Multi-Source Satellite Altimetry for Monitoring Storm Wave Footprints in the English Channel’s Coastal Areas

Highlights

What are the main findings?

• Merging multi-source altimeter SWH datasets is useful to investigate high-frequency wave variability in coastal basins during storms.
• Multi-source altimetry with SWOT captures ~75% of storms, resolving ~80% of wave features and coastal wave modulation–dissipation.

What is the implication of the main finding?

• Multi-mission altimetry-based waves enhance the reliability of numerical models via data assimilation and provide accurate estimates of storm waves.
• SWOT’s 2D wide swath would boost the valuable potential of using multi-source satellite altimeters to observe waves features with short wavelengths.

Abstract

Climate wave data, derived from significant wave height (SWH) altimetry, provide accurate information towards nearshore and coastal areas. Their use is crucial to enhance our capabilities of observing, understanding, and forecasting storm waves, even in complex coastal basins. In this study, SWOT nadir data were combined with nine existing altimeters for assessing waves and monitoring their evolution during storms in the English Channel, near UK–French coasts. Validation against wave buoys and numerical models shows high accuracy, with correlations around 95%, decreasing to 85% when buoy track offsets > 50 km, producing the largest errors. The multi-source approach enables depth-resolved monitoring, with SWH mapping revealing ~20–25% modulation in the Channel and ~36% dissipation near the Seine Bay during storms. Spectral analysis of multi-source altimeter-derived merged observations improve time-sampling, resolving high-frequency variability from monthly to daily scales and capturing ~75% of storms. Most storm wave features along altimetry tracks are resolved, with CFOSAT mapping nearshore areas and SWOT capturing coastal zones, both achieving ~80% variance. This temporal and spatial monitoring would be further enhanced with SWOT’s 2D wide swath. This finding provides a complementary, comprehensive understanding of coastal waves and offers valuable input for data assimilation, to improve storm wave estimates in coastal basins.

1. Introduction

Extreme climate events and storms are considered global high-risk hotspots and pose a significant threat to coastal communities [1,2,3]. Accurate monitoring of these events is crucial for assessing flood resilience in a changing climate [4,5], especially in nearshore and coastal environments characterized by complex geometries and physical processes [2,3]. Traditionally, such monitoring relies on in situ measurements collected by different networks, such as wave buoys, which are considered the most reliable tools for observing the distribution of energy and wave characteristics from the open ocean to shallow water areas and coasts [6]. However, observational data are relatively sparse, and their spatial distribution is extremely limited in areas where the deployment of in situ sensors is not feasible. Due to this limited spatial coverage (i.e., localized measurements), coastal waves remain poorly documented and are currently lacking at local scales [7]. The use of numerical models has increasingly been adopted as an alternative to overcome the shortcomings of in situ monitoring of extreme hydrodynamics, providing a scientific basis for issuing flood warnings. However, numerical modeling requires extremely high computational resources and a large amount of real-time spatio-temporal observations to improve predictive capacity and reduce numerical uncertainties.

The growth of satellite observations in recent decades, including altimetry datasets, has greatly improved the temporal and spatial coverage of nearshore and coastal areas by providing a range of physical measurements. These measurements can be used for timely and accurate monitoring of hydrodynamics in both global and local studies across different time scales, from short- to long-term [7]. Although significant efforts have been made to demonstrate their potential for assessing coastal changes [8], these applications are still in an early stage [5]. Research continues to focus on evaluating the accuracy of different satellite missions, including recent advances in radar altimetry technology, which allow for improvements in spatial and range resolution while reducing noise in conventional nadir altimetry [9,10]. Their use depends on the system of interest (open coasts, deltas, wetlands, estuaries, etc.) and is particularly important in complex environments, such as semi-enclosed basins or channels, for monitoring changes in wave characteristics [11] as well as for validating and calibrating numerical models [5]. Because individual satellite altimeters are limited in their spatial and temporal coverage, combining multiple datasets can help fill these gaps and are especially useful for monitoring wave changes during short-term events, such as storms, at local and regional scales. This multi-source altimetry approach has proven effective in providing observational data along China’s coastal areas [5,11] and in improving the predictive ability of numerical models. However, the accuracy of this integrated approach for monitoring the spatiotemporal evolution of hydrodynamics during storms and other high-energy periods has not yet been thoroughly investigated in different coastal environments influenced by multiple physical drivers and bathymetric changes [5,12].

This context introduces the main objective of the present work: to investigate the use of multi-source altimetry observations—combining SWOT nadir with nine conventional altimeters—for monitoring storm waves during the winter periods of 2022–2023 and 2023–2024, and for analyzing their evolution within a semi-open basin, namely the English Channel (NW France), where modulated hydrodynamic energy may intensify under extreme conditions, leading to increasingly frequent coastal flooding. Accurate SWH measurements gain more insights and become substantially more accessible when using data from Sar and InSar sensors [13,14,15], as in the case of SWOT-SAR-interferometry wide-swath altimetry mission [16]. This work is carried out in the framework of SWOT mission, launched in December 2022 for at least three years, which provides measurements from a single satellite pass [17] and global measurements of sea surface height (SSH) that can be used to map long swell parameters [18] and SWH [19,20].

