Previous Page Table of Contents Next Page


11. Integrating land-use change and evaluating feedbacks in global change models: The IMAGE 2 approach

RIK LEEMANS, GERT JAN VAN DEN BORN AND LEX BOUWMAN
Global Change Department, National Institute of Public Health and the Environment, RIVM, Bilthoven, The Netherlands


Importance of feedbacks and linkages
Model approaches
Concluding remarks
Acknowledgements
References

During the last decades concern about negative impacts of changing atmospheric conditions has increased. Initially most of this concern was focused on local and regional air pollution, but later global aspects were also considered. One of the major global concerns is the increase of 'greenhouse gases' (GHGs: H2O, CO2, CH4, CO, CFCs, N2O, tropospheric O3 and its precursors) that influence the earth's radiative balance. The atmospheric concentrations of these gases have significantly increased since the start of the industrial revolution. Such increase could well lead to large changes in regional and global climate (Houghton et al., 1992).

The sources and sinks of the different GHGs are heterogeneous and difficult to evaluate comprehensively on a global scale. The concentrations of these gases are influenced by a series of chemical processes, many of which are intrinsically linked to each other (Prinn, 1994). Besides these important chemical processes in the atmosphere, the final concentrations are strongly determined by several oceanic and terrestrial processes. For example, oceans take up atmospheric CO2 through diffusive and biological processes and the magnitude of both is determined by climatic properties, the oceanic energy balance and the ocean circulation (e.g., Sarmiento and Bender, 1994; Woods and Barkmann, 1993). Processes altering the terrestrial sources and sinks are important determinants of many different GHG fluxes between the atmosphere and biosphere (Leemans, 1996). Sources, sinks and fluxes of GHGs from the terrestrial biosphere are more difficult to quantify, because they are defined by local soil, vegetation and climatic properties and the actual land use. All these factors strongly differ geographically and interact with each other on specific time scales. Changes in these fluxes are currently dominated by human activities. This means that any comprehensive assessment of future atmospheric levels of GHGs requires incorporation of these heterogenous spatial and temporal properties, and of human activities affecting the terrestrial biosphere.

In this paper the focus is on processes in the terrestrial biosphere that determine the fluxes of GHGs. Firstly, the linkages and feedbacks between the different processes and compartments in the earth system are discussed with most emphasis on those processes which influence properties of the terrestrial biosphere. Then there is a short review of importance of land-use change, specifying its character and the problems involved in adequately evaluating its magnitude. Finally, the development of an earth system model (IMAGE 2; Alcamo, 1994) is presented that integrates climate change issues and evaluates their importance in a more comprehensive manner. Some applications of the model with respect to mitigation of GHGs will be presented and discussed.

Importance of feedbacks and linkages

PHYSICAL AND BIOGEOCHEMICAL FEEDBACKS

Feedback processes influence the exchange rates of GHGs between the atmosphere, biosphere and oceans and modify the residence time of those gases in the different compartments. Feedback processes largely determine the dynamic response of the earth system. These processes include not only the natural geophysical feedbacks, such as changes in climate and ocean circulation, and biogeochemical feedbacks, such as changes in biological activity, atmospheric chemistry and oceanic CO2 uptake, but also the important anthropogenic feedbacks such as changes in human activities, energy use and land use (see next section).

Geophysical feedbacks alter the radiative forcing characteristics of the atmosphere and include changes in characteristics and distribution of clouds, ice and snow. These feedbacks are incorporated in the global climate models and their strength determines the sensitivity of the climate system to changes in radiative forcing (e.g., Lashof, 1989; Houghton et al., 1990, 1992). Biogeochemical feedback processes involve changes in the sources and sinks of greenhouse gases and changes in surface properties, such as albedo and transpiration. Direct and indirect effects are distinguished.

Direct effects are those processes that influence atmospheric and ocean chemistry (Lashof, 1989; Prinn, 1994) and biospheric processes. The most important feedback process is probably the enhancement of plant growth by increased levels of atmospheric CO2 This so-called CO2 fertilization effect has been observed in many greenhouse and ecosystem experiments (e.g., Strain and Cure, 1985, Drake, 1992). Increased CO2 levels also improve the water-use efficiency (WUE) of plants and change plant growth under water-limited conditions (e.g., Morison, 1985).

The indirect biogeochemical feedbacks involve the consequences of climate change. A large number of those effects have been identified. Examples are changes in ocean circulation, shifts of vegetation zones, and changes in plant growth rates. A temperature increase influences photosynthesis and respiration in a complex manner (Larcher, 1980), enhancing plant growth in regions with temperature constraints. However, enhanced plant growth does not automatically lead to an increase in C storage within an ecosystem. Litter decay is simultaneously stimulated (Raich and Schlesinger, 1992).

Finally climate change alters, like an increased WUE, the distribution of vegetation and agricultural crops (Leemans and van den Born, 1994), Climate strongly defines the large-scale vegetation and crop distribution patterns and this correlation has be used to assess the global impacts of climate change (for a review see Leemans et al., 1996).

Some of these responses are illustrated in Figures 11.1 and 11.2, where the BIOME model (Prentice et al., 1992) is applied to assess changes in the distribution of potential vegetation, characteristic for prevailing stable climatic conditions. Human impacts and other (e.g., successional) influences are not taken into account. The effects of changes in WUE are simulated by shifting the drought tolerances in the BIOME model (Klein Goldewijk et al., 1994). The maps (Figure 11.1) clearly show the large regional differences. Vegetation shifts due to a warmer climate can mainly be observed in the boreal zone, while the effects of increased WUE are mainly located in the tropics, where forests and savanna ecosystems expand. These changes are summarized in Figure 11.2 and could have a pronounced effect on the global C cycle. It was often stated that under a warmer climate the resulting equilibrium biosphere could store significantly more C (e.g., Smith et al., 1992), but it becomes apparent that the transient vegetation response could well be different. Transient vegetation dynamics could temporally generate large CO2 emissions through a rapid decline of ecosystems (Neilson, 1993; Smith and Shugart, 1993), while the successive recovery could last many decades.

This short review indicates that both direct and indirect effects must be considered when evaluating the impacts and consequences of increasing levels of GHGs on ecosystems and agrosystems. This is especially true when scenarios for future trends are developed and evaluated with integrated assessment models for the earth system. Such models too often lack appropriate parameterization of such feedback processes and can therefore generate misleading results. Despite the limitations of these models, they are still the only tool to assess the whole range of climate issues comprehensively, from sources, through processes to impacts and their interactions. In order to create adequate assessments, the developers of global earth system models and disciplinary experts must communicate and try to understand on one hand the modeller's need for simplification and generalization and, on the other hand, the expert's need for adequate observations and experimentation.

IMPORTANCE OF LAND-USE CHANGE

Sustaining an increasing population requires a continuous flow of agricultural and forest products, such as food, fodder, timber and fuelwood. These products all require land to be produced. Increasing this flow requires more productivity. Productivity can increase through intensification of agricultural practices, improved cropping systems and increase in agricultural areas. Such changes influence the biogeochemical properties of land and are therefore important determinants of the size of sources, sinks and fluxes of GHGs. Up to now only the obvious and large-scale conversions of land (deforestation, reclamation of wetlands, etc.) have been addressed with respect to their importance for global C fluxes to the atmosphere.

Figure 11.1. Global maps with the distribution of potential vegetation, based on the BIOME model (Prentice et al., 1992) for different climates and biogeochemical feedbacks. (1) The distribution under current climate

Figure 11.1. Global maps with the distribution of potential vegetation, based on the BIOME model (Prentice et al., 1992) for different climates and biogeochemical feedbacks. (2) Distribution under a doubled CO2 climate according to the GFDL GCM (Manabe and Wetherald, 1987)

Figure 11.1. Global maps with the distribution of potential vegetation, based on the BIOME model (Prentice et al., 1992) for different climates and biogeochemical feedbacks. (3) Distribution under an enhanced WUE effect for doubled CO2

Figure 11.1. Global maps with the distribution of potential vegetation, based on the BIOME model (Prentice et al., 1992) for different climates and biogeochemical feedbacks. (4) Combined distribution of (2) and (3)

Figure 11.2. Shifts in distribution of potential vegetation for doubled CO2 conditions (GFDL: only climate change; WUE: enhanced WUE; GFDL and WUE: combined effect)

Tropical deforestation resulted in a large flux of CO2 and other GHGs. Forests are felled, burned and removed to create pasture and croplands. These land-use practices are common under swidden agricultural systems. During recent decades, however, deforestation rates have been accelerating and the resulting flux of CO2 accounted for 25% of the total increase of atmospheric CO2 (Watson et al., 1992). After a while large parts of the reclaimed land were abandoned, often degraded, and secondary forest developed. Recently more attention has been paid to the balance between deforestation and forestation in the tropics (e.g., Skole and Tucker, 1993). This balance is important with respect to the total global fluxes. The resulting heterogenic spatial and temporal patterns make a precise assessment of the total flux from deforestation difficult.

Many evaluations of the total 'land-use' flux are calculated through the total global C budget. Data are used from the observed atmospheric concentrations, the assumed (=modelled) uptake by the oceans and the well-defined emissions from fossil-fuel combustion. The global budget is then balanced with a derived flux from the terrestrial biosphere. The magnitude of the terrestrial flux needed from terrestrial biosphere is generally smaller than the deduced emissions from tropical deforestation. The conclusion of such a deconvolution exercise is that a significant sink is needed to balance the global C cycle. This missing sink is assumed to be located in the terrestrial biosphere, because the processes determining all other fluxes appear to be well understood. In recent years many studies claimed to have identified at least part of this missing sink (e.g., Kauppi et al., 1992; Wofsy et al., 1993; Fisher et al., 1994), but the actual location and the processes involved are difficult to localize and observe. The missing sink emphasizes the importance of land use, land cover and land management in assessing sources, sinks and fluxes of GHGs (Leemans, 1996).

The above discussion has mainly focused on C, but similar conclusions can be drawn for other greenhouse gases, such as CH4 and N2O (Bouwman, 1995). For example, changes in land cover alter the uptake of CH4 by soils, different agricultural practices lead to changed CH4 emissions, N2O emissions are influenced by the timing and amount of fertilizer applications. Such examples indicate that the spatial pattern of GHG emissions from the terrestrial biosphere is very heterogeneous and influenced by physical, biogeochemical, socio-economic and technical factors. Actual land use and its resulting land cover are important controls. Especially when mitigation policies are evaluated, globally aggregated assessments are no longer valid. State-of-the-art assessments should be dynamic, geographical and regionally explicit and include the most important aspects of the physical subsystem, the biogeochemical subsystem and land use and changes therein (Figure 11.3).

Model approaches

Many approaches to model the consequences of global climate change have been developed. Most of these models only cover one or a few aspects of the whole chain of emissions, processes and impacts. Emission models largely focus on the energy sector and are therefore often derived from macro-economic or technology models (see reviews in Nakícenovíc et al., 1994). Emissions from land use are not adequately covered in such emission models. More process-oriented models simulate aspects of atmospheric chemistry, climate and the biogeochemical cycles. The most advanced models are General Circulation Models (GCMs) that simulate the temporal and spatial characteristics of the climate system, and C-cycle models, that simulate the global C budget. The most diverse group of models addresses impacts. Many different sectors have their own model approaches which all simulate one or more aspects of changing atmospheric composition and/or climate change. In particular, agricultural, ecological and hydrological impact models have been developed (e.g. Parry et al., 1990; Cramer and Solomon, 1993; Kwadijk, 1993). Efforts are now under way to harmonize the large diversity in impact assessment approaches and probably an improved, standardized methodology will be agreed upon (Carter et al., 1994).

Figure 11.3. Flow diagram of the system earth (redrawn after IGBP, 1994)

Recently an approach has emerged that integrates different sources of GHG emissions with the processes in the atmosphere and, if evaluated as necessary, with different impacts. Among the first of such models was IMAGE (Integrated Model to Assess the Greenhouse Effect; Rotmans et al., 1990). This model simulated the dynamics of different GHG emission sources and computed dynamically the final atmospheric GHG concentrations considering the influence of atmospheric chemistry and the C cycle. These concentrations were fed into a radiative forcing model that computed a globally average annual climatic change. The only impact considered was sea-level rise. The model was applied to develop the first global IPCC scenarios (Leggett et al., 1992).

The innovative approach of this IMAGE model has now several competitors. The successor of IMAGE was ESCAPE, an integrated model that was developed for policy evaluation in the EU (Rotmans et al., 1994). The structure of IMAGE has been improved by including different impact modules in a geographically explicit way. The ESCAPE framework has been the basis for two further model developments. Hulme et al. (1994) concluded that the complete ESCAPE framework was too complex and developed a derived, highly parameterized global model that is suitable to determine the sensitivity of the climate system and rapidly evaluate the consequences of different policy options. However, feedbacks and linkages between subsystems are not adequately covered in this global model.

Currently several groups are involved in developing such integrated assessment models with a more adequate treatment of feedbacks and linkages, further integrating the structure of the ESCAPE framework. Some of them focus on only a single region, such as the Asian-Pacific region (the AIM model, Morita et al., 1994) or North America (GCAM, Edmonds et al., 1994), but include high-resolution impact modules and linkages between them, while the trends of other regions in GHG emissions are prescribed and the global biogeochemical and physical processes are simulated in a more aggregated manner. Other global models are being developed by several institutes (Dowlatabadi and Morgan, 1993). One of the few fully implemented and documented integrated global models is IMAGE 2 (Alcamo, 1994). Its innovation includes a dynamic simulation of land cover change, driven by socio-economic, atmospheric and climatic factors. The IMAGE 2 framework is especially developed to link all relevant processes comprehensively, while simultaneously retaining regional and local characteristics.

