In this Issue

National Antarctic Data Coordination Center
Products and Services
ARCTIC SYSTEM SCIENCE (ARCSS)
Hierarchical Data Format (HDF)
Meetings

NSIDC Notes

Issue 19, Fall 1996

NATIONAL ANTARCTIC DATA CENTER AT NSIDC

NSIDC has begun the development of the Antarctic Data Coordination Center (ADCC) at NSIDC. The ADCC will provide a U.S. focus for the collection of metadata for the Antarctic Master Directory (AMD) which is being hosted at the International Center for Antarctic Information and Research (ICAIR) in Christchurch, New Zealand. The National Science Foundation's (NSF) Office of Polar Programs is funding the ADCC activity at NSIDC and much of the AMD effort at ICAIR. The AMD will be closely linked to the NASA Global Change Master Directory.

NSIDC will identify Antarctic data sets (and cognizant individuals for these data sets) in all disciplines and prepare or assist in the preparation of data descriptions (DIFs) which will be the basis of the AMD. Already underway is an inventory of U.S. Antarctic data sites on the World Wide Web. We also have prepared a brochure for the ADCC. Recently, NSIDC briefed the National Research Council/National Academy of Sciences Polar Research Board on this activity. The Board provided much of the impetus for the coordination of metadata for the Antarctic through its involvement in the Scientific Committee for Antarctic Research (SCAR). NSIDC also met with the relevant program managers from NSF to review plans for the ADCC.

NSIDC is proposing to attend the National Antarctic Data Center (NADC) managers workshop to be held at ICAIR in 1997. Representatives from each of the participating countries will meet to coordinate their activities, discuss guidelines for operation of the NADCs, and become familiar with the data description software tools being developed by ICAIR for use across the AMD.

For further information contact NSIDC User Services.

PRODUCTS AND SERVICES

SSM/I F13 Data

Production of CD-ROMs containing SSM/I F13 brightness temperatures has begun at NSIDC, with delivery of the first three CDs to users expected to begin in January 1997. These three CDs include brightness temperature data for nine months, from May 1995 through January 1996.

New CD-ROM: Arctic and Antarctic Sea Ice Data, 1972 - 1994

We have just received a new CD-ROM from NIC, the National Ice Center, (formerly known as the U.S. Navy- NOAA Joint Ice Center - [JIC]), containing weekly digitized sea ice charts for Arctic West, Arctic East, and Antarctica, in the SIGRID (Sea Ice Grid) data interchange format of the World Meteorological Organization.

The data on this disk are corrected versions of the files that were formerly distributed by NSIDC and NOAA's National Climatic Data Center (NCDC) on 9-track tape. Extensive quality control and corrections have been completed by NIC and NCDC; this CD-ROM replaces all previous versions of the weekly NIC/JIC SIGRID data. The CD-ROM also extends the time period that was formerly available by four years; 1991 - 1994 are newly-released files. Complete documentation is provided on the CD-ROM, including the SIGRID format source document and a discussion of the types of errors leading to reprocessing.

The CD-ROM is available at no charge from NSIDC. Please contact NSIDC User Services if you wish to receive a copy from CIRES, 449 UCB, University of Colorado, Boulder, CO 80309-0449 USA.

Gridded SSM/I Brightness Temperature Data From NSIDC

NSIDC produces and distributes two gridded SSM/I brightness temperature data sets on CD-ROM. There are a number of significant differences between the two data sets, and we want to highlight those differences so that users may choose the product which is best suited for their application.

NSIDC began producing and distributing the DMSP SSM/I Daily Polar Brightness Temperature Grids in 1989. The data set currently covers data from the DMSP F-8, F-11 and F-13 platforms, for the period July 1987 through the present. (See F-13 Data update, above). This data set was intended primarily to calculate sea ice concentration and type. The geographic coverage reflects this purpose, and consists of the area around either pole which includes the maximum sea ice extent.

In 1995 NSIDC began production of the NOAA/NASA Pathfinder Program EASE-Grid (Equal Area SSM/I Earth Grid) Brightness Temperature data. The data on CD-ROM currently cover the F8 platform for the period August 1987 - December 1988, with processing underway to continue through the F8, F11, F13 platforms. This product was designed for broader applications not limited to polar research, and has global geographic coverage, comprising North or South polar azimuthal projections and a global cylindrical projection.

Although there are a number of differences between the gridded SSM/I data sets, there has been confusion among our users regarding the purpose and characteristics of the two SSM/I brightness temperature products.

Table 1. Primary differences between the two brightness temperature products.
SSM/I Polar Stereographic Grids EASE-Grids
Coverage limited to polar regions,
oriented towards sea ice applications
Global coverage
Individual files contain all
passes averaged daily
Individual files contain separate daily
ascending and descending passes;
no averaging
Binning based on "drop in the
bucket" method
Binning based on weighting functions derived
from actual antenna pattern
Data in Hierarchical Data (HDF)
Format; no compresion
Data in raw binary and compressed using
GNU gzip

Please see NSIDC for sample images and more information on both of these data sets, or contact NSIDC User Services.

Corrections to Buoy Data from International Arctic Buoy Program

Dear IABP Data User:

In May of this year, we sent out a letter detailing an error that had been uncovered in the computation of the second derivative of pressure, d2p/dxdy, reported in the Arctic Bouy datasets. This error has since been corrected, the reanalyzed datasets have been been rechecked, and are available from NSIDC, and via FTP or HTTP from the IABP data server, IABP.APL.WASHINGTON.EDU.

The pressure fields were reanalyzed using the procedure outlined in the 1982 buoy reports. Although the original pressure fields for 1979 - 1981 did not contain the error, these fields were analyzed using an older procedure outlined in the 1979 buoy report. For consistency, these years were also reanalyzed. Using the 1982 procedure, the estimated pressure at a grid point is optimally interpolated using the buoy pressure observations in combination with the NMC sea level pressure field.

Sincerely,

Ignatius Rigor, IABP Data Manager

A MORE COMPREHENSIVE VIEW OF THE DYNAMICS OF ARCTIC SEA ICE AS REVEALED THROUGH 85.5 GHz SSM/I IMAGERY

Tom A. Agnew * and Hao Le
Climate and Atmospheric Research Directorate,
Environment Canada
Downsview, Ontario, Canada

1. Introduction

The 85.5 GHz channels on the Special Sensor Microwave Imager (SSM/I) have historically not been used very much for estimating large scale geophysical properties of sea ice (Steffen et al., 1992) because of high attenuation due to atmospheric moisture and liquid water. During winter over the Arctic Basin, atmospheric moisture is typically low and atmospheric attenuation is not a serious problem. As a result, the higher pixel resolution of 85.5 GHz imagery compared to lower frequency channels can be used to advantage to observe the sea ice surface. It turns out that at pixel resolutions of 12.5 km, large multiyear ice floes and large leads which form over the Arctic Basin can be resolved. Animation of daily SSM/I imagery over periods of several month to several years has been done and the results reveal important dynamic behavior of the sea ice pack not observed before.

2. Quantitative Ice Motion

To quantify the ice motion seen in the animation, an image matching technique developed by Noetix Research Ltd. under contract to AES, Ice Branch, was used. The method is based on maximum cross correlation between windowed portions of a pair of SSM/I images. The method has been completely automated in a software package called Tracker (Hirose et al., 1991). The technique has been applied to 85.5 GHz SSM/I daily average imagery from the National Snow and Ice Data Center to obtain ice motion over the entire Arctic Basin for a contiguous two-month period. The accuracy of 85.5 GHz SSMI-derived ice motion is evaluated by comparing results with Arctic drift buoys. Ice motion estimated between January 13 and January 17, 1994, using the technique is shown in Figure 1. The vectors in white are ice motion estimated using the image matching technique. The small square marks the beginning of the displacement vector. The black vectors are ice motion estimated from Arctic buoy drift. The large circular region centred over the North Pole is a data void region of the sensor. The ice motion shows the very strong anti-cyclonic Beaufort Sea Gyre in the Canada Basin and large ice motion in Fram Strait (between Svalbard and Greenland). This is very consistent with the surface pressure pattern at the time which was characterized by a very intense anticyclone circulation over the Canada Basin/Beaufort Sea and strong pressure gradient over the north Greenland Sea. The average displacement of all the Arctic buoys over the Basin is 6.5 km per day. The average motion at the buoy locations using the imaging technique is 6.6 km per day.

A series of SSM/I image pairs over a two-month period 1 December 1993 to 31 January 1994, were used to estimate ice motion over the Basin and to compare results to Arctic buoy drift. A scatter plot of comparisons over the Canada Basin and the Makarov Basin is shown in Figure 2. A total of 391 comparisons were made and the correlation coefficient is .75. In general, results compare well with buoy drift but there is a tendency for the image matching technique to overestimate motion by about 10 per cent compared to the buoys.

3. Climate Applications

Sea ice motion is important for understanding climate processes at the sea ice-atmosphere-ocean interface. Differential ice motion determines the extent of open water within the ice pack which in turn greatly influences local sensible and latent heat exchange with the atmosphere, the rate of new ice production, brine production within the surface layers of the ocean, and ice ridging and thickness distribution. On longer time scales, sea ice motion determines the rate that sea ice is advected out of the Arctic Basin through Fram Strait and through the Canadian Archipelago and is therefore important in sea ice budget studies of the Arctic Ocean. This is also important in determining the rate of freshening of the North Atlantic and the intensity of thermohaline circulation of the global ocean (Aagaard and Carmack, 1989).

Results obtained for the two-month period indicate that except for the summer months when atmospheric moisture and liquid water content is high, obstruction due to cloud is not a serious problem with 85.5 GHz data. As a result, it should be possible to produce a database of large-scale ice motion over the Arctic Basin for the entire record of SSM/I imagery (September 1987 to the present). These data would provide much more spatial detail than is currently available from buoy data and would be useful in sea ice budget studies of the Arctic Basin and in sea ice model validation.

The large aggregate scale ice motion information over the entire Arctic Basin obtained from SSM/I 85.5 Ghz imagery complements ice motion obtained from other types of imagery such as synthetic aperture radar (SAR) and AVHRR which provides regional and local scale ice motion information.

* For more information contact NSIDC User Services. A more complete evaluation of the technique can be found in Agnew et al. (1996).

Acknowledgments

The authors would like to thank the National Snow and Ice Data Center for providing the processed SSM/I data on CD-ROM and the Polar Science Center, University of Washington, Seattle, Washington, for providing the Arctic buoy data for this study. This work is supported in part by the Canadian Panel on Energy Research and Development.

References

Aagaard, K. and E.C. Carmack, 1989. The role of sea ice and other fresh water in the Arctic circulation. Journal of Geophysical Research, 94 (C10): 14,485-14,498.

Agnew, Tom A., Hao Le, and Thomas Hirose. 1997. Estimation of large-scale sea-ice motion from SSM/I 85.5 GHz imagery. Annals of Glaciology 25: 305-311.

Hirose T., T. Heacock and F. Lee, 1991. Advances in Sea Ice Tracking Algorithms. Proceedings of the 14th Canadian Symposium on Remote Sensing, p. 292-295.

Steffen, K., J. Key, D. Cavalieri, J. Comiso, P. Gloersen, K. St. Germain and I. Rubinstein, 1992. Estimation of geophysical parameters using passive microwave algorithms, In: Carsey, F.D., ed. Microwave Remote Sensing of Sea Ice. Geophysical Monograph no. 68. Washington, D.C., American Geophysical Union, p. 201-231.

