openModeller - Open Source Spatial Distribution Modeller
CRIA - Centro de Referencia em Informacao Ambiental (SP - Brazil)
(Reference Center on Environmental Information)
FAPESP - Fundacao de Amparo a Pesquisa do Estado de SP
The original motivation of openModeller was to help automating the tedious and
complex tasks involved in species' potential distribution modelling regarding
georeferencing issues, resampling, existence of several raster formats and
algorithms, etc. This way:
a) One can spend time developing algorithms, testing, tunning or comparing
results, instead of clipping rasters, converting between different raster formats
or using different programs to run different algorithms. The same input data,
sampling methodology and ideally the same computing environment should be used
in a fair comparison between different algorithms.
b) Specialized components can help integrating data from distributed sources. For
example, one could look for species data in sources like the GBIF portal and for
environmental data in some remote source through WCS (OGC's Web Coverage Service).
Retrieving and integrating data from different sources using different protocols
and tools can be very complicated without using a single framework.
c) Different interfaces can be developed on top of the same modelling library.
d) Powerfull computers can be used to run heavy algorithms (like GARP)
in parallel (Cluster computing).
- Core modelling API implemented in ANSI C++ (compiles in GNU/Linux, Mac OSX and Windows).
- Algorithms dinamically loaded as plugins.
- Generic API to read & write raster data (GDAL and TerraLib drivers available).
- Generic API to load occurrence data (drivers available for tab-delimited text
files, serialized XML files, GBIF REST service and TAPIR/DarwinCore providers).
- Simple command-line/console interface.
- SOAP interface for remote modelling (server implementation available).
- Python binding.
The following libraries are necessary to compile and run openModeller.
If you installed openModeller from a pre-built package, you don't
need to worry about this (most dependencies should already be included
in the package).
* cmake (>= 2.6) if you need to compile the library.
* Expat XML parser
* Proj.4 (Cartographic Projections Library)
* GDAL (Geospatial Data Abstraction Library) version >= 1.1.9
* GSL (GNU Scientific Library) version >= 1.4 is needed to run
CSM (Climate Space Model) and ENFA algorithms.
* TerraLib (>= 3.2.0) is needed to enable the TerraLib occurrence and raster drivers.
* libX11 is needed to run the command line viewer (om_viewer).
* libcurl (>= 7.15.4) is needed to enable to TAPIR and GBIF occurrence drivers.
* SQLite3 is needed for the AquaMaps algorithm.
For compilation and installation instructions, please read:
For code documentation, see: http://openmodeller.cria.org.br/doxygen/
Besides the library itself, openModeller comes with a series of
command line tools. Please look at the README.txt file in the
"examples" directory for more information about each tool.
Windows users must use the DOS shell to run command line tools.
------------ x -----------