In the past 80 years, coastal Louisiana has lost 1,883 square miles of wetlands, an area roughly equal to the state of Delaware. According to a 2011 report by the U.S. Geological Survey (USGS), the average rate of swamp and marsh loss over this period has been about the size of an American football field every hour.
The coastal wetlands are an invaluable natural resource in the region supporting the livelihoods and recreational activities of a diverse population. Just as importantly, however, these marshes serve as a barrier that shields many Louisiana towns and cities from inundation during storms and hurricanes. Developing strategies to restore and protect this complex ecosystem has become a priority at both the state and federal level.
In early 2011, Woolpert conducted a 5,400-square-mile LiDAR mapping project under contract to the USGS as a vital step in delineating the wetlands and understanding the dynamic processes at work there. From a technical perspective, the project highlighted the challenges of mapping an enormous coastal zone where the changing tides influence water levels for miles inland. The effort also underscored the ability of LiDAR to accurately capture both elevation and land-cover information in a single dataset.
USGS awarded the Louisiana Region 1 and 2 LiDAR task order in September 2010 under Woolpert’s Geospatial Products and Services Contract. Each of the two regions covered approximately 2,700 square miles. Project requirements called for collection of LiDAR points at a nominal pulse spacing of 2 meters and a vertical accuracy of 12.5 cm RMSE. The LiDAR point cloud data was provided in LAS format classified into the Default (Class 1), Ground (Class 2), Noise (Class 7), Water (Class 9) and Ignored Ground (Class 10) classifications per USGS requirements.
Based in Dayton, Ohio, Woolpert is a national geospatial, architecture and engineering firm that has completed numerous regional LiDAR mapping projects for federal, state and county clients. The firm operates six aircraft and has several laser scanners. With such demanding specifications required for a large project area, Woolpert deployed two Leica ALS50-II airborne laser scanners to Louisiana aboard Cessna 404 aircraft.
The firm selected the Leica scanner for this project for two main reasons. First, the power of the unit enabled the aircraft to fly at a higher altitude without compromising the 2-meter post spacing. This allowed the firm to cover more terrain in a single flight line, improving overall project economics. Second, Woolpert has found the signal-to-noise ratio of the ALS50 LiDAR lends itself to accurate validation of features in areas of dense grass and low brush.
“The Leica ALS systems are extremely stable sensors that have less noise at the ground level compared to other LiDAR systems in the first three returns” says Chris Ogier, federal civilian practice leader for Woolpert. “That was a benefit for us in this project where one of our biggest challenges was determining where the water ended and land began.”
The two aircraft and sensors were deployed onsite in Louisiana during early January 2011. The LiDAR data was acquired at a flight altitude of 7,800 feet above ground level at a speed of 140 knots. The two project areas required 282 flights for a total of 8,772 flight line miles.
“The main challenge in this project proved to be the daily flight planning around the short acquisition windows caused by the tides to meet a maximum water level of 6 inches above mean low water requirement,” explains Woolpert Project Manager John Gerhard.
As is often required for wetlands mapping in intertidal zones, USGS requested LiDAR acquisition at the mean low water level so that maximum ground surface was exposed and mapped. The complexity of the topography and maze of tributaries through the uneven Louisiana coast made this difficult because the mean low water mark is not a straight line running across the wetlands. Instead, low tide occurs at different times of day throughout the marshes depending on their distance inland and terrain elevation.
“We couldn’t fly the entire area based on the same tide gauge and time of day,” says Gerhard. “We had to strategize based on multiple tide gauges because water levels varied throughout the project areas.”
Woolpert consulted Internet-accessible data from tide gauges operated by NOAA. The project team examined the location of the tide gauges and divided the entire project area into blocks where tide levels were relatively consistent. The best times of day, based on the mean low water requirement, were determined for each block. Early mornings, when fog forms over the coast, were usually deleted from the list. On a seven- to 10-day cycle, the flight crews studied weather forecasts for each block, again eliminating times when rain, clouds or haze were expected.
“Our crews had a spreadsheet created for each block to identify acquisition windows when the mean low water requirement coincided with good flying conditions,” says Gerhard. “Most windows were very short, and we had to have Plans B and C ready in case weather didn’t cooperate with Plan A.”
The primary, secondary and tertiary flight plans were entered each day into the Leica IPAS automated flight management and navigation system integrated with the ALS50. This guided the pilot over the correct flight lines and triggered scanner operation during the airborne operations. The integrated IPAS system also included the airborne GPS/inertial measurement unit (IMU) that collected precise aircraft position and attitude data through LiDAR point capture.
Collection of accurate ground control and check points was considered a critical aspect of the project for LiDAR processing and QA/QC. In each of the two project regions, USGS required a minimum of 60 check points. This translates to 20 each in the three main land cover classifications--bare-earth/low grass, high grass/crops and brush lands/low trees--found in the project regions. Woolpert field crews collected the points using real-time kinematic GPS with rover units in the field and a GPS base station established on a survey monument.
“As agreed upon with USGS, the points were not to be located in the swamps where the only access was by a boat or an airboat. To save effort in the field, the points were to be accessible by road,” says Gerhard.
The survey crews reviewed Google Earth imagery to determine the best access route to the control and check points locations in the three land cover classifications prior to mobilizing at the project site.
To the extent allowed by the terrain and accessibility, the control and check points were distributed so that they were spaced at intervals of at least 10 percent of the diagonal distance across the project areas, and at least 25 percent of the points were located in each quadrant of the project areas. The control points were located and surveyed on unobstructed, relatively flat, light-colored, hard, uniform surfaces. The control points were incorporated into the post-processing of the LiDAR data. The check points were also ground truth observations distributed within the land cover classifications and were used to independently verify the absolute and relative accuracy of the LiDAR data.
Processing the LiDAR data was carried out in two phases. The first was classification of the bare earth points using TerraScan software. Woolpert applied a series of proprietary filters that used the ground control data to assist with differentiating the LiDAR returns from the earth’s surface and overlying vegetation. These automated filters completed approximately 80 percent of the work in classifying the bare earth points. The remainder was carried out by manually comparing the LiDAR point cloud data with available imagery sources.
“From there, we had to take a different approach to creating breaklines for the water features, because the coastal zone is so flat, and this area has a very complex shoreline with a large expanse of wetlands,” says Ogier.
An automated procedure was developed to generate breaklines of water features, including rivers, streams and ponds. Next, they used LiDAR-grammetry, in which the 3D LiDAR points were viewed in stereo and with a two-dimensional orthophoto draped on top. The photogrammetric technicians manually examined the LiDAR points and compared them with the imagery to adjust the breaklines as needed to correctly delineate the water features.
This process successfully applied breaklines to water bodies including rivers wider than 100 feet, and lakes and ponds larger than two acres. The breaklines defined the water bodies as closed polygons so that hydrologic flattening could be applied. In this technique, the LiDAR technicians reclassified the LiDAR points within each polygon, rendering a gradient surface to the rivers and streams and a smooth surface for the ponds when the DEM was produced.
After wrapping up the flying portion of the Louisiana project in April 2011, Woolpert completed the processing and delivered the final products in October. The reaction to the accuracy and quality of the datasets has been very positive. The bare-earth DEMs and other products have been distributed by USGS to multiple interested parties working on hydrologic modeling and environmental monitoring in the area.
For more information about Woolpert, visit www.woolpert.com. Additional information about Leica ALS airborne laser scanners, visit www.leica-geosystems.us.