The European Centre for Medium-Range Weather Forecasts (ECMWF) Reanalysis version 5 (ERA5) is one of the most widely used atmospheric reanalysis datasets provided by the ECMWF. However, the transition in the sea ice dataset between 1978 and 1979 may introduce inconsistencies that affect various surface meteorological variables. This study evaluated ERA5 sea ice data in the Sea of Okhotsk (SO), focusing on the years 1978 and 1979. Furthermore, the impact of this sea ice transition on meteorological variables was evaluated. In 1979, ERA5 sea ice coverage reached the coastal areas of Hokkaido in the southern SO, with an extent of approximately 1.1 × 106 km². In contrast, an unrealistically low sea ice cover of approximately 0.5 × 106 km² was observed before 1978. This discontinuity in sea ice stems primarily from issues with assimilated sea ice data used in ERA5. In 1978, the unrealistic negative bias in sea ice cover arguably contributed to positive biases in significant wave height, sea surface temperature, surface air temperature, and surface winds. In the case of wave observations, from 1975 to 1978, ERA5 overestimated significant wave height by more than 60% compared to observations from February to April when sea ice was present. These findings highlight the need for caution when analyzing long-term changes in ice-covered areas when using ERA5 data, particularly for periods before 1979.
ERA5ocean surface wavesea iceSea of Okhotsksurface meteorological variablesThe author(s) declared that financial support was not received for this work and/or its publication.section-at-acceptanceCoastal Ocean ProcessesIntroduction
The Sea of Okhotsk (SO) is a seasonal sea-ice region, where sea ice forms from November to March before retreating. It is recognized as the world’s lowest-latitude sea-ice area (Figures 1A–G). Although sea ice in the SO exhibits significant interannual variability, its maximum annual extent has decreased by 3.4% per decade since the 1970s when considered over the long term1. Furthermore, sea ice is projected to continue decreasing in the future (Narita and Iwasaki, 2026). Sea ice is considered to act as a natural breakwater floating on the sea due to its wave-attenuating effect (Iwasaki, 2022). Recent studies have revealed that the reduction in sea ice in the SO has led to an increase in wave power in this region during winter, at a rate of 12–15% per decade (Iwasaki, 2023). Waves are not only indicators of climate but also critical factors influencing interactions between the atmosphere and the ocean at a climatic scale. Understanding long-term trends in sea surface waves is crucial for the effective management of coastal areas and the safety and efficiency of offshore operations. In particular, in the modern era, where the impacts of climate change are becoming increasingly pronounced, long-term changes in sea surface waves are receiving increasing attention.
(A–G) Monthly average sea ice concentration (%) from December to June based on 46 years (1979–2024) of ERA5 data, with contours indicating sea ice concentration at 10% intervals. (H) is an enlarged view of the red square in (G), showing the location of the NOWPHAS Monbetsu Port buoy (44.416°N, 143.432°E) marked by a red circle.
Seven regional maps (A–G) depict monthly sea ice concentration percentages from December to June, using blue gradients and contour lines for percentages. Panel H shows a detailed map highlighting Monbetsu Area in red with coordinates.
The role of reanalysis in climate monitoring applications is widely recognized (e.g., Keller and Wahl, 2021; Zandler et al., 2019). Among the widely used reanalysis datasets is the European Centre for Medium-Range Weather Forecasts (ECMWF) Reanalysis version 5 (ERA5), a next-generation product from ECMWF (Hersbach et al., 2020). Reanalyses provide consistent historical climate data by applying a fixed data assimilation system to all available observations, enabling the assessment of observational network changes, validation of model improvements, and evaluation of forecast skill against a homogeneous reference.
Studies on long-term wave characteristics using ERA5 have been conducted for various marine areas, including the coastal regions of China (Du et al., 2024), India (Anusree and Sanil Kumar, 2024), the Eastern Tropical Atlantic (Omonigbehin et al., 2024), the Mediterranean Sea (Barbariol et al., 2021), the South China Sea (Liu et al., 2022), coastal waters around Japan (Tanaka et al., 2023), and global ocean (Erikson et al., 2022; Patra et al., 2020). However, to date, no studies have focused on long-term wave changes in SO using ERA5. Furthermore, since ERA5 provides data from 1940 onward, it has the advantage of enabling wave analyses spanning more than 80 years. Iwasaki (2023) examined a 40-year period starting in the 1980s, leaving trends before that period unresolved.
When analyzing long-term wave fluctuations in sea-ice regions, such as the SO, using ERA5, the following concerns arise. Performing wave calculations for sea-ice regions requires both surface winds and sea-ice data. Although ERA5 uses sea-ice data as the sea-surface boundary condition, the data source varies by period (Hirahara et al., 2016). The earlier period (January 1940–December 1978) relies solely on HadISST2 (Titchner and Rayner, 2014). For the latter period (January 1979–August 2007), data created by the Ocean and Sea Ice Satellite Application Facilities (OSI-SAF) of the European Organisation for the Exploitation of Meteorological Satellites (OSI-409-a) are introduced (Eastwood et al., 2014). Due to differences in factors such as satellite data usage, the spatiotemporal resolution differs significantly between HadISST2 and OSI-409-a (HadISST2: monthly, 1°; OSI-409-a: daily, 10 km). Switching between these datasets raises concerns about introducing artificial discontinuities in sea ice, which could, in turn, affect wave fields that use sea ice as a surface boundary condition. Furthermore, sea ice is thought to influence not only wave fields but also various meteorological variables such as sea surface temperature, surface air temperature, and surface winds. Since the period before 1979 predates the satellite observation era, limited data is available for this timeframe. Therefore, results based on pre-satellite observation data should be interpreted with caution.
This study analyzes long-term changes in sea ice in the SO using ERA5 data from 1940 onward, with a particular focus on the 1978–1979 period, when the sea ice dataset was changed. Since ERA5 uses different sea-ice datasets around 1979, this period represents a key turning point for verifying the consistency of the reanalysis data. In addition, to evaluate whether the switch in sea ice data affects the reliability of wave fields, we compared ERA5 wave fields with observational data from the Nationwide Port and Coastal Wave Information Network (NOWPHAS). Furthermore, in addition to the impact of sea ice on wave fields, we also evaluated its contribution to sea surface temperature, surface air temperature, and surface winds.
Methods
In this study, we used ERA5 data for the entire SO (135°–165°E, 42°–63°N; Entire region in Figures 1A–G). The dataset includes sea-ice concentration, significant wave height (SWH) encompassing both wind waves and swells, mean wave period (MWP), sea surface temperature, 2-m air temperature, and surface winds (10-m winds). The data cover 85 years, from 1940 to 2024, with a temporal resolution of monthly averages. The spatial resolution is 0.5° × 0.5°for wave fields and 0.25° × 0.25°for other meteorological variables. However, for SWH and MWP, hourly data were used when comparing with NOWPHAS wave observations. The ERA5 wave parameters, SWH (Hs) and MWP (Tm), were determined using the following equations:
Hs=4m0Tm=m0m2
In Equations 1 and 2, mn is the moment of order n of the wave spectrum F(f, θ), which describes the distribution of wave energy as a function of frequency f and propagation direction θ. The mean period used in this study is based on the second moment of the wave spectrum.
The ERA5 wave model calculates wave generation and propagation only when sea-ice concentration is below 30%, solving wind-generated waves using a discrete spectrum of 24 directions × 30 frequencies. However, wave–sea-ice interactions are not explicitly parameterized (Hošeková et al., 2021). Therefore, when the sea ice concentration is ≥ 30%, wave field calculations are not performed.
Furthermore, for the validation of ERA5 wave data, observational data from NOWPHAS, operated by the Ministry of Land, Infrastructure, Transport and Tourism’s Port and Harbor Bureau, were used, particularly data from coastal areas in northern Hokkaido (Figure 1H). In this study, we used two-hourly SWH and MWP data obtained from NOWPHAS, which were calculated based on zero-crossing analysis. The observation is located at 44.416°N, 143.432°E, with an observation depth of 52 m and a distance from the coast of 9,590 m. Observational data for 27 years, from 1975 to 2001, were used. However, because this study focused on the period from 1978 to 1979, the analysis primarily used data from the preceding and following years, specifically from 1975 to 1982. Although the observational records do not cover the entire long-term period, data up to 2001 are available. Therefore, data up to 2001 were used for trend comparisons with ERA5. The observational data were measured using a bottom-mounted ultrasonic wave (USW) gauge, a widely employed instrument (jointly developed by the Kaijo Co. and PHRI in the1960s), which determines wave surface elevation precisely by measuring the time for an ultrasonic wave to travel between the seabed and water surface. Only normal data with flag “0” were used for the NOWPHAS observation data. Wave observation data are also available at a site located just south (44.318°N, 143.607°E) of the study observation point. However, this data covers the period after 2000 and therefore fall outside the focus period of this study and was not used.
As noted in the Introduction, the data sources for ERA5 sea ice differ around 1979. Therefore, to obtain the assimilated data for ERA5 sea-ice, HadISST version 2.1.0.0 (Titchner and Rayner, 2014) was used. The data period is 1978 and 1979, before and after the addition of OSI-409-a. Additionally, although not incorporated in ERA5, we used the latest update of HadISST for the same period (1978 and 1979), HadISST version 2.2.2.0 (https://www.metoffice.gov.uk/hadobs/hadisst2/data/download.html).
ResultsSea ice
Long-term changes in sea-ice area (1940–2024) based on ERA5 data showed that the 1978 sea-ice area was approximately 0.5×106 km², while in 1979 it increased to approximately 1.1×106 km², more than double (Figure 2A). Furthermore, the amplitude of interannual variations before 1979 was significantly smaller compared to that observed thereafter.
(A) Average time series of sea-ice area (blue line) and 2-m air temperature (red line) over the Sea of Okhotsk (135°E–165°E, 42°N–63°N) during winter (January−April), derived from ERA5. The period spans 1940–2024. The black vertical line indicates the year 1979. The correlation coefficient between sea ice and 2-m air temperature from 1940 to 1978 (1979 to 2024) are shown in the upper left (right) corner. The bold value denotes significant at a 99% confidence level. (B–D) Spatial distribution (%) of average sea-ice concentration during winter (January–April) based on ERA5. (B) 1940–1978, (C) 1979–2024, and (D) difference between the two periods (B, C).
Panel A shows a line graph with annual sea ice area in blue and two-meter air temperature in red from 1940 to 2020, with correlation values of negative zero point three six and negative zero point eight. Panel B is a map depicting sea ice concentration percentages across a region, shaded in blue gradients. Panel C is a similar map with different sea ice distributions by percentage. Panel D displays the difference in sea ice concentration between the two periods, using blue for decreases and red for increases, with values labeled in contours.
Sea-ice retreat has been reported to promote near-surface warming, and warming drives ice retreat (i.e., two-way), whereas sea-ice expansion leads to temperature decreases (Screen and Simmonds, 2010; Olonscheck et al., 2019). In the winter Arctic, a significant negative correlation exists between sea-ice extent and 2-m air temperature, with increases in surface temperature attributed to changes in sea-ice concentration (Huo et al., 2025). Indeed, the correlation coefficient between sea-ice extent and 2-m air temperature was −0.36 from 1940 to 1978, while from 1979 to 2024, it was −0.80 (significant with 99% confidence level). This indicates a strong correlation between sea-ice extent and 2-m air temperature after 1979, whereas the correlation was weak before 1979. As noted in the Introduction, this change is likely attributable to the change in the ERA5 sea-ice data in 1979.
Focusing on the spatial distribution of sea ice, in 1978, sea ice was nearly absent south of 54°N (especially in eastern Sakhalin) (Figure 2B). In contrast, in 1979, sea ice appeared extensively in the eastern waters off Sakhalin, reaching as far as the coastal areas of Hokkaido (Figure 2C). Comparing the two periods (Figures 2B, C), a difference exceeding 50% is observed south of 54°N (Figure 2D). According to observation reports from the Japan Meteorological Agency, drift ice was confirmed along the Monbetsu coast of Hokkaido even before 1979. Furthermore, drift ice has been reported over extended periods not only at Monbetsu but also at multiple coastal locations in northern Hokkaido (Toyoda et al., 2022). Additionally, a difference of −40% is observed north of 50°N along the western Kamchatka Peninsula (Figure 2D).
As expected, examination of the February sea-ice concentration for 1978 and 1979 in ERA5 showed that in 1979, when OSI-409-a was introduced, the effective resolution was higher, and sea ice extended south of 54°N compared to that of the previous year (Supplementary Figures 1A, B). This result is consistent with the findings presented in Figure 2. However, regardless of the presence of OSI-409-a, gaps in sea ice also occurred south of 54°N in HadISST version 2.1.0.0 (Supplementary Figures 1C, D). Therefore, the increase in sea-ice area in 1979 is attributed to consistency issues in the HadISST version 2.1.0.0 data rather than to the introduction of OSI-409-a.
In recent years, HadISST version 2.2.0.0 has been released as an update to HadISST version 2.1.0.0. This dataset incorporates several improvements in the reanalysis of sea ice and sea surface temperature. Specifically, it assimilates higher-resolution satellite observations and reevaluates ship and buoy data compared to those of previous versions, and applies an improved interpolation method that accounts for spatiotemporal correlation structures when filling missing data. Based on this dataset, sea ice exists south of 54°N even in 1978, and no significant discrepancies are observed in either 1978 or 1979 (Supplementary Figures 1E, F).
Wave fields
Such discontinuities in sea ice are also expected to affect wave fields (SWH and MWP) when calculated using this data as input. A comparison of SWH between NOWPHAS and ERA5 is presented in Figure 3. In this comparison, ERA5 data from the point closest to the NOWPHAS location were used. To examine the impact of discontinuities in sea ice, we compared the monthly averages by dividing the periods into before 1978 (1975–1978) and after 1979 (1979–1982). In this comparison, only periods with data available from both NOWPHAS and ERA5 were used.
Comparison of SWH between ERA5 (blue line) and NOWPHAS Mombetsu Port observations (red line). The relative error is defined as (Xs−Xo)Xo×100, where Xs represents the ERA5 data (SWH), and Xo represents the observed SWH from NOWPHAS. The periods considered are (A) 1975–1978 and (B) 1979–1982. Only times when both NOWPHAS and ERA5 data are available are included in the comparison.
Two-panel figure with monthly data comparing significant wave height (SWH, green bars, left y-axis) and relative error percentages (solid red and dashed blue lines, right y-axis) from July to June; panel A and panel B show differing seasonal patterns and error trends.
From these results, irrespective of the period, it can be confirmed that ERA5 SWH reproduces the seasonal variation observed in the measurements, with lower values in summer and higher values in winter (Figure 3). However, when investigating the period before 1978, between February and April, when sea ice existed around the observation area, ERA5 overestimated SWH, with relative errors reaching over 60% (Figure 3A). In contrast, from 1979 onward, the overestimation of ERA5 SWH between February and April was minor (Figure 3B).
As expected, the overestimation of ERA5 before 1978 is considered to result from the absence of sea ice in the southern SO. Examination of monthly data acquisition rates from NOWPHAS observations shows a low acquisition rate from February to April in both 1978 and 1979, due to the influence of sea ice (Supplementary Figure 2A); the data acquisition rate denotes the sampling rate of data (i.e., SWH) per month, and is computed as N2h/Xfull×100, where N2h(Xfull) denotes the number of 2-hourly data for observation or ERA5 (their number if there are no missing values). However, examination of the ERA5 data acquisition rates shows that from 1979 to 1982, similar to NOWPHAS observations, the acquisition rate decreases during winter. In contrast, from 1975 to 1978, the data acquisition rate remains at 100% throughout winter (Supplementary Figure 2B), likely because the influence of sea ice is not accounted for. As shown in Supplementary Figures 2A, the observation rate of NOWPHAS is low during the winter season. Rather than representing a four-year average, the data acquisition rate for NOWPHAS varies from year to year (Supplementary Figures 3A, B). A similar pattern is observed for ERA5 data after 1979 (Supplementary Figures 3C, D). Therefore, the results shown in Figure 3 reflect the data from 1976 for the earlier period (1975–1978) and from the data of 1979 and 1982 for the later period (1979–1982) (Supplementary Figure 4).
The impact of sea ice affects not only SWH but also MWP (Supplementary Figure 5). Sea ice contributes to the reduction of MWP as well as the attenuation of SWH (Iwasaki, 2022; Iwasaki and Otsuka, 2021). From 1975 to 1978, before 1979, ERA5’s MWP exceeded the observed values from February to April, with a relative error of over 15% (Supplementary Figure 5A). In contrast, during the period 1979–1982, ERA5’s MWP closely matched the observed MWP (Supplementary Figure 5B).
The spatial distribution of SWH and MWP during the winters (January to April) of 1978 and 1979 is shown in Supplementary Figure 6. In 1978, with sea ice limited to north of 54°N, few areas were missing data for SWH and MWP (Supplementary Figures 6A, D). However, in 1979, as the sea ice extended to the eastern part of Sakhalin, the missing areas for SWH and MWP expanded (Supplementary Figures 6B, E). Differences between the two years show that, in the largest affected areas, SWH reached 0.6 m, and MWP reached 1.0 s (Supplementary Figures 6C, F).
Discussion
In this study, we examined the discontinuity in ERA5 sea ice representation in the Sea of Okhotsk around 1978–1979 and assessed its impacts on wave variables. We found that the effects of sea ice discontinuity are spreading to sea surface temperature, surface air temperature, and surface winds (Figure 4). The extreme increase in sea ice from 1978 to 1979 was reflected not only in lower sea surface and air temperatures but also in reduced surface winds (Figure 4). This result indicates that the unrealistic negative bias in sea ice in 1978 imported a positive bias to sea surface temperature, surface air temperature, and surface winds. The reductions in surface winds are thought to result from reduced sea ice, causing a decrease in local atmospheric stability (i.e., more unstable). Therefore, sea ice and surface wind have a negative correlation (e.g., Alkama et al., 2020; Iwasaki, 2022). Indeed, the correlation coefficient between sea-ice extent and surface wind was 0.00 from 1940 to 1978, while from 1979 to 2024, it was −0.38 (Supplementary Figure 7).
Spatial distribution of [left panels; (A−C)] sea surface temperature, [middle panels; (D−F)] 2-m air temperature, and [right panels; (G−I)] surface winds for ERA5 during winter January−April) of (upper panels) 1978 and (middle panels) 1979, and (lower panels) their difference. Spatial distribution of the difference between 1979 and 1978 (i.e., 1979−1978).
Nine-panel scientific figure compares sea surface temperature, two-meter air temperature, and surface wind speed around a peninsula region, each column showing a different variable. Top and middle rows present absolute values for two different periods, and the bottom row shows differences between periods. Each map features color gradients, contour lines, latitudinal and longitudinal coordinates, and corresponding color bars for measurement units.
Based on the above results, caution is required when analyzing sea ice, wave fields (SWH and MWP), and meteorological variables (sea surface temperature, 2-m air temperature, and surface winds) using ERA5 data for the SO before 1979. For example, Supplementary Figure 8 shows the long-term trend in SWH and surface winds for two periods (1940–2024 and 1979–2024). For 1940–2024, SWH exhibited an overall positive trend, except for a slightly negative trend in eastern Sakhalin (Supplementary Figures 8A, color). Although not statistically significant, this negative trend coincided with the increase in sea-ice area in eastern Sakhalin (Supplementary Figures 8A, color and contour map). In contrast, for 1979–2024, SWH exhibited an overall positive trend (particularly pronounced in the northeast), and sea ice also showed a decreasing trend (Supplementary Figure 8B). The negative SWH trend observed in eastern Sakhalin for 1940–2024 is likely a spurious result caused by missing sea-ice data before 1979. In addition, the extreme increase in sea ice in eastern Sakhalin in 1979 also affected the decrease in surface wind (Supplementary Figures 8C, D). This observation indicates that an increase in sea ice and a decrease in surface wind have an indirect influence on the reduction in SWH. However, even if the sea ice data were accurate throughout the entire period, the trends for the two periods would not necessarily coincide. Despite the limited observation period from 1975 to 2001, a comparison with observations was conducted. The results show that excluding data before 1979 brings the ERA5 trend values into closer alignment with the observed trend (–0.14 m/decade for observation and –0.15 m/decade for ERA5) (Supplementary Figure 9).
As shown in the results, the discontinuity in Sea ICE is improved by the update to HadISST. In the future, incorporating HadISST v2.2.2.0 for sea-ice concentration in ERA5 could help reduce unrealistic sea-ice distributions and improve the associated wave fields and meteorological variables such as sea surface temperature, surface air temperature, and surface winds. Additionally, although the present study focused on the SO as one of the sea-ice areas, caution is warranted because gaps in sea-ice coverage may also occur in other regions. Furthermore, the data used for this ERA5 validation was limited to a single coastal location in northern Hokkaido. Given the ongoing concern about increased wave activity in the SO, associated with sea ice reduction, further expansion of long-term wave observations utilizing systems such as NOWPHAS and small GPS buoys (e.g., Shimura et al., 2022) is necessary. In addition, although this study focused on wave bias caused by sea ice, it should be noted that the coarse spatial resolution of ERA5 may also contribute the observed bias when compared with single-point observations in coastal areas. Our findings highlight the need for caution when conducting long-term analyses of sea-ice variability using ERA5 data. Moreover, while this study focused on ERA5, the lack of satellite data before 1979 is a common issue with other reanalysis data. Therefore, regardless of ERA5, extreme caution is required when using reanalysis data for periods before 1979.
Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.
This work was conducted as part of Theme 4 of the Advanced Studies of Climate Change Projection (SENTAN Program), Grant Number JPMXD0722678534, supported by the Ministry of Education, Culture, Sports, Science, and Technology (MEXT), Japan. We thank the editor and reviewers for their constructive comments.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Supplementary material
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmars.2026.1760724/full#supplementary-material
ReferencesAlkamaR.KoffiE. N.VavrusS. J.DiehlT.FrancisJ. A.StroeveJ.. (2020).
Wind amplifies the polar sea ice retreat. Environ. Res. Lett.15, 124022. doi: 10.1088/1748-9326/abc379AnusreeA.Sanil KumarV. (2024).
Mean wave direction and wave height in the ERA5 reanalysis dataset: comparison with measured data in the coastal waters of India. Dyn. Atmos. Oceans107, 101478. doi: 10.1016/j.dynatmoce.2024.101478, PMID: 41735180BarbariolF.DavisonS.FalcieriF. M.FerrettiR.RicchiA.SclavoM.. (2021).
Wind waves in the Mediterranean Sea: an ERA5 reanalysis wind-based climatology. Front. Mar. Sci.8. doi: 10.3389/fmars.2021.760614, PMID: 41728591DuW.ZhangX.ShiH.LiG.ZhouZ.YouZ.. (2024).
Long-term extreme wave characteristics in the water adjacent to China based on ERA5 reanalysis data. J. Ocean Univ. China23, 1–10. doi: 10.1007/s11802-024-5446-y, PMID: 41732346EastwoodS.LavergneT.TonboeR. (2014).
Algorithm Theoretical Basis Document for the OSI SAF Global Reprocessed Sea Ice Concentration Product OSI-409, OSI-409-a, OSI-430. Version 1.1 (
EUMETSAT OSI SAF). Available online at: https://osisaf-hl.met.no/sites/osisaf-hl/files/baseline_document/osisaf_cdop2_ss2_atbd_sea-ice-conc-reproc_v1p1.pdf (Accessed November 17, 2025).
EriksonL.MorimJ.HemerM.YoungI.WangX.-L.MentaschiL.. (2022).
Global ocean wave fields show consistent regional trends between 1980 and 2014 in a multi-product ensemble. Commun. Earth Environ.3, 320. doi: 10.1038/s43247-022-00654-9, PMID: 41725453HersbachH.BellB.BerrisfordP.HiraharaS.HorányiA.Muñoz-SabaterJ.. (2020).
The ERA5 global reanalysis. Q. J. R. Meteorol. Soc146, 1999–2049. doi: 10.1002/qj.3803, PMID: 41735180HiraharaS.BalmasedaM. A.de BoissesonE.HersbachH. (2016). Sea surface temperature and sea ice concentration for ERA5. ERA Report Series 26 (
ECMWF). Available online at: https://www.ecmwf.int/en/elibrary/79624-sea-surface-temperature-and-sea-ice-concentration-era5 (Accessed November 17, 2025)
HošekováL.EidamE.PanteleevG.RainvilleL.RogersW. E.ThomsonJ. (2021).
Landfast ice and coastal wave exposure in northern Alaska. Geophys. Res. Lett.48, e2021GL095103. doi: 10.1029/2021GL095103, PMID: 40890438HuoY.ZhangR.WangH.SweeneyA.FuQ.RaschP. J.. (2025).
