Modeling species habitat with environmental predictor variables

In this example, we demonstrate how to predict the occurrence of a marine animal from points where the animal was observed by sampling oceanographic images (bathymetry, SST, and chlorophyll density), fitting a multivariate statistical model, and then predicting probability of occurrence maps from the oceanographic images. This type of analysis is sometimes called habitat modeling or species occurrence modeling.

Watch the webinar

This demonstration was given as a "webinar" to the EBM Tools Network on 15 October 2008. It consisted of a short PowerPoint presentation followed by a longer demonstration showing how to perform the analysis in ArcGIS using geoprocessing tools in MGET. You can download the presentation in either Microsoft PowerPoint format or Adobe PDF format.

You can also watch a video recording (50 MB, WMV format) and listen to an audio recording (7 MB, MP3 format) of the presentation. We highly recommend this, because the PowerPoint does not explain any details of the analysis. All of the details are captured in the ArcGIS demonstration.

To do this most effectively, you must download both the video and the audio and play them back simultaneously. This will require you to run two different playback programs at the same time. To play the video, you should use Microsoft's Windows Media Player (built into Windows) or another player that can play WMV video files. To play the audio, you can use any program that can play MP3 files, such as WinAmp, iTunes, or RealPlayer. The audio is about three seconds ahead of the video, so start the audio, switch to your video program, and then start the video.

At the end of the presentation, the EBM Tools Network encountered a technical problem with its conference calling system and had to terminate the audio right when I called for questions. (If you were on the call, this is when the "Great Lakes" recording kicked in.) Thus you will hear the audio cut out while the presentation keeps going. If you have any questions about the end of the presentation, please contact me ( jason.roberts@duke.edu).

We realize having the video and audio separated is a pain. We're mixing them together into a single file that is not Microsoft-specific. When this is done, we'll replace the existing files with a new one and modify the instructions above. We will also eventually prepare an HTML version of this example, similar to the other MGET examples, so you can just view web pages rather than watching a presentation.

Download the demonstration and run it yourself

If you have ArcGIS 9.2 or 9.3, you can download the demonstration and run it yourself. This is supported for ArcGIS 9.2 with Python 2.4 and ArcGIS 9.3 with Python 2.5. Other combinations of ArcGIS and Python will not work. Please contact  jason.roberts@duke.edu if you need it for another combination.

1. Make sure you have MGET 0.6 or later installed. (Note: future releases of MGET are not guaranteed to work. If you have a later release and it won't work, please contact us for assistance.) Make sure you have the Python pywin32 and numpy packages installed, and R installed (see MGET's installation instructions details).

2. Download the file HabModExample2.zip (274 MB) and save it to C:\ directory.

3. Extract the file above to the directory C:\HabModExample?2. If you have WinZip, you can accomplish this by right clicking on the file in the C:\ directory, selecting WinZip and then Extract to here. If you don't, just open the file with Windows and follow the instructions. You might have to change some options to get it to extract to C:\HabModExample?2. When you are done, your directory structure should look like this from ArcCatalog:

4. The file above was so large because it included the satellite oceanography data. But it does not include the species observations. You must download these yourself from OBIS-SEAMAP. Go to  http://seamap.env.duke.edu/datasets/detail/5, click on the Download tab, and download the data as an ESRI shapefile. You will have to accept the OBIS-SEAMAP Terms of Use. Save the file to C:\HabModExample?2\seamap5.zip.

5. Unzip C:\HabModExample?2\seamap5.zip into the directory that contains it. Now, the directory structure should look like this from ArcCatalog (you may have to refresh the window):

6. You are now ready to run the geoprocessing models. In ArcCatalog, open the toolbox that is labelled for your version of ArcGIS and Python. You'll see five models. Right-click on the first one and select Edit. When the model diagram comes up, click the File menu and select Run entire model. This first model will convert the oceanography data contained in the C:\HabModExample?2\OceanographyFiles directory to ArcGIS binary grids in the C:\HabModExample?2\OceanographyRasters directory. This will take 30 to 60 minutes, depending on the speed of your computer.

7. After the first model completes, run each of the remaining models in sequence using the same method. Right-click, select Edit, click File and select Run entire model. These will run much faster than the first one.

If you have any questions, please email  jason.roberts@duke.edu.