ALBERT

Description / Table of Contents: PREFACE There can be little doubt that estuarine, coastal and shelf circulation modeling will assume increasing importance in the immediate future, as we work through the implications of industrialization for oceanic systems. These issues will place new and serious operational demands on available models, and the rapid increase in computational power we now enjoy makes it possible to respond with detailed simulations in many categories. As a result, we are witnessing an explosive growth in the quantity of model-generated information. Lacking, however, is a concomitant increase in its quality or even in quality control procedures. A single simulation exercise is easily capable of generating gigabytes of output in a matter of hours. Most of the data will necessarily go unexamined by its progenitors. Yet it is highly likely that disks full of simulation output will be used extensively as learning tools for students and researchers, as criteria for engineering design, as a basis for operational decision-making, and in the formulation of public policy. The purpose of this volume is to assemble and present what is known about the intrinsic quality of simulation output: its "correctness" for various purposes. We have operated on the twin premises that (1) every simulation has some intrinsic value and (2) every simulation has serious drawbacks. Between these two extremes lies a vast gulf of uncertainty and potential error, which must be bridged in a professional way if modeling is to achieve its potential in the coastal ocean. This is the basic challenge put to the authors of this volume. Essentially we seek to describe and consolidate approaches, theories, and practices for extracting information from models, and to understand the limits of their proper use. We have focused on quantitative approaches to the problem, from diverse physical and mathematical perspectives. Rather than pursuing abstract, philosophical analyses, we have sought contributions reflecting contemporary practice in government, academic, engineering, and research units where frontier model development is taking place. In a similar spirit, we have avoided imposing linguistic norms among authors, which we view as unproductive at this time. Instead, we have let the authors describe their approaches to this multi-faceted problem in their own terms, subject only to standards of clarity of expression. We are especially pleased to have balanced representation from European and North American authors. This volume was initially conceived at an informal workshop in Brussels, convened by our great friend and colleague, the late Bruno M. Jamart. That event first brought several of the contributing authors together over the skill assessment theme, and seeded a mutual vision of the present work. Bruno's energy, insight, integrity, collegiality, and boundless spirit has been an inspiration to all who knew and worked with him. His untimely death in 1990 represents a great loss which is deeply and widely felt. In his memory we dedicate this volume, a humble memorial to a great spirit and friend.