THE IMAGE 2 APPROACH

The overall objective of the IMAGE 2 model is to simulate, on the basis of political and socio-economic scenarios, plausible future trends of GHG concentrations in the atmosphere and to determine their impacts on physical, biogeochemical and human systems. To accomplish this, important scientific and policy aspects of global climate change must be linked. The time horizon of the simulations should be several decades in order to be able unambiguously to discriminate the effectiveness of policy measures. This objective leads to several scientific goals, such as to provide insights into the relative importance of different linkages and feedbacks in the earth system, to provide estimates of sources of uncertainty and to help identify gaps in our knowledge and data in order to help set the agenda for climate change research. The related policy-oriented goals are to provide a quantitative basis for analysing the societal cost and benefits of various measures to address climate change.

Here only the purpose of each IMAGE submodel is summarized. For a full description of all models and their linkages, reference is made to the papers in Alcamo (1994). The IMAGE model consists of three fully linked submodels (Figure 11.4): the Energy-Industry System (EIS), the Terrestrial Environment System (TES) and the Atmosphere-Ocean System (AOS). The input data of the socio-economic models are mainly prescribed economic, technologic and demographic trends combined with different control policies, while the environmental models are mainly driven by climatic, vegetation, crop and soil properties.

Figure 11.4. Framework of models and linkages in IMAGE 2 (after Alcamo, 1994)

The models making up EIS compute the emissions of GHGs in 13 larger, somewhat homogeneous regions (Table 11.1) as a function of energy consumption and industrial production. End-use energy consumption is computed from various economic driving forces. Sector-specific emission factors are obtained from literature and calibrated against data for global and regional emissions. The models are designed such that the effectiveness of improved energy efficiency and technological development on future emissions in each regions can be evaluated and that the set of models can be used to assess the consequences of different policies and socio-economic trends on future emissions.

TES simulates the changes in global land cover based on climatic and socio-economic factors. The roles of land cover and management are used to compute the fluxes of CO2 and other greenhouse gases from the terrestrial biosphere to the atmosphere. The structure of TES is discussed in more detail below.

The AOS computes the build-up of GHGs in the atmosphere and the resulting climate change averaged over latitudinal bands. AOS thus computes transient changes in climate resulting from changes in all GHG emissions. As a starting point the atmospheric concentration of CO2 is altered through CO2 uptake by the oceans. Then, temporal trends of the average tropospheric concentrations of GHGs are computed accounting for the chemical atmospheric reactions involving O3, OH and CO with other gases and sulphate aerosols. The final atmospheric levels of the GHGs (including HO) are used to compute the earth's energy balance. The cooling effect of backscattering of solar radiation on sulphate aerosols is included (Taylor and Penner, 1994). Surface heat exchange with land and oceans, and changes in albedo from land cover change, are considered in the final calculation of temperature change. Other climatic characteristics (e.g., precipitation) are crudely coupled by scaling the latitudinal temperature change towards the results from GCMs.

Table 11.1. Main characteristics of the regions represented in the IMAGE model and some scenario assumptions. The figures given here are taken from the conventional wisdom scenario, which is derived from the IPCC baseline scenario (IS92a; Leggett et al., 1992). (The yield increase index is relative to 1990)


Population (x106)

Economic growth (%)

Yield increase (%)

Region

1990

2100

1990-2025

2025-2100

2025

2100

Canada

27

27

2.06

1.31

1.83

1.91

USA

250

295

2.09

1.25

1.62

1.66

Latin America

448

877

1.85

2.20

2.14

2.95

Africa

642

2875

1.57

2.39

1.50

1.73

Western Europe

378

388

2.06

1.31

1.45

1.57

Eastern Europe

123

148

1.87

1.18

1.42

1.56

CIS

289

347

1.87

1.18

1.98

2.31

Near East

203

937

1.36

1.98

1.41

1.67

India and South

1 171

2644

2.97

2.84

2.34

2.98

Asia

China and centrally planned countries

1248

1963

4.23

3.07

1.50

1.93

East Asia

371

837

2.97

2.84

2.00

2.81

Oceania

23

24

2.71

1.28

1.64

1.84

Japan

124

130

2.71

1.28

1.02

0.99

Global

5 297

11492





TES and EIS are linked to each other by the demand for biofuels (fuelwood and biomass). This linkage allows IMAGE 2 to evaluate the land cover consequences for future demands for biofuels. EIS and TES are linked in several ways with AOS. The most important linkages are through the emissions of GHGs and changed albedo of land cover patterns. AOS provides EIS and TES with a changed climate that is directly used to determine the energy demands for heating and cooling, compute potential vegetation patterns, crop distribution and yields, and modify the rates of N2O emissions. The final concentrations of CO2 also directly influence plant productivity. These dynamic linkages allow IMAGE 2 to be directly applied to assess the sensitivity of the system for different feedback processes.

All submodels have their specific domain with their own spatial and temporal resolution, scale and dimensionality. Despite this heterogeneity, we have tried to integrate all submodels through their obvious linkages. All socio-economic models (EIS and the agricultural demand and land-use models) distinguish 13 major regions. These models are thus regionally explicit and can be calibrated and validated with a large series of different tabular databases for countries and regions. All environmental models for the terrestrial biosphere are implemented on a 0.5x0.5° grid. Each grid cell is characterized by its climate (Leemans and Cramer, 1991), soil (Zobler, 1986), and land cover (Olson et al., 1985). Spatial heterogeneity in environmental characteristics and local environmental influences on biogeochemical processes are thus adequately covered. Innovative rule-based systems are used to integrate all these dimensions comprehensively.

Modelling land cover change

Modelling land cover change is difficult. Although many models simulate some effect of land-use change, such changes involve often only prescribed tropical deforestation. Contraction, expansion and shifts of agricultural lands, intensification, and different crops and forest management practices are rarely included. Such changes have to be included in global emission models because their cumulative effects are significant. We have developed a land cover change model that is forced by a heterogeneous set of physical, biogeochemical and socio-economic variables. Although, many other models and studies (e.g., FAO, 1993a; Rosenzweig and Parry, 1994) stress the importance of macro-economic driving forces, we have not emphasized those.

Our approach is based on a set of well-established ecological and agricultural models. Potential vegetation patterns are modelled with the BIOME model (Prentice et al., 1992). This model uses climatic variables to delimit the distribution of different Plant Functional Types (PFTs, cf. deciduous broad-leaved trees, evergreen needle-leaved trees, C4 grasses). Each PFT is characterized by its specific climatic constraints defined by the accumulated heat during the growing season, drought, and frost tolerances. Biomes are assemblies of these PFTs (Figure 11.1). BIOME allows us to evaluate shifts in vegetation due to climatic change and changes in water-use efficiency (Figures 11.1 and 11.2) and the consequences of these shifts on total C sequestration.

