G11B-01 08:00h
Defining and Verifying Research Grade Airborne Laser Swath Mapping (ALSM) Observations
The first and primary goal of the National Science Foundation (NSF) supported Center for Airborne Laser Mapping (NCALM), operated jointly by the University of Florida and the University of California, Berkeley, is to make "research grade" ALSM data widely available at affordable cost to the national scientific community. Cost aside, researchers need to know what NCALM considers research grade data and how the quality of the data is verified, to be able to determine the likelihood that the data they receive will meet their project specific requirements. Given the current state of the technology it is reasonable to expect a well planned and executed survey to produce surface elevations with uncertainties less than 10 centimeters and horizontal uncertainties of a few decimeters. Various components of the total error are generally associated with the aircraft trajectory, aircraft orientation, or laser vectors. Aircraft trajectory error is dependent largely on the Global Positioning System (GPS) observations, aircraft orientation on Inertial Measurement Unit (IMU) observations, and laser vectors on the scanning and ranging instrumentation. In addition to the issue of the precision or accuracy of the coordinates of the surface points, consideration must also be given to the point-to-point spacing and voids in the coverage. The major sources of error produce distinct artifacts in the data set. For example, aircraft trajectory errors tend to change slowly as the satellite constellation geometry varies, producing slopes within swaths and offsets between swaths. Roll, pitch and yaw biases in the IMU observations tend to persist through whole flights, and created distinctive artifacts in the swath overlap areas. Errors in the zero-point and scale of the laser scanner cause the edges of swaths to turn up or down. Range walk errors cause offsets between bright and dark surfaces, causing paint stripes to float above the dark surfaces of roads. The three keys to producing research grade ALSM observations are calibration, calibration, calibration. In this paper we discuss our general calibrations procedures, give examples of project specific calibration procedures, and discuss the use of ground truth data to verify the accuracy of ALSM surface coordinates.
G11B-02 08:15h
Implementation of Waveform Digitization In A Small Footprint, Airborne Lidar Topographic Mapping System
Accurate mapping is critical for applications ranging from geodesy, geomorphology, and forestry to urban planning and natural hazards monitoring. While airborne lidar (Light Detection and Ranging) has had a revolutionary impact on three-dimensional imaging of the earth's surface, there is great potential for developing new capability by replacing the laser range and backscatter intensity information recorded by conventional lidar systems with full waveform digitization. The University of Texas at Austin (UT) owns and operates an Optech ALTM 1225, a small footprint lidar system. In response to an initiative from UT, Optech has developed a module which samples the analog waveform of a laser pulse and converts these samples into digital measurements. The waveform digitizer specifications include a 1-nanosecond sampling interval, 440 samples per return laser waveform (approximately 65 meters of vertical extent), and waveform digitization at the 25kHz laser pulse repetition rate. The digitizer also records the initial T0 pulse that starts the timing cycle. The digitizer unit is an independent module supported by a Pentium-4 computer, two hard drives, and a high-speed data recording system. The digitizer is integrated into the ALTM system so that both full waveform and the conventional first and last returns are recorded for each transmitted laser pulse. This unique capability allows for conventional lidar data to be directly compared to the full waveform. We present examples of full waveform lidar mapping over different environments and discuss future applications.
G11B-03 INVITED 08:30h
Morphometric differences in debris flow and mixed flow fans in eastern Death Valley, CA
Geomorphological features are best examined through direct measurement and parameterization of accurate topographic data. Fine-scale data are therefore required to produce a complete set of elevation data. Airborne Laser Swath Mapping (ALSM) data provide high-resolution data over large spatially continuous areas. The National Center for Advanced Laser Mapping (NCALM) collected ALSM data for an area along the eastern side of Death Valley extending from slightly north of Badwater to Mormon Point. The raw ALSM data were post-processed and delivered by NCALM in one-meter grid nodes that we converted to one-meter raster data sets. ALSM data are used to assess variations in the dimensions of surficial features found in 32 alluvial fans (21 debris flow and 11 mixed flow fans). Planimetric curvature of the fan surfaces is used to develop a topographic signature to distinguish debris flow from mixed flow fans. These two groups of fans are identified from field analysis of near vertical exposures along channels as well as surficial exposures at proximal, medial, and distal fan locations. One group of fans exhibited debris flow characteristics (DF), while the second group contained a mixture of fluid and debris flows (MF). Local planimetric curvature of the alluvial fan surfaces was derived from the one-meter DEM. The local curvature data were reclassified into concave and convex features. This sequence corresponds to two broad classes of fan features: channels and interfluves. Thirty random points were generated inside each fan polygon. The length of the nearest concave-convex (channel-interfluve) couplet was measured at each point and the percentage of convex and concave pixels in a 10m box centered on the random point was also recorded. Plots and statistical analyses of the data show clear indication that local planimetric curvature can be used as a topographic signature to distinguish between the varying formative processes in alluvial fans. Significant differences in the lengths of the channel and interfluve couplets and the percent pixels per unit area have been identified between the two fan groups. In general, shorter distances were found in the mixed flow fans. This finding can be attributed to primary and secondary erosional processes leading to a higher degree of dissection in the MF fan surfaces than was identified in the DF fans. The clast-rich deposits of the DF fan are more resistant to secondary erosional processes and overtopping of the levees likely leads to filling of older incised debris flow channels.
