Chapter 1

Chemicals, Models, and GIS: Introduction

In this book, we discuss concepts and algorithms for the analysis of environmental data and the development of chemical fate and transport models, which can easily be applied using a number of different geographic information systems (GIS) packages having conceptually similar structure and functionality. Most of these packages are freely available as open software, while a few are upmarket, commercial software broadly used in industry, government, and academia.

This chapter introduces some fundamental concepts that are subsequently dealt with in specific chapters. After identifying certain open issues in developing and using spatially explicit models, a tiered approach is proposed for chemical fate and transport modeling, examining separately absolute quantities, temporal dynamics, and spatial patterns. Finally, we will discuss the strengths and weaknesses, along with the opportunities and limitations of using simple GIS based models for different chemical modeling problems. The issue of choosing an optimal model complexity level, in light of the model scope and data availability, is a common thread that is considered throughout the book.

1.1 Chemistry, Modeling, and Geography

The title of this book recalls three distinct subjects: chemicals, environmental modeling, and geographic information systems (GIS). These may sound like they go well together but, if we examine them in more detail, we must admit there is something odd in this triplet.

Geography is about what one can see in the world out there. It is a discipline often associated with the humanities, descriptive in nature, that only in relatively recent times has introduced the use of quantitative tools [1, 2]. GIS technology and software are not just tools for geographers: their use is widespread in everyday life, and everyone has some experience of it (e.g., optimal routing or place finding using web-GIS applications such as Google Maps). In this book, however, we do not introduce the use of GIS technology and software as a mere ancillary tool for the handling of spatial data: rather, we describe how spatial analysis supported by GIS may help in setting up chemical fate and transport models that are simple, transparent, and at the same time spatially explicit and geographically meaningful, and enable us to make decisions on chemicals at a level that could only be addressed traditionally through complex numerical models.

In contrast with geography, chemistry has always been a science of the hardly visible, dealing with mysterious objects beyond common perception and only accessible through knowledge perceived by the layman as closer to magic than to geographic descriptions, or to the geometric clarity of Galilean physics. Not by chance, the word to denote this discipline is a minor variation on the original alchemy. Contemporary environmental chemistry was born because of the concerns of humankind for the consequences of a number of types of molecules—something like 105—that were deliberately created by the industry, partly in response to societal needs, and released into the environment, often without questioning their possible consequences. Chemicals cannot be seen, but their effects have been noted and are sometimes worrying. Environmental chemistry appears extremely complex because the phenomena with which it deals involve an overwhelming number of variables: the behavior of chemicals in the environment is difficult to predict from laboratory tests, as environmental drivers of fate and transport vary in space and time, and their comprehensive description is still a challenge.

Environmental chemistry is inherently a multidisciplinary arena, where causal relationships between observed chemical concentrations or fluxes in the environment, human activities, and environmental processes are continuously examined by scientists using necessarily limited data.

In such an arena, however, cultural attitudes are rather varied, and the modeling of chemical fate and transport has always been to some extent a story of misunderstanding. On the one hand, mathematicians and system engineers seem convinced, following Galileo's claim of mathematics as the language in which the book of the world is written [3], that good equations with a sufficient number of parameters will lead us to accurate predictions of chemical fate and transport; on the other hand, chemists with a precise understanding of the complexity of chemicals, but often with limited critical expertise with models, are eager to see their experience-based beliefs reflected in nice, smooth, but still realistic graphs or maps of chemical concentrations in air, soil, water, or vegetation. Traditionally, and obviously with several exceptions, chemists tend to be “experiment oriented” and to reason by examples capitalizing on their direct experience in measurements; they regard models as “tools” that may be useful or even essential but should not be questioned in too much detail about their grounds, merits, and limitations. And, most importantly, they should not be taken too seriously until robust evidence of agreement with observations is available. There is some truth in the old saying that “everybody believes in data except the one who measured them, nobody believes in models except the modeler”.

1.2 Mr. Palomar and Models

The original challenge of modeling is to describe reality as exactly and comprehensively as possible, and the last few decades have witnessed significant advances toward this goal, especially in such fields as meteorology and oceanography, paving the way for a generation of Earth system models [4].

Models of the Earth as a single (albeit extremely complex) system tend to be in line with the Digital Earth concept [5, 6], that is, a single repository of information on whatever can be observed on Earth. Earth system modelers would like to see this huge amount of information actually used in their models, which should tend to be more and more reliable so they could be used to make critical decisions for “what-if” type scenario simulations.

This greed for modeling has not spared the domain of chemical fate and transport in the environment, despite the apparent complexity of environmental chemistry. It was noted long ago that simple phenomena easily observed in practice—such as the dilution of a drop of dye in a pond, or the transport of a spill of oil over the sea, or the flume of smoke from an isolated industrial chimney—lend themselves to be interpreted through a neat mathematical scheme—a Gaussian plume or puff, for example—suggesting that complex environmental systems might obey, if not analytical solutions, at least numerical solutions of elegant governing equations.

Modelers, often coming from nonchemical disciplines such as meteorology, oceanography, or hydrology, tend to use their traditional interpretive schemes, implicitly assuming that models are basically adequate and problems are mostly in the input data, which and will be resolved as soon as new data are available. They tend to confront discrepancies between model predictions and field measurements by making models more comprehensive: a frequent temptation of modelers is to overemphasize the needs of a detailed and accurate description of reality, for example, by including more process equations in models.

Novelist Italo Calvino reflects on the use of models to describe reality in a collection of stories having as the main character a man with a capacity for observation that recalls a telescope: Mr. Palomar.

In Mr. Palomar's life there was a period when his rule was this: first, to construct in his mind a model, the most perfect, logical, geometrical model possible; second, to see if the model is adapted to the practical situations observed in experience; third, to make the corrections necessary for model and reality to coincide. The procedure, developed by physicists and astronomers, who investigate the structure of matter and of the universe, seemed to Palomar the only way to tackle the most entangled human problems, those involving society, first of all, and the best way to govern [7].

Although not easy to admit, many modelers still live in something like Mr. Palomar's first period. Our mathematical models still fall short of being able to accurately predict the spatiotemporal environmental distribution of chemicals in many cases of practical interest, even when they include a very fine description of virtually all known relevant processes.

But when comparing the sublime geometric harmony of the first period's ideal model, with the contorted and deformed lines of reality, Palomar obviously perceives the need of continuous and gradual adjustments to both the model and reality: “if the model can't change reality, reality may change the model” [7]. Palomar is an icon of what Donald Schön calls the “reflective practitioner” [8], who continuously questions the body of knowledge he/she refers to in action and recognizes the need to zoom out of the perspective of simply “applying tools” when tackling problems that require a combination of intuition, experience, creativity, and technical skills. A reflective practitioner soon realizes that a single line of modeling cannot be suitable for all purposes. Again Calvino's point:

Mr. Palomar's rule had gradually altered: now he needed a great variety of models, perhaps interchangeable, in a combining process, in order to find the one that would best fit a reality that, for its own part, was always made of many different...