Crop distributions are modelled using the agro-ecological zone approach (Brinkman, 1987; Leemans and Solomon, 1993). Crops require a minimum growing period (the period with adequate moisture and heat supply) to grow and mature. This growing period, together with several other climatic parameters (see Leemans and Solomon, 1993), defines the potential rainfed distribution of crops. Yield is computed with a simple model (FAO, 1987) that distinguishes different crop types (C3 or C4, tropical or temperate, legumes or non-legumes) and estimates biomass production for each crop under the prevailing climatic conditions during the growing period. The economic yield is a fraction of this production and is adjusted for soil quality. The result of these two models is thus a potential vegetation and crop yield for each 0.5x0.5° gridcell.

Transformation of land cover is governed by the local, regional and global demand for agricultural commodities (grain, meat, wood and biofuels). This demand is satisfied by specific land use in croplands, pasture lands, rangelands and (managed) forests. Land use is strongly related to the potential productivities and technically achievable yields for each gridcell. We developed an agricultural demand model to estimate the societal demand for agricultural products. These calculations are performed for each world region and use the same socio-economic input as EIS (Table 11.1).

Regional agricultural demand is calculated from the per caput human consumption for different crops and meat products based on a assumed elasticity relationship between consumption and per caput income. These elasticity coefficients are estimated from 1970-1990 data of FAO (1991), and are the main adjustable parameters in this demand model. Total human (as opposed to animal) demands for these products are then computed for a given income and population scenario. A similar procedure is used to compute the total regional meat demand. This demand is then converted into numbers of livestock, and their feed requirements (concentrates from crops and residues and roughage from range and pasture lands). These feed requirements are added to the total regional human crop demand. This demand is then adjusted for trade between regions. The roughage demand is converted into a demand for grassland by multiplying productivity per animal and the average required rangelands per animal.

The total regional demand for agricultural products is reconciled with the potential local distribution and yields of crops, computed earlier. Land cover is initialized for 1970 conditions using an aggregated version of the database of Olson et al. (1985). The agricultural crop demands are assumed to be satisfied in the single class 'agricultural land', while the demand for rangelands can be satisfied by several types of grasslands, depending on what occurs locally. The basic idea of the dynamic land cover change model is to change land cover until the total regional demands are satisfied. The model thus generates future land cover logically consistent with demand and production potential. This is probably not realistic but it is congruous with available global data sets and some basic driving forces (Turner et al., 1990). Improvement of the analyses, incorporating different land-use systems, political structures and cultural characteristics, is far beyond the current requirements and analysis in IMAGE 2. However, we are eager to learn from a better understanding from emerging projects (e.g. Turner et al., 1993) and experiences with other models (e.g., Fischer et al., 1988).

Actual changes of future land cover patterns are difficult to estimate, because the actual driving forces are heterogeneous, differ per region and are in general poorly understood (Turner et al., 1993). Our approach is therefore to prescribe a set of transparent logical rules to match regional land demand with its potential. These rules, together with a management factor (which takes into account the yield increase by fertilizer application and technological improvement, cf. Table 11.1), define future patterns. For example, there are two weighted rules that guide the assignment of new agricultural areas: (1) new areas should occur adjacent, if possible, to existing agricultural land, because of the availability of infrastructure, transport and population; (2) the new land should develop in those areas of the highest production potential. Similar sets of rules are specified to satisfy rangelands and fuelwood demand, but the demand for cropland is satisfied first. Land not under agricultural management is assumed to succeed towards the potential vegetation as determined by the BIOME model. This land cover model results in a dynamically changing land cover on the terrestrial grid. The land cover patterns are consistent with a changed climate and are used to drive the terrestrial C cycle. One major improvement of this approach over earlier models is that an increase of agricultural area does not per se lead to deforestation, but it is a function of the agro-ecological suitability of all types of land. Agriculture could thus expand into grasslands as well as forests.

Carbon fluxes

The C-cycle model estimate the C sources, sinks and fluxes resulting from natural processes and land cover change. The C-cycle model computes, for each land cover type, Net Ecosystem Productivity (NEP), which is the difference between the Net Primary Productivity (NPP) and soil respiration. NPP is a characteristic of each land cover type, but its actual value is corrected according to local (temperature and moisture availability) and global (atmospheric CO2 concentration) conditions and the time passed since the last land cover change. NPP and soil respiration are modified under the influence of climate change, assuming characteristic response functions for photosynthesis, plant respiration and soil respiration (e.g., Larcher, 1980). CO2 fertilization is a function of available PFTs in each land cover type (such as C3 C4, annual, perennials or trees), temperature, moisture, soil fertility and altitude. The simulated NEP is consistent with the simulated climatic conditions.

NPP is partitioned into land cover specific biomass components (leaves, branches, stems and roots), each with a specific longevity, and non-living matter (litter, humus and charcoal), each with a specific turnover time. We have initialized the model starting from 1900. Most IMAGE 2 simulations start therefore (in 1990) with NEP conditions that are close to equilibrium with prevailing climatic and atmospheric conditions. This near-equilibrium condition shift when the environment changes or land is converted into another type. If natural vegetation is converted, a fraction of the biomass is burned on site, immediately releasing CO2, while the remainder is allocated to the non-living pools. Such change results in an additional long-term low CO2 flux from these sites. If agricultural land is abandoned, we assume that the potential natural vegetation returns. However, such succession takes time and we assume that NPP slowly builds up following a logistic function.

Land-use emissions

The Land Use Emissions model relates global land use to the flux of emissions of CH4, N2O, CO2 NOx, and Volatile Organic Carbons (VOCs, excluding CH4) (cf. Table 11.2). In addition, the model estimates the emissions resulting from biotic processes unrelated to human activity, such as N2O emissions from soils in unmanaged forests, and trace gas emissions from aquatic systems. The calculations are important in determining land use and cover related GHG emissions and can be used to evaluate strategies for reducing these emissions.

Table 11. 2. Sources described in the IMAGE 2 Land Use Emissions Model, gas species emitted and the type of calculation and presentation. GE = geographically explicit; R = regional total; G = global total. Adapted from Kreileman and Bouwman (1994)

Source

Species

Type of calculation

Wetland rice fields

CH4

GE

Natural wetlands

CH4

GE

Landfills

CH4

R

Domestic sewage treatment

CH4

R

Animals

CH4

R

Animal waste

CH4, N2O

R

Termites

CH4

G

Methane hydrates

CH4

G

Aquatic sources

CH4, N2O

G

Biomass burning

CH4, CO, NOx, N2O, VOC



- Deforestation


GE


- Savanna burning


GE


- Agricultural waste burning


GE

Natural soils

N2O

GE

Agricultural fields

N2O

GE

Deforestation

N2O

GE

This model is based on current estimates for the various sources and species. Given the limitations of the available data, the model presents grid-based estimates for a number of different sources, including CH4 emission from rice, wetlands, emissions of CH4, CO, NOx, N2O and VOC from deforestation, savanna burning and agricultural waste burning and N2O from natural soils, N fertilizer and deforestation (Table 11.3; Bouwman, 1995). For some of the sources (landfills, domestic sewage treatment, termites, CH4 hydrates and aquatic sources) geographically explicit calculations are not yet possible because of data limitations (Table 11.2).

