92nd American Meteorological Society Annual Meeting (January 22-26, 2012)

Monday, 23 January 2012
Developing 21st Century Climate Services for the Pacific Northwest
Hall E (New Orleans Convention Center )
Kathie D. Dello, Oregon State Univ., Corvallis, OR; and A. Wiggins, P. W. Mote, and M. Bailey

State climate offices respond to requests daily regarding past events and current weather and climate with its long-term context. Some of these requests require many different sources and examining huge amounts of data to get the answered. The general public and media however, are not interested in large amounts of data or spending the time sifting through it. The typical customer only wants the information they requested and they usually want it in short order. Regional Climate Services (RCS) was developed as a solution to thinly funded state climate offices to allow users to visualize and explore many different data sets at any time without using the resources of the climate offices. The RCS system also engages the increasingly tech-savvy user, cutting out arduous data processing and offering visualizations, graphs and other interactive applications. The developed web application for the RCS project currently tries to balance the need for vast amounts of data to be able to compare all facets of the data sets, with tools to pre-process and compress the information to quickly deliver informative content to users and maintain an excellent experience. RCS employs the use of prefetching of data for user experience, pre-processing of tiled images and data comparisons, memory caching server side, and data manipulation client side using shaders and the gpu for data culling and scaling to give clients a great experience and limit the amount of data transfer and recalculation from the server. Allowing the client to do data manipulation client side we are saving terabytes of preprocessed images and data that would have to be stored to maintain the user experience as well as allowing them to create useful visualizations and explore the data sets completely.

Supplementary URL: