We are five miles off the New Jersey beaches — although the water is only 50 feet deep — the twin outboard motors idling as our boat rolls in the swell. Somewhere beneath us a cluster of scientific instruments takes continual measurements of what is happening on the bottom, sending the information through a buried fiber-optic cable to a computer room in an old Coast Guard station. Also down there, in water so turbid that visibility is only three to four feet, is a docking cone for what looks like a small torpedo; the vehicle is packed with more scientific instruments and can roam 12 miles farther out to sea and back, taking measurements all the way.

The idea is that the "torpedo" will move into the docking cone like a shuttle rendezvousing with the Mir space station, except that acoustic signals are used to align torpedo and cone. Once docked, the vehicle will disgorge all its data ashore via the same cable, then recharge its batteries before setting off again. Aboard a 24-foot Zodiac known as the Minnow, we learn by radio from the shore station where the vehicle is and what it is doing. This day it is having trouble: it does not get far enough away from the cone to give itself enough time to line up when it heads for the cone again, and it keeps straying off to the south of the proper alignment. It corrects as it approaches, but never quite enough.

Roger Stokey, an engineer from the Woods Hole Oceanographic Institution in Massachusetts, talks by radio to the graduate student ashore whose program the torpedo will be using today. At one point they give up for the moment and bring the torpedo to the surface with an acoustic signal. Stokey and a colleague swing it into the boat and plug a modem into it; a thick black cable receives a new program over the Minnow's radio modem and loads it into the torpedo's computers. The white-and-yellow vehicle is then swung back into the water; a propeller on its tail spins, and it drives off and disappears. Alas, the vehicle still has trouble lining up with the cone, and finally we send five pings, it surfaces, and we scoop it up and head for home.

The instrument package on the bottom, known as Node A, lies southeast of Tuckerton, offshore from Great Bay, the estuary at the mouth of the Mullica River, which in turn brings exceptionally unpolluted water from the New Jersey Pinelands National Reserve. A second scientific station lies farther out at Node B; extending out from Node B is a line of six buoys that convert acoustic signals from the torpedo, heard underwater, into airborne radio transmissions, so that the shore station always knows where the torpedo is.

The system of instruments on the bottom and in the shore lab is known as LEO-15 (Long-term Ecosystem Observatory at 15 meters depth) and is largely funded by the National Science Foundation. The instruments are operated by New Jersey's Rutgers University Institute of Marine and Coastal Sciences, headed by J. Frederick Grassle. The torpedo, an autonomous underwater vehicle known as a REMUS (Remote Environmental Monitoring UnitS), is the work of the Oceanographic Systems Laboratory at Woods Hole, led by Christopher von Alt. The goal is to produce a picture of the underwater environment as detailed as that which a human has of his or her terrestrial environment, with the information presented on time and distance scales important to individual organisms.

LEO-15 transmits video, sound and data on light, temperature and salinity levels, currents, wave height and period, sediment transport, plankton blooms and a variety of chemical characteristics. (Continual observations, at frequencies from seconds to decades, are possible, Grassle says.) LEO-15 uses acoustic Doppler current profilers to measure currents at intervals of 10 to 20 inches from the bottom to the surface. Measurements at the surface are augmented by data from a 230-foot meterological tower, Doppler radar and satellites. Ken Able, director of the Tuckerton station, is correlating current data with his work on the movements of juvenile and adult fish in and out of the estuary.

Grassle has been an innovator before. He led the first biological expedition to hydrothermal vents in 1979. And he conducted the most intensive census yet of the very uncharismatic worms and other invertebrates in the mud off the New Jersey and Delaware coasts. He was able to show that even though environmental conditions appeared much the same throughout, the number of new species per mile suggests a biodiversity nearly as high as that of coral reefs and tropical rain forests. His main research now is focused on the communities found on the continental shelf at LEO-15.

