9 - DEM Workflows with globato

globato connects the CUDEM workflow. It combines fetchez for data discovery, access, and source-level preparation; transformez for vertical datum and reference-frame transformation; and CUDEM tools for source filtering, weighting, stacking, interpolation, and output generation.

The result is a reproducible, recipe-based process that transforms diverse source datasets into a commonly referenced, composite DEM. It turns DEM building into a workflow that can be exported, inspected, modified, shared, rerun, and version-controlled.

In Day 1, we generated the Fiji test DEM and reviewed the YAML recipe used to define the workflow. In Day 2, we will focus on two primary globato commands:

globato cudem build    Generate a DEM workflow from command-line options
globato cudem run      Execute a saved YAML recipe

The main concept is:

build = define a DEM workflow using command-line options
run   = execute a saved YAML recipe

globato can also draw on module bundles that collect multiple source modules and their default settings under a single name, covered in the Module Bundles section below.

Basic Help Commands¶

Start with:

globato --help

Inspect the CUDEM workflow commands:

globato cudem --help

View the available build options:

globato cudem build --help

View the recipe-run options:

globato cudem run --help

The build command is useful when you want to define a workflow from a region, resolution, target reference frame, interpolation method, and source bundle.

The run command is useful when you already have a saved YAML recipe and want to execute or modify that workflow.

Module Bundles¶

Before defining a DEM workflow, it is useful to understand module bundles.

A module bundle is a preconfigured collection of source modules and processing settings. Instead of manually listing many individual data sources, a bundle allows a predefined group of sources to be referenced by one name.

In this workshop, we use:

global-bathy-topo

This bundle brings together global topographic and bathymetric sources into one reusable source collection.

The main idea is:

Instead of manually listing many source modules, use one named bundle.

Inspect Available Module Bundles¶

View the module-bundle commands:

globato modules bundles --help

List available bundles:

globato modules bundles list

Example bundles may include:

crm_alaska
crm_standard
global-bathy-topo
us_coastal_stream

Each bundle is designed for a different DEM-building use case.

Inspect the global-bathy-topo Bundle¶

Print the YAML definition of the global-bathy-topo bundle:

globato modules bundles dump global-bathy-topo

This displays the source modules, arguments, hooks, weights, and other settings included when the bundle is used in a DEM workflow.

The global-bathy-topo bundle may include source modules such as:

copernicus
copernicus_marine
mbdb
sysu
margrav
csb

It may also include processing hooks such as:

unzip
filename_filter
set_datatype
range_z
set-srs
stream_reproject
rq

The bundle provides a practical starting point, but it is not a black box. Its contents can be inspected, exported into a complete YAML recipe, and modified for a specific region or project.

Export a Fiji DEM Recipe¶

The Fiji YAML recipe used in this workshop can be generated from a globato cudem build command using the --export option.

Run:

globato cudem build -R 178.25/178.65/-18.30/-17.95 -E 3s -O fiji_test -P epsg:4326+3855 -M ms_binary_cudem:algos=interp_gmt -X 0:5 global-bathy-topo --export

This command defines a Fiji coastal DEM workflow using the global-bathy-topo bundle and writes the generated workflow to a YAML recipe.

The --export option exports the recipe instead of running the DEM build.

Command Breakdown¶

Expected recipe output:

fiji_test_recipe.yaml

At this stage, the workflow has been exported but the final DEM has not yet been generated.

Why Export the Recipe?¶

The exported YAML file records the workflow decisions used to define the DEM.

This may include:

• Project name

• Region

• Target reference system

• Source modules

• Source weights

• Per-source hooks

• Global hooks

• Stacking settings

• Interpolation settings

• Output files

• Execution settings

This is what makes the workflow reproducible. The command defines the workflow, and the exported YAML preserves it in a form that can be inspected, modified, shared, version-controlled, and rerun.

Add Local Data with local_fs¶

The local_fs module searches a local directory for files that match a specified extension and adds them to the CUDEM workflow.

Example for local GeoTIFF files:

modules:
  - module: local_fs
    args:
      path: /path/to/local/data
      ext: .tif
      datatype: raster
      weight: 1.0
      uncertainty: 0.0
      use_cache: true

Example for local XYZ point data:

modules:
  - module: local_fs
    args:
      path: /path/to/local/xyz
      ext: .xyz
      datatype: xyz
      weight: 1.0
      use_cache: false

Hooks can be added to describe and prepare the local dataset before it enters the larger workflow:

modules:
  - module: local_fs
    args:
      path: /path/to/local/xyz
      ext: .xyz
      datatype: xyz
      weight: 1.0
      use_cache: false
    hooks:
      - name: set-srs
        args:
          srs: EPSG:4326+3855
      - name: spatial_crop
      - name: range_z
        args:
          max_z: -0.1

This allows local lidar, multibeam, chart soundings, XYZ files, rasters, or other project-specific datasets to be combined with the remote source modules already included in the exported recipe.

General pattern:

export recipe
→ add local_fs module
→ define local path and file type
→ assign source reference information
→ add optional filtering hooks
→ run the modified recipe

Run a Saved YAML Recipe¶

Once a recipe has been exported, use globato cudem run to execute it.

General pattern:

globato cudem run RECIPE.yaml

Fiji example:

globato cudem run fiji_test_recipe.yaml

This command executes the saved workflow and generates the DEM and associated outputs.

Expected final DEM:

fiji_test_final.tif

The recipe can now be rerun without reconstructing the original command.

Override Recipe Settings¶

Parts of the saved recipe may be overridden from the command line.

Override the region:

globato cudem run fiji_test_recipe.yaml -R WEST/EAST/SOUTH/NORTH

Override the resolution:

globato cudem run fiji_test_recipe.yaml -E 1s

Override the output name:

globato cudem run fiji_test_recipe.yaml -O fiji_test_v2

This makes a saved recipe reusable across different regions, resolutions, and output configurations.

Inspect the Output¶

After the recipe finishes, inspect the DEM metadata:

gdalinfo fiji_test_final.tif

Useful information in the gdalinfo output includes:

• Driver

• Raster size

• Coordinate system

• Origin

• Pixel size

• Corner coordinates

• Band information

• NoData value

• Minimum and maximum values

Create a Hillshade¶

Create a quick visual-inspection product using globato perspecto:

globato perspecto hillshade fiji_test_final.tif fiji_test_hillshade.tif

This creates a georeferenced hillshade that can be used for visual inspection without opening an external GIS application.

Optional hillshade settings:

globato perspecto hillshade fiji_test_final.tif fiji_test_hillshade.tif --cmap etopo --exag 3 --blend soft_light

What Does -X 0:5 Do?¶

The -X option extends the processing region before gridding.

In this workshop, we use:

-X 0:5

This adds a 5 percent processing buffer around the specified output region.

For coastal DEM generation, the buffer can help reduce edge effects by including additional source data beyond the final DEM boundary. These surrounding data provide better support for interpolation near the edge of the requested output area.

The processing region is extended, but the final DEM remains associated with the requested output extent.

Main Takeaway¶

globato cudem build defines a DEM workflow from command-line options.

globato cudem build --export writes that workflow to a reusable YAML recipe instead of running the DEM build.

globato cudem run executes a saved YAML recipe and generates the DEM outputs.

Together, these commands support a reproducible workflow:

choose a region
→ select a source bundle
→ export the recipe
→ inspect the YAML
→ modify the workflow
→ run the recipe
→ inspect the outputs

globato is the CUDEM workflow engine that brings source modules, transformations, filtering, stacking, interpolation, and output generation together in one inspectable, recipe-based process.