The NSIDC Web site and data services are currently having intermittent problems and may be unavailable. We are working to restore these services as soon as possible and apologize for any inconvenience this may cause. Please contact NSIDC User Services for assistance.
In this Issue
Greenland Airborne Precise Elevation Survey (GRAPES)
The Greenland Airborne Precise Elevation Survey consists of a series of survey lines distributed over the large ice sheet that covers most of the interior of Greenland. These surveys, which were conducted as part of NASA's Mission to Planet Earth, are designed to provide a precise baseline from which net changes in the volume of the ice sheet can be accurately measured. Effects of global climatic change on the Greenland Ice Sheet and resulting changes in world-wide sea level are not well understood and have been the subject of considerable debate within the scientific community. Future airborne and/or satellite altimeter measurements of the ice surface elevation over these survey lines yield definitive estimates of the net gain or loss in ice volume.The map of Greenland (Figure 1) shows the locations of the individual survey lines. It shows laser-altimetry flight lines flown in 1993 in solid, and in 1994 in dashed lines. The arrow indicates the location of a series of passes collected in 1991 that have been processed into a DEM.
Figure 1. Greenland, showing laser-altimetry flight lines flowin in 1993 in solid, and in 1994 in dashed lines. The arrow indicates the location of a series of passes collected in 1991 that have been processed into a DEM.
The survey was made with the NASA Airborne Topographic Mapper (ATM) and its predecessor, the Airborne Oceanographic Lidar (AOL), during field deployments conducted each year from 1991 to 1994. These sensors are pulsed laser altimeters which scan a laser beam over the ice surface beneath the aircraft with a nutating mirror, producing an elliptical sampling pattern. The off-nadir angle of the scan mirror was set to yield a swath of altimeter measurements approximately 140 m wide. A schematic illustration of an ATM survey over a previously occupied survey track is shown in Figure 2 which shows a conceptual illustration of the application of the scanning Airborne Terrain Mapper (ATM) for ice sheet elevation mapping. The illustration shows the ATM acquiring elevation measurements along a scan swath which overlaps data collected during a previous flight. The conical scanning pattern of the ATM is also depicted in the figure. An expanded discussion of these sensors is provided by Krabill et al. (1995).
Figure 2. A conceptual illustration of the application of the scanning Airborne Terrain Mapper (ATM) for ice sheet elevation mapping. The illustration shows the ATM acquiring elevation measurements along a scan swath which overlaps data collected during a previous flight. The conical scanning pattern of the ATM is also depicted in the figure.
The precise vertical and horizontal location of the aircraft was determined using Global Positioning System (GPS) receivers, employing kinematic differential GPS techniques (Krabill and Martin, 1987). The aircraft heading and attitude information necessary to locate the laser footprint on the ice surface were acquired by a ring laser gyro inertial navigation system. The initial goal was to provide elevation data to a root mean square (RMS) accuracy of 20 cm.
In order to ensure that repeat surveys would obtain data over the 140 m wide data swaths from earlier surveys, the aircraft must be navigated along prescribed routes to an accuracy of ± 100 m or better. Aircraft guidance systems capable of meeting this requirement were not commercially available at the time the program began. Consequently, a PC-based program was developed which utilized the real-time output from one of the aircraft GPS receivers to determine the magnitude and direction of cross-track error from the desired flight path. This information was then fed directly to the aircraft guidance system. Analysis of resulting flight tracks indicates that the NASA-built aircraft navigation system was able to maintain track to within ± 100m of the desired flight path for more than 90% of the time.
Accuracy assessment was performed by:
Results confirm data consistency in 1991 at the 20 cm RMS level, improving to 10 cm RMS in 1993.
The data sets exist in three forms:
Full Data Set - This data set contains all of the individual laser observations within the approximately 150 m swath. Maximum along-track spacing between data points is approximately 20 m and cross-track spacing is less than 4 m. Each record contains the latitude, longitude, and elevation information for each laser footprint on the ice-sheet surface, the signal strengths of the transmitted and reflected laser pulses, aircraft pitch and roll, and the scan angle. These data are in the original conical scanning format, and may be useful for the analysis of very small-scale features and for areas of high relief, such as regions near the calving fronts of outflow glaciers.
Users may find the conical scan data difficult to work with for applications not requiring this level of detail, and NASA is developing profile and gridded data sets, and a capability for users to select subsets of these data for delivery to their sites by ftp. They will be available on anonymous ftp on gdglas.gsfc.nasa.gov. Users may access this by using their e-mail address as their password. The file "README.GRAPES" will give the current status of these data sets. We plan to have the 1993 profile data sets completed by April 1, 1995 and the gridded data sets by May 1, 1995. These higher-level data sets are derived from conically scanned data.
