An Introduction in 2008 to

DOMES This page was modified on
A Distributed Observatory
Many Earth Sciences
by Douglas Moreman

The paths in 2005 of hurricanes Katrina (on the left) and Rita through the oil platforms (the many black dots) in the Gulf of Mexico.

The DOMES question was this:
What, for science, can be built upon the oil platforms of the Gulf of Mexico by installing on each of them a suite of sensors, a small computer, and a small radio? This question was asked, around 2006, of someone at one Louisiana university of someone at another. As the question was asked of students and teachers and people off-campus, the reply grew increasingly multifaceted and wonderful and, remained impractical. But recent events in the Gulf seem to have likely changed the calculations of what is practical. Today we have a window of opportunity to attract the ears of those who will form the social components of the observatory: owners of oil platforms, politicians, scientists, and engineers. The answer to that question can be: "The Largest Observatory the World Has Ever Seen" - but of earthly and not of astral phenomena.

Three thousand, nine hundred and ninety (more or less) oil platforms stand in coastal waters of the United States' part of the Gulf of Mexico. One can sail along the 900 or so miles of that coast from Alabama's border with Florida to the border of Texas with Mexico and never be out of sight of such a structure. In the Gulf, natural events unfold over vast regions - usually slowly, but sometimes with the speed of a hurricane. A number of large and important phenomena are natural subjects for a new "observatory." Consider just three subjects of expanded science : weather, "mud-tectonics,", and large populations of valuable and fascinating creatures.

Experience suggests that most Americans of the Gulf, asked what the observatory could be used to study, respond immediately with "hurricanes." Distributed from Florida to, and eventually along the coast of, Mexico, the observatory's weather sensors could nearly encircle a hurricane in the Gulf. The data coming in from a storm in the Gulf could be fed to a "grid" of super-computers (that already exists) running predictive models. The extra data gathered during a hurricane's approach will improve predictions of location, time, and strength of impact. The extra information gathered during hurricanes will allow computer-using scientists to improve the modeling of these monster storms. Between storms, the meteorological sensors can help improve understanding and forecasting of day-to-day weather. Sensors never before used in predicting weather will increase the kinds of, as well as the quantity of, information. For example, acoustic sensors on legs of platforms can allow, for the first time, triangulation of precise locations of lightning strikes. The distribution and timing of bolts might allow inference of patterns of drafts in a storm overhead, perhaps teaching us a new indicator of the probability of a tornado. The observatory will allow study of the complete life of a thunderstorm in a way not now possible anywhere in the world.

Water carries sound much faster and with less attentuation than does air, with which we are more familiar. Sounds can be digitized and fed into computers that classify them. Sources of sounds can be computed by "triconication," that is analogous to triangulation but useS conic surfaces in its math.

Proven, but not yet deployed, acoustic methods can image schools of shrimp, or mile-long schools of commercially valuable fish. Rather then occasional deployment on a pair of boats, the system can be deployed permanently on platforms. Rather than the present occasional, short-term study, the system can continuously follow populations. Imaging will allow several schools' numbers to be estimated, their behavior studied, their health monitored.
Some individual generators of sounds can be indentified - they have "signatures." Each dolphin has its identifying "signature-whistle." Even individual machines can be identified by their sounds. This can be done by computers that keep track of locations of individuals.
Of immense sentimental or intellectual value by many Americans, dolphins and several species of their relatives (including sperm whales) live most of their lives withing 20 miles of an oil platform in the Gulf of Mexico. Acoustic sensors on the legs of platforms will feed the sounds of the sea into the computer network. Our computer system can be programmed to learn, over time, the calls of the individual dolphins and, so, to track them throughout most of their lives. Following dolphins for even a few of days has seldom been possible before. Motion of dolphins along the Gulf coast and occasional close-up television shots can be shown on the Internet. Students of these creatures will be attracted by the studies made possible; but, there might be direct practicalities. Perhaps a knowledge of the wide-spread movements of dolphins will aid in understanding movements among commercially valuable populations of fish. Perhaps dolphins or even noisy shrimp, by being absent from some region, will signal the presence of a leak (natural or not) of oil from the earth below.

It is not widely known that a natural oil leak on the order of that of the BP leak of 2010 was reported in a special issue of Scientific American in 1910. There was no television to show close-ups of oil to everyone and no national governmental presence to treat the leak as crisis worthy of massive payments in reparation. There was less fishing going on and the beaches had no where near the number of human visitors as today.
Oil has been leaking, naturally, into the Gulf of Mexico for perhaps millions of years. Lesser oil leaks provided tar, washed up on beaches in sufficient quantities that it was used for caulking boats by Indians and, ships by Spaniards.
Sensors on the legs of oil-platforms might allow the study of both natural and man-caused leaks, some of which might fail to reach the surface.

