The Natural History of the Present, Chapter 18

Chapter 18: The Ocean

So we come to the ocean, into which everything flows. The sea is enormous and delicate. Ocean waters vary in temperature, density, salinity, the presence of faint electrical fields, dissolved gases and minerals, scents: subtle chemical variations over horizontal and vertical scales. Sea turtles may follow the scent of specific ocean basins, as well as the earth’s magnetic lines, in their thousands-mile migrations along ocean currents. The Gulf Stream travels up to 3.5 miles per hour and is 2000 times the flow of the Misissippi. Young fish and larvae from Caribbean waters are caught up in it and deposited by its turbulent swirls along the south coast of Long Island, where they don’t survive the winter. Deep waters upwell as currents pass over seamounts, or collide with coasts, fertilising the sunlit surface, whose organisms are the basis of their productivity. Where southward flowing (or in the Southern Hemisphere northward flowing) nutrient-rich coldwater currents meet oppositely flowing warm ones, plankton grows—an engine of productivity that runs most of the year off the coasts of Maine and Atlantic Canada, where the southward flowing Labrador Current meets the warm waters of the Gulf Stream—and fish (along with seabirds, turtles and marine mammals) thrive; forage fish and invertebrates like squid feed on the plankton and are food for cod, whales, seals, swordfish, bluefin tuna, porpoises and Atlantic salmon; strong swimmers like tuna cross huge expanses of open ocean to reach such places. Flows of fresh water also deliver nutrients to surface waters — the Amazon off South America, the Mississippi in the Gulf of Mexico, the Niger and the Congo in the Gulf of Guinea; and before they were dammed, the Yellow River in the Yellow sea, the Nile in the eastern Mediterranean, the Colorado in the Gulf of California. Such places are also productive, but productivity varies with the flow of fresh water. The meeting places of currents shift: marine reserves that protected the meeting places of currents (or the migrations of turtles) would have to move with them. The sea is also noisy. Whales sing low songs across great distances, the noise of sea urchins grazing on rocky reefs peaks just before dawn and just after dusk, a chorus of shouting fish greets the dawn off the California coast.

Biologists divide the sea into biogeographical provinces that depend on local processes delivering nutrients to sunlit surface waters; but the boundaries of such provinces are dynamic. A large wave breaking over a mud flat removes the chemical signature that attracts the larvae of the sedentary animals of the flat (polychaete worms, clams) to settle, dig burrows, build feeding tubes, oxygenate the sediments, grow. Without the presence of that chemical scent, the larvae drift off elsewhere. Larvae must settle within a given period of time (their development is time-limited). Thus chance is involved in the colonization of a flat; but the signature is restored in a day. Rivers bring down silt, sand and other nutrients into estuaries; the flow of fresh water, against the tides, determines the level of salinity in a given spot. Salinity moves up and downstream with the tides and seasons and is used as a developmental cue by many plants and animals. How deep the mud flats lie below mean high tide is determined by silt loads and tidal flushing. The water depth over the flats determines what animals live there. Rapid siltation raises the flats, suffocates the existing benthos and means different suites of animals (or for a time, none), a slow re-colonization, a gap in the food chain. The silt sent down the Sacramento and San Joaquin Rivers by the gold miners of 1849-1852 raised the bed of the Sacramento River by 10 to 30 feet and reshaped much of San Francisco Bay. Hundreds of square miles of farmland along both rivers was flooded, as was downtown Sacramento. As a result hydraulic mining was curtailed, but not before most of the gold was gone. (To date about 35% of the bay has been lost to sedimentation and land reclamation.)

Ditching and flood-gating for mosquito control on East Coast marshes in the 1930s also changed marsh habitat. A major problem was the pile of spoil left at the side of the ditch by the dredges, which prevented complete drainage of the marsh at low tide. (New dredges that send the spoil flying out over the marsh avoid this.) In the stagnant pools, reeds began to replace the stands of smooth cordgrass, which was eaten by wintering ducks and geese. Detritus from the grass, whose beds delimited the reach of high tide, fed micro-organisms, plankton, fish and crabs; the reed beds were much less useful to the birds. About 50% of East Coast marshes were lost to invasive phragmites through draining and diking. Similarly, in San Francisco Bay many organisms have been introduced, either intentionally, such as striped bass and shad, or unintentionally: the latter often arrive in the ballast water of ships, which ought to be pumped out and exchanged in midocean, but which, because of the time involved (a cost to shipping companies), is not. Half the fish in San Francisco Bay are alien species along with the majority of the plants and animals on the bay’s floor. In parts of San Francisco Bay 99% of the plants and animals are non-native. In healthy ecosystems, most introduced aliens find a small niche among the natives; some however, released from their usual predators and parasites, find a competitive advantage, expand geometrically, and take over the habitat; an event more common in stressed ecosystems.

In the late 1940s and 1950s, DDT, sprayed to control mosquitoes, tremendously reduced the crustaceans (crabs, shellfish) and other invertebrates of the marshes, whose populations never really recovered. The oceans and their top predators, the marine mammals, are sinks for DDT and other persistant organic pollutants. DDT was banned in most industrial countries in the 1970s but is still manufactured in the U.S. and is used in many underdeveloped countries. The DDT burden in marine mammals fell through the 1980s, but then leveled off, as the chemical was recycled in the environment and as new supplies eroded off uplands, seeped out of dumps, condensed out of the air (from current use), was released in melting glacial ice, or rose on currents from the depths of the sea. Levels of persistant organic chemicals in male marine mammals tend to rise throughout their lives but those in female mammals level off when they reach reproductive age and begin transferring the chemicals to their offspring during pregnancy and nursing. Persistant organic pollutants such as DDT and the PCBs have been linked to reproductive failure, lowered immune function, and skeletal abnormalities in marine mammals such as seals, walrus, whales and dolphins. Lowered immune function probably explains recent massive die offs of dolphins along the East Coast of the United States, and of seals in the Mediterranean and North Sea, from infective viruses.

The land meets the sea in coral reefs, salt marshes, sea grass meadows, mangrove forests, beds of kelp. These aquatic forests and grasslands are the nurseries of the sea. Along the Gulf of Maine a brown kelp forest grew near the coast, while further out the rocky bottom was covered with shaggy red algae. Kelp forests are maintained by a balance between the invertebrates that graze on them, such as sea urchins, and the animals that eat the grazers (sea otters, cod, lobsters). If predation pressure on the grazers is reduced by trapping out sea otters or by overfishing cod, the kelp forests disappear. Kelp forests provide nursery habitat and shelter for many fish and marine invertebrates. Coral reefs lie between land and sea in nutrient-poor tropical and sub-tropical waters, drawing in nutrients from land and sea in a very efficient recycling system. A symbiotic algae that lives in the coral polyp secretes sugars for the polyp, whose nitrogen rich wastes are used by the algae. The nitrogen, a building block in all proteins, came originally from land or from nitrogen-fixing cyanobacteria floating on the ocean’s surface. Perhaps as a result of their low-nutrient regime, which encourages the exploitation of small niches, coral reefs contain 25% of marine species.

Mangrove forests and coral reefs, like the meadows and islands in the Mississippi Delta in the Gulf of Mexico, protect the coast from high tides and storm surges. Mangroves are as effective as sea walls costing $300,000 or more per kilometer; are an important nursery for fish; and provide approximately10% of the organic carbon that enters the ocean from the land. This carbon is resistant to breakdown and so constitutes carbon removed from the atmosphere and withheld from the oceans; carbon put in medium-term storage (making mangrove forests candidates for a carbon storage payment). Such coastal areas are often used for aquaculture. Mangrove forests are cut to dig shrimp ponds, replacing the natural environment of coastal wetlands, deltas, lagoons, tidal flats, with man-made brackish ponds one third of whose water (half of that fresh) must be replaced every day. So levels of fresh ground water in coastal regions with shrimp ponds drop rapidly. As fresh water is pumped out, salt water intrudes into the aquifers. Waste water from the ponds, full of nutrients and chemicals, contaminates surface and ground water, overfertilizes offshore coral reefs and ruins coastal fisheries. While the existing habitat of coral reefs and mangrove forests are capable of self-sustaining production of timber, fish, sand and rock if not overexploited, shrimp ponds, because of the build-up of shrimp diseases, are abandoned after five years, leaving an unproductive landscape behind. Mangrove forests are also cleared for shoreline development, the stands of trees (which grow in coastal shallows) replaced with beaches. About half of coastal mangrove forests worldwide have been cut (about 70% in the Phillipines). The benefits of mangrove forests are not linear; they can be exploited. Perhaps 20-30% of a stand can be cut without losing its benefits (as a fish nursery, source of carbon, protection from storms).

Coral reefs, situated between the intertidal and the deep sea, mediate between land and sea, concentrating nutrients in a low-nutrient environment. Built of calcium, they do well in waters supersaturated with that mineral. When overwhelmed by nutrient runoff from the land, they turn into algae-covered rocks. The algae grow over the corals, smothering them. Sugars from the algae fertilise pathogenic bacteria that infect the corals and kill them. (The bacteria are always present but multiply more rapidly in the presence of algal sugars.) Algae have always been present on coral reefs but reefs were formerly grazed of algae much more heavily by turtles and fish. Between 35 and 100 million green sea turtles once grazed Caribbean coal reefs, accompanied by 33 to 39 million hawksbill turtles. Adult green sea turtles turtles weigh 220 to 500 pounds. They eat crustaceans, seaweeds, starfish and mollusks. They move slowly in shallow water and like to bask in the sun near the high tide line, waiting to be refloated by the tide. Their grazing, along with that of herbivorous fish, especially the larger fish, keep the corals clean of algae. The grazers (especially the fish) recycle the productivity of the reef. As nutrient runoff over the reefs increased during the twentieth century, there weren’t enough grazers to keep the algae in check. By the 1750s most of the turtles in the Caribbean had been fished for food for slaves on the islands’ sugar plantations. Manatees, the other large (and docile) vertebrate of Caribbean reef systems, were more or less gone as organisms of ecological importance by 1800. Both animals also ate seagrass and so helped maintain the shallow underwater meadows that serve as fish nurseries. But the reefs maintained themselves until modern times, when increased nutrients running off the land, along with increased fishing pressure, overwhelmed them. An example of such reef deterioration comes from Jamaica. As prosperity increased in Jamaica after the Second World War, live coral near the north coast of the island fell from 60% of the reef in the 1960s to less than 5% in 2000. Kelp and other seaweeds grew over the coral. The grazing turtles were mostly gone, algae eating sea urchins died of a disease in the 1980s, grazing reef fish were fished more and more intensively for restaurants. A market for sharks developed in Asia. Fishing for sharks (which reproduce slowly) let more mid-sized predatory fish like groupers increase; the groupers ate the smaller grazers (the smaller parrot-fish); at the same time the larger grazing fish (the larger species of parrotfish, whch might have helped control the algae) and the groupers were still fished down to supply restaurants. The seaweeds were nourished by silt and nutrients washing off the land from golf courses, resort development, sewage. A final effect in this cascade is a continuing loss of beach sand. Sand in the Caribbean is renewed by several species of coralline algae that assemble carbonate grains out of seawater; when the algae die, the grains become sand that help renew the beaches of white carbonate sand. Such algae are adapted to low-nutrient situations and are overgrown by bacterial slimes in the presence of too many nutrients. So Jamaican beaches also lose sand.

The effects of overfishing can be reversed through the use of ocean reserves, where fishing is limited or forbidden. In protected Bahamian reefs, groupers increased by 7 times and ate the smaller parrot fish, those less than 6 inches, reducing their numbers considerably. Large parrot fish, too big for the groupers to swallow, increased in numbers and (being larger) caused a net doubling of reef grazing. The result was a four-fold reduction in the seaweed on the reef. Jamaica is of course not alone in its loss of marine habitat. Deforestation in the highlands of the Dominican Republic sends silt into Samara Bay, degrading one of the most important fish nurseries in the Caribbean. The estuary, the Caribbean’s largest, produces 40% of the fish catch of the Dominican Republic and is a sanctuary for humpback whales. Silt also runs off the coast of southern Florida, smothering its coral reefs; water saturated with lawn fertiliser and septic tank effluent seeps into Florida bays, over-fertilising them and the reefs that edge them, with the result that many, if not most, Florida reefs are dying or dead.

The pastures of the sea are microbial. Despite their vertical extent, compared to terrestrial pastures they are sparse: standing biomass of terrestrial plant life is 200 times that of marine plants. (But the mass of microbial life on the seafloor may dwarf that of terrestrial plants.) Cyanobacteria (formerly blue-green algae, a member of the pico-plankton) are among the chief photosynthesizers and cycle both carbon and nitrogen. They are eaten by single-celled animals (the protists) or filtered out by shellfish. Among the known bacteria and viruses floating on the surface of the sea are thousands or tens of thousands of unknown species, most in very small numbers, that represent vast stores of genetic diversity, with the ability to take over planetary functions after massive global changes (increases or decreases in temperature, changes in the relative abundance of atmospheric gases), that increase their competitive advantage. (One theory of past extinctions is that catastrophic changes in climate or in the chemical composition of the atmosphere, let microbes take over much of the planet, replacing—outcompeting—for a time multicellular organisms.) Photosynthetic activity occurs throughout the sunlit 600 feet of the upper ocean but much it occurs on the water’s thin surface skin. Fish also concentrate in the top 600 feet of the sea, especially in the area over the continental shelves. About 90% of fish are caught here and catches are two orders of magnitude (a hundred times) over those in the open ocean. Winds drive currents, and upwelling of nutrient-rich deep waters along coasts or seamounts or in the wake of hurricanes and typhoons, as well as the daily action of the tides, returns the nutrients depleted by the photosynthesizers to surface waters. Winter storms stir up nutrients from the bottoms of shallow seas, such as Europe’s North Sea, setting the stage for the plankton growth of spring. Much of the deep upper ocean, constantly depleted by microbial photosynthesis, is nutrient-poor. Populations of microbes respond rapidly to changes in the nutrient situation, expanding and contracting with changes in their food supply. Domoic acid is a neurotoxin produced by a photosynthetic diatom that accumulates harmlessly in fish and shellfish but kills seabirds and mammals (sea lions, whales, dolphins, people). Increasing amounts of urea (from fertiliser and sewage) and copper (from boat paint) in seawater, along with rising water temperatures, seem to stimulate the algae to produce more of the chemical. Domoic acid first was found off the California coast in 1991 and has appeared regularly since 2001. Thousands of sea mammals have died and some fish and shellfish are unsafe to eat. Similarly, increasing nutrients and warming temperatures increase the relative numbers of the dinoflagelates that cause red tides, toxic to people and fish; and increase the incidence of the fish-eating bacteria Pfiesteria piscida that causes fish kills in east coast rivers polluted by run-off from hog farms. (Pfiesteria toxins also affect the nervous system of people.) The micro-organisms are responding to environmental changes that favor their competitive position.

