, if necessary, take a shot and attribute the structure.At this stage, field managers also queried the data for connectivity and condition-assessment issues indicated by field crews and worked to resolve these problems.Once the field manager was satisfied with the results of the automated review, the orthophoto review, and connectivity- and condition-assessment reviews, the revised data was reforwarded to the GIS staff. At this point, GIS took ownership of the data set. Any digital revisions to the database were done strictly by GIS staff and required field manager authorization. GIS staff did the following among other additional QA/QC checks: They ran programs to check for voids in, correctness of, and presence of conflicting data.They ran a program to check if the open and closed systems were connected and to verify that they were flowing in the same direction.They ran a program to check for negative slopes in pipes. If the program found a negative slope, GIS staff would notify the field manager, who would have crews recheck the pipe. (Some pipes were verified to have reverse grades.) They captured screen shots of anything that did not make sense and sent the screen shot plus a report to the applicable field manager.In all, 132 different QA/QC programming checks were performed on the data.The District’s QA/QCThe district selected random structures for QA/QC. Instead of relying on a third-party vendor, the district worked in conjunction with Woolpert to perform the QA/QC process. In the field, a Woolpert field manager would measure random structures and immediately consult with a district representative, who had the collected inventory data loaded as planimetric maps in ArcPad, to see if the measurements were identical. All measurements were checked, except for rechecking of x and y data. Ninety-five percent of materials, depths, and so on had to be correct in the GIS. Pipe diameters had to be 90% correct. The district and Woolpert performed checks on 59 of the 256 sub-basins (23%) that were surveyed, which translated into checking nearly 1,700 of 55,000 points (3%). These points included catch basins, manholes, pipe inlets, pipe outlets, culverts, and ponds. Everything passed except for one sub-basin, which was resurveyed.Having Woolpert perform QA/QC on the data with a district staff member in the field eliminated the need to train a third party in project particulars, which saved time and money. In addition, the results were instantaneous; the project team knew immediately if a sub-basin passed or failed.The district also ran programs to ensure that features did not have duplicate numbers and that they had the correct names. The Project ResultThe project was submitted in five deliverables to the district based on watersheds. The district now has an extremely valuable tool that will serve as the foundation of its new stormwater management program. At least 10 people at the district initially will use the GIS in ArcView 3.2. In the short term, they will use this asset management tool to develop the illicit discharge detection and elimination program as required by the NPDES/KPDES permit;query the system about water quality;
assist in emerg
