Using the Species Distribution Workflow Dan Higgins – NCEAS Prepared for: Ecoinformatics Training for Ecologists LTER (Albuquerque) January 8-12, 2007 http://www.kepler-project.org
Jan 05, 2016
Using the Species Distribution Workflow
Dan Higgins – NCEAS
Prepared for:
Ecoinformatics Training for Ecologists
LTER (Albuquerque)
January 8-12, 2007
http://www.kepler-project.org
Goal : Predict Species Locations Based on Known Locations and Climate/Geographic
Spatial Data
• Goal is to correlate spatial data with known locations to predict other locations where a species might exist (Environmental Niche Modeling)
• Requires manipulation of geographically referenced point data and GIS raster data from variety of sources
Example Species Data
• Mephitis mephitis (skunk) – 814 locations• Zapus trinotatus (Pacific Jumping Mouse ) - 387
locations• Orthogeomys cuniculus (Oaxacan Pocket Gopher) – 2
locations
• Pappogeomys gymnurus (Llano Pocket Gopher) – 9
locations
Species Examples suggested by Towne Peterson
GDAL - Geospatial Data Abstraction Libraryhttp://www.remotesensing.org/gdal/
GDAL translatorlibrary connected toKepler actors via JNI
GDAL can input ~40different raster formats
Conversion betweendifferent projectionspossible
File format conversionsalso possible
Java-based actors created to read and manipulate ASCII grid files
ImageJ image processing package from NIH added as an actor to display and manipulate images
ImageJ has macro capabilities & numerous plug-ins
ImageJ - http://rsb.info.nih.gov/ij/
Java Actors for Handling ASC Grid Files
Convex Hull algorithm relatively easy to implement in Java
Java routines for the Convex Hull convert the polygon to a shape which can be ‘scaled’ in size
Scale Factor = 1Scale Factor = 2
Java Routines for Convex Hull Calculation and Rasterization
Java Actors for Rescaling and Merging
Java actors can rescale ASCII grid files, changing resolution and extent
Both nearest-neighbor and Inverse-Distance-Weighted algorithms implemented
Disk-based code allows very large grids to be manipulated (i.e not limited by RAM)
Grid rasters can be ‘merged’ with various operations on data in overlapping pixels
Start
Rescale and Clip
Merge
Finish
Museum Specimen Data (Digir)
34 ‘hits’ for ‘mephitis’ located in seach
Drag to workflow area to create a datasource
Search for species name (“mephitis”)
Specimen information can be ‘filtered’ using a built-in SQL database
SQL Filter Dialog
Fields included in Digir data
SQL filter specification to returnonly location data
i.e. (longitude, latitude)
Workflow Displaying Filtered Data
Single GARP Calculation of
Presence/Absence
Single-Species Best RulesetWorkflow
Predicted Occurrence LocationsUsing ‘Best’ Rulesets
Acknowledgements•This material is based upon work supported by:
•The National Science Foundation under Grant Numbers 9980154, 9904777, 0131178, 9905838, 0129792, and 0225676.
•Collaborators: NCEAS (UC Santa Barbara), University of New Mexico (Long Term Ecological Research Network Office), San Diego Supercomputer Center, University of Kansas (Center for Biodiversity Research), University of Vermont, University of North Carolina, Napier University, Arizona State University, UC Davis
•The National Center for Ecological Analysis and Synthesis, a Center funded by NSF (Grant Number 0072909), the University of California, and the UC Santa Barbara campus.
•The Andrew W. Mellon Foundation.
•Kepler contributors: SEEK, Ptolemy II, SDM/SciDAC, GEON