A CyberGIS Framework for the Synthesis of Cyberinfrastructure, GIS, and Spatial Analysis

Abstract

Cyberinfrastructure (CI) integrates distributed information and communication technologies for coordinated knowledge discovery. The purpose of this article is to develop a CyberGIS framework for the synthesis of CI, geographic information systems (GIS), and spatial analysis (broadly including spatial modeling). This framework focuses on enabling computationally intensive and collaborative geographic problem solving. The article describes new trends in the development and use of CyberGIS while illustrating particular CyberGIS components. Spatial middleware glues CyberGIS components and corresponding services while managing the complexity of generic CI middleware. Spatial middleware, tailored to GIS and spatial analysis, is developed to capture important spatial characteristics of problems through the spatially explicit representation of computing, data, and communication intensity (collectively termed computational intensity), which enables GIS and spatial analysis to locate, allocate, and use CI resources effectively and efficiently. A CyberGIS implementation—GISolve—is developed to systematically integrate CI capabilities, including high-performance and distributed computing, data management and visualization, and virtual organization support. Currently, GISolve is deployed on the National Science Foundation TeraGrid, a key element of the U.S. and worldwide CI. A case study, motivated by an application in which geographic patterns of the impact of global climate change on large-scale crop yields are examined in the United States, focuses on assessing the computational performance of GISolve on resolving the computational intensity of a widely used spatial interpolation analysis that is conducted in a collaborative fashion. Computational experiments demonstrate that GISolve achieves a high-performance, distributed, and collaborative CyberGIS implementation.