.gif)
JAN ROTMANS, HANS DE BOOIS and ROBERT J. SWART
National Institute of Public Health and Environmental Protection (Rijksinstituut voor Volksgezondheid en Milieuhygiëne) P.O. Box 1, 3720 BA, Bilthoven, The Netherlands.
Abstract. This paper describes a simulation policy model of the combined greenhouse effect of trace gases. With this model, the Integrated Model for the Assessment of the Greenhouse Effect (IMAGE) scenarios for the future impact of the greenhouse effect can be made, based on different a assumptions for technological and socio-economic developments. The contribution of each trace gas can be estimated separately.
Basically the model, consisting of a number of coupled modules, gives policy makers a concise overview of the problem and enables them to evaluate the impact of different strategies. Because the model covers the complete cause-effect relationship it can be utilized to derive allowable emission rates for the different trace gases from a set effect related targets. Regular demonstration sessions with the simulation model have proven the importance of such science based integrated models for policy development.
Four different scenarios are worked out for the most important trace gases (CO2, CH4, N2O, CFC-11 and CFC-12). (One of these scenarios can be regarded as a growth scenario unrestricted by environmental concerns. The others are based on different strategic policies. After the simulation of future trace gas concentrations global equilibrium temperature increases are computed. Finally the sea level rise, the most threatening effect of the greenhouse problem for the Netherlands, is estimated.
Simulation results so far emphasize the importance of trace gases other than CO2. The Montreal Protocol on reduction of CFC is found to stabilize the relative contribution of these substance to the greenhouse effect
Introduction
Until recently the greenhouse problem was primarily attributed to the emission of CO2 resulting from the combustion of fossil fuel. At the Villach Conference in 1985 the role of other trace gases was considered to be equally important, advancing the data the GCM show for doubled atmospheric concentration of CO2. To policy makers the relation to various issues such as economic development, changes in biogeochemical cycles (especially that of carbon), the melting of ice caps and sea level rise is not always clear. Many of these topics are dealt with in independent studies. An integration of knowledge is needed to answer practical 'what if' questions.
As a reference institute of the Dutch government. RIVM has therefore developed a simulation model of the greenhouse problem. The primary objective of this model is threefold: firstly to offer policy agencies a concise overview of the quantitative aspects of the greenhouse problem, secondly to identify uncertainties or cru-
cial gaps in current knowledge, and thirdly to calculate the effects of several policy options concerning the greenhouse problem. In the long run it is meant to be an interactive tool for policy makers.
The model is not yet complete, but is continually updated, improved and extended. In this paper the results of the first project phase are presented. The feasibility of a policy oriented model based on scientific principles and consisting of a combination of modules has been investigated. Demonstrations given for high level Dutch policy officials and in Parliament have proven the model to be a helpful tool to improve the understanding of the greenhouse problem.
Model Description
The model is based on a large variety of data derived from both an extensive study of literature and knowledge transfer resulting from consultations of specialized experts. In this way the problem could be dealt with by combining different fields of research. The integrated information in the model yields insights which cannot be obtained from scattered information.
We modelled the greenhouse problem as a dynamic system with discrete time steps of one year and a simulation time of 200 yr, from 1900 to 2100. The system has a global character, entailing absence of spatial dimensions. The components (modules) of the model are given in Figure 1, in which an arrow from one component to another represents a driving influence from the first component to the second. The associated computer model, the Integrated Model for the Assessment of the Greenhouse Effect (IMAGE), consists of independent modules which are linked together, the modular structure allowing improvements to be implemented gradually without affecting the basic structure.
The framework of IMAGE consists of emission modules, a concentration module, a module converting concentration increases into temperature rise and finally modules deriving the sea level rise from temperature changes. The modules are linked in a simple way: the output of one module serves as input for the next. The modules mentioned are highly aggregated with a dynamic structure. The model, excluding the energy module, contains some 1500 equations, of which approximately 400 are for the carbon cycle.
