Computational Facilities for Biodiversity Research for Biodiversity Research e-Infrastructures Wouter Los Wouter Los University of Amsterdam (institute of Biodiversity and Ecosystem Dynamics) Computer-assisted photo identification Computer assisted photo identification Eric Pauwels Centre of Mathematics and Informatics, Amsterdam (Signals and images)
Richard Leakey (Nov 2008) (Nov 2008) “Th “The cost of the melt-down of Wall Street is t f th lt d f W ll St t i the next day in your news paper; do we know the costs of the melt down of our do we know the costs of the melt-down of our planet’s nature?” CompSust09
Bi di Biodiversity it Species (organisms and their populations) >10 7 species; species with 10 2 to 10 12 individuals Genes and DNA Genes and DNA 10 6 to 10 9 nucleotides in a DNA molecule Ecosystems habitats with 10 4 to 10 6 species, and manyfold interactions CompSust09
The big questions in The big questions in biodiversity research E Ecosystems t Time and evolution evolution Species DNA, proteins Scale and genes d @ Robert Guralnick & Andrew Hill (Univ Colorado) CompSust09
Even more pressing questions E i i Biodiversity loss, added to climate change, requires entirely new approaches and requires entirely new approaches and mitigation strategies. We need forecasts and measures of future changes and their uncertainty CompSust09
Understanding, predicting and managing change in biodiversity, landscapes and ecosystem services Landscapes are highly modified by human activities modified by human activities Multiple drivers and Multiple drivers and pressures affect the state of biodiversity Research to understand, predict and manage biodiversity and its changes y g CompSust09
Experimentation on a few The biodiversity system is complex parameters is not enough: and cannot be described by the simple sum of its components and relations f it t d l ti Limitations to scaling up Li it ti t li results for understanding of system properties We need new technologies to support the generation and the analysis th ti d th l i of large-scale data-sets on biodiversity. CompSust09
Understanding of the biodiversity g y system and its functions requires the analysis and modeling of large data sets to identify patterns and underlying processes underlying processes. This defines an infrastructure with • distributed observatories/ sensors, • interoperable databases, • computational capability, • and computational capacity. CompSust09
Building blocks of the research infrastructure Applications Analysis & Ecosystems LTER modelling g Species Interoperability GBIF Genes Observatories GenBank data functions CompSust09
A Architecture hit t Collaboration • Common Exploratory Environment • Collaborative Virtual Organisations Users Semantic annotation Semantic annotation E-Infrastructure Analysis and processing • Integration of resources • Integration of resources • Documented, shared workflows Composition • Grid computation Data • measurements, Resources observations & sensors • other infrastructures Statistical software Distributed computing power CompSust09
CompSust09 Collaboration Architecture t hit A E-Infrastructure Resources Composition Users
CompSust09 e-Infrastructure Biodiversity
Year 2012 A researcher has the innovative idea to combine distributional, genetic, ecological, phylogenetic, earth, and climatic data together in an statistical analysis to “predict” statistical analysis to predict not native species invasions, with special attention to the horizontal transfer of health related parasites in the host l d i i h h species. Year 2013 Year 2013 Year 2014 Year 2014 The WHO starts a campaign Our researcher builds her infrastructure work with a funding programme to space and attracts dozens of collaborators sustain the project as a main inventing additional functions Data providers inventing additional functions. Data providers health service also jump in. CompSust09
A Architecture hit t Users Semantic annotation Semantic annotation E-Infrastructure Analysis and processing Composition Resources CompSust09
Spatial interpolation: Kriging Uncertainty Spatial interpolation: Kriging Uncertainty Uncertainty light green: low g g dark green: high EcoGrid CompSust09
CompSust09 A prototype example
CompSust09
Color legend for all slides Africa Asia Central America Australia North America Europe South America Oceania CompSust09
Why not organising the CompSust en e-Biosphere Why not organising the CompSust en e Biosphere conferences back to back in the same location? London 1 -3 June 2 0 0 9 CompSust09
A Architecture hit t Users E-Infrastructure Composition Resources Data -> Observations & Sensors Statistical software Distributed computing power CompSust09
Thank you w.los@uva.nl y
Recommend
More recommend
