Presentation and delivery of the data -- some notes
Some notes regarding the presentation and delivery of the data:
The colour scale of the plots
The colour scale of the SO2 vertical
column plots shows the colour coding for the concentration in Dobson Units (DU).
The regular colour bar is from 0.0 to 2.0 DU. If measurements are between 2.0 and 10.0 DU the color is orange.
A special colour (red) is used for really strong SO2 signals -- usually
these associated with volcanic eruptions -- for measurements higher than 10 DU.
Colour bar for SO2 vertical column densities.
The reason to end the main scales at 2 DU is to make sure that low
level SO2 concentrations are visible, concentrations that can be great
interest too, e.g. because they signify small-scale volcanic eruptions.
Furthermore, it is most interesting to show where large concentrations of
SO2 appear, without being too specific about the exact values, especially
because large SO2 concentrations do not occur that widespread or that often.
Furthermore, representing all data using the same colour makes it easy to
compare results for different periods and regions.
Note that the scale for the SO2 plots shown on the website varies linearly.
For special purposes, however, the scale may have other ranges and/or may be
The colour scale of the (effective) cloud fraction varies linearly from zero
(cloud free; dark blue) to one (fully clouded; white). Values below zero and
above one cannot occur. The snow/ice mode of the cloud data products is
represented by very ligth grey (indicated simply as "snow" in the colour
Colour bar for the effective cloud cover fraction.
Plots of overlapping orbits
At high latitudes, above 65 degrees or so, satellite orbits overlap since
the swath width of the instrument has a fixed size at ground level
(960 km in normal mode for SCIAMACHY, about 2000 km for
OMI and GOME-2). When plotting the daily data at orbit coordinates, this
plotting is done with the orbits listing in ascending time. As a result, the
values shown in the plot at a given location are those of the latest orbit
of that day. This can be seen clearly in the SCIAMACHY-based picture shown
For the gridded data this is not a problem: all data falling in a given grid
cell is averaged for that grid cell and therefore each grid cell has a
unique data value.
Plots for the website made with IDL
The plots of the SO2 data and the associated maps are all made with IDL
(version 6.3). Some notes:
For plots of the daily data at orbit coordinates, rectangles made up of the
ground pixel corners as given in the data product are filled with the
appropriate colour. Since these rectangles are made up of straight lines,
they do not exactly match those north and south of them: there are little
openings between the rectangles, because the earth's surface is curved.
This can, for example, be seen in the example image shown above and below.
It does not show up in the images of the gridded data.
The IDL plotting routines include the possibility to draw lines along
continents and to colour the continents. IDL appears to have some problems
correctly colouring the continent Antarctica. The reason for this is that
the information for colouring the continents and for drawing the borders is
taken from two different databases, which evidently differ at the border of
Antarctica. The line around the continents shows the real border.
Unfortunately, the lakes drawn on the maps by IDL do not get the same colour
as oceans do: lakes get the land colour, whereas one would expect lakes to
have the sea (water) colour. This is especially clear in the plots of
Eastern Europe, where the Black and Kaspian Sea and several lakes are given the land colour.
On the maps of the regions, boundaries of the countries are drawn.
These boundaries are the political boundaries from 1993.
IDL's database unfortunately does not have all borders correct;
for example, in Arabia boundaries are missing. The IDL developers are
aware of this problem.