Chemistry Reference
In-Depth Information
ZEP
PAL
SMR
ASP
BIR
PLA
VHL
WAL
BOE
MPZ
OBK
KPO
CBW
HPB
HWL
BEO
SLL
ZSF
MHD
JFJ
ISP
PDD
CMN
FKL
low land stations
mountain stations
EUSAAR
GUAN
EUSAAR
GUAN
Fig. 2 Stations used in measurements of aerosol number size distributions.
Black symbols
are
EUSAAR stations,
white
GUAN (MPZ was in both networks).
Triangles
denote high-altitude
mountain stations (over 1,000 m from mean sea level). Figure adapted from figure published in
[
18
], which also has more details on the locations and types of the stations
concentrations measured at these stations should represent the overall regional
background concentrations and should be thus representable of a large footprint
area around them [
19
].
3 Levels and Variability of Aerosol Number Concentrations
3.1 General Properties of Number Size Distributions
The particle number size distributions can be, as previously noted, often considered
to be a combination of several aerosol subpopulations or modes. These modes are
often log-normally distributed in the diameter space. Interestingly, the concentra-
tion histograms (i.e., frequency of detecting specific concentration) are typically
also log-normal. Thus, using linear measures of concentration or size, such as
arithmetic mean, can be strongly influenced by outlier values. This is the reason
why either order-based metrics (such as percentiles and medians) or geometric
properties (e.g., geometric mean or geometric standard deviation) of distributions
are better mean properties to study if one is interested in typical concentration levels
or mean particle diameters. The logarithmic occurrence spectrum is not only a
feature of particle number concentrations. There are many reasons to expect similar
behavior from the particle mass-based metrics [
12
], even though arithmetic means
are the traditional method of averaging in the air quality contexts.