At present the model includes five trace gases: CO2, CH4, N2O, CFC-11 and CFC-12. In the emission modules current estimates of historical emissions are implemented for the period 1900 to 1985. For the period 1985 to 2100 four sets of scenarios were chosen. An emission scenario is defined as a possible future development without pretending probability being a prediction. Furthermore, we define a set of scenarios as a similar, consistent development for all trace gases. The underlying scenario assumptions which will be discussed in more detail later, are based on a study of the sources of trace gas emissions: the anthropogenic sources being primarily energy supply, agriculture and industry. The module is run with global, annual emission figures (see Figures 2, 3, 4 and 5).
The emission modules provide the input for the concentration module. The emission and concentration of CO2 is linked to an ocean module and a terrestrial biota module, together reflecting the C-cycle. The latter has been modelled according to Goudriaan and Ketner (1984). As a consequence of the coupling of the emission-, concentration- and terrestrial biota modules the airborne fraction has been defined in a specific sense: the airborne fraction is the fraction of the total CO2 emissions from fossil fuels and biosphere changes that remains in the atmosphere. In formula:
The modular structure for the other trace gases is quite different from that of CO2. The removal of the trace gases concerned by atmospheric chemical processes is an essential feature. Spatial dimension of the simulated trace gas concentration is zero. Generally, the trace gas concentrations of CH4, CO2, N2O, CFC-11 and CFC-12 are expressed as
.gif)
The concentration of methane is derived from the global CH4-CO-OH cycle by simulating the main atmospheric chemical prosesses influencing the global concentrations of these trace gases. The removal rates of CH4 and CO are determined by the uptake-, transport- and oxidation rates of these trace gases, the latter being
dependent on the OH-concentration (Thompson and Cicerone, 1986: Khalil and Rasmussen, 1985: Brühl and Crutzen, 1988; Isaksen and Høv, 1987; Rotmans and Eggink, l988; Swart, 1988).
For N2O and CFC the removal rate is supposed to be inversely proportional to the atmospheric lifetime (Rotmans, 1986). For CFC production figures are the input for the emission module, taking into account the delay time between the production and emission for different fractions of CFC-uses (Miller and Mintzer, 1986).
The calculated trace gas concentrations serve as an input for the climate module. The total change in radiative forcing [Delta]Qtot) resulting from concentration changes of CO2, CH4, N2O, CFC-11 and CFC-12 is modelled according to Wigley (1987):
.gif)
The resulting global mean equilibrium surface temperature rise can be calculated from (3) by multiplying the radiative forcing by a climate feedback factor in which the water vapour factor is explicitly taken into account (Dickinson, 1986; Wigley, 1985; Tricot and Berger, 1987; Ramanathan, 1985; Health Council, 1983).
The effects of global warming on the potential sea level are determined by four processes: thermal expansion of ocean water, melting of alpine glaciers, and ablation or accumulation of the Greenland and Antarctica ice caps:
.gif)
The information necessary for the description of the various complicated aspects of the phenomenon sea level rise is derived from Barnett (1983), Gornitz et al.
(1982), Meier (1984), Revelle (1983), Robin (1985), United States Department of Energy (1985), Van der Veen (1986), Barth and Titus (1984), Oerlemans (1987), and has been integrated and aggregated to a high abstraction level.
The thermal expansion effect is divided into a uniform expansion for the mixed layer (0-75 m) of the ocean module of Goudriaan and Ketner (1984) and, by differential equations, a delayed expansion effect for the layers below (75-1000 m), determined by the depth, atmospheric temperature increase, thermal expansion coefficient and a time lag of 25 yr after equilibrium temperature rise (Barth and Titus, 1984).
For the contributions of the glaciers, Greenland and Antarctica differential equations have been incorporated, containing input-output factors (in mm/[Delta]T* °C).
Uncertainties and Deficiencies of the Model
The simulation model is a reflection of the current state of knowledge. Current knowledge being far from complete, the model has structural limitations.
Scenarios for future trace gas emissions, as presented in this paper, are only indicative, while sources are not fully known or quantified. Positive feedback on microbiological emissions of trace gases by temperature rises is not yet taken into account.