Pith. sign in

REVIEW 3 major objections 5 minor 23 references

The impact of using reconditioned correlated observation error covariance matrices in the Met Office 1D-Var system

T0 review · 3 major / 5 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 Correlated observation-error matrices converge faster than the operational diagonal one in 1D-Var, and reconditioning increases the number of observations that pass quality control.

desk verdict The convergence comparison is undermined by a scale-dependent stopping criterion, but the operational QC results are real and the paper deserves a revision rather than rejection. read the letter →

arxiv 1908.04071 v1 pith:65AQO6RQ submitted 2019-08-12 physics.ao-ph

Jemima M. Tabeart , Sarah L. Dance , Amos S. Lawless , Stefano Migliorini , Nancy K. Nichols , Fiona Smith , Joanne A. Waller This is my paper · ORCID
classification physics.ao-ph
keywords dataassimilationobservationerrorcovarianceIASI1D-Varreconditioningridgeregressionqualitycontrolconvergence
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 asks whether a correlated observation-error covariance matrix, repaired by ridge-regression reconditioning, can replace the diagonal matrix currently used in an operational one-dimensional variational retrieval system for IASI satellite radiances. Testing seven candidate matrices, it finds that the operational diagonal matrix requires the most iterations to converge, while every correlated choice converges faster and stronger reconditioning converges faster still. It also finds that reconditioned correlated matrices let more observations pass the ten-iteration quality-control gate. Retrieved skin temperature, cloud fraction and cloud top pressure change by only small amounts for most observations, but a few percent show very large differences, so the paper concludes that quality control would need to be retuned alongside any change of matrix.

What carries the argument

The central object is ridge-regression reconditioning: for a diagnosed observation-error covariance matrix $\mathbf{R}$, form $\mathbf{R}_{\mathrm{RR}} = \mathbf{R} + \delta\mathbf{I}$ with $\delta = (\lambda_{\max}(\mathbf{R}) - \lambda_{\min}(\mathbf{R})\kappa_{\max})/(\kappa_{\max}-1)$, where $\kappa_{\max}$ is a user-chosen target condition number. This raises every eigenvalue, so the small eigenvalues that hurt convergence are lifted while off-diagonal correlations are reduced. The paper uses bounds on the condition number of the Hessian $\mathbf{S} = \mathbf{B}^{-1} + \mathbf{H}^T\mathbf{R}^{-1}\mathbf{H}$ to explain why the minimum eigenvalue of $\mathbf{R}$ controls the speed of conjugate-gradient convergence, and shows empirically that the qualitative prediction survives in a nonlinear operational retrieval.

What would settle it

Rerun the same 1D-Var experiments using an OEC matrix diagnosed from 1D-Var background and observation statistics, as the paper notes was not done; if the operational diagonal matrix then converges in no more iterations than the correlated matrices, or if reconditioning no longer increases quality-control acceptance, the central convergence-and-throughput claim collapses.

Watch

Extended reading notes

Core claim

On the paper's own terms: replacing the diagonal observation-error covariance matrix for IASI in a 1D-Var pre-processing system with a correlated matrix reconditioned by ridge regression improves both convergence and throughput. The current diagonal matrix is the slowest-converging of the seven choices tested; increasing the amount of reconditioning—raising the minimum eigenvalue while lowering the target condition number from 1500 to 67—monotonically reduces iteration counts and Hessian condition numbers. That matches theoretical bounds derived for linear observation operators even though the radiative-transfer observation operator here is nonlinear. An inflated diagonal matrix converges fastest of all, showing that variance inflation is a separate lever. More observations pass quality control with the reconditioned correlated matrices, and for the majority of observations the changes to skin temperature, cloud top pressure and cloud fraction stay within retrieved-standard-deviation envelopes, although a few percent of retrievals show very large differences that the paper treats as 1D-Var failures.

Load-bearing premise

The load-bearing premise is that error-correlation estimates taken from the larger four-dimensional assimilation system are representative enough for the one-dimensional system where the retrievals actually run; the paper explicitly notes this is not theoretically consistent, and if the mismatch is large the convergence and quality-control gains may not survive.

Editorial extensions

If this is right

  • If a correlated OEC matrix is adopted in 1D-Var, the convergence budget can be tightened, for example from ten to eight iterations, saving computation in a procedure that runs every six hours.
  • More observations pass the quality-control gate, changing the observation set passed to the main four-dimensional assimilation, so forecast impacts would need monitoring.
  • The theoretical result that raising the minimum eigenvalue of $\mathbf{R}$ improves conditioning extends qualitatively to a nonlinear observation operator in an operational system.
  • Retrieved uncertainties increase when correlations are introduced, so the demonstrated benefit is computational and in quality-control throughput rather than in reduced retrieval error.
  • Quality-control rules must be retuned: extreme skin-temperature differences of more than 20 K appear for a small number of observations and are best treated as retrieval failures rather than physical signals.

Reading between the lines

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

  • The surprising slowness of the diagonal matrix suggests that the variance pattern, not just the minimum eigenvalue, drives convergence; a controlled comparison holding variances fixed while toggling correlations would isolate the mechanism.
  • The benefits may depend on the paper's choice of a 4D-Var-derived correlated matrix; if error correlations estimated within the 1D-Var system itself are much smaller, the convergence and quality-control gains could shrink or vanish.
  • The same reconditioned-matrix comparison could be run for other hyperspectral infrared sounders to test whether this convergence advantage is a general property of correlated IASI-type radiance assimilation.
  • A stricter iteration cap combined with a reconditioned correlated matrix is a concrete, testable way to convert the faster convergence into operational savings while keeping or improving the quality-control acceptance rate.
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 / 5 minor

Summary. The paper investigates whether ridge-regression reconditioning of correlated observation error covariance (OEC) matrices benefits the Met Office 1D-Var system for IASI observations. Seven OEC matrices are compared: the operational diagonal matrix, an inflated diagonal matrix, the raw diagnosed correlated matrix, and four reconditioned versions with target condition numbers 1500, 1000, 500, and 67. Using a common set of 97,330 observations from a single date, the paper reports iteration counts to convergence, Hessian condition numbers, retrieval differences for temperature and humidity, quality-control pass rates, and changes in retrieved skin temperature, cloud fraction, and cloud top pressure. The central claim is that the current uncorrelated matrix requires more iterations to converge than any correlated/reconditioned matrix, and that reconditioned matrices increase the number of observations passing quality control.

Significance. If the convergence and quality-control findings are valid, the paper provides a practically important result: correlated OEC information can be used in an operational 1D-Var system without the expected computational slowdown, and can even reduce iteration counts and increase observation throughput. The study has notable strengths: a large observational sample (97,330 observations), seven OEC matrices, use of a real operational assimilation framework, and clear figures and tables. The convergence and QC results are empirical measurements rather than consequences of the authors' own theory, and the agreement with earlier theoretical bounds is a useful operational check. However, the headline convergence measurement is the crux of the paper, and the scale-dependent convergence criterion used in Section 4.1 is a serious load-bearing concern. If that concern is not resolved, the abstract's central claim about computational benefits is not supported.

major comments (3)
  1. [§4.1 and §3.1] Section 4.1 defines convergence for the reported iteration counts by the absolute difference between successive state estimates being smaller than 0.4σB. This criterion is not scale-invariant: a Gauss-Newton iteration started from the background produces increments that shrink as the observation weight decreases, so inflating R (as in Einfl) or adding δI via ridge regression can satisfy the threshold in fewer iterations even when the cost function is no closer to its minimum. Section 3.1 states that the operational 1D-Var convergence criterion is based on the cost function and normalized gradient. Because the abstract's central claim that correlated/reconditioned matrices are computationally beneficial rests on these niter counts, the claim is not currently established. Please report iteration counts under the operational cost-function/gradient criterion, or demonstrate that the step-size criterion yields the same ordering when evaluated on J and its gradient.
  2. [§3.2] The correlated matrix Rest is diagnosed from 4D-Var background and observation statistics, and the text acknowledges that this is 'not theoretically consistent' with the smaller error correlations previously estimated for the 1D-Var problem. Since the paper's title and conclusions concern the 1D-Var system, the magnitude of this mismatch matters. If the true 1D-Var correlations are considerably weaker, the convergence and QC benefits found here may not persist. The authors should quantify this sensitivity, for example by repeating the comparison with a 1D-Var-consistent diagnostic, or explicitly restrict the scope of the claims to the 4D-Var-style statistics used here.
  3. [§5.1 and Table 4] The prediction and explanation that reconditioned correlated matrices increase the number of observations passing quality control are built directly on the Section 4.1 iteration counts. If those counts are biased by the scale-dependent step-size convergence criterion, then the QC conclusion in Table 4 also needs to be re-established using the operational cost-function/gradient criterion. The current presentation does not make clear whether the acceptance counts in Table 4 use the operational criterion or the modified step-size criterion.
minor comments (5)
  1. [Abstract and §6] The phrase 'as the reconditioning parameter is increased' is ambiguous, because the experiments vary the target condition number κ_max, which decreases as more reconditioning is applied; please specify whether the parameter is δ or κ_max.
  2. [§3.2] In the list of experiments, 'E1500E1000' appears without a separator; this seems to be a typo for 'E1500, E1000'.
  3. [Figure 6(c)] The tick label 'Ra' in Figure 6(c) appears to be a typo for 'est' (the Eest experiment).
  4. [§4.1, discussion of Table 3] The text says that increasing λ_min(R) results in a decrease in the maximum value of κ(R), but Table 3 reports κ(S), the Hessian condition number; the symbol should be corrected.
  5. [§3.2 and §4] The paper states that results were 'similar across all trials' for several dates between December 2015 and June 2016, but only 16 June 2016 is shown; a brief summary of the other dates would make this claim checkable.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the convergence and quality-control results are direct empirical measurements, and the self-cited matrix theory is not used to force them.

full rationale

The paper's central claims are empirical: Figure 3 and Tables 3-6 report measured iteration counts, Hessian condition numbers, quality-control pass rates, and retrieval differences for seven prescribed observation error covariance matrices. These numbers are not derived from the authors' prior theory; they are outputs of the operational 1D-Var system. The correlated matrices are constructed by the ridge-regression definition (Section 2.2) and the DBCP diagnostic (Section 3.2), but the convergence and quality-control statistics are compared against the operational diagonal matrix, so the comparisons are not fitted to the conclusions. The cited results of Tabeart et al. (2018, 2019) motivate the experiments and supply theoretical bounds, but the empirical rankings do not reduce to those bounds; the paper itself notes that the nonlinear observation operator means the linear bounds need not apply (Section 3.1). The acknowledged limitation that the OEC matrix is diagnosed from 4D-Var statistics rather than 1D-Var (Section 3.2) is a validity concern, not a circularity. A separate methodological caveat outside circularity is that the niter metric uses an absolute 0.4 sigma_B step-size threshold (Section 4.1), which is scale-dependent and may favour inflated variance matrices; this could weaken the operational conclusion, but it does not make the measurements equivalent to the inputs by construction.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

No invented entities. The paper's claims are empirical, so the ledger is small: the main external inputs are the diagnosed OEC matrix and the operational B matrix, and the main hand-chosen settings are the target condition numbers.

free parameters (1)
  • Target condition number κ_max for ridge regression = 67, 500, 1000, 1500
    User-specified reconditioning targets in the ridge regression method (Section 2.2, Table 2). Results are reported per choice; these are experimental design choices, not constants fitted to data.
assumptions (4)
  • domain assumption The DBCP diagnostic (Desroziers et al. 2005) provides a reliable estimate of interchannel observation error correlations for IASI.
    The correlated matrix Rest is obtained by symmetrizing the DBCP diagnostic output (Section 3.2). The accuracy of the diagnostic depends on the initial background and OEC matrices used.
  • domain assumption Observation error statistics diagnosed from the 4D-Var system are applicable to the 1D-Var system.
    Section 3.2 states the OEC matrix was estimated using 4D-Var background and OEC matrices, which is not theoretically consistent with 1D-Var error statistics. The paper relies on this for all correlated experiments.
  • domain assumption Qualitative theoretical bounds for linear observation operators (Tabeart et al. 2018) carry over to the nonlinear IASI radiative transfer model.
    Section 2.2 presents bounds on κ(S) derived for linear H; Section 4.1 interprets the nonlinear 1D-Var results as confirming these qualitative conclusions.
  • standard math Cauchy interlacing theorem: deleting rows and columns of a symmetric positive definite matrix does not increase its condition number.
    Section 3.1 uses this to bound the condition number when cloud-affected channels are removed from the OEC matrix; standard linear algebra result cited to Bernstein (2009).

how reviews work

0 comments
Cite this review

Pith. "Pith review of The impact of using reconditioned correlated observation error covariance matrices in the Met Office 1D-Var system." pith.science (2026). https://pith.science/paper/65AQO6RQ

@misc{pith2026190804071,
  author       = {Pith},
  title        = {Pith review of: The impact of using reconditioned correlated observation error covariance matrices in the Met Office 1D-Var system},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/65AQO6RQ}},
  note         = {Machine review of arXiv:1908.04071}
}
read the original abstract

Recent developments in numerical weather prediction have led to the use of correlated observation error covariance (OEC) information in data assimilation and forecasting systems. However, diagnosed OEC matrices are often ill-conditioned and may cause convergence problems for variational data assimilation procedures. Reconditioning methods are used to improve the conditioning of covariance matrices while retaining correlation information. In this paper we study the impact of using the 'ridge regression' method of reconditioning to assimilate Infrared Atmospheric Sounding Interferometer (IASI) observations in the Met Office 1D-Var system. This is the first systematic investigation of how changing target condition numbers affects convergence of a 1D-Var routine. This procedure is used for quality control, and to estimate key variables (skin temperature, cloud top pressure, cloud fraction) that are not analysed by the main 4D-Var data assimilation system. Our new results show that the current (uncorrelated) OEC matrix requires more iterations to reach convergence than any choice of correlated OEC matrix studied. This suggests that using a correlated OEC matrix in the 1D-Var routine would have computational benefits for IASI observations. Using reconditioned correlated OEC matrices also increases the number of observations that pass quality control. However, the impact on skin temperature, cloud fraction and cloud top pressure is less clear. As the reconditioning parameter is increased, differences between retrieved variables for correlated OEC matrices and the operational diagonal OEC matrix reduce. As correlated choices of OEC matrix yield faster convergence, using stricter convergence criteria along with these matrices may increase efficiency and improve quality control.

Figures

Figures reproduced from arXiv: 1908.04071 by the authors.

Figure 1
Figure 1. Standard deviation values for the operational backgrou [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. Correlation matrices for the operational background er [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Number of iterations required for convergence of the min [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Background minus retrieved profiles from observation at [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]
Figure 5
Figure 5. Figure 5: Differences in retrievals between Ediag and E67 for trial on 16th June 0000Z for (a) tem￾perature and (b) ln(specific humidity) for 97330 observations. Dashed lines and solid lines give the mean RSD values for Ediag and E67 respectively. Dashed lines with dots denote t…
Figure 6
Figure 6. Figure 6: Box plot showing differences between retrieved variables fo [PITH_FULL_IMAGE:figures/full_fig_p016_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

23 extracted references · 21 canonical work pages

  1. [1]

    Bathmann

    K. Bathmann. Justification for estimating observation-error cov ariances with the Desroziers diagnostic. Quarterly Journal of the Royal Meteorological Society , 144(715):1965–1974,

    work page 1965

  2. [5]

    URL https://doi.org/10.1175/MWR-D-16-0240.1

    doi: 10.1175/MWR-D-16-0240.1. URL https://doi.org/10.1175/MWR-D-16-0240.1 . G. Chalon, F. Cayla, and D. Diebel. IASI: An advanced sounder for o perational meteorology. In Proceedings of IAF, Toulouse, France ,

    work page doi:10.1175/mwr-d-16-0240.1

  3. [11]

    doi: 10.1002/qj.3183. P. Gauthier, P. Du, S. Heilliette, and L. Garand. Convergence issue s in the estimation of interchannel correlated observation errors in infrared radiance data. Monthly Weather Review , 146 (10):3227–3239,

    work page doi:10.1002/qj.3183

  4. [12]

    doi: 10.1175/MWR-D-17-0273.1. G. H. Golub and C. F. Van Loan. Matrix Computations . The John Hopkins University Press, third edition,

    work page doi:10.1175/mwr-d-17-0273.1

  5. [13]

    doi: 10.1002/qj.3130. 20 C. Lupu, C. Cardinali, and A. P. McNally. Adjoint-based forecast se nsitivity applied to observation-error variance tuning. Q. J. R. Meteorol. Soc. , 141:3157–3165,

    doi:10.1002/qj.3130

  6. [14]

    URL https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/qj.3249

    doi: 10.1002/qj.3249. URL https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/qj.3249. J. Nocedal. Numerical optimization. Springer series in operations research and financial engineering. Springer, New York ; London, 2nd ed. edition,

    work page doi:10.1002/qj.3249

  7. [18]

    L. M. Stewart, J. Cameron, S. L. Dance, S. English, J. Eyre, and N . K. Nichols. Observation error correlations in IASI radiance data. Mathematics report series, 1/ 2009, University of Reading, Reading, UK,

    work page 2009

  8. [21]

    URL http://dx.doi.org/10.1002/nla.2165. J. M. Tabeart, S. L. Dance, A. S. Lawless, N. K. Nichols, and J. A. W aller. Improving the conditioning of estimated covariance matrices

    doi:10.1002/nla.2165

Show all 23 references
  1. [22]

    J. A. Waller, S. P. Ballard, S. L. Dance, G. Kelly, N. K. Nichols, and D. S imonin. Diagnosing Horizontal and Inter-Channel Observation Error Correlations f or SEVIRI Observations Using Observation-Minus-Background and Observation-Minus-Analysis Statistics. Remote Sensing , 8 ...

    doi

  2. [23]

    doi: 10.1002/qj.3117. T. Wang, J. Fei, X. Cheng, X. Huang, and J. Zhong. Estimating the c orrelated observation-error characteristics of the chinese fengyun microwave temperature s ounder and microwave humidity sounder. Advances in Atmospheric Sciences , 35(11):1428–1441,

    doi

  3. [24]

    doi: 10.1007/s00376-018-8014-9

    ISSN 1861-9533. doi: 10.1007/s00376-018-8014-9. P. Weston. Progress towards the implementation of correlated ob servation errors in 4D-Var. Forecasting research technical report 560, Met Office, Exeter, UK,

    doi

  4. [877]

    Bormann, A

    N. Bormann, A. J. Geer, and P. Bauer. Estimates of observation- error characteristics in clear and cloudy regions for microwave imager radiances from numerical weat her prediction. Q. J. R. Meteorol. Soc., 137:2014–2023,

    2014

  5. [1989]

    doi: 10.1002/qj.49711548902. A. M. Fowler, S. L. Dance, and J. A. Waller. On the interaction of obs ervation and prior error correlations in data assimilation. Quarterly Journal of the Royal Meteorological Society , 144(710): 48–62,

    doi

  6. [2001]

    A. D. Collard. Selection of iasi channels for use in numerical weather prediction. Quarterly Journal of the Royal Meteorological Society , 133(629):1977–1991,

    1977

  7. [2007]

    URL https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/qj.178

    doi: 10.1002/qj.178. URL https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/qj.178. A. D. Collard, A. P. McNally, F. I. Hilton, S. B. Healy, and N. C. Atkinso n. The use of principal component analysis for the assimilation of high-resolution infrared s ounder observations f...

    doi

  8. [2008]

    doi: 10.1002/qj.243. C. Prates, S. Migliorini, L. Stewart, and J. Eyre. Assimilation of tran sformed retrievals obtained from clear-sky IASI measurements. Q. J. R. Meteorol. Soc. , 142:1697–1712,

    doi

  9. [2010]

    doi: 10.1002/qj.701. M. Cordoba, S. L. Dance, G. A. Kelly, N. K. Nichols, and J. A. Waller. D iagnosing Atmospheric Motion Vector observation errors for an operational high resolut ion data assimilation system. Q. J. R. Meteorol. Soc. ,

    doi

  10. [2013]

    URL http://dx.doi.org/10.3402/tellusa.v65i0.19546. L. M. Stewart, S. L. Dance, N. K. Nichols, J. R. Eyre, and J. Camer on. Estimating interchannel observation-error correlations of IASI radiance data in the Met O ffice system. Q. J. R. Meteorol. Soc., 140:1236–1244,

    doi

  11. [2014]

    doi: 10.1002/qj.2218. E. G. Pavelin, S. J. English, and J. R. Eyre. The assimilation of cloud-a ffected infrared satellite radiances for numerical weather prediction. Quarterly Journal of the Royal Meteorological Society , 134(632):737–749,

    doi

  12. [2016]

    doi: 10.1002/qj.2925. G. Desroziers, L. Berre, B. Chapnik, and P. Poli. Diagnosis of obser vation, background and analysis-error statistics in observation space. Q. J. R. Meteorol. Soc. , 131:3385–3396,

    doi

  13. [2017]

    doi: 10.1002/qj.3097

    ISSN 1477-870X. doi: 10.1002/qj.3097. URL http://dx.doi.org/10.1002/qj.3097. D. S. Bernstein. Matrix mathematics : theory, facts, and formulas . Princeton University Press, Princeton, N.J. ; Oxford, 2nd ed. edition,

    doi

  14. [2018]

    URL https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/qj.3395

    doi: 10.1002/qj.3395. URL https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/qj.3395. G. V. Bennitt, H. R. Johnson, P. P. Weston, J. Jones, and E. Pott iaux. An assessment of ground-based GNSS Zenith Total Delay observation errors and th eir correlations using the Met Office ...

    doi

  15. [2019]

    doi: https://doi.org/10.1002/qj.3592. L. M. Stewart. Correlated observation errors in data assimilation . PhD thesis, University of Reading,

    doi

Pith tools

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