1 Introduction

Since the early days of altimetry, estimating absolute dynamic topography (ADT) accurately has been an issue (Rio 2010). The absence of an accurate geoid at spatial scales corresponding to the spatial resolution of altimetry data along tracks, i.e. 7km at 1Hz and 300m at 20Hz, does allow to reconstruct absolute dynamic topography accurately. So most scientific studies of the ocean have used the temporally variable part of sea level relative to the mean over a reference period: the sea level anomaly (SLA). And this has improved our knowledge of the dynamics of mesoscale structures. However, many scientific and operational activities require accurate absolute sea level. For the study of eddy-mean interactions, a positive sea level anomaly can be due to different processes such as anticyclonic eddies, the strengthening of a quasi-permanent anticyclonic eddy, the weakening of a cyclonic eddy, or the displacement of a current or eddy (Rio et al. 2007; Pegliasco et al. 2020). Pegliasco et al. (2020) shows that it is more appropriate to use absolute dynamic topography rather than SLA to track eddies. For the correct assimilation of altimetry data into models, Hamon et al. (2019) have highlighted the need for an accurate MDT as well as its associated error. Finally, absolute dynamic topography provides access to geostrophic currents, useful data for monitoring ocean currents, and different application as maritime security, ocean pollution.
Furthermore, with the arrival of new swath observations from the SWOT (Surface Water and Ocean Topography) satellite launched in December 2022 (Fu et al. 2012), which provides sea level observations over swaths 120 km wide with a resolution of 2 km, an accurate MDT at a spatially finer resolution and defined close to the coast is needed.
To access the absolute dynamic topography, it is necessary to estimate an accurate Mean Dynamic Topography (MDT ; ADT = MDT + SLA). Since the launch of the ESA GOCE satellite (Gravity and Ocean Circulation Experiment; Pail et al. 2011), the Earth’s geoid has been measured with centimetric accuracy at a spatial resolution of 100km. In addition, the accumulation of altimetry data, improved processing and, in recent years, the special processing applied to leads (fractures in ice) have led to an improvement in the Mean Sea Surface (MSS) and its estimation over ice-covered areas such as the Artic Ocean (Schaeffer et al. 2023). The ”geodetic” approach (Bingham et al. 2008), which consists of estimating the MDT by subtracting the geoid from the MSS, then applying a reliable filter, provides accurate solutions for spatial scales greater than 100km. To estimate spatial scales shorter than 100km, it is necessary to add information to these scales. A first method is to use altimetry data to add finer scales to the geoid. These geoids are called combined geoids, Eigen6c4 (Förste et al. 2014) and XGM2019e (Zingerle et al. 2020) are examples. From these combined geoids, it is possible to estimate a geodetic MDT such as MDT DTU22 (Knudsen et al. 2022). Another approach is to use a large-scale satellite-only geodetic MDT and add the small scales from in-situ ocean data (temperature and salinity profiles, velocities estimated from drifting buoys or High Frequency radars). These in situ data need to be processed to extract only the physical content corresponding to the MDT. This approach is used to estimate the various CNES-CLS MDTs (Rio and Hernandez 2004; Rio et al. 2011; Rio et al. 2014; Mulet et al. 2021) and in this study.
This paper presents the new CNES-CLS22 Mean Dynamic Topography (MDT) solution. Improvement has been made possible by the recent availability of updated time series of drifter and in situ temperature and salinity profiles data, improved MSS and geoid model. The method is reminded in section 2, while data used in the computation and validation are presented in section 3. The new CNES-CLS22 MDT is described and validated in section 4. Conclusions and discussion are provided in section 5.