Delta Lidar

In a fertile delta in northern California, water resource officials used aerial lidar to map levees, creating a host of usable data in many forms.
By Joel Dudas, P.E.

The levees of the Sacramento-San Joaquin Delta protect farming communities and critical infrastructure against ever-present flood threats. Like levees everywhere, these must hold back runoff floodwaters produced by large rainfall events. However, the delta is a tidal estuary, and much of the land behind delta levees lies below sea level due to more than a century of subsidence of drained organic soils. Consequently, the delta levees hold back water all the time, not just during major storms. With the hydrodynamics involved, whether individual levees fail often becomes a matter of inches. Since reclamation began in the late 19th century, levee breaches have occurred more than 150 times, with ever-growing consequences as more is invested in and dependent on the delta.

It becomes obvious that with earthquake dangers factored in, delta levees face a wider variety of risks than most levee systems. And the delta is a unique and interesting estuary for another critical reason. The flood-control levees of the delta double as a major water supply conveyance system, serving agriculture in the valley and more than 23 million people in urban areas throughout California.

In recent years, the California Department of Water Resources (DWR) has undertaken a comprehensive study to assess risk levels, consequences of failures, and risk mitigation for the levees. The study results support comparative decision-making for how and where to fund projects. Engineers are devising new levee designs. High-accuracy baselines for geometric assessment of the levee system and high-resolution subsidence monitoring are being established. The baselines will be compared against future studies to objectively assess bond-funding progress towards meeting improvement and mitigation requirements.

Elevation Surveys Crucial

For all this technical work, elevation surveys are a crucial data requirement. But in the delta, accurate survey data cannot be taken for granted, as it ranks as one of the more dynamic non-volcanic areas on Earth. When levee construction projects have placed a foot of fill to raise a levee, it is not unusual for that levee (built on peat soils) to settle back to the original elevation within a couple of years. Subsidence of up to thirty feet in the last century is an ongoing problem.
Plus, in some places, half a foot is the difference between a typical mean high water stage and a 100-year flood stage, and levee freeboard is frequently less than a foot above the 100-year stage. Data accuracy needs are therefore on the order of inches, not feet. So for objective, comparative analysis, a single snapshot of high-accuracy data that covers the entire delta is required.

Unfortunately, the delta suffered from the lack of anything like a comprehensive, high-accuracy survey. Delta-wide digital elevation model (DEM) surfaces included the USGS National Elevation Dataset and a couple of interferometric synthetic aperture radar (IFSAR) surveys, but these sources not only lacked the accuracy needed for hydrodynamic simulations, but in general characterized the levee system very poorly. As for field surveys, the situation was a hodgepodge. Between various public agencies and local flood districts, some areas had been surveyed well, but many areas were not. Even the well-surveyed areas were from many vintages and often obsolete. In some cases, levees had not been surveyed at all since the 1980s.

After passage of flood project bonds, it quickly became clear that a solution to the elevation data problem was needed. Since the area is several hundred thousand acres in size, the costs loomed potentially large. After reviewing available options, DWR decided to go with a lidar acquisition. Lidar was chosen because if offered the best combination of accuracy and cost-effectiveness for such a large area. Under the right circumstances, with the right equipment, and maybe some good luck, lidar can produce raw data that supports development of 1-foot contours.

For the project, DWR used a risk assessment contract with URS Corporation, who awarded the production work to Fugro-EarthData and a separate independent QA/QC aspect to Spectrum Mapping. Fugro-EarthData in turn hired Airborne 1 Corporation to conduct the aerial survey work before handoff back to Fugro-EarthData for post-processing and delivery.

DWR’s experience with lidar projects indicated it would not be feasible to collect points of sufficient accuracy to generate 1-foot contours with significant vegetation present during the survey flight. In the delta, trees begin leafing out, field weeds sprout, and widespread agriculture gets going in March, so March 7 was set as the allowable flight acquisition window closing date. This is a critical part of any lidar project specification, indeed, probably the most important variable a client can request to ensure the reliability of the entire survey. Flying during leaf-off conditions can greatly expand the amount of the project footprint that meets specification.

Other aspects of the project specifications and work plan were used to improve the deliverable quality. Since the edges of scan swaths produce the greatest vertical error, the allowable maximum scan angle was halved. This slightly compromised vegetation penetration and increased the flight costs by requiring more flight lines. But DWR felt the cost tradeoff was worthwhile given the project’s importance, and that by flying in winter, the vegetation penetration benefit was outweighed. In addition, USGS came to DWR’s assistance by upclocking the GPS broadcast from 1 Hz to 5 Hz at control stations during the flights, which reduced errors induced by interpolating GPS solutions from 1 Hz data.

One option DWR did not choose was acquiring the data only at low tide. This method is often proposed in tidal areas to maximize the number of laser returns from the intertidal levee waterside slopes. While it would have been ideal to do this, acquisition would have become prohibitively expensive.

Hit It Lucky

The first survey flights occurred in January and February 2007. As fate would have it, this was a fortuitous period to acquire lidar data. The winter was dry to that point, unlike the wet conditions experienced a year prior, and it was also very cold. The contractors acquired data during cold, stable air masses, reducing atmospheric errors. And though the delta can get socked in with fog much of the winter, those same air masses kept things dry and fogless. In a little more than three weeks, acquisition of the million-acre project footprint was complete.

Unfortunately, during processing, it became evident that some areas had experienced problems. Initial fixes did not succeed, and it was decided to re-fly about nine percent of the area. To produce the deliverables in a timely manner, initial re-flights occurred outside the prescribed acquisition window, during June. But independent QA/QC revealed that the June data didn’t comply with project specifications because of the leaf-on conditions, so another flight was needed. The area was flown yet again in early March 2008.

The data were re-delivered in fall 2008. A few minor issues remain with the overall dataset, including integration of the data from the original flight and the re-flight, but none of these should require any more re-flights. Final revision work, however, is currently suspended, pending issues with the State of California’s financial situation. When work resumes, it is hoped the final product will be ready within a few weeks.

Interim products are available in the public domain, as will be the final product. Point data are available as raw data, bare earth products, and for first returns. One-foot contours and processed ESRI DEM grids were also produced. Intensity imagery to assist interpretation of site ground conditions is also part of the deliverables. These data are available in a variety of digital formats to support use in various software platforms.

Many Uses for the Data

For DWR, our first use of the data is to support a variety of flood management data needs. Hydrodynamic modelers are using the data to satisfy model needs for island elevations to conduct delta island post-levee breach scenarios. The total volume of water that will fill an island in response to a flood is now more precisely known. DWR is currently undertaking a much more technically robust assessment of the delta levees’ FEMA levee geometry standard compliance and is also assessing costs for upgrades to certain standards for each of the delta islands. A large, DWR-funded floodplain re-mapping effort throughout the entire Central Valley also uses the delta lidar data.

The data are also supporting basic site-specific project needs for topo data for a variety of levee, subsidence reversal, and ecosystem restoration projects. With NASA and USGS, DWR is proposing fusing the lidar with plane-based permanent scatterer-InSAR techniques to develop a method for short-turnaround, high-resolution, high-accuracy delta-wide elevation products. These would prove useful for studying seasonal subsidence and subsidence reversal patterns, determining whether the “spongy” organic soil delta islands float up and down with the tides, and monitoring levee deformation during prolonged high-water events.

Beyond DWR, the data already have been used in many different ways by a wide variety of agencies, consulting firms, universities, and non-profits. Most prominently, the data are considered a baseline for future subsidence and subsidence reversal assessments. Reclamation district engineers have used the data for levee project planning and are looking at re-routing drainage features on seemingly flat delta islands. More broadly, a multi-agency group looking to restore tens of thousands of acres of ecosystem is using the data to satisfy the fundamental role elevation plays in habitat planning at the regional scale. UC Davis researchers have even been trying to count each individual tree in the delta using these data, with an eye toward modeling effects of vegetation removal from levee slopes on water temperatures.

For GIS professionals working in the delta, if a project or study needs a DEM, it will likely use the delta lidar data. The data in all its delivered formats are readily usable in GIS and support the integrative nature of GIS analysis as well as any other data types. Delta maps and spatial analysis products that use these data should become abundant. However, it is critical to understand that these are not perfect for every application or in every part of the survey area. Once the product is finalized, extensive documentation for the project will be compiled. These documents will include supplemental surveys and accuracy statistics for 18 land-cover types commonly found throughout the region. The documentation is vital to understanding when and where the data should be used.

Looking to the future, DWR has a few ideas in mind for how to do more with lidar. To better standardize the products coming from this growing and exciting industry, we are researching new types of lidar project specifications. For cited DEM accuracy, to avoid the problems inherent in the apples-and-oranges comparisons that arise from comparing spot elevations to interpolated grid cell values, using ground-based lidar allows comparing grids to grids. DWR is exploring better quantitative methods of specifying bare earth artifact removal, also using ground-based lidar as the basis. This work is occurring in consultation with USGS and University of California experts.

We hope the industry and ASPRS will be receptive to the ideas that come from this work, and DWR intends to communicate findings to these entities. DWR also intends to resurvey the delta on an ongoing basis. Because the delta is a place of dynamic elevation, in a sense, elevation surveys go out of date as soon as they are finished. Using a glass-half-empty-or-full analogy, this can make the work frustrating, but it also makes it exciting.
Joel Dudas is a California-licensed engineer with the California Department of Water Resources. He focuses on GIS, engineering, lidar, and home brewing.

*This article first appeared in the California Surveyor, the quarterly publication of the California Land Surveyors Association, Spring 2009, Issue #158.

» Back to our Aerial Mapping March 2010 Issue

Website design and hosting provided by 270net Technologies in Frederick, Maryland.