Since the launch of TOPEX/Poseidon (T/P) in the summer of 1992, the CNES Precise Orbit Determination (POD) Team gained experience on altimetry missions jointly with dedicated multi-institutional analysis centers. In order to meet past and present altimeter mission requirements, exhaustive efforts were undertaken to characterize short-term and long-term orbit errors impacting mean sea level change estimation. The precise orbit determination on the Jason series and Sentinel-3A/B series of missions now reliably verifies radial orbit accuracies of 7 - 8 mm RMS, making use of the three complementary tracking systems, DORIS (Doppler Orbitography and Radiopositioning Integrated by Satellite), GNSS (Global Navigation Satellite Systems) and SLR (Satellite Laser Ranging).

Today, the requirements have become more stringent: long time series of regional sea surface height and mean sea level should be maintained with a stability of 0.5 mm/year (Ablain et al., 2015). This implies that the radial position of the satellite must be known to better than ~0.1 mm/year at regional scales. Many challenges are thus still ahead of us to achieve the next leap in accuracy and long-term orbit stability. As a consequence, POD standards still need to be adapted periodically to minimize or mitigate the remaining sources of orbit errors (e.g., terrestrial reference frame, systematic signatures in the tracking data, mismodelling of surface forces, incomplete modeling of time-variable gravity). To this end, SLR retroreflectors allow independent verification of orbit accuracy since the version E of CNES POD processing standards (DORIS+GNSS-only orbits are now computed), provided the requirement on SLR data is at least 1 cm RMS short-term stability, with long-term stability well below 1 mm/year (e.g., for identifying geographically correlated orbit errors at inter-annual to decadal time scales).

This paper gives an overview of the activities led by the CNES POD Group, current advances in POD modeling, as well as anticipated future work. The recent improvements for the Sentinel3-A and Sentinel3-B missions, performed in the frame of the Copernicus POD Quality Working Group (QWG), will be detailed. The participation of many analysis centers to the QWG is the key to all altimeter missions success, as for Jason (OSTST) and IGS for example.

GNSS-based Precise Orbit Determination (POD) of Low Earth Orbiters (LEOs) has become a standard application for high-quality GNSS products as provided by the IGS, or its individual analysis centers. For many LEOs, e.g., for altimetry missions, the centimeter absolute orbit precision and accuracy achieved by carrier phase-based GNSS positioning is mandatory and challenging to achieve by other tracking data.

Up to now, the majority of single receiver LEO POD computations are performed by estimating the GNSS carrier phase ambiguities as float values. However, resolving the ambiguities to their integer values is known to inherently stabilize solutions and has been confirmed in many applications. In case of LEO POD, this has been shown extensively, e.g., by double difference processing of space or space-ground baselines for relative POD, allowing for ambiguity resolution without using any dedicated GNSS products in addition. While the processing of space baselines is feasible only for constellations of LEOs in formation flight (e.g., GRACE), the processing of space-ground baselines is very costly in computational terms if all correlations shall be modeled correctly.

Single-receiver or undifferenced ambiguity resolution requires the knowledge of GNSS satellite phase biases which cancel out when forming baselines. Such products have been provided by the CNES/CLS (Centre National d'Etudes Spatiales/Collecte Localisation Satellites) analysis center of the IGS since a few years. Recently, the Center for Orbit Determination in Europe (CODE) IGS analysis center has established the generation of a high-quality signal-specific phase bias product and a fully consistent ambiguity-fixed clock product within its rapid and final IGS-related processing.

On the basis of this new CODE bias and clock products, we demonstrate the performance of single-receiver ambiguity resolution for the absolute POD of GRACE, as well as Sentinel-3, and Swarm. The impact of ambiguity resolution on orbit quality is assessed in terms of standard POD quality measures, independent Satellite Laser Ranging (SLR) validation, and, in case of GRACE, ultra-precise inter-satellite K-band ranging validation.

The Copernicus satellites Sentinel-1, -2 and -3, on both A&B units, carry each two dual-frequency geodetic grade GPS receivers manufactured by RUAG. These receivers are able to track up to 8 GPS satellites simultaneously and are tracking nominally C/A and P(Y) signals. In addition, the B units of these satellites are also able to track C/A and L2C signals; a configuration that has been tested extensively on Sentinel-3B during the commissioning phase.
However, this scenario is going to change in the near future with the C & D units of these satellites. They will carry the new dual-frequency PODRIX receivers that include, among other new features, the capability to track Galileo E1 and E5a signals, in addition to GPS C/A, L2C, L5 and P(Y) signals. Additionally, this receiver will be able to track up to 18 satellites simultaneously, a big increase from the previous 8 channels.
It is still not clear which configuration will have this new receiver, as there are many options (GPS only, Galileo only, GPS+Galileo, GPS with C/A + P(Y), or with C/A + L2C, etc.) Assuming that the receivers are configured to track GPS+Galileo, this will have an impact on the current GNSS POD concepts of LEO satellites, which are typically based on GPS only.
The impact on the current GNSS POD concept used on the Copernicus POD Service, and its supporting infrastructure will be presented, including the impact of Differential Code Biases (DCB), Inter-System Biases (ISB), different quality of the new and legacy signals, or sources for accurate GPS+Galileo products (orbits, clocks and biases).