It should be noted that there are sections of data that were collected in the presence of clouds, ice fog, and blowing snow below the aircraft. They are characterized by early laser reflections from the clouds, etc., which yield a surface profile with sections that are anomalously high and very rough. These data have been retained because they may be useful for some applications, and users should make detailed plots of the data of interest to detect these sections. In doing so it will be apparent that a relatively trivial editing scheme can be used to remove the bulk of the cloud and ice fog measurements. Some form of a running RMS calculation will locate these regions, along with areas of extensive crevassing. The questionable data will also have large standard deviations which will be available on the record. These can also be used to edit the data.
All elevations are above the WGS-84 ellipsoid.
Profiled Data - Elevation along the flight track is given at 25 m horizontal separation (Figure 3a). Values are obtained by averaging all measurements within a distance of 12.5 m from the data point using a 3 sigma edit to remove outliers. This data record comprises: average elevation; date and average time (UT) of observation; standard deviation before editing; standard deviation for the data used (after editing); number of measured elevations averaged; and latitude and longitude of the data point.
Figure 3a. Examples of profile and gridded data. The profile includes a section along the centerline of the gridded area. The horizontal area at the low-elevation end of the gridded swath is a meltwater lake.
Gridded Data - Elevation gridded into 25 x 25 m grid (Figure 3b) for the full swath width (~150 m). The grid square, which is uniquely defined on the same stereo-graphic projection used for satellite-radar altimetry data. This is a tangent polar stereographic projection where the plane of projection is located at the geographic North Pole and is normal to the earth's axis. The 25 m interval is an average throughout Greenland and will vary by several cm depending on location. Thus, at locations where one laser flight track crosses another, data will be gridded into identical grid squares, facilitating inter-comparison.
Figure 3b. Examples of profile and gridded data. The profile includes a section along the centerline of the gridded area. The horizontal area at the low-elevation end of the gridded swath is a meltwater lake.
The data set will comprise: average surface elevation; standard deviation before editing; standard deviation after editing; the number of data points in the grid square; a measure of data distribution within the grid square; latitude and longitude and polar stereo graphic coordinates of the grid centerpoint; and the date and averaged time (UT) at which data were collected.
Figures 3a and 3b are examples of profile and gridded data. The profile includes a section along the centerline of the gridded area. The horizontal area at the low-elevation end of the gridded swath is a meltwater lake.
The Profile and Gridded products derived from 1993 flights is available on line to all interested users. These products retain virtually all the information contained within the conically scanned data, except where the surface is extremely rough, and they are far more compact and easy to use. Nevertheless, investigators who find the higher-level products inadequate to their applications will be able to obtain subsets of the scanned data.
A Digital Elevation Model (DEM) created using five parallel flight lines from the 1991 survey is also available on line. The flights, separated laterally by about 800 m, were taken on three consecutive days (September 18-20) in the area denoted in Figure 1. A 100 km section at the southern end of this transect was processed to a 4-km wide DEM by interpolating between the adjacent data swaths. The data are provided on a 50 x 50 m grid using the same projection for the gridded product de-scribed above so that they can be easily intercompared. The accuracy of elevation measurements in 1991 was approximately ± 20cm, and the interpolation process further degrades the accuracy. Consequently, it is difficult to quantify the absolute accuracy of the DEM, but it is probably at the few tens of cm level. These data are included to allow users to obtain a three-dimensional view of the ice surface in a region where there are many undulations and summer lakes. In addition to this intepolated grid, the individual flight lines that comprise the DEM are also available in the standard formats described above.
NASA encourages scientists who are interested in the Greenland Ice Sheet to obtain GRAPES data for their areas of interest. Limited funding may be available to support research using these data particularly by graduate students and recent post docs.
For more information regarding such support contact: Robert H. Thomas, Manager, Polar Research Program, Code YSG., NASA Headquarters, Washington, DC 20546, or Phone: (202) 358-1154, fax: (202) 358-2771, or e-mail: email@example.com
GRAPES data are available on anonymous ftp at gdglas.gsfc.nasa.gov. Users may access this by using their e-mail address as their password.
Krabill, W. and C. Martin. 1987. Aircraft Positioning using Global Positioning System carrier phase data. Navigation 34 (1): 1-21.
Krabill, W., R. Thomas, C. Martin, R. Swift, and E. Frederick. (1995). Accuracy of airborne laser altimetry over the Greenland ice sheet. International Journal of Remote Sensing 16 (7) (May): 1211-1222.