Multivariate-Clustering-ArcGIS

This repository includes Jupyter notebooks outlining the workflow used to generate regional benthic habitat classification maps using ArcGIS Pro geoprocessing tools. This workflow uses similar non-hierarchical multivariate clustering methods as those described in McQuaid et al. (2023).

This content was used during a hands-on training course as part of the Deep Ocean Observing Strategy "Deep Ocean Collective Solution Accelerator" in October 2023 at Scripps Institution of Oceanography, La Jolla, CA, USA.

This workflow was run using ArcGIS Pro 3.1.2. The ArcGIS Pro Project file can be downloaded at this link: https://arcg.is/1DOOXr0. Note that the URL needs to be copy and pasted in a browser.

Below is an overview of what each notebook covers:

Step1_Data_Access.ipynb:

• Accessing GEBCO bathymetry data, clipping that data to a region of interest, and masking that data to exclude values above sea level.
• Running the Benthic Terrain Modeler
• Accessing Bio-ORACLE and NEMO-MEDUSA raster layers

Step2_Run_Clustering.ipynb:

• Aggregating data and converting either to hex bins or points
• Creating feature class containing all layers
• Run Multivariate Clustering
• Visualizing clustering outputs using box plots and different mapping techniques

Want to get involved in the DOOS Habitat Mapping working group?

📧 Reach out to us: [email protected]