The land is the ultimate source of nutrients for the sea. Estuaries, which receive relatively large volumes of river-borne nutrients, are the only marine biome that competes with the land in productivity. Two-thirds of marine life begins in shallow coastal waters. (Perhaps 98% of commercially important fish in the Gulf of Mexico begin their lives in the gulf’s estuaries.) Estuarine plants and animals are adapted to seasonal changes in the levels of silt, nutrients, salinity, pulses of fresh water. Developing the terrestrial landscape upstream changes the levels and timing of the pulses of nutrients and fresh water and the old systems break down. Florida Bay, a major nursery for fish of the Gulf of Mexico, is fed by the Everglades, the great swampland, in fact a shallow river six inches deep and 40 to 60 miles wide, with a drop of 2-3 inches a mile, that flowed the 80 miles from Lake Okeechobee (15 feet above sealevel) across southern Florida to Florida Bay. Wet prairies bordered the sawgrass flats. On the west coast the land slid down into coastal mangrove forests fed by the water, while on the slightly higher and rocky east coast the land met the sea in barrier islands. The sawgrass rooted in the marl precipitated out of the calcium rich water (limestone lies under the Everglades) by the periphyton (a mix of algae and some zooplankton) that lived on the roots of the grass. The dead leaves of the sawgrass formed peat, in which bay trees and willows sometimes rooted. These plants produced more peat, whose acid decay dissolved the limestone and deepened the pools in which they lived, forming bayheads and willow heads. The pools varied the habitat and stored water for the dry season. Deeper pools were occupied by cypress trees. Mounds of peat that accumulated above water level, sometimes thrown up by alligators weeding their pools, became dry hummocks with pines, hardwood trees and palms. Waterlevels in the Everglades deepened in autumn as the summer rains from central Florida filtered into them, letting their fish populations grow, and delivering a slow pulse of fresh water to Florida Bay, and then fell in winter, concentrating the fish in small pools, where they were easy prey for nesting wading birds. Most of the pools were dug by alligators as refuges for the dry season. (The modern Everglades would probably not have worked without alligators.) Water from the chain of marshes and lakes in central Florida flowed into Lake Okeechobee down the wide floodplain of the Kissimmee River, then spilled out of the lake into the Everglades. The Kissimmee Basin, Lake Okeechobee and the Everglades were a 9000 square mile hydrological system. The slow autumn pulse of fresh water maintained Florida Bay’s mangroves and seagrasses, and their associated bacteria, plankton and fish (all organisms adapted to a mix of salt and fresh water). The seagrass meadows were the nurseries for the growing fish. The Everglades are a low nutrient system, whose characteristic phytoplankton and sawgrasses are overtaken by other plants (especially cattails) when water levels are stabilized and nutrient levels in the water rise. The Everglades sawgrass is a plant of nutrient-poor, hydrologically unstable regimes. Its roots support the mix of algae and zooplankton—the paraphyton—that is the basis of the Everglades’ food chain. Shrimp graze on the paraphyton, fish eat the shrimp, and wading and diving birds and alligators eat the crustaceans and fish. Cattail roots do not support the layer of paraphyton.

Everglades National Park occupies 20% of the historic area of the swamp. Half the Everglades has been drained for agriculture, much of the rest diked off to use for housing for the expanding populations of southeast Florida. Phosphorus from fertiliser used on sugarcane grown on the several million acres of drained swampland near Lake Okeechobee, along with phosphorus and nitrogen from agriculture and from urban areas carried down the Kissimmee from central Florida, now enter the Everglades. After floods in 1928 Lake Okeechobee was leveed off and its excess water drained away into the Atlantic, or used for irrigation and water supply. Five million people depend on water from Lake Ocheechobee. The Kissimmee River, once a meandering 103 mile stream was straightened and turned into a series of five pools with locks, making the river usable for recreational motorboats and its formerly swampy floodplain for cattle pasture. Channelization destroyed the filtering capacity of the riverside marshes. The numbers of wintering waterfowl on the floodplain fell by 90%. (Channelization of the Kissimmee cost $35 million. Restoration, which began almost immediately, as the extent of the disaster to the Everglades became clear, cost $20,000 an acre or $512 million for restoring two-thirds of the former Kissimmee wetlands. But the restoration seems to have been successful.) Further diking and draining and the construction of a highway across the Everglades, turned the Everglades into a series of managed pools, rather than a continuous shallow flow. Water levels were more stable, which, along with the raised nutrient levels, favored cattails. In years when water is scarce, water is delivered to farms and cities rather than let flow into the Everglades. Water bird populations in south Florida have fallen 95% from the 1930s, and the fisheries that depend on Florida Bay have also fallen. The effects of prolonged water shortages and nutrient pollution are cumulative and in the early 1990s Florida Bay turned from a clear water ecosystem dominated by sea grasses and manatees to an ecosystem of turbid water dominated by algal blooms. The underwater meadows suffered extensive mortality from diseases. The meadows functioned as nurseries for fish of the gulf, and as habitat for organisms on which the small fish feed. Their roots oxygenate the muds and water. Excess nutrients and increasing salinity because of the lack of fresh water let algae outcompete the grasses, and their supporting suite of organisms that eat the algae. Some writers speculate that grazing (by sea turtles, waterfowl or manatees), which renews the meadows, may—especially in the present situation—be necessary for their long-term health. Green sea turtles favor the tender tips of young grasses and clip away and discard the tough older tips as part of their feeding behavior. This stimulates the sprouting of buds lower down on the stem and keeps the grasses fresh and productive. Such trimming may be especially useful under high-nutrient regimes. But green sea turtles and manatees are no longer present in numbers that matter ecologically; they are functionally absent.

About 75% of marine fish in the Gulf of Mexico begin life in the wetlands of the Mississippi Delta. Until the 1950s these wetlands grew, against the constant erosive action of the sea, from silt brought down by the river. Dams on the Mississippi’s tributaries, especially the Missouri, but also the Arkansas and the Red, reduced silt loads at the Mississippi Delta by half. (The Ohio, once a wide clear shallow stream, with mussels in its riffles—la Belle Rivière of the French—is now a narrow deep muddy one thanks to channelization, agricultural runoff and deforestation. It carries 10 times the silt of formerly, but its contribution doesn’t make up for that lost from the prairie streams.) A major shipping channel (rarely used) that led directly to New Orleans from the gulf, and 10,000 miles of oil exploration canals opened up the delta marshlands to wave erosion. As the tides penetrated further and further inland, the freshwater marshes turned more and more salty, and the land washed away. The freshwater marshes were feeding grounds for overwintering waterfowl. Some supported cypress and oak forests, which are more efficient than marshlands at reducing the height of storm surges and at protecting the coastline from storms. Sending silty Mississippi water, with its fertilisers, pesticides and industrial chemicals directly out into the gulf, instead of letting it flow over the natural river levee and into the Louisiana marshes, turned the silt and fertiliser into something toxic, rather than something that would build new land and grow useful biomass (trees, fish, crayfish, ducks). Those newly built marshes would further protect the shoreline and release cleaner water downstream. Cores from the seafloor off the Mississippi Delta show a long period of stability in algal biomass before the Europeans arrived (though for several hundred years of this time large sections of the Mississippi Valley near the river were being farmed by Native Americans), an increase in algal growth from the 1850s as the Middle West was settled by European Americans, then algal blooms and seafloor hypoxia on a regular basis from the mid-1950s on. Nitrogen use increased 6 times in the United States from 1955 to 1980 and the concentration of dissolved nitrogen and phosphorus in the Mississippi doubled. New levees after the Second World War narrowed the river’s floodplain and eliminated much riverside swampland. (The Mississippi is separated by levees from 90% of its floodplain.) The dead zone of summer hypoxia in the Gulf of Mexico doubled in size after the 1993 floods (which brought down huge amounts of nutrients) to 7000 square miles (an area about 80 by 85 miles). Excess nutrients accumulate in marine or lakebottom sediments and are recycled again and again into the water column, especially under low oxygen conditions, making such conditions somewhat intractable, but usually much more intractable in lakes than in the ocean, whose waters are churned by tides and storms and mixed by currents. Currently, perhaps afraid of criticisms that might follow damage from a major hurricane on a coastline less and less protected by marshlands, the Corps of Engineers is considering siphoning some of that silt-laden water over the levee into the marshes. Some may not help; and in order to protect the delta as a whole, some settled lands may have to be abandoned.

Nutrient-related seasonal hypoxias (areas of the ocean in which there is too little oxygen for oxygen-breathing organisms like fish and shellfish to survive) are now found at the mouth of the Mississippi, in large portions of Chesapeake Bay, about New York City (including the western end of Long Island Sound), in the Adriatic, the North and Baltic Seas, the Inland Sea of Japan, the Yellow Sea off the Chinese coast, and the Persian Gulf. There are about 200 such areas worldwide, most in places important for spawning marine fish. Nitrogen fertilisation, like disturbed soils or pioneer ecosystems, or rising levels of carbon dioxide, is a sign of human occupation. Dams on rivers that feed these seas reduce the silicaeous sand needed by diatoms (a base of the food chain), and shift plankton populations to the potentially more toxic dinoflagellates, which thrive on the increased nitrogen. Dams also block fish spawning runs, greatly reducing or eliminating many species and shortening the food chains of estuaries and bays. Shorter food chains, with their fewer prey species, make populations of predatory fish more vulnerable to yearly variations in the supply of plankton. (Plankton production is dependent on the weather. The catch of alewives, or river herring, in Long Island Sound in the early 2000s was 3% that of the 1960s. The causes of the decline are probably overfishing and the dams that block nearly every alewife nursery stream entering the sound. Alewives are forage fish for larger fish and an important food fish for ospreys, whose colonial nesting sites—some held several thousand birds—on Long Island have never really recovered from the birds’ poisoning with DDT.) Dams change the amount and timing of the pulses of fresh water to which the life cycles of fish are adjusted. Algal blooms from the excess nutrients smother seagrass beds and shut off their light. Diseases decimate the beds. Shellfish populations collapse from lack of oxygen and introduced or stress related diseases. The frequency of red tides from blooms of some species of dinoflagellates increases. Red tides occur naturally but in modern times are also a sign of excess nutrients (from land, from aquaculture, from shoreline reclamation) or of unusual warmth. A red tide off the west coast of Florida from January to October of 2005 covered 2000 square miles (an area 200 miles by 10 miles). The toxins in the dinoflagellates kill fish and the rotting fish deplete the oxygen in the water and supply more nutrients for the bacteria. In Florida, hundreds of tons of dead fish washed up on the beach. The toxins from the dinoflagellates come ashore on the wind, so people living near the coast experience respiratory problems (coughing, sneezing, itchy eyes, difficulty in breathing).

Human activity degrades marine habitat in many ways. Traffic from pleasure boats in coastal waters increases shoreline erosion. Turbidity from boat traffic is a major problem in busy estuaries. Large, fast ships increase underwater noise, a problem for some species, especially marine mammals. (It may be involved with beaching whales.) Underwater noise from ships has increased 10 times off southern California since the 1960s. Anti-fouling paints used on boats and docks to kill the marine organisms that colonize ships’ hulls or bore into docks, kill many non-target marine organisms and contaminate harbor sediments with tributyl tin. (Tributyl tin is also used as a slimicide in the cooling systems of power plants and so contaminantes the rivers from which power plants draw their water.) Industrial chemicals, oil, metals, and organochlorines wash off the land and also pollute harbors. Much of this ends up in the sediment, which in extreme cases (such as New Haven Harbor on Long Island Sound) becomes toxic to the burrowing organisms of the benthos, such as polychaete worms, which in healthy environments oxygenate the sediments and recycle their nutrients. The larvae of the worms may settle but in penetrating the sediment they ingest chemicals that kill them. Most oil pollution in North American coastal waters comes from land and not from the occasional (but spectacular) spills at sea, or from the 180 ships a year that sink at sea. (However chronic low-level releases of oil from vessel operations, including oil tankers, may have the greatest effect on seabirds, and similar small spills by the oil industry worldwide average more than one per day.) Of the 29 million gallons of oil that enters North American waters yearly, 85% comes from end-users, such as trucks, car owners and service stations, not from the oil industry. The oil is put down drains, washes off roads and is carried by rivers to the sea. The presence of oil reduces the survival of eggs and larvae of many marine species, including fish. Effects may be long-term. Herring populations in Prince William Sound collapsed four years after the Exxon Valdez spill there and fish living near old oil spills (several decades old, the oil, little degraded in the anoxic environment, now several inches down in the revegetated mud) show elevated levels of liver enzymes associated with chronic oil pollution. Minute amounts of polycyclic aromatic hydrocarbons (PAHs) in the oil cause deformities in developing fish embryos that kill many of them before they reach adulthood.

Tiny plastic pellets, feedstock for the plastics industry, dumped or lost at sea, are dispersed throughout the oceans, at a density of 1 to 4 per square meter. The pellets are mistaken for fish eggs or zooplankton by seabirds, sea turtles, and fish. The pellets cause ulceration in the stomachs of seabirds and reduce the functional capacity of their gizzards. PCBs and other persistant organochlorines are adsorbed on the surfaces of the pellets at concentrations up to a million times those in open water and ingested with them by fish and seabirds: the chemicals are then absorbed into their bodies (and absorbed into ours, if we eat them). Smaller pellets, so-called micro-litter, used in products like exfoliant creams, wash through sewers and rivers into the sea, are carried back onto the beach, where they clog the digestive systems of organisms like sand fleas and lugworms. Larger pieces of plastic, blown or washed off the land, or dumped with other refuse at sea, cover the ocean’s tropical gyres (approximately a quarter of the planet’s surface) with a mass weighing 6 times that of the zooplankton on the ocean’s surface—or more recently (2008), in the North Pacific Subtropical Gyre, a mass of water twice as extensive as the continental United States, weighing 46 times as much. Such bits—of toothbrushes, cigarette lighters, plastic containers, plastic bags, all of which also adsorb toxic hydrocarbons—are eaten by seabirds and turtles, who starve to death with their stomachs full of shards of plastic. As a writer has remarked, plastic, like salt, or hydrocarbons, is now a component of ocean water.

