Pith. sign in

REVIEW 3 major objections 4 minor 38 references

Convergence Tests for Transdimensional Markov Chains in Geoscience Imaging

T0 review · 3 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Mapping variable-size MCMC models to fixed-dimension scalars or grids lets standard convergence tests run unchanged.

desk verdict Useful practical extension of rjMCMC convergence diagnostics, but the central evidence is weakened by an unaddressed feature-ordering dependence and lack of code/data. read the letter →

arxiv 1908.11652 v1 pith:UPSPUJQZ submitted 2019-08-30 physics.geo-ph physics.comp-ph

Márk Somogyvári , Sebastian Reich This is my paper · ORCID
classification physics.geo-phphysics.comp-ph MSC 65C0562F1586A22
keywords transdimensionalinversionreversible-jumpMCMCconvergenceassessmentdiscretefracturenetworksGewekediagnosticGelman-Rubingeologicalprojectionscalarmodelindicator
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Reversible-jump MCMC produces ensembles in which the number of model parameters changes from sample to sample, so classic convergence diagnostics, which track fixed parameters across a chain, cannot be applied directly. The paper shows that three fixed-dimension conversions of each model realization, a scalar model indicator, a per-parameter-type vector of such indicators, and a rasterized geological projection, make the resulting sequences analyzable by autocorrelation, Geweke, and Gelman-Rubin diagnostics without any modification. On a simple synthetic fracture-network problem and a field-based tracer-tomography example, the modified scalar-vector conversion and the geological projection give similar diagnostic curves. In the field case both indicate convergence roughly between 30,000 and 40,000 iterations, meaning the previously used 50 percent burn-in discards more of the chain than needed. The practical payoff is that transdimensional inversions could be stopped earlier or retain longer usable ensembles.

What carries the argument

The load-bearing object is the scalar model indicator $\nu(\theta_i)=1+\sum_{j=1}^{k} \theta_i(j)d^{j-1}$, which encodes a discretized $k$-parameter model as a single integer in $\{1,\dots,d^k\}$, where $d$ is the cardinality of each parameter's value set. Since a single scalar becomes dominated by whichever parameter type has larger numerical range, the paper modifies it into a vector whose components are computed separately for each parameter type, giving a fixed-size sequence for every model in the chain. The geological projection, by contrast, rasterizes each model onto a common grid, producing a tensor. Each conversion reduces the transdimensional chain to a fixed-dimension series, which is exactly what allows the unmodified autocorrelation, Geweke, and Gelman-Rubin diagnostics to operate.

What would settle it

Re-run the field-based diagnostic analysis after permuting the order in which fractures are listed in each realization before applying the scalar indicator formula; if the resulting Geweke Z-scores cross the $\pm 1.96$ band or the Gelman-Rubin PSRF crosses the 1.1 cutoff at different iterations, the claimed order-independence of the scalar-vector conversion fails. This is a direct computation on the same chains, requiring no new data.

Watch

Extended reading notes

Core claim

The paper's central claim is that transdimensional Markov chains can be audited by ordinary convergence-assessment tools after a fixed-dimensional replacement of each model realization. For discretized variable-selection problems, the scalar model indicator is a bijection mapping each model to an integer; for feature-based models such as fracture networks, one scalar per parameter type forms a fixed-size vector; for imaging problems, rasterization onto a common grid yields a geological projection. Applied to a simple synthetic DFN model and to a field-based DFN ensemble from tracer tomography, the modified scalar indicator and the geological projection produce Geweke and Gelman-Rubin curves with the same qualitative features, and both place convergence of the field-based chains at 30,000 to 40,000 iterations, earlier than the 50 percent burn-in used previously. The paper thus establishes that appropriately chosen scalar conversions can retain the statistical characteristics of the ensemble without the cost of rasterization.

Load-bearing premise

The scalar indicator is a unique code only when the model's parameters are listed in a fixed order; in feature-based fracture models the order of features is arbitrary, and the paper does not test whether reordering the features changes the convergence curves.

