The ice-surface-temperature data generated by J. Comiso were reviewed for possible distribution by NSIDC. The data are monthly mean temperature data for January and July for the 17-year period 1979-1996 for the continental ice sheet of Antarctica, using AVHRR resampled to 6.25-km grid spacing (in the SSM/I polar stereographic projection). An extensive comparison of the data with air temperature data from the Antarctic stations indicated a slight bias and a standard deviation of ~ 3K. However, it is speculated that this rather high standard deviation (relative to ocean surface temperature measurements using AVHRR) may be a result of a combination of differences between skin and air temperature, the difficulties in comparing point data with 6-km pixels (particularly in the vicinity of coastal stations), and some cloud contamination. A comparison of the temperature data with more-remote AWS stations was suggested. It was agreed that the data set would be a valuable addition to the products currently offered by NSIDC, and a complement to the upcoming Polar Pathfinder products. T. Scambos undertook to contact J. Comiso to acquire the products.

The context for the MODIS snow and ice products within the framework of other MODIS standard products (approximately 44 products developed by 32 Principal Investigators at 20 institutions, and processed at 3 DAACs) was summarized. The snow and ice products will be available in orbital and gridded formats, and as daily, 8-day, or monthly composites. Current improvements to the snow and ice products include sea ice-surface temperature and use of a vegetation index to improve performance of the snow algorithm in boreal forests. A snow albedo product may be implemented after the launch of the instrument. NSIDC is preparing for launch by participating in science software integration and test activities and operations test scenarios, and by developing operations agreements with external interfaces. Facilities upgrades are in progress. Production will be phased in at the DAACs over several months. Quality assurance of products will be split, depending on the type of activity, among the MODIS Team and the DAACs. The MODIS Team and NSIDC are discussing the provision of the snow and ice products to selected operational users on a near-real-time basis. The MODIS Team has requested that the snow and ice products be available in the EASE-polar grid. ESDIS Project must approve the needed processing and network resources before this can happen.

NSIDC was charged with handling the distribution and some generation of Polar Pathfinder data sets from within existing resources.

A review of the NSIDC budget revealed that the standard Polar Pathfinder products (passive microwave, AVHRR, and TOVS) could be generated (as needed) and distributed with existing resources for the next 2-3 years.

The funding for most Polar Pathfinder projects ends in September 1998. There are limited funds in the Polar Program for extensions. A new Pathfinder NASA Research Announcement (NRA) is expected to be released in summer 1998 for funding in spring 1999.

There was a general discussion of costs in relation to the media. It was generally recognized that charging users for reproduction of products was not practical. However, it was recognized that NSIDC needed to make sure that users receiving CD-ROMs regularly communicate their desire to continue receiving these products. NSIDC was encouraged to explore alternative cost-effective approaches for distributing products including FTP and DLT tapes.

There was extensive discussion of the need for a major effort to publicize the existence and availability of the Polar Pathfinder data sets to broaden the user community. Strong support was expressed for the idea of a sampler CD-ROM illustrating the capabilities of the standard Polar Pathfinder data sets on the same coarse resolution grid (100 km), the so-called P-cube, and including a representative sample of high-resolution products. Access tools should be included on the CD-ROM. The CD-ROM should be distributed at the American Meteorological Society's Fifth Conference on Polar Meteorology and Oceanography at Dallas, Texas, in January 1999. The International Union of Geodesy and Geophysics meeting at Birmingham, United Kingdom in July 1999 was another desirable distribution venue. Papers in journals and presented at scientific meetings are other publicity activities. The need for a long-time series from P-cube was emphasized.

Altimetry data for the polar ice sheets is an unresolved Polar Pathfinder problem and is not considered in the present budget scenarios for the Polar Pathfinders. Program Manager Prasad Gogenini will assemble a study team to review the current altimetry data processing, archiving, and distribution efforts by Dr. J. Zwally and colleagues.

RAMP (Radarsat Antarctic Mapping Project) data were not considered because of evolving limitations on usage.

Level 2 products will be sea ice concentration (21 km resolution), sea ice temperature (58 km), and snow depth on sea ice (21 km). Level three products will be produced on a polar stereographic projection (SSM/I grid) and include daily averaged brightness temperatures at three resolutions (25, 12.5, and 6.25 km), daily sea ice concentrations and sea ice temperatures at 25 km, and 5-day averaged snow depth on sea ice maps at 25 km.

The objectives were to assess which data assimilation techniques are applicable to regional sea ice models and/or sea ice parameterizations in atmospheric and oceanic GCMs (general circulation models), and to guide NSIDC in the development of data sets for sea ice monitoring in both Polar Regions.

The primary recommendations were to take an incremental approach to assessment of the impact of data assimilation on sea ice simulations, to conduct a pilot study using simple data assimilation models, and to investigate the coordination of research and operational efforts to make more efficient use of resources. Secondary recommendations included: use of data assimilation only in conjunction with accurate error statistics; more emphasis on error assessment of radiances, data and models; much of the assimilatable data are available at 1-3 day intervals, and archival systems will need to be reconfigured if more-frequent inputs are needed; data sets must carry time tags for each pixel because actual observation time is essential in data assimilation; and a passive microwave test data set with error assessment and time tags for each orbit should be made available on FTP by NSIDC.

The project is in the strongest state ever, and is the focus of cryosphere-climate studies in Canada. Radarsat-2 is scheduled for launch in 2001 as a primarily private sector effort. There are unresolved data access questions for interested government agencies. A CD-ROM containing extensive historical snow cover and snow depth observations in Canada is about to be released. The cost will be around $C 150. Finally, the World Climate Research Programme is developing a global cryosphere and climate project including, for example, permafrost and mountain glaciers in addition to the Polar Regions.

There was discussion of a possible joint meeting of CRYSYS and PoDAG in Montreal during October 1998.

• Data Acquisition
  The map of global land ice compiled by GLIMS is reasonably complete. The ASTER Scheduler for data acquisitions will rely on a simple threshold algorithm to detect cloudiness in acquired images. Automatic re-acquisition will be scheduled for cloudy images. This approach will likely fail for snow- and ice-covered areas, and manual cloud assessment will be used for these areas. The ASTER Scheduler is complicated and may not be efficient; prioritization of glacier targets will probably be needed, and input from the glaciological community will be sought.

• Organizational Status
  Collaborators (11 at present), either in the form of Regional Centers or Stewards, continue slowly to sign up.

• Technical Status
  Software development for image classification and delineation of glacier boundaries is behind schedule because of other demands. Development continues on the GLIMS database, which will hold GLIMS analysis products at NSIDC. Prototype software has been developed to generate digital elevation models from ASTER stereo imagery.