Volume No. 10, 2001
As a part of the Southwest Regional Gap Analysis Project, the Colorado Division of Wildlife (CDOW) and U.S. Bureau of Land Management (BLM) conducted an evaluation of helicopter-based methods for collecting ground-truth reference information and compared this methodology to collecting data via automobile and on foot. These data are used for classifying Landsat-7 Enhanced Thematic Mapper satellite imagery in developing a land cover map of a five-state region in the Southwest. It was found that although more expensive than traditional ground-based collection of field data, the helicopter method had some advantages.
The Southwest Regional Gap Analysis Project (SW ReGAP) is attempting to create a high-resolution, seamless land cover map of Arizona, Colorado, Nevada, New Mexico, and Utah using Landsat-7 TM satellite imagery, field data, digital elevation models, and other spatial data. Thematic categories are based on the National Vegetation Classification System (NVCS) (Anderson et al. 1998, Grossman et al. 1998, Jennings et al. 2002). To evaluate efficient methods for collecting training site data for the land cover classification process, we compared training site data collection by helicopter, as has been used by the BLM in Alaska, to a more traditional method of travel by automobile and on foot to visit sample sites of each land cover type. Mission planning time, mission execution logistics, methodological efficiencies, and cost considerations were evaluated and compared.
The study site for the helicopter data collection evaluation was
within the Southern Piedmont Mapping Zone, a SW ReGAP defined
ecoregion (Manis et al. 2000) on the southeastern plains of
Colorado near La Junta. Since Landsat TM-7 imagery was not
yet available for use in the evaluation, 1995 Landsat-5 TM was
utilized. The helicopter method evaluation was conducted over
three days from June 11-13, 2001.
This method of data collection was compared to data collected by
traveling by automobile and on foot throughout the summer of 2001
(July through October) over the entire Southern Piedmont and some
adjacent high plains mapping zones.
First, for the helicopter protocol, a "ground school" was held
to provide aircraft safety training for participating field
personnel.
Then, field sites to be visited were selected using an ArcViewII
Avenue script (O'Brien and Schrupp 2001) designed to randomly
select 10 field sites per each of 58 spectral cluster classes of a
minimum size (2 ha) and of a specified distance either close to or
far from roads (50 meters). The 58 cluster classes were
generated from an unsupervised classification of a two-date,
six-band, merged data set of Landsat-5 imagery (June 21, 1995, and
September 25, 1995) and delineated a set of spectrally homogeneous
land cover patches. Once selected, the target sites were
transferred to 1:100,000 scale BLM Surface Management Series Status
maps, and digital files of the geographic coordinates of the
centers of each site were uploaded to a GPS receiver to aid in the
helicopter navigation. Field personnel from the Colorado
component of SW ReGAP and the Colorado Vegetation
Classification Project collected the site data. The same
computer programs and laptops that were used for collecting site
data via ground methods were used for recording data from the
helicopter.
The helicopter method for accessing ground control points was
similar to one used by agencies in Alaska. The helicopter
would travel to each target field site and either land or hover
over the site, depending upon landowner access considerations.
Each "mission" was typically less than 2 hours of air time,
including team rotation and refueling. Coordinates of the
target field sites selected for each mission were loaded onto a GPS
unit before each crew rotation.
Following BLM aircraft safety guidelines, the helicopter's ground
movement was shadowed by ground teams that provided for
air-to-ground communication between the pilot, the aviation fuel
manager, and the Safety Management crew.
GPS units were used to navigate to each field site where data were collected. These data were later used to classify the site to the Alliance level of the NVCS. One to four digital photos were taken at each site, from either right or obtuse angles, generally at heights of 90 m and 30 m above the site. Descriptive information for each site was catalogued on a field form, and associated photo numbers were catalogued on the field form, the navigator's map, or both.
The methods for collecting site data by travelling by automobile and on foot were similar to the helicopter methods, except that, obviously, we could not hover over a site. As much of the land on the plains of Colorado is in private ownership, crews were prevented from walking out onto many of the sites, and land cover had to be described from the roadside.
Costs for the helicopter protocol were tracked via BLM's standard "Aircraft Services Reimbursement" procedures for helicopter costs. Costs included ferry time of the helicopter from its home base in Englewood, Colorado, to the study site in La Junta, aircraft time while conducting field sampling, personnel time of both the pilot and aviation fuel manager, and per diem for the air crew. There were additional costs for field crew time, per diem, and vehicles. Both CDOW and BLM contributed personnel time towards the evaluation.
