Amin tayyebi: Big Data and Land Use Change Science
- 1. 1
- 2. University of Tehran, Iran Survey and Geomatic Engineer (GIS, Remote Sensing, Photogrammetry and Geodesy) University of Tehran, Iran GIS Purdue University, USA Department of Forestry and Natural Resource University of Wisconsin-Madison, USA Wisconsin Energy Institute Landscape Ecology University of California-Riverside, USA Center for Conservation Biology Department of Botany and Plant Sciences 2
- 3. Introduction Land cover and land use History of land change science Sustainability Big data and land use change science Software development -> LTM-HPC Summary of other projects 3
- 4. Introduction Land cover and land use History of land change science Sustainability Big data and land use change science Software development -> LTM-HPC Summary of other projects 4
- 5. Land cover and land use One third to one-half (Turner, 1995) Land use cover change (Foley et al., 2005) Land change science 5
- 6. Moving to urban areas In 1900 (<10%) In 2050 (>50%) Occurred on <3% Doubles every 30 years Approach 10% by 2070 78% of carbon emissions 60% of water use … 6
- 7. Conversion of forest to agriculture in the Amazon Local temperature Carbon dioxide 7
- 8. 8
- 9. Fragmentation of natural habitats Richness and abundance 9
- 10. Earth as a system Sustainability Current and future needs Land change science (Turner, 2007) Future and historical land use map 10
- 11. Introduction Land cover and land use History of land change science Sustainability Big data and land use change science Software development -> LTM-HPC Summary of other projects 11
- 12. Using Big Data to Simulate Land Use Change at a National Scale: An Application of Land Transformation Model-High Performance Computing (LTM-HPC) 12
- 13. Managing the nation’s fish habitat at multiple spatial and temporal scales in a rapidly changing climate Land use Climate change Fish habitat Research team Scientists from the USGS, University of Missouri, Michigan State University, and Purdue University Series of meeting (3 years) 13
- 14. Limitations Discrete time periods Particular regions Coarse spatial resolution Multiple land uses Understand global process Forecasting annual multiple land use changes at continental scale -> 2000 to 2100 14
- 15. Modeling land use change Big data (GIS and remote sensing) Data mining (Artificial neural network) Calibration Validation Forecasting Products and applications Software Programming (Python, C#, C++ and batch) Communication (XML) Parallel processing (High performance computing)
- 16. 16
- 17. Big data (Python) Create pattern file (C# executable) Data mining Artificial neural network (C++ executable) Calibration Land use change quantity (C# executable) Suitability map (C# executable) Simulated map (C# executable) Validation (C# executable) Forecasting (C# executable) Products and applications 17
- 18. Drivers in time 1 Land use change 18
- 19. Workstation Quantity of files Size of files Server Parallel processing … 19
- 20. 20 Drivers in 1992 Land use map Results Distance to road NLCD 1992 Pattern file Distance to urban NLCD 2001 Suitability map Distance to stream Change map Simulated map Distance to city center ---- Error map Distance to highway ---- ---- Slope ---- ---- Gross domestic product ---- ---- Exclusionary zone (existing urban, water, state parks and others) ---- ----
- 21. 20722 × 11 = 227942 ~ 228K 20722 × 4 = 82888 ~ 83K 21
- 23. 23
- 24. 20722 × 11 = 227942 ~ 228K 20722 × 4 = 82888 ~ 83K 24
- 25. 25
- 26. 26
- 27. Big data (Python) Create pattern file (C# executable) Data mining Artificial neural network (C++ executable) Calibration Land use change quantity (C# executable) Suitability map (C# executable) Simulated map (C# executable) Validation (C# executable) Forecasting (C# executable) Products and applications 27
- 29. 29
- 30. 30
