Professional Documents
Culture Documents
Inversion Assgn 4 PDF
Inversion Assgn 4 PDF
in the geosciences
Malcolm Sambridge
Research School of Earth Sciences,
Australian National University
Canberra, ACT, Australia.
A 3-D image of the Earth’s mantle using multi-frequency seismic shear waves
−20˚ −20˚
−30˚ −30˚
−40˚ −40˚
−40˚ −40˚
140˚ 150˚
3
Irregular parametrizations
Static optimized meshes in regional and global studies have become popular
Chou & Booker (1979); Tarantola & Nercessian (1984); Abers & Rocker (1991); Fukao et al. (1992); Zelt & Smith (1992); Michelini (1995); Vesnaver
(1996); Curtis & Snieder (1997); Widiyantoro & van der Hilst (1998); Bijwaard et al. (1998); Bohm et al. (2000); Sambridge & Faletic (2003).
4
A probabilistic framework
Posterior PDF Likelihood
mT = (m1 , . . . , mk )
k
!
m(x) = mj φj (x)
j=1
5
Sampling and optimization
Maximum
Markov chain
Parameter Monte Carlo sampling
space
Gradient optimization
6
Transdimensional inversion
'Withfour parameters I can fit an elephant and
with five I can make him wiggle his trunkʼ
J. von Neumann
k
!
m(x) = mj ψj (x) Uncertain parametrization
j=1
! N
#
1 1 " |d − g(m)|2
L(d|m) = √ exp − Uncertain Likelihood
(σ 2π)N 2 i σ2
Hierarchical Bayes
M-H ! "
p(m2 |d)q(m1 |m2 )
Acceptance α=1∧ Example
Rejection
Algorithm
Issues:
Efficiency M-H
Dimension Algorithm
8
Transdimensional inversion
Variable dimension
! "
p(d|m2 , k2 )p(m2 |k2 )p(k2 )q(m1 , k1 |m2 , k2 )
α=1∧ |J|
p(d|m1 , k1 )p(m1 |k1 )p(k1 )q(m2 , k2 |m1 , k1 )
k2 = k1 ± 1 k
!
m(x) = mj φj (x)
j=1
9
Example regression
' Make everything as simple as possible, but not simplerʼ
Einstein
Birth-Death example
60
1-D Regression Data
40 Partition modelling
20
!60
0 2 4 6 8 10
Properties of the ensemble x
may be examined
True
Uncertainty in
Ensemble
Average
model dimension
Data
10
Inversion for 1-D earth models
Using Airborne EM to constrain
subsurface conductivity mT = (c, x, k)T
Conductivity
0.01 0.1
0
10
20
30
Depth (m)
40
50
Uncertainty in
60
discontinuity position
70
80
Marginals in interface
position
o o o
9 S 9 S 9 S
o o o
18 S 18 S 18 S
o o o
27 S 27 S 27 S
o
36 S o
36 S
o
36 S
o
o o
120 E 130 E 140oE 150 oE o o
45 S110o E o o
120 E 130 E 140oE 150 oE o o
120 E 130 E 140oE 150 oE
45 S110o E 45 S110o E
o
0
o
9 S
Birth-Death McMC with unknowns mT = (f , x, k)T
o
18 S
o
36 S
o o o
130 E 140oE
Effective mesh resolution increases with number of
120 E 150 oE
20 Voronoi meshes
12
Classic tomography:
single models with a fixed grid
Damping
Smoothing
!40˚ !40˚ !40˚ !40˚
−20˚ −20˚
−40˚ −40˚
50 x 35
!40˚ !40˚
B-splines nodes
!40˚ !40˚
−20˚ −20˚ −20˚ −20˚ 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4
−30˚ −30˚
min ||δd − Gδm|| + α||m − mo ||2 + β||∇m||2 2
m
−30˚ −30˚
13
After Bodin et al. (2009)
−40˚ −40˚ −40˚ −40˚
Ensemble mean
1750 cells + regularization 62
Number of cells
110˚
120˚ 130˚ 140˚ 150˚
110˚
120˚ 130˚ 140˚ 150˚
60
vs
Number of Cells
!10˚ !10˚ !10˚ !10˚ 58
iterations
56
54
50
48
0 0.5 1 1.5 2 2.5 3 3.5 4
!30˚ !30˚ !30˚ !30˚
Iterations 5
x 10
110˚
120˚ 130˚ 140˚ 150˚
110˚
120˚ 130˚ 140˚ 150˚
2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4
−10˚ −10˚ −10˚ −10˚
0 10 20 30 40 50 60 70 80
−20˚ −20˚ −20˚ −20˚
Number of cells
14
−30˚ −30˚ −30˚ −30˚
True Data
Uncertainty estimates
for free
True Data
Ambient noise imaging
Transdimensional Hierarchical Bayes
mT = (v, λi , x, k)T
2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4
p(n|d)
0 !20˚ !20˚
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Number of Cells
1 1 1
120˚ 130˚ 140˚ 150˚
110˚
−10˚ −10˚
p(!|d)
p(a|d)
p(b|d)
!30˚ !30˚
0 0 0
−20˚ −20˚ 0.2 0.3 0.4 0.6 0.8 1 1.2 0 0.5 1 1.5
! a b
110˚
120˚ 130˚ 140˚ 150˚
Velocity cell density Output: Error in the ensemble mean
120˚ 130˚ 140˚ 150˚
110˚ 130˚ 140˚
120˚ 150˚
140˚ 150˚ !10˚ !10˚ 110˚
!10˚ !10˚
!20˚ !20˚
!20˚ !20˚
!30˚ !30˚
!30˚ !30˚
−40˚ −40˚
!40˚ !40˚
!40˚ !40˚
140˚ 150˚
17
Hierarchical Bayes
! N
#
1 1 " |d − g(m)|2
L(d|m) = √ exp − 2
(λσ 2π)N 2λ i σ2
!estimated = 10 !60
60
0 2 4
x 6 8 10
B)
40
B)
20 Model complexity OK 1
Estimated Model
0 Data fit good
complexity is accurate
!20
!40
0 2 4 6 8 10 0
0 10 20 30 40 50
!estimated = 30 Number of Partitions
60
C) C) 1
40 True
20
Estimated Data Estimated
Model too smooth noise is accurate
0 Data
Data under fit
!20
0
!40 5 10 15 20 25
0 2 4 6 8 10 !
18
Joint inversion of multiple
classes of data
Receiver function data
mT = (v, λi , x, k)T
38
.(*(+'("&2$0*,+#0&70#+6(8
.(*(+'("&/$0*,+#0
.(*(+'("&2$0*,+#0&70#&0#+6(8
=:;
9:< 1$"23*(&43'(&5+6%("6+#0&70#+6(8
1$"23*(&43'(&5+6%("6+#0&70#&0#+6(8
9:;
9; >; B; C; D;;
?("+#5&768
20
The end
21
Some papers on Trans-dimensional inversion
Electrical resistivity inversion
Malinverno, A., 2002. Parsimonious Bayesian Markov chain Monte Carlo inversion in a
107 nonlinear geophysical problem, Geophysical Journal International, 151(3), 675–688.
Malinverno, A. & Briggs, V., 2004. Expanded uncertainty quantification in inverse prob1029
lems: Hierarchical Bayes and empirical Bayes, Geophysics, 69, 1005.
Climate histories
Hopcroft, P., Gallagher, K., & Pain, C., 2007. Inference of past climate from borehole temperature data using
Bayesian Reversible Jump Markov chain Monte Carlo, Geophysical Journal International, 171(3), 1430–1439.
Hopcroft, P., Gallagher, K., & Pain, C., 2009. A Bayesian partition modelling approach to resolve spatial variability in
climate records from borehole temperature inversion, Geophysical Journal International, 178(2), 651–666.
Thermochronology
Gallagher, K., Charvin, K., Nielsen, S., Sambridge, M., & Stephenson, J., 2009. Markov chain Monte Carlo (MCMC)
sampling methods to determine optimal models, model resolution and model choice for Earth Science problems,
Marine and Petroleum Geology, 26(4), 525–535.
Stratigraphic modelling
Charvin, K., Gallagher, K., Hampson, G., & Labourdette, R., 2009a. A Bayesian approach to inverse modelling of
stratigraphy, part 1: method, Basin Research, 21(1), 5–25.
22
Some papers on Trans-dimensional inversion
Geochronlogy
Jasra, A., Stephens, D., Gallagher, K., & Holmes, C., 2006. Bayesian mixture modelling in geochronology via Markov
chain Monte Carlo, Mathematical Geology, 38(3), 269–300.
General
Gallagher, K., Bodin, T., Sambridge, M., Weiss, D., Kylander, M., & Large, D., 2011. Inference of abrupt changes in
noisy geochemical records using Bayesian Transdimensional changepoint models, Earth and Planetary Science
Letters, (October 2011) doi:10.1016/j.epsl.2011.09.015
Data inference in the 21st Century: Some ideas from outside the box,
Sambridge, M., Bodin, T., Reading, A., and Gallagher, K., 2010. Australian Soc. of Exploration Geophysics 21st International
Geophysics Conference and Exhibition, Extended Abstracts, 22-26 August, Sydney, Australia.