satellite geometry
Schematic overview of the three primary data modalities: while AIRS aboard NASA's Aqua satellite scans the atmosphere directly below the satellite (at nadir), Envisat's MIPAS instrument uses a so called limb geometry, i.e. it performs a tangential scan. For each MIPAS detection (yellow dot), we provide CLaMS trajectories, which are started at the location of the detection.

The data consists of three main data modalities and two support data sets. All datasets are registered in common geographical frame of reference defined by global latitude, longitude, and time. Where applicable, data points feature additional information on their altitude above mean sea level. Time is given in Julian seconds since 00:00:00, January 1, 2000 UTC. For reference, use the following table that pinpoints several key dates:

Date Julian seconds Comment
01.06.2011 00:00 UTC 360,201,600 data start
04.06.2011 00:00 UTC 360,460,800 day of the Puyehue-Cordón Caulle eruption
13.06.2011 00:00 UTC 361,238,400 day of the Nabro eruption

Specific days can be deduced from these by adding or subtracting 86,400 seconds per day.

The data are provided by the Simulation Laboratory Climate Science at the Jülich Supercomputing Centre and the Intitute of Energy and Climate Reserach, both at Research Center Jülich, Germany within the JARA - High Performance Computing collaboration. The data are provided for free use within the scope of this visualization contest. Please acknowledge the original data source in any related publication:

Griessbach, S., Hoffmann, L., Spang, R., and Riese, M.: Volcanic ash detection with infrared limb sounding: MIPAS observations and radiative transfer simulations, Atmos. Meas. Tech., 7, 1487-1507, doi:10.5194/amt-7-1487-2014, 2014. (Link)

Lars Hoffmann, Sabine Griessbach, and Catrin I. Meyer: Volcanic emissions from AIRS observations: detection methods, case study, and statistical analysis, Proc. SPIE 9242, Remote Sensing of Clouds and the Atmosphere XIX; and Optics in Atmospheric Propagation and Adaptive Systems XVII, 924214 (October 21, 2014), doi:10.1117/12.2066326. (Link)

If you would like to use this data for atmospheric research please contact Sabine Grießbach at s.griessbach(at) to get background information on the data products and the latest version of the data.


Envisat MIPAS

Depiction of a single MIPAS orbit. Note that the satellite comes rather close to the poles which results in the quasi-horizontal track and causes a wrap-around artifact at the bottom of the image.
A close-up view of the corresponding detections over the Americas shows that sample profiles are not perfectly vertical due to the satellite's continuous motion.

The Michelson Interferometer for Passive Atmospheric Sounding (MIPAS) aboard the European Space Agency's Envisat, which was launched in 2002, measured nearly 10 years until April 2012. MIPAS measures infrared spectra over a broad spectral range from 4.15 µm to 14.6 µm. It uses a so called limb geometry, i.e. it samples the atmosphere tangentially, as shown in Figure 2. Profiles are measured between 5-70 km of altitude along the orbit track. In the upper troposphere and lower stratosphere the vertical sampling is about 1.5 km. The MIPAS field of view is approximately 3 km. This means that radiances from up to 1.5 km below and above the tangent altitude contribute to the signal measured at the specific tangent altitude. MIPAS provides a good daily coverage with about 14 orbits per day. Due to the limb geometry, MIPAS has - in addition to its very good vertical resolution - a high sensitivity towards aerosol and hence allows for long time tracing of aerosols. With recently developed detection methods for volcanic ash and sulfate aerosol it is not only possible to detect enhanced aerosol events, but also to discriminate between different particle types. Background information on MIPAS is publicly available through the ESA's Envisat mission site.

For this contest, the MIPAS data is provided for the time frame from June to August 2011. Measurements form a time-dependent point data set, which contains information on the latitude, longitude, altitude, and time for each discrete measurement. The point set is given in VTK's legacy binary format and described as cell-less vtkPolyData where each vertex corresponds to a single MIPAS measurement. Per point, the following attributes are given:

The data has been pre-filtered by domain scientists in order to contain only meaningful measurements with respect to the tasks provided below, as well as the clear air points for reference. Sulfate aerosol detections are only included in the range between 90N and 20S. In the southern hemisphere between 20S and 90S they have been filtered out in order to avoid confusion with non-ice polar stratospheric clouds, to which the method is also sensitive to. Please note that measurements are not always spaced evenly in time, with possible gaps appearing in the data. This is due to different measurements modes of the MIPAS instrument.


AIRS / Aqua

AIRS coverage from 00:00 to 12:00 UTC on June 10, 2011. The measurements from a 12 h period yield almost global coverage.
AIRS ash measurements for June 10, 2011. In the southern hemisphere, ash from the Puyehue-Cordón Caulle is clearly visible. In contrast, detections over northern Africa are not from volcanic origin, but relate to dust clouds over the Sahara desert.

The Atmospheric Infrared Sounder (AIRS) is one of six instruments aboard NASA's Aqua satellite, which was launched in May 2002. AIRS measures the thermal emissions of atmospheric constituents in the nadir and sub-limb observation geometry, i.e. it scans the area directly below the satellite (at nadir) and to both, left and right of the satellite's track (sub-limb). A rotating mirror is used to carry out scans in the across-track direction. AIRS has two main advantages: excellent horizontal resolution and nearly global coverage.

During one day AIRS measures about 2.9 million Infrared radiance spectra. The globe is covered by 14.5 orbits per day. We combine the measurement data from 12 h periods into single data files (named _am_ and _pm_, respectively), which each provide nearly global coverage. However, at low and mid latitudes there are small data gaps between successive tracks for which no AIRS data are available.

A single AIRS scan consists of 90 individual footprints in the across-track direction and covers a distance of 1.765 km on the ground. The AIRS aperture is 1.1°, corresponding to a footprint size of 13.5 km × 13.5 km at nadir and 41 km × 21.4 km at the scan extremes. Each footprint essentially corresponds to a pixel in the final satellite image. While each footprint contains precise information on measurement time, latitude, and longitude, vertical information is not available. Example images for a half-day period are shown here.

For the detection of sulfur dioxide and volcanic ash with AIRS, well-established index methods are used. The data set has been pre-processed to contain two scalar index values, being proportional to the total column amount of SO2 and ash, respectively.

The data set is given as vtkPolyData and stored in legacy, binary vtk format. It containes the following data attributes:

Data points from successive tracks have been connected by quad strips in order to facilitate interpoaltion. This may also enable processing such as e.g. region growing or isoline extraction. Please note that only cells that do not span the wraparound from -180 to 180 degrees in the longitudinal direction have been included. As such, the provided cell structure is only of limited use for a 3D, spheric depiction of the given data. If you intend such a visualization, you'll have to re-wire the cells yourself. You can assume that each AIRS track features 90 data points, which have been inserted into the data set in successive order, and there are no missing points.


CLaMS Trajectories

Selection of CLaMS trajectories for the southern hemisphere.
A close-up of the trajectories on the left reveals their altitude extent.

The Chemical Lagrangian Model of the Stratosphere (CLaMS) was developed at the Institute for Energy and Climate Research - Stratosphere. It integrates a hierarchy of computational atmospheric models ranging from a simple box model to a 3-dimensional global chemistry transport model. CLaMS contains separate modules for transport, mixing, chemistry, and microphysics that can be included or excluded.

To prepare the dataset for this contest, only the trajectory module, which is based on a fourth-order Runge-Kutta scheme, was used to calculate trajectories for individual air parcels, including vertical transport via the time derivative of pressure or diabatic heating. The underlying meteorological fields for the trajectory calculation are based on ERA interim data.

Trajectories are included for sulfate aerosol detections from 20 S to 90 N and volcanic ash detections from 0 S to 90 S of the provided MIPAS data set. Two sets of trajectories are provided:

Both data sets are seeded at the locations of MIPAS aerosol detections. Trajectories stretch back in time until the start of the data period, and forward in time for five days following w.r.t. the respective seed detection. They provide the opportunity to track detected aerosol back to its origin, namely to a specific eruption event, and forward in time to e.g. relate them to later detections. The trajectories are a primary option to establish temporal relationships between data items and/or other spatial locations.

The data is also provided in legacy vtk format and it is composed as follows. For each MIPAS detection, the vtkPolyData in the data set contains a single polyline. The line is seeded at the location of the MIPAS detection and extends both forward and backward in time. Points are connected to each other according to monotonically increasing time stamp. The index of the seed, i.e. the original detection, is given as an integer cell attribute that provides a point reference w.r.t. the current line.

The following data attributes are available per point:

The data features the following cell attributes per trajectory:


Tropopause Altitudes

Visualization of the altitude of the first tropopause.
Depiction of the 2nd tropopause, where it exists. Please note that we have nulled all data below 20S because of spurious detection in the southern polar region.

The tropopause defines the boundary between the Earth's troposphere and the stratosphere. Aerosol in the troposphere is rather short-lived and is washed out due to clouds and rain within several weeks. In contrast, aerosol in the stratosphere is long-lived, because it cannot be washed out. Hence, it has an impact on climate because it takes months or even years to sediment out. There is a variety of different definitions for actually determining the altitude of the tropopause. The thermal tropopause is defined by means of the temperature gradient. In contrast, the definition of the dynamical tropopause is based on potenitial vorticity - values larger than 2 indicate stratospheric air. Another measure to separate between troposphere and stratosphere is the potential temperature. In the tropics, values larger than 380 K indicate stratospheric air. In some cases, there are so-called "double-tropopause" situations, i.e. there are two such booundary layers stacked on top of each other.

The tropopause data we provide here has been derived from ERA-interim data and is subject to conditions of ECMWF. The first tropopause is the thermal tropopause according to the WMO definition. The second tropopause has been computed acoording to the definition presented in Randel (2007) for model data.

The time-dependent data is given as a series of files each of which represents a uniform grid covering the northern hemisphere. It contains the following data variables: