Prologue: Even in the high-tech world of GIS there is still nothing exciting, glamorous or interesting about locating manholes on utility maps. Ergo, the focus of this opus is an attempt to give the reader an appreciation for the value of these seemingly mundane tasks in the grand scheme of things.
GIS Utility MappingOne of the things GIS does very well is to provide a spatial framework for projects that cover large areas and involve the interactions of many diverse groups with a common interest. Project Clean Water in southern California fits that description. The overall plan administers “waste discharge requirements for discharges of urban runoff from the municipal storm sewer system.”
Stormwater Pollution Monitoring
Initially, the individual political subdivisions were matched to the watersheds impacted by their respective runoff volumes and coordinated rainfall data. Rainfall data from 190 rain gaging stations was modeled with NOAA data to produce Isopluvial charts.
The GIS part of this project is essentially to map the effect of urban runoff in all major watersheds in the San Diego region. More than 20 government agencies are involved in this action. And a number of private firms were contracted to provide additional data and services.
Mapping Stormwater Conveyances
One of the requirements spelled out in the permit was “Stormwater Conveyance Mapping.” We were provided with a definition: a discernable, confined and discrete conveyance or system of conveyances, including but not limited to: natural channels, storm drains, gutters, ditches, pipes, tunnel, conduit, highways and roads under the jurisdiction of a city, town, borough, county, parish, or other public body.”
GIS staff members were tasked to examine and evaluate the data currently contained in the GIS to determine whether or not it met the mapping requirements of the permit. This group quickly realized a prodigious chore lay ahead. Historically, utility mapping had lagged behind other GIS layers in terms of accuracy and completeness. There were a variety of reasons for this, chief among them is the magnitude of the required effort. During the years the San Diego regional GIS was in its development and construction phase there were numerous attempts to find cost-effective methods of performing large scale utility mapping.
The history of getting utility locations in a GIS parallels the general evolution of the technology itself. The first step was to examine existing inventories. In our case many of the utilities were actually surveyed and appeared on various drawings. Many more were located on hanging “flat files” in the form of “as-constructed” drawings. But most were “warehoused” in a variety of disparate inventory systems of both the digital and hard copy variety.
Initially, whatever attributes were contained in the legacy data were converted into GIS data. Items that had survey values (x,y coordinates) included were easily registered. The hard copy maps were digitized and the digital and tabular data were geocoded* with mixed results. In most cases the source data was provided for inclusion by the putative users. At best, this method produced a patchwork of various utility themes. The need for some uniformity and a more comprehensive approach quickly surfaced as other parties became interested in the data.
Conventional surveying data collection techniques were more or less ruled out of the mix early on for a variety of reasons. Early attempts proved time-consuming and not cost-effective when applied to large areas. Although, existing survey networks did provide a sound framework for all of the base mapping activities.
Our first serious attempt to implement GPS on a broad scale GIS utility location was the Highway Emergency Callbox Project in 1995. We employed survey grade Ashtech LD12 and M12 dual-frequency receivers, a base station and a rover. Preplanning indicated seven minutes of occupation would yield horizontal positions accurate to about 10 meters. What we achieved after post-processing was closer to three meters, but we were only able to collect about 300 of the 700 call boxes spaced over 560 miles of roads and freeways with the available funding. Attributes were collected on data sheets and entered by hand into the database. The remaining boxes were collected with hand-held GPS units with Selective Availability active. The resulting positional accuracy was in the neighborhood of 50 meters** and required a great deal of “on-screen” editing to match other features.
In 1997, we moved onto an RTK (Real-Time Kinematic) GPS approach to collect stormwater conveyances over a limited area using SOKKIA Models GSR 2200 and GSR 2300RTK. This operation was conducted with both vehicle mounted and bipod held rovers and achieved near decimeter horizontal accuracy with some post-processing. TDS Husky Data Loggers allowed enhanced digital attribute entry in the field. But rover operation was effectively limited to a maximum of about 10 kilometers from the base station. The great improvement in precision and attribute collection with RTK was offset by increases in collection costs.
Light Poles, A Tale of Two ApproachesShortly thereafter DGPS (Differential Global Positioning Systems) utilizing a correction signal advanced to the real-time sub-meter level. Ray Miller, a maintenance engineer for the city of Escondido devised a method to mount a SOKKIA/Ashtech GIR1000 on a motorcycle to resolve a conflict the city was having with a local utility over the number of street lights on line.
The 5500 lights were mapped over a three-month period with all attribute data collected in a Husky Data Logger. Ray combined using DGPS (Coast Guard Correction Beacon) with a dedicated base station and some post-processing. Occupations of 30 seconds (about the time required to enter the attribute data) were used and produced solid sub-meter results. Ray used this same approach in 1999-2000 to collect the stormwater conveyances for the city of Escondido.
Our department had mapped the county street lights largely by using a form of geocoding. This method is often described as “event” mapping. In theory, geocoding is very similar to profiling and cross sectioning techniques in surveying. When the user has a “route,” such as a road center line, his position (event) is determined by a distance or a distance and offset from a fixed point. The resulting accuracy of the positioning is dependent on the quality of the base map and the method of measurement.
In an effort to improve the facility mapping accuracy in 1999, the county of San Diego contracted with Datria Systems of Aurora, Colo. The pilot project was to map a variety of features for the road maintenance division. Datria had recently completed several large projects including mapping the signage for the cities of Toronto, Canada; Austin, Texas; and Aurora, Colorado. They had also mapped large utility systems for a number of municipalities.
Datria utilized a voice recognition system called VoCarta to collect San Diego county data. The voice recognition software runs on a laptop computer. Data was collected with the Trimble Pro XR receiver. With the receiver mounted on a vehicle road and utility data, which included many of those geocoded street lights, was collected for a week in a designated area and then overlaid on the base map. Several “check ins” to known positions were made, and based on that data and some post-processing, we estimate most of the assets were mapped to a horizontal positional accuracy of ±1.5 foot. DGPS had arrived.
The PresentBy the summer of 2000, several local cities were well along with collecting their stormwater runoff facilities for GIS mapping. Chula Vista completed mapping over 3,000 assets in about three months using a SOKKIA AXIS3 and a Husky Data Logger.
In early 2001, the consulting firm of Nolte and Associates completed Phase I Flood Control Master Plan for the county of San Diego Flood Control District. Several miles of open channels and other drainage assets in the county were mapped in this effort. Maintenance staff completed the physical inventory of 12,000 various storm drain structures in an MS Access database.
The facilities in the Access database are located by the “Mile Post Method” from known road intersections. Our experience indicated the time required to collect these all with our SOKKIA AXIS3 DGPS system with the Omnistar correction could not meet the product delivery date specified. So we used a combination of GPS data collection using geocoding. The “pseudo route” geocoding methods produced match rates about 94 percent.** DOQQ overlays were used to analyze mismatches.
What are we going to do with all of this data once the mapping is completed? The plan is clearly to locate sources of pollution and potential sources of pollution and then effect remedies. We need to identify outfalls and potential sites that could improve water quality in urbanized areas of San Diego County. To help accomplish this, “Field Screening” stations will be selected based on the modeling of the collected data.
Most of the mission oriented data sets polled for this project will remain in the jurisdictions that developed them. The regional data sets will be used to help produce the reports necessary to comply with the conditions of the state discharge permit.
No, utility location surveys aren’t often glamorous or exciting. But they can and do make a major contribution to maintaining a healthy environment.