Increasing ultraviolet radiation (from the Arctic and Antarctic ozone holes) causes mortality among the juveniles of some species of fish, such as the northern anchovy, an important forage fish. Human disturbance limits animals that breed on the beach: sea turtles, terns, piping plovers. Harvesting horse-shoe crabs for fishing bait has tremendously lowered their population in Delaware Bay, once the center of their distribution on the North American East Coast. The crabs were trawled from the bottom of the bay and also collected by the truck load from the beach. The female crabs were sold as bait to commercial fisherman. Horse-shoe crab eggs, laid on the beach at spring high tides, in densities of 100,000 or so to the square meter, are a major food for shorebirds migrating north. Flocks of red knots (a shorebird) arrive on the beach after a 7000 mile flight over the ocean from the bulge of Brazil; the nonstop flight takes a week and the birds land in a state of exhaustion and hyperphagia, or extreme hunger: they need immediate, easily available food. Their feeding barely reduces the abundant supply of eggs but the eggs must be abundant for the birds to survive. (They also need to put on weight for the next leg of their trip north.) Despite all this, the nutrients that flow into estuaries from growing human settlements is one of their main problems; and overfishing the main problem for ocean fisheries.

* * *

“If you lose the hills, you lose the sea,” a scientist has remarked. The sea also has problems specific to itself. Early forms of trawl netting were opposed in England 600 years ago because of fear of damage to the ‘flower of the sea’: the plants and sessile animals that grow on the ocean bottom, waving blooms of pink, green and brown, in many nearshore locations. Living in these tangles were healthy populations of juvenile fish and of invertebrates, food for adult fish. As in aboriginal North America, fish in Carolingian Europe were caught by hook and line, in seines, in dip nets or traps. River fish, especially anadromous fish, which spend a major part of their lives in the sea, were first exploited, but by the 1050s mill dams and siltation from expanding agricultural settlement had reduced their habitat and, along with overfishing, their populations. Weirs and nets across rivers let people catch most of the migrating fish. At the same time the human population was growing and the demand for fish increasing. Diking the Rhine Delta eliminated much of the breeding habitat of North Sea sturgeon, once a mainstay of the northern European diet (studies of middens indicate sturgeon made up to 70% of the fish eaten in the Baltic States in the 700s). Christians were obliged to avoid meat 130 to 150 days a year, so fish was a major part of the diet of pious Europe. Fish production turned to fishponds (often established in the still water behind dams), and the sea, which was also fished with hook and line, traps (such as those for migrating tuna along the shores of the Mediterranean, which date from Phoenician times) and nets. The trawl was first mentioned in a complaint to the English king in 1376. The petitioners argued the trawl nets destroyed the plantations on the sea bottom and thus the little fish and other animals the big fish ate. While the trawl was spectacularly successful at catching fish, dragging nets behind ships destroyed the bottom habitat of rock outcroppings, boulders and cobbles; the shell-like structures of algae, worms, brachiopods and bryozoans; the beds of mollusks (now harvested by specialized trawl); the vertical structures of anemones, sea pens, cold-water corals, rooted algae. (Deep cold water corals extend for tens of kilometers along oceanic gravel ridges, reach 180 feet in height and shelter 1000 or more species of organisms.) Trawling also eliminates the cycling of seafloor sediments by a wide variety of worms (sediments smother the worms) and releases nutrients to the upper waters. The North Sea originally had many oyster beds and extensive reefs of tube building worms (shallow waters along the German and Dutch coasts still have these) and was much more clear than now. Despite 800 years of fishing, until the 1870s European seas still were full of life; then powerful, steam-driven trawls began to transform their bottoms and catch too many fish. Much of the North Sea was dry land during the last ice age and is shallow, easy to trawl and productive. Fish catches (measured by effort put into fishing) soon fell. Later the absolute size of the catch began to fall. By the 1920s the effects of motorized trawlers were felt worldwide. The average size of cod landed in the Gulf of Maine decreased by 66% (from one meter to one third of a meter) following mechanized fishing in the 1920s.

Every year shallow banks are reworked by storms and must be recolonized by worms and other benthic organisms but below about 80 feet (less near sheltered coasts) most animals and plants survive the storms. Coastal banks are also scoured by tides. Tides increase the flow of oxygen, organic matter and plankton over the banks, making them good habitat for filter feeders, and their sediments and gravels can be stabilized by animals like corals, sea fans and crinoids. Stabilization of the sediments lets other plants and animals establish themselves: starfish, snails, sponges, sea squits, crabs, lobsters, sea anemones, prawns. In shallow northern seas such as the North Sea or the Gulf of Maine, whose cold, nutrient rich, well oxygenated waters make them excellent places for fish, elaborate reefs of cold water corals covered the rubble left by melting glaciers. Trawling dug up the bottom, broke up the oyster beds and the crusts of shells scallops live among, scooped up young fish, crayfish, and other invertebrates the cod and haddock eat. In the North Sea, 16 pounds of marine invertebrates are killed for every 1 pound of marketable sole. Currents also sweep over seamounts, bringing nutrients and oxygen, creating another favorable environment for fish and for filter feeders of the bottom. In the Tasman Sea, corals and crinoids cover 90% of pristine seamounts. After trawling, the figure drops to 5% and the seamount loses half its biomass, and much of its potential as a fishery. Recovery takes 50 years or more. (With 30,000 seamounts in the Pacific and 6,000 in the Atlantic, such fisheries can go on for some decades.) Modern bottom trawls trap and kill almost all fish, mollusks, and invertebrates they encounter. Much of this is unwanted or too small to keep legally and is thrown back dead or dying into the sea. (Trawlers sorting their catch leave a trail of dying fish and feeding gulls.) Large bottom trawls, pulled across the seabed at 4 miles per hour, leave trails of mud visible from space. The fertilized water column above the trawl is good for algae, and the muddy bottom good for breeding shrimp. (In an early argument trawls were said to “plow the sea” and so increase fish production; they may increase production of shrimp.) Trawls are set to run along the bottom. Two boats fishing a rough bottom (rough, say, with boulders or underwater corals) drag a heavy chain between them over the bottom to level it and then fish. Scallop trawls are set to excavate the bottom for scallops, flounder trawls to dig out the fish, which lie half buried in the mud. So the habitat left by trawling is not good for fish: their prey is gone; and the mud clogs their gills, interferes with their vision and causes algal blooms in the waters above. Global positioning systems and competing fishing fleets mean likely places for fish are trawled much more often. In 1900 it was estimated every trawlable part of the North Sea (an area of 100,000 square miles) was trawled twice a year. Near the end of the Grand Banks fishery for cod each spot on the banks was being trawled every four months. Small hills were levelled and the seabed turned into a vast mudflat. When fishing stopped, the cod population was perhaps 0.3% of the original population: in Canadian waters the original cod population was something like 7 million tonnes, while the population when trawling stopped, perhaps 22,000 tons—for a fisher who encountered a school, still a lot of fish. With the cod gone, invertebrate populations exploded and a lucrative fishery for snow crab, northern prawns, lobsters, rock crab and sea urchins—all once prey of the cod—appeared. The legally saleable bycatch of cod from fisheries now allowed in Canadian cod waters amounts to probably 90% of the cod population, which shows no sign of recovery.

Heavy fishing changes the structure of fish populations and simplifies food webs in the oceans. Under the selective pressure of heavy fishing, fish become smaller and breed earlier. More complex food webs provide ecosystems with a greater degree of resilience, by providing more sources of prey for predators. Biological production and fish catches change by a factor of 10 by trophic level (from phytoplankton to zooplankton to small fish to still bigger fish to the fish we eat). Most large fish eat at several trophic levels. Marine food webs are dynamic and the fish taken every year (by other fish, whales, seals, weather, people) are replaced at varying rates. If one animal (fish, crab, squid) becomes scarce, others are available as prey; and a change in its favored prey by a predator allows another population of animals time to recover. (Human fishers are just one predator in the sea; fish and sea mammals still eat several times more fish than people, though the proportion is falling). Fishermen in heavily fished waters now catch fewer large fish, fewer predatory species of fish (such as cod and tuna), more fish at the middle of food chains (so-called forage fish, such as pollock). Industrial fishing has reduced populations of large predatory fish (blue marlin, tuna, swordfish, sailfish, cod) in the oceans by 90%. Populations of predatory fish tend to stabilize at about 10% of their preharvest populations under industrial fishing. Newly fished populations take 10 to 15 years to crash. (Such newly fished populations include many slow growing deepwater fish of the continental slopes, such as the orange roughy, or Chilean sea bass, a fish of the sunless depths that takes 20 to 30 years to mature, and lives to 150; orange roughy are fished in their spawning aggregations—deepwater fish must be, if the fishery is to be profitable—and could probably be sustainably harvested at 1-2% a year, an economic impossibilty.) Along heavily fished coasts, which are all coasts in the developed world, large predatory fish such as cod, jewfish, swordfish, sharks and rays are functionally or entirely absent, with implications that run down the trophic levels to the plankton that feed the sea; such fish join the large reptiles and mammals—whales, sea turtles, manatees, dugongs, sea cows, monk seals, salt water crocodiles—as ecological ghosts. Fished fish become smaller. The overfishing of cod under industrial trawling in the 1920s, reduced the proportion of large fish, decreased the length of fish of a given age, and decreased the size at which fish spawned (thus reducing the total production of eggs and the fertilisation rate). Tuna weigh half of what they did 20 years ago, marlin one quarter. Fishing for so-called forage fish, some of which remain abundant, reduces the food available to large predatory fish, the fish people in general prefer to eat. The forage fish are processed into animal feed and oil, some of which is fed to farmed fish, like Atlantic salmon, which convert the processed fish pellets to salmon flesh at a theoretical efficiency of about 33%; that is, at a loss of about 67% of the fish protein, which ends up concentrated below densely stocked salmon cages as a pollutant. (For many reasons, actual losses from forage fish to salmon flesh are much larger; all intensively farmed carnivorous finfish and shrimp are net consumers of protein and require 2-5 times the protein they produce; if our desire was food, and not the taste of shrimp or salmon, we would be better off eating the forage fish: as a nutritous broth, for instance.)

Large fish of any kind have become rare partly because industrial fishing catches them all, and partly because catching all the large fish creates a selection pressure for smaller, earlier maturing fish; such fish, because they are smaller, have several times fewer eggs and sperm, making recovery of the population from fishing more difficult. (Large old fish have larger and healthier eggs, and more eggs, which hatch into faster-growing young.) Size selection in a population of Atlantic silversides (a minnow), in which the larger 90% of the fish were removed before breeding, led over 4 generations to a population of fish half the size of a population that was anti-fished (the smaller 90% of the population removed before breeding). The biomass of the fished population also fell, to about half that of the anti-fished population. The same process seems to happen in the wild. So fish in an over-fished population become smaller and the biomass of the population declines. The results of such selection pressures are not limited to fish. In an isolated population of bighorn sheep in Alberta, 25 years of trophy hunting, involving the taking of 57 animals (2 per year), reduced the mean body weight of 4 year old rams from 200 to 160 pounds, and their horn lengths from 28 to 20 inches. Most of the rams taken were 8 years old or younger; they had horns with four-fifths of a curl or more; most of them probably had not yet bred. Presumably smaller animals with less perfect horns did breed and so the population got smaller. The speed of the effect is startling. But populations of seed-eating finches in the Galapogos show small yearly variations in bill sizes depending on what seeds are available; two years of large seeds is reflected in a change in the mean size of their bills. So populations of animals respond rapidly to selection pressures.

Weather and the size of the parental population determine reproductive success in most marine species, more than the availability of resources. Environmental factors in any one year can overwhelm the effect of population size on recruitment of young fish—one reason the theory of maximum sustainable yield, which used the size of the parental population as the major determining factor in managing populations, didn’t work. Summer storms bring up nutrients from deep water, fertilising the surface waters and improving the survival of algae-eating young fish, making for good catches of Bering Sea pollock a few years later. Salmon populations in the North Pacific are in general favored by stormy winters, and extensive ice cover improves the survival of larval snow crabs in the Labrador Sea. Fish like cod and herring prefer cool water, so do better in the southern parts of their range (off the Massachusetts coast, off the coasts of Holland or Sweden) when winters are windy and bitter, while the opposite is true of cod at the northern limits of their range in the Lofoten Islands of Norway or the Labrador Sea. Herring eat plankton, the size of whose bloom varies with weather and sea conditions. And cod eat herring, perhaps as many as 29 billion of them in the North Sea in the mid-1800s, or 10 times more than people. Most fish and marine invertebrates release their eggs and sperm into the sea. Fertilisation is by chance. Since it helps if eggs and sperm are abundant, many species gather in groups to spawn. Spawning locations are traditional and partly determined by currents that carry the fertilized eggs and larvae back toward good juvenile habitat near shore. After spawning, the eggs and larvae float to the surface and drift back to inshore waters, where the juveniles mature. (So gravid female lobsters congregate where currents will wash their larvae back to the shallow, cobbled bays juvenile lobsters favor.) As they grow larger, the young fish begin moving offshore. Such fish generally remain over the continental shelf, in waters less than 600 feet deep.

When fish congregate to breed, the abundance and health of animals and of spawn matter. In terms of energy allottment (the use to which food is put), small fish are growing, large fish are reproducing. Big old fish play an important role in the ecosystem. A 25 inch female red snapper produces 200 times as many eggs as a female 16 inches long, or two-thirds her length. Among Pacific rockfish, older fish produce 10 times the eggs, and the survival rates of their larvae are nearly 3 times higher and their growth rates 3.5 times faster. (The larvae are larger when they hatch from the eggs, which probably contain more nutrients.) Fish eggs and larvae lead very uncertain lives. Fish and birds eat fish eggs, including fish that are normally prey. (People do too, but kill the adults to get them.) So abundant populations of fish must produce abundant spawn. Herring spawn along the English coast up through the 1800s covered gravelly bottoms in drifts 3 to 6 feet deep. Haddock eating them acquired a distinctive flavor.