This research is structured into five sections. Section 2 presents the available data in the English Channel and the methods used. Section 3 evaluates the accuracy of multi-source altimetry observations, against in situ measurements and numerical models, for monitoring storm evolution and assessing short-scale changes in wave features. Section 4 summarizes the key findings, discusses their limitations and considers their implications. Finally, Section 5 presents the main conclusions and outlines future perspectives on integrating SWOT’s 2D view to improve wave estimates in nearshore and coastal basins.

2. Data and Methods

Storm waves during the winters of 2022–2023 and 2023–2024 were monitored in the English Channel (Figure 1), from nearshore to coastal areas, using multi-source altimeter SWH observations, including SWOT nadir, with in situ measurements and numerical simulations.

2.1. In Situ Observations

Four recent storms between 2023 and 2024, following the launch of the SWOT mission, were selected with varying durations (

Table 1. Selected storm events during both winters 2022/2023 and 2023/2024 and their physical characteristics within the Channel (duration, significant wave Height Hs, mean period Tm, and the wave energy Estorm).

Scheme 2.	Duration	Hs (max), Tm	E Storm (m2/h)
Mathis 31/03 (4 a.m.)–01/04 (6 a.m.)-2023	26 h	5 m, 9 s	360.42
Ciaran 01/11 (10 p.m.)–03/11 (4 a.m.)-2023	30 h	5.8 m, 11 s	530.75
Karlotta 10/02 (10 a.m.)–11/02 (7 a.m.)-2024	21 h	5.27 m, 8 s	517.93
Pierrick 09/04 (10 a.m.)–09/04 (3 p.m.)-2024	13 h	4.8 m, 8 s	173.1

): Mathis (April–July 2023) and Ciaran, Karlotta, and Pierrick (since November 2023). Storms were identified using the method proposed by [17], based on the wave height threshold, duration threshold, and storm independence criterion. The wave height threshold was defined as the 95th percentile of significant wave heights [21], and the duration threshold was set at 12 h to capture events capable of inducing substantial changes in hydrodynamics. Storms lasting more than 12 h were characterized by maximum significant wave height (Hs, max), peak wave period (Tp), mean wave direction, and storm energy (E_storm). Storm energy was calculated as the time integral of wave heights exceeding the threshold, multiplied by seawater density and gravitational acceleration [21].

An 18-month wave time series (January 2022–July 2024) was recorded at 1 h intervals. Three wave buoys—one near Brest, France, and two in the central Channel—were obtained from the Copernicus Global Ocean In Situ Near-Real-Time Observations database (https://data.marine.copernicus.eu/product/INSITU_GLO_PHYBGCWAV_DISCRETE_MYNRT_013_030 (accessed on 30 July 2024)), providing hourly time series of significant wave height, peak period, and mean wave direction (buoy 6200069 only). Water levels were recorded by tide gauges with 15 min (UK, BODC: https://www.bodc.ac.uk/data/hosted_data_systems/sea_level/uk_tide_gauge_network (accessed on 30 July 2024)) and 10 min (France, SHOM–REFMAR: https://data.shom.fr/donnees/refmar (accessed on 30 July 2024)) delays. The buoy data were used to validate multi-source altimeter significant wave heights and to analyze wave period distributions.

2.2. Hydrodynamic Modeling

Simulations were performed using the numerical model Delft3D, which models two- and three-dimensional flow and wave interactions in coastal environments. A two-dimensional hydrodynamic model of the English Channel, developed by [21], used extensive hydrodynamic and bathymetric data with spatial resolutions ranging from 4 km in deep waters to 500 m near the coast. The model was validated against in situ wave and water level measurements, showing good agreement, and has been applied previously to assess coastal flooding along the Normandy coast [22,23]. It was employed here to simulate waves during the selected storms.

2.3. Altimetry Observations: SWOT and Conventional Missions

2.3.1. SWOT Altimetry Data

SWOT uses two SAR antennas on a 10 m mast to provide the first two-dimensional, high-resolution measurements of sea surface height from space. Its radar system covers a 120 km swath (with a 20 km nadir gap), enabling direct observations of ocean topography and inland water surfaces, with pixel averaging reducing noise while preserving small-scale signals [17].

The Level 2 (L2) Low Rate (LR) “Expert” products from KaRIn [24] were used. Each pass covers 10–60 km on either side of the nadir, providing SSH, sea surface height anomaly, and significant wave height (SWH) on a 2 × 2 km² swath-aligned grid. SWH was derived from interferometric KaRIn data using independent radar return combinations, applying interferometric correlation to reduce observational uncertainties [19]. These L2 LR Expert products, covering SWOT Cal/Val and Science Orbit periods, are available via CNES-AVISO (SWOT_L2_KaRIn_SSH_Expert: https://doi.org/10.24400/527896/a01-2023.015 (accessed on 30 July 2024)).