APPLICATIONS OF IMAGE 2

The IMAGE 2 model has been tested and calibrated against data from 1900 to 1970 and can for this period reproduce trends in regional energy consumption and energy related emissions, land cover, terrestrial flux of CO2 and emissions of other GHGs into the atmosphere. The time horizon for the actual simulations extends from 1970 to 2100 and each simulation is characterized by a specific set of input options that describe policy and socio-economic trends (Table 11.1). Here we present a scenario which is an adaptation of the intermediate IPCC scenario (IS92a, Leggett et al., 1992). The scenario makes 'conventional' assumptions about future demographic, economic and technological driving forces (Table 11.1; Alcamo, 1994). There are no climate-related policies, except those that were already agreed upon (e.g., Montreal Treaty on CFC emissions). This scenario further assumes an increase in the energy use efficiency and a regionally specific vehicle utilization. This baseline scenario can be used as a baseline against which to compare other scenarios.

Land-use change and emissions

We compared the emissions calculated for the years 1970, 1990 and 2050 for CH4 and N2O with different reference estimates for selected land-use-related sources (Table 11.3; Watson et al., 1992). From these reports either the methodology has been adopted, giving by definition similar results, or emission factors have been adapted to arrive at equal global totals for 1990. In some cases our results slightly differ from those of the reference studies, for example for CH4 emissions from animal waste. Major differences are found for all emissions from deforestation. The global deforestation rate for the baseline scenario is 155 000 km2/year for the period 1970 to 1980 and 140 000 km2/year for 1980 to 1990. Official inventories of deforestation by FAO (1993b) show that rates may have been respectively 115 000 km2/year and 155 000 km2/year. Reference estimates for biomass burning are based on these FAO deforestation estimates for the early 1980s and result therefore in lower emissions than our baseline scenario (Table 11.2).

Our calculations are sensitive to assumptions of demographic, technological and economic development and the productivity of land (crop yields and animal productivity). Slight changes in the balance between agricultural demand and production could cause large differences in land use and associated emissions. There is a disagreement between the baseline scenario and the data from Watson et al. (1992) for future agricultural emissions, as shown for rice (Table 11.3). The assumed increase in rice yield (t/ha) in the baseline scenario for South and East Asia and the Centrally Planned Asian countries is 1.1% per year for the period 1990-2010, or a 25% increase over these 20 years. The assumed annual yield increases by Watson et al. (1992) are respectively 1.2% and 0.5%. Both estimates are far less optimistic than those of FAO (1993b) that assume a global rice yield increase of 1.5% per year for the same period leading to a total yield increase of about 35%. This illustrates the uncertainty in the forecasts of rice productivity and area and the consequences for the associated CH4 emissions (Table 11.3). In general, our baseline also assumes far lower animal productivities than FAO (1993b).

Table 11.3. Estimated global emissions for 1970,1990 and 2050, and reference estimates for 1990 and 2050. Reference estimates for 1990 are from Watson et al. (1992) unless indicated otherwise. Reference estimates for 2050 are from Pepper et al. (1992), scenario IS92a. Emissions are expressed as Tg CH4/yr and Tg N2O-N/yr. The range of uncertainty of the various estimates is not indicated here, but can be found in Watson et al. (1992). Adapted from Kreileman and Bouwman (1994)

Source

Emission (Tg/yr)

1970 IMAGE

1990 IMAGE

1990 REF.

2050 IMAGE

2050 REF.

a. Sources of CH4

Wetland rice fields

53

59

60

52

87

Animals

66

79

80

161

173

Animal waste

12

14

14 a

28

54

Biomass burning





34


Deforestation b,c

16

14

6 d

12

14


Savanna burning

17

16

13 d

6

-


Agricultural waste burning

7

8

9 d

-


b. Sources of N2O

Fertilizer induced

0.4

0.9

2.2 e

2.0

4.2 e

Deforestation 1

0.2

0.2

0.4 f

0.1

1.1

Animal waste

0.5

0.6

0.4 g

1.5


a Gibbs and Woodbury (1993).

b Estimate for deforestation excludes shifting cultivation, which contributes about 10 Tg CH4/yr (Crutzen and Andreae, 1990).

c For 1970 no values are available, since this calculation involves the history of deforestation over a period of some years. For deforestation effects on soil N2O emission as well as for direct biomass burning the year 1975 is presented.

d Crutzen and Andreae (1990).

e Pepper et al. (1992). This estimate of total N2O loss, including fertilizer-induced and background losses from arable land. From this number the amount of circa 1 Tg N2O-N/yr needs to be subtracted to arrive at the 1990 fertilizer-induced loss.

f Bouwman (1995).

g Khalil and Rasmussen (1992).

The increase in crop productivity is coupled to increases in the use of agricultural inputs, represented in the model by N fertilizer and a management factor (Table 11.1). The annual increase in global N-fertilizer use in our baseline scenario is about 2% for the period 1990-2025 and 0.3% for 2025 to 2050. These assumptions lead to a doubling of N-fertilizer use between 1990 and 2025. The associated N2O losses are directly related to fertilizer use. It is not unrealistic to assume that fertilizer use will continue to increase along with the projected increases in crop productivity. However, the projected increase in both the baseline and Watson et al. (1992) for the coming decades is somewhat slower than for the period 1970-1990.

The baseline scenario generates globally a large increase in the extent of agricultural land (Table 11.5). Although the global population more than doubles, the increase in agricultural area is only around 15%. This is comparable to the observed increase during the last half century (Plucknett, 1994). This relatively small increase in extent is accompanied by a large increase of agricultural productivity on existing agricultural land.

Although the global increase is modest, there are large regional differences. Agriculturally highly developed regions (especially North America and Europe) increase their productivity significantly, while the population is largely stable. Part of the increase is exported, but there is still a large amount of land that is abandoned. We assume that this land returns slowly to its potential vegetation (often forest), which creates a significant C sink in these regions. The decrease in agricultural land starts immediately in the simulation. Latin America follows a similar path during the next century, but initially there are still significant amount of deforestation.

Other regions (especially Africa, India and China) display different patterns; the increase in population and income (and its linkage to changed dietary preferences) amplify agricultural demand. This demand cannot be satisfied by an increase in productivity and imports and leads to a large increase in agricultural land. By the end of next century almost all suitable agricultural land will be occupied. This will result in deforestation and consequently large GHG fluxes from these regions. The model thus simulates a shift of the major loci of deforestation from Latin America towards Africa and Southeast Asia. The strongly increased human pressures on land in these regions could have severe consequences for the development of sustainable land management, food security and biodiversity.

Forestation as a mitigation option

The model allows not only the evaluation of the consequences of different socio-economic trends, but can also be used to evaluate different mitigation options. There are many different mitigation options (cf. Nakícenovíc et al., 1994), but their effectiveness or suitability can differ per region. It is therefore required to develop a comprehensive capability to evaluate such options. IMAGE 2 provides some possibilities, from which we present two examples.