ARCTIC SYSTEM SCIENCE (ARCSS)

GISP-2 Data Acquisitions and CD-ROM

The ARCSS Data Coordination Center at NSIDC is currently compiling data from the Greenland Ice Sheet Project (GISP2) for dissemination on CD-ROM and the WWW. We expect all data from the GISP2 project to be downloaded to NSIDC by early 1997. A CD-ROM of data from both GISP2 and the European Greenland Ice Core Project (GRIP) will be released in 1997 in conjunction with the publication of a special issue of the Journal of Geophysical Research devoted to the GISP2/ GRIP projects. The data will also be available via FTP from ARCSS.

Arctic Radiation Data CD-ROM

Mark Serreze, an ARCSS-funded investigator at NSIDC, is compiling a new monthly gridded climatology of arctic downwelling short-wave radiation. In conjunction with this effort, we are currently assembling all of NSIDC's Arctic radiation data for inclusion on a CD-ROM. This will include short-wave, long-wave and net radiation values from stations above 50°North as well as the new Serreze climatology. The data will also be available via FTP from ARCSS.

For information on GISP2 or radiation data, contact NSIDC User Services.

HIERARCHICAL DATA FORMAT (HDF)

During August we conducted a survey of users of the data products that we distribute in Hierarchical Data Format (HDF) as part of our NASA EOSDIS project. We thank the 74 respondents for taking the time to provide detailed comments and suggestions. We also plan to devote a portion of each issue of this newsletter to delivering information about HDF and its effective use, including sources of applicable software, and "tips and tricks" for NSIDC's HDF products. As we re-organize our Web site we will include this same type of information there, with links to other sites that provide software or helpful information pertaining to HDF.

Survey respondents were in the following categories, determined from either the e-mail domain or the affiliation listed by the respondent:

U.S. college or university40.5%
Non-U.S. college or university17.5%
U.S. government12.1%
Non-U.S. government10.8%
Unknown9.5%
U.S. Commercial (.com domain)4.1%
Non-U.S. commercial2.7%
Kindergarten through 12th grade1.4%
U.S. other (.org domain)1.4%

We learned from the survey that you use a wide variety of commercial software, freeware and shareware to read and analyze our HDF products, the SSM/I polar stereographic grids, AVHRR browse images, and TOVS Pathfinder Arctic atmospheric grids. IDL, PVWAVE, EASI-PACE, Terascan, Spyglass, and Windows Image Manager were listed most frequently among the commercial applications, with NCSA's Collage as the most popular freeware. About one-third of you indicated you use your own software either designed especially to read the NSIDC HDF products or adapted to do so. Your responses indicated, not surprisingly, a wide variety of computer platforms in use with our data: SUN, SGI, HP, PC Linux, and other unspecified types of UNIX platforms; DOS, Windows 3.x, Windows 95 and Windows NT, and Macintosh were listed. A total of 75 responses in the "Unix" category, 32 in the "PC" category, and 6 in the "Macintosh" category remind us that our user community is not uniform!

We asked for specific examples of your experience with our HDF products, whether positive or negative, and we received many detailed comments. Here is a selection, both positive and negative:

"HDF is easy to view quickly (given the right software); climate model I am using also uses HDF format."

"No problems. NSIDC's packaging of the data in HDF is simple, straightforward, and usable."

"IDL handles the technical aspects. I just had to write some simple programs to plot the data once it had been extracted."

"You can see any result so fast with the HDF. It is easy to read and transform data using conversion programs like (hdftor8, etc.)."

"Once set up - easy to use."

"It is very straightforward to view the files with NCSA Collage on the Mac. To actually use the data I convert the format. It's a clumsy procedure but I want to be able to use the data with my other data sets."

"We had to change our software for the F8 CDs in order to read the F11 CDs. No major problem though, since all the necessary information is in the User's Guide!"

"A bit more of a hassle to read HDF (SSMI-F11) format than it was to read the SSMI-F8 format. But, this is not a big deal."