In this study, SWOT’s on-board conventional nadir radar altimeter was used to assess significant wave height (SWH), rather than the Ka-band radar interferometer (KaRIn), the mission’s primary innovative payload. Previous studies indicate that SWOT nadir measurements closely match in situ observations, with accuracies of 10–40 cm [17,24]. This limitation is partly due to the unavailability of KaRIn SWH data at certain locations along the Channel, though CNES is improving its processing. The nadir observations used in this work were obtained from the Copernicus dataset.

2.3.2. Conventional Altimetry Data

This study used nine conventional altimetry missions (Low Resolution Modulation and SAR modes) between January 2022 and May 2024—Saral, CryoSat-2, Jason-3, Sentinel-3A/B (S3A, S3B), Sentinel-6 (S6A), Hai Yang-2B/C, and CFOSAT-(

Table 2. General information of the conventional altimetry satellites, used in this work.

Satellite Altimetry	Agency	Mission Duration	Altitude (km)	Inclination (°)	Repeat Cycle (Days)
Saral	ISRO/CNES	2023. 02	800	98.55	35
Cryosat-2	ESA	2010. 04	717	92	30
Jason-3	NASA/CNES	2016. 01	1336	66	10
Sentinel-3A/B	ESA	2016. 02/2018. 04	814.5	98.65	27
Sentinel-6A	ESA	2020. 11	1336	66	10
Hai Yang-2B	CNSA	2018. 10	973	99.3	14
Hai Yang-2C	CNSA	2020. 09	955.58	66	10
CFOSAT	CNES/CNSA	2028. 10	500	97	13
SWOT	NASA/CNES	2022. 12	890	77.6	21
Cal/Val April. 2023
Scientific Orbit November. 2023

). SWH observations were extracted from the Copernicus platform (https://data.marine.copernicus.eu/products (accessed on 30 July 2024)), with sampling frequencies along the track of ~1 Hz for all altimeters except CFOSAT where SWIM (Surface Waves Investigation and Monitoring) instrument provides two processing of SWH (1 and 5 Hz) along the nadir.

2.4. Methodology

2.4.1. Multi-Source Altimetry Coverage in the English Channel

SWH observations from ten altimetry missions were extracted from the Copernicus platform (https://marine.copernicus.eu/about/producers/wave-tac (accessed on 24 July 2024)), where the consistency of data between the different altimeters was assessed through DUACS (Data Unification and Altimeter Combination System). The overall observations were calibrated with Jason-3, used as a reference for inter-altimeter conformity [25].

The reachable tracks from each mission were used to evaluate the spatial coverage of the English Channel, from the Atlantic to Normandy and the UK coasts; they are displayed in Figure 2a,b according to depth as follows: 34 from Sentinel-3A-B (S3A-B, 17 each from A and B), 5 from Sentinel-6A (S6A), 12 from SWOT, and 8 from CFOSAT. The multi-mission tracks (Figure 2e) provide good coverage from deep ocean to nearshore and coastal areas. A series of ten locations, along the Channel at different depths, were selected across ascending and descending passes to maximize the number of observations (white points in Figure 2). For each altimeter, an 18-month SWH dataset (January 2022–July 2024) was recorded along the Channel at specific locations near wave buoys within a radius ranging between 5 km and 50 km (P_S3L, P_S6A, P_CL close to the buoy 6200305; P_S3F, P_S6B, P_CF, P_SE close to the buoy 6200305; P_S3Y, P_CX, P_SG close to the buoy 6200069, as shown in Figure 2). Bathymetry data were obtained from the EMODnet Bathymetry Portal [26] and the latest digital terrain model (DTM), as explored in [22].

SWH reliability was assessed through a substantial process of comparison, with three uniformly distributed buoys (Figure 1) used to evaluate measurements from Sentinel-3A/B, Sentinel-6, CFOSAT, and SWOT. Results of this comparative analysis are presented in the following section.

2.4.2. Spectral Analysis of Altimetry Data

The multi-source SWH observations, assessed at different locations between January 2022 and July 2024, have been analyzed using a non-stationary approach of a Continuous Wavelet Transform (CWT) to explore the spectral content of the SWH signal. The typical analyses of wave and sea-level variability have been performed for investigating the multi-timescale changes and their connection with extreme events [25,27]. The CWT is a well-known method that has been used over the past decade for data analysis in hydrology, geo-physics, and environmental sciences [27,28]. The CWT produces either a timescale or time period with the means of the Fourier transform contour diagram on which time is indicated on the x-axis, period, or scale on the y-axis, and amplitude (called also variance, or power, etc.) on the z-axis.

