My WebLink
|
Help
|
About
|
Sign Out
Home
Browse
Search
WSP02238
CWCB
>
Water Supply Protection
>
Backfile
>
2001-3000
>
WSP02238
Metadata
Thumbnails
Annotations
Entry Properties
Last modified
1/26/2010 12:35:30 PM
Creation date
10/11/2006 11:00:00 PM
Metadata
Fields
Template:
Water Supply Protection
File Number
8220.101.10
Description
Colorado River-Water Projects-Glen Canyon Dam/Lake Powel-Glen Canyon Adaptive Management
Basin
Colorado Mainstem
Water Division
5
Date
1/1/2004
Author
Phillip Davis
Title
Review of Results and Recommendations from the GCMRC 200-2003 Remote Sensing Initiative for Monitoring Environmental Resources Within the Colorado River Ecosystem
Water Supply Pro - Doc Type
Report/Study
There are no annotations on this page.
Document management portal powered by Laserfiche WebLink 9 © 1998-2015
Laserfiche.
All rights reserved.
/
78
PDF
Print
Pages to print
Enter page numbers and/or page ranges separated by commas. For example, 1,3,5-12.
After downloading, print the document using a PDF reader (e.g. Adobe Reader).
Show annotations
View images
View plain text
<br />00846 <br /> <br />Some vegetation alliances were found difticult to discern using the 4-band ISTAR data. <br />which raises an imponant issue for future vegetation mapping. Is a sensor that provides <br />additional useful wavelength bands at the expense of spatial resolution (and derived texture) <br />better for future vegetation inventories than a sensor that provide a few bands but higher spatial <br />rcsolution? Before the next data collection for vegetation inventory, this question needs to be <br />answered so that mapping can be less time consuming and more accurate. Higher order <br />multispeclral data (refcrred to as hyperspectral data) were acquired for small areas within the <br />CRE in order to evaluate these data. HYDICE (an Navy experimental sensor) acquired 206 bands <br />at 1.5-m resolution within Glen Canyon, but the data proved to be too noisy (poorly calibrated) to <br />use for vegetation classification. A VIRIS (a NASA instrument) acquired 220 bands at 4-m <br />resolution, but occurrences of some imponant vegetal ion alliances are much smaller than an <br />A VIRIS picture element. We hope that a well-calibrated, hyperspectral data set can be acquired <br />in the near future at a few differcnt resolutions in order to resolve this issue, preferably using a <br />commercially available system so that we could confidently employ the system once collection <br />specifications were determined. <br /> <br />3.2.1 Canopy elevation and volume <br /> <br />Currently, vegetation stand volume is estimated from the stand's area, which can be <br />oblained from correctly onhorectified imagery, and from spot vegetation height measurements. <br />which may be limited to the more accessible parts of a stand. Two remote-sensing approaches. <br />photogrammetry and L1DAR, were investigated that potentially could produce more accurate <br />(representative) stand volumes. For this evaluation, we examined photogrammetric data <br />produced from I :4,000-scale photography and two sets of L1DAR data acquired using different <br />L1DAR sensors that collected points at a 1.5-m and 3.75-m spot spacings. The L1DAR data were <br />evaluated as a potential method for mapping canopy volumes because neither L1DAR data set <br />provided ground elevations within our CRE vegetated test arcas. We therefore thought that these <br />L1DAR data might at least provide canopy elevations. [Our assessment of L1DAR for ground <br />elevation is discussed in the following section within the physicaf resource program.] Based on <br />our assumption that stand volume could be under-lover-represented by 20% using current field <br />sampling technique, we set this level of accuracy as the minimum accuracy for the remote- <br />sensing data. <br /> <br />Our analyses of the photogrammetric and low-to-moderate resolution L1DAR elevation <br />data (Davis et aI., 2002a and submitted) showed that photogrammetry is much more accurate for <br />mapping canopy elevalions than are either of the two L1DAR surveys. We found thaI (I) 67% of <br />the photogrammetric spot elevations were within 20% of ground-surveyed canopy elevation:;;-(2) <br />only 38% of the high-resolution (I-m spot spacing) L1DAR data metthis accuracy, and (3) less <br />than 5% of the moderate-resolution (3.75-m spot spacing) L1DAR data met this criterion (Figure <br />20). The new ISTAR airborne technology that we employed in June of2002 for the entire CRE <br />produced a I-m digital surface model (DSM), which is a digital elevation model for the retlected <br />surface. that should include the vegetation canopy within dense stands. These data are currently <br />being evaluated for canopy heights, which will require that the DSM data accurately represent <br />both the canopy and surrounding bare ground. These DSM data were produced without human <br />intervention using automated softcopy photogrammetry and therefore the cost is relatively low <br />($625 per river km) compared to more conventional photogrammetric analyses. On the other <br />hand, we are currently evaluating very high-resolution L1DAR data for topography and canopy <br />height for vegetated areas. We have already found these data to be extremely accurate on bare <br />surfaces (8 cm) and hope this also holds for the vegetation. However, the cost for these L1DAR <br />data ($6,200 per river mile) may preclude its use for large-area volume estimates. A less- <br /> <br />17 <br />
The URL can be used to link to this page
Your browser does not support the video tag.