SIPPI (v1.5 - Aug 19, 2016 )

Thomas Mejer Hansen

Knud Skou Cordua

Aug 19, 2016


Table of Contents

About
1. Installation
Install SIPPI
Install SIPPI manually from github
SGeMS (optional)
2. Setting up SIPPI
prior: The a priori model
Types of a priori models
Sampling the prior
Sequential Gibbs sampling / Conditional Re-sampling
data: Data and data uncertainties/noise
Gaussian measurement noise
Gaussian modeling error
forward: The forward model
Validating prior, data, and forward
3. Sampling the a posteriori distribution
The rejection sampler
The extended Metropolis sampler
Controlling the step length
The independent extended Metropolis sampler
Annealing schedule
Parallel tempering
4. Examples
Examples of A priori models
Multiple 1D Gaussian prior model
Multivariate Gaussian prior with unknown covariance model properties.
Polynomial line fitting
The forward problem
Reference data, data, forward
The prior model
Setup and run the Metropolis sampler
Setup and run the rejection sampler
Cross hole tomography
Reference data set from Arrenęs
Travel delay computation: The forward problem
AM13 Gaussian: Inversion of cross hole GPR data from Arrenaes data with a Gaussian type a priori model
AM13 Gaussian, accounting for modeling errors
AM13 Gaussian with bimodal velocity distribution
AM13 Gaussian with unknown Gaussian model parameters
Probilistic covariance/semivariogram indeference
Specification of covariance model parameters
Inferring a 2D covariance model from the Jura data set - Example of point support
Bibliography
5. Reference
SIPPI
sippi_adjust_step_size
sippi_anneal_temperature
sippi_compute_acceptance_rate
sippi_compute_modelization_forward_error
sippi_forward
sippi_forward_linear
sippi_get_resim_data
sippi_get_sample
sippi_least_squares
sippi_likelihood
sippi_mcmc_init
sippi_metropolis
sippi_prior
sippi_prior_cholesky
sippi_prior_dsim
sippi_prior_init
sippi_prior_mps
sippi_prior_plurigaussian
sippi_prior_set_steplength
sippi_prior_sisim
sippi_prior_snesim
sippi_prior_snesim_std
sippi_prior_uniform
sippi_prior_visim
sippi_prior_voronoi
sippi_rejection
sippi_set_path
SIPPI plotting commands
sippi_colormap
sippi_get_posterior_data
sippi_plot_current_model
sippi_plot_data
sippi_plot_defaults
sippi_plot_loglikelihood
sippi_plot_model
sippi_plot_movie
sippi_plot_posterior
sippi_plot_posterior_2d_marg
sippi_plot_posterior_data
sippi_plot_posterior_loglikelihood
sippi_plot_posterior_sample
sippi_plot_prior
sippi_plot_prior_movie
sippi_plot_prior_sample
sippi_plot_set_axis
wiggle
SIPPI toolbox: Traveltime tomography
calc_Cd
eikonal
eikonal_raylength
eikonal_traveltime
kernel_buursink_2d
kernel_finite_2d
kernel_fresnel_2d
kernel_fresnel_monochrome_2d
kernel_multiple
kernel_slowness_to_velocity
mspectrum
munk_fresnel_2d
munk_fresnel_3d
pick_first_arrival
plot_traveltime_sr
sippi_forward_traveltime
sippi_forward_traveltime_example
tomography_kernel
SIPPI toolbox: 2D Full Waveform GPR modeling
FDTD_fwi
LoadTraceM
ant_geometry
disc_freq
distribute_to_cores
freq_int_diff_1D
fwi_ant_geometry
fwi_execute
fwi_execute_parfor
image_snapshots
iseven
load2Dbin
load_wavelet
resize
save1Dbin
save2Dbin
save_wavelet
setup_input_parameters
shift_signal
sippi_forward_gpr_fd
sippi_gpr_fd_example
sippi_plot_data_gpr
velocity_to_eps
write_parameters_to_screen
SIPPI toolbox: Covariance inference
sippi_forward_covariance_inference

List of Figures

4.1. Location of boreholes AM1, AM2, AM3, and AM4 at Arrenęs.
4.2. Ray coverage between wells left) AM1-AM3, middle) AM2-AM4, right) AM1-4.
4.3. AM13: One sample (15 realizations) of the prior (Gaussian) model.
4.4. AM13: Data response from one realization of the prior.
4.5. 5 realizations from a FFTMA prior model type with top) Gaussian and b) Bimodal distribution
4.6. Distribution of one realization using a Gaussian Bimodal target distribution
4.7. A sample from a FFTMA type prior model with varying range_1, range_2, and ang_1.
4.8. Distribution of one sample of a 1D Gaussian distribution describing range_1, range_2, and ang_1