Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Ensemble Kalman Filter for Data Assimilation coupled with low-resolution computations techniques applied in Fluid Dynamics

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Running the Ensemble Kalman Filter on downsampled fluid fields, then rebuilding them with low-cost SVD, cuts computation time and memory by an order of magnitude while keeping errors close to the full-resolution baseline.

desk verdict New EnKF+lcSVD coupling with useful benchmarks, but the lcSVD reconstruction basis may be learned from the truth—the reported accuracy numbers are in-sample until the authors say otherwise. read the letter →

arxiv 2507.00539 v2 pith:AFS3VDZD submitted 2025-07-01 cs.CE physics.flu-dyn

Paul Jeanney , Ashton Hetherington , Shady E. Ahmed , David Lanceta , Susana Saiz , José Miguel Perez , Soledad Le Clainche This is my paper · ORCID
classification cs.CEphysics.flu-dyn
keywords ensembleKalmanfilterdataassimilationlow-resolutioncomputationslcSVDreduced-ordermodelcomputationalfluiddynamicsstatereconstruction
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

This paper claims that data assimilation with the Ensemble Kalman Filter (EnKF) can be run on spatially downsampled fluid fields and then rebuilt to full resolution with a low-cost Singular Value Decomposition (lcSVD), without losing the accuracy of the full-resolution computation. If true, this would let EnKF-based state estimation be applied to large-scale or real-time CFD problems where the full-resolution computation is too expensive. The paper demonstrates the idea on three flows: a laminar cylinder wake at Re=100, an experimental turbulent cylinder wake at Re=2600, and a turbulent jet LES. On the turbulent cylinder, the low-resolution route runs 13.7 times faster and uses 90.9% less RAM than the high-resolution reference, with a relative root-mean-square error of 2.6% versus 0.8% at 5% measurement noise.

What carries the argument

The mechanism that carries the argument is lcSVD, a low-cost variant of SVD for reconstructing a full-resolution snapshot matrix from a reduced one. The algorithm applies SVD to the downsampled snapshot matrix, re-orthonormalizes the modes with QR factorization, then recovers the full-resolution spatial modes and temporal coefficients from semi-reduced snapshots via Eqs. (3.10)-(3.12), and finally rebuilds the full dataset. This reconstruction step is what turns a cheap EnKF analysis on subsampled fields back into a high-resolution estimate, and it is the part of the pipeline that the paper claims is new to data assimilation.

What would settle it

Re-run the turbulent cylinder case at CR_ub=15.9 with the lcSVD reference matrix built only from the noisy background ensemble and the sparse observations, with no access to the True field; if the RRMSE rises well above 2.6%, the claimed accuracy is not an independent estimate of the True state.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central claim is that downsampling the background and observation fields, performing the EnKF analysis in that reduced space, and reconstructing the analyzed state with lcSVD yields an estimate of the True flow that is nearly as accurate as the high-resolution EnKF, while cutting computation time and memory by an order of magnitude or more. The claim is quantified through twin experiments in which Gaussian noise is added to a known True dataset to create synthetic CFD and experimental fields. For the turbulent cylinder at Re=2600, the LR configuration with compression rate CR_ub=15.9 gives a speed-up of 13.7, a RAM compression of 90.9%, and an RRMSE of 2.61%, compared to 0.83% for the HR reference at 5% noise. The authors further argue that lcSVD acts as a noise filter: at high compression it retains only the dominant coherent structures, which for the jet LES case actually produces lower RRMSE than the unfiltered high-resolution result.

Load-bearing premise

The reported LR accuracy presupposes that the lcSVD reconstruction is given a full-resolution (or semi-reduced) reference snapshot matrix, and the paper does not state where that matrix comes from in the twin experiments; if it comes from the same True data used to compute RRMSE, the errors are in-sample rather than independent estimates of the True state.

Editorial extensions

If this is right

  • Data assimilation can be applied to flows whose full-resolution state vector is too large for a routine EnKF, by moving the EnKF to a downsampled representation and reconstructing afterward.
  • At moderate compression (CR around 4-16), the LR route reproduces the HR accuracy while keeping the memory footprint small enough for single-node computation; e.g., the turbulent cylinder case drops from 37.35 GB to 3.40 GB RAM.
  • The lcSVD reconstruction provides implicit denoising, so in noise-dominated regimes the LR estimate can match or beat the HR estimate, as reported for the turbulent jet LES.
  • The compression rate is a tunable knob trading accuracy against cost: high compression (CR above 1000) can degrade RRMSE sharply, so practical use would choose CR to sit near the knee of the accuracy-cost curve.

