LETKF-based Ocean Research Analysis version 2.0 for a quasi-global domain (LORA-QG): Validation and intercomparison with eddy-permitting global ocean reanalysis datasets

Ohishi, Shun; Miyoshi, Takemasa; Kachi, Misako

doi:10.5194/egusphere-2026-2277

Preprints

https://doi.org/10.5194/egusphere-2026-2277

Preprints

23 Apr 2026

| 23 Apr 2026

LETKF-based Ocean Research Analysis version 2.0 for a quasi-global domain (LORA-QG): Validation and intercomparison with eddy-permitting global ocean reanalysis datasets

Shun Ohishi, Takemasa Miyoshi, and Misako Kachi

Abstract. We previously produced the local ensemble transform Kalman filter (LETKF)-based Ocean Research Analysis (LORA) version 1.0 datasets for the western North Pacific and Maritime Continent regions (LORA-WNP and LORA-MC, respectively) during the period from August 2015 to January 2024. However, these limited domains and periods constrain their applicability. Therefore, we developed a new eddy-permitting LETKF-based ocean data assimilation system and produced LORA version 2.0 for a quasi-global domain (LORA-QG) from June 2002, when the Advanced Microwave Scanning Radiometer (AMSR) series, a series of space-borne microwave imagers, began providing sea surface temperature observations and the Argo program substantially expanded in situ temperature and salinity measurements. We validated LORA-QG using observations from surface drifter buoys, tide gauges, and ocean climate stations, and compared the results with those of three eddy-permitting global ocean reanalysis datasets (GLORYS2V4, ORAS5, and C-GLORSv7). Although these observations are independent of LORA-QG, they may not be entirely independent of the other three reanalysis datasets. The validation results show that LORA-QG agrees well with the observations and has the second-highest accuracy among the four datasets in terms of overall root-mean-square deviations relative to the observations, thus achieving sufficient accuracy for geoscientific research and practical applications. LORA-QG provides features unavailable in conventional global reanalysis products, including ensemble-based uncertainty estimates and individual terms of the heat and salinity budget equations. These features make LORA-QG a valuable dataset for ensemble-based ocean forecasting and process-based studies. However, room for improvement remains, as LORA-QG exhibits significant warm biases in the tropics, particularly in the western tropical Pacific, and its sea surface salinity representation is likely limited due to relatively strong salinity nudging toward a climatological dataset in the mixed layer.

Received: 21 Apr 2026 – Discussion started: 23 Apr 2026

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Shun Ohishi, Takemasa Miyoshi, and Misako Kachi

Status: final response (author comments only)

RC1:
'Comment on egusphere-2026-2277', Anonymous Referee #1, 03 Jun 2026
Review
The manuscript provides a nice description and assessment of the new LORA-QG reanalysis in comparison with other state-of-the-science eddy permitting reanalyses, positioning LORA-QG in the envelope of these and providing a summary of strengths and weaknesses. The manuscript is also well written and clear, and I also liked the efforts in summarizing the results (e.g., Figure 19 and the text across the manuscript).
I recommend the manuscript for publication after a few points are better framed and discussed, which likely require some more discussion (and maybe a few additional analyses).
Two main points
The argument that the reanalysis provides uncertainty estimates appears in various parts of the manuscript (notably abstract and conclusions), and it is unique to the ensemble DA scheme adopted, but is not really validated, especially because the reanalysis proves under-dispersive. I suggest answering the questions (either through a discussion, and if possible, corroborated by results): is the choice of the (quite large) ensemble size (>100) really required, adding signal to the spread? Even though the reanalysis is under-dispersive, is the temporal evolution of the spread consistent with that of the RMSE (namely, net of the systematic offset), for some diagnostics (e.g., correlation of spread and RMS in Fig. 11)? Can the authors suggest reasons/ways to improve the under-dispersion despite the large ensemble size (i.e., lack of stochastic physics? Suboptimal inflation scheme in the LETKF?)

While bias/RMS metrics may be, in general, difficult to explain in detail, some more efforts could be made to explain the large-scale biases (i.e., in the comparison with climatology), at least at speculation level. For instance, large surface warm biases in the tropics (could this be ascribed to the use of microwave SST, i.e., assimilating subskin SST, while all other reanalyses rely on skin/infrared SST?) or perhaps the warm/salty biases at depth (maybe depend on Antarctic bottom water formation that is limited by the global domain extension and the use of LBCs from SODA?)

Minor points
Abstract L22 and later in the text (L170): I think you should be more precise about the dependence/independence of the validating dataset. I think the reanalyses you included do not assimilate any drifter/tide gauge data; the climate station data dependence may depend on whether these are included in the in-situ dataset assimilated (EN4, CORA), but, if possible, this needs to be clarified.

I think the short summary should be self-contained, so I suggest rewriting it to avoid wording like “the second-most” etc., which assumes the reader to already know the work

Tab 1: Please check, I think SODA also released an eddy-resolving version recently

L58 Adding a concise explanation of advantages (all weather) and disadvantages (resolution) compared to infrared SST could be good for non-experts

L68, maybe say already there what you mean by quasi-global?

L125 I am skeptical that you improve computational efficiency by converting routines from Fortran 77 to Fortran 90, given that compilers are typically the same. Did you really obtain more efficiency, or did you mean something else?

L132 not clear, may you mean “to prevent ensemble filter convergence/collapse”?

L139 is only OpenMP parallel, or the OpenMP tiles MPI parallelization

The validation description seems to be missing if you use the ensemble mean for your reanalyses, and the temporal frequency (daily, monthly, in some cases?) for all validated reanalyses

All other sections and analyses are well structured and clear for me (besides the need to try to understand the reason for systematic large-scale biases).
Citation: https://doi.org/10.5194/egusphere-2026-2277-RC1
RC2:
'Comment on egusphere-2026-2277', Anonymous Referee #2, 22 Jul 2026
Review of "LETKF-based Ocean Research Analysis version 2.0 for a quasi-global domain (LORA-QG): Validation and intercomparison with eddy-permitting global ocean reanalysis datasets" by Ohishi et al.
This manuscript presents the development of the LETKF-based Ocean Research Analysis version 2.0 (LORA-QG), a quasi-global ocean reanalysis covering the period from June 2002 to January 2024. The product is validated against existing ocean reanalyses, including C-GLORSv7, GLORYS2V4, and ORAS5, and evaluated using drifter observations of sea surface temperature (SST) and surface currents, as well as temperature and salinity observations from the KEO and Papa mooring stations.
The development of a new ocean reanalysis is an important contribution to the ocean modelling and data assimilation community. However, in its current form, the manuscript does not clearly demonstrate the scientific advances or added value of LORA-QG relative to existing state-of-the-art ocean reanalysis products. The validation is limited, the diagnostics are not sufficiently comprehensive, and the discussion does not adequately explain several notable differences between LORA-QG and the other products. Consequently, the manuscript does not yet provide sufficient evidence of the advantages or unique capabilities of the proposed reanalysis. I therefore do not recommend publication in its present form.
The following comments are intended to help improve the manuscript.
Major Comments
Ocean reanalysis datasets (Table 2)

Table 2 is incomplete and not up to date. Several widely used operational and research ocean reanalysis products are missing, including:
GloSea5 (UK Met Office)

NCEP Global Ocean Data Assimilation System (NCEP-GODAS, NOAA/NCEP, USA)

IGORA v1 (INCOIS Global Ocean Reanalysis Version 1, India)

Other recent operational ocean reanalyses listed at:

https://reanalyses.org/ocean/overview-current-reanalyses

Therefore, the statement that the table represents the "best of the authors' knowledge" is not justified and should be revised.
Ensemble size

The rationale for selecting 128 ensemble members is not explained. Was a sensitivity analysis performed? Why was this number chosen instead of 64, 96, or 256 members? Please justify this choice and discuss its impact on analysis quality and computational cost.
Mean Dynamic Ocean Topography

The manuscript states that the mean dynamic ocean topography (MDOT) is estimated from simulated sea surface height averaged over 1997–2001. Please discuss the reliability of this approach. Why was a model-derived MDOT used instead of an observation-based product such as CNES-CLS MDT?
Surface restoring

The manuscript clearly states that sea surface salinity (SSS) is relaxed toward observations. However, it is unclear whether SST is also restored or relaxed during the model integration. Please clarify the surface boundary conditions applied to SST.
Large subsurface bias in the equatorial region

Figure 4 shows considerably larger temperature and salinity biases in the equatorial ocean compared to the other reanalysis products. The explanation provided (Lines 223–225) appears insufficient.
Please discuss whether these larger biases could also result from:
the relatively short model spin-up period,

deficiencies in the data assimilation configuration,

model physics,

vertical mixing parameterization, or

sparse subsurface observations.

A more comprehensive discussion is needed.
Inconsistency between text and figure

The discussion in Lines 257–259 (Page 18) does not appear to agree with the results shown in the corresponding figure. Please revise either the figure interpretation or the text.
Regional differences in RMSD

The manuscript states:
"The RMSDs of LORA-QG are significantly larger than those of the other three datasets in the tropics, especially in the western tropical Pacific, while they are significantly smaller around the western boundary currents and the Antarctic Circumpolar Current."
This is an important result but is not adequately explained. Please discuss the physical or methodological reasons behind these contrasting regional performances.
Accuracy of drifter observations

Please provide the accuracy and uncertainty of the drifter observations used for validating SST, zonal current, and meridional current. This information is necessary to assess the significance of the reported differences.
Sea Level Anomaly (SLA)

The methodology for computing SLA is not adequately described.
Please clarify:
How was SLA calculated for LORA-QG?

Were identical reference periods used for all reanalysis products?

Were all datasets processed consistently before comparison?

In addition, spatial maps alone are insufficient. Time-series comparisons at representative locations should be included together with statistical metrics such as:
Mean

Standard deviation

Bias

RMSE

Correlation coefficient

KEO and Papa buoy validation

Very little information is provided regarding the KEO and Papa buoy observations.
Please include:
observation period,

sampling frequency,

number of available profiles,

quality control procedures.

The validation should also include comparisons of:
mean temperature profiles,

mean salinity profiles,

standard deviation,

correlation coefficient,

Evaluation over climatically important regions

The manuscript would be considerably strengthened by evaluating the product over major climate-sensitive regions.
Examples include:
Niño-3.4 region (Pacific Ocean),

Thermocline Ridge/Dome region (Indian Ocean),

western tropical Pacific warm pool.

Time-series comparisons with drifter observations or independent datasets should be presented for SST and surface currents, together with statistical measures (mean, standard deviation, bias, RMSE, and correlation coefficient).
In addition, it would be valuable to demonstrate how well LORA-QG reproduces important climate indices such as the Niño-3.4 SST index.
Subsurface validation using EN4

The manuscript evaluates subsurface temperature and salinity only against observations from two mooring sites.
A more comprehensive assessment should also include comparisons with widely used objective analysis products such as EN4 over climatically important regions. This would provide a more robust evaluation of the basin-scale performance of LORA-QG and facilitate comparison with existing ocean reanalysis products.
Overall Recommendation
The manuscript describes the development of a new ocean reanalysis product with a long analysis period, which is potentially valuable for the oceanographic community. However, the current validation is insufficient to demonstrate a clear improvement over existing operational reanalysis products. More comprehensive validation, improved diagnostics, deeper scientific interpretation, and clearer justification of the methodology are necessary before the manuscript can be considered for publication.
Citation: https://doi.org/10.5194/egusphere-2026-2277-RC2

Shun Ohishi, Takemasa Miyoshi, and Misako Kachi

Viewed

Total article views: 324 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
205	102	17	324	26	19

HTML: 205
PDF: 102
XML: 17
Total: 324
BibTeX: 26
EndNote: 19

Views and downloads (calculated since 23 Apr 2026)

Month	HTML	PDF	XML	Total
Apr 2026	91	33	9	133
May 2026	57	46	2	105
Jun 2026	16	5	2	23
Jul 2026	39	18	4	61
Aug 2026	2	0	2

Cumulative views and downloads (calculated since 23 Apr 2026)

Month	HTML	PDF	XML	Total
Apr 2026	91	33	9	133
May 2026	57	46	2	105
Jun 2026	16	5	2	23
Jul 2026	39	18	4	61
Aug 2026	2	0	2

Viewed (geographical distribution)

Total article views: 310 (including HTML, PDF, and XML) Thereof 310 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 04 Aug 2026

Short summary

We developed a new eddy-permitting local ensemble Kalman filter (LETKF)-based Ocean Research Analysis (LORA) version 2.0 for a quasi-global domain (LORA-QG) from June 2002. Validation results show that LORA-QG is the second-most accurate among the four quasi-global and global ocean analysis datasets and has sufficient accuracy for scientific and practical applications.


Total:	0
HTML:	0
PDF:	0
XML:	0