Eight crew rotations were performed during the helicopter evaluation on June 12 and 13 (two on the 12th and six on the 13th). Problems were experienced with the Trimble Geo-Explorer III GPS unit, which took about half the day of the 12th to resolve. Ultimately, the pilot's Garmin unit was used for navigation to the field sites. In summary, 9.2 hours (3.5 hours on 06/12/01 and 5.7 hours on 06/13/01) at $750/hour were spent aloft, visiting 48 sites over the two days of site description activity (13 on the 12th and 35 on the 13th). Costs for visiting field sites by helicopter averaged $228/point over the two days. Helicopter costs alone (the most significant component of the project) averaged $265/point for the first day and $145/point for the second day.
By comparison, it would have taken approximately 54 hours to visit these sites by automobile and on foot. Costs of traditional data collection were extrapolated from costs of 39 field days spent during four months of field work, from July to October of 2001. The average number of sites visited during these trips was nine per day. Costs for visiting these sites by automobile and on foot averaged $72/point.
About half a day was wasted dealing with GPS and site coordinate
problems, while money was being spent for helicopter personnel
time.
This increased the overall cost of each field point collected
using the helicopter method. Once these problems were
resolved on the second day, the costs per site visit came down to
what we feel should be expected for this type of operation.
The costs for the helicopter method were higher compared to
traditional methods; however, many more sites were visited per day,
and better land cover classifications were obtained through better
access to the sites and the ability to view sites from above, as
the satellite does, and make better cover estimates.
1. A synoptic view of the field site; more in keeping with the view-angle of the satellite than of ground-based field crews.
2. Access to field sites that could not have been visited from the ground, given the sparseness of roads in southeastern Colorado and the amount of privately held land.
3. Efficiencies of travel time to and from field sites.
1. Verify the coordinates of field points to be loaded to the navigational GPS unit and test the procedures for doing so.
2. Make sure the coordinate system used on the GPS unit are the same as those used by the helicopter pilot.
3. Make sure the GPS equipment has a robust antenna system and all field crew members are versed in its operation. Have a hard copy of the GPS operator's manual in hand.
4. Download and catalogue digital photos each evening.
5. Upload and check the next day's field targets the evening before.
While the helicopter data collection methodology is relatively expensive, it affords some benefits not achievable with a ground-based methodology, and the cost/benefit ratio may be improved through careful planning. The BLM and the US Forest Service often post helicopters at remote locations for readiness in the event of wildfires throughout the fire season, and there may be cost benefits realized by scheduling such craft when they are not being used to fight fires. Even at the standard rate, helicopter use to visit a subset of field sites may be the most efficient way to build a high-quality photo-interpretation key to the land cover types being classified. This research did not include a cost-benefit evaluation of using aerial photographs or videos taken from fixed-wing aircraft, in combination with ground reconnaissance.
Anderson, M., P. Bourgeron, M.T. Bryer, R. Crawford, L. Engelking, D. Faber-Langendoen, M. Gallyoun, K. Goodin, D.H. Grossman, S. Landaal, K. Metzler, K. D. Patterson, M. Pyne, M. Reid, L. Sneddon, and A.S. Weakley. 1998. International classification of ecological communities: terrestrial vegetation of the United States. Volume II. The National Vegetation Classification System: list of types. The Nature Conservancy, Arlington, Virginia.
Grossman, D.H., D. Faber-Langendoen, A.S. Weakley, M. Anderson, P. Bourgeron, R. Crawford, K. Goodin, S. Landaal, K. Metzler, K. D. Patterson, M. Pyne, M. Reid, and L. Sneddon. 1998. International classification of ecological communities: terrestrial vegetation of the United States. Volume I. The National Vegetation Classification System: development, status, and applications. The Nature Conservancy, Arlington, Virginia.
Jennings, M., O. Loucks,
D. Glenn-Lewin, R. Peet, D. Faber-Langendoen, D. Grossman, A.
Damman, M. Barbour, R. Pfister, M. Walker, S. Talbot, J. Walker, G.
Hartshorn, G. Waggoner, M. Abrams, A. Hill, D. Roberts, and D.
Tart. 2002. Standards for associations and alliances of
the U.S. National Vegetation Classification. The
Ecological Society of America, Vegetation Classification Panel.
Version 1.0, May 2002.
Manis, G., C. Homer, R.D. Ramsey, J. Lowry, T. Sajwaj, and S. Graves. 2000. The development of mapping zones to assist in land cover mapping over large geographic areas: A case study of the Southwest ReGAP Project. Gap Analysis Bulletin No. 9. USGS Gap Analysis Program, Moscow, Idaho.
O'Brien, L.E., and D.L. Schrupp. 2001. ArcView random site selection tool (script). Natural Resource Ecology Laboratory, Fort Collins, Colorado.
Return to Table of Contents