Damage to ecosystems by commercial fishing is examined. This type of damage leads to a loss of opportunity to study and understand intact marine communities.

Traditional management of fisheries aims to optimize the catch of certain economically important species by commercial fishing boats, but this goal often eventually results in the collapse of the targeted species itself. We have excellent long-term data documenting this process for some fisheries. (See report on page 860 and news story on page 809.) But what is happening to the hundreds of noncommercial species taken incidentally or by poaching or ghost fishing by lost or abandoned gear? There is virtually no information. In addition, thousands of square kilometers of benthic habitat and invertebrate communities have been obliterated by trawling (see the figure), and recreational fishing and poaching have massively altered many coastal marine communities.

One irreparable consequence of this widespread damage is the loss of the opportunity to study and understand intact communities: In most cases there are no descriptions of the pristine habitats. The damage is so pervasive that it may be impossible ever to know or reconstruct the ecosystem. In fact, each succeeding generation of biologists has markedly different expectations of what is natural, because they study increasingly altered systems that bear less and less resemblance to the former, preexploitation versions. This loss of perspective is accompanied by fewer direct human experiences (or even memories) of once undisturbed systems. The effects of humans sometimes result in cascading ecological changes-a void often in part filled with introduced or inappropriate impostors that replace and mask the traces of the former natural system-but the species often simply disappear, leaving no conspicuous effect on the community. As with the loss of human cultures and languages after the passing of the elders with their wisdom, so too is humanity losing the evolutionary wisdom found in intact ecosystems.

How can society stop the alteration of these previously diverse and productive habitats? It is first necessary to recognize a fundamental problem: Unlike other effects of private interests on the resources of the general public, fishing often is considered a right not a privilege. Regulations often are barely tolerated by the fishing community, and poaching is rampant and minimally penalized. Management of fisheries has typically aimed to maximize the number of fish caught, while allowing little safety margin for assessment error, interannual variability in recruitment of young fish, or other factors such as El Nino and diseases. The countless species incidentally killed are usually ignored-unless they are also of commercial or recreational value, or are protected by the Endangered Species Act or the Marine Mammal Protection Act. Even the marginal protection afforded by these regulations are impeded by controversies and may take more than a decade to implement.

The challenge to management of any wild resource is to provide a buffer for uncertainties to safeguard the future health of the population or ecosystem. Appropriate application of available statistical techniques could allow such buffering. The statistical method of power analysis is appropriate and well understood, but rarely used. In this analysis, the consequences of making errors of two types (type I and type II) when testing the null hypothesis of no effect are clearly determined and stated. If the conclusion is that there is an effect when in fact none exists, a type I error results. However, if the null hypothesis is not rejected when in fact an impact does exist, a type II error results (1).

Enlarge 200% Enlarge 400%

Consider a proposal to restrict trawling from some areas in the Gulf of Maine to protect benthic habitat. If the proposal is accepted and fishing is restricted, when in fact it has no serious impact, it would be a type I error; however, if the proposal is rejected and trawling results in habitat destruction, a type II error is made. Current management focuses on reducing the type I error because this kind of error results in catching fewer than the maximum number of fish and is therefore highly visible to politicians and the fishing industry; management virtually ignores the type II error, principally because the deleterious effects are not immediately obvious. But ignoring the type II error results in failure to recognize and avoid serious long-term damage such as the collapse of the fisheries or environmental destruction. Scientific advice should be explicit about both types of errors, and most importantly, it must articulate the consequences to the ecosystem of making each type of error. The environmental consequences from type II error are much more serious because of the great time lags in the recovery of ecosystems or animal populations. Type I errors usually result only in short-term economic costs.

Proper management that weights both types of error has proven difficult because those profiting from the public resources are not required to prove that their actions cause no damage. The only mechanism available to society to protect these resources is somehow to prove actual or potential serious impact. This is virtually impossible for many reasons. Those defending the profiteering can argue endlessly over the accuracy of statistics that are virtually impossible to verify without an observer on each fishing boat or exorbitantly expensive sampling programs to generate independent data. Implementation of restrictions may be delayed by creating imaginative alternative explanations for the ecological damage and demanding that these be negated before restricting exploitation. Finally, even when presented with excellent data, regulators can simply assert that the data are inadequate and ignore serious environmental impacts. Resource management officials face strong economic barriers to risk-averse strategies. These policies cannot be expected to be implemented until the burden of proof is placed on exploiters of public marine resources to prove that they do not cause damage rather than simply assuming this to be the case until demonstrated otherwise. Similar commercial use of land resources requires extensive environmental impact studies and is carefully regulated. Continued monitoring is required, and all data are readily accessible to the public. Our marine resources need the same careful protection and stewardship.

In other contexts, particularly those involving human health and safety, we routinely place the burden of proof that the intervention will not cause damage on those hoping to exploit public resources. This need to protect against the serious type II errors is obvious for the Nuclear Regulatory Agency, which demands an extremely high margin of safety for the building and operation of nuclear power plants, for example. The Food and Drug Administration too demands a large margin of safety before approving the use of drugs in humans. Extensive testing to demonstrate the safety of new pesticide products is required before they can go to market, and air pollution regulations are expressly written to include an adequate margin of safety to protect human health. If society's environmental needs are to be protected so that future generations can also enjoy, learn, and profit from marine ecosystems, this legal burden of proof must be applied to our marine resources so that those hoping to exploit them must demonstrate no ecologically significant long-term changes. If the public hopes to preserve our marine environment, they must act quickly to change the relevant regulations and reverse the burden of proof.