This study focuses on short frequencies of SWH from days to months, representing oscillation periods under half the record length and a significant contribution to the signal variance. Spectral analysis of SWH along satellite tracks, from the coastline to deeper waters, was used to identify the dominant frequencies of wave height variations.

3. Results

3.1. Monitoring of Waves During Storm Events

This section examines the evolution of nearshore and coastal waves during four storms in the winters of 2022–2023 (Mathis) and 2023–2024 (Ciaran, Karlotta, and Pierrick), using a multi-source dataset (ten missions including SWOT). Mathis (Figure 3a) was captured by eight tracks, including Jason-3 (two tracks), CFOSAT, Saral, Sentinel-3A/B, HY-2B/C, and SWOT Cal/Val. Karlotta (Figure 3b) was monitored by seven tracks, including CryoSat-2, Sentinel-6A (two tracks), Sentinel-3A/B, HY-2C, and SWOT Science Orbit. Pierrick (Figure 3c) was observed by eight tracks, including Sentinel-3A/B, Sentinel-6A, HY-2B/C, and SWOT. Ciaran (Figure 3d) was covered by eleven tracks from most altimeters, with more than two tracks for S3A, S6A, and HY-2B.

The spatio-temporal coverage, provided by reachable altimeters across the English Channel during each storm, is displayed in Figure 3. Long storms, such as Ciaran (lasting over 24 h), were sampled more frequently, with each altimeter recording multiple measurements.

The track distribution within the Channel and near coastal zones offers coverage every 1–4 h, enabling mapping of peak wave heights and their dissipation near the coastline. In some cases, two altimeters were operated simultaneously at different locations, e.g., CFOSAT and Saral, or S3A and S6A during Mathis (31 March, 18:00) and Ciaran (2 November, 21:00).

The distribution of SWH along the Channel, as highlighted by Figure 3a–d, exhibits significant changes in waves with depth and between both sides of the Channel (UK and French coasts). For example, S3B track records higher SWH near Weymouth (UK) than Roscoff (France) during both Storm Mathis (31 March, 10:00) and Storm Pierrick (9 April, 07:00), whereas waves are bigger near the French coast during Karlotta (10 February, 10:00).

The SWH distribution along the S6A track during Karlotta (10 February, 18:00) and Ciaran (2 November, 21:00) shows a decrease in wave energy from the Atlantic toward Cardiff (Wales). HY-2B, crossing the Channel from the Atlantic to Devonport (UK), displays significant variations in SWH along the track during the early stage of Ciaran (1 November, 05:00), with notable dissipation in shallow areas and near the shore; this variation appears more uniform at the start of Storm Mathis (31 March, 07:00). Figure 3e illustrates the evolution of SWH over a five-day period (30 October–4 November 2023), encompassing the storm event of Ciaran and two days before and after the event. An enhanced coverage of the Channel is observed by combining 45 satellite tracks: 7 from S3A, 5 each from S3B and S6A, and 4 each from Jason-3, Saral, CFOSAT, and SWOT.

The integration of multi-source satellite observations provides comprehensive temporal and spatial coverage, extending from the North Atlantic to the Channel entrance and along its axis toward the North Sea.

Beyond capturing extreme wave events, the multi-source altimetry observations provide a spatio-temporal mapping of waves and reveal the distribution of energy increasingly modulated along the basin during the storm and dissipated in shallow water afterward.

A comparative analysis of SWH between the entrance and the center of the Channel during Storm Ciaran reveals ~20–25% of energy modulation. After Ciaran, waves from the centrum of the Channel (~75 m of depth) exhibit ~36% of dissipation at reduced depth (~35–20 m) close to the Seine Bay (Figure 3e).

3.2. Time Evolution of Nearshore and Coastal Waves

The multi-source SWH records were compared to buoy measurements (Figure 4): buoy 6200305 with CFOSAT, SWOT, S-3A, S-3B, and S-6A; buoy 6200103 with CFOSAT, SWOT, S-3A, and S-6A; and buoy 6200069 with CFOSAT, SWOT, S-3A, and S-3B.

Overall, altimetry SWH observations reproduce significant wave heights accurately, including extreme waves (Hs > 3 m). Comparisons with S-3A and S-3B yield a mean RMSE (Root Mean Square Error) of 0.75 m and SD (Standard Deviation) of 0.62 m, with moderate waves (Hs < 2 m) overestimated by 15%. Strong agreement is also observed between in situ measurements and altimetry SWH from CFOSAT, SWOT, and S-6A, with RMSE and SD below 0.42 m.

