Geospatial Data Sciences Presented by Ranga Raju Vatsavai

Geospatial Data Sciences Presented by Ranga Raju Vatsavai Auroop Ganguly Budhendra Bhaduri Geographic Information Science and Technology Computational Sciences and Engineering

HP Visualization Terabytes of spatially referenced data per day Knowledge Discovery Geospatial data sciences Data mining Spatial analysis Machine learning Pattern recognitio n Statistics Disaster mapping and monitoring Actionable insights/ knowledge Nonlinear dynamics Efficient storage and retrieval methods Geospatial intelligence discovery Extremes Social networks National security HPC Errors and uncertainty Geo. Databases Damage Petabytes of spatiotemporal data archives 2 Total population Senior 100% 791, 361 90, 773 75% 1, 594, 806 187, 677 50% 1, 505, 196 184, 372 25% 1, 701, 593 224, 279 Total 5, 592, 956 687, 101

Pattern detection and decision support Consequences Additional data hypothesis generation Action End-user Process modeling Offline Predictive analysis Pattern detection Analys t Decisio n Body of actionable knowledge HIT Decision support Online Decision making Data integration Stream of new data 3 Event summary

Spatial classification and prediction · Spatial autoregressive regression (SAR) SAR W matrix 2000 60, 000 Overlap 3, 600, 000 Node 1 4 · Challenge Node 2 - Large neighborhood matrices - Overlapping computation 3, 600, 000 60 km IKNOS 1 m 60, 000 TM 30 m 2000 60 km · Markov random fields (MRF)

Extremes · A new measure to describe the relationship among extremes in space and time: - Simultaneous occurrence of 100 -year return levels: · Perfect dependence 100 year event. · No dependence 10000 year event. Correlation Tail Dependence Rainfall observations in South America Correlation Tail Dependence Simulations from CCSM 3 Climate Model · Challenge Spatiotemporal dependence estimation is computationally expensive (1000 grid cells results in scanning more than half a million pairs). 5

Managing uncertainty in spatiotemporal environments (MUSE) Knowledge Discovery · Reasoning · Data mining · Spatial analysis · Visualization n io t ga pa o Pr n Geo. DB io at nt e s re p Re s e at d Up Distributed asynchronous update framework · Ontology · Semantic · Reasoning ce r ou S Modeling · Conceptual · Logical Data collection source of uncertainty · Challenge - Expensive query processing and spatial analysis 6

Contacts Ranga Raju Vatsavai Geographic Information Science and Technology Computational Sciences and Engineering (865) 576 -3569 vatsavairr@ornl. gov Auroop Ganguly Geographic Information Science and Technology Computational Sciences and Engineering (865) 241 -1305 gangulyar@ornl. gov Budhendra Bhaduri Geographic Information Science and Technology Computational Sciences and Engineering (865) 241 -9272 bhaduribl@ornl. gov 7 Bhaduri_GDS_SC 07