The sea is subdivided by currents. Sea turtles migrate between the west coast of North America and Japan on undersea currents (remaining mostly within the top 20 to 120 feet, which makes it possible to set up a protected zone). The currents that deliver the larvae of marine animals to inshore waters depend on oceanographic conditions; currents that move lobster larvae in the Gulf of Maine are influenced by ice melting in the Arctic, by cloud cover and by winds. (Ice, by damping waves, changes the movement of currents.) Some populations of Caribbean reef fish that live near each other as adults show genetic differences indicating they are from separate breeding populations. The different populations spawn in different places and different currents carry their larvae inshore to nearby, but separate, rearing grounds. Tides and winds are somewhat predictable, but the subdivisions of the sea are dynamic. While local currents, dependent on the weather, are more or less dependable, storms and sudden shifts in currents wash many fish larvae out to sea. The very great majority don’t survive. Red sea urchins along the Pacific coast, fished by hand by divers, went into a decline under controlled fishing. Red sea urchins need a sufficient density of adults for efficient fertilisation. Young that settle near adults also survive better (the so-called nursery or canopy effect). To maintain the fishery it was suggested that each area be fished once every 3 years; and that the largest 20% of the population be left (they are the most fecund, and provide the most canopy habitat), along with the smallest 20% (for the next harvest). In such animals, small amounts of random variation in recruitment lead to highly variable future populations. Low levels of exploitation can cause a continual decline in the population, with no stabilization. In many echinoderm species (and marine species in general) periods of relatively low recruitment are followed by years of abundance; the population must be able to take advantage of the abundant years to maintain itself. (That is, when conditions are favorable, there must be enough large, old individuals to breed abundantly.)

The collapse of cod populations across the North Atlantic was a signature event of twentieth century fisheries; cod from the North Sea and the North American banks was a cheap source of protein in Europe for over a thousand years. The original cod populations had many large fish. Cod caught off New England in 1602 were larger than those then caught off the Grand Banks of Newfoundland, which had been fished by Europeans for 100 years. These large fish produced enormous amounts of eggs and sperm. Their reproductive potential allowed for recovery of the population from declines caused by fishing or bad weather. Many species of animals that gather in groups to breed (such as lekking birds, like sage grouse) will not breed unless enough individuals come together. Original spawning aggregations of cod at the edge of the North American continental shelf probably consisted of hundreds of millions of fish (the Grand Banks population is thought to have been several billion fish). Old cod knew the way to traditional spawning sites, following channels of warm oceanic water through canyons in the continental shelf. The underground forest of an undisturbed bottom probably aids in the survival of the fertilized eggs and larvae of cod, which are large zooplankton (and food for many of the fish that will eventually become their prey). After spawning, the adult cod followed capelin, a major prey species, inshore to feed over the summer. Most bays along the New England coast had their local populations of cod, which were connected to specific spawning sites. These populations were fished out from the 1930s to the 1950s.

Mechanized fishing for cod (steam trawling) began in the 1920s. Trawling was aimed at large old fish and thus selected for faster-growing, earlier maturing fish. Reducing the numbers of old fish reduced in geometric proportion the abundance of eggs and sperm and made maintaining the population more difficult. By the 1960s trawling had reduced the proportion of large fish in the population, decreased the length of fish of a given age and decreased the age at which fish bred. These are all ominous signs, but were not recognized as such by fisheries biologists, who believed half the population of a fish like cod could be removed every year indefinitely: the high rate of predation would make the cod population respond and grow faster. This theory of maximum sustainable yield ignored the effect of environmental factors on fish recruitment and the effects of fishing for large fish on the size of individual fish. Some inshore populations of fish (those fished more heavily and whose spawning sites were known) began to disappear. As cod declined in the southern Grand Banks in the 1950s and 1960s they were replaced by flatfish (flounder). Those were fished down with specialized trawls in the 1980s and 1990s. The collapse of the cod, the apex predator of these northern seas, may have improved the lobster fishery (one of the few modern sustainable fisheries) in the Gulf of Maine (cod eat young lobsters). The decline in cod ended 5000 years of stability between cod and kelp forests. Cod ate the sea urchins and other invertebrates that grazed on the kelp. When cod were abundant the invertebrates fed only at night. When the cod disappeared, the increasing sea urchins and other invertebrates along the Maine coast grazed down the kelp forests. After the sea urchins were fished out for the Japanese market, the vegetation grew back, but it was now dominated by introduced species (thus the bright yellow green of the modern intertidal). These had been there all along, and took advantage of the reduction in the native plants to make their move. The new seaweeds changed the camoflage background and probably the nutritional status of the seashore.

The primary reason cod off eastern North America have not recovered is probably that they are still overfished. Once fish stocks fall below 10% of their unfished populations (some would say 15%-20%) recovery becomes difficult. Predation and competition from other fish may prevent recovery. Meanwhile other problems appear. A warming climate creates various mismatches among connected species — hummingbirds migrating across the Mexican desert arrive in the high altitude gardens of the Rockies before the flowers have opened; the late freezing of sea ice keeps polar bears marooned on shore in Hudson’s Bay, losing weight, out of reach of their prey, the seals that live on the ice. The time of algal blooms in the ocean is determined by daylength, while the development of fish and invertebrate larvae is determined by temperature, so many mismatches are possible. In the North Sea the quantity and quality (calculated by size) of zooplankton available to larval cod has declined since the 1980s. The sea has warmed by almost 1º C. While the spring bloom of diatoms and dinoflagellates (the photosynthesizers) is determined by daylength, and occurs more or less at the same time, the large zooplankton (such as larval cod and the larvae of copepods, a food of larval cod) emerge in response to temperature and now hatch up to 2 months earlier. Young cod are faced with food too big or too small for rapid growth. Unless evolution corrects this mismatch between producers and consumers the local cod population may disappear. In the North Atlantic, the collapse of the cod population has let its prey species (such as shrimp, crabs and herring) increase. These species feed on large zooplankton, including juvenile and larval cod, and their predation may be helping prevent recovery. Also, forage fish, such as capelin, on which cod feed, are still overfished, among other things, for food for farmed salmon. Finally, most of the adult cod that are left are small. The females produce many fewer eggs and less well nourished eggs than large old fish.

There are many explanations for the lack of recovery in collapsed fisheries but in almost all cases fisheries reduced below an economically exploitable level (so breeding stock was, say, 1%-5% of unfished levels) have not recovered during the period of observation (until now, 15 to 25 years). In general, there is little evidence of recovery in fish stocks fished down to 10% of their reproductive biomass, after a period of 15 years. All major herring fisheries in the North Atlantic and Pacific collapsed in the 1960s and 1970s; after 25 years there has been some recovery in Norwegian stocks, but little or none in the others. Catches of Peruvian anchovies fell from 11 million tons in the late 1960s to 100,000 tons in 2000 and remains there. (Herring and anchovies, like salmon, flounder and cod, are opportunistic species capable of rapid reproduction, that can better stand fishing pressure than slower growing species like marlin, shark, or grouper.) Healthy populations of fish can survive climate cycles, such as periods of cooling or warming, if they don’t last too long. Not all periods of warming are due to us. Salmon are colonizing rivers on the Bering coast of Alaska north of their former range. Cod are also extending their range north and if trawling is stopped, may do well there. With a warming climate, the fish that stayed in the old range would survive on other prey (perhaps favored by an earlier bloom) that would be available in the healthy food webs of an unfished ocean and, if not, follow the changing water temperatures north or south. Populations of any large animal (cod, ospreys, passenger pigeons, moose) probably build up through a series of lucky events; several good years for recruitment; good years for survival and growth of the young animals; that is, through a historical process, which over enough time converges in many abundant populations. The population reaches a size that can buffer itself against normal environmental perturbance. But once reduced from that size (by fishing, hunting or natural catastrophe) it may no longer be able to buffer itself; its breeding stock may be too small, its individuals too scattered (the case with Pacific abalone), its individuals too poor at reproducing, their numbers too low for cultural behaviors such as breeding aggregations to work well. Its food supply may be limited or its habitat less favorable; perhaps polluted. Predation that the population could once withstand (and that had evolutionary benefits) may reduce the population further, perhaps to zero (the effect of wolves on a failing population of elk; perhaps of herring on larval cod). All this is true of cod populations (whose habitat for instance, has been diminished by trawling, and whose larvae are being eaten by their former prey), but their primary problem is that they are still overfished.

Fishing quotas in the North Sea are a political matter on which the governments and fisherman cannot agree, so limits are set too high. The same is true in most commercial fisheries. The Canadian Grand Banks have been closed to cod fishing for more than a decade, with little apparent recovery, but cod caught during legal fishing for shrimp, flounder and skate are thought to amount to 90% of the breeding cod population. (Such fish can be sold, an improvement over previous regulations, since few caught fish survive.) Industrial fishing is very efficient. In general, in new fisheries, it reduces a population of large predatory fish by 90% in 10 to 15 years. So this is the life of new fisheries, such as that for orange roughy (or Chilean sea bass), a deepwater fish that matures at 20 to 30 years and lives to 150; or for round nosed grenadier, another deepwater fish that matures at 8 to 10 years and lives to 75. Such fish could probably be fished at 1-2% of the population sustainably. Long life and low fecundity are typical of deepwater fish, which, living in the dark waters of the continental slope, depend on the rain of nutrients from the sunlit ocean above. Modern deepwater fisheries (there are several) are both extremely profitable and unsustainable. Cod, a fish of the sunlit waters above 600 feet (like most fished fish), matures at 5 to 6 years and lives to 20, reaching up to 90 pounds in weight. Cod, along with anchovies, sardines, herring, salmon and flounder, are so-called opportunistic fish; these are species with short maturation times, that produce a large number of young. Since under favorable conditions they have a capacity for rapid recovery, they can be more heavily exploited. During the industrial fishery after World War II, fisheries were managed for so-called maximum sustainable yield. It was thought half the population could be taken in any year. This would allow the remaining individuals, with more food available, to grow and reproduce more quickly and increase the yield of the stock. Unfortunately such fishing, with its minimum size limits, selectively removed the large old fish, the best breeding stock. Assessments of the size of stocks were also often too high, influenced by the ability of fishermen to find fish in an emptying sea, political pressure, and wishful thinking. The effects of weather and currents, which can greatly reduce reproductive success in fish, and also that of bycatch (most of the young fish thrown back died), were not taken into account. Probably too many variables affect the size of fish stocks for such simple calculations to work; or catch limits would have to be set much, much lower. (It is now thought 20-30% of the stock is a more reasonable catch limit.) So-called competitive species of shallow water fish, such as marlin, shark, grouper, sturgeon and halibut, are slower to mature. Their populations recover less quickly from exploitation and they withstand fishing pressure less well. Most stocks of such fish are seriously depleted or becoming so. Deepwater species can barely withstand any fishing pressure; their continuing presence on restaurant menus comes from locating new stocks. A recent study of data on fish and invertebrate catches from 1950 to 2003 in 64 large marine ecosystems (which together comprised 83% of global fisheries yield over the last 50 years) points to total collapse (stocks fished down to 10% of previous levels) by 2048. (The results surprised the researchers, who had not expected such a grim result.) Close to 30% of fished species have already collapsed.

Abundant populations of animals may be more vulnerable than their abundance indicates. A rule of thumb is that any animal population that declines by 20% over 10 years is at risk of further depletion. A high standing biomass of plants or animals may be associated with a low capacity for renewal, that is, a low reproductive rate, which makes the population vulnerable to a rapid reduction in size under some new stress. Some populations of land animals (moose, for instance, which have a relatively high capacity for renewal) seem to flip between populations of high and low abundance; once low, the population remains low. At European contact, passenger pigeons comprised 25-40% of the biomass of terrestrial birds in the United States: perhaps 3 to 5 billion birds. Pigeons ate mast (oak, walnut, elm, chestnut); they preferred beech, an oily nut of a size easy for them to handle. Beech is a common species of eastern and middle western primary forest. Like many nut trees, beech produces crops every other year, huge crops (allowing good survival and recruitment for pigeons) every 3 to 7 years. (The birds also ate berries, grasshoppers, insects and fruits.) The pigeons bred from April to June, on overwintered mast. Nut trees produce crops at periodic intervals, as a defense against animals that eat mast, and often those in a large area — thousands of square miles — produce in the same year, so the previous year’s breeding location will be unusable. The pigeons came north in huge flocks soon after the snow melted. The flocks spread out over a wide front. This was perhaps a strategy to find the quantities of mast a breeding aggregation needed. Breeding among the pigeons was synchronous: 3 days of courtship, 3 days of nest building, 1 egg, 13 days of incubation, 14 days of nestling care, abandonment of the site as a group. After 3 or 4 days the squabs also left in one flock. Losses during breeding were high. Branches collapsed under the weight of roosting and nesting birds; eggs and squabs fell out of the flimsy nests. The large breeding flocks attracted predators from far away. (This probably explains why the breeding process was so choreographed and so fast.) Most tribes of Native Americans would not hunt the birds when they were breeding, and pigeon bones in Indian middens are scarce. (But the Senecas performed a dance to celebrate the pigeons’ return.) Losses in the juvenile flocks were probably also high. Passenger pigeons laid 1 egg and took a relatively long time (perhaps 10 years) to reproduce themselves, that is, for each adult to produce another breeding adult. Adding an additional predation rate from Euro-American hunters of 10-20% might have sent them into a decline, which would have accelerated as the population fell and demand for pigeons remained high. (Pigeons were commonly taken by American market hunters from a nesting.) It was however habitat loss that finally made their lives impossible. The middle western hardwood forests on which passenger pigeons depended were being turned into cornfields in the nineteenth century at a rate of about 5% a year. Approximately 95% of Ohio was forested in 1800, 10% in 1910. Beech trees begin to yield mast when 40 years old. Clearing began in earnest in the Middle West in the 1820s and peaked in the 1880s. The last big nesting by pigeons was in the 1880s in Michigan. There was less food in the second growth forests for the birds and their colonial habits made them vulnerable to continued hunting. When their numbers were reduced below a certain point, they likely ceased breeding.

Men with shotguns and plows did in the passenger pigeons. Flightless seabirds and island tortoises were eliminated by men with clubs, populations of whales by men with harpoons in wooden boats. Whales in the Bay of Biscay were eliminated as an economic population between 1000 and 1400 by Basque whalers in rowboats, right and bowhead whales in the Strait of Belle Isle off Newfoundland between 1500 and 1600, bowhead whales off Spitzbergen between 1607 to 1670. In the 1760s, ships sailed in summer from New England to the islands off the Labrador coast where colonies of nesting seaducks took advantage of the lack of predators and abundance of food in the sea to breed. The hunters arrived during the molt of the adult bird’s flight feathers, which occurs when the ducklings are still in the nest. Unable to fly, the adults were herded into stone pens and clubbed to death, their feathers plucked to supply the demand for featherbeds in the American market. Within a decade the seaduck populations had collapsed (one species, the Labrador duck, went extinct), and the voyages were no longer worthwhile. Great auks, another flightless seabird of islands, originally ranged from Florida and the Mediterranean to Norway and Greenland. Great Auks were eliminated from the Mediterranean and from much of the European coastline early, but in 1500 they were still found on offshore islands along the North American coast from the Carolinas to Labrador. The feathers of the great auk were also used for bedding and its fat for lamp oil. The plucked bodies were boiled to extract the oil, what was left used to fuel the fire. Its flesh was salted away in barrels and sold to the poor in place of pork. (Like many island birds, it also went extinct.)