One of the most often presented mitigation options is forestation. This option increases the C sequestered in forests and remains effective if the total extent of forests continues to increase (NEP of old-growth forests are in balance with respect to the C cycle), or if the resulting forest products are used in a durable way (construction wood), replace other CO2 sources (woody biomass could replace fossil fuel) or both (construction wood could replace cement).

The simulation of IMAGE 2 satisfies the demand for agricultural products first. This demand drives land cover patterns and the remainder is natural vegetation with its specific C dynamics and storage. Only those regions that climatologically can be forests are suitable for the forestation mitigation option, but a large portion of these regions is used for other purposes. Early estimates of forestation potential were therefore probably much too optimistic because, to offset large amounts of fossil-fuel-related CO2 emissions, large areas have to be forested annually (Sedjo and Solomon, 1989). Forestation in shelterbelts, along fields and roads and in degraded lands is, although important for other reasons, alone not sufficient.

Table 11.4 lists the suitable and available extent for forestation under the baseline scenario. From these figures it is apparent that large regional differences occur. Forestation is a possible option in the high and mid latitudes of developed regions. In the high latitudes there is, in 2050, even a large potential increase due to the polewards shift of boreal and temperate forests due to climatic warming (Figure 11.1). This phenomenon is a negative feedback, which actually makes the forestation option more effective. In Latin America forestation possibilities also improve somewhat, but in many other regions only few opportunities are available. In these regions, C sequestration in shelterbelts and on degraded lands is probably the only valid option.

Biofuels as a mitigation option

Another mitigation option is the use of biomass for energy generation. An efficient use of renewable biofuels and biomass for generation of energy could decrease the dependence on fossil fuels. Such a programme is already being implemented in Brazil, where ethanol from sugar cane is used to replace petrol. Biofuels applied for transportation are derived from oil crops or from fermentation of sugar or starch. The techniques involved are not yet competitive with energy derived from fossil oils. However, new technologies could make very efficient use of biomass as an energy source (Johansson et al., 1993). For example, the foreseen development of fuel cells, in which biomass is converted into a hydrogen-rich gas, promises the possibility of small- and large-scale electricity production with an efficiency ranging from 50 to 80% with few environmental impacts. Many future scenarios assume therefore large amounts of energy derived from renewable biomass. Especially in the second half of the next century when fuel-cell systems should be widely operational, biomass could be the major source of renewable energy (Johansson et al., 1993; Kassler, 1994).

Table 11.4. Land suitability and availability (not used for other purposes) for the forestation mitigation option, according to IMAGE 2 simulation for the conventional wisdom scenario

Region

Available in 1990 (106 km2)

Suitable in 1990 (106 km2)

Available in 2050 (106 km2)

Suitable in 2050 (106 km2)

Canada and USA

6.1

10.2

7.9

11.7

Latin America

6.2

11.1

6.5

14.3

Africa

2.5

4.5

0.2

6.3

Europe

1.3

3.9

1.4

4.2

CIS

8.5

11.5

9.6

14.1

India and South Asia

0.5

0.9

0.0

1.3

China

1.9

4.5

0.2

5.5

East Asia

1.3

3.2

0.5

3.6

We make several assumptions for the use of biomass as an energy source in the baseline scenario. The use of such biomass increases from 5 EJ now (1018 J: current global energy use is 300 EJ) to 75 EJ in 2050 and 210 EJ in 2100. These figures are based on reasonable mixes of energy carriers, the annual growth in energy demand induced by the increasing global population and its income, a decline in oil and natural gas resources and the preference for renewable energy sources over fossil fuels. We assume for the baseline that the required biomass is simply available from crop and forest residues and waste. This amount of biomass thus does not require new cropland. However, the biofuels used in the transportation sector (ethanol) are still derived from sugar cane and maize but this relatively small additional demand has been included in the simulated extent of agricultural land.

We derived two different scenarios from the baseline scenario. These scenarios only differ with respect to their source of biomass-energy level. In the second scenario, biomass crops, we assumed that such large amounts of biomass are not simply available and that a large portion should be grown specifically in the form of short-rotation forests and other biomass crops, such as Miscanthes spp. These crops are effective sources, because they generate a high yield of 40-60 t/ha/year and have a wide potential distribution. We assumed that 40% of the total biomass-related energy demand is satisfied by such crops. The third scenario, no-biomass fuel, is the opposite approach. Renewable energy from biomass is not feasible and the total modem biomass fuels are replace by fossil oils.

Table 11.5 presents some of the results of these three scenarios. The baseline and the biofuel-crop scenario result in lower atmospheric concentrations of CO2 but significantly higher CH4 concentrations, than the no-biofuel scenario. From such results we must conclude that the effectiveness of mitigation options should be evaluated simultaneously for all GHGs. The final concentration of a single GHG is not a good indicator because reducing one species could enhance another. CH4 has a much larger global warming potential. Despite the differences in atmospheric concentrations between the scenarios, the global temperature increase is therefore not significantly different between the scenarios. Most differences are buffered through changes in land cover and the accompanying climatic and biogeochemical feedbacks.

The differences in the simulated extent of agricultural lands are large (Table 11.5). The large amount of additional demand for biomass crops result in a 30 or 65% increase in, respectively, 2050 and 2100. This increase leads to both a larger deforestation and expansion into more arid regions, such as steppes and savannas. Here the importance for a comprehensive evaluation with respect to land use becomes apparent. The increasing land requirements in the biofuel-crop scenario result in a smaller total sink in the terrestrial biosphere. This has consequences for the simulated atmospheric CO2 concentration and is the main reason that this scenario yields higher concentrations than the baseline, where it was assumed that biomass was freely available. The consequences of biomass energy scenarios must be carefully evaluated and land-use issues should not be neglected in such assessment.

Table 11.5. Agricultural extent under different scenario options for the generation of energy. The difference between the scenarios is the source of a large fraction of energy carrier (1. All biomass conies from residuals. 2. Biofuel crops. 60% of the biomass conies from biofuel crops. 3. Biomass is replaced by oil)

Property

Scenario 1
2050

Scenario 2
2050

Scenario 3
2050

Scenario 1
2100

Scenario 2
2100

Scenario 3
2100

Atmospheric CO2 concentration (ppmv)

522

534

539

777

821

857

Atmospheric CH4 concentration (ppmv)

2.5

2.6

2.4

2.3

2.4

1.7

Average surface temperature(°C)

+1.2

+1.3

+1.2

+2.1

+2.3

+2.2

Change in agricultural area (26.7 106 km2)

+9%

+30%

+9%

+14%

+65%

+15%

Change in forest area (47.2 106 km2)

-26%

-32%

-26%

-27%

-31%

-27%

Concluding remarks

The added value of integrated assessment models is thus their ability to generate insights that cannot be easily derived from individual natural or social science component models that have been developed in the past. The current integrated assessment models with respect to global climate change all have a mix of policy and scientific goals. This has lead to a large diversity of different, but complementary approaches. Only a few models include a simulation of land use and land cover change. This omission limits a comprehensive evaluation of feedbacks and effective linkages among economic, energy, agriculture and forestry sectors and withholds adequate evaluation of many C cycle or land cover related mitigation options. The above examples illustrate the usefulness of these integrated models in creating a greater understanding of the linkages and relationships among the climate system, the biosphere and human activities. Models such as IMAGE 2 (Alcamo, 1994) and AIM (Morita et al., 1994) will be enhanced continuously. More consistent scenario developments, better algorithms and better initialization and validation data will improve future applications. A major conclusion from such analysis should be that from the integration of a series of disciplinary expertise and insights emerges a better understanding, which could further disciplinary and multidisciplinary research.

Without such understanding, integrated global change assessments will continue to emphasize energy-industry-transport emissions, a globally aggregated C cycle with limited land-use change, and the climate system, while natural ecosystems and agrosystems will only be addressed in impact studies. Excluding them from integrated assessments of climate change will lead to non-sustainable climate policies. Fortunately, major international research programmes (Diversitas, WCRP, IGBP and HDP) are now addressing land use and land cover change more systematically, which means that we now enter an exciting new era in global change research.

Acknowledgements

The preparation of this paper was funded by the Dutch Ministry of Housing, Planning and the Environment under contract MAP410 to RIVM and the National Research Programme 'Global Air Pollution and Climate Change'. The Terrestrial Environment System of the IMAGE 2 model is an official contribution to IGBP-GCTE core research.

References

Alcamo, J. (ed.). 1994. IMAGE 2.0: Integrated Modeling of Global Climate Change. Kluwer Academic Publishers, Dordrecht.

Bouwman, A.F. 1995. Compilation of a Global Inventory of Emissions of Nitrous Oxide. Ph.D. Thesis. Agricultural University, Wageningen.

Brinkman, R. 1987. Agro-ecological characterization, classification and mapping. Different approaches by the international agricultural research centres. In: Agricultural Environments. Characterization, Classification and Mapping. A.H. Bunting (ed.). CAB International, Wallingford, UK. pp. 31-42.

Carter, T.R., Parry, M.L., Harasawa, H. and Nishioka, S. 1994. IPCC technical guidelines for assessing impacts of climate change. IPCC Special Report CGER-1015-'94. Intergovernmental Panel on Climate Change, WMO and UNEP, Geneva.

Cramer, W. and Solomon, A.M. 1993. Climatic classification and future global redistribution of agricultural land. Clim. Res. 3: 97-110.

Crutzen, P.J. and Andreae, M.O. 1990. Biomass burning in the tropics: impact on atmospheric and biochemical cycles. Science 250: 1669-1678.

Dowlatabadi, H. and Morgan, M.G. 1993. Integrated assessment of climate change. Science 259: 1813 and 1932.

Drake, B.G. 1992. The impact of rising CO2 on ecosystem production. Water Air Soil Pol. 64: 25-44.

Edmonds, J., Wise, M. and MacCracken, C. 1994. Advanced energy technologies and climate change: an analysis using the global change assessment model (GCAM). Report PNL-22. Pacific Northwest Laboratory, Washington DC.

FAO. 1987. Report on the agro-ecological zones project. Vol. 3. Methodology and results for South and Central America. World Soil Resources Report 48/3. FAO, Rome.

FAO. 1991. Agrostat PC. Land Use. Computerized Information Series 1/7. FAO, Rome.

FAO. 1993a. Agriculture: Towards 2010. Conference Report C93/24. FAO, Rome.

FAO. 1993b. Summary of the final report of the forest resources assessment 1990 project for the tropical world, 11th session of the Committee on Forestry. Report D1/U9798. FAO, Rome.

Fischer, G., Froberg, K., Keyzer, M.A. and Pahrik, K.S. 1988. Linked National Models: A Tool for International Food Policy Analysis. Kluwer, Dordrecht.

Fisher, M.J., Rao, I.M., Ayarza, M.A., Lascano, C.E., Sanz, J.I., Thomas, R.J. and Vera, R.R. 1994. Carbon storage by introduced deep-rooted grasses in the South American savannas. Nature 371: 236-238.

Gibbs, M.J. and Woodbury, J.W. 1993. Methane emissions from livestock manure. In: Proceedings of the International Workshop on Methane and Nitrous Oxide. Methods for National Inventories and Options for Control. A. Van Amstel (ed.). National Institute of Public Health and Environmental Protection, Bilthoven. pp. 81-91.

Houghton, J.T., Jenkins, G.J. and Ephraums, J.J. (eds.). 1990. Climate Change: The IPCC Scientific Assessment. Cambridge University Press, Cambridge.

Houghton, J.T, Callander, B.A. and Varney, S.K. (eds.). 1992. Climate Change 1992: The Supplementary Report to the IPCC Scientific Assessment. Cambridge University Press, Cambridge.

Hulme, M., Raper, S.C.B. and Wigley, T.M.L. 1994. An integrated framework to address climate change (ESCAPE) and further developments of the global and regional climate modules (MAGICC). In: Integrative Assessment of Mitigation, Impacts and Adaptation to Climate Change. N. Nakicenovic (ed.). IIASA, Laxenburg. pp. 112-134.

IGBP. 1994. Global modelling and data activities 1994-1998. IGBP Report 30. International Geosphere-Biosphere Programme, Stockholm.

Johansson, T.B., Kelly, H., Reddy, A.K.N. and Williams, R.H. (eds.). 1993. Renewable Energy: Sources for Fuels and Electricity. Island Press, Washington DC.

Kassler, P. 1994. Energy for Development. Shell Selected Paper, Shell, London.

Kauppi, P., Mielikäinen, K. and Kuusela, K. 1992. Biomass and carbon budget of European forests, 1971 to 1990. Science 256: 70-74.

Khalil, M.A.K. and Rasmussen, R.A. 1992. The global sources of nitrous oxide. J. Geophys. Res. 97: 14651-14660.

Klein Goldewijk, K., van Minnen, J.G., Kreileman, G.J.J., Vloedbeld, M. and Leemans, R. 1994. Simulating the carbon flux between the terrestrial environment and the atmosphere. Wat. Air Soil Pollut. 76: 199-230.

Kreileman, G.J.J. and Bouwman, A.F. 1994. Computing land use emissions of greenhouse gases. Wat. Air Soil Pollut. 76: 231-258.

Kwadijk, J. 1993. The impact of climate change on the discharge of the River Rhine. Nederl. Geogr. Stud. 171: 1-201.

Larcher, W. 1980. Physiological Plant Ecology. Springer-Verlag, Berlin.

Lashof, D.A. 1989. The dynamics greenhouse: Feedback proceses that may influence future concentrations of atmospheric trace gases and climatic change. Clim. Change 14: 213-214.

Leemans, R. 1996. Incorporating land-use change in Earth system models illustrated by IMAGE 2. In: Global Change and Terrestrial Ecosystems. B. Walker and W. Steffen (eds.). Cambridge University Press, Cambridge (in press).

