312 | IAASTD Global Report

were developed for, their integration from the perspective of the IAASTD's objective is still rather low. In particular, feedbacks from ecological changes to socioeconomic drivers are limited, with some exceptions on the impacts of food production and climate policy on socioeconomic drivers.

     Processes that change ecosystems and their  services mostly occur at highly disaggregated levels. Models therefore require regional specificity. A tendency to increase the level of explicit geographic information in models, for instance by using a detailed grid, can be seen in the literature. Un­derstanding interregional links, but also regional differences will be an important research issue for integrated modeling in the coming years. A nested approach to integrated assess­ment modeling could be a helpful way forward, in which global models provide context for detailed, regional (eco­logical) models.

     Uncertainties are a key element in IAMs, given the high complexity and its focus on decision-making. These uncer­tainties include, for example, variability of parameters, in­accuracy of model specification or lack of knowledge with regard to model boundaries. Although the existence of un­certainties has been recognized early in the process of devel­oping IAMs, uncertainty analysis is typically included only partially or not at all.

5.2.1.3 CGE models

CGE models are widely used as an analytical framework to study economic issues of national, regional and global di­mension. CGE models provide a representation of national economies, next to a specification of trade relations between economies. CGE models are specifically concerned with re­source allocation issues, that is, where the allocation of pro­duction factors over alternative uses is affected by certain policies or exogenous developments. International trade is typically an area where such induced effects are important consequences of policy choices. These models provide an economy-wide perspective and are very useful when:

•     The numerous, and often intricate, interactions between various parts of an economy are of critical importance. As for agriculture, such interactions occur between ag­riculture sectors themselves (e.g., competing for limited productive resources including various types of land) as well as between agricultural sectors with other sectors/ actors which either service agricultural sectors or oper­ate in the food and fiber chain including downstream processors, traders and distributors, final consumers and governments (e.g., public policies).

•     The research objective is to analyze counterfactual poli­cy alternatives and/or plausible scenarios about how the future is likely to evolve. Examples could include the implications for agriculture of likely multilateral trade liberalization in the future, the implications for agri­culture of future growth in food demand and shifts in consumer preference, or the role of bioenergy in climate change mitigation and implications for agriculture.

For analyzing such issues, the modeling of sectoral interac­tions is fundamental (e.g., among agriculture, energy, pro­cessing and manufacturing as well as services), trade (do­mestic and international), and existing policies. Given their economy-wide coverage, some variant of this type of mod-

 

els has become a part of the Integrated Assessment models (e.g., IMAGE; Eickhout et al., 2006).

     A strength of CGE models is their ability to analyze the interaction between different sectors such as agricultural sectors, manufacturing sectors and services. In their con­ventional usage, CGE models are flexible price models and are used to examine the impact of relative price changes on resource allocations (of goods and factors) across a range of economic agents. Thus, in addition to providing insights into the economy-wide general equilibrium effects of policy changes, CGE models allow key interindustry linkages to be examined. However, CGEs are poor in addressing distribu­tional issues within the regions: only average adjustments in the regional economies are simulated. Moreover, CGE models should be handled with care for long-term projec­tions since fundamental changes in the economic structure of a region cannot be simulated by a CGE model. Therefore, CGE models are only used in this assessment for assessing the global economic consequences of trade liberalization.

5.2.1.4 Marine biomass balance models

Fisheries models, such as EcoOcean, allow managers to explore how marine systems, especially fisheries, might re­spond to policy changes at the scale of the ocean basin or region not addressed by most other fisheries models. This model reduces what is a highly complex and dynamic sys­tem that covers 70% of the Earth's surface to 19 regions and describes the world's fisheries for the last 50 years with reasonable accuracy (often with 10% or less variation of what is recorded by FAO between 1950 and 2003). A com­plete marine system is modeled that ranges from detritus to top predators including marine mammals and seabirds, and provides sufficient detail to assess changes but avoids complexity so that it is computationally possible. The pred­ator-prey relationships between functional groups are also accounted for in the model. Because EcoOcean is based on the Ecopath suite of software and uses a trophic structure as well as predator-prey relationships, consumption rates and fishing effort, it provides a description of the ecological dy­namics of the system and an indication of how the diversity of the fisheries will change over time.

     The models have  some weaknesses.  The  functional groups used in EcoOcean are broad groupings of marine organisms, which limit their ability to describe in detail how a particular species or groups of species may respond to a specific policy intervention. The model is based on biomass from published time series studies and does not necessar­ily include a comprehensive suite of species to provide an estimate of the biomass for each functional group. The FAO regions used in the model are broad and cannot include climate or oceanographic features. This limitation makes it difficult to accurately model the small pelagic fish group (e.g., anchoveta) which is highly influenced by changes in oceanographic conditions as seen in the offshore upwelling system in Peru. The tuna groups do not differentiate be­tween long-lived slow-growing species such as bluefin tuna and short-lived ones such as yellowfin. This can result in overestimation of tuna landings as well as resilience. Effort, based on seven fleets, is the driver of the model and while some effort is gear-specific, such as tuna long-line and tuna purse seiners, effort for the demersal fleet is based on a range