iGlobe is open-source software built on NASA World Wind virtual globe technology. iGlobe provides a growing set of tools for weather science, climate research, and agricultural analysis. Up until now, these types of sophisticated tools have been developed in isolation by national agencies, academic institutions, and research organizations. By providing an open-source solution to analyze and visualize weather, climate, and agricultural data, the scientific and research communities can more readily advance solutions needed to understand better the dynamics of our home planet, Earth.

iGlobe provides a flexible interface for sophisticated analysis and highly interactive visualization of NetCDF (Network Common Data Format) data. NetCDF, the data format typically used for weather and climate data, is large and complex in nature. Even the simple act of accessing NetCDF data is a computation- and data-storage-intensive undertaking. iGlobe is there for the international community to advance collectively solutions that address issues of concern to all.

iGlobe is a 4D virtual globe application using NASA World Wind visualization technology (www.goworldwind.org ). iGlobe integrates analysis of climate model outputs and remote sensing observations, combined with demographic and environmental data sets, to understand global and regional phenomena better, and provides impact analysis on a critical national resource, our agricultural industry. iGlobe allows seamless access to remote data repositories, allows users to run sophisticated data analysis algorithms on the server side, and provides accelerated statistical analysis on the client side via a thin client analytic engine able to incorporate server- side processing power.

iGlobe server-side analysis provides support for different data analysis algorithms purposed to identify patterns in spatial-temporal data, i.e., change detection, anomaly detection, clustering, and frequent-pattern analysis. The iGlobe client-side analysis also provides support for statistical operations on selected regions using any number of spatialtemporal data layers and parameters, i.e., spatial mean, median, variance, auto-correlation, etc.

This work was done by Patrick Hogan of Ames Research Center. ARC-15166-1A