The Snow Data System at NASA JPL

Tuesday, 16 December 2014
Ross Laidlaw1, Thomas H Painter1, Chris A Mattmann1, Paul Ramirez2, Kat Bormann1, Mary J. Brodzik3, Annie Bryant Burgess4, Karl Rittger5, Cameron E Goodale2, Michael Joyce1, Lewis John McGibbney1 and Paul Zimdars2, (1)NASA Jet Propulsion Laboratory, Pasadena, CA, United States, (2)Jet Propulsion Laboratory, Pasadena, CA, United States, (3)University of Colorado at Boulder, Boulder, CO, United States, (4)University of Southern California, Computer Science, Los Angeles, CA, United States, (5)National Snow and Ice Data Center, Boulder, CO, United States
NASA JPL’s Snow Data System has a data-processing pipeline powered by Apache OODT, an open source software tool. The pipeline has been running for several years and has successfully generated a significant amount of cryosphere data, including MODIS-based products such as MODSCAG, MODDRFS and MODICE, with historical and near-real time windows and covering regions such as the Artic, Western US, Alaska, Central Europe, Asia, South America, Australia and New Zealand.

The team continues to improve the pipeline, using monitoring tools such as Ganglia to give an overview of operations, and improving fault-tolerance with automated recovery scripts. Several alternative adaptations of the Snow Covered Area and Grain size (SCAG) algorithm are being investigated. These include using VIIRS and Landsat TM/ETM+ satellite data as inputs. Parallel computing techniques are being considered for core SCAG processing, such as using the PyCUDA Python API to utilize multi-core GPU architectures. An experimental version of MODSCAG is also being developed for the Google Earth Engine platform, a cloud-based service.