Laserfiche WebLink
<br />grain size is also easily seen in that data. as evidenced by the image mosaic of a channel segment <br />south of the Lower Colorado River (LCR) conlluence (Figure 26). Although these data will be <br />examined in detail this summer, our preliminary evaluation ofthcse data show thaI it requires <br />about onc hour to georectify each image to an existing controlled image base. At 15-cm <br />resolution, there are about 1000 overlapping images that cover the first 100 miles of the CRE. <br />Thus, without accurate GPS and Inertia Measurement Unit (IMU) data for each image to allow <br />automated rectitication, the true cost of use able data needs to include one-half year of salary for <br />the image rectification. We are in the process oftest;ng software that is especially dcsigned for <br />more automated image rectification. which is also relevant for rectifying some of the more useful <br />historical data sets. Another approach is the use of image autocorrelation software. <br /> <br />4.1.3 Main-stem water elevation <br /> <br />During 2002, water resources personncl (Tucson, Arizona) noticed a correlation between <br />one of the LI DAR elevation data sets (acquired in 2000) over the CRE main stem and the <br />historical main-stem elevations that were measured by a ground survey in the early 1900's. They <br />also noticed at some locations that the L1DAR main-stem elevations were at least I m higher than <br />the hislorical water profile and that these locations coincided with riffles or rapids. The L1DAR <br />data used in that comparison were acquired with a NIR laser system. Although L1DAR is not <br />commonly used to measure water surface elevation, the SHOALS L1DAR does use a NIR laser <br />source to measure water-surface elevation, as well as a green-wavelength laser to measure <br />substrate elevation. Thus, water surface elevations might be obtained from conventional L1DAR <br />data. In order to verify this observation, Davis et al. (2002b) compared the March 2000 L1DAR <br />elcvation data over the main stem channel with corresponding water-edge elevations obtained <br />from coincident ground surveys at four long-term monitoring sites. They found a high degree of <br />correlation between the L1DAR water elevations and the surveyed water-edge elevations. The <br />L1DAR elcvations over the main-stem channel showed a vertical root-mean-square error (RMSF) <br />value of30 cm (Figure 27). assuming that ground-surveyed water-edge elevations accurately <br />rellected the water's surface elevation at the center of the channel. Thus, the L1DAR data were <br />found to closely represent water-surface elevations and may in fact be useful in monitoring <br />changes in submergent debris tlows or cobble bars. The majority of the L1DAR elcvations (90%) <br />at most sites were within 40-50 cm of the survcyed water-edge elevations (Figure 27) and, . <br />therefore, LI DAR elevation differences between any two time periods would have to exceed 40- <br />50 cm before being considered significant. <br /> <br />4.2. Terrestrial Env;roW-1enl <br /> <br />On land, the physical resource program monitors change in fine- and coarse-grained <br />sediment storage, which is represented by debris tlows. cobble bars, river terraces, and different <br />types of sand bars. The basic parameters measured consist of geomorphology and topography. <br />These parameters are monitorcd at active sand bars in the New High Water Zone. the return- <br />current channels within fan/eddy complexes (backwaters), and the pre-dam river terraces with <br />cultural resources, both on a historical (annual) basis and on a short-term, experimental-event <br />basis. Study sites number about 35 between river mile 6 and 225. but many of these are camp <br />sitcs and camping beaches; there are II main moniloring sites for sediment that average 2 miles <br />in length. Geomorphology is interpreted and mapped using stereo imagery. Before 2000. <br />scientists within the program used aerial photographic data that had no pointing or positional <br />information, which made rectification and orthorectificat;on very difficult. time-consuming. and <br />error-prone. The remote-sensing PEP recommended that future data be acquired with GPS and <br />IMU data. This panel also recommended an image resolution of 1-5 m for sand-bar mapping. but <br /> <br />22 <br />