Here we compare the results of these measurement programs. In order
to facilitate these comparisons:
Absorption versus time is presented in "bap
Time Series.jpg". Absorption of aerosols <10 µm aerodynamic
diameter (CMDL) versus absorption of all aerosols (UM) is presented in
"<10 µm bap UM vs CMDL.jpg".
Similarly, absorption of aerosols <1 µm aerodynamic diameter (CMDL)
versus absorption of all aerosols (UM) is presented in "<1
µm bap UM vs CMDL.jpg". The error bars illustrate the magnitude
of the uncertainty in the accuracy of these measurements. The linear regression
of the <10 µm absorption (CMDL) versus total absorption (UM) shows
the measurements to be highly correlated (R^{2} = 0.94) with the
CMDL measurement averaging nearly 20 % higher than the UM measurement.
The linear regression of the <1 µm absorption (CMDL) versus total
absorption (UM) shows the measurements to also be highly correlated (R^{2}
= 0.90) with the CMDL measurement averaging only about 2 % higher than
the UM measurement. We expected the <10 µm absorption measurement
by CMDL to be closest to the UM measurement of total absorption. However,
that was not the case. Nevertheless, given the uncertainties in the measurements,
the differences between the CMDL and UM measurements of absorption appear
not to be significant.
Scattering versus time is presented in "<10 µm bsp Time Series.jpg" and "<1 µm bsp Time Series.jpg". Scattering of aerosols <10 µm aerodynamic diameter from both CMDL and UM instruments is presented in "<10 µm bsp UM vs CMDL.jpg". The measurements are highly correlated (R^{2} = 0.97). The CMDL measurements average approximately 10 % lower than those by UM. The error bars, which represent one standard deviation of uncertainty in accuracy, indicate that the two measures of scattering are not significantly different.
Scattering of aerosols <1 µm aerodynamic diameter from both
CMDL and UM instruments is presented in "<1
µm bsp UM vs CMDL.jpg". The measurements are correlated (R^{2}
= 0.86). The CMDL measurements average approximately 10 % higher than those
by UM. The error bars represent one standard deviation of uncertainty in
accuracy. There appear to be two populations of data points for scattering
values above 50 Mm^{-1}. One group of points lies very near the
1:1 line. In the other group of points, the CMDL measurements appear to
be systematically higher than the UM data. At present, we have no explanation
for this ~20% discrepancy.