Hundreds of kilometers above Earth, precipitation satellites fly by, observing clouds and rainfall using passive and active sensors. Some of them measure the temperature of the atmosphere and surface, from which information can be gleaned about how much rain and snow are falling at that moment. Others bounce microwaves off the clouds, assessing what size water droplets and snowflakes are falling based on how those microwaves come back.

Data from these satellites are available at the NASA STORM website, where scientists can search for observations from over a dozen platforms going back over twenty years. When users are exploring which files they need, they have a couple of ways to understand the information contained before downloading them. These include static images, methods for searching by geographic area, and a two-dimensional variable explorer called THOROnline. Now, they have a way of visualizing the information in three dimensions, thanks to Cesium, in STORM Virtual Globe.

Matt Lammers, web analyst/developer for the NASA Precipitation Processing System (PPS), has created a system through which data can be read directly from the raw HDF5 files, converted into JSON, and visualized in Cesium without storing any post-processed files. This is enabled by the STORM website being built on top of the Java Tomcat service, which allows the creation of servlets written in compiled Java to be exposed online. STORM VG performs Ajax queries to grab the satellite data from the servlet.

Depending on the file being explored, STORM VG presents the user with two different ways of exploring the precipitation rate information. Level 3 precipitation data (meaning it is stored in a non-orbit-based way) simply displays the entire grid at once. Each grid point can be moused over for the raw integer precipitation rate value. Level 3 GPROF files feature 0.25 degree gridded precipitation rate data from a single satellite aggregated over an entire day. Level 3 IMERG files combine information from a number of satellites with surface observations on a half-hour basis. Level 2 data (stored in an orbit-based way), being often higher resolution, is made available fifteen minutes at a time. Using a scroll element, the user can look at any portion of the orbit, with both camera motion and a polygon making apparent where the new segment will appear.

Radar products (2ADPR and 2BCMB) provide information both in the horizontal and the vertical about precipitation rate, and so data points are visualized both on the surface and through the lower atmosphere, giving information about the overall structure of storms. Tropical cyclones stand out in these radar profiles, which is why the link provided here is for a perfect flyover by the Global Precipitation Mission (GPM) core satellite of the eye of Super Typhoon Vongfong in October 2014.