Abstract's details
First quality assessment of SWOT data in the Gironde estuary
Event: 2025 SWOT Science Team Meeting
Session: Deltas, Estuaries and Coasts
Presentation type: Oral
Estuaries are part of the so-called land/ocean continuum, thereby making the link between the continent and the ocean. They are places for intense sediment exchanges between muddy river fluxes and the adjacent continental shelf as well as for mixing between freshwater and salty water. In addition, these environments are in many cases highly anthropized with diverse human activities (marine trade, agriculture, aquaculture and fisheries, industries,…), so that monitoring water quality (turbidity, salinity…), as well as discharge and water level variations is a major concern. However, understanding the estuarine dynamics remain quite challenging as they are driven by both hydrological (i.e river discharge, …) and oceanic processes (tides, surges, waves,…) which interact between each other. In particular, the river discharge acts as a tide amplitude and a phase modulator and may create strong tidal signal distortion, making tide prediction difficult to achieve. Yet, tides and river discharge along with their interactions occur on a wide range of temporal (from minutes to both seasonal and interannual variabilities) and spatial scales, thereby requiring a large number of measurements to capture the whole dynamics.
In this sense, the SWOT mission represents an incredible opportunity to get additional and unique observations, as well as new insights on these dynamics. By carrying an innovative altimetry radar which combines SAR and interferometry techniques, SWOT is able to provide two dimensional snapshots of water level at high spatial resolution over 50 km wide swaths, which represents a substantial improvement with respect to conventional altimetry in terms of sampling and coverage for coastal and estuarine regions.
A first step consists to verify the quality and the relevance of the existing SWOT products for the ocean (LR products) and for continental waters (HR products), especially as none of these products were specifically designed for estuaries. For this purpose, the Gironde estuary, including the adjacent continental shelf of the Bay of Biscay, is a very suitable area since it is monitored by several tide gauge stations. Besides, realistic simulations of water level and 2D dynamics by the hydrodynamic model T-UGOm are available to us (see Lyard et al. presentation). In this study, we focus more particularly on the validation of the Level 2 Unsmoothed product at 250 metres posting, which corresponds to the LR product with the highest resolution. Indeed, it both has a sufficient spatial resolution for an estuary that is maximum 12km wide compared to the 2km LR product and already includes correction of sea state bias unlike HR product, which remains important in areas undergoing significant waves. As this product is generated through algorithms primarily designed for ocean applications, some processing, corrections or flags may be not fully appropriated in coastal regions.
The first objective of our study is therefore to assess the relevance of the SWOT quality flag by examining total water level (‘ssh’ variable) and backscatter (‘sigma0’ variable) measurements and their corrections. We compute dynamic land/sea masks which change according to the phase of tides based on two different methods. The first one can be used to provide a dynamic surface classification including the identification of intertidal areas, while the second one appears to be more robust to edit spurious anomalies in the data. We find that the ssh quality flag provided in the LR products can be too severe in the upper estuary by removing a too large number of ssh measurements while not capturing certain well-known anomalies, such as swell/wave induced decorrelation at the inner edges of the swaths. In addition, the presented SWOT editing methodology is found to be easier to use than the quality flag and could be applicable in other estuaries throughout the world.
This editing is therefore applied to provide a first SWOT sea surface height error budget by comparing with in-situ and model data. Overall, SWOT deviations remain of the order of ten centimetres along the whole estuary which is quite impressive. Moreover, the comparison with the model enabled to detect some potential errors in the model that could not have been revealed in tide gauge comparisons alone. Yet, there is still a work in progress to explain the remaining bias between SWOT and in-situ/model data which may be linked to vertical movement of the Earth’s crust or/and to residual errors from altimetry corrections. In particular, we start to explore the sensitivity to the cross-track correction which is required to avoid deviations that can reach several tens of cm exceeding sometimes one metre.
This study is a contribution to the SCOEPUS (SWOT Coastal Ocean and Estuarine Products Usability Study) project of the SWOT Science Team.
Back to the list of abstractIn this sense, the SWOT mission represents an incredible opportunity to get additional and unique observations, as well as new insights on these dynamics. By carrying an innovative altimetry radar which combines SAR and interferometry techniques, SWOT is able to provide two dimensional snapshots of water level at high spatial resolution over 50 km wide swaths, which represents a substantial improvement with respect to conventional altimetry in terms of sampling and coverage for coastal and estuarine regions.
A first step consists to verify the quality and the relevance of the existing SWOT products for the ocean (LR products) and for continental waters (HR products), especially as none of these products were specifically designed for estuaries. For this purpose, the Gironde estuary, including the adjacent continental shelf of the Bay of Biscay, is a very suitable area since it is monitored by several tide gauge stations. Besides, realistic simulations of water level and 2D dynamics by the hydrodynamic model T-UGOm are available to us (see Lyard et al. presentation). In this study, we focus more particularly on the validation of the Level 2 Unsmoothed product at 250 metres posting, which corresponds to the LR product with the highest resolution. Indeed, it both has a sufficient spatial resolution for an estuary that is maximum 12km wide compared to the 2km LR product and already includes correction of sea state bias unlike HR product, which remains important in areas undergoing significant waves. As this product is generated through algorithms primarily designed for ocean applications, some processing, corrections or flags may be not fully appropriated in coastal regions.
The first objective of our study is therefore to assess the relevance of the SWOT quality flag by examining total water level (‘ssh’ variable) and backscatter (‘sigma0’ variable) measurements and their corrections. We compute dynamic land/sea masks which change according to the phase of tides based on two different methods. The first one can be used to provide a dynamic surface classification including the identification of intertidal areas, while the second one appears to be more robust to edit spurious anomalies in the data. We find that the ssh quality flag provided in the LR products can be too severe in the upper estuary by removing a too large number of ssh measurements while not capturing certain well-known anomalies, such as swell/wave induced decorrelation at the inner edges of the swaths. In addition, the presented SWOT editing methodology is found to be easier to use than the quality flag and could be applicable in other estuaries throughout the world.
This editing is therefore applied to provide a first SWOT sea surface height error budget by comparing with in-situ and model data. Overall, SWOT deviations remain of the order of ten centimetres along the whole estuary which is quite impressive. Moreover, the comparison with the model enabled to detect some potential errors in the model that could not have been revealed in tide gauge comparisons alone. Yet, there is still a work in progress to explain the remaining bias between SWOT and in-situ/model data which may be linked to vertical movement of the Earth’s crust or/and to residual errors from altimetry corrections. In particular, we start to explore the sensitivity to the cross-track correction which is required to avoid deviations that can reach several tens of cm exceeding sometimes one metre.
This study is a contribution to the SCOEPUS (SWOT Coastal Ocean and Estuarine Products Usability Study) project of the SWOT Science Team.