Lecture Series on Environmental Systems Analysis

Using Multi-Compartment Models to Investigate the Distribution Dynamics of Environmental Pollutants

Dr. Martin Scheringer, Institute for Chemical and Bioengineering, ETH Zürich
Multi-compartment models consist of several interconnected boxes representing different environmental media and spatial domains such as stretches of a river or larger geographical regions. These models provide a useful quantitative framework to describe the environmental fate of chemicals, and they serve several purposes: to evaluate the impact of individual processes and chemical properties on the model results, to set up a mass balance of a chemical in the model domain, and to estimate environmental concentrations to be expected for a chemical. In the presentation, applications of multi-compartment models to regional-scale problems such as nanosilver in the Rhine River, to the long-range transport of perfluorinated chemicals to the Arctic, and to the global distribution of the insecticide DDT are presented. For all cases, model results are compared to field data, and the performance of the models is discussed. Finally, an overview of multi-compartment model applications in the context of international chemicals assessment and management is given.

Paths towards Energy-efficient Mobility: Microsimulation and Agent-based Modeling of Incentives, Social Pressure and Technology Innovation

Dr. Peter de Haan, Institute for Environmental Decisions (IED) Natural and Social Science
Interface (NSSI), ETH Zürich
As purely technical improvements will not be sufficient to reduce our energy needs, policy tools to influence the demand are intensely debated. For example revenue-neutral incentive schemes target the end consumer, but will affect the whole market for the energy service in question. The common procedure is to assess the purely monetary effect of such policies, yielding reductions in environmental load and the costs thereof. This ignores that consumers do not show a purely priceelasticity-driven behavior towards any incentive, let alone governmental incentives. Both short-term psychological effects and medium- to long-term feedbacks on preferences and norms are to be expected.

We assume that people adopt resource-efficient behavior not primarily for monetary aspects, but because environmental-friendly behavior has pronounced symbolic values, i.e., it is important to those consumers to “do something”. Instruments like cash rewards create attention and emphasize the importance of the potential behavioral change. They have both direct monetary (price elasticity) effects, and indirect effects, through the creation of norms and the social influence on preferences (neighbor effects).
We present two different modeling techniques to address the direct and indirect effects for the case of incentives (or tax cuts) for the purchase of energy-efficient new cars. For short-term effects, accounting for possible non-linear market characteristics, we use a microsimulation framework with census data and highly detailed market data. The embedded discrete choice model is enhanced by element  of bounded rationality (decision heuristics) and by psychological effects from prospect theory. In addition, we present a simple agent-based model that allows for the assessment of how social influence might promote altruistic behavior. The model shows that green consumption patterns may emerge even if individuals’ hedonistic preferences favor environmental harmful products. To model does not “lock in” green consumption permanently, new technologies with other characteristics might cause consumers to “crowd out” of green consumption.

Are catchments complex ecosystems?

PD Dr. Holger Lange, Norwegian Forest and Landscape Institute, As, Norway
The hydrology and biogeochemistry of catchments is a focal area of ecosystem research since decades. On one hand, it is convenient to work with systems constrained by a (theoretically) closed balance for water and solute fluxes; on the other hand are these systems usually considered as complex and difficult to manipulate experimentally. However, a concise answer to whether these systems are complex requires rigorous concepts for the notion and quantification of complexity. In this presentation, we will first develop a framework for information and complexity analysis, using information theory, nonlinear dynamics and multivariate statistics, and apply it to time series of precipitation, runoff and other elements of the hydrochemical cycle. We will demonstrate that in this framework (a) complexity is clearly dependent on the time scales considered, (b) on short time scales, runoff data exhibit a common structure which is difficult to reproduce with conventional time series (autoregressive) models, and (c) for a given environmental variable, the relationship between information and complexity can be characterized by a family of one-parametric curves. We then proceed to longer time scales using methods capable of extracting (long-term) components from time series. It will be shown that runoff data are scaling (persistent) and in addition contain multi-annual modes which are partially synchronized over larger regions, which has implications e.g. for extreme events or in the context of climate change. It will also be discussed that the complexity of catchments is “modeller-dependent” – the presence of interacting biological degrees of freedom might require a new class of models, different from both process-oriented as well as empirical-statistical ones, which we term behavioural or interactive. An outlook on the potential of these models for catchment classification will be given. Although there will be no definite answer to the title question, complexity investigations may serve as a guiding principle for a research program towards a new classification of catchments, e.g. according to the degree of biotic influence on the water cycle.