The Earth Observing System
By James C. Dodge, Ph.D

The Earth Observing System's Direct Broadcast program will offer exciting new opportunities to users requiring access to near real-time data from the EOS spacecraft, including Terra and EOS PM-1. In contrast to EOSDIS (see article by Timothy Gubbels and Martha Maiden in this issue), which will feature standard product turn around times of approximately 24 to 48 hours, EOS DB will provide near real-time access. It is significant to note that there is also much commercial interest in this capability, with a number of commercial suppliers now advertising complete DB ground stations that are EOS compatible.   

   Direct broadcast is not new; NOAA has had a DB system for their meteorological and environmental imaging sensors for many years. The principal difference between these systems and the EOS DB system will be the amount of data sent to the ground and the bandwidth required for transmission. The data rate for the DB system on Terra will be an impressive 13.125 million bits per second. For the Terra satellite, the EOS DB system will transmit on the X-band frequency of 8.2125 Gigahertz and will be dedicated solely to the MODIS instrument's data stream. For the EOS PM platform, the EOS DB system will be extended to all of the instruments.
    The massive data rate of the downlinked signal requires a sophisticated three-part encoding scheme that includes pseudo-random, convolutional, and error correction Viterbi encoding. Traditionally, high speed decoding of this type of signal has been handled by specialized hardware. Recent improvements in processor speed and modern software have allowed this task to be handled exclusively by software in some solutions. Other hybrid solutions combine hardware- and software-based methods. (See http://rsd.gsfc.nasa.gov/eosdb)
    The size of the antenna required to receive EOS DB is dependent upon the purpose of the ground station. For stations dedicated to receive EOS DB, a dish as small as 3.2 meters in diameter is sufficient. For stations that require the antennae and reception hardware to receive other contemporary data streams such as Landsat 7 or ERS for SAR data, in addition to EOS DB, the antenna needs to be slightly larger, around 4m in diameter. A larger antenna is also required for stations that utilize a radome for adverse weather protection or desire reception all the way to the horizon.
    Because the downlinked data rate is large and the specific implementation technology is still young, the present cost of EOS DB systems is fairly high. Although the initial generation of NASA-sponsored DB systems are associated with development costs in the range of $500,000, it is expected that technology trends will allow cost to drop over the next few years, as numerous commercial companies offer EOS DB compatible systems.
    NASA's present plans for U.S. territory call for developing a few strategically located stations for regional coverage and distribution of their data and derived geophysical products via the Internet. Currently, NASA is developing three DB ground stations for nearly complete coverage of the continental U.S. and an additional one in Hawaii to cover a wide region of the Pacific Ocean. One will be located at the University of Wisconsin, another at the University of South Florida, and a third at a site to be selected somewhere in the Pacific Northwest. NASA's GSFC will maintain a reception capability for occasional use in evaluating the performance of low-cost systems. All data and derived products are planned to be available within 1 hour of reception on individual Internet sites at each ground station, as well as in a mosaic form at the previously listed Web-site.
    The first EOS DB ground station has been completed at the University of Hawaii on the island of Oahu. Their 5m antenna has been tested on various high-bandwidth satellite data transmissions, including JERS-SAR, and found to have very low error rates. It is also adequate for all types of transmissions from EOS in the foreseeable future.
    The potential uses of an EOS DB ground station are many and vary geographically based on region- or sector-specific remotely-sensed informational needs. The unique characteristics of the MODIS instrument offer exciting remotely-sensed information potential. MODIS is a 36 band, calibrated scanning spectroradiometer which has been designed to extend the environmental observing capabilities of the AVHRR instrument on NOAA operational satellites. MODIS will observe the entire Earth every 2 days with moderate resolution (250m to 1km). After careful post-launch calibration and validation of MODIS data using in situ measurements, MODIS data will set the new standard for accuracy in a wide range of environmental observations. MODIS data will be used to measure Earth system parameters such as: land use/land cover change, land/cloud boundaries and properties, ocean color and related phytoplankton/biochemistry characteristics, atmospheric water vapor, land and ocean surface temperatures, and a variety of cloud characteristics, including cloud top altitude, temperature, and type.
    In its initial stages of operation, the EOS DB network will focus on Earth characteristics that can be calculated quickly. Some of these initial EOS DB operational products include regional ocean color and chlorophyll measurements, vegetation indices, cloud types and properties, fire locations, and storm cloud structure. Later EOS DB products will include sea-surface temperatures, aerosol concentrations, and land-surface temperatures requiring more refined corrections for the intervening atmosphere.    

   Developers of new EOS DB products will be able to obtain documentation concerning the MODIS sensor, its interpretation algorithms, and planned calibration and validation activities. (See http://eospso.gsfc.nasa.gov) In addition, the various independent EOS DB Web-sites will also likely contain the individual spectral band data and browse images from each of the 36 MODIS bands and a number of regionally determined test products using fast algorithms suitable for creation within the 1-hour period after each satellite overpass.
    The production of real-time products faces significant technical challenges. Effective reception of the EOS DB signal is only one challenge. Rapid, high-volume processing capability is another big challenge. The MODIS orbital period of 90 minutes places an upper limit on production of that orbit's data, roughly one Gigabyte of data. A production time of 1-hour per orbital swath is therefore a reasonable goal; this baseline requires very fast processing machines and very efficient real-time algorithms to avoid a processing bottleneck or backlog. Numerous strategies are evolving to solve this problem, including powerful UNIX workstations and networks of PCs in a UNIX-controlled "Beowulf" configuration for high-speed parallel computation.
    The results of the product generation will be placed on the individual EOS DB Web-sites and shared with a central site at the GSFC for the preparation of combined real-time products and visualization over the entire U.S. territory. There is also considerable interest in developing or upgrading foreign ground stations in some countries to be able to receive EOS DB data at their locations. Preparations are underway in Australia, China, Japan, Scotland, and Sweden, and there is interest in developing stations in Italy, Germany, New Zealand, Thailand, and Indonesia.
    The Terra satellite is only the first in a long sequence of advanced Earth viewing satellites that will be broadcasting their data in real-time, continuously for all potential users to receive freely. The second EOS satellite observatory will be the EOS-PM satellite, scheduled for launch in December, 2000. It will have not only the MODIS that will be on Terra, but also the Advanced Microwave Scanning Radiometer (AMSR-E), the Advanced Microwave Sounder Unit (AMSU), the Atmospheric Infrared Sounder (AIRS), the Humidity Sounder for Brazil (HSB), and the sensor for Clouds and the Earth's Radiant Energy System(CERES). The data from all of these sensors will be broadcast from the EOS PM satellite for the world to receive.
    Since the types of data go beyond imaging to atmospheric temperature and moisture sounding, some of the operational meteorological agencies around the world are interested in receiving the data and determining if its use in real-time can help to improve their weather forecasts (see article by Dave Jones in this issue).
    In the future, a continuation mission called the National Polar-orbiting Operational Environmental Satellite System Preparatory Mission is planned. An X-band DB system is scheduled for this mission, as well as for the eventual converged U.S. NPOESS. Thus, the community is likely to see a series of high-bandwidth U.S. environmental satellites over the next decade and beyond that will be delivering real-time geophysical observations and measurements to world-wide users on a free and open basis. The proliferation of onboard data processing technology will further increase the efficiency of DB systems, reduce the associated costs, and lead to increased use of real-time data products.

About the Author:  
Dr. James C. Dodge works for the research division in the Office of Earth Science at NASA Headquarters in Washington, D.C.

Back