vi
Foreword
needs, we have increasingly harmed the sea. One reason is that, in a geochemical sense , the sea is downhill from the land; far more nutrients and
toxics flow from land to sea than vice versa. Moreover, the sea is where
land-dwellers conduct the last great hunt for Earth's wildlife, our marine
fisheries.
Unfortunately, individuals, companies and governments seldom employ
the precautionary principle (essentially, "Don't act unless you can be confident of doing no consequential harm") in dealing with the rest of the
world. Thus, deciding to cease and reverse harm requires an after-the-fact
understanding of how we are affecting marine processes. However, our understanding is complicated by statistical confounding, the difficulty in unraveling strands of cause and effect when many human activities are affecting the sea concurrently. For example, we know that North Atlantic right
whales are critically endangered even though whaling for them was
banned more than 60 years ago. But is this because undersea noise prevents them from finding mates, persistent organic pollutants reduce their
reproductive success, the food webs that support them have been altered,
they are experiencing demographic imbalances or inbreeding depression,
some other agent that we haven't yet recognized is harming them, or all
these factors in combination? In this and many other cases, it is difficult to
reach sound conclusions, yet we desperately need more insight to protect,
restore and sustainably use the living sea. In essence, humans are performing a vast unplanned and uncontrolled experiment on our planet 's lifesupport systems, and the fact that we are utterly dependent on their functioning for our survival suggests the value of tools that help us understand
key cause-effect relationships.
Modeling the dynamics of single populations, interacting species, ecosystems and human impacts is a powerful means of penetrating the haze
caused by multiple variables behaving in different ways. Modeling allows
marine conservation biologists to describe components of systems quantitatively and to assemble them into larger systems whose complexity exceeds
our unassisted predictive capacity. Thus, it can reveal results that we might
not readily infer, results whose assumptions can be examined, challenged,
modified and re-examined until they represent the broadest and deepest
understanding we can create. Once modeling was done in the realm of intimidating mainframe computers and equally intimidating programming
languages . Now the exponential increase of computing power and access
to it has democratized modeling . At the same time, models have become
more realistic and user-friendly, allowing a growing number of established
scientists and students to use them to test hypotheses, explore dynamics
and weigh sensitivities.
Of course , a model is not a panacea. It is a potent tool in a time when the
questions facing us are dauntingly complex and policy makers require
timely guidance. A model is a simplified vision of nature, and its validity depends on the degree to which it identifies, portrays and connects relevant
Précédent

- 7/461

Suivant