How Did We Tile Greenland?

This gallery was created for Earth Science Week 2015 and beyond. It includes a quick start guide for educators and first-hand stories (blogs) for learners of all ages by NASA visualizers, scientists and educators. We hope that your understanding and use of NASA's visualizations will only increase as your appreciation grows for the beauty of the science they portray, and the communicative power they hold. Read all the blogs and find educational resources for all ages at: http://ift.tt/1OKWxQV. I am a visualizer at NASA's Scientific Visualization Studio. It is my job to combine a variety of data to generate scientific visualizations of changes affecting our planet. I regularly use Earth science datasets from a variety of sources. Some data is from NASA's Distributed Active Archive Centers (the public data centers that store NASA's satellite data.) Data also comes from NASA scientists or scientists at universities or other research institutions. But whether I'm visualizing sea ice in the Arctic and Antarctic, glacier retreat, or seasonal snow cover and fires, the steps needed to handle the data are for all of these are similar. Data comes in all sorts of shapes and sizes. For example, a dataset for the same geographic region may be in different geographic projections. Or the same data may have been collected and processed at different resolutions. I put these datasets together to show them in layers. I need to handle the data in such a way that they correctly overlay one on to another. For example, I would not want our topography data showing the terrain around a mountain lake to be shifted or misregistered from the image data showing the blue region covered by the water. That would obviously be wrong! We use software that follows the exact mathematical rules defined for each type of projection. We can then align the geographic features of multiple datasets, transforming all of the datasets so that they accurate match one another. I'm sure many of you are familiar with the difference in the quality of pictures taken by cell phone cameras. The higher number of megapixels* (MP) captured by the camera usually means that the camera takes higher quality pictures. The same is true with data collected by satellite. Satellites take their "pictures" from directly overhead looking down as they orbit the Earth. But can you take a picture of a friend standing on a mountain hundreds of kilometers away? Of course not! Satellites gather many different kinds of data as they circle our planet from hundreds of kilometers above the surface. But how much surface detail can they obtain from such a distance as this? Satellite instruments may not cover a small area (like your back yard) in as much detail as your camera phone, but they are vastly improving. For many satellite instruments, a pixel value represents every 6, 10 or 12 kilometers, but for higher-resolution data a pixel value could represent as little as 15, 20 or 30 meters across. When we examine an area as large as Greenland, that covers more than 2 million square kilometers, this amount of data is enormous! To map Greenland, we used the topography, ice mask, and ocean mask datasets from the Greenland Ice Mapping Project. Each one of these were broken into 36 tiles that were 16,620 by 30,000 pixels in resolution or 498.6 MP per tile. That turns out to be 17,949.6 MP per dataset or 17.9 gigapixels (GP.)** We used the ocean and ice datasets to color RadarSat data from the Canadian Space Agency. Each of these were broken down into 25 tiles each about 421.8 GP. Every year generates a whopping 10,545 MP or 10.5 GP. But we had 6 years plus a mosaic dataset, which came out to 73.8 GP! Combining this data tile by tile into a single visualization would take a very, very long time for one person (or even a team!) to correct, so we developed a method that automated this process. We created a computer program that automatically extracts the coordinate and projection information from the geotiff images. For each image tile, this routine then passes the correct parameters to a projection routine that accurately positions the related texture tile. Using this method, our team successfully created a high-resolution Greenland visualization and accurately mapped 87 GP of data. This visualization was created for SIGGRAPH 2015.For more information, see: http://ift.tt/1VVqGMq -- Cindy Starr, Scientific Visualizer at Global Science & Technology / NASA GSFC (SVS) * 1 megapixel (MP) = 1 million pixels ** 1 gigapixel (GP) = 1 billion pixels

from NASA's Scientific Visualization Studio: Most Recent Items http://ift.tt/1OH001s
via IFTTT