Qatar Petroleum’s Rob Ross showed the extent to which the ArcSDE Geodatabase can be used in a geological context. GIS can capture a vast range of data types—from soil samples to geophysical horizons, digital terrain models and satellite imagery. The quality and source of input datasets can be captured as metadata and geoprocessing techniques can be applied to data. Ross showed how surface geological studies revealed a 2 meter high stand in the Holocene. GIS was used to visualize the ancient sea. In contrast, a 100 meter low stand during the Ice Age dried up the Persian Gulf completely. Such techniques have been applied to autoclassification of satellite imagery for geotechnical studies including mapping of near surface karsts. Ross believes that the key to GIS is in the judicious exploitation of metadata.
Sidney Pereira unveiled Petrobras’ pipeline Land Property Management System, an ArcGIS Server running on top of an Oracle database. The LPMS addresses the appraisal, negotiation, legal and accounting process. Development of the system benefitted from close cooperation with business users. Understanding and debugging workflows before starting to code was also key. Systems should assume minimal GIS knowhow from end users. Petrobras is now working on a Google Earth Enterprise front end for the LPMS integrating high resolution imagery, search and 3D visualization.
Paola Peroni showed off Exprodat’s exploratory spatial data analysis and uncertainty workflows around ESRI Geostatistical Analyst 10 (GA10). Much geoscience activity revolves around deriving a 2D surface from sparse data. The resulting uncertainty may not be effectively captured and carried through the workflow—with results of uncertain quality. Using test data (the Johansen data set from the Norwegian SINTEF CO2 Sequestration program—links/1009_9) Peroni demonstrated that many common assumptions about data (stationarity, normal distributions) are not true. The investigation highlighted areas of high uncertainty and used ‘stochastic concepts’ to manage uncertainty throughout the modeling process. GA10 ‘compares favorably with established E&P mapping packages’ although this was somewhat at odds with another of Peroni’s findings—that GA has a limited capacity for handling faults.
Larry Phillips (San Antonio Water Systems) demonstrated ArcGIS Data Interoperability (DI) module, pulling data from CAD (Microstation), land (Access) and accounting systems. DI provides support for hundreds of data formats, and ETL* functionality. A canvas lets users graphically assemble data workflows between sources and applications with scripting for transformations and ad-hoc conversions. Phillips offered some compelling metrics as to the merits of bulk transformations of GIS data sets with DI as opposed to manual changes.
ArcGIS 10 novelties announced at the show include how the notion of time is now embedded in GIS data for time lapse visualization and analysis. ArcGIS is now also a ‘true’ 3D GIS, offering 3D data models, editing, analysis, and visualization on the local machine. Users can now do virtually everything they do in a 2D environment in a 3D environment. Python scripting was highlighted for automating common tasks and analyses. Python lets users combine ArcGIS functionality with other scientific programming to extend the solutions. More from www.esri.com.
* Extract, transform and load.
© Oil IT Journal - all rights reserved.