1.4 Using Python to Expand MesoWest's Utilization of Surface Weather Observations

Monday, 7 January 2013: 2:15 PM
Room 12B (Austin Convention Center)
Chris Galli, Univ. of Utah, Salt Lake City, UT

Software to access, archive, and disseminate surface weather observations around the United States has been developed continuously over the past 15 years as part of MesoWest (http://mesowest.utah.edu). Data and metadata from ~20,000 active stations from ~100 distinct data streams have been obtained, stored in relational data bases, and disseminated through many different web accessible applications. This software has relied on a mix of shell scripts, perl, MySQL, and various graphical and scientific packages (e.g., Matlab and IDL). Further development within this disparate software environment was becoming increasingly difficult and costly.

A concerted effort has been underway during the past year to migrate core components of the MesoWest software to Python to more efficiently: (1) receive data streams from diverse sources using Python servers; (2) store metadata and data using PyTables and hdf5 storage formats; (3) improve real-time quality control algorithms; and (4) provide access to the metadata and data using a web API and improved graphical tools. Software development with Python was facilitated by a short tutorial course to the MesoWest software team. The speed at which both novice and experienced programmers transitioned legacy codes as well as developed new software using Python was remarkable.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner