| AIRBORNE GPS: GPS Flies High in Midwest Flood Study
The Mississippi River Project demonstrates viability of large-area airborne GPS-controlled mapping.
By William G. Fry, P.E. Rivers are one of nature's greatest storytellers. By following a waterway across a map or aerial photograph, one can see how time and topography have shaped its course, how communities have grown up along its banks, and how people have come to depend on rivers for agriculture, industry, and recreation.
Unfortunately, rivers occasionally spread tales of death and destruction. In the spring and early summer of 1993, a sad story was repeated across the upper Midwest when the rain-swollen streams of the Mississippi and Missouri River systems caused the region's worst flood disaster in generations. For weeks, unrelenting floodwaters swept away houses, washed out bridges and railroad lines, and sent torrents of silt-laden water and debris through weakened levees across millions of acres of farmland. Along with paralyzed communities and billions of dollars in lost property, the floods left government officials with scores of questions about the factors that contributed to the disaster, and how to lessen the threat of other, potentially more dangerous sequels.
The search for answers led to one of the largest airborne GPS mapping projects ever, a year-long undertaking that produced a wealth of precision geographic data for over two million acres of floodplain along the Mississippi, Missouri, and Illinois rivers in parts of six states. In addition, the project demonstrated how emerging applications of differential and kinematic GPS can help overcome the time and logistical challenges of large-area mapping, making such ventures both technically and economically feasible.
The effort, known to those associated with it as the Mississippi River Project, was designed to provide detailed information about the region's topography and hydraulics to the Scientific Assessment and Strategy Team (SAST). Under the leadership of the U.S. Geological Survey, this multi-agency panel of scientists and engineers had been tasked by the White House Office of Science and Technology Policy to study alternatives for flood mitigation, recovery, and river basin management in the flood-impacted region. Because highly accurate digital elevation information was lacking, the Bethesda, Md.-based Spatial Data Institute (SDI), a nationwide consortium of six leading mapping and survey firms, was selected by USGS through a Corps of Engineers-administered contract to develop a photogrammetrically derived, airborne GPS controlled digital elevation models (DEMs) for seven separate study areas totaling 3,125 square miles.
"The existing data for the region was simply not suitable for the panel members to accurately model the river flows under flood conditions," says Dan Canfield, the SAST DEM manager at USGS' Mid-Continent Mapping Center. "We examined a number of new technologies that would produce the DEM with the desired accuracy level. Airborne GPS offered the highest potential savings in both time and effort."
In consultation with the government's project coordinators, SDI determined that airborne GPS was the most viable method for controlling the aerial photography so that DEMs could be produced with a majority of postings having an accuracy of 1.33 ft. RMSE. In addition, elevation data for over 800 miles of flow obstructing features was to be collected using kinematic GPS. The accuracy limitation for these points was .67 ft. RMSE. The horizontal spacing of the elevation posts was set at 5 meters.
"As we planned the project's scope and specifications, we had to continually weigh a number of factors to accommodate the diverse accuracy requirements of the SAST member agencies within our cost constraints," says Dennis Morgan of the U.S. Army Corps of Engineers' Technical Center of Expertise for Photogrammetric Mapping, Saint Louis District,which administered the contract. "SAST agencies were interested in everything from the rivers' hydraulic characteristics to how plants and animals were affected by the flood. We could have had an even tighter accuracy specification, but economics would have limited the amount of area we could study."
Project timing was particularly crucial. Because the DEMs were to help the scientists recreate the conditions of 1993, data needed to be gathered before land and structures were permanently altered. To meet SAST's deadlines, SDI would have to complete the DEMs within 12 months of contract award. "It's hard to imagine how long it would have taken to do a project of this size without the current technology," says Jack Dozzi,vice president of operations of SDI-team member Horizons Inc., and the consortium's project manager for the Mississippi River Project. "Even then, only a few combinations of firms could do a job of these proportions in the time required by SAST. We were fortunate to have an existing entity with the resources and expertise necessary to address the project's many technical and logistical challenges." Racing the Clock
Once the project's technical processes were determined, the study areas were photographed using airborne GPS under a separate government contract. The flights produced thousands of stereoscopic models that SDI needed to get the Mississippi River Project underway in May 1995. The team's first step was to establish a ground network to bridge the airborne GPS data. Using 108 government-supplied points plus others selected by SDI's photogrammetry firms, the field crews worked their way down the Missouri River from Falls City, Neb., to each of the designated study areas.
Among the locations were some of the most hard-hit areas from 1993, including Rulo, Neb.; Kansas City; Glasgow, Mo.; and St. Louis. The crews also followed the Mississippi River from its confluence with the Illinois near Jerseyville, Ill., to just below the mouth of the Ohio River. Despite the fact that study areas were separated by hundreds of miles, SDI's quality control checks with NGS monuments confirmed the network's accuracy to within a tenth of a foot.
For the 800 miles of kinematic surveys, trucks and all-terrain vehicles equipped with Trimble 4000 SSI GPS units helped surveyors measure critical cultural features, with reliable data acquired at speeds up to 55 miles per hour. Backpack units were used for locations inaccessible to the vehicles. "We were really pushing the technology to the limit," says Steve Hebert of SDI-member firm Vernon F. Meyer & Associates,which handled the project's field work. "Never before had kinematic GPS been used to cover so much territory. But we couldn't argue with the results. Less than one percent of the data points were inaccurate, and many of those discrepancies were easily resolved through computer checks."
As field data came in, SDI's mappers and photogrammetrists performed aerotriangulation calculations to control the aerial photographs. For the most part, comparisons of the kinematic profiles with those created through the photogrammetric process showed outstanding results. In fact, only a minimum number of ground points were necessary to adjust 3,050 stereoscopic models to the required accuracy standards. "Even though airborne GPS was the primary control from the outset, we had anticipated more extensive use of the ground network," notes Dozzi. "As it turned out, the ground data served mainly as a quality check for the airborne GPS data."
SDI's photogrammetrists used conventional and innovative compilation methods such as "Softcopy" to translate the photographic and GPS data into the DEMs. Even with the project's vast size and demanding accuracy requirements, the only major conflicts were due to hidden errors in one portion of the airborne GPS data, which complicated work in the project's southernmost section.
"Softcopy gave us a 'virtual pugging' capability that was instrumental in producing a high-accuracy DEM," says Warren McKean of SDI member firm Towill Inc. "We can use the best reference point to adjust the model, rather than simply measuring from the edges."
Weather proved to be a complicating factor throughout the project, however, as heavy rains and persistent mud sidelined SDI's field crews several times. These and other unforeseen difficulties led the Corps to grant a three-month contract extension for SDI to complete its work and finalize the deliverables. In a matter of weeks, the team provided the DEMs generated on a preset grid interval of 5 meters along with a network of random points supplemented with the elevation break points coded by type. In addition, SDI provided 1,000 sheets of 1:5000 scale check plots, complete documentation of all aerotriangulation calculations and ground control points, and 1:16,000 scale black and white diapositives.
"Delivering products of such demanding accuracy was certainly an accomplishment," says Larry Huston, CEO of SDI team member Bohannan-Huston Inc., "but even more important was the fact that we were able to avoid potential logistical and operational hurdles always lurking around a project this large and diverse." Laurels and Lessons
Indeed, the success of the Mississippi River Project opens a wealth of opportunities for both providers and users of digital mapping data, as well as many lessons on how future projects should be conducted. "The knowledge and experience gained from this project will help advance technologies used throughout the world to collect and use digital elevation data," says Canfield. "We've also expanded our knowledge of the capabilities of airborne GPS."
Another exciting outcome is the fact that the convergence of several technologies has created a clear economy of scale that benefits both mapping professionals and data users. Certainly, GPS and high-speed computers have given engineers, surveyors, and mappers powerful new tools to delineate the features of nature and the built environment. Projects need not be performed in piecemeal fashion and risk inconsistencies from differing accuracy requirements and changes in surface features over time. A vast amount of precision data can be gathered over a larger area in a relatively short period, then applied to different issues as needs dictate.
And the same spirit of collaboration among government agencies and mapping firms that helped make the Mississippi River Project possible can also enhance the value of its applications. Using GPS's inherent cost efficiencies, federal, state, and local governments should strive to take a regional approach to the challenges of land use, resource management, and disaster planning and recovery by sharing both the costs and benefits of large-area mapping.
Still, the project also illustrated that this cutting-edge technology is not without its limitations. As Dozzi notes, even with the extensive planning and preparation, the project provided a fair number of headaches. "Given that both study areas and deliverable file sizes are growing, there is a strong need to carefully coordinate the aerial photography planning with the aerotriangulation and compilation efforts, including provisions for contingencies and data verification," he says.
For example, the lack of a coordinated program between the airborne GPS surveys and SDI's contract resulted in a number of inconsistencies (e.g., split and crossed flight lines, gaps in airborne GPS data due to the distance from base stations). These problems required SDI to recalculate and reverify the data on several occasions. In another case, a ground receiver had lost lock with the satellites on one channel, leading to weeks of frustrating recalculations. "For airborne GPS, it is imperative that the base stations and targets be surveyed together using the same procedures and datum," says Dozzi. "While we always had the option of providing additional control points to supplement these gaps, we also spent a lot of time and effort to tie the separate surveys together."
Dozzi maintains, however, that the benefits of airborne GPS as a primary control for aerial photography far outweigh its challenges. "You get rock solid control at half the cost of conventional surveys," he says. "I think most people would agree that's a worthwhile payoff."
That dependability also proved to be essential to the project's success. Due to the lack of landmarks and significant topographic features in many locations, the traditional method of using photo IDs to locate many of the selected ground control points was ineffective. Still, Hebert stresses that ground network is very important. "In addition to serving as a quality control check, ground surveys can easily supplement missing airborne data," he says, citing a monument that had been used as an airborne GPS base receiver station. "When we came through with the ground surveys, we found that the monument had been destroyed. We got around the problem by performing additional extensive surveys with known NGS monuments in the area to complete the network." Better Data for Better Planning
Data from the Mississippi River Project is already being put to work by the engineers and scientists of SAST to rewrite the ongoing story of man's vulnerability to the mighty rivers of the Midwest. But the project's most important legacy will unfold in how the information benefits flood-prone communities in the Mississippi-Missouri-Illinois River system. Plans call for the project's base data to be imported into a variety of mapping programs that can generate detailed 3-D representations of rivers and adjoining topography for a variety of studies.
"The project was designed to satisfy a number of different research needs, making the range of applications virtually unlimited," Morgan says. "Used in concert with new or existing GIS programs, the region's flood-prone areas have a powerful tool for resource management, land-use planning, and hazard mitigation."
Information from the Mississippi River Project will soon be available to municipalities and the public in a variety of formats. It may be accessed by contacting USGS, the EROS Data Center (605/594-6511), or through SAST's home page on the Internet. (URL: http//edcww2.cr.usgs.gov/sast-home.html). SDI is also committed to assisting communities utilize the existing data and develop additional products such as orthophotography, planimetric mapping, and development of vector mapping for GIS applications.
While the data may not guarantee a happy ending for every flood-prone part of the upper Midwest, Morgan asserts that the Mississippi River Project provides a valuable tool that will help states and localities prepare for the floodwaters' inevitable return. "We're eager to see more people take advantage of data resources made possible by this project," he says. "It is a major investment in technology and research that will pay off for years to come." About the Author:
William G. Fry, P.E., is chief executive officer of Spatial Data Institute, headquartered in Bethesda, Md. SDI member firms include Aerial Data Service, Inc., Tulsa, Ok.; Bohannan-Huston Inc., Albuquerque, N.M.; Horizons, Inc., Rapid City, S.D.; Photo Science, Inc., Gaithersburg, Md.; Towill, Inc., San Francisco, Ca.; and Vernon F. Meyer and Associates, Inc., Sulphur, La. For more information, call 301/652-9477. Back |
