This document describes PARASID, an approach for near-real time monitoring of habitat change using neural networks and MODIS data. PARASID aims to monitor changes at continental to global scales with a turnaround time of less than 3 months. It uses machine learning to identify anomalies in NDVI values compared to expected values based on climate and site characteristics, indicating potential human impacts. The approach has been tested on sites in Colombia, Paraguay, and the Amazon, detecting deforestation rates with 76% coverage of Colombia. PARASID provides an early warning system for broad-scale habitat conversion monitoring but not detailed local analyses. Further methodological development and adoption by countries and institutions is planned.
3. Objectives of PARASID HUmanImpactMonitoring And Natural Ecosystems Providenear-real time monitoring of habitatchange (<3 monthturn-around) Continental – global coverage (forests AND non-forests) Regularity in updates
4. The Approach The change in greenness of a given pixel is a function of: Climate Site (vegetation, soil, geology) Human impact
5. Machine learning Wetherefore try tolearnhoweach pixel (site) respondstoclimate, and anyanomolycorrespondstohumanimpact Machine learning (or neural-network), is a bio-inspired technology which emulates the basic mechanism of a brain. It allows To find a pattern in noisy dataset To apply these patterns to new dataset
12. Methodology – Bayesian NN To detect novelties, Bayesian Neural Networks provide us two indicators The predicted value The probability repartition of where the value should be The first one allows us to detect abnormal measurements The second one allows us to say how sure we are a measurement is abnormal.
13. The Processing For South Americaalone, firstcalculationsapproximated 10 years of processingforthe NN tolearn: A map of 30720 by 37440 pixels 1,150,156,800 vectors 23 vectors per year 26,453,606,400 NDVI values to manage per year 9.5 years of data 251,309,260,800 individual data points Through various processes, optimizations and hardware acquisitions reduced time to 2 weeks for NN learning Detection takes 1 day
14. The Bottom-Line 250m resolution Latin American coverage (currently) 3 week turnaround from data being made available (4 week delay in MODIS going to NASA ftp) (3+4 = 7 weeks) Report every 16 days Measurement of scale of habitat change (0-1) and probability of event
16. Introduction Different test cases with different vegetation and climate types All the test are done with the same parameters Training parameters From 2000 to the end of 2003 Detections parameters From 2004 to May 2009 A detection map is created each 16 days within this period The process is close to be fully automated
17. Colombia – Río Caquetá Size 480 * 300 [km2] 14400000 [ha] Vegetation type Tropical forest
28. PARASID - Colombia Direct usage for developing negoatiation position of Colombia in Copenhagen September 2009 Colombia were going to COP15 with a figure of 100,000Ha/year deforestation PARASID analysis predicting MINIMUM 180,000Ha/year, most likely 250-300,000Ha/year Resulted in change in negotiation plan, and increased relevance of expansion of Chiribiqueti NP Discussions underway for PARASID to become a 1st tier monitoring tool for National Parks
29. 76% coverage of country Approx. 250,000Ha/year average 90% increase in deforestation rate 2004 - 2009
31. What PARASID is…. 1st tier monitoring tool for looking at broad-scale patterns of habitat conversion National and regional platform for consistent measurement of habitat conversion Suitable early-warning system Important policy-influencing tool
32. What PARASID is not….. Detailed monitoring tool for examining local-scale impacts and changes – 2nd and 3rd tier analyses are needed A system for monitoring steady degradation
33. Outlook and next steps Three major pushes right now: Methodological development Long wish list…. Getting it out there Adoption by countries Adoption by institutions Website and online data Writing it up Methodological paper imminent submission Latin American patterns in habitat change Effectiveness of Pas across the continent + many more…