G11B-04 08:45h
Classifying and Analyzing Spatial Patterns of Surface Features in Debris Flow Fans using Curvature Parameters
A modified method of calculating profile and planform curvature is used to identify homogenous slope units (surface features, such as debris flow levees, lobes, or plugs) within alluvial fans. High-resolution DEMs were generated from Airborne Laser Swath Mapping techniques. One-meter grid nodes processed by the National Center for Airborne Laser Mapping were converted to raster datasets. ALSM data are used to assess the spatial patterns in surficial features identified in 21 debris flow fans. The debris flow fans are identified from field analysis of near vertical exposures along channels as well as surficial exposures at proximal, medial, and distal fan locations. Our observations indicate the debris flow fans are generated from clast-rich debris flow deposits. They exhibit several characteristic debris flow features that include channels, levees, interfluves, lobes, and plugs. Initial analyses indicate the length-scales appropriate for quantifying surface curvature are different for planform and profile curvature. The longitudinal profile of the fan surface contains a series of lobe-like features that have a significantly different length-scale than the channels and interfluves identified in the planimetric profile of the fans. The proper length scales are determined by minimizing the frequency of rectilinear curvature values. The statistical properties of the curvature values are then used to quantify the anisotropic variability of alluvial fan morphometry. The spatial variability is compared with the fan area, drainage basin area, fan slope and drainage basin geology to compare the variability within the debris flow fans.
G11B-05 INVITED 09:00h
Using airborne laser swath mapping to quantify sediment production and transport processes
Since Gilbert, Agassiz, and other pioneers of geomorphology viewed the shape of landscapes as a reflection of surficial processes, earth scientists have embraced the process-form linkage as an important tool in the quantification of sediment production and transport processes. Testing the rich pool of geomorphic models has been difficult due to the lack of topographic data at the process scale. Emerging datasets acquired from airborne laser swath mapping (ALSM) are enabling earth scientists to marry state-of-the-art theoretical models with high-resolution topographic information. Here, we describe examples of how ALSM data can be used for: 1) geomorphic model testing and calibration, 2) prediction of erosion rates, and 3) deciphering the chronology of mass movement over millennial timescales. Models describing the movement of soil on hillslopes generate distinctive morphologic predictions. Analyses of ALSM datasets illustrate that soil-mantled hillslopes in the Oregon Coast Range (OCR) become increasingly convex with steepness, consistent with proposed nonlinear, slope-dependent transport models. Given assumptions regarding background erosion rates or field-based measurements, such transport models can be calibrated and used to predict the spatial distribution of erosion rates using ALSM datasets. Our analysis of post-fire erosion on OCR hillslopes demonstrates that subtle variations in topographic form are important for determining the spatial pattern of soil stripping and bedrock emergence. The seemingly jumbled and chaotic topographic form of large landslide complexes contains a wealth of information regarding the history and mechanics of slide movement. Statistical analyses of ALSM datasets for slide-dominated terrain can identify unstable areas and distinguish the chronology and style of internal deformation features. These projects emphasize the need to both carefully consider how we estimate standard topographic properties (such as slope and curvature) and explore new methods for quantifying the form of the earth's surface.
G11B-06 09:15h
The use of Airborne Laser Swath Mapping Data in Watershed Analysis to Guide Restoration Priorities: the Napa River Watershed Study
A necessary step in the management and restoration of ecosystem functions in a watershed is to quantify the linkages between landuse practices and channel habitat. With the advent of widely available digital elevation data, increasing numerical skills, and increasing insight about physical and ecological processes, models are being built that explore these linkages. Development and application of these models, however, strongly depends on the resolution of the toporgraphic data. Critical details of hillslope topography are not captured by the highest resolution USGS data (10 m), and, impotantly, channel banks are not a topographic feature in the digital elevation model. Instead the position of the main channels are delineated from hand mapped "blue lines" of USGS topographic quadrangle and then the smaller channels are typically estimated to occur at grid cells receiving drainage area exceeding some critical amount. High-resolution airborne laser swath mapping data (ALSM)) captures much of the finer scale topography, including that of channel banks, but introduces new challenges in both accuracy determination and GIS applications. As part of work to guide development of a Total Maximum Daily Load Analysis of the Napa Watershed, ALSM data were acquired for the entire 1100 km2 Napa River (California) watershed by the University of Florida. The filtered bare earth data set exceeded 1 billion points and gave an average data density of 1.5 m with areas in grasslands dropping below 1 m. Many GIS tools exist to analyze digital elevation data, but we have found many of them inadequate for the large, detailed data set. A central goal of the data acquisition was to create an accurate delineation of the channel network and to estimate channel morphology and grain size to help define the extent of available habitat for salmon. Over 400 on-channel dams and 4000 channel road crossings were identified, which create topographic barriers of significance to modeling and watershed management efforts. Each barrier had to be eliminated to pass drainage area downstream. Furthermore, standard GIS tools treat channels as single cells bearing some drainage area in excess of a critical amount. Such tools are inappropriate where the grid cell is much smaller than the channel width and the channel banks are topographically distinct. A new method of channel delineation needs to be developed. Nonetheless, the single cell approach allowed us to estimate extent of habitat using a channel classification based on a slope and an estimate of median grain size from slope and hydraulic geometry relationships. Mapping and modeling revealed potential salmonid habitat is constrained to the mainstem Napa River and the mainstem's of its major tributaries. Reservoirs, which reduce this habitat, have been built at a rate of about 20 per year for 50 years, leading to increasing conflict between habitat needs of fish and the demand for water. The 1m data of the entire Napa watershed are available at http://calm.geo.berkeley.edu/website/NapaDwnldRaw/ as part of the NSF-supported Center for Airborne Laser Altimerty (NCALM).
http://calm.geo.berkeley.edu/website/NapaDwnldRaw/
G11B-07 09:30h
Evaluation of LiDAR Imagery as a Tool for Mapping the Northern San Andreas Fault in Heavily Forested Areas of Mendocino and Sonoma Counties, California
We are mapping in detail active traces of the San Andreas Fault in Mendocino and Sonoma Counties in northern California, using recently acquired airborne LiDAR (also known as ALSM) data. The LiDAR data set provides a powerful new tool for mapping geomorphic features related to the San Andreas Fault because it can be used to produce high-resolution images of the ground surfaces beneath the forest canopy along the 70-km-long section of the fault zone encompassed by the data. Our effort represents the first use of LiDAR data to map active fault traces in a densely vegetated region along the San Andreas Fault. We are using shaded relief images generated from bare-earth DEMs to conduct detailed mapping of fault-related geomorphic features (e.g. scarps, offset streams, linear valleys, shutter ridges, and sag ponds) between Fort Ross and Point Arena. Initially, we map fault traces digitally, on-screen, based only on the geomorphology interpreted from LiDAR images. We then conduct field reconnaissance using the initial computer-based maps in order to verify and further refine our mapping. We found that field reconnaissance is of utmost importance in producing an accurate and detailed map of fault traces. Many lineaments identified as faults from the on-screen images were determined in the field to be old logging roads or other features unrelated to faulting. Also, in areas where the resolution of LiDAR data is poor, field reconnaissance, coupled with topographic maps and aerial photographs, permits a more accurate location of fault-related geomorphic features. LiDAR images are extremely valuable as a base for field mapping in this heavily forested area, and the use of LiDAR is far superior to traditional mapping techniques relying only on aerial photography and 7.5 minute USGS quadrangle topographic maps. Comparison with earlier mapping of the northern San Andreas fault (Brown and Wolfe, 1972) shows that in some areas the LiDAR data allow a correction of the fault trace location of up to several hundred meters. To date we have field checked approximately 24 km of the 70-km-long section of the fault for which LiDAR data is available. The remaining 46 km will be field checked in 2005. The result will be a much more accurate map of the active traces of the northern San Andreas Fault, which will be of great use for future fault studies.
http://quake.usgs.gov/research/geology/lidar/
G11B-08 INVITED 09:45h
Relative geomorphic surface chronology measured by high-resolution LIDAR topography
LIDAR is a potential tool to establish the relative chronology of geomorphic surfaces over a broad spatial area, with benefit to Quaternary research. Diffusive processes, such as creep, progressively smooth landscapes through time. Detailed elevation data can quantify this smoothing over different topographic wavelengths, and chronologically equivalent surfaces could thus be identified and mapped. Unlike spectral imaging, which relies on weathering changes, LIDAR chronosurfaces could be identified over areas of differing lithology. However, LIDAR chronosurfaces may be adversely impacted by large grain size differences that influence diffusion rates. To accomplish the goal of quantifying surface roughness over time requires first that the relative error within a LIDAR swath and between LIDAR swathes be better understood to determine the minimum resolvable surface feature amplitude and wavelength. One approach to measuring roughness is to sample data collected in the same swath, which minimizes error in instrument location but has the disadvantage of a fixed orientation relative to topographic features; the actual amplitude would need to be calculated in an oblique transect. The other approach is to sample data along an optimal azimuth, which could result in a mismatch in elevations between individual swaths of LIDAR data. Combining LIDAR chronosurface maps with the traditional tools of fieldwork and numerical dating would provide an age model over a broader area than that mapped in the field, along with an additional chronology check. Absolute dating methods (e.g. $^{14}$C, cosmogenic nuclides) are expensive; the chronosurface map provides a framework to develop the most parsimonious dating strategy. Sparser initial sampling for absolute dates reserves resources to more densely sample areas with ambiguous chronosurfaces and to revisit areas with unexpected numerical dates.