The correlation coefficient R² between altimetric and buoy-measured SWH was quantified at multiple locations near buoys 6200305 and 6200069 (P_S3L, P_S6A, P_CL close to the buoy 6200305; P_S3F, P_S6B, P_CF, P_SE close to the buoy 6200305; P_S3Y, P_CX, P_SG close to the buoy 6200069 as shown in Figure 2) at distances ranging from 5 to 50 km. SWH estimates from five altimeters (Sentinel 3-A, 3-B, S-6A, CFOSAT, SWOT) show reasonable agreement with buoy measurements, with a mean correlation coefficient of approximately 0.90. Correlation exceeds 0.96 in sites located within a 10 km radius of the buoys and declines to ~0.86 when this radius is around 50 km (

Table 3. Correlations coefficients R² between SWH observations, from five altimeters, wave buoy measurements (buoys 6200305, 6200103, 6200069). The radius range, distance between the altimetry track and the wave buoy, is also specified.

Altimeters	Buoy 6200305	Buoy 6200103	Buoy 6200069
S3A		PS3F R < 10 km R2 > 95%	PS3Y R ~ 50 km R2 ~ 86–89%
S3B	PS3L R ~ 50 km R2 ~ 86–88%		PS3Y R ~ 50 km R2 ~ 86–89%
S6A	PS6A R ~ 15 km R2 ~ 90–93%	PS6B R ~ 50 km R2 ~ 86–88%
CFOSAT	PCL R < 10 km R2 > 95%	PCF R < 10 km R2 > 94%	PCX R < 10 km R2 > 96%
SWOT		PSE R ~ 15 km R2 > 88%	PSG R ~ 10 km R2 > 90%

).

Scatter-point density plots comparing SWH observations (S3, S6, and CFOSAT for 6200305; S3, S6, CFOSAT, SWOT for 6200103; S3, CFOSAT, SWOT for 6200069) are shown in Figure 5. Results indicate that SWH closely matches buoy-measured wave variances.

Near the Seine Bay, coherence between buoy 6200305 and altimetry (S6A) shows slight bias, with SWH overestimated in moderate seas and underestimated in energetic states. Discrepancies are strongest in complex, shallow areas like the semi-enclosed Channel, where wave reflection between French and UK coasts intensifies during storms [26]. They also reflect buoy–track distance, as altimetry data were extracted from the nearest point offshore. Then, accuracy of buoy–altimeter SWH comparisons decreases with separation distance, and most differences are attributable to the buoy–track offset (>50 km), producing large discrepancies.

Altimeter SWH data from S3A-B (combined S3A and S3B missions), S6A, CFOSAT, and SWOT were used to examine temporal wave variability across the Channel at varying depths (locations defined in Figure 2), from its entrance (~100 m) to nearshore (~20 m) and coastal areas. Individual mission time series are shown in Figure 6a–d, and the integrated multi-mission dataset in Figure 6e. The multi-source SWH dataset provides high-resolution temporal coverage, capturing wave multi-timescale variability, from daily to seasonal and annual scales, with moderate waves from May to October and energetic winter waves from November to April. SWH estimates, from individual altimeters, are limited in accuracy but improve considerably when combined, capturing most short-scale variability, including extreme events. Depth-resolved SWH time series reveal ~35% dissipation at 80–65 m, followed by an additional ~10% attenuation at 25–35 m (Figure 6e).

Spectral analysis of an 18-month wave time series (January 2022–July 2024) from combined altimeters was conducted to assess the ability of multi-source data to capture variability across timescales, including high-energy short periods. Continuous Wavelet Transform (CWT) analysis revealed spectral variability ranging from short (high-frequency) to long (low-frequency) scales. The multi-source SWH signal was analyzed at two locations as follows: (1) deep areas near the Channel entrance and Channel Islands (P_S3T for S3A-B, P_S6B for S6A, P_SE for SWOT; Figure 2, top panel of Figure 7), and (2) shallow areas near the coast (P_S3R for _S3A-B, PS6A for S6A, P_SA for SWOT; Figure 2, bottom panel of Figure 7). Frequency comparisons were then carried out for CFOSAT (P_CA), S3A-B (P_S3G), and S6A with SWOT (P_Z; Figure 2e), as shown in the top and bottom panels of Figure 8.

Figure 7 and Figure 8 present the CWT of SWH, with variance shown from blue (low) to red (high). The black line marks the cone of influence, where power may be underestimated due to zero-padding, and thick black contours indicate 99% confidence limits from Monte Carlo simulations (>95% significance). Each diagram depicts variance power across time (x-axis) and frequency/scale (y-axis). Analysis focused on short frequencies, corresponding to oscillation periods less than half the record length, which contribute strongly to total variance. Significant variance appears at short scales, with wavelengths of ~12 days to ~3 months, extending to ~6 months in some cases depending on record length.

CWT diagrams (Figure 7) were used to compare SWOT, S3A-B, and S6A in capturing SWH variability across timescales. Wavelet plots reveal multiple wavelength bands (frequencies), each representing a power range (variance). All three altimeters show similar high-variance regions (>20%, red in the diagram) over wavelengths of ~12 days to ~3–6 months, though with heterogeneous energy structures and distributions. This confirms that SWH variability spans inter-daily (~12–22.5 days) to inter-monthly (~1.5–6 months) scales, shaping distinct wave patterns.

Energy concentration at short timescales (days) may be partly influenced by the repeat cycles of Sentinel-3A/B, S6A, and SWOT. The heterogeneous power distribution across timescales brings about differences in sampling frequency (higher sampling captures higher frequencies) and record length (longer series capture lower frequencies). Low-frequency variability (~6 months) is absent in SWOT data due to the limited length of data (only recorded during 2023–2024).

CWT diagrams (Figure 7) for 2022–2024 show energy concentrated in winter (October 2022–March 2023 and November 2023–April 2024), with high-variance spots linked to storm events, clearly captured by S3A-B and S6A. These patterns highlight SWH variability at inter-monthly and seasonal scales.

Figure 8 shows the spectral content of SWH from multi-source altimetry, assessed individually for each altimeter and jointly in combination, at two Channel entrance sites (P_CA and P_S3G in top panel; P_Z in bottom panel) near Roscoff in the Channel Islands. Results reveal two variability modes: high frequencies (~12–22.5 days) and low frequencies (3–6 months), with low-frequency variance < 16%. Combining S3A-B and CFOSAT increases variance at inter-monthly to seasonal scales to 32%, yielding a clearer variability pattern. Inter-daily frequencies are better resolved in the combined dataset, capturing short, high-energy events.

S3A-B captures the frequency distribution of SWH variability, further enhanced when combined with S6A, CFOSAT, and SWOT, yielding higher variance and clearer energy patterns across timescales. These patterns peak during the energetic winters of 2023–2024, with ~1.5- and ~3-month frequencies well resolved in the combined dataset, demonstrating SWOT’s improved accuracy in capturing multi-frequency SWH features when integrated with other altimeters.

Enhanced detection of SWH variability across scales from multi-source altimetry arises from increased temporal sampling, capturing short, high-energy events often missed by individual altimeters. Four energetic storms were identified during 2023–2024 (Table 1). Individual altimeters could capture at most one event (e.g., Storm Mathis by S3A but not CFOSAT), whereas the multi-source approach effectively captured three storms—Mathis, Ciaran, and Karlotta—covering 75% of events. Conventional altimeters are further limited by spatial gaps between cross-tracks [29,30], partly mitigated by SWOT’s wide swath, which provides two-dimensional coverage, reduces data voids, and improves spatial completeness.

3.3. Spatial Evolution of Wave Features Along the Altimeter Tracks

This section analyzes SWH measurements along altimetry tracks to assess the spatial variability of wave features from the nearshore to the coast, using four tracks from Sentinel-3A, S6A, CFOSAT, and SWOT that crossed the Channel at different times during Storm Ciaran (1 November 22:00–3 November 04:00 2023) as follows: S3A (2 November 21 h) and CFOSAT (2 November 7 h) sampled the Channel’s midsection during the storm, while S6A (3 November 4 h) and SWOT (3 November 1 h) covered its later stages (Figure 9a).

Wavelet spectral analysis of SWH along the tracks was calculated from the coast to deeper zones (following each ascending or descending pass; Figure 9a). The CWT reveals four spatial variability modes (>14 s, 12–14 s, 10–12 s, and 8–10 s; Figure 9b), highlighting distinct distributions of wave energy from the open ocean to the Channel and nearshore areas. S6A, mapping SWH in the Atlantic near the UK’s southwest peninsula, captured energetic wave classes dominated by >14 s and ~12–14 s. At the Channel entrance, S3A recorded higher-frequency waves, limited to ~14 s, with two high-variance bands (~12–14 s and ~10–12 s) and reduced variance at ~8–10 s. Near Roscoff (Channel Islands), CFOSAT recorded SWH up to 7–8 m (Figure 9a) with energetic waves of ~10–12 s and <10 s periods (Figure 9b). Within the Channel, at the east of the Seine Bay, SWOT captured reduced waves at the end of Storm Ciaran, with shorter periods of ~8–9 s (Figure 9b).

Observed wave features were compared with in situ buoy measurements of wave period (Figure 10), which summarizes period distributions from buoys 6200305 (red crosses) and 6200103 (green crosses), shown in Figure 9a, during Ciaran. The probability density function (PDF) of wave periods near to Roscoff indicates a dominant mode of waves with ~9 s (Figure 10a; wave buoy 6200103), decreasing to ~7–8 s at the east of the Seine Bay (Figure 10b; wave buoy 6200305).

Wave modeling from the Atlantic to the English Channel and adjacent coasts was performed with DELFT3D during Storm Ciaran. Simulated Hs (significant wave height) and Tp (peak period) maps were extracted at the dates and times of the altimeter passes, with track positions shown in Figure 11. Comparisons between altimeter-derived SWH and wave simulations were performed in time (altimeter pass dates) and space (along tracks). Results show good agreement, with a mean RMSE of 0.25 cm and SD of 0.2 cm. Spectral analyses (CWT, Figure 9) reveal multiple wave modes (~8–10 s, 10–12 s, 12–14 s, >14 s), which are well reproduced by modeled Tp with RMSE < 1.2 s.

Comparative analysis shows that SWH estimates from CFOSAT and SWOT reproduced over 82% of simulated spatial wave changes along tracks. In shallow waters and near the coast (Seine Bay and Normandy), only SWOT reached higher variances (>87%), declining to ~78% for S3A and S6A. Simulated wave periods were reasonably reproduced by altimetry (>80% agreement), with improved performance from SWOT and CFOSAT nadir.

The multi-source approach provides substantial insight into wave feature changes from the Channel entrance to its center, consistent with numerical modeling (Figure 11). As noted by [21,22], this shift is mainly driven by energy modulation from wave refraction near the seabed, which intensifies during storms with significant reflection between the UK and French coasts.

Previous studies [21,22] demonstrated the complexities of the English Channel arising from nonlinear interactions of waves with tidal currents, water level, and storm surges. These processes intensify during storms, producing wave modulation within the basin and energy reflection between the UK and French coasts.

Small-scale wave variability is evident around the Channel Islands in CFOSAT data, reflecting nearshore–coastal interactions. These interactions become nonlinear where wave transformations coincide, especially in complex bathymetry (Figure 1). S3A data over the Channel entrance (~100–120 m depth) and across the UK island reveal small-scale variability near the coast, with short waves (~6–8 s). At the end of Ciaran, SWOT observed modulated waves with more homogeneous structures (~8–9 s).

Spectral analysis of SWH provides insight into wave features and their bathymetric modulation, clarifying fluctuations in frequency, wave groups, and height gradients. Assessing these features from satellite altimetry is complex and requires retracking techniques [31]. Identifying dominant frequencies controlling wave fluctuations improves the separation of swell from the total SWH signal, particularly in delay/Doppler altimetry with high burst repetition frequency [32]. Combined with aliasing effects, common in SAR data, this underscores the need for high-rate along-track sampling to fully capture long waves, including in shallow and coastal areas [33].

4. Discussion

In this work, a merged multi-source SWH dataset from 10 satellite missions was developed to monitor the evolution of coastal storm waves at a regional scale (English Channel) and a local scale (French and UK coasts), in combination with in situ measurements and numerical simulations [21,22]. This integrative approach proves more effective than single-altimeter observations, particularly in complex coastal environments such as the English Channel, where wave energy is strongly reflected between coasts during storms.

Multi-source altimetry effectively maps short-scale wave features from the Atlantic into the English Channel and toward coastal areas, particularly during storms and high-energy periods. It enhances spatial and temporal coverage of SWH variability from nearshore to coastal zones, with S3A/B, S6A, HY-2B/2C, Saral, Jason-3, and CFOSAT contributing to the Seine Bay, and CryoSat-2 to the Channel entrance and near the Channel Islands (UK).

Significant wave heights from five altimeters (CFOSAT, S3A, S3B, S6A, SWOT) were compared with buoy data under time-matching conditions, showing strong agreement with an average correlation of 95%. Discrepancies mainly reflect spatial mismatches, with R² < 0.9 beyond 50 km, and increase in nearshore and shallow waters.

In fact, nearshore and coastal waves in semi-open basins undergo physical transformations that drive modulation in some zones and dissipation in others, depending on seabed topography and shoreline geometry. These modifications can be highly localized (meters to kilometers), complicating comparisons with buoys which are distant from altimetry tracks (>50 km, as shown here). In addition, buoy-derived wave frequencies must be correlated and aligned with satellite-based SWH to ensure valid comparisons. This requires further investigation, as highlighted by [34], who corrected buoy SWH using data matching, error statistics, and linear least-squares fitting. Their results showed time-varying corrections consistent with 10 m wind, with an average adjustment of 3.8 m.

Multi-source satellite observations provide broad temporal and spatial coverage, mapping wave energy from the Atlantic into the Channel and North Sea. Depth-resolved SWH time series show ~35% dissipation at 80–65 m, followed by ~10% attenuation at 25–35 m near the Seine Bay. During storms, such as the case of Ciaran, the energy is strongly modulated within the basin (~20–25% from the Channel entrance to its center at ~75 m), with ~36% dissipated near the Seine Bay in the storm’s final stages.

The time evolution of SWH was analyzed via spectral analysis of multi-source altimetry to identify variability across inter-daily to inter-monthly scales. Multi-timescale modes at daily scales, typically associated with storms and high-energy periods, were captured; however, these short-scale signals may also reflect satellite repeat orbits and should be interpreted with caution. During 2023–2024, multi-source altimetry captured three of four major storms (75%), versus at most one detected by individual altimeters. Longer time series (>18 months), such as those combining SWOT with conventional altimeters, are needed for more consistent estimates.

Spatial changes in wave features during Storm Ciaran, one of the most energetic storms in the English Channel (winter 2023–2024), were analyzed using altimetry (CFOSAT, SWOT, S3A, S6A) and DELFT3D simulations. Spectral analysis along the tracks identified waves with periods ~8–10 s, 10–12 s, 12–14 s, and >14 s at different times from the Channel entrance to the coast. SWH from CFOSAT and SWOT reproduced >82% of simulated spatial wave changes, with SWOT exceeding 87% variance in shallow waters compared to ~78% for S3A and S6A. Wave periods were also well captured (>80% agreement), with best performance from SWOT and CFOSAT nadir.

Similar storm-driven wave evolution has been reported in the English Channel [21,22], driven by abrupt bathymetric changes and strong winds, such as in Raz Blanchard. SWOT has also revealed short-wavelength features during severe storms in the Indian Ocean (June 2023), attributed to wave-direction fluctuations and nonlinear interactions over shallow topography [19]. These processes highlight gaps in understanding wave evolution, beyond the scope of this study. Comparable spatial changes in storm surges have been detected in the western Pacific using multi-sensor altimetry and in situ data, showing strong modulation during high-energy events [5,35].

The integrative multi-source altimetry approach is valuable for mapping bathymetric influences and assessing fluctuations in wave frequencies, groups, and height gradients—crucial for understanding wave modulation, transformation (e.g., refraction, reflection), and long-wavelength features during storms in semi-enclosed basins. However, capturing long waves can be limited by reduced quality of altimetry-derived parameters due to narrow along-track sampling [36,37], with distortions arising when ocean wavelengths approach the delay/Doppler sampling scale [38]. Developing accurate processing techniques is therefore essential for resolving wave components and their complex interactions, a key challenge in oceanography.

Overall, our results show that multi-source, altimeter-derived wave climate data are more valuable than the sparse spatial coverage of individual altimeters. While single missions can capture hydrodynamic features, combining altimeters greatly improves spatial and temporal coverage, enabling the detection of storm waves in coastal environments with complex bathymetry and nonlinear shallow-water processes [32]. Accurate exploitation of geophysical patterns and their temporal evolution requires high time-sampling, achievable only through multi-altimeter integration. As individual repeat orbits exceed 10 days, a single track may sample a storm shorter than 24 h only once, but multi-altimeter datasets provide multiple passes for longer storms (>5 days).

Satellite altimeters not only monitor local hydrodynamic features but also provide accurate information on the spatial and temporal evolution of wave dynamics. This integrative approach improves the reliability of numerical wave models during storms and supports real-time operational data assimilation, reducing uncertainties in simulations. Previous studies have highlighted the value of multi-source altimetry [5,35,39,40]. Over 25 years of multi-source altimetry and coastal tide gauge data have been used to monitor storm surges in China’s coastal regions, demonstrating that combined altimeters account for ~26% of tide-gauge-identified events and capture both extreme signals and surge dynamics [5]. At the global scale, assimilation of ERS-2 and Jason-1 altimetry into the WAM model improved wave forecasts more effectively with two altimeters than with one [39]. Similarly, assimilation of SARAL/AltiKa, Jason-2, and Jason-3 using Optimal Interpolation improved wave prediction reliability in the Indian Ocean by ~15%, compared with ~4–6% for wind forecasting [40]. Long-term impacts have also been assessed as follows: a regional 3D wave model for 1970–2100 under SSP5–8.5, improved with altimetry assimilation, showed the strongest effects of sea-level and wave changes on wide continental shelves and macro-tidal areas, such as Mont-Saint-Michel Bay, France [35].

5. Conclusions and Perspectives

This study demonstrates the value of multi-source altimeter-derived merged observations for monitoring the spatial and temporal evolution of storm waves in coastal basins, as shown in the case of the English Channel and along the French–UK coasts where energy modulation is strong over complex shallow topography. Data from ten satellite missions, including SWOT, were integrated to assess high-frequency wave variability from inter-daily to monthly scales, including short events of storms. This integrative approach captured ~75% of storms with strong agreement to in situ data (correlation > 0.9). SWH mapping revealed storm waves showed~20–25% storm wave modulation within the Channel and ~36% dissipation near the Seine Bay; with most storm features resolved by multiple altimeters and accurately mapped by CFOSAT in nearshore and SWOT close to the coast (~80% variance). This finding offers a valuable input to enhance the reliability of numerical models via data assimilation and provide accurate estimates of storm waves.

This work introduces SWOT’s contribution alongside conventional altimeters for monitoring storm wave footprints, though further investigation is needed to fully exploit its Ka-band Radar Interferometer (KaRIn), the mission’s primary innovation. SWOT enhances temporal sampling of wave patterns and variability, while its high-resolution 2D observations provide new spatial coverage not yet fully explored. The strong performance of this multi-source approach highlights the need for further accuracy assessments using diverse products, including high-resolution data, to better resolve complex coastal features. Addressing this challenge will support more reliable estimates of coastal change and enhance numerical modeling at regional and local scales.