Developing a Generic Framework to Support Multi-dimensional Earth Observing System Data in GIS Applications



Jiang, Yunfeng

Journal Title

Journal ISSN

Volume Title



Earth Observing System (EOS) data are expanding at an unprecedented rate due to the fast development of advanced data acquisition technologies. These data provide valuable, long-term record of change and dynamics about our Earth, and therefore are paramount in addressing key national and global challenges in climate change, water use and quality, natural disasters, weather forecasting and warnings, renewable energy, agriculture, forestry and natural ecosystems, coasts and oceans, and national security. As a result, they have been increasingly used in various GIS applications by both government and science communities. However, many varied formats and standards have been defined to organize and store the EOS data that are highly tailored for different applications by different organizations over the past decades. Many of these data are in old formats, and specialized geospatial tools are required to interpret and use them. This makes it difficult to incorporate EOS into GIS and it very ineffectively to analyze in either commercial or open-source GIS tools. On the other hand, most GIS systems cannot comprehensively process and utilize all types of EOS data, and there are always unexpected issues and errors while importing and manipulating EOS data. To reconcile the conflicts between EOS data and GIS systems, initiatives have been made for developing a general methodology to solve EOS data compatibility in GIS using common standards. However, no solutions are currently available to support the processing of all types of EOS data products. The objective of this research is to explore the barriers and strategies of integrating various types of EOS data in GIS applications. Specifically, the research investigates and solves three key technical problems including: (i) designing a generic and heuristic plug-in framework for consuming different types of EOS data; (ii) developing a series of functions to fix the problem occurring when using EOS data in GIS applications; (iii) optimizing HDF4/HDF5 data drivers of GDAL for enhancing its capability of handle EOS data; and (iv) developing an open source GIS extension to enhance the capability of GIS systems in accessing EOS data. One research result of this thesis is optimized source code of Geospatial Data Abstraction Library (GDAL) commonly used in most GIS systems for handling geospatial raster and vector data, without impacting the original function on reading non-EOS data products. The optimized GDAL fixes the issues of HDF4 and HDF5 data drivers used to access HDF datasets and overcome limitations in processing multiple dimensional datasets posed by the current GDAL version. Finally based on the optimized GDAL, an open source extension is developed to support the access of more EOS data of different types and fill in the gap between GDAL and commercial GIS software (e.g. ArcGIS) or open source GIS projects (e.g. QGIS). A series of EOS data products collected from NASA’s Atmospheric Scientific Data Center (ASDC) are selected as study cases for demonstrating the effectiveness and applicability of the proposed framework and tools. The enhanced GDAL and GIS extension enable and encourage more GIS users to use EOS data in GIS software for different research or applications. Additionally, a series of Application Programming Interfaces (APIs) are provided to allow other developers in GIS communities to integrate these interfaces into their GIS applications. It is concluded that the proposed plug-in framework can be effectively applied to different domains for handling the current problems or limitations of interpreting multi-dimensional dataset, without compromising their original functions.



Plug-in framework, EOS data, GIS, GDAL, HDF