Editorial extensions

If this is right

  • rjMCMC ensembles in geoscience imaging can be monitored with standard diagnostic software, with no change to the sampling algorithm.
  • The modified scalar model indicator vector reproduces the convergence messages of the rasterized geological projection at a fraction of the computational and storage cost.
  • For the field-based fracture-network ensemble, convergence occurs by iteration 30,000 to 40,000, so the common practice of discarding the first 50 percent of the chain can be relaxed, retaining more samples.
  • Where feature counts are large enough to overflow floating-point scalars, the geological projection remains a viable fallback for convergence assessment.
  • The evolution of model dimensionality can itself be treated as a scalar chain and assessed with the same diagnostics.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A permutation-invariant version of the scalar indicator, for instance sorting features by location or using a multiset summary, would remove the ordering sensitivity identified in the weakest-assumption note and could be tested on the same chains.
  • The same conversion strategy should carry over to other transdimensional samplers, such as birth-death point-process MCMC, where parameter count varies but models are naturally ordered by feature coordinates.
  • Because the scalar conversion is cheap, it could be computed online during sampling to stop runs adaptively once the diagnostic curves stabilize, rather than fixing chain length in advance.
  • The paper's comparison suggests that when only convergence timing is needed, scalar fingerprints may serve as a proxy even for continuum models, where rasterization is the conventional choice.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper addresses convergence assessment for reversible-jump Markov Chain Monte Carlo (rjMCMC) ensembles whose models vary in dimensionality. It proposes three conversions of each transdimensional model to a common-dimensional representation: the scalar model indicator of Brooks et al. (2003), a modified scalar model indicator vector computed separately for each parameter type, and a geological projection obtained by rasterizing the model onto a fixed grid. The conversions are applied to two discrete fracture network (DFN) inverse problems, a simple synthetic grid-based case and a field-based tracer tomography case from Somogyvári et al. (2017). Three standard diagnostics—autocorrelation, Geweke's Z, and the Gelman-Rubin PSRF—are applied to the converted chains. The main claim is that the modified scalar model indicator and the geological projection yield similar diagnostic curves, indicating convergence after 30,000–40,000 iterations in the field case, earlier than the 50% burn-in previously used.

Significance. If the central claim is robust, the paper offers a practical and low-cost route to applying classical convergence diagnostics to transdimensional geoscience inversions, a real gap in the rjMCMC literature. The idea of converting each parameter type separately and comparing against rasterized projections is intuitive and likely useful to practitioners. The paper does not ship machine-checked proofs or reproducible code, but all diagnostics are standard and the algorithmic steps are described clearly enough to reimplement. The significance is currently conditional: the key scalar-versus-projection agreement rests on visual comparison, the scalar conversion has an unresolved ordering sensitivity for feature-based models, and the PSRF does not reach the stated cutoff in the field example.

major comments (3)
  1. [Section 3, Eq. (3)] The scalar model indicator in Eq. (3) is defined for an ordered, discretized model vector, and the paper adapts it to feature-based DFN models by computing one scalar per parameter type. However, a DFN realization is a set of fractures, and the posterior distribution is invariant under permutation of the fracture list, while the scalar indicator is not: reordering the same set of fractures changes the value of ν(θ_i). The geological projection, by contrast, is invariant under feature reordering because it is defined on a fixed raster grid. The paper's central evidence—that the modified scalar indicator and the geological projection give similar Geweke and Gelman-Rubin curves—therefore compares an order-dependent summary with an order-invariant one. A permutation test is needed: recompute the modified scalar indicator for several random orderings (or a canonical ordering) of the fracture list in each realization and demonstrate that the autocorrelation, Geweke Z, and PSRF curves are unchanged. Without this test, the claimed convergence point of 30,000–40,000 iterations and the broader recommendation to use scalar conversions may not be robust.
  2. [Section 6, Figure 7 (e,f)] The paper states that 'both curves saturate at values higher than one' for the Gelman-Rubin diagnostics, meaning the PSRF never reaches the 1.1 cutoff adopted in Section 5.3. The conclusion that the chains converge after 30,000–40,000 iterations is instead drawn by visually comparing the Geweke plots and the trends of the PSRF curves. This is a weaker claim than 'could be analyzed using standard convergence assessment techniques, without modification,' because in the field example a standard absolute PSRF cutoff is not satisfied for either conversion. The authors should either state explicitly that a modified, trend-based criterion is being used and define it, or provide a quantitative rule (for example, stabilization of the PSRF curve within a tolerance) so that the claimed convergence point is reproducible.
  3. [Figures 4, 5, 7] The core claim that the modified scalar model indicator and the geological projection give comparable diagnostics is supported only by visual inspection of plotted curves. The text asserts that the Geweke Z values are 'very close' (Section 5.2) and that the diagnostics are 'similar' (Section 6), but no quantitative similarity measure is reported. Because the central conclusion depends on this agreement, the authors should report a numerical comparison, such as the correlation or mean absolute difference between the Geweke Z sequences after aligning them on the same iteration index, or an analogous summary for the PSRF curves. This would also make it possible to judge whether the disagreements at specific iteration ranges (e.g., the sub-10,000 and 30,000–40,000 outliers in Figure 7 c,d) are within an acceptable tolerance.
minor comments (4)
  1. [Section 3, Eq. (3)] The exponent in Eq. (3) is rendered as 'd j−1' in the text; it should be d^{j−1} to match the bijection definition in Brooks et al. (2003).
  2. [Section 4] The phrase 'which is the table with the number and location of the fractures' is unclear; it presumably means the row and column sums are stored as observation vectors, but the wording should be revised.
  3. [Section 6] The paragraph explaining why the classic scalar model indicator is not computed for the field case is placed before the discussion of the modified indicator and projection; moving this explanation to the description of the field setup would improve readability.
  4. [Section 5.2] The statement that the Geweke Z values of the modified scalar indicator and the projection are 'very close' while 'the signs may differ' is puzzling: if signs differ, the curves are not as close as implied. Clarify whether sign differences occur only in regions where |Z| is small.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the conversions are explicit, the diagnostics are applied post hoc, and the scalar/projection agreement is empirical.

full rationale

No load-bearing step in this paper reduces to its own inputs. The scalar model indicator (Eq. 3) is taken from Brooks et al. (2003) and is a stated bijection independent of the convergence diagnostics; the modified scalar indicator vector is obtained by computing that same scalar separately for each parameter type, and the geological projection is produced by rasterization. Neither representation is fitted or calibrated to make the Geweke, Gelman-Rubin, or autocorrelation curves agree. The agreement between the modified scalar indicator and the geological projection in Sections 5 and 6 is reported as an empirical observation, not as a consequence of any fitted parameter. Citations to Somogyvári et al. (2017) supply the rjMCMC sampler, the field-based DFN data, and the previously used 50% burn-in convention, but they are not invoked as evidence that the new conversion methods work. The paper explicitly notes its scope limitation regarding assessment of whether the sampler captures the full posterior shape (Section 7), which is a stated incompleteness rather than a hidden circularity. The skeptic's concern that Eq. (3) depends on the arbitrary ordering of features in a DFN realization is a legitimate robustness question about the scalar encoding, but it is not a circularity: the ordering is not defined in terms of the diagnostic results, and the paper's central comparison remains an empirical claim that could in principle fail under reordering.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The central claim rests on borrowed conversions (Brooks et al. 2003 scalar indicator, Bodin and Sambridge 2009 rasterization) and on the assumption that classic diagnostics applied to converted chains reflect convergence of the original transdimensional chain. The only ad hoc addition is the per-parameter-type scalar vector, whose statistical equivalence to the ensemble is asserted from qualitative plot agreement rather than proven.

free parameters (4)
  • Rasterization grid resolution
    The geological projection requires choosing a grid; the paper does not state cell size or study sensitivity to it. This choice affects the vector of pixel values and hence the diagnostics, but it is not fitted to the data.
  • Number of intervals in Geweke test = 100
    The first half of the chain is split into 100 intervals; this is a user choice and can affect which outliers are detected.
  • PSRF cutoff = 1.1
    Used as convergence threshold following common practice; affects interpretation but not the conversion itself.
  • Thinning factor for field data = 100
    Every 100th member is kept to reduce autocorrelation, a user choice that changes the effective chain length.
assumptions (6)
  • standard math The rjMCMC chain is irreducible, aperiodic, and positive recurrent, so it converges to a stationary distribution.
    Assumed by the convergence theory in Section 2 and required for the diagnostics to be meaningful.
  • domain assumption The parameter space in the applications is fully discretized, so the Jacobian in the MHG acceptance ratio is 1.
    Stated in Section 2: 'In this paper problems with fully discretized parameter space are presented'.
  • standard math The scalar model indicator (Eq. 3) is a bijection between model states and integers for ordered, discretized parameter spaces.
    Taken from Brooks et al. (2003); the paper relies on this to map each model to a unique scalar.
  • domain assumption Classic convergence diagnostics remain valid when applied to the converted scalar/vector/projection chains.
    This is the central enabling assumption of the paper; it is demonstrated by example, not proven.
  • ad hoc to paper The modified scalar model indicator vector preserves the statistical properties of the transdimensional ensemble.
    This is the paper's key claim for the vector conversion; it is supported only by qualitative visual agreement with projection-based diagnostics.
  • domain assumption The geological projection via rasterization adequately represents the model for the purpose of convergence assessment.
    Rasterization is assumed to capture the relevant statistical behavior; large zero-filled regions damp variances as noted in Section 6.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Convergence Tests for Transdimensional Markov Chains in Geoscience Imaging." pith.science (2026). https://pith.science/paper/UPSPUJQZ

@misc{pith2026190811652,
  author       = {Pith},
  title        = {Pith review of: Convergence Tests for Transdimensional Markov Chains in Geoscience Imaging},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UPSPUJQZ}},
  note         = {Machine review of arXiv:1908.11652}
}
read the original abstract

Classic inversion methods adjust a model with a predefined number of parameters to the observed data. With transdimensional inversion algorithms such as the reversible-jump Markov Chain Monte Carlo (rjMCMC), it is possible to vary this number during the inversion and to interpret the observations in a more flexible way. Geoscience imaging applications use this behaviour to automatically adjust model resolution to the inhomogeneities of the investigated system, while keeping the model parameters on an optimal level. The rjMCMC algorithm produces an ensemble as result, a set of model realizations which together represent the posterior probability distribution of the investigated problem. The realizations are evolved via sequential updates from a randomly chosen initial solution, and converge toward the target posterior distribution of the inverse problem. Up to a point in the chain, the realizations may be strongly biased by the initial model, and have to be discarded from the final ensemble. With convergence assessment techniques, this point in the chain can be identified. Transdimensional MCMC methods produce ensembles which are not suitable for classic convergence assessment techniques because of the changes in parameter numbers. To overcome this hurdle, three solutions are introduced to convert model realizations to a common dimensionality while maintaining the statistical characteristics of the ensemble. A scalar, a vector and a matrix representation for models is presented, inferred from tomographic subsurface investigations, and three classic convergence assessment techniques are applied on them. It is shown that appropriately chosen scalar conversions of the models could retain similar statistical ensemble properties as geologic projections created by rasterization.

Figures

Figures reproduced from arXiv: 1908.11652 by the authors.

Figure 7
Figure 7. For the autocorrelation plots, we show 5 independent chains, for Geweke [PITH_FULL_IMAGE:figures/full_fig_p014_7.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 25 canonical work pages

  1. [1]

    Journal of Geophysical Research: Solid Earth 123(11):9324--9340, doi:10.1029/2018JB016438

    Afshari Moein MJ, Somogyv \' a ri M, Valley B, Jalali M, Loew S, Bayer P (2018) Fracture Network Characterization Using Stress-Based Tomography . Journal of Geophysical Research: Solid Earth 123(11):9324--9340, doi:10.1029/2018JB016438

    work page doi:10.1029/2018jb016438 2018

  2. [2]

    Geophysical Journal International 181(2):858--872, doi:10.1111/j.1365-246X.2010.04530.x

    Agostinetti NP, Malinverno A (2010) Receiver function inversion by trans-dimensional Monte Carlo sampling . Geophysical Journal International 181(2):858--872, doi:10.1111/j.1365-246X.2010.04530.x

    arXiv 2010

  3. [3]

    Convergence diagnosis and output analysis software for Gibbs sampling output Manual Version 030 MRC Biostatistics Unit, Cambridge, UK

    Best NG, Cowles MK, Vines SK (1996) CODA . Convergence diagnosis and output analysis software for Gibbs sampling output Manual Version 030 MRC Biostatistics Unit, Cambridge, UK

    work page 1996

  4. [4]

