xrspatial.diffusion.diffuse#

xrspatial.diffusion.diffuse(agg, diffusivity=1.0, steps=1, dt=None, boundary='nan', name='diffuse')[source]#

Run forward-Euler diffusion on a 2D scalar field.

Solves du/dt = alpha * laplacian(u) using an explicit 5-point stencil for steps time steps.

Parameters:

  • agg (xarray.DataArray) – 2D raster representing the initial scalar field (temperature, concentration, etc.).

  • diffusivity (float or xarray.DataArray) – Thermal/mass diffusivity. A scalar applies uniformly; a DataArray of the same shape allows spatially varying diffusivity.

  • steps (int, default 1) – Number of time steps to run. Capped at 100,000 to prevent a single call from pinning a CPU indefinitely.

  • dt (float or None) – Time step size. When None the largest stable step is chosen automatically: dt = 0.25 * dx**2 / max(alpha).

  • boundary (str, default 'nan') – Edge handling: 'nan', 'nearest', 'reflect', or 'wrap'.

  • name (str, default 'diffuse') – Name for the output DataArray.

Returns:

Scalar field after steps diffusion steps.

Return type:

xarray.DataArray