|
Data and Services version history -- version 1.0.3
This page provides an overview of the new features of version 1.0.3
with respect to the previous version (0.9.0) in the processing and
the data presentation on the web pages. The description is split into
the following subsections:
For details on the different parts of the processing, consult the product
information pages using the menu on the left.
General description
Due to the interference between the absorption signals of SO2 and ozone,
there were quite a lot of artifacts at high SZA, giving rise to "false
notifications" in version 0.9.0. The background correction of version 0.9.0 (which
was based on SO2 as a function of SZA) was doing a reasonably good job for
situations where the ozone distribution along a given longitude is more or
less constant. In situations with large variations in the ozone
concentrations, which in particular arise at high latitudes in
Feb. - April at the Northern Hemisphere, the background correction
was performing rather badly.
The new background correction is based on a relationship between the
absorption signals of ozone and SO2 and should therefore take into account
automatically variations in the ozone concentration as function of
longitude. This in itself already reduced the number of artifacts at high
SZA considerably. A further improvement was found in some changes in the
settings for the retrieval of the SO2 slant column density.
The implementation of both the new retrieval settings and the new background
correction implies that all data has to be reprocessed. This presented an
ideal opportunity to implement some more improvements and additions to the
service, such as the inclusion of SO2 vertical column data and changes in
the presentation of the data on the website.
The new background correction is doing a better job than the previous one,
but there are -- and always will be -- artifacts at high SZA due to the
interference in absorption signals. With the retrieval method currently in
use (DOAS), it is probably not possible to make a more significant
improvement to get rid of these artifacts and the corresponding "false
notifications".
SO2 vertical column density
Before only the SO2 slant column density (SCD) was presented, but now we present
only the SO2 vertical column density (VCD). The reason for adding the
VCD is on the one hand that the VCD is a quantity that is more familiar to
most people (it represents the total column density above the measurement
point) and on the other hand to facilitate comparison with the results of
other satellite-based measurements.
Please note that the values given for the SO2 concentrations should not be
taken as hard numbers. For the moment the SO2 data -- in particular the VCD
-- has not undergone a rigorous validation: only a qualitative and
small-scale validation has been carried out so far.
In order to compute the VCD from the SCD, a so-called Air-Mass Factor (AMF)
is used: VCD=SCD/AMF. And in order to determine AMF, information on the
(effective) cloud cover fraction (CCF) is needed.
The AMF used here is based on pre-calculated look-up tables with a number
of variables as entry points, such as viewing and zenith angles, as well as
the height of the SO2 plume. But the latter is not known from the retrieval
in an automated processing, so an assumption must be made. It has been
decided that the VCD is computed for three different assumed plume heights:
- 2.0 km above local ground level
- 6.0 km above local ground level
- 14.0 km above local ground level
The first one represents passive degassing of volcanoes and anthropogenic
acitivities, the second one moderate volcanic eruptions and the third one
explosive eruptions. These numbers corresponds to SCIAMACHY instrument. These agree more or less with the choise made for GOME-2 and by the OMI team for their SO2 data retrieval and delivery (see AMF info according the different instruments).
All three VCDs based on these assumptions are available in the data files
(which required a change in the data format of both the ASCII and the HDF
data files), whereas the middle one is used for the plots presented on these
web pages. The SCD can also be plotted and accessed under "other plots".
The value of the AMF depends on the circumstances, but in most cases it will
be larger than one, which means that the numerical value of the VCD will in
most cases be lower than the numerial value of the SCD. For this reason in the NRT and archive service, the 3 types of VCD( deduced from the three types of AMF) are presented: (Low, Middle, High AMF). At high SZA the AMF can be smaller that one, so that the VCD is larger than the SCD.
Note that if no cloud cover information is available for a given ground
pixel, the AMF and VCD of that ground pixel cannot be computed. In that
case, the entries in the data files will get the "no data" value. Only
the AMF for the clear-sky part will be given a value, as that can in
principle be computed always.
Presentation on web pages
The pages presenting the data now show images of the SO2 vertical column and
the cloud cover fraction, and a plot of the region.
The diffenret instruments (SCAMACHY, GOME-2, OMI and IASI) can be selected. To the left of the plots, the instrument name is mentioned.
The colour scale of the plots has been changed. The scale now starts at
0.5 DU, rather than zero, the reduce the low-level scatter of the
background SO2. The standard scale goes up to 2 DU, as before. But to
easier identify really large SO2 peak values, an additional colour has been
added for values larger than 10 DU.
Apart from the plots for the set of geographic regions defined for the
service, the previous version showed a plot for the whole world. As such a
plot does not show the data near the poles very clearly, a "polar view"
region has been added in the menu: it shows both poles in one plot,
above (below) +30 (-30) degrees latitude.
Near-real time and Notification Service specific features
The dedicated notification web pages point to have been given a new layout.
A dedicated plot is made: a 30 by 30 degree region is drawn with the notification nadir state right in the middle, and the plot shows only the data of the orbit that nadir state is part of. Below that a plot of the volcanoes (if there are any) in that dedicated region is shown. If a notification is confirmed by an other instrument the level of the notification is increase by 1 (maximum is level 4 for a notification for SCIAMACHY, GOME-2, OMI and IASI instruments).
If no cloud cover information is available, it is not possible to compute
the AMF, and so the ground pixel in question would be marked as "no data".
To have at least a rough idea of the VCD of that ground pixel, the NRT
processing assumes that the ground pixel is cloud-free and uses the
clear-sky AMF (which can be computed always) for the plots on the web page
(not in the data file). Whether this happened can be seen from the absence
of data in the plots of the cloud fraction.
Archive Service specific features
In the NRT service a VCD is shown assuming clear-sky conditions in case
cloud cover data is missing (see above). This is not done for the archive
service: if no cloud data is available, no AMF or VCD is available (the
data files do contain - if possible - the value of the clear-sky AMF, but a
"no data" value for the other parameters).
The archive service now also contains an overview of the notification cases of per
month. This shows for which nadir states in the archive the criteria for
"exceptional SO2 concentrations" are met. The main table of the overview page
is generated automatically, leaving room for remarks about the notification, e.g.
about volcanic eruptions that may be related to the observed SO2, based on
external information. These remarks need to be added by hand, which takes
time and so these pages are in a way always "under construction".
|
|