Module `live_dashboard`

This IGM output module provides a real-time visualization dashboard for glacier simulations. It supports two rendering modes selectable via the mode parameter:

2D mode (matplotlib)

A three-panel planview dashboard updated live during the simulation:

Ice Thickness (left): ice thickness over a hillshaded bedrock, using the cividis colormap. The colorbar range is controlled by thk_max.
Velocity (right): vertically-averaged ice speed on a log scale over hillshaded bedrock, using the inferno colormap. The colorbar range is controlled by vel_max. An ice-edge contour is overlaid in blue.
Time series (bottom, full width): dual y-axis plot showing the Equilibrium Line Altitude (ELA, left axis) and ice volume (right axis) over time.

The current simulation time is displayed as a centered title above the panels. A status bar above the time series shows dt, volume, area, max velocity, and time.

ELA computation: If the smb_simple module is active (detected via state.smbpar), the full ELA forcing curve is shown as a dashed line from the start. Otherwise, the module computes a live ELA estimate at each save step as the mean surface elevation where \(|SMB| < 0.5\) m/yr and ice thickness \(> 1\) m.

The figure size adapts automatically to the domain's aspect ratio: the configured width (figsize[0]) is used, and the height is computed so that the two map panels fill their space without dead space.

If save_frames is set to true, PNG frames are saved at each update step (named dashboard_XXXXXX.png).

3D mode (PyVista)

A GPU-accelerated 3D visualization with:

Bedrock rendered as a semi-transparent grey surface.
Ice surface colored by speed using the inferno colormap, with a persistent horizontal colorbar at the bottom.
Time title displayed prominently at the top center.
Stats overlay (upper left) showing volume, area, max speed, dt, and wall time.
Inset chart (bottom left) showing the same ELA/volume time series as in 2D mode.

The initial camera is automatically positioned at 300% of the elevation range above the glacier, offset to the south and looking north toward the glacier center. The view is interactive and can be rotated/zoomed during the simulation.

Requirements

2D mode: matplotlib with TkAgg backend (interactive window).
3D mode: pyvista (GPU rendering via VTK).

Usage

Activate it in your params.yaml:

outputs:
  - live_dashboard

Parameters

Default configuration file (live_dashboard.yaml):

live_dashboard:
  mode: "2d"              # "2d" (matplotlib planview) or "3d" (PyVista GPU)
  update_freq: 1          # update every N save events (1 = every save)
  figsize: [14, 9]        # figure size in inches (2d mode only)
  dpi: 100                # figure resolution (2d mode only)
  vel_max: 500.0          # velocity colorbar max (m/a)
  thk_max: 800.0          # thickness colorbar max (m)
  save_frames: false      # also save PNG frames
  window_size: [1200, 800] # 3d mode window size

Description of the parameters:

Name	Description	Default value	Units
`mode`	Rendering mode: '2d' for matplotlib planview, '3d' for PyVista GPU.	2d	—
`update_freq`	Update the dashboard every N save events.	1	—
`figsize`	Figure width in inches (2D only); height is auto-computed from domain aspect ratio.	[14, 9]	inches
`dpi`	Figure resolution in dots per inch (2D only).	100	—
`vel_max`	Maximum value for the velocity colorbar.	500.0	m/a
`thk_max`	Maximum value for the thickness colorbar.	800.0	m
`save_frames`	Save PNG frames at each update step (2D only).	False	—
`window_size`	Window size in pixels (3D only).	[1200, 800]	pixels

Module live_dashboard