    Experimental Observation of Vortex Rings in a Bulk Magnet

    Bodin T, Sambridge M (2009) Seismic tomography with the reversible jump algorithm . Geophysical Journal International 178(3):1411--1436, doi:10.1111/j.1365-246X.2009.04226.x

    work page Pith review arXiv 2009

  5. [5]

    Journal of Geophysical Research: Solid Earth 117(2):1--24, doi:10.1029/2011JB008560

    Bodin T, Sambridge M, Tkal \' c i \' c H, Arroucau P, Gallagher K, Rawlinson N (2012) Transdimensional inversion of receiver functions and surface wave dispersion . Journal of Geophysical Research: Solid Earth 117(2):1--24, doi:10.1029/2011JB008560

    doi:10.1029/2011jb008560 2012

  6. [6]

    Hydrogeology Journal 21(6):1265--1274, doi:10.1007/s10040-013-1006-z

    Brauchler R, B \" o hm G, Leven C, Dietrich P, Sauter M (2013) A laboratory study of tracer tomography . Hydrogeology Journal 21(6):1265--1274, doi:10.1007/s10040-013-1006-z

    work page doi:10.1007/s10040-013-1006-z 2013

  7. [8]

    Chapman & Hall/CRC Handbooks of Modern Statistical Methods, Chapman and Hall/CRC, doi:10.1201/b10905

    Brooks S, Gelman A, Jones G, Meng XL (eds) (2011) Handbook of Markov Chain Monte Carlo . Chapman & Hall/CRC Handbooks of Modern Statistical Methods, Chapman and Hall/CRC, doi:10.1201/b10905

    doi:10.1201/b10905 2011

  8. [9]

    Journal of Computational and Graphical Statistics 7(4):434--455, doi:10.1080/10618600.1998.10474787

    Brooks SP, Gelman A (1998) General Methods for Monitoring Convergence of Iterative Simulations . Journal of Computational and Graphical Statistics 7(4):434--455, doi:10.1080/10618600.1998.10474787

    arXiv 1998

Show all 38 references
  1. [10]

    In: Handbook of Geomathematics, Springer Berlin Heidelberg, Berlin, Heidelberg, pp 865--889, doi:10.1007/978-3-642-01546-5_29

    Caers J, Park K, Scheidt C (2010) Modeling Uncertainty of Complex Earth Systems in Metric Space . In: Handbook of Geomathematics, Springer Berlin Heidelberg, Berlin, Heidelberg, pp 865--889, doi:10.1007/978-3-642-01546-5_29

    2010 doi

  2. [11]

    Journal of the American Statistical Association 91(434):883--904, doi:10.1080/01621459.1996.10476956

    Cowles MK, Carlin BP (1996) Markov Chain Monte Carlo Convergence Diagnostics: A Comparative Review . Journal of the American Statistical Association 91(434):883--904, doi:10.1080/01621459.1996.10476956

    1996

  3. [12]

    Journal of Statistical Computation and Simulation 64(1):87--104, doi:10.1080/00949659908811968

    Cowles MK, Roberts GO, Rosenthal JS (1999) Possible biases induced by MCMC convergence diagnostics . Journal of Statistical Computation and Simulation 64(1):87--104, doi:10.1080/00949659908811968

    1999 doi

  4. [13]

    Computational Statistics and Data Analysis 38(4):475--485, doi:10.1016/S0167-9473(01)00073-1

    Denison DG, Adams NM, Holmes CC, Hand DJ (2002) Bayesian partition modelling . Computational Statistics and Data Analysis 38(4):475--485, doi:10.1016/S0167-9473(01)00073-1

    2002 doi

  5. [14]

    Journal of the Acoustical Society of America 132(4):2239--50, doi:10.1121/1.4746016

    Dettmer J, Dosso SE (2012) Trans-dimensional matched-field geoacoustic inversion with hierarchical error models and interacting Markov chains. Journal of the Acoustical Society of America 132(4):2239--50, doi:10.1121/1.4746016

    2012 doi

  6. [15]

    Advances in Water Resources 97:120--129, doi:10.1016/j.advwatres.2016.09.004

    Djibrilla Saley A, Jardani A, Soueid Ahmed A, Raphael A, Dupont J (2016) Hamiltonian Monte Carlo algorithm for the characterization of hydraulic conductivity from the heat tracing data . Advances in Water Resources 97:120--129, doi:10.1016/j.advwatres.2016.09.004

    2016 doi

  7. [16]

    Inverse Problems 30(11):114018, doi:10.1088/0266-5611/30/11/114018

    Dosso SE, Dettmer J, Steininger G, Holland CW (2014) Efficient trans-dimensional Bayesian inversion for geoacoustic profile estimation . Inverse Problems 30(11):114018, doi:10.1088/0266-5611/30/11/114018

    2014 doi

  8. [17]

    Mathematical Geosciences 44(5):515--543, doi:10.1007/s11004-012-9397-2

    Elsheikh AH, Jackson MD, Laforce TC (2012) Bayesian Reservoir History Matching Considering Model and Parameter Uncertainties . Mathematical Geosciences 44(5):515--543, doi:10.1007/s11004-012-9397-2

    2012 doi

  9. [18]

    Fonnesbeck CJ (2013) PyMC Documentation

    2013

  10. [19]

    Geomorphology 238:1--14, doi:10.1016/j.geomorph.2015.02.022

    Fox M, Bodin T, Shuster DL (2015) Abrupt changes in the rate of Andean Plateau uplift from reversible jump Markov Chain Monte Carlo inversion of river profiles . Geomorphology 238:1--14, doi:10.1016/j.geomorph.2015.02.022

    2015 doi

  11. [20]

    Statistical Science 7(4):457--472, doi:10.1214/ss/1177011136

    Gelman A, Rubin DB (1992) Inference from Iterative Simulation Using Multiple Sequences . Statistical Science 7(4):457--472, doi:10.1214/ss/1177011136

    1992

  12. [21]

    Chapman and Hall/CRC

    Gelman A, Stern HS, Carlin JB, Dunson DB, Vehtari A, Rubin DB (2013) Bayesian data analysis . Chapman and Hall/CRC

    2013

  13. [22]

    Bayesian Statistics 4 pp 169--193, doi:1176289

    Geweke J (1992) Evaluating the accuracy of sampling-based approaches to the calculation of posterior moments . Bayesian Statistics 4 pp 169--193, doi:1176289

    1992

  14. [23]

    Biometrika 82(4):711--732, doi:10.1093/biomet/82.4.711

    Green PJ (1995) Reversible jump Markov chain monte carlo computation and Bayesian model determination . Biometrika 82(4):711--732, doi:10.1093/biomet/82.4.711

    1995 doi

  15. [24]

    Journal of Hydrology 281(1-2):36--54, doi:10.1016/S0022-1694(03)00199-9

    Illman WA, Neuman SP (2003) Steady-state analysis of cross-hole pneumatic injection tests in unsaturated fractured tuff . Journal of Hydrology 281(1-2):36--54, doi:10.1016/S0022-1694(03)00199-9

    2003 doi

  16. [25]

    PhD thesis, doi:http://hdl.handle.net/10012/7642

    Jalali M (2013) Thermo-Hydro-Mechanical Behavior of Conductive Fractures using a Hybrid Finite Difference – Displacement Discontinuity Method . PhD thesis, doi:http://hdl.handle.net/10012/7642

    2013

  17. [26]

    Mathematical Geosciences 49(7):845--869, doi:10.1007/s11004-017-9693-y

    Jeong C, Mukerji T, Mariethoz G (2017) A Fast Approximation for Seismic Inverse Modeling: Adaptive Spatial Resampling . Mathematical Geosciences 49(7):845--869, doi:10.1007/s11004-017-9693-y

    2017 doi

  18. [27]

    Water Resources Research 52(5):3966--3983, doi:10.1002/2015WR017922

    Jim \' e nez S, Mariethoz G, Brauchler R, Bayer P (2016) Smart pilot points using reversible-jump Markov-chain Monte Carlo . Water Resources Research 52(5):3966--3983, doi:10.1002/2015WR017922

    2016 doi

  19. [28]

    J Am Stat Assoc 91(433):154--166

    Johnson VE (1996) Studying convergence of Markov Chain Monte Carlo algorithms using coupled sample paths . J Am Stat Assoc 91(433):154--166

