Ecological Niche Modeling (GARP) Workflows


A number of  Ecological Niche Modeling (GARP) workflows have been created using Kepler. Most are just pieces of a final, complete workflow.. Brief descriptions of these 'samples'  are given below

1. Baseline 3-Actor GARP - Browser Display
This is the basic starting point for many of the ENM workflows. It consists of Desktop GARP code wrapped into 3 actors with inputs from static (pre-defined) files that are already in the exact formats required by GARP (e.g. layers in *.raw format). The final occurrence distribution map is displayed using a browser.
2. Baseline 3-Actor GARP - ImageJ Display
This workflow is the same as the previous one except the actor for displaying the resulting map is now the ImageJ actor, based on the ImageJ image processing system developed at NIH. This actor allows the user to carry out a variety of manipulations on the resulting image.
3. GARP with Occurrence Data from Digir/Ecogrid
This workflow gets the GARP occurrence data from a DarwinCore data search through the Ecogrid. It illustrates how Kepler/SEEK technology can supply dynamic data sources.
4. GARP with Occurrence Data and Layer Integration
This example extends the previous model by adding the dynamic integration of a set of geographic layers that are in *.ASC grid format. In this case, the ASC grids all have the same extent and grid sizes. These layers are feed through an actor which converts them to the raw files needed for GARP and automatically creates the *.dxl file which summarized the layers for GARP.
5. GARP with Occurrence Data and Ecogrid Layer Integration
This example obtains occurrence data and one layer of ASC spatial data from the Ecogrid. The spatial data layer is regridded and integrated with other ASC layers to create the *.RAW and *.dxl files needed by GARP. This example thus illustrates how one can add to spatial data with new source. (Note: spatial data is low resolution to minimize download and run times.)
6. GDAL Input and ReProjection
This workflow shows an example of using the GDAL system for input of specialized GIS format files and reprojection of spatial data into desired forms.
7. GARP Omission Commission Example
The example calculates omission and commision values from the output of a GARP prediction map and a 'test set' of occurence longitudes and latitudes. Results are written to a local file as a table.
8. GARP Best Subset Example (using R)
This is an example of the RExpression actor with a custom R script for calculating the 'best subset' of a number of GARP runs. The example takes the omission/commission table as an input.
9. Overall ENM Workflow
This is a high level view of an ENM workflow constructed to give a view of all of the high level actions that need to be carried out. It started as a simple conceptual workflow built with Composite actors. Note that one can "Look Inside" many of these composite actors to see details of their actions.
10. Climate Layer Data Import and Formatting
IPCC climate data has been described using EML and saved on the Ecogrid/SRB. It can thus be obtained using a Kepler data search and the result dragged to the screen as a data source. Each file, however, contains data for each month and that data needs to be separated/combined to create annual/winter/spring/summer/fall data sets in ASC grid formats that can be combined to create GARP geographic layers. This is a test workflow which creates and displays a ASC file.
11. Convex Hull Operations (GRASS-based code)
Actors for calculating a Convex Hull and related GIS operations are shown in this test workflow. The actors used code from the GRASS GIS system recompiled using JNI interfaces to Java
12. Convex Hull Operations
(Java-based code)

This is an example of Convex Hull and related operations implemented completely in Java.
13. Rescaling Example
Example workflow that changes the grid size and extent of input *.ASC grid files
14. Grid Merging Example
This workflow starts with a grid image of the entire world, creates rescaled images of North and South America, and then recombines the 2 rescaled grids into a new grid of the western hemisphere using the MergeGrid actor.
15. ASC to RAW,  DXL Example
Workflow which takes a set of *.ASC grid files and creates corresponding *.RAW filed for GARP input along with the *.dxl layer summary file
16. IPCC Base Climate Layers
This workflow copies all 10 of the IPCC historical climate information files from the Ecogrid to the local cache and converts them to *.ASC files. The default files are annual averages; user can change to minimums or maximums and change time periods.
17. Hydro1K Data Preparation
This workflow shows how one inports Hydro1K digital elevation model derived spatial data and processes it for use in a GARP model. GDAL actors are used to import and re-project the spatial grid data.
18. GARP Single Species Best RuleSet - Local Files
This workflow runs the GARP algorithm multiple times and calculates omission and commission values to determine 'best' rulesets. Spatial and location data are read from local files. It is thus assumed that this input data has been previously prepared.
19. GARP Single Species Best RuleSet - Local Spatial Files with DigiR Location Data
This workflow runs the GARP algorithm multiple times and calculates omission and commission values to determine 'best' rulesets. Spatial data is read from local files. It is thus assumed that this input data has been previously prepared. Location data is read from a DigiR ecogrid data source, making it easy to change the input species by just doing another data search.
20. Prepare Climate Change Prediction Data
This is an example of how an IPCC climate change dataset is processed by combining with the historical data to produce a grid that can be used for ENM/GARP future predictions.
21. Convert Baseline Layers (local files) to GARP raw files and create DXL summary
This example converts a set of local *.asc layer fiiles to GARP RAW files and creates the *.dxl file needed by GARP.