The stories scientists tell change. The collapse of the California sardine fishery in the 1950s was thought at the time to be from overfishing. In 1936, 726,000 tons of sardines were landed from Monterey Bay, and the sardine fishery there was the largest fishery on the North American west coast. (The anchovy fishery off Peru, now also mostly gone, was 15 times as large). The sardines didn’t disappear; they became too few to be worth catching. Their place in the food web was taken over by a larger fish and fisheries biologists thought recovery was prevented by competition with that fish, since predation and competition commonly prevent recovery of drastically reduced populations of animals. More recently, it was suggested that the sardines of Monterey Bay didn’t recover because of the draining of the wetlands that encircled the bay. Several hundred thousand acres of wetlands were drained in the 1940s and 1950s for farmland. The wetlands had supplied iron to the bay; iron is a limiting nutrient for phytoplankton, the algae at the bottom of all sunpowered oceanic food chains. Reducing the iron in the bay would lower its productivity, reduce all its fish populations overall and perhaps preclude the recovery of the sardines. However by 2004 the sardine population had somewhat recovered and 50,000 tons of sardines were landed. Monterey Bay has warmed by 3º C. over its long-term average in the last few decades. The warming of the bay may or may not be connected with global warming (which would make the warming greater and faster). It is however associated with a 60-year cycle of cold and warm periods in the bay that was unrecognized in the 1930s. Sardines do better during warm periods; anchovies during cold ones. (Similarly, a 20 to 30 year temperature cycle between the north coast of California and the Gulf of Alaska seems to be associated with a variation in salmon abundance between northern California and the Gulf of Alaska. A period of strong, persistent low pressure systems in the Aleutians diverts more of the cold North Pacific current into the Alaskan gyre, less down the coast of California. This means warmer ocean temperatures and less upwelling of nutrient-rich water off the California coast, fewer krill and small forage fish for the salmon, more competition and predation by warm-water fish, and fewer salmon off California, more in the Gulf of Alaska.) All the explanations for the crash in sardines may be correct, and declining amounts of iron in the water together with competition from other fish and commercial fishing may keep the sardines from recovering to anywhere near their former abundance.

The ocean is the final receptacle. Its saltiness and proportion of minerals is determined by what has washed off the land during the last 4 billion years, as well as by what has been stored away in salt and mineral deposits. In modern times oil also washes off the land, carried down to the sea by rivers and spilled directly from ships. The oil weathers into tar balls, that (when small) resemble fish eggs and are eaten by juvenile sea turtles as they enter the Gulf Stream off the Florida coast. The tar balls are concentrated by currents; young sea turtles make the same mistake with the plastic pellets used as feedstock by the plastics industry, while their parents mistake floating plastic bags for jellyfish, with disastrous consequences. For several decades fields of tar balls have concentrated in ocean gyres like the Sargasso Sea. Ocean gyres off the west coast of the United States and Mexico are also full of bits and pieces of weathered plastic, probably carried on currents from Japanese waters. Most seabirds found dead on the beach have bits of plastic filling their stomachs. Some floating junk is useful. A load of sneakers lost off Korea was used by oceanographers to track currents. The shoes float well and were still being found on North American beaches years later.

A fluid like the air, the sea also absorbs carbon dioxide, which is slowly increasing its acidity. The sea is thought to have absorbed half of the anthropogenic carbon dioxide released since 1800. It now absorbs about 2 billion tons a year, 20 times more than its net absorbtion of 100 million tons under natural (current, post-glacial) conditions. This carbon dioxide will eventually be neutralized by calcium carbonate on the ocean floor but because ocean circulation is slow, this will take thousands of years. In the meantime the ocean’s acidity rises, a tenth of a pH unit so far in its surface waters (a 30% rise in acidity). Tall seamonts in the ocean are covered with a calcium carbonate ooze, like snow, from the constant rain of shells from phytoplankton. Below 4000 meters, where the ocean becomes more acidic, the ooze dissolves. As the acidity of the upper ocean rises, these deposits will also dissolve, and organisms that make shells of calcium carbonate will have more trouble making them; the process will be energetically more difficult. Creatures with calcium carbonate shells or exoskeltons are common in the ocean and include algae, shellfish and other crustaceans, and corals. As acidity rises further, the shells of shellfish will tend to dissolve.

The ocean also absorbs heat from the atmosphere; this makes it grow in size, increases the speed of its chemical reactions, melts its ice, and changes the flow of its currents. The freezing of sea water in the Arctic and Antarctic leaves behind a layer of very cold, dense, salty water that sinks. Along with the circumferential Antarctic winds (the so-called “roaring forties,” that circumscribe the globe unimpeded by land), and variations in saltiness and temperature caused by other factors (such as winds and input of fresh water from rivers), the freezing and thawing of sea ice is thought to drive deep ocean circulation. Oceanic circulation drives the upwelling of nutrients from deep water and oxygenates the deep sea. Antarctic krill, which hatch in the abyssal depths, rise to the surface as larvae and feed on the algae that grows under the ice; the abundant krill (now overfished for salmon food) are what make the seas off Antartica so rich in fish, seabirds and sea mammals. The ecosystem on the soft, slushy underside of the ice, driven by a thick coat of algae (along with nematodes, bacteria, ciliates, rotifers, copepods), is much richer than that on the sea surface: ice supports the abundance of Arctic and Antarctic waters. (Algae live on the top, sides, bottom and within ice floes, hanging on during the winter, growing and dividing as the light returns.) Upwelling, caused partly by ice, and also by winds, replenishes the nutrients in surface waters. The daily upward movement of krill in Norwegian fjords at dusk to feed on plankton causes some mixing of the nutrient-rich deeper waters with the surface ones. The freezing and thawing of sea ice about Antarctica once involved 7 million square miles of ice and was the greatest global climate cycle. Shrinking of the ice sheet began in the mid 1950s but wasn’t recognized until scientists began to study the logbooks of whalers. The ice had shrunk by 25% by 1970. Perennial Arctic sea ice has shrunk about 8% per decade since the 1970s. There was a 14% drop from 2004-2005. The Arctic Ocean is expected to be icefree in summer by about 2050 (some say 2030), except for areas about northern Greenland and the Canadian Arctic Archipelago. Ocean productivity, measured by chlorophyll content, has fallen by about 6% since the 1980s, perhaps from reduced mixing of seawater, perhaps from higher temperatures in the polar oceans.

Open ocean in the Arctic reached its greatest extent in modern times in the summer of 2007. The increase in the rate of melting of Arctic ice was the result of a confluence of events, none of them predictable: the sort of historical accident that can also result in the rise or fall of animal populations. The Arctic ice pack normally rotates around the Pole, where it thickens as it ages. Thick ice resists summer melting, especially at high latitudes. In 1989 a periodic flip in the Arctic Oscillation changed the normal pattern of winds and air pressures over the Arctic. The weather settled into a phase that carried sea ice out into the Atlantic rather than circulating in a gyre around the Pole. The proportion of 10 year old ice fell from 80% in the spring of 1987 to 2% in the spring of 2007. New sea ice, being thinner, melts more easily. The same shifts in air pressure increased the flow of warm water from the Bering Sea into the Arctic Ocean, with a corresponding flow of cold water out, and increased the flow of the deep warm currents that run north from the Atlantic near Scandinavia. In the summer of 2007 a high pressure system settled over the Arctic and caused unusually sunny skies in June and July. Warm winds from Siberia pushed the melting ice floes offshore, where currents and winds carried them out of the Arctic Ocean. These events worked together with the increasing rain of soot (from burning coal, oil, cow dung, forest fires, wood) and dust (from construction, desertification, agriculture) that now falls on Arctic ice, and which, by absorbing solar radiation, increases its rate of melting; with the warmer air temperatures from ongoing warming of the Arctic; and with the feedback loop of more open ocean absorbing more solar radiation. Once ice starts melting, it melts faster as the meltwater on its surface absorbs heat. For all these reasons the loss of sea ice in the summer of 2007 was the greatest ever measured. Such losses set the stage for more positive feedback (less ice in the fall, thinner ice, more open ocean absorbing more heat) and a rapidly increasing loss of sea ice in the coming years.

A healthy ocean needs all its parts. Diverse ecosystems are more resilient, recover more quickly from environmental stresses and can better withstand nutrient pollution. (A generalization with exceptions.) It is thought one reason the North Sea withstands such heavy fishing pressure while the North American cod fisheries have all collapsed is that the North Sea is a more diverse habitat. Without its full compliment of sea mammals (the keystone predators), sea birds, large predatory fish, forage fish, zooplankton, benthic organisms and other invertebrates, the ocean becomes a very different place. Some think ecosystems, once formed, evolve as wholes, to states of greater abundance and resiliency, with benefits for all the ecosystem’s inhabitants. The health of an ecosystem should be a primary consideration in any management decision affecting that ecosystem. In the 1950s marine mammals caught 7 times more fish than the human fishery, this dropped to 3 times in the 1990s, a result largely of whaling and overfishing. The majority of the species eaten by marine mammals are not fished by humans, though there is some overlap in human fisheries with those of marine mammals and baleen whales off the northeast coast of the Americas and in the North Sea. The minke whale, the most abundant species of baleen whale, feeds on small pelagic fish that are not targeted by human fisheries.

One can see the effects of eliminating keystone species in the marine ecosystems of the northeast Pacific. When whaling resumed there after World War Two, the biomass of fin and sperm whales in the North Pacific was reduced from 30 million to 3 million tons. In the waters off the Aleutians 500,000 whales were killed from 1949 to 1969. Killer whales, or orcas, once preyed on all the great whales. During the period of the hunt, killer whales may have increased in number by feeding on injured or dead and abandoned whales. They would have been directed to the slaughter by the sounds of ships and exploding harpoons, similar to the way seabirds follow trawlers. When whaling ceased, this food resource ended. Live whales were also now much less abundant. (When great whales were common in the oceans, their bodies were not only food for killer whales but a huge source of organic matter at the ocean bottom. Their carcasses supported a variety of specialized invertebrates, such as polychaete worms that live on the fat in whalebone. Besides being a general source of nutrients for the deep sea, dead whales may have supported entire novel ecosystems, in the same fashion as oil seeps and hydrothermal vents. Before whaling, 850,000 whale carcasses may have been on the sea bottom at any one time.) Orcas first turned to eating harbor seals, which began declining in the 1970s, then to fur seals and Stellar sea lions, both of which began declining in the late 1970s. It is thought once such animals reach 1% of the orcas’ diet they decline. (One writer has estimated that 40 orcas eating sea lions would have caused their observed decline.) Hunted almost to extinction early in the twentieth century, then managed successfully for their pelts by the native Aleuts of the Pribolof Islands, fur seals still continue to decline. The decline in seals and sea lions was probably also helped along by a decline in their preferred prey of oily fish, such as perch and herring. These species were overfished and also faced a warming in the North Pacific that began in the late 1970s. (At this time Pacific salmon began colonizing rivers north of their traditional range.) With the warming seas, perch and herring met competition from an increasing number of pollock, a less oily (and therefore less nutritious) fish. The pollock may have been partly responding to the decline in whales, as both whales and pollock feed on plankton. At any rate, as the fishery for pollock developed, they also began to decline. In the 1990s the killer whales turned to eating sea otters, a very undesireable food for them. By 1998 they had eliminated 90% of the sea otters in a 1000 kilometer stretch of the central Aleutians. Sea otters eat sea urchins and sea snails, which graze on coastal kelp forests. Unlike the plants of warmer waters, kelp have no chemical defenses against grazing. Without sea otters to eat the herbivores, kelp forests decline, and with them the fish and invertebrates that depend on the forests (such as abalone, now heavily reduced by overfishing) and the predators of these animals (bald eagles, seagulls, seaducks).

This story began with whales and ended with sea otters and kelp. The more traditional story about sea otters and kelp begins with the trapping out of the sea otters along the northwest coast of the Americas by Russian and American trappers in the 1700s and 1800s for the Chinese fur trade. Sea otters were once extremely abundant in the Aleutians and along the northwest coast of the Americas down through much of California. The Europeans who wintered on Bering Island in 1741-42 found them tame enough to be killed with clubs. Like cod along the east coast, sea otters along the west coast of the Americas let the great kelp forests of the North Pacific Basin, which go from southern California to Korea and Japan, maintain themselves.

The Black Sea is a lesson in the future of ocean ecosystems. The Black Sea is an enclosed basin, fed by huge rivers from the wet north, and draining out through the Mediterranean. It is naturally anoxic below a few hundred meters. (Hydrogen sulfide increases with depth.) It occupies an area of about 160,000 square miles and is up to 7000 feet deep. Modern vertebrate life in the Black Sea (its fish, birds and mammals) depends on the kelp forests and seagrass meadows of its shallow northeastern shelves. Red kelp (Phyllophorra) once covered 5800 square miles of the shelves (about 3% of the sea’s total area). The kelp forests were a source of food and a nursery for many fish species; they were also the principal source of oxygen on the shelf, producing about two million cubic meters of oxygen a day. The tops of the kelp were harvested for agar, which is used as a thickener in processed foods, such as ice cream. Beds of sea grass (Zostera) and oysters ringed the sea, with beds of mussels below, to the level of the anoxia. Sea grasses also oxygenate the bottom muds and the water. Schools of anchovies circumnavigated the sea, wintering in the warmer southern waters, and fattening on the northeastern shelves. The anchovies were food for bonito, mackeral, tuna and dolphins, fish of open water. Twenty-six species of fish were caught commercially in the 1970s. The Black Sea was famous for its sturgeon.