"For my work, there have been no real benefits. I usually just strip the data down straight away using hdftor8 and convert it into a format that we can easily use here." [Note: "Stripping" out the HDF information must be accomplished using the HDF utilities "hdftor8" or "hdftor16", because the HDF is not always just a header; in some files it is embedded in the file structure itself. Contact NSIDC User Services if you are not using "hdftor8" or "hdftor16" for this purpose.]

"Would like to manipulate HDF SSMI files on my Mac Powerbook 5300 using NIH Image or GIF Converter, but cannot because no way to get out of HDF format." [Note: At press time, a functioning Macintosh version of "hdftor8" was still not available from NCSA.]

"I find HDF overly complex - would prefer netCDF."

"Negative: Expense, time learning new software packages to handle HDF. Too many different data formats out there."

"HDF software was difficult to obtain and difficult to use."

"Required revision of existing "C" language and IDL software to handle HDF files. Required installing HDF libraries solely to handle the NSIDC CDROMs."

"HDF is a pain.... It is large and hard to use. HDF does not provide the supposed "easy to read" capability advertised when you want to use your own utilities. It is especially hard to write."

Then we asked, "In general, is it an overall positive or negative experience to use NSIDC data products in HDF?"

Responses were 55.4% "positive", 18.9% "negative", 5.4% "neutral", and 20.3% had no response to this question. In reading the survey forms, we noted a general trend that the "positive" respondents were more likely to be people who had contacted us to ask for assistance, while the "negative" respondents were more likely not to have contacted us. This is not statistically significant, but gives us the general idea that we can provide the type of support to our users that enables them to work more effectively with our data products. So if you are having difficulty and have not asked for help, perhaps you should give us a try!

One of the primary reasons we undertook the HDF survey was to find out whether there is a community of scientists and other serious data users who rely on PC Windows and Macintosh platforms for their data work. From the platforms listed above, it seems there is a PC Windows constituency out there, and there are even some loyal Macintosh users. This information has been provided to the NASA EOSDIS project in the hope that some of the tools development work that is now underway for EOSDIS Version 1 can be extended from purely Unix to include some PC and Macintosh applications as well. We specifically asked whether users would "vote" for development of PC and Macintosh versions of "EOS View", the forthcoming Unix-based data visualization utility package designed for use with data in HDF-EOS from the EOS instruments. There were 74 respondents who indicated an interest in trying EOS View for Unix, while 47 "voted" for development of a PC version and 25 for a Macintosh version. Some respondents indicated a vote for more than one platform version.

Secondarily, we asked some questions about the User's Guide for our SSM/I polar stereographic gridded brightness temperature and sea ice product, and learned that many of you were not aware of our Web site nor that the User's Guide is available online. So we remind all of you that we are on the Web. Our newly re-organized site will debut in time for the Fall Meeting of the American Geophysical Union (see related note below); we hope you will take a look and let us know what you think.

If you did not receive a copy of the HDF survey, either by postal mail or by browsing our Web site during August and September, and if you have comments, complaints, suggestions, or other feedback regarding our HDF data products and related support services, we would be happy to hear from you at any time. Contact NSIDC User Services; we'll listen and respond to the best of our ability.

MEETINGS

NSIDC at AGU

NSIDC staff will participate in the Fall American Geophysical Union meeting in San Francisco, 15-19 December, with a booth in the Exhibit Hall and several poster presentations.

Look for us at Booth 321. We will have our updated data catalog available and also will be unveiling our revamped Web site which offers access to our data and information services.

Look for NSIDC at the poster sessions. Topics include:

Snow trends within the Former Soviet Union - C. Haggerty

Snowmelt runoff modelling of the Upper Amu Darya Basin, FSU - D. Bedford

Northern Hemisphere weekly snow cover and sea ice extent, 1978- 1995 - R. Armstrong and M.J. Brodzik

Atmospheric controls on Eurasian snowfall - M. Clark

Improved AVHRR imagery of ice sheets via data cumulation - G. Kvaran, T. Scambos and M. Fahnestock.

See you there!