for orest est mon monitor itorin ing g for or action
play

For orest est Mon Monitor itorin ing g For or Action Action - PowerPoint PPT Presentation

For orest est Mon Monitor itorin ing g For or Action Action David Wheeler, Center for Global Development Robin Kraft and Dan Hammer, University of Maryland Scipy 2010 rkraft@umd.edu cgdev.org/forest Forest Monitoring for Action (FORMA)


  1. For orest est Mon Monitor itorin ing g For or Action Action David Wheeler, Center for Global Development Robin Kraft and Dan Hammer, University of Maryland Scipy 2010 rkraft@umd.edu cgdev.org/forest

  2. Forest Monitoring for Action (FORMA) • Set of algorithms to identify deforestation using satellite imagery quickly and cheaply Outline • Why deforestation? • What is FORMA? • Python and FORMA

  3. Caveats • Economist/geographer, not a programmer • Parallel computing broadly speaking – not necessarily in robust compsci terms

  4. Why do we care about deforestation? • Biodiversity • Local and regional environment • Climate change

  5. How do we deal with deforestation? • Address economic drivers • Monitor outcomes

  6. Monitoring: Issues with traditional methods • Slow turnaround • Relatively opaque methodologies • Data black box • Local accuracy vs. wide coverage • Needed software tools expensive and hard to use • Research aimed at scientists more than people on the ground. • BUT – it’s still really cool stuff!

  7. Forest Monitoring for Action • Set of algorithms to identify deforestation using rapidly updated, free satellite imagery. • Prototype: monthly maps at 1km resolution • Early-warning system to complement hi-res approaches • Special features – Rapid (monthly) updates: 2006-present – Potential pan-tropical application – Indonesia-wide prototype is unique

  8. The (very) basic intuition • Forested areas look green • Forest + fires + browning = deforestation? • FORMA algorithms search for telltale patterns of fires and “browning”

  9. FORMA Methodology • Identify relevant patterns and time trends in “greenness” and fires • Train algorithms using historical data (2000-5) • Apply parameters from training to subsequent data • Output: probability of forest clearing, by pixel

  10. Original workflow for Indonesia: Doesn’t scale • Two dual core Windows desktops • One 1tb hard drive that got quite full • One ArcGIS license ($3000) for GIS data and imagery • One Stata license ($1500) for statistical modeling • Python 2.5 as glue • Start to finish: 4 weeks • Start to finish on 4 cores: ~1+ week

  11. Pan-tropical application @ 250m resolution • ~130x more raw data • Data comes split into pieces – thanks NASA! • “Easy” to parallelize if you can get rid of pricey software and manage all the data …

  12. New workflow: Homebrew “Map/Reduce” in the cloud • Amazon Web Services (AWS): – EC2: small Linux spot instances – approx. $0.06/hr – SQS: fault tolerant job management – S3: persistent storage for intermediate and output data • Boto: interact with AWS • GDAL/OGR: load images as Numpy arrays; re-project images; GIS data management • Numpy: – Slice images into pieces, stack slices as time series in tabular format – Statistical modeling by pixel, pixel neighborhood or ecoregion means plenty of room for parallelization

  13. The cloud/FOSS difference for Indonesia prototype • 20 Linux instances: – 10x the compute power for $1.20/hour, with no software costs • Easily scalable with SQS – just re-run queue processing script on new instance • Per-image processing significantly faster – Numpy arrays are great! • Parallel pre-processing in the cloud – RUNNING FOR THE FIRST TIME THIS MORNING! – Preprocessing time drops from from 3 days to <5 hours • Allows us to experiment with different algorithms and data much more easily

  14. Future plans • Matplotlib + basemap for visualization • Numpy to mask out water and non-forest pixels • PiCloud? StarCluster?

  15. Maps of forest clearing: monthly time series for Riau, Indonesia

  16. Forested in 2000 270 km 167 mi Forest in 2000

  17. Cleared by 2005 270 km 167 mi Forest in 2000 Cleared 2000 - 2005

  18. What happened after 2005?

  19. Cleared by 10/2009 Probability 50 – 60% 60 – 70% 70 – 80% 80 – 90% > 90% 270 km 167 mi Forest in 2000 Cleared 2000 - 2005

  20. 11/2005 Forest in 2000 Cleared 2000 - 2005

  21. 12/2005

  22. 1/2006

  23. 2/2006

  24. 3/2006

  25. 4/2006

  26. 5/2006

  27. 6/2006

  28. 7/2006

  29. 8/2006

  30. 9/2006

  31. 10/2006

  32. 11/2006

  33. 12/2006

  34. 1/2007

  35. 2/2007

  36. 3/2007

  37. 4/2007

  38. 5/2007

  39. 6/2007

  40. 7/2007

  41. 8/2007

  42. 9/2007

  43. 10/2007

  44. 11/2007

  45. 12/2007

  46. 1/2008

  47. 2/2008

  48. 3/2008

  49. 4/2008

  50. 5/2008

  51. 6/2008

  52. 7/2008

  53. 8/2008

  54. 9/2008

  55. 10/2008

  56. 11/2008

  57. 12/2008

  58. 1/2009

  59. 2/2009

  60. 3/2009

  61. 4/2009

  62. 5/2009

  63. 6/2009

  64. 7/2009

  65. 8/2009

  66. 9/2009

  67. 10/2009

  68. Cleared by 10/2009 Probability 50 – 60% 60 – 70% 70 – 80% 80 – 90% > 90% 270 km 167 mi Forest in 2000 Cleared 2000 - 2005

  69. 12/2005

  70. 1/2006

  71. 2/2006

  72. 3/2006

  73. 4/2006

  74. 5/2006

  75. 6/2006

  76. 7/2006

  77. 8/2006

  78. 9/2006

  79. 10/2006

  80. 11/2006

  81. 12/2006

  82. 1/2007

  83. 2/2007

  84. 3/2007

  85. 4/2007

  86. 5/2007

  87. 6/2007

  88. 7/2007

  89. 8/2007

  90. 9/2007

  91. 10/2007

  92. 11/2007

  93. 12/2007

  94. 1/2008

Recommend


More recommend