- 32. Big data (Python) Create pattern file (C# executable) Data mining Artificial neural network (C++ executable) Calibration Suitability map (C# executable) Land use change quantity (C# executable) Simulated map (C# executable) Validation (C# executable) Forecasting (C# executable) Products and applications 32
- 33. Quantity of change between times 1 and 2 Simulated map in time 2 Sort suitability values Reference map (time 2) Status 1 (Non-Urban) Status 2 (Urban) Reference map (time 1) Status 1 (Non-Urban) A B Status 2 (Urban) C D Drivers in time 1 Suitability Map 33
- 34. B 34
- 35. 20722 × 11 = 227942 ~ 228K 20722 × 4 = 82888 ~ 83K 35
- 36. 36
- 37. 37
- 38. Big data (Python) Create pattern file (C# executable) Data mining Artificial neural network (C++ executable) Calibration Land use change quantity (C# executable) Suitability map (C# executable) Simulated map (C# executable) Validation (C# executable) Forecasting (C# executable) Products and applications 38
- 39. Reference map (time 2) Status 1 (Non-Urban) Status (Urban) Simulated map (time 2) Status 1 (Non-Urban) True Negative (TN) False Negative (FN) Status 2 (Urban) False Positive (FP) True Positive (TP) Future Scenario 39
- 40. 20722 × 11 = 227942 ~ 228K 20722 × 4 = 82888 ~ 83K 40
- 41. 41
- 42. 42
- 43. 43
- 44. 44
- 45. Big data (Python) Create pattern file (C# executable) Data mining Artificial neural network (C++ executable) Calibration Land use change quantity (C# executable) Suitability map (C# executable) Simulated map (C# executable) Validation (C# executable) Forecasting (C# executable) Products and applications 45
- 46. 46
- 47. 20722 × 11 = 227942 ~ 228K 20722 × 4 = 82888 ~ 83K 47
- 48. 48
- 49. 49
- 50. 50
- 51. 51
- 52. 52
- 53. Big data (Python) Create pattern file (C# executable) Data mining Artificial neural network (C++ executable) Calibration Land use change quantity (C# executable) Suitability map (C# executable) Simulated map (C# executable) Validation (C# executable) Forecasting (C# executable) Products and applications 53
- 54. In Great Lakes area, LaBeau et al., (2014), used the future land use maps (between 2010-2050) Land use (agriculture and urban) and phosphorus delivery Increase P loadings by 3.5–9.5% 54
- 55. 55
- 56. 56
- 57. Developing a model to simulate land use change at continental scale LTM-HPC Sustainability Climate, water quality and biodiversity Big data and land change science Land use legacy 57
- 58. Tayyebi, A., Pekin, B. K., Pijanowski, B. C., Plourde, J. D., Doucette, J. S., and D. Braun. (2013). Hierarchical modeling of urban growth across the conterminous USA: Developing meso-scale quantity drivers for the Land Transformation Model. Journal of Land Use Science, 8(4), 422-442. Pijanowski, B. C., Tayyebi, A., Doucette, J., Pekin, B. K., Braun, D., and J. Plourde. (2014). A big data urban growth simulation at a national scale: Configuring the GIS and neural network based Land Transformation Model to run in a High Performance Computing environment. Environmental Modelling & Software, 51, 250-268. Tayyebi, A., Pekin, B. K., and B. C. Pijanowski. (In review). Urbanization trends across the conterminous of USA from 1900 to 2100: Lessons learned from studies in 11 mega-regions. Regional Environmental Change. 58
- 59. Advisor -> Bryan C Pijanowski Post Doc -> Burak K Pekin GIS Specialist -> Jarrod Doucette GIS Specialist -> James Plourde IT Specialist -> David Braun 59
- 60. Introduction Land cover and land use History of land change science Sustainability Big data and land use change science Software development -> LTM-HPC Summary of other projects 60
- 61. SmartScape™: A web-based decision support system for strategic agricultural land use policy development 61
- 62. 62
- 63. 63
- 64. Urban Heat Island Variation across a Dramatic Coastal to Desert Climate Gradient: An Application to Los Angeles, CA Metropolitan Area 64
- 65. 65
- 66. Video time 66
- 67. 67
- 68. 68
- 69. 69
- 70. Mike Batty 70