Laserfiche WebLink
<br />I <br /> <br />1501 <br /> <br />I <br /> <br />EV ALVA nON OF LIDAR AND PHOTOGRAMMETRY FOR MONITORING VOLUME <br />CHANGES IN RIPARIAN RESOURCES WITIDN THE GRAND CANYONt ARIWNA <br /> <br />I <br /> <br />P. A. Dan., S. N. Mlelz, K. A. Koh~ M. R. Roslek, F. M. GonZales <br />U.S. Geological Survey <br />Flagstaff. Arizona 8600 I. USA <br /><pdavis, srrietz, kkohl, mrosick,fmgonz)@usgs.gov <br /> <br />M. F. Manooe, J. E. Haze~ M. A. KapUmld <br />Departmenl of Geology <br />Northern Arizona University <br />Flagstaff, Arizona 860 II. USA <br />(Mark.Manone.Joseph.Hazel,Mall. Kaplinski)@nau.edu <br /> <br />ABSTRACf <br /> <br />Elevation data from high-resolution photograrnmetry and different resolution and replicate collections of LlDAR were <br />compared with ground survey elevations of bare- and vegetated-sediment deposits within the Colorado ecosystem to <br />determine their ability to map ground topography at an accuracy of =20 em and to estimate sediment volume at an <br />accuracy of =3%. We found high-resolution L1DAR and photogrammctry to meet this vertical accuracy requirement <br />on bare-sand surfaces, but not on vegetated-sand surfaces. Photogrammetry and LIDAR second-return data produced <br />the best venica\ accuracies on vegetated-sand surfaces, but accuracies were only about 35 cm. Photogrammetry <br />cons4stently produced the highest volumetric accuracies on bare ~3%) and vegetated ground (4-9%). Overall, <br />photogrammetry provides more accurate monitoring data than LlDAR, although more careful photogrammetric <br />processing is necessary to attain monitoring accuracy requirements within vegetation. <br /> <br />KEYWORDS: LIDAR, Photogrammetry, Accuracy, Precision, Ecosystem, Monitoring <br /> <br />INTRODUCTION <br /> <br />The Grand Canyon Monitoring and Research Center (GCMRC) of the U.S. Geological Survey studies the effects of <br />Glen Canyon dam water release on the Colorado River ecosystem between Lake Powell to Lake Mead, Arizona. One <br />GCMRC program studies the effects of water release on terrestrial sediment deposits and routinely monitors the area, <br />topography, and volume change relative to dam flow conditions to determine their relations to dam water release and to <br />develop wateHelease protocols to restore, or at least to maintain, the ecosystem (Hazel et aI., 1996). Collection of <br />these data in this canyon system currently involves time-consuming (expensive) ground surveys. A recent increase in <br />the number of study sites prompted an examination of alternative methods for data collection that might provide <br />acceptable data over a larger region of the river corridor, in a shorter time frame, and possibly at a cost comparable to <br />current ground surveys. For GCMRC monitoring purposes an acceptable range for vertical accuracy and precision is <br />\5-20 em (Schmidt et at, 1999). Sedlment volume estimates derived from these topographic data need to be accurate <br />to within 3% to meet current ground accuracy (Beus et aI., L 992). <br />The most commonly used airborne technologies for ilpographic mapping are photogrammetry and LIDAR and <br />several studies have evaluated their vertical accuracies. LIDAR assessments on large, uniform surfaces found vertical <br />accuracies 10 be 15-20 em (Krabill el al., 1995. 2000). Gomez Pereira and Wicherson (1999) found the vertical <br />accuracy of LlDAR data to be about 14 cm on flat, homogenous surfaces, whereas vertical accuracy of (1 :4,000-scale) <br />photogrammetric data was about 25 cm. Blank (2000) found this same photogrammetric accuracy for a small area <br />within the Colorado River ecosystem. Several studies have found LIDAR elevations to be offset from true ground <br />elevations. Kraus and Pfeifer (1998) found LlDAR data to have a +20 cm vertical offset within the Vienna Woods, <br />Austria: correction of the data for this offset reduced the mean square error from 26 cm to 10 cm. A LlDAR accuracy <br />assessment of beaches along the U.S. Atlantic coast (Shrestha et aI., \999) found LlDAR vertical offsets to be +20 cm <br />to -10 cm with LIDAR accuracies of 15-20 em only obtained after offset correction. A more recent UDAR evaluation <br />along the Green River in Utah found an overall LIDAR offset of -44 cm on bare sand, cobble bars, brush-covered areas, <br />and rock outcrops (Bowen and Waltermire, 2002). Adjustment of the LIDAR data by this offset resulted in a vertical <br />accuracy of 22 cm. <br />Results from these studies suggest that LIDAR or photogrnmmetry might provide acceptable elevation data for <br />GCMRC monitoring requirements, but these studies did not address some important issues for GCMRC monitoring. <br /> <br />EV ALUA nON OF LIDAR AND PHOTOGRAMMETRY FOR MONfIORlNG VOLUME CHANGES IN <br />RIPARIAN RESOURCES WI1lIIN TIlE GRANDCANYON,ARIWNA <br /> <br />Pecora IS/Land Satellite Information IV/ISPRS Commission 1IF1EQS 2002 Conference Proceedings <br />