Changes in sea ice concentration explain half of the winter warming of the Arctic surface. Commun. Earth Environ.6, 775. doi: 10.1038/s43247-025-02548-y, PMID: 41725453IwasakiS. (2022).
Role of sea ice on winter wave power and its interannual variability in the Sea of Okhotsk: natural breakwater modified by surface wind changes. Prog. Oceanogr.210, 102944. doi: 10.1016/j.pocean.2022.102944, PMID: 41735180IwasakiS. (2023).
Increase in the wave power caused by decreasing sea ice over the Sea of Okhotsk in winter. Sci. Rep.13, 2539. doi: 10.1038/s41598-023-29692-9, PMID: 36781948IwasakiS.OtsukaJ. (2021).
Evaluation of wave–ice parameterization models in WAVEWATCH III® along the coastal area of the Sea of Okhotsk during winter. Front. Mar. Sci.8. doi: 10.3389/fmars.2021.713784, PMID: 41728591KellerJ. D.WahlS. (2021).
Representation of climate in reanalyses: An intercomparison for Europe and North America. J. Clim.34, 1667–1684. doi: 10.1175/JCLI-D-20-0609.1, PMID: 40797873LiuJ.LiB.ChenW.LiJ.YanJ. (2022).
Evaluation of ERA5 wave parameters with in situ data in the South China Sea. Atmosphere13, 935. doi: 10.3390/atmos13060935, PMID: 41725453NaritaD.IwasakiS. (2026).
Past and future changes in sea ice in the Sea of Okhotsk: analysis using the future ocean regional projection dataset. J. Mar. Sci. Eng.14, 52. doi: 10.3390/jmse14010052, PMID: 41725453OlonscheckD.MauritsenT.NotzD. (2019).
Arctic sea-ice variability is primarily driven by atmospheric temperature fluctuations. Nat. Geosci.12, 430–434. doi: 10.1038/s41561-019-0363-1, PMID: 41725453OmonigbehinO.EresanyaE. O.TaoA.SetordjieV. E.DaramolaS.AdebiyiA. (2024).
Long-term evolution of significant wave height in the eastern tropical Atlantic between 1940 and 2022 using the ERA5 dataset. J. Mar. Sci. Eng.12, 714. doi: 10.3390/jmse12050714, PMID: 41725453PatraA.MinS.-K.SeongM.-G. (2020).
Climate variability impacts on global extreme wave heights: Seasonal assessment using satellite data and ERA5 reanalysis. J. Geophys. Res. Oceans125, e2020JC016754. doi: 10.1029/2020JC016754, PMID: 40890438ScreenJ. A.SimmondsI. (2010).
The central role of diminishing sea ice in recent Arctic temperature amplification. Nature464, 1334–1337. doi: 10.1038/nature09051, PMID: 20428168ShimuraT.MoriN.BabaY.MiyashitaT. (2022).
Ocean surface wind estimation from waves based on small GPS buoy observations in a bay and the open ocean. J. Geophys. Res. Oceans127, e2022JC018786. doi: 10.1029/2022JC018786, PMID: 40890438TanakaY.ManawasekaraC.SuzuyamaK. (2023).
Long-term wave climate variability over the global oceans and around Japan based on ERA5 reanalysis. J. Jpn. Soc Civ. Eng.79, 23–18006. doi: 10.2208/jscejj.23-18006TitchnerH. A.RaynerN. A. (2014).
The Met Office Hadley Centre sea ice and sea surface temperature data set, version 2: 1. J. Geophys. Res. Atmos. 119, 2864–2889. doi: 10.1002/2013JD020316, PMID: 41735180ToyodaT.KitamuraY.OkadaR.MatsumuraK.KomatsuK. K.SakamotoK.. (2022).
Sea ice variability along the Okhotsk coast of Hokkaido based on long-term JMA meteorological observatory data. Okhotsk Sea Polar Oceans Res.6, 27–35. doi: 10.57287/ospor.6.27ZandlerH.HaagL.SamimiC. (2019).
Evaluation needs and temporal performance differences of gridded precipitation products in peripheral mountain regions. Sci. Rep.9, 15118. doi: 10.1038/s41598-019-51666-z, PMID: 31641198