A_fake(temp, noise=false)

Fake law establishing a theoretical relationship between ice viscosity (A) and long-term air temperature.

SIA!(dH, H, context)

Compute an in-place step of the Shallow Ice Approximation PDE in a forward model

SIA(H, A, context)

Compute a step of the Shallow Ice Approximation UDE in a forward model. Allocates memory.

avg(A)

4-point average of a matrix

avg_surface_V(H, B, temp)

Computes the average ice velocity for a given input temperature

avg_x(A)

2-point average of a matrix's X axis

avg_y(A)

2-point average of a matrix's Y axis

batch_iceflow_PDE(climate, gdir, context)

Solve the Shallow Ice Approximation iceflow PDE for a given temperature series batch

batch_iceflow_UDE(θ, H, climate, context)

Solve the Shallow Ice Approximation iceflow UDE for a given temperature series batch

diff_x(A)

2-point differential of a matrix's X axis

diff_y(A)

2-point differential of a matrix's Y axis

fillNaN!(x, fill)

Convert empty matrix grid cells into fill value

generate_ref_dataset(temp_series, H₀)

Generate reference dataset based on the iceflow PDE

get_NN()

Generates a neural network.

get_initial_geometry(glacier_gd)

Retrieves the initial glacier geometry (bedrock + ice thickness) for a glacier.

iceflow!(dH, H, context,t)

Runs a single time step of the iceflow PDE model in-place

iceflow_NN(H, θ, t, context, temps, UA)

Runs a single time step of the iceflow UDE model

Initialize_ODINN(processes, python_path)

Initializes ODINN by configuring PyCall based on a given Python path. It also configures multiprocessing for a given number of processes.

inn(A)

Access the inner part of the matrix (-2,-2)

loss_iceflow(θ, context, UA, PDE_refs::Dict{String, Any}, temp_series)

Loss function based on glacier ice velocities and/or ice thickness

predict_A̅(UA, θ, temp)

Predicts the value of A with a neural network based on the long-term air temperature.

predict_iceflow(θ, UA, context, temp_series)

Makes a prediction of glacier evolution with the UDE for a given temperature series

retrieve_context(context::ArrayPartition)

Retrieves context variables for computing the surface velocities of the PDE.

retrieve_context(context::Tuple)

Retrieves context variables for computing the surface velocities of the UDE.

simulate_iceflow_PDE(H, context, solver)

Make forward simulation of the SIA PDE.

smooth!(A)

Smooth data contained in a matrix with one time step (CFL) of diffusion.

train_iceflow_UDE(H₀, UA, θ, train_settings, PDE_refs, temp_series)

Train the Shallow Ice Approximation iceflow UDE