Appraising the Direct Impacts of Aerosol on Climate
Regional analysis of GlobAEROSOL data
This page summarises a regional assessment of the GlobAEROSOL AATSR
product against AERONET data. The regions used for this analysis are
shown here:
Monthly statistics of the GlobAEROSOL dataset for 2006 are given in
the following table for each of
these regions, with links to maps of the mean AOD for each month
(click month names) and time series of the AOD for each region
(click region names). [Click image for full version]
The following table shows the statistics of AERONET comparisons for
each of these regions. The procedure for these comparisons was:
Level 2 version 2 AERONET column AOD data from 2002 - 2007 was used.
GlobAEROSOL AATSR ORBIT files were used, with the aerosol class set
by the "BestType" variable provided with the data.
A comparison was made if:
- GlobAEROSOL results had passed the
standard quality controls (i.e. the GlobAEROSOL pixel had a
"PixelQuality" value of 1)
- Valid GlobAEROSOL AOD retrievals were
available within a radius of a given AERONET station
- Valid AERONET results were available
within 30 minutes of the satellite overpass time
- Both GlobAEROSOL and AERONET AODs
passing the above criteria had a standard deviation of < 0.15.
The mean of the GlobAEROSOL and AERONET data meeting these
criteria were then used to as the comparison value for each
overpass
Click on the first column (the region name) to see scatter plots of
the comparisons. Note that these plots include linear fits to the
data (weighted by the standard deviation of the AOD from both
GlobAEROSOL and AERONET). Due to the differences behaviours of the
comparisons at different stations, these lines often do not provide
a very reasonable estimate of the true relationship between the two
datasets, and thus their parameters have not been included in the
table.
The final table shows the mean bias of GlobAEROSOL AATSR AOD
against Maritime Aerosol Network (MAN) measurements. MAN is a
project run in conjunction with AERONET, in which sun photometer
measurements are made from ships, often in the remote ocean. Due to
the sparcity of MAN measurements, and the fact that data are only
available from 2007 onwards, AERONET style point by point
comparisons with GlobAEROSOL are not practical, so the table
presents mean differences between PDFs of
log10AOD550 from the GlobAEROSOL record and 4
years of MAN data.