The ecosystem of the Black Sea was weakened by algal blooms that prevented light from reaching the sea grasses and kelp forests, sharply reducing nursery and feeding areas for fish and reducing the production of oxygen. The blooms were caused by nutrients brought to the sea by its rivers, which flowed down through industrialized Central Europe and western Russia. Half the nutrients entering the sea come from the Danube, largely from agricultural runoff and sewage. Riverworks that bypassed large swamplands in Hungary and in the Danube’s delta eliminated much of the function of nutrient removal that these wetlands had formerly performed. Besides nitrogen and phosphorus, the rivers brought down oil, mercury (present in high levels in marine mammals of the Black Sea), and organochlorines (also found in the sediments, fish and marine mammals). Dams for hydropower and river transport removed sand from the river water, so the algal blooms in the sea shifted from diatoms to dinoflagellates. Road construction and hotel development degraded the coastline, adding nutrients and silt to the water. Bottom trawling for fish caused siltation and destruction of the bottom fauna. Microbes in untreated sewage infected marine mammals. A comb jelly (Mnemiopsis, a predatory ctenophore) from the east coast of North America was accidently introduced in ships’ ballast water in the late 1960s and rapidly began to dominate the ecosystem. Mnemiopsis feeds on the eggs and larvae of fish. Freed from its usual predators and competitors, Mnemiopsis reached densities of one kilogram per square meter in the open ocean, five kilograms per square meter over the shelves: huge numbers for a jellyfish. Pelagic fish populations crashed in the late 1970s. Mnemiopsis has since declined (another introduced jellyfish feeds on it), but only six fish species were still commercially fishable in the 1990s. The sturgeon populations (fish famous for their flesh and eggs) have collapsed, partly from overfishing, partly from loss of spawning habitat. Without rehabilitation, the future of the Black Sea is of a sea of jellyfish feeding on phytoplankton. Because of the input of nutrients, overall productivity in the sea is probably up. (Of course, both jellyfish and plankton can be harvested for food.)

* * *

In an ideal world low-trophic animals would be used for aquaculture: edible seaweeds, bivalves such as oysters and mussels, sea urchins, sea cucumbers, sea snails, fish like mullet that graze on bacterial films. (Mullet were cultivated in tide pools by the pre-contact Hawaiians). Growing the more valuable shrimp and carnivorous fin fish (such as salmon and tuna) in pools and pens causes many problems. Aquaculture works by keeping costs down: by crowding the fish (saving on cage space), treating the salmon with chemicals (to control diseases that spread through the crowded and stressed fish and the sea lice that thrive on them), by letting pollutants sink to the seafloor. A typical Maine salmon farm has 250,000 salmon in 20 pens. Each pen produces 2 metric tonnes of waste a day, or 40 tonnes total, the waste of a small city. This sinks to the bottom suffocates the plants and animals there. Its decomposition produces hydrogen sulfide. Carnivorous fin fish require 70% of their diet in fish oil and meal; as a rule, they consume 2 to 5 times the protein they produce. So large numbers of so-called forage fish (such as sand eels in the North Sea or capelin in the northwest Atlantic) are caught to feed them, 10 to 20 times the weight of the salmon (much more than wild salmon would have eaten on their own). Such forage fish are an important link in the food chain of wild fish, seabirds and sea mammals. Catching forage fish puts us in direct competition with wild fish like salmon and cod, reduces the food available to them and makes their populations more vulnerable to yearly changes in the abundance of plankton. Concentrating the forage fish in fish meal and oil raises the level of persistant organic pollutants in farmed salmon to 10 times the level in wild fish. A writer has suggested that salmon be raised on insect larvae cultivated in grain wastes, instead of on fishmeal (this would eliminate much of the problem with persistent organic chemicals in salmon, and reduce the problems associated with fishing so far down the trophic levels). Another has suggested that rapeseed oil, which contains the omega-3 fats salmon need, could be substituted for much of the fish oil. Raising them in better cages would minimize escaping fish, which swamp the declining wild stocks with attractive large males whose genes do not adapt them for survival out of the cage. (But salmon are survivors: fish escaping from Chilean farms have dispersed around the tip of South America and colonized rivers on the eastern side of the continent, which formerly had no salmon.) Raising salmon in conjunction with oysters and mussels would use some of the escaping nutrients (ideally salmon numbers would be adjusted so the bivalves took up most of the nutrients: this may or may not be economically possible). The marine pollution problems caused by salmon farms might be solved by raising them in saltwater ponds on land; and recirculating the purified seawater; a notion with its own problems — where to put the ponds, how to supply the seawater. Raising fish in concrete pens can work: Calvisius caviar comes from white sturgeon raised in warm water from the steel mills in Brescia. (All of which brings us back to the integration of aquaculture and horticulture in the small farms of the Pearl River Delta, where the wastes from the fish are used to grow vegetables.)

Artificial reefs are a way of recycling biological production in high-nutrient (nutrient polluted) areas (which include all the coasts of developed nations). They are an alternative to raising fish in cages. The reefs concentrate fish by creating new feeding grounds and nursery areas. Thus they enlarge fish biomass. The fish feed partly on the plankton blooms of the nutrient-rich waters, partly on the algae and other organisms growing on the reefs, partly on each other. Artificial reefs also support sessile communites of marketable filter feeders (oysters, mussels, sponges); these filter the water column of algae and bacteria directly. Up to a point, such structures (like delta wetlands, or the oyster reefs of shallow estuaries) can turn nutrients from pollutants into useful biomass. In Japan concrete reef structures have been used to create entirely new fishing grounds. (Similarly, crushed stone of the appropriate size can be laid down on river beds to replace salmon spawning gravels that have been silted over.) Paid for by the state, or by organizations of fishermen, in the interest of making estuaries more productive, but populated by wild fish, such structures (netted plastic bags suspended in the water also work; filter feeders colonize the inside of the bags, their protuding parts nipped off by feeding fish) require a different look at who owns wild stocks of fish and shellfish.

Biological production, and fish catches, change by a factor of 10 by trophic level. So phytoplankton (green plant plankton: the base of all sunlight driven marine food chains) have 10 times the mass of zooplankton (the small animals of the sunlit surface); zooplankton 10 times the mass of the small forage fish that eat them, small fish 10 times the mass of larger predatory fish that eat them, and so on. Trophic levels are more an intellectual device for understanding differing biomasses of predator and prey than an exact delineation of the natural habitat. Fishing at lower trophic levels (typically for fish of low value to be used for oil or meal: that is, animal or fish food) reduces the forage fish available to larger fish and makes them more dependent on the smaller fish that feed directly on plankton (or in some cases on the plankton itself). This exposes larger fish more directly to seasonal and yearly changes in plankton abundance. The ecosystem as a whole loses resilience and the population of large fish fluctuates more strongly. Large predatory fish (tuna, cod, sharks, rays) and marine mammals (dolphins, whales, seals, sea otters, walruses) play keystone roles in marine ecosystems. Predators promote species richness by holding down competition among prey species; and help maintain a balance between herbivorous animals and the plants they eat. Removing too many top predators causes a cascade of changes through the ecosystem. So overfishing of cod, a large, predatory fish, let sea urchin populations off the New England coast explode and graze down the kelp forests which were one of the bases of the ecosystem. Over-fishing of sharks lets mid-size predatory fish increase — such as groupers, a fish of Caribbean coral reefs. Groupers prey on the parrot fish that keep the reefs clean of algae. (And it turns out the size of the parrot fish, especially in the absence of large reef grazers like turtles, is important.) Similarly, killing wolves lets coyotes increase. The coyotes reduce the numbers of mid-level predators, such as foxes, oppossums and skunks.

Fishing, like forestry or farming, should try to fit into the ecosystems it uses: that is, to catch the right kinds and sizes of fish, while avoiding other kinds and sizes, and avoiding dolphins, seabirds, turtles, whales (run into by ships, or caught in purse seines for tuna) and damage to the seabed. The Maine lobster fishery is one of the few healthy industrial fisheries. (The Alaskan salmon fishery may be another.) Fishermen are limited to 800 traps each. Small lobsters and lobsters large enough to breed must be thrown back. Any female with eggs must be thrown back and her tail notched; any female with a notched tale must be thrown back. So the focus is on catching lobsters of the right size and sex to let the population thrive. Recent videos of traps on the seafloor seem to indicate that lobsters go in and out of baited traps at will and are caught only if the trap is pulled up when they are inside; so lobsters are caught by chance, and the fishery may be feeding lobsters. (Eight hundred traps is a lot. Lobstermen in northeastern Canada use half that number of traps, close the season for several months a year and catch virtually the same numbers of lobster per man per year.) Small-scale fisheries (2 men in a boat) catch more fish for human consumption, rather than for meal or oil, catch higher value fish and expend more labor per fish. (So the sea supports more fishermen. The catch of forage fish for processing is about equal to that of fish caught for human food, but the table catch produces 94% of fishery revenues.) More of the value of the fish stays with the fisher. Rather than going to shipbuilders, equipment manufacturers, and fuel companies, more of the money goes back into the community as mortgage payments, car payments, payments for food, clothes, schooling, medical care. Because small-scale fisheries are coastal and local, they are more amenable to effective management and regulation. For one thing, everyone knows everyone else. (The North American lobster fishery is largely self-regulating. This works because the lobstermen know each other’s traps and because the area where a lobsterman sets his traps is considered his.) Such fisheries also use passive gear (traps, seines, dip nets, hook and lines of limited length) which, while effective (traps at river mouths will catch virtually all migrating salmon), tend to catch the right fish and do not harm the ocean bottom. (Harvesting shellfish involves digging up the bottom. This may or may not be harmful depending on where and how it is done.)

Fish should be protected during their spawning and nursery stages, in places where they gather before migration, and on the seamounts and along the thermoclines (meeting places of warm and cold water, sought out by schools of anchovies and sardines), where fish of the open ocean, such as tuna, sharks and rays, gather to feed on the forage fish. Trawling, except on historically muddy, gravelly bottoms where recovery of the bottom animals is not expected, should be banned; or perhaps eliminated entirely. Purse seining for species like tuna involves tremendous by-catches of both juvenile tuna and other species. The other species often include large slow-growing fish which are slow to reproduce (such as sharks and rays), as well as turtles, dolphins and whales. Schooling fish like tuna, when drifting along a convergence zone, tend to hang out under floating objects. Once these would have been drift logs, many hundreds of thousands or millions of which sailed through the oceans. Now they tend to be man-made structures equipped with fish-finding sonar and satellite beacons, which notify the fishing boat of the presence of fish. Tuna seines are 6000 feet in circumference and 850 feet deep. Each haul of the net brings in fish worth $250,000 to $750,000, and many fish and sea mammals besides tuna. Long lines in the Pacific, up to 60 miles in length and holding 30,000 hooks, catch endangered albatrosses, turtles and sharks. Longline fisheries catch about half the loggerhead and leatherback turtles in the Pacific each year. They have significantly reduced populations of albatross. Adult leatherback turtles weigh 1500 pounds and eat jellyfish. They are useful animals in the modern ocean. The numbers of females returning to nesting beaches fell from 90,000 in 1980 to fewer than 5000 in 2006, a 95% reduction in 25 years. The by-catch of industrial fishing (turtles, dolphins, underage fish, endangered fish, unwanted fish) is generally about a third of the catch. Most of it is abandoned dying. Trawling brings up the bottom itself: stones and mud with the scallops, shoveled off the deck; a ton of deepwater coral for every 2.5 tons of orange roughy. In some fisheries, such as those for tuna and cod, much of the by-catch is juvenile fish of the target species. These fish are thrown back dead, or in the case of bigeye tuna, a threatened species, end up in the can with the legal species. Making fishermen keep all their catch, and stop fishing when they reach their quota of rare or underage fish, is one way to eliminate bycatch.

Closing areas permanently to fishing in marine reserves is the best way to restore marine ecosystems. Unlike fishing equipment or quotas, reserves are relatively easy to police. To restore fisheries, large areas may have to be in reserves: from 20-40% of coral reefs, for instance; and 20-40% of continental shelves. Marine reserves outperform fished areas in egg production by 10 to 100 times (the spawning fish are larger and produce more eggs) and fish in reserves increase in biomass several times over their biomass in fished areas. Without fishing, the bottom habitat can recover. The selective pressure for small, early-maturing fish is eliminated. Fish from reserves will populate the areas about them; these areas can be sustainably fished. In birds it is thought that only 10% of the population produces an excess of surviving young; these birds breed in specific places; so the places to protect are important. (Places that produce an excess of young may shift.) Something similar may be true for fish, with weather on the spawning grounds or during larval development the determing factor. It is thought that fishing effort in the North Atlantic must be reduced by a factor of 3 to 4 to allow fisheries to recover; and that protecting 40% of the North Sea from fishing would produce the most fish for the least fishing effort, and thus the most profit for the fishers. The fishing fleet worldwide has something like 2.5 times the fishing capacity the seas can withstand, with much of the surplus resulting from government subsidies. To reduce fishing effort much of the fleet would have to be bought out and sunk, not re-sold. Scuttled boats make good reefs. Buyout payments would go into retirement accounts, not new ships. Trawling would end (so much of the fleet would be obsolete). The number of fishermen would not necessarily decrease, as more labor went into catching bigger fish. Some fisherman might be hired to restore fisheries, such as oyster reefs in the North Sea or local populations of cod in the bays off the New England coast. The cost of restoring fisheries would not be more than the current subsidy to the fishery, which amounts to 25-35% the value of the fish caught. The story is the same as for forests: the largest long-term profit is made by catching fewer, larger fish.

Overfishing by sport fishermen is also a serious problem. In some fisheries, the sport fishery may equal or exceed the commercial catch, but catches by sport fishermen are not included in the fishery statistics used to set catch limits. (From 1960 to 1980 commercial fishermen took 1.5 million pounds of fish per year from Long Island Sound, recreational fishermen 23 million pounds, including 15 million pounds of bluefish. The recreational fishery was worth several billion dollars.) From 1972 to 1988 the average weight of 5 sport fish in the Southeastern United States—red snapper, gag, snowy grouper, scamp, speckled hind—fell by 75%. The only place to catch trophy fish in Florida at the present time (2008) is near the defacto reserve about the space station at Cape Canaveral, closed to fishermen for reasons of safety.

* * *

In the United States, the heraldic fish connecting the land with the ocean is the Pacific salmon: the totem fish. Salmon are born and mature in rivers but gain most of their size and weight in the ocean. I think salmon have become so symbolically important partly because they were extinguished so recently, and so deliberately. Every person sees as normal the state of the world, the trees, fish and birds, of his childhood and salmon were still abundant half a lifetime ago. Thus we adapt to a less and less abundant wild world. The sea and the land mingle in the Pacific Northwest; along unsettled coasts, forests come down to the shore; after storms, stranded salmon sometimes dangle from the branches of firs. In California, Oregon and Washington, salmon are now at 6-7% of their former abundance. But salmon are an opportunistic species. Under good conditions 4 to 6 fish return to a river for each spawning adult. They are thus capable of doubling or tripling their population with each spawning run. Their needs are simple: access to healthy rivers, a congenial ocean and protection from overfishing.

