A wide range of release inventories have been developed, for different sources, contaminants and study areas.  In the EU, the most definitive source is the European Pollutant Release and Transfer Register, which provides information on releases of 91 pollutants into the air, water and soil from ca. 24,000 industrial facilities covering 65 economic sectors.    Many countries and municipalities also maintain national inventories, offering data and mapping capability at a higher spatial resolution.   In addition,  specialist inventories have been developed, both regionally and globally, targetted at specific sources (e.g. road transport, airports), release pathways (e.g. air, water) or pollutants (e.g. greenhouse gases, mercury).  Links to many national and international inventories are provided by the PRTR.net portal, and the OECD Centre for PRTR Data, as well as by the Opasnet website (see link below). 

Inventory methodology

The information contained within these inventories is derived in a number of different ways.  A broad distinction can be made between 'bottom-up' and 'top-down' approaches.  In the former, inventories are built up by aggregating detailed information from individual sources (e.g. industrial facilities, road links) or small areas (e.g. residential zones); in the latter, broader (often national) data are disaggregated to a more local level.  While the former methods can sometimes make use of direct measurements of releases (e.g. at major point sources), the majority of the data are derived from modelling.  Bottom-up inventories generally involve the use of the standard emission formula (see link to Modelling contaminant releases in panel to left), in which a measure of source activity is multiplied by an emission factor for each locality (e.g. each grid cell or administrative area).  In top-down inventories, similar formulae are typically applied at an aggregated (e.g. national) level, and then the estimated emissions total reapportioned to a more local scale using data on relevant proxies, such as population density, GDP, employment or land cover.   

Uncertainties in emission estimates

The data provided by emissions inventories are inevitably uncertain.  Major sources of uncertainty include:

• biases in the sample of sites for which direct measurements are available (often the larger and better regulated facilities);
• uncertainties in the emission factors;
• uncertainties in the statistical or other data used to represent source activity;
• poor characterisation of the emission control technologies and their actual efficiency;
• errors in model formulation;
• inadequate recognition of the effect of local factors (e.g. facility management and maintenance) on emission rates.

The levels of uncertainty vary, depending on the pollutants and source, and the spatial scale of analysis. Understanding of the uncertainties is probably best for atmospheric emissions.  Typical levels of uncertainty at national level in these range from <10% for major criteria pollutants such as sulphur dioxide and oxides of nitrogen to 20-50% for particulates and benzene, and as much as 100% (or even more) for dioxin and benzo(a)pyrene.  For releases into other media, for which the inventories tend to be less well-developed, the uncertainties may be greater.  Likewise, larger uncertainties may be expected for many estimates at more local scales.