Earth scientists, inundated by volumes of data from space satellites, will need whole new strategy to process the additional one-trillion bits per day expected from NASA's Earth Observing System (Eos) in the 1990s, two NASA scientists said.
Ralph Kahn of the Jet Propulsion Laboratory in Pasadena, Calif., and Henning Leidecker of Goddard Space Flight Center in Greenbelt, Md., have laid out strategy for answering the Eos data and information challenge.
To illustrate the problem, Kahn and Leidecker said it would take 10,000 Washington, D.C. telephone books (white pages) to match the amount of data from one Eos day.
Put another way, the Eos archive will accumulate as much data about every few weeks as is currently stored in the 19 million bound books in the Library of Congress.
The two scientists outlined their program in detailed paper, entitled "The Crush of New Data Knocking at Our Door," published in the winter edition of Renewable Resources Journal.
The Eos project includes polar-orbiting, instrument-laden platforms that will sense ocean-atmosphere interactions, geologic processes, agricultural use, ice distribution, forestry and weather patterns among other studies for wide range of geophysical disciplines.
About 30 instruments are currently planned for the Eos system and some will have average data rates of 5 million bits to 20 million bits per second. The primary conduit of Eos information to Earth is currently through the Tracking and Data Relay Satellite System (TDRSS) which sets limit on the amount of data the Eos program can receive on Earth.
Still, it will take less than week for the Eos data sets to exceed the sum of all present Earth science data bases.
As result of the wealth of information EOS will make available, the project will include Data and Information System which is "unprecedented in its scope," the scientists said
Fortunately, they said, storage technology continues to develop rapidly and there are hopes that new developments will reduce the size of the Eos data storage problem to manageable proportions.
Some access to the data bases will be through electronic computer networks. Data sets one-million bits in size easily move through the networks now in place, the scientists said.
"While the operational tasks posed by Eos are challenging, they do not seem impossible to perform in the Eos time frame (by the mid-1990s)," Kahn and Leidecker said.
Performing the science analyses task with all the Eos data may prove even more challenging. Three-dimensional display of data objects is envisioned in the new data analysis environment using technology which in some aspects has existed since the late 1960s, but which has not yet been put to use for these purposes.
The scientists said one aspect of the display would include matching the speed of the display to human perception rates because processes that are several times faster or several times slower are not perceived.
They said an example is time-lapse photography which has opened to human perceptions world of slow processes such as the blooming of flower. Also, high-speed photography has shown how the wings of birds actually function in flight.
These technologies are technically demanding and expensive, they said, and the display speed is usually fixed in advance so it cannot conveniently be varied over wide range.
But combination of computer and video technologies now in development promises relatively inexpensive displays capable of wide degree of time-scaling.
"We can expect these will be as productive of discoveries in the 'time domain' as microscopes and telescopes have been in the 'space domain'," they said.
One additional problem in the Eos data stream planning, they said, is policy, which includes getting the data properly archived and made accessible to users across wide spectrum of disciplines. That means common lexicon of terms and keywords, and directory that any researcher interested in Eos data can access with relative ease.
"A user should be able to search many databases without having to find new abbreviation for 'latitude' for example, in order to access each one," they said. The computer language must be as recognizable to forestry scientist as to geologist, and to an agronomist as to an oceanographer.
Work on the paper was performed by Ralph Kahn, staff scientist at the Jet Propulsion Laboratory, California Institute of Technology, under contract with NASA, and by Henning Leidecker, materials scientist at NASA's Goddard Space Flight Center.
818-354-5011