Many dangers threaten great potential damage, but with unknown immediacy. Asteroid impacts and the next great quake of the New Madrid fault under the Mississippi River are dangers known to many Americans. Statistics of the larger meteoroid-impacts in the Gulf might, perhaps, be gathered via our acoustic sensors and triconication. Seismic sensors provided by drilling firms might be aimed at and monitor various faults. But unknown to many, even among those doing business in the Gulf of Mexico, is another danger: that of a catastrophic avalanche of mud on the long slope of the submerged delta of the great river. Around the world, gashes in the deltas of major rivers seem to have been scoured in catastrophic avalanches of mud. The stronger hurricanes are known to trigger slides in the Gulf of Mexico because some of these have damaged petroleum pipelines. A mega-gouger, as indicated in "fossil" canyons, but not yet recorded in history, might rip away the top structures of several wells in one go and make the BP leak of 2010 seem small. It seems to be unknown whether a tsunami could result from any probable slide. An improved science of mudslides on continental slopes might better alert us to real danger and improve planning of petroleum structures so as to reduce the chances of calamity.

Obstacles exist which could abort the observatory before its birth or doom it to mediocrity in its childhood. Two categories of challenge are Technical and Permissions.

The technical difficulties begin with the need for a new, or at least the latest, self-organizing, adaptive network of computers. An easier example to understand, yet representative of many, comes from the "Dead Zone" on the continental shelf westward from the delta of the great river. In most summers, oxygen levels drop so low, in the bottom of waters 10 to 200 feet deep, that fish abandon the area. At least one study by ship, lasting a week or two, maps the low oxygenated region each year. This region lies entirely in amongst oil platforms. It is natural to think that oxygen sensors on those platforms could provide for continuous monitoring of the oxygen levels at various depths. But, oxygen sensors last only about a month between cleanings to remove organic growths. A visit to an oil platform requires a boat and crew and are expensive, relative to small jobs. A lesson in this is that the gear of the observatory must, it seems, be designed so that costs of maintenance are practical. Simple sensors of salinity might last longer and proved information useful in the study of hypoxia - which tends to stay below the "pycnocline" boundary between saltier water below and fresher water above. Underwater acoustic sensors, good for studying animals, boats, and weather also last longer than do oxygen sensors before organic growth take them down. Developing appropriate sensors will require new engineering. Twenty two years ago David Packard, a co-founder of Hewlett-Packard Corp and a former Secretary of Defense, brought engineering into intimate association with marine science through his Monterey Bay Aquarium Research Institute. He housed some of each of these two breeds of technologist in the same building and charged the engineers with inventively supporting the scientists. Such an organization, given some business guidance, might serve the economies of the states that participate in the observatory.

Platforms in the Gulf are private property. Granting access to ones property entails liability risks. Strong positive incentives will be needed to overcome the risks. For large oil corporations today, the subtly enhanced good will of public and government might nearly over-ride prudent fear of risk. Governments can provide financial incentives but possibly of more importance, the Federal Government can guarantee owners participating in the observatory:
* the right to withdraw their permissions.
* insurance against liability,
* protection against unexpected, non-trivial costs, and
* that the real business of the owners will never be hampered.

A team of graduate students (about 4 from Andra Pradesh, one from Mumbai, one from China, and one from Botswana), a varying number of American undergraduates, and Dr. Moreman worked on the following:
Suppose there is, distributed over some part of the Earth (such as, on the oil platforms of the GoM) some "pods." Each "pod" houses a computer equivalent in power to a laptop PC of the day, a radio with a line-of-sight capability, and some sensors. All of the pods have exactly the same software, which we are to design. At least one of the pods has a means to communicate with some other network, G, of computers (such as the "Grid" worked on at LSU in Baton Rouge). Design the software so that the system will "wake up" at some designated time and organize itself so that data will flow to G from the sensors of our network. The system would re-organize after hurricanes. While every node runs a copy of the one software, differences in behavior of nodes can result from their perceived facts about themselves. Our approach to this problem (you can approach it however you want to), did NOT include reading what other people had written about it. We invented everything from scratch. And were encouraged by our results.

DOMES, as discussed above, is a specific, possible instance of a more general idea. In other forms, an observatory might be distributed on cell-phone towers and even on cellphones.