IMAGE framework/A brief history of IMAGE

Jump to navigation Jump to search

A brief history of IMAGE

IMAGE 3.0 is the most recent, operational model framework progressively developed since in 1980s. Over the years the model was enhanced and extended continuously through incremental changes, and more profound revisions.


IMAGE 1.0 (Rotmans, 1990) was developed as a single region, integrated global model to explore interactions between human activities and future climate change. As one of the first Integrated Assessment Models to address climate change, IMAGE contributed to raising awareness of the long-term consequences of human development. In the absence of regional or spatially explicit algorithms, the model operated on trends in global total and average parameters, such as world population and averaged emission factors per unit of activity.

IMAGE 2.0 to 2.2

In the 1990s, the new generation IMAGE 2, was developed with regional drivers of global change and gridded, process-oriented modelling of the terrestrial biosphere, land cover and land use (Alcamo, 1994). IMAGE 2.0 comprised three subsystems:

  • 13-region Energy-Industry System (EIS);
  • Terrestrial Environment System (TES) operating at 0.5x0.5 degrees grid-scale;
  • Atmosphere-Ocean System (AOS) to compute the resulting changes in the composition of the atmosphere leading to climate change.

Further refinements and extensions were implemented in IMAGE 2.1 (Alcamo et al., 1998) to enhance model performance and broaden its applicability to issues other than climate change.

The enhanced capabilities of IMAGE 2.2 were demonstrated in the contribution to the IPCC Special Report on Emissions Scenarios (IMAGE-team, 2001). The earlier zonal-mean climate-ocean model was replaced by a combination of the MAGICC climate model and the Bern ocean model. The resulting global average temperature and precipitation changes were scaled using temperature and precipitation patterns generated by complex Global Circulation Models (GCMs) to provide spatially explicit climate impacts and feedback. For economy and energy, the EIS of version 2.0 was replaced with the TIMER energy model, which also improved linkage with the macroeconomic model Worldscan.


A range of developments were implemented stepwise in intermediate versions, leading to the release of IMAGE 2.4 (MNP, 2006). In close cooperation with the agro-economic research institute LEI, links were established with agro-economic modelling in IMAGE 2.4. This ensured the inclusion of biophysical conditions in modelling future agricultural production based on intensification of production and expansion of agricultural area.

Furthermore, to align closer with policy discussions, the number of regions was increased to 24 to reveal the position of major global players. Other extensions include a link with the global biodiversity model GLOBIO to study impacts of global change drivers on natural and cultivated land.

Many of the modules were enhanced, including the energy model TIMER, emission modelling and the carbon cycle. Experimental links with an intermediate complexity climate model were discontinued and the simple climate model MAGICC with a strong feature to represent uncertainties in the climate system was adopted as the default. IMAGE 2.4 has played a key role in supporting various international environmental assessment studies.

Towards IMAGE 3.0

After publication of the IMAGE 2.4 book and a subsequent progress review by the IMAGE Advisory Board, the framework has been further developed. These developments were published in journal articles and conference papers, but no new versions were officially released. For example, representation of energy demand was improved by more bottom-up modelling of household energy systems in TIMER for rural and urban population by income level. Selected industries were better represented in more technical detail to underpin energy demands and emissions. The forestry sector included forestry management options in addition to clear-cutting.

In cooperation with Wageningen University (WUR) and the Potsdam Institute for Climate Impact Research (PIK, Germany), the IMAGE natural vegetation and crop modules were replaced with the LPJmL Global Dynamic Vegetation Model. This enabled modelling of linked carbon and water cycles, and adding a global hydrology module to IMAGE, which was not available in previous versions. Modelling biodiversity impacts was extended to cover freshwater systems as well as terrestrial biomes.


The following new developments have been incorporated in IMAGE 3.0:

  • Energy demand modules to address household energy demand and energy carrier preferences for urban and rural populations, and per income level in developing and emerging economies. Energy demand also includes selected energy-intensive industries using technological production alternatives with their costs and efficiencies in delivering energy services.
  • Forestry management module covering different production systems per region. Management systems include clear-cutting, selective cutting (conventional and reduced impact logging) and dedicated forest plantations. Wood products are also retrieved from areas deforested for agriculture and other purposes.
  • Plant growth and carbon modelling by LPJmL, coupled to IMAGE. LPJmL simulates plant growth as a function of soil properties, water availability, climatic conditions and plant and crop growth parameters. Carbon stocks and fluxes, biomass yields and water surplus are integrated and internally consistent.
  • Global hydrological modelling, linked with natural vegetation and crop growth in LPJmL. The balance of precipitation and evapotranspiration in each grid cell feeds a routing network of rivers and natural lakes. Man-made reservoirs for hydropower production, irrigation, and mixed use are included and alter river flows.
  • Nutrient (N, P) soil budgets for natural and anthropogenic land use to assess nutrient cycles in agricultural and natural ecosystems, fertiliser use, and efficiency and integration of manure in crop production systems. In addition to these non-point sources of nutrients, point sources of urban wastewater are modelled. The fate of the nutrients in the river systems determines the load in coastal waters at a river mouth, creating risks of hypoxia and algal blooms.
  • Landscape composition on a 5x5 minutes resolution, instead of the 0.5x0.5 degrees grid used in all IMAGE 2.x versions. Depending on the modules, 5 minute information is processed either directly or translated into fractional land use at the 0.5 degree scale.
  • MAGICC 6.0, which updates the climate model and associated data, is a simple climate model that estimates global average temperatures as the result of net greenhouse gas emissions, carbon uptake, and atmospheric concentrations of climate forcing agents. The global average temperature is used to scale grid-based climate indicators emerging from complex climate model studies.
  • Additional impact modules, providing information on aquatic biodiversity, flood risks, soil degradation, ecosystem services, and human health.
  • Optimal greenhouse gas emission reduction pathways for overall climate policy goals are explored under assumptions for participation timing, rules and emission targets under global strategies. A cost-benefit analysis tool has been added to test the net economic outcome of mitigation efforts, adaptation costs and residual climate change impacts at different levels of forcing, subject to cost and damage assumptions found in the literature.