    1996

  20. [29]

    Journal of the American Statistical Association 93(441):238--248, doi:10.1080/01621459.1998.10474105

    Johnson VE (1998) A coupling-regeneration scheme for diagnosing convergence in markov chain monte carlo algorithms . Journal of the American Statistical Association 93(441):238--248, doi:10.1080/01621459.1998.10474105

    1998

  21. [30]

    Methods in Ecology and Evolution 3(1):112--115, doi:10.1111/j.2041-210X.2011.00131.x

    Link WA, Eaton MJ (2012) On thinning of chains in MCMC . Methods in Ecology and Evolution 3(1):112--115, doi:10.1111/j.2041-210X.2011.00131.x

    2012 arXiv

  22. [31]

    Advances in Water Resources 33(3):241--256, doi:10.1016/j.advwatres.2009.10.010

    Mondal A, Efendiev Y, Mallick B, Datta-Gupta A (2010) Bayesian uncertainty quantification for flows in heterogeneous porous media using reversible jump Markov chain Monte Carlo methods . Advances in Water Resources 33(3):241--256, doi:10.1016/j.advwatres.2009.10.010

    2010 doi

  23. [32]

    Water Resources Research 54(5):3758--3766, doi:10.1029/2017WR022073, ://onlinelibrary.wiley.com/doi/abs/10.1029/2017WR022073

    O'Malley D, Karra S, Hyman JD, Viswanathan HS, Srinivasan G (2018) Efficient Monte Carlo With Graph-Based Subsurface Flow and Transport Models . Water Resources Research 54(5):3758--3766, doi:10.1029/2017WR022073, ://onlinelibrary.wiley.com/doi/abs/10.1029/2017WR022073

    2018 doi

  24. [33]

    Cambridge University Press, doi:10.1017/CBO9781107415324.004, arXiv:1011.1669v3

    Robert CP, Casella G (1998) Monte Carlo Statistical Methods , vol 53. Cambridge University Press, doi:10.1017/CBO9781107415324.004, arXiv:1011.1669v3

    1998 arXiv

  25. [34]

    John Wiley & Sons

    Sahimi M (2011) Flow and transport in porous media and fractured rock: from classical methods to modern approaches . John Wiley & Sons

    2011

  26. [35]

    Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 371(1984):20110547--20110547, doi:10.1098/rsta.2011.0547

    Sambridge M, Bodin T, Gallagher K, Tkalcic H (2012) Transdimensional inference in the geosciences . Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 371(1984):20110547--20110547, doi:10.1098/rsta.2011.0547

    2012

  27. [36]

    Geophysical Journal International 204(2):918--931, doi:10.1093/gji/ggv466

    Saygin E, Cummins PR, Cipta A, Hawkins R, Pandhu R, Murjaya J, Masturyono, Irsyam M, Widiyantoro S, Kennett BLN (2016) Imaging architecture of the Jakarta Basin, Indonesia with transdimensional inversion of seismic noise . Geophysical Journal International 204(2):918--931, doi...

    2016 doi

  28. [37]

    Water Resources Research 53(6):5104--5123, doi:10.1002/2016WR020293

    Somogyv \' a ri M, Jalali M, Jimenez Parras S, Bayer P (2017) Synthetic fracture network characterization with transdimensional inversion . Water Resources Research 53(6):5104--5123, doi:10.1002/2016WR020293

    2017 doi

  29. [38]

    Advances in Water Resources 32(3):391--400, doi:10.1016/j.advwatres.2008.12.001

    Zhu J, Cai X, Jim Yeh TC (2009) Analysis of tracer tomography using temporal moments of tracer breakthrough curves . Advances in Water Resources 32(3):391--400, doi:10.1016/j.advwatres.2008.12.001

    2009 doi

  30. [39]

    Tectonophysics 626(1):1--20, doi:10.1016/j.tecto.2014.03.017

    Ziegler M, Loew S, Bahat D (2014) Growth of exfoliation joints and near-surface stress orientations inferred from fractographic markings observed in the upper Aar valley (Swiss Alps) . Tectonophysics 626(1):1--20, doi:10.1016/j.tecto.2014.03.017

    2014 doi

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.