The first part of this contest is based on general visualization tasks to get started and to develop a first understanding - a feel - for the data and the atmospheric processes it describes. Following that are three groups with additional and equally challenging problems. A complete submission would provide answers for the general tasks and for at least one group (A, B or C) of your choice. You may complete as many - or as few - of the visualization tasks as you want, but the more tasks are completed, the greater your chances are of winning this competition. Yet most importantly, have fun with this contest and enjoy exploring our data, while learning new things about clouds and precipitation processes.
The tasks are roughly arranged with an ascending level of difficulty. Some visualization tasks can be completed out-of-the-box using standard visualization software, others require a little more programming and data mining skills. Several references that might be useful are provided at the bottom of this page.
Figure 1. Clouds over southern Europe. (Source: DLR)
Visualize the three scalar cloud quantities CLW, CLI and QR and create a short animation showing the progress and the development of various rain bands along with cloud ice and precipitation. Annotate your visualization with topographic information.
Highlight several differences and extrema in the data, for instance, highlight those areas that either receive the most or the least amount of precipitation. Other examples could be to highlight those areas that are coolest, and those which are warmest, or those air masses that are driest or wettest. Are there any connections between those parts and the formation/evolution of the cloud systems?
Quantitatively visualize the amount of rain that reaches the ground. Ideally, create an animation to visualize the progress over time.
Extend your visualization from the beginning by adding local wind properties (direction, magnitude) for individual time steps. Create a visualization that combines wind properties with the cloud structures and highlight regions of special interest (e.g., regions of high/low magnitudes, strong vertical winds up/downward). What statements can you derive from your visualization about the relation of wind direction, magnitude and cloud structures in the ICON model?
To make our data more interesting and realistic, EUROCONTROL Network Management, provided us with all flight trajectories within Europe for April 26, 2013. Please annotate your visualization from the beginning with airplane trajectories for some of Germany's airports. Although our HD(CP)² data originates from a computer simulation, it is based on an observation based weather reanalysis for that particular day. Questions are: Do the partially severe weather conditions possibly influence the flight paths? Try to find those flights and visualize the flight trails, where a plane might have deviated from its original track due to the severe weather conditions. (Hint: The data is stored in UTC time.)
Recreate your visualization from Task 1.1 for one time step using the original ICON grid for domain 3, i.e. for the highest resolution. Concentrate on creating the best visual quality possible. Is interpolation at some point required? How interactive is your visualization?
With your visualizations and the things you have learned about clouds and cloud evolution, please create a scientific TV news report forecasting the weather for April 26, over Germany. (in English)
This last task is less about visualization and more about classic computer graphics. Please create renditions that show the clouds and the rain as realistic as possible. Include at least one image that shows the clouds from below (i.e. standing on the ground looking up).
Group A – Visualization of Wind
The interaction of the atmospheric flow field with other cloud quantities is fundamental for the formation of clouds and their development over time. The goal of the tasks in this group is to create visualizations that support a better understanding of their complex interaction.
Typical measures to describe turbulence are e.g., vorticity and divergence. Provide a visualization of at least one of those measures together with the existing cloud visualization. What information can be gained from visualizing those fields and what can you derive about the relation between those turbulence measures and the cloud systems? Visualize the temporal development of those turbulence measures during cloud evolution. How can they be effectively combined with the previous visualizations and what is their additional value to capture temporal behavior?
Trajectories are a powerful tool to describe long-term interactions between the wind field and clouds. To visualize trajectories either use the short (20 timesteps) or long (240 timesteps) timeseries to compute different types of field lines (e.g., streamlines, pathlines, streaklines) in the wind field. What additional information (compared to tasks 2.1. and 2.2.) can you derive about the long-term interaction of the wind field and the simulated cloud structures? Can you visualize the relation between trajectory properties and cloud development? As trajectories are often prone to create visual clutter; can you come up with improved visualizations to reduce this problem?
One very important cloud-related wind property is the up- and downdraft. Up- and downdraft is encoded in the vertical component of the wind field, but also in the vertical displacement of trajectories. Create a visualization that displays both, the updraft, as well as the downdraft in one of the larger cumulonimbus cloud systems. Your visualization may use the local vertical component of the wind vector, as well as the height change of trajectories from task 4.3. What statements can be derived about the up- and downdraft within the clouds systems? What is the relation of local up- and downdraft, trajectories, and the development of the cloud system?
Although there is not much precipitation over the Alps in this specific simulation, there is a lot of orography induced turbulence over the Alps. Create a visualization that highlights the relation of wind field properties and the orography. Visualize those parts that exhibit a strong topography induced circulation - How do they differ from the turbulence within larger clouds and clouds that are less affected by such ground effects?
Group B – Comparing Resolutions
Our simulations are not only performed for one fixed spatial resolution, but with three very different resolutions for three domains. Furthermore, 2D variables are written out with a higher frequency than 3D variables. The goal of the following tasks is to compare the various 2D and 3D variables in each domain in terms of precision, visibility of features and effectiveness.
For this visualization task, use a selection of 2D/3D variables and compare each of them in all three domains and visualize the differences, as well as the similarities. When are the lower resolutions sufficient, and when do we really need the finest resolution available? Please use the 2D/3D ICON data for all three domains for solving this task.
Does a finer grid resolution just resolve more details, or do even new features emerge? Are the cloud patterns and wind field structures self-similar? Please use the 2D/3D ICON data for all three domains for solving this task.
The simulation produces two types of output: A 2D average of the entire domain with high temporal and spatial resolution, as well as a lower resolution 3D output that is stored less frequently, but contains the data stored at individual height levels. Please create visualizations to visually combine both outputs. Please use the 2D ICON data and the 3D lon/lat low resolution data, both for Domain 2. Focus thereby especially on CLW and CLI for 3D and CLWVI and CLIVI for 2D. The latter ones are vertically integrated data from CLW and CLI. Can the additional temporal resolution of the 2D output be used to gain more or the same information about the evolution of 3D cloud structures? What additional information can be gained through the 3D output (e.g., like multiple layers of clouds)? You may also use the additional variables available stored within the 2D ICON netCDF files.
Group C – Cloud Classification and Tracking
The primary goal of the HD(CP)² project is to simulate and study clouds and precipitation processes. A wide variety of clouds exist, which can be classified using Luke Howard's cloud classification scheme . The overall goal of the tasks in this group is to detect and classify individual clouds, as well as to track and study their evolution over time.
Which properties characterize clouds besides their water content and geometrical dimensions? Can you create a visualization where all cloud properties are shown simultaneously?
By combining different cloud properties it is possible to classify clouds in different categories. Is it possible to distinguish different cloud types/regimes by using their phase-space? Luke Howard established a cloud classification by "observing the sky" in the beginning of the 18th century. In which regard is his classification still valid, when confronting it with 3D high-resolution simulations?
Besides classifying clouds in individual time steps, the temporal evolution of the structures plays a crucial role. Propose suitable methods to visualize the temporal evolution of those cloud structures. What are good visual methods to communicate the temporal evolution of certain structures? Are there additional methods (temporal filtering or tracking), that can help to improve the understanding of the temporal evolution of those cloud structures? Can you make use of the higher temporal resolution in the 2D output?
If the cloud tracking was successful we can derive quantitative information about the lifetime and the evolution of cloud systems. Therefore, using the results of the the previous visualization tasks, create a quantitative visualization that displays the life and the evolution of three individual cloud systems of your choice.