Environmental pollution, like other stressors, originates in sources somewhere in the environment.   These sources, no less than the hazards themselves, take many different forms.  Many, though not all, relate to human activities and the infrastructure (e.g. roads, industrial plant, houses) used in the process.   Some, however, are entirely natural (e.g. radon emissions and the gases and particles released by volcanoes).  Moreover many other types of environmental hazard exist, such as earthquakes, floods and storms, which are essentially natural in origin - though some, it has to be admitted, are sometimes triggered or exacerbated by human interference.  

Information on the sources of environmental pollution (and other stressors) is clearly vital in integrated impact assessment.  It is needed as the starting point for modelling environmental contamination, to estimate how this might change in response to changes in technology or policies affecting socio-economic activities, and ultimately to help identify where preventative action can best be taken.  In some cases, also, the exposure-response functions available from epidemiological studies relate not to exposures per se, nor even to environmental concentrations, but to measures of source activity.  Examples include the health effects of by-stander exposures to pesticides (which have often been based on pesticide application rates or practices), health risks from landfill sites (which have mainly been based on distance from source), and many estimates of health risks from indoor pollution (which have typically been based on presence/absence of different cooking or heating sources in the home).  In these cases, therefore, the assessment has to follow suit, and base its exposure estimates on the same metric. 

Fortunately, a great deal of information on sources tends to exist, primarily because of their social and economic importance - which means that they are subject to close monitoring and/or regulation (see links to Characterising sources in the panel to the left).  Nevertheless, not all the data we require is usually already available, and for many assessments we need to undertake modelling, either to fill in gaps in the data that do exist, or to estimate how sources may change under the assessment scenario.  Links to methods for Modelling source activity are also given to the left.