Leemans, R. and Cramer, W. 1991. The IIASA database for mean monthly values of temperature, precipitation and cloudiness on a global terrestrial grid. Research Report RR-91-18. International Institute for Applied Systems Analysis, Laxenburg.

Leemans, R. and Solomon, A.M. 1993. The potential response and redistribution of crops under a doubled CO2 climate. Clim. Res. 3: 79-96.

Leemans, R. and van den Born, G.J. 1994. Determining the potential global distribution of natural vegetation, crops and agricultural productivity. Wat. Air Soil Pollut. 76: 133-161.

Leemans, R., Cramer, W. and van Minnen, J.G. 1996. Prediction of global biome distribution using bioclimatic equilibrium models. In: Effects of Global Change on Coniferous Forests and Grassland. J.M. Melillo and A. Breymeyer (eds.). John Wiley, New York (in press).

Leggett, J., Pepper, W.J. and Swart, R.J. 1992. Emissions scenarios for the IPCC: an update. In: Climate Change 1992. The Supplementary Report to the IPCC Scientific Assessment. J.T. Houghton, B.A. Callander and S.K. Varney (eds.). Cambridge University Press, Cambridge. pp. 71-95.

Manabe, S. and Wetherald, R.T. 1987. Large-scale changes in soil wetness induced by an increase in carbon dioxide. J. Atmosph. Sci. 44: 1211-1235.

Morison, J.I.L. 1985. Sensitivity of stomata and water use efficiency to high CO2 Plant Cell Environ. 8: 467-474.

Morita, T., Matsuoka, Y., Kainuma, M., Kai, K., Harasawa, H. and Dong-Kun, L. 1994. Asian-Pacific Integrated Model to assess policy options for stabilizing global climate. AIM Report 1.0. National Institute for Environmental Studies, Tsukuba, Japan.

Nakícenovíc, N., Nordhaus, W.D., Richels, R. and Toth, F.L. (eds.). 1994. Integrative Assessment of Mitigation, Impact, and Adaptation to Climate Change. International Institute for Applied System Analysis, Laxenburg.

Neilson, R.P. 1993. Vegetation redistribution: a possible biosphere source of CO2 during climatic change. Wat. Air Soil Pollut. 70: 659-673.

Olson, J., Watts, J.A. and Allison, L.J. 1985. Major World Ecosystem Complexes Ranked by Carbon in Live Vegetation: A Database. Report NDP-017. Oak Ridge National Laboratory, Oak Ridge, Tennessee.

Parry, M.L., Porter, J.H. arid Carter, T.M. 1990. Climatic change and its implications for agriculture. Outlook on Agriculture 19: 9-15.

Pepper, W., Leggett, J., Swart, R., Watson, J., Edmonds, J. and Mintzer, I. 1992. Emission scenarios for the IPCC: an update. Assumptions, methodology and results. Report IPCC Working Group I. Geneva.

Plucknett, D.L. 1994. Science and agricultural transformation. IFPRI Lecture Series. International Food Policy Research Institute, Washington DC.

Prentice, I.C., Cramer, W., Harrison, S.P., Leemans, R., Monserud, R.A. and Solomon, A.M. 1992. A global biome model based on plant physiology and dominance, soil properties and climate. J. Biogeogr. 19: 117-134.

Prinn, R.G. 1994. The interactive atmosphere: global atmospheric-biospheric chemistry. Ambio 23: 50-61.

Raich, J.W. and Schlesinger, W.H. 1992. The global carbon dioxide flux in soil respiration and its relationship to vegetation and climate. Tellus 44B: 81-99.

Rosenzweig, C. and Parry, M.L. 1994. Potential impact of climate change on world food supply. Nature 367: 133-138.

Rotmans, J., de Boois, H. and Swart, R.J. 1990. An integrated model for the assessment of the greenhouse effect: the Dutch approach. Clim. Change 16: 331-356.

Rotmans, J., Hulme, M. and Downing, T.E. 1994. Climate change implications for Europe: An application of the ESCAPE model. Glob. Env. Change 4: 97-124.

Sarmiento, J.L. and Bender, M. 1994. Carbon biogeochemistry and climate change. Photosynth. Res. 39: 209-234.

Sedjo, R.A. and Solomon, A.M. 1989. Climate and forests. In: Greenhouse Warming: Abatement and Adaptation. N.J. Rosenberg, W.E. Easterling III, PR. Crosson and J. Darmstadter (eds.). Resources for the Future, Washington DC. pp 105-119.

Skole, D. and Tucker, C. 1993. Tropical deforestation and habitat fragmentation in the Amazon: Satellite data from 1978 to 1988. Science 260: 1905-1910.

Smith, T.M. and Shugart, H.H. 1993. The transient response of terrestrial carbon storage to a perturbed climate. Nature 361: 523-526.

Smith, T.M., Leemans, R. and Shugart, H.H. 1992. Sensitivity of terrestrial carbon storage to CO2 induced climate change: Comparison of four scenarios based on general circulation models. Clim. Change 21: 367-384.

Strain, B.R. and Cure, J.D. 1985. Direct Effects of Increasing Carbon Dioxide on Vegetation. Report DOE/ER-0238. US Department of Energy, Washington DC.

Taylor, K.E. and Penner, J.E. 1994. Response of the climate system to atmospheric aerosols and greenhouse gases. Nature 369: 734-737.

Turner, B.L. II, dark, W.C., Kates, R.W., Richards, J.F., Mathews, J.T. and Meyer, W.B. (eds.). 1990. The Earth as Transformed by Human Action: Global and Regional Changes in the Biosphere over the Past 300 Years. Cambridge University Press, Cambridge.

Turner, B.L., Moss, R.H. and Skole, D.L. 1993. Relating Land Use and Global Change: A Proposal for an IGBP-HDP Core Project. IGBP Report No. 24 and HDP Report No. 5. International Geosphere-Biosphere Programme and the Human Dimensions of Global Environmental Change Programme, Stockholm.

Watson, R.T., Meira Filho, L.G., Sanheuza, E. and Janetos, A. 1992. Greenhouse gases: sources and sinks. In: Climate Change 1992. The Supplementary Report to the IPCC Scientific Assessment. J.T. Houghton, B.A. Callander and S.K. Varney (eds.). Cambridge University Press, Cambridge. pp. 27-46.

Wofsy, S.C., Goulden, M.L., Munger, J.W., Fan, S.M., Bakwin, P.S., Daube, B.C., Bassow, S.L. and Bazzaz, F.A. 1993. Net exchange of CO2 in a mid-latitude forest. Science 260: 1314-1317.

Woods, J. and Barkmann, W. 1993. The plankton multiplier - positive feedback in the greenhouse. J. Plankton Res. 15: 1053-1074.

Zobler, L. 1986. A World Soil File for Global Climate Modeling. Technical Memorandum. NASA, New York.


Previous Page Top of Page Next Page