In 1600 the northeastern United States and Canada also had abundant populations of Atlantic salmon. These faded away in the nineteenth century. Salmon were said to be commercially extinct in New England by 1850 but by the 1770s New England fisherman were sailing to Labrador for their (more abundant) salmon. Atlantic salmon along the east coast of North America were never as abundant as Pacific salmon. The riverine habitat was less favorable (though some spawning fish reached six feet in length); and salmon also shared east coast streams, and the nearby ocean, with abundant runs of other fish like herring and shad. Aboriginal fishing pressure on salmon was light. In pre-contact times in eastern North America it is thought salmon outnumbered people by a thousand to one (for 2.5-5 million fish). As Euro-American settlement expanded, sedimentation from logging and agriculture buried spawning gravels and smothered eggs and larval fish. Dams blocked rivers. (Small mill dams numbered in the tens of thousands.) By 1850 half the original salmon habitat in eastern North America had been cut off by dams. Nutrient pollution from sawdust, silt, sewage and manure reduced the oxygen in river water. Rivers flowing through cleared land warmed. High water flows coming off cleared land scoured out salmon nests, as did water released from temporary dams (so-called splash dams) used in log drives. Travelling downriver with the spring high water, the logs scoured out the gravels and killed juvenile and adult fish. In Oregon, log-driving, which ended in 1954, scoured some small riverbeds to bedrock and made them useless to fish. (Without gravel, there is no place to spawn. So gravel mining in riverbeds also reduces salmon habitat.) Europe’s populations of Atlantic salmon, which were also enormous, were reduced by the eleventh century, though in the 1700s German apprentices still complained of being fed salmon several days a week in season. In the late twentieth century a few salmon were still making it up the Rhine to Switzerland. The long reduction in the English and Scottish fisheries was accompanied by laments and much good advice; when the advice was followed, the fishery would temporarily recover. Traps set at the entrance to English salmon rivers were very effective at catching salmon. Atlantic salmon fell before overfishing, dams and development. While salmon are born in rivers, they gain most of their weight in the sea. The final blow to them was the discovery of their main foraging area in the Labrador Sea, north of the Grand Banks, between Labrador and Greenland. Salmon from both North America and Europe converged here, where the mixing of cold and warm currents produced a continual growth of plankton and forage fish that also supported the North American cod. The salmon fed on capelin, sand eels, herring, squid and amphipods (shrimp-like crustaceans) for one to two years before returning to their natal rivers to spawn. Netting salmon there from the 1950s-60s on reduced returning populations of fish to very low levels. This ocean fishery was regulated in 1984 but the catch was set too high to do much good. Salmon requirements were well known when the Pacific species were discovered by Euro-Americans.

Pacific salmon die after spawning and their bodies contribute nutrients to streams and streamside forests. Scavengers remove half the dead coho salmon from small streams on the Olympic Peninsula and carry them up to 200 yards from the stream; scavengers like black bears leave about half the carcass uneaten; the marine nutrients in the remains of the salmon and the urine and faeces of the scavengers feed the forest. Trees grow 3 times faster along healthy salmon streams. Nutrients from the sea once formed the bones of California grizzly bears. Up to 30% of the nitrogen in valley bottom forests with salmon streams is marine in origin. In lakes that hold sockeye salmon up to 90% of the nitrogen in the algae on the lake bottom and up to 70% of the nitrogen in the lake’s plankton and in the juvenile sockeye come from decaying adult salmon. (Juvenile sockeye have been seen nibbling on the carcasses of their elders.) Salmon and their river valley habitat constitute a pool of shared nutrients brought by the fish from the sea. Spawning salmon re-arrange river beds, carve away gravel bars, even out the stream bed, turn over the gravels; their yearly efforts change the width and shape of a stream. One could say the shapers of a river valley with salmon are streams, trees, weather and salmon. Abundant runs of salmon make small streams overflow. They rearrange the streams by digging nests, filling them with fertilized eggs, then covering the eggs over with gravel, so the eggs and larval fish develop within the protection of the stones. Nest building by female sockeye salmon move virtually the same amount of sediment as that moved by currents. Nests are dug below the usual level of winter high water scour. Such ‘knowledge’ of how deep to put the eggs is set by evolution in the different races of salmon that use a watershed and evolves as conditions change. (Eggs set at the correct depth survive.) The size of the fish also determines the depth of the nest, and the size of the gravel it can deal with. Small fish can only spawn in small gravel streams, with relatively low winter flows; large fish can spawn anywhere there is room for them. The five species of Pacific salmon reach different sizes and return to spawn at different times, from late summer to early winter; thus they fully exploit the varied riverine habitat. Chinook and coho spawn in rivers, sockeye in lakes or in streams draining lakes, chum salmon in small channels near estuaries, pink salmon in estuaries. Juvenile salmon live in fresh water for several years, then spend from 1 to 4 years in the ocean, where they gain 90% of their weight. Atlantic salmon are less differentiated than Pacific salmon, but vary somewhat in ‘body style’ from stream to stream.

Settlement changes rivers. Undeveloped rivers in the Pacific Northwest are bordered by huge old trees. When the trees fall, some get caught in the river channel. Drift logs accumulate against them and the deflected river excavates a deep pool under them, a resting place for migrating salmon. The river also excavates a channel around the jam on the shore side; the slower water there is good habitat for young salmon. Gravel accumulates at the tail of a pool, creating spawning habitat. As a river meanders, pools and riffles are excavated at the outside of bends, and cut-offs and side channels provide still-water habitat, but snags greatly increase the number and variety of pools and side channels, especially in steeper river valleys. Pools dug around snags are 2 to 4 times deeper than those dug by the current alone. Sometimes soil accumulates on the log jams and they become islands. Over time, such natural rivers develop complex patterns of channels in their lower reaches. The main channel is used mainly by adults for migration and spawning, the side channels by young salmon. The more pools and the deeper they are, the more room there is in the river for large fish, and the more downpool spawning habitat. Chinook salmon (the largest of the five subspecies of Pacific salmon) spawn in the main channels of large rivers, such as the Columbia. The abundance of chinook and coho salmon in a river (coho spawn in smaller, steeper streams) is related to the number of pools. Over the last 150 years log jams have been removed and streamside trees cut for fuel or timber, to make rivers navigable, or their banks settleable, and 65% of the deep pools in rivers draining into Puget Sound have been lost. Thus much habitat for fish and for large fish in particular is gone. A stable log raft with trees 2 to 3 feet in diameter growing on it blocked the mouth of the Skagit river when American settlers arrived. The river flowed under it. In winter the log jam flooded 150 square miles of the valley: this expanded rearing habitat for juvenile salmon but prevented human settlement. Before development, most rivers in forested landscapes had stretches where they flowed as a complex network of channels. Those in parts of the Rhine valley were famous; infamous to boatmen. Interaction among the flowing water, the topography of the floodplain and riverside trees created the riverine habitat. The Nisqually River in Puget Sound (an undisturbed river) has 2000 logs per mile of channel (one every 2 to 3 feet), most of them in jams, and a complex network of channels in its lower reaches. The Army Corps of Engineers removed 65,000 logs from the Willamette River in Oregon (880 logs per mile) from 1870 to 1950 to improve navigation. A million snags were removed from the lower Mississippi and 180,000 streamside trees were cut from 1864 to 1884 to prevent them from becoming snags. (Many, many more trees were cut for fuel for steamboats.) Steadily-flowing, single channel rivers, of a certain depth, without side-channels, backwaters or swampy riverside wetlands, are a modern creation. (Sometimes inadvertent, as with former mill streams in the eastern U.S.) Fishing reduced the number of salmon in the Pacific Northwest, but land development reduced the ability of the landscape to support salmon.

Heavy fishing for the largest fish reduces over time the number of large fish, which are the more successful breeders. In the Pacific Northwest in 1700, 50,000 Native Americans along the Columbia lived on salmon, catching 20 to 40 million pounds of salmon a year, or 1 to 2 million fish. This amounted to 5-20% of the pre-contact run of 11 to 16 million fish. Industrial fishing after the 1880s took about 90% of the runs. The salmon maintained their population under this regime for 40 years. Their numbers started to fall in the 1920s, shortly before dam construction began on the Columbia, perhaps because of a change in ocean conditions to those less favorable to salmon, perhaps a result of cumulative changes in the river habitat, perhaps from growing competition from the introduced shad, another anadromous fish. Most likely, all three influenced salmon numbers. Forty years of heavy, size-selective fishing would have an inevitable evolutionary effect on a population of fish, making them both smaller and less abundant. With salmon, this reduces the fitness of the fish for larger streams, higher water flows and larger gravel sizes. Removing log jams from rivers reduces the size of the salmon habitat. The runoff associated with logging and land development increases siltation in gravel beds, making them less suitable for fish, raises river beds overall (as silt accumulates), reduces the number of pools, and increases flooding and the size of winter flows. The increased flows scour out spawning gravels more deeply, destroying salmon nests (dug less deep by smaller fish). Silty runoff from logging buries spawning gravels and smothers salmon eggs and larvae. Unscreened irrigation diversions lead juvenile salmon (migrating downstream to the sea) into cornfields, where they die at so many to the acre. Logging as little as 5% of a watershed increases streamflows by 10-55%. (Steep slopes and road ditches emptying directly into streams account for the larger numbers.) Five years after a clearcut, stream flows are typically up 50%; streamflows remain 25-40% higher for 25 years; and summer water temperatures remain high for several decades. Logging, together with selecting for smaller breeding-age salmon, would have had a long term effect on salmon populations. In rivers about Puget Sound affected by urbanization and agriculture, and thus with high winter flows, fall-spawning salmon tend to be replaced with spring-spawning cutthroat trout—there is less rain and runoff and thus less scouring in spring and summer, when cutthroats develop.

Salmon are associated with cold, clear, forested streams. They became abundant in the Pacific Northwest 2000 to 3000 years ago. By then, the warm, muddy, post-glacial rivers had cleared the glacial debris out of their channels, the climate was cooling somewhat, and forests were expanding. The expanding forests retarded the spring snowmelt, extending the spring rise in many rivers and lessening its force, and also shaded the streams, cooling them in summer, and releasing cooler ground water into them. (This effect of forests can be dramatic, reaching 10º F. in summer in a shaded reach.) The forests reduced runoff and sedimentation. Streamside trees fell in the rivers and, anchored by their roots, forced the water to excavate pools around them. This was true even in major rivers like the Columbia. During all this time, people lived in the Pacific Northwest. All the reasons for the salmon’s association with forests are not clear, but salmon do better in forested suburban streams than in unforested rural ones. (Since salmon productivity depends on water temperature, this is likely because, especially in the more southerly ranges of the Pacific salmon, unshaded streams commonly reach 70° F., too warm for the fish; though the difference could also be caused by increased siltation of spawning gravels in rural areas, or the effect of farm chemicals on the invertebrate life of the streams.)

Salmon also vary in abundance with ocean conditions. The Pacific Northwest and Alaskan stocks of salmon tend to alternate in abundance in cycles of 20 to 30 years. Good years for salmon in the Pacific Northwest occur when a strong North Pacific current runs south along the California coast, causing an upwelling of cold, nutrient-rich water, and creating an abundance of krill and small forage fish for the salmon to eat. Such conditions keep fish of warmer waters, competitors of the salmon (such as the Pacific mackeral, which eat young salmon) south of the main feeding grounds of the salmon. Alaskan stocks of salmon are favored when strong Aleutian lows, associated with strong El Nino events, divert more of the North Pacific Current into the Alaskan gyre, and less down the coast of North America. (Strong El Nino events are becoming more and more common with the warming of the atmosphere.) Then the Pacific Northwest stocks have less food, more competition and suffer more predation, while the Alaskan stocks prosper. A warming climate may make life more difficult for the southern stocks of Pacific salmon.

Dams for electric power, flood control and navigation on salmon rivers usually mean the end of salmon. Bonneville Dam, the first on the Columbia’s main stem, was completed in the 1930s, part of the Works Progess Administration’s effort to revive the American economy during the Depression. While salmon on the Columbia were already in strong decline from overfishing, the fact is that dams block salmon runs. Mill dams across English salmon streams were required to leave gaps as big as a well-fed three year old pig, put sidewise to the current, to let the fish pass. Laws requiring fish passage in New England in the early 1700s exempted existing mill dams; they also weren’t much enforced. Fish ladders (‘salmon stairs’) were invented in the 1820s in Scotland. If dams are not too high, fish ladders let the salmon surmount them. (Ladders also work for some other species of anadromous fish, such as shad and alewives.) Some dams are too high for ladders. Grand Coulee on the upper Columbia blocked the runs above it. (The dam is 550 feet high and shut off 1500 miles of spawning streams, but the fish that used that water amounted to only 5% of the Columbia’s salmon run.) Dams change the river habitat into a series of still pools, whose water temperatures and chemistries are more like those of lakes. Summer temperatures rise in the pools to levels near the tolerance of salmon and oxygen levels fall. Warm water from the upper levels of the reservoirs flows into the fishways, with the result that the temperature in many fishways is near the upper limit salmon can stand. (This may be one reason salmon have trouble locating the ladders: they are avoiding the warm water—an easily fixable problem.) Dams flood the stretches of spawning streams above them, the former gravelly deltas of creeks in the river’s main stem. In the now slow water, spawning gravels in those creeks silt over, as do those in the slow sections of the main channel, where the largest salmon of the largest run spawned. Gravels downstream of dams erode away. Thus much spawning habitat disappears. About 55% of the area and 31% of the stream miles of salmon and steelhead spawning habitat in the Columbia Basin has been eliminated by dams.

