Although a wealth of information on human activities exists, assessments still have to rely to a large extent on methods for modelling sources of environmental contaminants.  Even where data on emission sources are available, they are often too highly aggregated to show the more local effects of individual sources, so the data have to be disaggregated for the purpose of asssessments.  More generally, the scenarios that underpin many assessments involve some sort of change in emission sources.  Sometimes these changes are explicit: for example, the scenario may specify that the use of certain pesticides will be banned.  In these cases the need to model how these changes take place (e.g. what alternative forms of pest control will be used) and their consequences is inherent in the assessment.  Often, however, they are implicit: an emission reduction scenario may specify the amount by which emissions will decline but not give any indication of how this will be achieved.  Even in this case, however, the implied changes in source activity need to be analysed, for otherwise any collateral effects (e.g. associated with the new technologies that have to be introduced) would be ignored, and the assessment would be biased.         

Two main approaches to modelling can therefore be identified, in relation to source activities:

• Spatial modelling is required to estimate the geographic distribution of source activities.  One of their main applications is to disaggregate broad scale socio-economic data to a more local level, in order to allow a more detailed analysis of impacts.  National or regional data on energy consumption or traffic flows, for example, may be disaggregated to a local level, in order to simulate the spatial distribution of emission sources, and thus to model local variations in exposures and health impacts. 
• Change models are required to simulate the way in which policy or other (e.g. technological) developments feed through the socio-economic system (e.g. via prices) to affect specific socio-economic decisions and practices (e.g. production, consumption).  A range of different modelling approaches may be used for this purpose, including statistical, econometric, engineering or biophysical models, or (where quantitative modelling is inappropriate) the use of expert elicitation methods.   

These approaches are not necessarily alternatives; often a combination of methods will be needed - for example, to simulate macro-economic changes and then downscale these to the local level, or to model changes in the spatial distribution of sources (e.g. urban areas) in response to policy or other drivers. 

Links to examples of modelling source activity are given below.