Reading between the lines

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

  • Our inference: if the reference snapshot matrix for lcSVD is drawn from the same True dataset used for error evaluation, then the reported RRMSE values are in-sample reconstruction skills; an out-of-sample test would be needed to confirm the accuracy claim for genuinely unseen flows.
  • Our inference: the framework suggests a natural extension—choosing the downsampling points by optimal sensor placement rather than uniform subsampling could lower the compression rate needed for a given accuracy.
  • Our inference: the same downsampling-plus-reconstruction pipeline could be applied to other sequential estimators such as 3D-Var or particle filters, because the downsampling and reconstruction steps are independent of the EnKF update.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The manuscript proposes to couple the Ensemble Kalman Filter (EnKF) with low-resolution (LR) computations and a low-cost Singular Value Decomposition (lcSVD) reconstruction step. In a twin-experiment setup, noisy, downsampled surrogate CFD and experimental data are generated from three benchmark flow databases; EnKF is run at low resolution, and lcSVD is used to recover a high-resolution estimate. The paper reports large reductions in computation time and RAM usage with modest accuracy loss, e.g., Table 3, Case 28 for Re=2600: speed-up 13.7, RAM compression 90.9%, RRMSE 2.61% versus 0.83% for the high-resolution reference. The central methodological question is whether the lcSVD recovery step has access to full-resolution reference data during evaluation, which the manuscript never states explicitly.

Significance. If the reported accuracy is out-of-sample, the contribution is significant: it offers a practical route to real-time data assimilation in fluid dynamics using a standard EnKF formulation and a previously proposed lcSVD method, with resource comparisons reported on three public benchmark datasets. The paper also deserves credit for documenting computation times and RAM usage, and for using established test cases. However, the validation protocol as written does not establish that the LR reconstruction is independent of the evaluation target; until the source of the lcSVD reference matrix is clarified, the quantitative accuracy claims are conditional.

major comments (4)
  1. [§3.2.2, Eq. (3.10); §4, Eq. (4.2)] The recovery step W_rec = (\bar{V}^{K,\bar{K}}_1) \bar{T}(\bar{\Sigma})^{-1} requires a full-resolution or semi-reduced snapshot matrix with full spatial dimension J. Nowhere does the paper state how this matrix is obtained in the twin experiments: the EnKF pipeline described in Section 4 operates on downsampled u_b and w, so the assimilated state u_a exists only at the low resolution \bar{J}. If \bar{V}^{K,\bar{K}}_1 is assembled from the same True data used to define RRMSE in Eq. (4.2), then every reported LR RRMSE, including the headline 2.61% in Table 3 Case 28, is an in-sample reconstruction error rather than an independent estimate of the True state. Please state the origin of this matrix explicitly and enforce a genuine train/test split: learn the reconstruction basis from an independent time interval or from full-resolution data not used in the RRMSE evaluation, and report both in-sample and out-of-sample errors.
  2. [§3.2.2, mode-selection rule] The number of retained modes is fixed as N_modes = N_{s,ub} × K × 0.20 for all cases and all noise levels. No sensitivity study is provided to show that this percentage is appropriate for the different flow regimes (laminar wake, turbulent wake, turbulent jet), and the choice directly controls the reconstruction error because lcSVD projects onto this basis. Please provide a sensitivity analysis of RRMSE versus N_modes for each test case, or replace the fixed percentage with a criterion analogous to Eq. (3.5).
  3. [§6, Tables 2–4] The reported RRMSE and MAE values are single realizations. Since the noise added to u_b and w is sampled from a Gaussian (Section 4), different draws will change every accuracy metric, yet only computation time is stated to be averaged over ten runs (Section 4, after Eq. (4.4)). Without error bars, confidence intervals, or explicitly fixed random seeds, differences such as MAE 3.31% versus 3.23% in Table 3 cannot be assessed for significance. Please report statistics over repeated noise/ensemble realizations or state the seeds used.
  4. [§6.3, Table 4] The evaluation does not include a baseline that separates the EnKF contribution from the smoothing effect of lcSVD. In the jet case the LR reconstruction has a lower RRMSE than the HR reference (e.g., 0.31% versus 0.41% at 5% noise), which is attributed to lcSVD filtering. However, the same filtering could be obtained by applying lcSVD directly to the noisy low-resolution u_b without any assimilation step. To support the claim that EnKF improves the estimate, please compare EnKF+lcSVD with (i) lcSVD applied to u_b alone and (ii) EnKF at low resolution without reconstruction.
minor comments (5)
  1. [§5.2] The text describes the flow as 'turbulent three-dimensional' but the database is planar (N_x × N_y = 301 × 111, two velocity components); clarify whether the cylinder is spanwise three-dimensional or the database is two-dimensional.
  2. [§5.3] The nozzle temperature ratio is written as 'NPR = P_t/P_∞ = 1.15', reusing the abbreviation NPR for a temperature ratio; this appears to be a typo for T_t/T_∞.
  3. [Figure 27 caption] The caption states CR_ub = 16 for Case 46, while Table 1 and the surrounding text assign Case 46 CR_ub = 4; the caption and the main text must be aligned.
  4. [§3.2.2] The symbol N_{s,ub} is used in the definition of N_modes before it is introduced in Section 4; define it earlier or add a forward reference.
  5. [§6.3] The heuristic that 'fewer experimental points reduce exposure to noise' is not supported by a controlled comparison; the error reduction could equally be explained by the lcSVD projection onto a small number of modes.

