1 All data in the Database (relative mass and its uncertainty) are reported as mass of the given species relative to the PM total mass.
2 Not detected species are identified as 'not detected' as the minimum detection limit is in general not known.
3 The uncertainty is set to the original value and when missing is because was not reported in the original publication.
4 In the few cases where the original data are reported as absolute mass (x) the uncertainty of the relative mass is determined by the propagation of uncertainty u² of the ratio x/PM following JCGM (2008), paragraph 5.1.2, equation 10. Here the assumption of not correlated variables gives an overestimation of the uncertainty. In this case the round-off of the uncertainty is made as in JCGM (2008), paragraph 7.2.6, taking two significant digits for uncertainty, and adjust the relative mass to the same decimals of the uncertainty.
5 Each species is identified with a unique internal ID, in most cases corresponding to US-SPECIATE ID. The CAS number and the AIRBASE code are also reported, when available.
6 Each profile is identified with a unique internal ID, which should be reported in the citations of the profiles to prevent ambiguity in the identification.
7 The 'name' of the profile is derived from but not equal to the original data and is not unique. The same goes for the ‘original name’, which is reported to help the user in going back to the original data (in the original paper or technical note) if necessary.
8 Source profiles are either 'O' original , 'C' composite , 'D' derived (as form PMF outputs), 'S' simplified (as in SPECIATE) or 'T' theoretical (e.g. stoichiometric ratios).
9 Source profiles may refer to the total mass of PM2.5, PM10, PM50, TSM, PMraw or any other PM size as specified in the original data.
10 There is a hierarchy in source categories that ranges from the most general to the most specific ones. For that reason one profile may belong to more than one category.
JCGM, 2008. Evaluation of measurement data - Guide to the expression of uncertainty in measurement