From the LEO-15 data, scientists will be able to make physical and biological forecasts analogous to those their counterparts make on land. Last July the oceanographers at Tuckerton were delighted to hear over a local television station their forecast for a cold-water upwelling along that section of the coast. (The LEO-15 instruments had detected water from the outer continental shelf — colder than the coastal surface water by nearly 30 degrees Fahrenheit — moving in along the bottom to take the place of water blown offshore by winds out of the south and southwest.) Upwellings are important because they bring with them very high nutrient levels, which can lead to phytoplankton blooms in the water and eventual oxygen depletion as bacteria graze on the dead microscopic plants. In 1976, a low-oxygen event killed millions of surf clams, and caused half a billion dollars' worth of real and potential losses to the shellfishing industry. Pollution was suspected at first, but oceanographers now believe persistent upwelling during an unusually hot summer led to phytoplankton blooms and then hypoxia. Since 1976 oceanographers have found that low-oxygen events tend to occur in evenly spaced places along the Jersey coast where the bottom topography traps the upwelling water, creating a circular surface current called a gyre.

The LEO-15 instruments include optical sensors to detect florescence in phytoplankton. Oscar Schofield of Rutgers, the "plankton guy," can determine not only the density of phytoplankton but its identity. As early as the year 2000, a LEO-like observatory set up in the Gulf of Maine will begin trying to predict red tides, the sudden bloom in the water of a kind of microscopic organism that can kill marine creatures and injure humans. Despite satellites, we still sometimes learn of red tides only when people call their doctors.

People are also sometimes the detectors for a far more deadly presence in the sea. Among the Navy's methods for finding mines in potential landing-craft lanes in very shallow water is the use of Navy SEALS, who roll out of rubber boats in the dark of night, swim along possible routes and ever so carefully map the explosives. Chris von Alt of Woods Hole is working with the Naval Special Warefare Command to demonstrate that an adapted REMUS vehicle could do some of the swimming now done by humans. Far more exotic plans for REMUS-type vehicles are in the works. Jupiter's moon Europa may have an ocean beneath a layer of ice (Smithsonian, May 1997). An autonomous underwater vehicle is just what NASA will need when the time comes to explore the Europan ocean. Von Alt thinks of the work at Tuckerton in part as "ground-truthing" REMUS; the next step toward Europa will be to test new types of REMUS vehicles in Lake Vostok in Antarctica.

LEO-15 is a first-generation instrument system, easily serviced by divers. Once all the bugs are worked out, the Rutgers group plans more ambitious observatories, at 100 feet, 330 feet and 8,200 feet down. Similar stations are being installed elsewhere: a global Seafloor Observatory is taking shape, especially in the waters off Japan and the West Coast of the United States, where packages of seismographic and oceanographic instruments placed on the bottom are serviced by remotely operated vehicles. One of the latest is a package put on the bottom of Monterey Bay, 3,300 feet down, by an international collaboration led by the Monterey Bay Aquarium Research Institute in Moss Landing, California.

In the LEO-15 room at the shore station, computer monitors report real-time conditions out at Node B, and others track whichever REMUS is in motion at the moment. In another room, computer wizards turn mountains of data into "movies" of cold-water upwellings along New Jersey beaches. Von Alt prepares to return to Woods Hole; Grassle packs for a trip to Washington, D.C. to testify before a House subcommittee. The 48-foot Arabella, the 30-foot Caleta and the chartered Northstar 4 return to their docks after spending the day lowering and towing instruments to supplement the LEO-15 data. After censusing fish populations in the back bays and the Mullica, five small boats unload their nets. Scientists, graduate students and postdocs, and undergraduate interns joke about the trials and tribulations of another day at sea. Little by little they disperse, walking down a long causeway to where their cars are parked. The station quiets. Five miles out at sea, however, odd-looking scientific instruments continue to pour data down the fiber-optic pipeline. LEO-15 never sleeps.

By John P. Wiley, Jr.