Circularity Check

1 steps flagged · score 6.0 of 10

lcSVD reconstruction (Eq. 3.10) needs a full-resolution snapshot matrix that the paper never sources; in the twin experiments only the True ground truth has full resolution, so the reported LR RRMSE values are in-sample reconstructions, not independent predictions.

  1. self definitional [Section 3.2.2, Step 4 (Eq. 3.10); Section 4 methodology; Eq. 4.2]
    "Step 4: Recover SVD modes. The SVD modes expanded to their full spatial dimensions W from eq. (3.4) are reconstructed as: W≃W rec = ( ¯V K,J ¯K 1 ) ¯T( ¯Σ)−1, where W rec ∈R J× ¯N. ... The final step involves reconstructing the LR assimilated data into the original HR data using lcSVD [21] described in section 3. ... RRM SE= ∥V K 1 −V K,rec 1 ∥2 / ∥V K 1 ∥2, where ... V K 1 the original snapshot matrix (eq. 3.1), and V K,rec 1 the reconstruction of the original dataset (eq. 3.12)."

    In Eq. 3.10 the reconstruction basis W_rec is built from V^{K,J¯K}_1, a semi-reduced snapshot matrix with full spatial dimension J. In the twin experiments the only full-resolution dataset is the True ground truth: the CFD stream ub is generated by adding Gaussian noise to True and then downsampling, the experimental stream w is generated by downsampling True, and the assimilated state u_a is explicitly low-resolution ('at a low-resolution in this stage'). The paper never states where V^{K,J¯K}_1 comes from. If, as the setup implies, it is taken from the same True snapshot matrix used as the reference in Eq. 4.2, then W_rec is a projection onto a basis learned from the target itself. Every reported LR RRMSE (e.g.

full rationale

The EnKF formulation (Section 2) is standard and externally grounded in the DA literature, and the lcSVD algorithm itself is fully specified in Section 3, so those components are not circular. The circularity burden falls on the empirical validation of the central claim that LR computations preserve accuracy. The reconstruction step Eq. 3.10 requires a full-resolution semi-reduced snapshot matrix V^{K,J¯K}_1, but the described pipeline only ever produces low-resolution assimilated fields u_a from downsampled, noisy True data. The paper does not state the source of this full-resolution matrix; the only full-resolution dataset in the twin experiments is the True ground truth, which is also the reference in the RRMSE definition (Eq. 4.2). If that matrix is derived from True, then the reported LR RRMSE values are in-sample projection errors onto a basis constructed from the target field, not independent estimates of reconstruction or DA accuracy. This makes the headline speed-up/accuracy trade-off (e.g. Table 3) self-referential as written. The hand-tuned 20% mode fraction and absence of a train/test split reinforce the concern but are secondary. Overall score 6: the central accuracy claim reduces by construction unless the reference-matrix source is clarified and an out-of-sample protocol is enforced.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central claim rests on two standard pillars (EnKF and SVD) plus two hand-tuned choices (20% mode truncation and the compression-rate sweep) and one unstated data-provenance assumption (the full-resolution basis for lcSVD). No invented physical entities are introduced.

free parameters (3)
  • Nmodes fraction = 0.20
    Mode truncation for lcSVD is fixed at 20% of Ns,ub x K in all test cases (Section 3.2.2), rather than derived from the tolerance criterion in Eq. 3.5. Reconstruction accuracy depends directly on this choice.
  • Compression rate CR,ub and CR,w = CR,ub swept from 3.10 to 17413; CR,w fixed per case at 101 to 317
    Explicitly called a tuned parameter in Section 4. The favorable accuracy/speed trade-offs highlighted in Section 6 are selected from this sweep.
  • Ensemble size E = 25 for Re=100; 45 for turbulent cases
    Chosen by hand in Section 5. Ensemble size affects both accuracy and computational cost of the EnKF.
assumptions (5)
  • domain assumption The ensemble covariance B in Eq. 2.1 approximates the true background error covariance.
    Central to EnKF, but the approximation is only valid when the ensemble is representative of the error distribution.
  • domain assumption Observation and process noise are Gaussian and independent, with covariances R and Q.
    Invoked in Section 2.4 for perturbed observations and in Section 2.3 for process noise.
  • domain assumption Synthetic noise with sigma_N = eta * sigma_T added to True data mimics real CFD and experimental errors.
    Used in Section 4 to build the twin experiments; the realism of the noise model is not validated against actual measurement error statistics.
  • domain assumption Flow fields are low-rank enough that 20% of the available SVD modes capture the dynamics needed for accurate reconstruction.
    Invoked in Section 3.2.2 as the rule Nmodes = Ns,ub x K x 0.20; no convergence study justifies this fraction for each test case.
  • ad hoc to paper A full-resolution or semi-reduced reference snapshot matrix is available to lcSVD for the recovery step in Eq. 3.10.
    Never stated where this matrix comes from. Without it, Eqs. 3.10-3.12 cannot be executed from the sensor data alone.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Ensemble Kalman Filter for Data Assimilation coupled with low-resolution computations techniques applied in Fluid Dynamics." pith.science (2026). https://pith.science/paper/AFS3VDZD

@misc{pith2026250700539,
  author       = {Pith},
  title        = {Pith review of: Ensemble Kalman Filter for Data Assimilation coupled with low-resolution computations techniques applied in Fluid Dynamics},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AFS3VDZD}},
  note         = {Machine review of arXiv:2507.00539}
}
read the original abstract

This paper presents an innovative Reduced-Order Model (ROM) for merging experimental and simulation data using Data Assimilation (DA) to estimate the "True" state of a fluid dynamics system, leading to more accurate predictions. Our methodology introduces a novel approach implementing the Ensemble Kalman Filter (EnKF) within a reduced-dimensional framework, grounded in a robust theoretical foundation and applied to fluid dynamics. To address the substantial computational demands of DA, the proposed ROM employs low-resolution (LR) techniques to drastically reduce computational costs. This approach involves downsampling datasets for DA computations, followed by an advanced reconstruction technique based on low-cost Singular Value Decomposition (lcSVD). The lcSVD method, a key innovation in this paper, has never been applied to DA before and offers a highly efficient way to enhance resolution with minimal computational resources. Our results demonstrate significant reductions in both computation time and RAM usage through the LR techniques without compromising the accuracy of the estimations. For instance, in a turbulent test case, the LR approach with a compression rate of 15.9 can achieve a speed-up of 13.7 and a RAM compression of 90.9% while maintaining a low Relative Root Mean Square Error (RRMSE) of 2.6%, compared to 0.8% in the high-resolution (HR) reference. Furthermore, we highlight the effectiveness of the EnKF in estimating and predicting the state of fluid flow systems based on limited observations and low-fidelity numerical data. This paper highlights the potential of the proposed DA method in fluid dynamics applications, particularly for improving computational efficiency in CFD and related fields. Its ability to balance accuracy with low computational and memory costs makes it suitable for large-scale and real-time applications, such as environmental monitoring or aerospace.

Figures

Figures reproduced from arXiv: 2507.00539 by the authors.

Figure 1
Figure 1. Schematic representation of the methodology to model a (b) CFD database [PITH_FULL_IMAGE:figures/full_fig_p012_1.png] view at source ↗
Figure 2
Figure 2. Methodology scheme of the twin-experiment coupling DA code with low [PITH_FULL_IMAGE:figures/full_fig_p012_2.png] view at source ↗
Figure 3
Figure 3. Streamwise (left) and normal (right) velocities of a representative snapshot [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
Figures from the paper (26 more)
Figure 4
Figure 4. Figure 4: Streamwise (left) and normal (right) velocities of a representative snapshot [PITH_FULL_IMAGE:figures/full_fig_p016_4.png]
Figure 5
Figure 5. Figure 5: Streamwise velocity of a representative snapshot of the turbulent jet large [PITH_FULL_IMAGE:figures/full_fig_p017_5.png]
Figure 6
Figure 6. Figure 6: Case 12 in table 1. Cylinder dataset at Re = 100 results with 30% of noise level applied on both ub and w, HR method. Ns,ub × K = 139300 × 151 (CR,ub = 1) and Ns,w × K = 440 × 151 (CR,w = 317). From the top to the bottom: True data, CFD data, experimental data, DA reco…
Figure 7
Figure 7. Figure 7: Case 13 in table 1. Cylinder dataset at Re = 100 results with 50% of noise level applied on both ub and w, HR method. Ns,ub × K = 139300 × 151 (CR,ub = 1) and Ns,w × K = 440 × 151 (CR,w = 317). From the top to the bottom: True data, CFD data, experimental data, DA reco…
Figure 8
Figure 8. Figure 8: Velocity tracking point (1st red dot from the left side column 1 in figures [PITH_FULL_IMAGE:figures/full_fig_p022_8.png]
Figure 9
Figure 9. Figure 9: Velocity tracking point (2nd red dot from the left side column 1 in figures [PITH_FULL_IMAGE:figures/full_fig_p022_9.png]
Figure 10
Figure 10. Figure 10: Case 7 in table 1. Cylinder dataset at Re = 100 results with 5% of noise level applied on both ub and w, with lcSVD reconstruction method. Ns,ub ×K = 1800×151 (CR,ub = 77) and Ns,w×K = 440×151 (CR,w = 317). From the top to the bottom: True data, CFD data, experimental…
Figure 11
Figure 11. Figure 11: Case 10 in table 1. Cylinder dataset at Re = 100 results with 5% of noise level applied on both ub and w, with lcSVD reconstruction method. Ns,ub × K = 45000 × 151 (CR,ub = 3.10) and Ns,w × K = 440 × 151 (CR,w = 317). From the top to the bottom: True data, CFD data, e…
Figure 12
Figure 12. Figure 12: Velocity tracking point (1st red dot from the left side, column 1 in figures [PITH_FULL_IMAGE:figures/full_fig_p025_12.png]
Figure 13
Figure 13. Figure 13: Velocity tracking point (2nd red dot from the left side, column 1 in figures [PITH_FULL_IMAGE:figures/full_fig_p026_13.png]
Figure 14
Figure 14. Figure 14: Case 31 in table 1. Turbulent cylinder dataset at Re = 2600 results with 30% of noise level applied on both ub and w, HR method. Ns,ub × K = 66822 × 151 (CR,ub = 1) and Ns,w × K = 660 × 151 (CR,w = 101). From the top to the bottom: True data, CFD data, experimental da…
Figure 15
Figure 15. Figure 15: Case 32 in table 1. Turbulent cylinder dataset at Re = 2600 results with 50% of noise level applied on both ub and w, HR method. Ns,ub × K = 66822 × 151 (CR,ub = 1) and Ns,w × K = 660 × 151 (CR,w = 101). From the top to the bottom: True data, CFD data, experimental da…
Figure 16
Figure 16. Figure 16: Velocity tracking point (1st red dot from the left side, column 1 in figures [PITH_FULL_IMAGE:figures/full_fig_p030_16.png]
Figure 17
Figure 17. Figure 17: Velocity tracking point (2nd red dot from the left side, column 1 in figures [PITH_FULL_IMAGE:figures/full_fig_p030_17.png]
Figure 18
Figure 18. Figure 18: Case 25 in table 1. Turbulent cylinder dataset at Re = 2600 results with 5% of noise level on both ub and w, lcSVD reconstruction method. Ns,ub × K = 150 × 151 (CR,ub = 445) and Ns,w × K = 660 × 151 (CR,w = 110). From the top to the bottom: True data, CFD data, experi…
Figure 19
Figure 19. Figure 19: Case 29 in table 1. Turbulent cylinder dataset at Re = 2600 results with 5% of noise level applied on both ub and w, lcSVD reconstruction method. Ns,ub ×K = 16800 × 151 (CR,ub = 3.98) and Ns,w × K = 660 × 151 (CR,w = 101). From the top to the bottom: True data, CFD da…
Figure 20
Figure 20. Figure 20: Velocity tracking point (1st red dot from the left side, column 1 in figures [PITH_FULL_IMAGE:figures/full_fig_p034_20.png]
Figure 21
Figure 21. Figure 21: Velocity tracking point (2nd red dot from the left side, column 1 in figures [PITH_FULL_IMAGE:figures/full_fig_p035_21.png]
Figure 22
Figure 22. Figure 22: Case 48 in table 1. Turbulent jet large eddy (streamwise velocity) results with 30% of noise level applied on both ub and w, HR method. Ns,ub ×K = 6825×151 (CR,ub = 1) and Ns,w ×K = 27×151 (CR,w = 253). From the top to the bottom: True data, CFD data, experimental dat…
Figure 23
Figure 23. Figure 23: Case 49 in table 1. Turbulent jet large eddy (streamwise velocity) results with 50% of noise level applied on both ub and w, HR method. Ns,ub ×K = 6825×151 (CR,ub = 1) and Ns,w ×K = 27×151 (CR,w = 253). From the top to the bottom: True data, CFD data, experimental dat…
Figure 24
Figure 24. Figure 24: Velocity tracking point (1st red dot from the left side in figures [PITH_FULL_IMAGE:figures/full_fig_p038_24.png]
Figure 25
Figure 25. Figure 25: Velocity tracking point (2nd red dot from the left side in figures [PITH_FULL_IMAGE:figures/full_fig_p039_25.png]
Figure 26
Figure 26. Figure 26: Case 45 in table 1. Turbulent jet large eddy results with 5% of noise level, CR,ub = 16, with lcSVD reconstruction method. Ns,ub × K = 440 × 151 (CR,ub = 16) and Ns,w × K = 27 × 151 (CR,w = 253). From the top to the bottom: True data, CFD data, DA, experimental data, …
Figure 27
Figure 27. Figure 27: Case 46 in table 1. Turbulent jet large eddy results with 5% of noise level, CR,ub = 4, with lcSVD reconstruction method. Ns,ub × K = 1760 × 151 (CR,ub = 16) and Ns,w × K = 27 × 151 (CR,w = 253). From the top to the bottom: True data, CFD data, DA, experimental data, …
Figure 28
Figure 28. Figure 28: Velocity tracking point (1st red dot from the left side in figures [PITH_FULL_IMAGE:figures/full_fig_p043_28.png]
Figure 29
Figure 29. Figure 29: Velocity tracking point (2nd red dot from the left side in figures [PITH_FULL_IMAGE:figures/full_fig_p043_29.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. GenDA: Generative Data Assimilation on Complex Urban Areas via Classifier-Free Diffusion Guidance

    cs.LG 2026-01 conditional novelty 6.0 of 10

    A sensor-conditioned graph diffusion model reconstructs urban wind fields from <1% of mesh nodes, beating supervised GNN and SVD baselines on held-out altitude slices of one RANS dataset.

Reference graph

Works this paper leans on

47 extracted references · 37 canonical work pages · cited by 1 Pith paper

  1. [1]

    L., & Trucano, T

    Oberkampf, W. L., & Trucano, T. G. (2002). Verification and validation in computational fluid dynamics.Progress in Aerospace Sciences, 38(3), 209-272

    work page 2002

  2. [2]

    E., Borgoltz, A., Devenport, W

    Gargiulo, A., Duetsch-Patel, J. E., Borgoltz, A., Devenport, W. J., Roy, C. J., & Lowe, K. T. (2023). Strategies for computational fluid dynamics validation experiments.Journal of Verification, Validation and Uncertainty Quantification, 8(3), 031004.https://doi. org/10.1115/1.4063639

    work page doi:10.1115/1.4063639 2023

  3. [3]

    Wu, X., Zhang, W., & Song, S. (2017). Uncertainty quantification and sensitivity analysis of transonic aerodynamics with geometric uncertainty.International Journal of Aerospace Engineering, 2017(1), 8107190.https://doi.org/10.1155/2017/8107190

    work page doi:10.1155/2017/8107190 2017

  4. [4]

    Barzegar, R., Shafee, S., & Khalilarya, S. (2013). Computational fluid dynamics sim- ulation of the combustion process, emission formation and the flow field in an in- direct injection diesel engine.Thermal Science, 17(1), 11-23.https://doi.org/10.2298/ TSCI111218108B

    work page 2013

  5. [5]

    Khan, M., Odemark, Y., & Fransson, J. H. M. (2017). Effects of inflow conditions on wind turbine performance and near wake structure.Open Journal of Fluid Dynamics, 7(1), 105-129.https://doi.org/10.4236/ojfd.2017.71008

    work page arXiv 2017

  6. [6]

    Blocken, B. (2015). Computational fluid dynamics for urban physics: Importance, scales, possibilities, limitations and ten tips and tricks towards accurate and reliable simulations. Building and Environment, 91, 219-245.https://doi.org/10.1016/j.buildenv.2015. 02.015

    work page doi:10.1016/j.buildenv.2015 2015

  7. [7]

    Parente, A., Gorlé, C., van Beeck, J., & Benocci, C. (2011). A comprehensive modelling approach for the neutral atmospheric boundary layer: Consistent inflow conditions, wall function and turbulence model.Boundary-Layer Meteorology, 140(3), 411-428.https: //doi.org/10.1007/s10546-011-9621-5

    work page doi:10.1007/s10546-011-9621-5 2011

  8. [8]

    García-Sánchez, C., Vitalis, S., Paden, I., & Stoter, J. (2021). The impact of level of detail in 3D city models for CFD-based wind flow simulations.The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 46, 67-72. https://doi.org/10.5194/isprs-archives-XLVI-4-W4-2021-67-2021

    work page doi:10.5194/isprs-archives-xlvi-4-w4-2021-67-2021 2021

Show all 47 references
  1. [9]

    Meldi, M., & Poux, A. (2017). A reduced order model based on Kalman filtering for sequential data assimilation of turbulent flows.Journal of Computational Physics, 347, 207-234.https://doi.org/10.1016/j.jcp.2017.06.042

    2017 doi

  2. [10]

    Moldovan, G., Lehnasch, G., Cordier, L., & Meldi, M. (2021). A multigrid/ensem- ble Kalman filter strategy for assimilation of unsteady flows.Journal of Computational Physics, 443, 110481.https://doi.org/10.1016/j.jcp.2021.110481

    2021

  3. [11]

    Aristodemou, E., Arcucci, R., Mottet, L., Robins, A., Pain, C., & Guo, Y.-K. (2019). Enhancing CFD-LES air pollution prediction accuracy using data assimilation.Building and Environment, 165, 106383.https://doi.org/10.1016/j.buildenv.2019.106383

    2019

  4. [12]

    S., & Ridley, A

    Gillijns, S., Barrero Mendoza, O., Chandrasekar, J., De Moor, B., Bernstein, D. S., & Ridley, A. (2006). What is the ensemble Kalman filter and how well does it work? In Proceedings of the 2006 American Control Conference(pp. 6-pp). IEEE.https://doi. org/10.1109/ACC.2006.1657419 46

    2006 arXiv

  5. [13]

    D., Sandu, A., & Deng, X

    Nino-Ruiz, E. D., Sandu, A., & Deng, X. (2018). An ensemble Kalman filter implementa- tion based on modified Cholesky decomposition for inverse covariance matrix estimation. SIAM Journal on Scientific Computing, 40(2), A867-A886.https://doi.org/10.1137/ 16M1097031

    2018

  6. [14]

    Mons, V., Chassaing, J.-C., Gomez, T., & Sagaut, P. (2016). Reconstruction of unsteady viscous flows using data assimilation schemes.Journal of Computational Physics, 316, 255-280.https://doi.org/10.1016/j.jcp.2016.04.022

    2016 doi

  7. [15]

    H., Cessna, J

    Colburn, C. H., Cessna, J. B., & Bewley, T. R. (2011). State estimation in wall-bounded flow systems. Part 3. The ensemble Kalman filter.Journal of Fluid Mechanics, 682, 289- 303.https://doi.org/10.1017/jfm.2011.222

    2011 doi

  8. [16]

    M., Glumac, A

    Villanueva, L., Valero, M. M., Glumac, A. Š., & Meldi, M. (2024). Augmented state estimation of urban settings using on-the-fly sequential data assimilation.Computers & Fluids, 269, 106118.https://doi.org/10.1016/j.compfluid.2023.106118

    2024

  9. [17]

    Chandramouli, P., Mémin, E., & Heitz, D. (2020). 4D large scale variational data as- similation of a turbulent flow with a dynamics error model.Journal of Computational Physics, 412, 109446.https://doi.org/10.1016/j.jcp.2020.109446

    2020

  10. [18]

    R., & Holm, H

    Brodtkorb, A. R., & Holm, H. H. (2019). Real-world oceanographic simulations on the GPU using a two-dimensional finite-volume scheme.arXiv preprint arXiv:1912.02457. https://doi.org/10.48550/arXiv.1912.02457

    2019 doi arXiv

  11. [19]

    A., Bode, M., Ferrier, N., Fischer, P., Göbbert, J

    Mateevitsi, V. A., Bode, M., Ferrier, N., Fischer, P., Göbbert, J. H., Insley, J. A., Lan, Y.-H., Min, M., Papka, M. E., Patel, S., Rizzi, S., & Windgassen, J. (2023). Scaling com- putational fluid dynamics: In situ visualization of NekRS using SENSEI. InProceedings of the SC ...

    2023 doi

  12. [20]

    J., & Tong, X

    Majda, A. J., & Tong, X. T. (2018). Performance of ensemble Kalman filters in large dimensions.Communications on Pure and Applied Mathematics, 71(5), 892-937.https: //doi.org/10.1002/cpa.21722

    2018 doi

  13. [21]

    Hetherington, A., & Le Clainche, S. (2023). Low-cost singular value decomposition with optimal sensor placement.arXiv preprint arXiv:2311.09791.https://doi.org/10. 48550/arXiv.2311.09791

    2023 doi arXiv

  14. [22]

    Sirovich, L. (1987). Turbulence and the dynamics of coherent structures. I. Coher- ent structures.Quarterly of Applied Mathematics, 45(3), 561-571.https://doi.org/10. 1090/qam/910462

    1987

  15. [23]

    Hetherington, A., Corrochano, A., Abadía-Heredia, R., Lazpita, E., Muñoz, E., Díaz, P., Corrochano, S., García-Vallejo, D., Ferrer, E., & Le Clainche, S. (2024). ModelFLOWs- app: Data-driven post-processing and reduced order modelling tools.Computer Physics Communications, 301...

    2024

  16. [24]

    E., Pawar, S., & San, O

    Ahmed, S. E., Pawar, S., & San, O. (2020). PyDA: A hands-on introduction to dy- namical data assimilation with Python.Fluids, 5(4), 225.https://doi.org/10.3390/ fluids5040225

    2020

  17. [25]

    M., Lakshmivarahan, S., & Dhall, S

    Lewis, J. M., Lakshmivarahan, S., & Dhall, S. (2006).Dynamic data assimilation: A least squares approach(Vol. 13). Cambridge University Press. 47

    2006

  18. [26]

    Evensen, G. (1994). Sequential data assimilation with a nonlinear quasi-geostrophic model using Monte Carlo methods to forecast error statistics.Journal of Geophysical Research: Oceans, 99(C5), 10143-10162.https://doi.org/10.1029/94JC00572

    1994 doi

  19. [27]

    J., & Evensen, G

    Burgers, G., van Leeuwen, P. J., & Evensen, G. (1998). Analysis scheme in the ensemble Kalman filter.Monthly Weather Review, 126(6), 1719-1724.https://doi.org/10.1175/ 1520-0493(1998)126%3C1719:ASITEK%3E2.0.CO;2

    1998

  20. [28]

    Evensen, G. (2003). The ensemble Kalman filter: Theoretical formulation and practical implementation.Ocean Dynamics, 53, 343-367.https://doi.org/10.1007/ s10236-003-0036-9

    2003

  21. [29]

    Lumley, J. L. (1967). The structure of inhomogeneous turbulent flows.Atmo- spheric turbulence and radio wave propagation, 166-178.https://cir.nii.ac.jp/crid/ 1574231874542771712

    1967

  22. [30]

    Le Clainche, S., Izbassarov, D., Rosti, M., Brandt, L., & Tammisola, O. (2020). Coherent structures in the turbulent channel flow of an elastoviscoplastic fluid.Journal of Fluid Mechanics, 888, A5.https://doi.org/10.1017/jfm.2020.31

    2020 doi

  23. [31]

    Le Clainche, S., & Vega, J. M. (2017). Higher order dynamic mode decomposition. SIAM Journal on Applied Dynamical Systems, 16(2), 882-925.https://doi.org/10. 1137/15M1054924

    2017

  24. [32]

    Parente, A., & Sutherland, J. C. (2013). Principal component analysis of turbulent combustion data: Data pre-processing and manifold sensitivity.Combustion and Flame, 160(2), 340-350.https://doi.org/10.1016/j.combustflame.2012.09.016

    2013 doi

  25. [33]

    L., Terragni, F., & Vega, J

    Rapún, M. L., Terragni, F., & Vega, J. M. (2017). LUPOD: Collocation in POD via LU decomposition.Journal of Computational Physics, 335, 1-20.https://doi.org/10. 1016/j.jcp.2017.01.005

    2017

  26. [34]

    Deng, Z., He, C., Wen, X., & Liu, Y. (2018). Recovering turbulent flow field from local quantity measurement: Turbulence modeling using ensemble-Kalman-filter-based data assimilation.Journal of Visualization, 21, 1043-1063.https://doi.org/10.1007/ s12650-018-0508-0

    2018

  27. [35]

    Jackson, C. P. (1987). A finite-element study of the onset of vortex shedding in flow past variously shaped bodies.Journal of Fluid Mechanics, 182, 23-45.https://doi.org/10. 1017/S0022112087002234

    1987

  28. [36]

    Barkley, D., & Henderson, R. D. (1996). Three-dimensional Floquet stability analysis of the wake of a circular cylinder.Journal of Fluid Mechanics, 322, 215-241.https: //doi.org/10.1017/S0022112096002777

    1996 doi

  29. [37]

    mcs.anl.gov/

    Nek5000.(2023).Nek5000: Open source spectral element CFD solver.https://nek5000. mcs.anl.gov/

    2023

  30. [38]

    A., Hess, D., Watz, B

    Mendez, M. A., Hess, D., Watz, B. B., & Buchlin, J.-M. (2020). Multiscale proper orthogonal decomposition (mPOD) of TR-PIV data—A case study on stationary and transient cylinder wake flows.Measurement Science and Technology, 31(9), 094014. https://doi.org/10.1088/1361-6501/ab82be 48

    2020 doi

  31. [39]

    A., Jordan, P., Jaunet, V., Le Rallic, M., Cavalieri, A

    Brès, G. A., Jordan, P., Jaunet, V., Le Rallic, M., Cavalieri, A. V., Towne, A., Lele, S. K., Colonius, T., & Schmidt, O. T. (2018). Importance of the nozzle-exit boundary- layer state in subsonic turbulent jets.Journal of Fluid Mechanics, 851, 83-124.https: //doi.org/10.1017/...

    2018 doi

  32. [40]

    Jain, N., Bravo, L., Kim, D., Murugan, M., Ghoshal, A., Ham, F., & Flatau, A. (2020). Massively parallel large eddy simulation of rotating turbomachinery for variable speed gas turbine engine operation.Energies, 13(3), 703.https://doi.org/10.3390/en13030703

    2020 doi

  33. [41]

    A., Bose, S

    Brès, G. A., Bose, S. T., Emory, M., Ham, F. E., Schmidt, O. T., Rigas, G., & Colonius, T. (2018). Large-eddy simulations of co-annular turbulent jet using a Voronoi-based mesh generation framework. In2018 AIAA/CEAS Aeroacoustics Conference(p. 3302).https: //doi.org/10.2514/6....

    2018 doi

  34. [42]

    Vreman, A. W. (2004). An eddy-viscosity subgrid-scale model for turbulent shear flow: Algebraic theory and applications.Physics of Fluids, 16(10), 3670-3681.https://doi. org/10.1063/1.1785131

    2004 doi

  35. [43]

    Freund, J.B.(1997).Proposedinflow/outflowboundaryconditionfordirectcomputation of aerodynamic sound.AIAA Journal, 35(4), 740-742.https://doi.org/10.2514/2.167

    1997 doi

  36. [44]

    Mani, A. (2012). Analysis and optimization of numerical sponge layers as a nonreflective boundary treatment.Journal of Computational Physics, 231(2), 704-716.https://doi. org/10.1016/j.jcp.2011.10.017

    2012 doi

  37. [45]

    Zaman, K. B. M. Q. (1998). Asymptotic spreading rate of initially compressible jets—Experiment and analysis.Physics of Fluids, 10(10), 2652-2660.https://doi.org/ 10.1063/1.869778

    1998 doi

  38. [46]

    Zaman, K. B. M. Q. (1999). Spreading characteristics of compressible jets from nozzles of various geometries.Journal of Fluid Mechanics, 383, 197-228.https://doi.org/10. 1017/S0022112099003833

    1999

  39. [47]

    (2023).ModelFLOWs-app.https://modelflows

    ModelFLOWs Research Group. (2023).ModelFLOWs-app.https://modelflows. github.io/modelflowsapp/ 49

    2023

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.