Young salmon are adapted to traveling downstream on the spring highwater (facing upstream, steadying themselves with their caudal fins). During their downstream ride they begin the physiological transformations necessary for them to survive in salt water. These changes are timed and finish in the brackish water of the estuary. Young salmon must reach salt water at the right (evolutionarily determined) time. In a dammed river, young salmon move downstream through waters that are warm and unshaded. The more favorable cooler waters deeper in the pools are low in oxygen. Since the spring flow is controlled, they move more slowly, under their own power, not the river’s, and must stop to feed. A journey that once took two weeks now takes two months. Their travels expose them to predation by birds, mammals and other fish. Counter intuitively, dams also make it more difficult for salmon to move upstream. Coho swim up the Columbia at 1 to 2 miles an hour for their 1000 mile journey to Redfish Lake in Idaho. Fish can move relatively effortlessly upstream through turbulent water by using slight movements of their heads to let opposing currents push them forward. In the long pools between dams they must swim, using their main swimming muscle. Salmon usually don’t feed during their upstream migration and this additional effort uses up more of their stored fat. As they congregate below the fishways, they are exposed to predators like seals. Juvenile salmon moving downstream are killed by their passage through the turbines in dams. From 10-15% are killed at each dam; so passage through 10 dams means a cumulative loss of 66-80% of the fish. While losses of juvenile salmon are always high, these are not small numbers. Losses through dams can be mitigated by new construction; problems in the pools can be reduced by making the river flow closer to normal conditions (so-called run of the river power generation, which means loss of some summer generation capacity).

So the physical conditions of modern rivers are less favorable to salmon, partly because of changes in the rivers themselves, such as dams, less spawning habitat, fewer pools, warmer water, lower flow, partly because of changes in the watershed that affects rivers, such as less forest cover, higher flows of water entering streams, levees blocking off valley habitat, more silt entering the stream, caused by agriculture, logging, mining, commercial and urban development. The waters of modern rivers also contain many industrial chemicals, some of which affect salmon. Nonylphenol is a surfactant used in detergents, insect sprays and various industrial products and processes. It is common in sewage effluent and the effluent from paper mills and textile factories. Its use in spruce budworm sprays strongly correlates with a decrease in returning Atlantic salmon in Atlantic Canada. Nonylphenol seems to interfere with the changes young salmon undergo to adapt to salt water (so-called smoltification). It also acts as a hormone on larval fish, turning males into females. Other harmful chemicals include the copper-based algaecides and fungicides that leach from the treated lumber in docks and other river structures. These compounds interfere with the young salmons’ vasonomeral system (a primitive sense related to smell) and makes them more susceptible to predation. Chemicals diffusing from broken salmon skin cells (released by the strike of a kingfisher’s bill or the bite of a mink) cause nearby young salmon to freeze (and thus be less likely seen). In the presence of the copper-based compounds used in treated lumber this warning system doesn’t work.
Then there is fishing. Native Americans and early modern Europeans fished for salmon (Atlantic and Pacific) with traps at the mouths of rivers. Fences of stakes and brush led the fish into enclosures in which they could be caught. Both peoples also used nets, dip nets, hook and line, and spears. They built weirs of stone and brush in the rivers to herd migrating fish. The different races of Pacific salmon are stream-specific, and the different strains are adapted to different parts of their rivers. The fish suffer tremendous natural mortality when young: this is part of the evolutionary lability that lets them adapt to new conditions. Fishing at the entrance to a stream or in the stream itself can be managed to ensure a constant supply of fish in that stream, while fishing in the ocean, where all stocks mix, is another matter. Salmon vary in abundance with the weather during their lifetimes. Weather helps determine good spawning conditions in rivers and good growing conditions in the ocean. The size of a given year-class of a salmon stock depends on the reproductive success of the salmon in their natal stream that year: that is, on the number of breeders, their size and vigor, and the weather. The number of fry hatching out sets an upper limit for the numbers of fish that will return, though returnees will be orders of magnitude less. Conditions in a given year vary from river to river. River conditions for the several years the juveniles are growing in the river, and ocean conditions for their years in the ocean determine how many salmon survive to return to a given stream. In the ocean the different stocks and year-classes of salmon mix. The result is that limits on ocean fishing are impossible to set: there is no way to determine what strain of fish or what year-class of that strain one is catching. One can’t know how successful that particular class of that strain has been until they return to spawn. Ocean fish, not yet full grown, are also smaller. Most native peoples along the Pacific coast waited several days or weeks after catching the first, celebratory salmon, before beginning to fish. This let many large fish escape upriver. They also caught less than half the fish.

Finally, salmon must compete with introduced species of fish, like shad and walleye pike. In 1990 the Columbia River produced about 16 million pounds of shad and 20 million pounds of salmonids (salmon and trout). One historical estimate of salmon production on the Columbia is 50 million pounds. So perhaps 36 million pounds of fish is what the river can now produce. This however supposes salmon and shad (another anadromous fish) compete directly and are not merely complimentary inhabitants (like prairie dogs and buffalo) of the same habitat. Salmon and shad, migrating upstream together, crowd each other at the fishways below the dams, where both are eaten by seals and sealions, and delay each other’s passage upstream. Perhaps a less degraded Columbia (or one with man-made spawning channels) could produce 16 million pounds of shad and 40-50 million pounds of salmonids. The introduced walleye pike, a large freshwater predatory fish, eats young salmon as they migrate downstream through the dammed pools. Smolts that make it to the ocean normally spend some time in the Columbia’s delta, where many—estimates are 30-40%—are eaten by a growing population of gulls and terns, who nest on new islands of dredge spoil in the delta. Diking and draining the delta wetlands for agriculture, as well as improvements to it for river navigation, have made the habitat of the delta less favorable to salmon and more favorable to shad. Many of the water plants and the animals that fed on them—food and shelter for salmon—have disappeared and been replaced by smaller floating plankton, which shad eat.

The State of Alaska has one of the few healthy industrial salmon fisheries, especially that of Bristol Bay. Part of this is by accident: most of Alaska’s rivers are undeveloped and in a more or less natural state. Fishing is allowed only at the mouths of rivers (thus catching fish specific to the river) and only for set times. The goal is to let 50% of that year’s fish escape upriver. The large escapement and the restricted timing of fishing let salmon escape the evolutionary pressure associated with catching very large percentages of large fish. (So-called split regulations, which require the release of fish smaller and larger than a certain size, are supposed to do this for recreational fisheries.) The State of Alaska also does not allow pens for farmed salmon in its coastal waters, since the tame fish escape and breed with wild fish, to the stocks’ genetic detriment. The penned fish also become infested with sea lice, which attach to, and cause severe mortality in, young native salmon that pass by on their way to the sea. Under somewhat similar management, Icelandic runs of Atlantic salmon have increased during the late twentieth century. The Icelanders limit the times of netting and rod fishing and forbid ocean fishing in their territorial waters. Lately an organization of the owners of fishing rights on Icelandic, European and North American rivers has been buying out commercial ocean fisheries for Atlantic salmon. It was thought drift netting for salmon in their foraging areas off Greenland and the Faeroes and in ocean waters near their natal rivers has helped reduce Atlantic salmon to their current very low levels. Three years after netting near the Faeroes was stopped, the number of salmon returning to Icelandic and European rivers doubled. A serious plan for salmon recovery in the Pacific Northwest would buy out the ocean fleet and sink it. It would also take account of the salmon caught in the recreational fishery, which is worth several times the commercial one. The current decline of chinook salmon in the Columbia River is thought to be a result of ocean fishing as much as of dams. (Chinook spawn in the main channel and large parts of the main channel of the Columbia are still suitable for spawning.)

In the United States the idea of the new world still lurks behind the Puritan origin myths, but the actual aboriginal New England landscape of Abenaki villages, great flocks of migrating shorebirds, cornfields, clam beds, whales, forests of fire-managed trees, migrating shad and Atlantic salmon, has been largely forgotten. That once-settled landscape has been replaced by the Puritan villages and fields of a later myth. In this myth the welcoming Abenaki walk out of the empty forests bearing gifts: in fact the few coastal natives left in a landscape emptied by European diseases were trying to negotiate a political arrangement with the newcomers, to prevent their land being taken over by tribes to the west. But the idea of the wild land remains. On the west coast the Pacific salmon still hang on as wild fish, with surprising political power. In large part their political power has translated into hatcheries. Hatcheries were once viewed as compensating for the loss of river habitat (caused by siltation, logging, channeling of rivers for navigation, dam construction, water diversions for agriculture). They were supposed to let people develop the rivers for economic use and also have salmon. Unfortunately, hatcheries help only under very specific conditions. If used sparingly, they can help augment wild runs in the early stages of recovery. Hatcheries eliminate the evolutionary pressure salmon face in developing from egg to fry to smolt (a young salmon of a size capable of returning to the sea). They select for fish adapted to hatcheries: fish that can deal with a mob of other fish, that have a rapid response to being fed, and that have no awareness of predators. Most hatchery fish put in rivers die. They are larger than wild fish of the same age and compete with them for food and eat them, but they lack the genetic and learned adaptations to life in a stream. The story is the same for trout: a river with wild trout into which stocked trout are placed will end up with 50% fewer trout. The stocked trout harass and compete with the wild ones, but don’t know how to feed or to conserve energy or avoid predators. The result, after a year, is fewer trout. Several states gave up stocking viable trout streams in the 1950s, Canada in the 1930s, but modern management of American trout streams generally consists of stocking hatchery trout. Most of the fish are caught by fisherman in a few months. More hatchery fish are stocked the next spring. The cost of an adult hatchery Pacific salmon, returning from the sea, now varies from $10 to $100, depending on the run. Hatcheries were abandoned on British Columbian rivers a long time ago, but in the continental United States they have remained, as a way of avoiding hard political decisions.

* * *

But salmon runs can be restored. In the 1990s the winter run of chinook salmon on the Sacramento River shrank to less than 200 fish and the run was classified as threatened under the Endangered Species Act. (The original salmon runs on the Sacramento were second only to those on the Columbia.) As a result, two large irrigation districts and pumping stations at state water projects in the Sacramento Delta installed modern fish screens on their intakes. This cost money they had seen no reason to spend on protecting salmon before. Pumping schedules at state water projects in the delta were adjusted to better accomodate downstream migrating smolts (the pumps are powerful enough to reverse the current in the river). The operators of a diversion dam also changed their practices to accomodate migrating smolts. Obsolete dams on two former spawning tributaries were removed. Hatchery practices at Shasta Dam (which blocks any further upstream migation and closes off much upriver spawning habitat) were altered to use eggs and milt from captured wild fish. After 10 years 3000 to 7000 winter-run chinook were returning to the river. The population was still increasing. This is an increase of 10 to 20 times in a decade. If suitable spawning habitat were available, the salmon population could reach 300,000 to 2,800,000 fish in 30 years. A run of 200,000 to 500,000 fish would provide an excellent sport fishery. Such an increase is unlikely without further habitat restoration and control over ocean fishing.

Salmon are a test of our environmental seriousness. Salmon need cold, unpolluted water, without excessive run-off from logged or developed areas. They need undeveloped, forested floodplains so rivers can interact with trees and salmon to create salmon spawning and rearing habitat. If levees are moved back and rivers allowed to occupy their former flood plains (the “migration zone” over which the river moves as sedimentation and downcutting alter its bed and flow), the rivers and forest will re-create themselves. Such floodplains are places for both salmon and people. The mere presence of people doesn’t bother salmon, as it does grizzly bears, wood turtles and North American wolves. Urban runoff can be controlled with many small constructions (one for each parking lot, each roof), the same way it was increased; the cost of each construction is not great (a few hundred dollars per roof). Dam operations can be changed to let salmon pass and to release water when salmon need it. Making water available for salmon probably means growing varieties of irrigated crops that need less water; and generating less electricity. (Run-of-the-river dam operations on eastern trout rivers increase the populations of fish, of the invertebrates and insects the fish eat, of freshwater mussels. The absolute number of animals, their population density and their growth rates all increase. Power generation under such regimes is dictated by the natural flow of the river, rather than by demand for electricity.) Obsolete dams, or dams whose benefits are worth less than a restored fishery, can be removed. Removal of a dam is usually cheaper than repairs. (High dams, with lots of accumulated sediment behind them, can be a problem.) Protecting salmon pays off in the long run, in a capitalist world. Property values rise near natural rivers, as well as near reservoir impoundments. Salmon runs on the Snake River, a tributary of the Columbia, were obliterated by four dams built to facilitate river transportation between the State of Idaho and the sea. The barge channel, built and maintained by the government (barge fees don’t even pay for maintenance), is a cheaper means of moving Idaho logs and irrigated corn from the high plains to the Pacific coast than railroad, and thus made logging and irrigated agriculture more economic. Removal of the four navigation dams on the Snake would amount to the cost of two to three years maintenance (such costs include the ongoing costs of salmon conservation associated with the dams). A small transportation rebate would make up the difference between river and railroad transportation. The recreational salmon fishery on the upper Snake, maintained for free by the river and the salmon, is worth more than the value of the barge industry and the electricity produced by the dams. Unlike modern logging and farming, a sustainable salmon fishery is not an extractive activity, and will remain productive as long as the habitat remains.

Tule Lake is a natural lake in northern California that was turned into a reservoir by a dam on the Klamath River. Klamath River and (Tule) Lake are managed for electricity, irrigation water, waterfowl habitat and salmon. The water is over-subscribed and in dry years there is not enough water for the irrigation district and the salmon. In their natural state the seasonal marshes about the lakeshore flooded through early summer with the spring runoff. As the water receded in summer, food plants important to migratory waterfowl grew on the wet mud. The stabilized water levels of the reservoir eliminated this cycle. Falling lake levels were no longer seasonal but corresponded with the need for irrigation water and electricity. Water management also eliminated the natural disturbances of drying, flood and fire that wetlands need to renew themselves. The reservoir also stored silt carried into it from upstream. From 1958 to 1986 sedimentation in Tule Lake amounted to about 14 inches. This reduced floodwater storage and deepwater fish habitat. It reduced the water depth in the emergent marsh areas, eliminating most of the nesting habitat for diving ducks and colonial nesting waterbirds. The quality of the agricultural lands about the lake also declined. Soil nutrient levels fell, soil tilth worsened, while root-knot nematodes and fungal diseases increased. Looking for a way to deal with these problems, the refuge managers began introducing rotations between cropland and wetland about the lake. The rotations recreated the juvenile marsh habitat, broke agricultural pest cycles, and improved soil fertility and condition. The rotations are of 2 lengths. One rotation consists of 3 to 4 years in marsh followed by 3 to 4 years in crops. The second consists of 15 to 30 years in each. The short rotation creates early marsh, the longer one late succession marsh. The short rotation eliminates pests and adds nitrogen and phosphorus to the soil. Little additional fertiliser is needed to grow satisfactory crops. The long rotation adds nutrients and organic matter to the soil. Combining such agricultural rotations with water management that favors salmon lets one have crops, fish, migratory waterfowl and water supply, all under some semblance of human control. (Ideally yearly water withdrawals should be set at 20-25% of longterm average river flow, with a provision to reduce use in dry years. This is precisely the sort of limitation capitalist management abhors.)