pith. sign in

arxiv: 2606.31957 · v1 · pith:WHFW6TTWnew · submitted 2026-06-30 · 📊 stat.AP · physics.ao-ph· stat.ME

Locally stationary Argo ocean heat content estimates: Modeling, validation and uncertainty quantification

Thea Sukianto , Mikael Kuusela , Donata Giglio , Anirban Mondal , Pulong Ma , Douglas W. Nychka This is my paper

Pith reviewed 2026-07-01 02:08 UTC · model grok-4.3

classification 📊 stat.AP physics.ao-phstat.ME
keywords Argoocean heat contentGaussian processuncertainty quantificationspatio-temporal statisticslocally stationary processcross-validation
0
0 comments X

The pith

Argo ocean heat content can be mapped with tractable global estimates and correlated uncertainties by treating temperature profiles as locally stationary Gaussian processes.

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

The paper establishes an end-to-end statistical framework for turning Argo float temperature measurements into ocean heat content anomaly maps. It represents the vertically integrated profiles as a locally stationary Gaussian process over space and time, with decorrelation scales estimated directly from the data. This produces computationally feasible maps while local conditional simulation ensembles supply spatially and temporally correlated uncertainty fields. Cross-validation confirms that adding a climatological time trend to the mean and time to the covariance improves performance, and a new paired cross-validation checks the uncertainty calibration. The result is an open-source, modular codebase that delivers 2004-2022 maps together with downstream climatological quantities and their uncertainties.

Core claim

The framework models vertically integrated Argo temperature profiles as a locally stationary Gaussian process defined over space and time. This enables computationally tractable OHC anomaly maps based on data-driven decorrelation scales estimated from the Argo observations. Uncertainty is quantified using local conditional simulation ensembles that produce principled spatially and temporally correlated uncertainty quantification. The modeling choices are validated using statistical cross-validation demonstrating the importance of a climatological time trend in the mean field and accounting for time in the covariance function, together with a new paired cross-validation technique for the unce

What carries the argument

locally stationary Gaussian process over space and time with data-driven decorrelation scales, implemented via local conditional simulation ensembles for uncertainty

If this is right

  • Enables production of Argo-based OHC anomaly maps for 2004-2022 together with spatially and temporally correlated uncertainties.
  • Cross-validation establishes the necessity of a climatological time trend in the mean and time dependence in the covariance.
  • A new paired cross-validation technique directly validates the uncertainty quantification.
  • The open-source modular codebase supports reproducible extension to other periods or variables.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The correlated uncertainty fields could be propagated into trend estimates of Earth energy imbalance to obtain more defensible confidence intervals.
  • The local simulation approach may transfer to other sparse in-situ ocean datasets where spatial error correlation matters for integrated quantities.

Load-bearing premise

Vertically integrated Argo temperature profiles can be adequately represented by a locally stationary Gaussian process over space and time whose decorrelation scales can be reliably estimated from the observations themselves.

What would settle it

If cross-validation on held-out Argo profiles shows that the local stationarity assumption produces systematic misfit in decorrelation scales or that the conditional simulation ensembles fail to cover withheld observations at the nominal rate, the mapping and uncertainty framework would be falsified.

read the original abstract

Argo profiling floats measure seawater temperature and salinity in the upper 2000 meters of the ocean. These floats are uniquely capable of measuring the global Ocean Heat Content (OHC), a quantity that is of central importance for understanding Earth Energy Imbalance. Yet, producing Argo-based OHC estimates with reliable uncertainties is statistically challenging due to the complex structure and large size of the Argo dataset. Here we present an end-to-end mapping and uncertainty quantification framework for Argo-based OHC estimation using state-of-the-art methods from spatio-temporal statistics. The framework is based on modeling vertically integrated Argo temperature profiles as a locally stationary Gaussian process defined over space and time. This enables us to produce computationally tractable OHC anomaly maps based on data-driven decorrelation scales estimated from the Argo observations. Our modeling choices are validated using statistical cross-validation, which demonstrates the importance of including a climatological time trend in the mean field and accounting for time in the covariance function. We quantify the uncertainty of these maps using local conditional simulation ensembles, a novel approach that leads to principled spatially and temporally correlated uncertainty quantification. A new paired cross-validation technique is presented to validate these uncertainties. The mapping framework is implemented in an open-source codebase that is designed to be modular, reproducible and extensible. To demonstrate the mapping and uncertainty quantification capabilities of this approach, we present new Argo OHC maps with uncertainties for 2004-2022 and report on various downstream climatological estimates and their uncertainties.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 0 minor

Summary. The paper presents an end-to-end framework for Argo-based ocean heat content (OHC) anomaly mapping that models vertically integrated temperature profiles as a locally stationary Gaussian process over space and time. Decorrelation scales are estimated directly from the observations, a climatological time trend is included in the mean, and time is accounted for in the covariance. Uncertainty is quantified via local conditional simulation ensembles, with validation through cross-validation (including a new paired technique) and an open-source modular implementation. Results include OHC maps for 2004-2022 along with downstream climatological estimates.

Significance. If the local conditional simulation produces ensembles whose joint distribution preserves the claimed spatio-temporal correlations at all scales, the framework would offer a practical advance for uncertainty quantification on large Argo datasets while remaining computationally tractable. The data-driven estimation of decorrelation scales, explicit cross-validation of modeling choices, and open-source code are positive features that support reproducibility.

major comments (1)
  1. [Methods (local conditional simulation)] The abstract and methods description of local conditional simulation claim 'principled spatially and temporally correlated uncertainty quantification,' yet the locality of the windows raises the possibility that covariances between points separated by more than one window are zero or near-zero by construction unless neighboring windows are explicitly coupled through shared conditioning data or a global taper. This directly affects the central uncertainty claim and requires explicit verification (e.g., via ensemble covariance diagnostics across window boundaries).

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their constructive review and positive assessment of the framework. We address the single major comment below.

read point-by-point responses

  1. Referee: [Methods (local conditional simulation)] The abstract and methods description of local conditional simulation claim 'principled spatially and temporally correlated uncertainty quantification,' yet the locality of the windows raises the possibility that covariances between points separated by more than one window are zero or near-zero by construction unless neighboring windows are explicitly coupled through shared conditioning data or a global taper. This directly affects the central uncertainty claim and requires explicit verification (e.g., via ensemble covariance diagnostics across window boundaries).

    Authors: We appreciate the referee's careful attention to the correlation properties of the local conditional simulation. Our implementation uses overlapping windows with shared conditioning observations drawn from the full dataset within each local neighborhood; these shared data points explicitly couple neighboring windows and induce non-zero cross-boundary covariances. The resulting ensemble therefore preserves the claimed spatio-temporal correlations at the scales relevant to Argo sampling. To make this explicit, we will add ensemble covariance diagnostics across window boundaries (as suggested) to the revised manuscript. revision: yes

Circularity Check

0 steps flagged

No significant circularity; estimates and validation are data-driven

full rationale

The paper estimates decorrelation scales directly from Argo observations and validates both the mean field (including climatological trend) and the uncertainty quantification via statistical and paired cross-validation on held-out data. These steps ground the OHC maps and ensemble uncertainties in external observations rather than reducing them to fitted inputs or self-citations by construction. The local conditional simulation is presented as a tractable approximation whose correlation properties are empirically checked, with no load-bearing self-citation chains or self-definitional reductions visible in the derivation.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The approach rests on the standard Gaussian process modeling assumption and on parameters estimated directly from data. No new physical entities are postulated.

free parameters (1)
  • decorrelation scales
    Estimated from Argo observations and used in the covariance function of the locally stationary process.
axioms (1)
  • domain assumption Vertically integrated temperature profiles follow a Gaussian process with locally stationary covariance
    Core modeling choice stated in the abstract for the spatio-temporal field.

pith-pipeline@v0.9.1-grok · 5825 in / 1383 out tokens · 43867 ms · 2026-07-01T02:08:23.874203+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

12 extracted references · 8 canonical work pages · 1 internal anchor

  1. [1]

    Allison, L. C., C. D. Roberts, M. D. Palmer, L. Hermanson, R. E. Killick, N. A. Rayner, D. M. Smith, and M. B. Andrews, 2019: Towards quantifying uncertainty in ocean heat content changes using synthetic profiles.Environmental Research Letters,14 (8), 084

    2019

  2. [2]

    Argo, 2025: Argo data products

    Argo, 2023: Argo float data and metadata from Global Data Assembly Centre (Argo GDAC) - Snapshot of Argo GDAC of January 10th 2023.SEANOE, https://doi.org/10.17882/42182# 98916. Argo, 2025: Argo data products. URL https://argo.ucsd.edu/data/argo-data-products/. Bamston, A. G., M. Chelliah, and S. B. Goldenberg, 1997: Documentation of a highly ENSO- relate...

    work page doi:10.17882/42182 2023

  3. [3]

    Good, S. A., M. J. Martin, and N. A. Rayner, 2013: EN4: Quality controlled ocean temperature and salinity profiles and monthly objective analyses with uncertainty estimates.Journal of Geophysical Research: Oceans,118 (12), 6704–6716, https://doi.org/10.1002/2013JC009067. Gouretski, V., 2018: World Ocean Circulation Experiment – Argo Global Hydrographic Cl...

    work page doi:10.1002/2013jc009067 2013

  4. [4]

    State of the Climate in 2023

    Proceedings of the Sixth Valencia International Meeting, Oxford University Press, 761–768. 46 Hosoda, S., T. Ohira, and T. Nakamura, 2008: A monthly mean dataset of global oceanic tem- perature and salinity derived from Argo float observations.JAMSTEC Report on Research and Development,8, 47–59. Hughes, T. P., and Coauthors, 2017: Global warming and recur...

    work page doi:10.1017/9781009157896 2008

  5. [5]

    Meyssignac, B., and Coauthors, 2019: Measuring global ocean heat content to estimate the Earth energy imbalance.Frontiers in Marine Science,6, https://doi.org/10.3389/fmars.2019.00432. Meyssignac, B., and Coauthors, 2023: How accurate is accurate enough for measuring sea- level rise and variability.Nature Climate Change,13 (8), 796–803, https://doi.org/10...

    work page doi:10.3389/fmars.2019.00432 2019

  6. [6]

    Hammerling, M

    Nychka, D., D. Hammerling, M. Krock, and A. Wiens, 2018: Modeling and emulation of nonsta- tionary Gaussian fields.Spatial Statistics,28, 21–38, https://doi.org/10.1016/j.spasta.2018.08

    work page doi:10.1016/j.spasta.2018.08 2018

  7. [7]

    Quantum marginal inequalities and the conjectured entropic inequalities

    Park, B., M. Kuusela, D. Giglio, and A. Gray, 2023: Spatiotemporal local interpolation of global ocean heat transport using Argo floats: A debiased latent Gaussian process approach.The Annals of Applied Statistics,17 (2), 1491–1520, https://doi.org/10.1214/22-AOAS1679. Porcu, E., R. Furrer, and D. Nychka, 2021: 30 years of space–time covariance functions....

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1214/22-aoas1679 2023

  8. [8]

    Nature Climate Change,6 (2), 138–144

    von Schuckmann, K., and Coauthors, 2016: An imperative to monitor Earth’s energy imbalance. Nature Climate Change,6 (2), 138–144. von Schuckmann, K., and Coauthors, 2023: Heat stored in the Earth system 1960–2020: where does the energy go?Earth System Science Data,15 (4), 1675–1709, https://doi.org/10.5194/ essd-15-1675-2023. Walchessen, J., A. Lenzi, and...

    2016

  9. [9]

    Zammit-Mangion, R

    51 Walchessen, J., A. Zammit-Mangion, R. Huser, and M. Kuusela, 2025: Neural conditional simu- lation for complex spatial processes. arXiv:2508.20067 [stat.ME]. Wilson, A. G., Z. Hu, R. Salakhutdinov, and E. P. Xing, 2016: Deep kernel learning.Proceedings of the 19th International Conference on Artificial Intelligence and Statistics, A. Gretton, and C. C....

    work page doi:10.3389/fmars.2020.00700 2025

  10. [10]

    1290, WMO, Geneva, URL https://library.wmo.int/idurl/4/56300

    WMO-No. 1290, WMO, Geneva, URL https://library.wmo.int/idurl/4/56300. World Meteorological Organization, 2023:State of the Global Climate

    2023

  11. [11]

    1316, WMO, Geneva, URL https://library.wmo.int/idurl/4/66214

    WMO-No. 1316, WMO, Geneva, URL https://library.wmo.int/idurl/4/66214. World Meteorological Organization, 2024:State of the Global Climate

    2024

  12. [12]

    1347, WMO, Geneva, URL https://library.wmo.int/idurl/4/68835

    WMO-No. 1347, WMO, Geneva, URL https://library.wmo.int/idurl/4/68835. Yarger, D., S. Stoev, and T. Hsing, 2022: A functional-data approach to the Argo data.The Annals of Applied Statistics,16 (1), 216–246. Zhang, C., D. Wang, Z. Liu, S. Lu, C. Sun, Y. Wei, and M. Zhang, 2022: Global gridded Argo dataset based on gradient-dependent optimal interpolation.Jo...

    work page doi:10.3390/jmse10050650 2022