Precise Orbit Determination (POD) is indispensable for many spaceborne applications that monitor the Earth’s climate system, including Global Navigation Satellite System (GNSS) Radio Occultation (RO). RO is a highly valuable observation technique providing accurate atmospheric profiling measurements in the troposphere and stratosphere regions with global coverage, long-term stability, and virtually all-weather capability. During an occultation event the GNSS signals scan the atmosphere in limb sounding geometry and arrive with a time delay at the receiving RO satellite, which is due to the signal’s refraction in the Earth’s atmosphere. Essential Climate Variables (ECVs) such as temperature, pressure, and tropospheric water vapor, can in turn be retrieved from the raw measurements. In order to enable high accuracy of RO derived ECVs, highly accurate orbit positions and velocities as well as clock estimates of the GNSS transmitter satellites and RO receiver satellites in low Earth orbit (LEO) need to be determined.

Using GPS orbit and clock data from the GNSS orbit data archives of CODE and IGS, and employing Bernese (v5.2) and Napeos (v3.3.1) software packages for the LEO orbit determination of RO satellite missions, we perform a mutual consistency check including estimates of systematic uncertainty bounds and propagated random uncertainties. Furthermore, the obtained orbit products are compared to external orbit solutions from other processing centers and, as possible, to radial position estimates from satellite-laser ranging. Resulting monthly statistics in general show a consistency within RO orbit uncertainty target specifications of 5 cm in position and 0.05 mm/s in velocity for MetOp-A/-B, GRACE, and CHAMP missions. However, degraded observation data, detected by the rigorous consistency evaluation of the quality of the calculated orbits and clocks, can at times lead to decreased accuracy estimates. In particular COSMIC RO satellite orbits show decreased accuracy estimates near 20 cm in position and 0.2 mm/s in velocity, due to less favorable attitude behavior and restrictions in processing observations from two antennas.

Geocenter coordinates, Earth rotation parameters (length of day, polar motion), station coordinates as well as the realization of the International Terrestrial Reference Frame are typically determined by using Satellite Laser Ranging (SLR) observations to passive geodetic satellites, such as LAGEOS-1/2. In the case of active satellites such as GNSS, TerraSAR-X, TanDEM-X, Sentinels, GRACE, or GOCE, the SLR technique is commonly used for the validation of GNSS-based orbits.
Sentinel-3A/3B are ESA and EUMETSAT active satellites launched on February 16, 2016, and April 25, 2018, respectively. Satellites are equipped with laser retroreflectors, thus, can be tracked by SLR stations which provide accurate range measurements. Sentinel-3A/3B integrates thus GNSS and SLR techniques onboard. Mission requirements of Sentinel-3A/3B satellites request precise orientation data and high-quality orbits with the threshold of 3 cm radial 1-sigma RMS and the target of 2 cm to fulfill mission objectives.
In this study, we show that SLR observations and precise GNSS orbits for active LEO satellites, i.e., Sentinel-3A/B, can be used for deriving high-quality parameters, such as station coordinates, Earth rotation parameters, and geocenter coordinates. We present different types of solutions used for the parameter determination, including different solution lengths and different SLR ground network constraining to investigate the best solution pattern. In this study, we employ SLR observations to precise orbit products based on GNSS data and provided by the Astronomical Institute of the University of Bern. Solutions based on SLR tracking of Sentinel provide the length of day parameter with the accuracy of 0.06 ms/d and polar motion with the accuracy of 0.32 mas when compared to the IERS-14-C04 series, and geocenter coordinates with the accuracy of 0.3 cm when compared to LAGEOS-1/2 solutions. Our results show that SLR observations to active satellites can be used for the combination with geodetic satellites, including LAGEOS-based solutions. Combination of observations improves the station coordinates especially for those stations which are dedicated to tracking specific constellations, i.e., providing more SLR observations to LEO satellites than to LAGEOS.

With three identical satellites flying in low Earth orbits, the Swarm mission can deliver very useful information about the thermosphere. Unfortunately, the onboard accelerometers that are dedicated to retrieve information about the atmospheric density and wind, suffer from a variety of unexpected disturbances, which severely affects their usefulness. Therefore, an alternative approach has been developed, where the non-gravitational accelerations experienced by the satellites are derived from the high-quality GPS data, instead of directly measured by the accelerometers. These estimated accelerations are used to generate low resolution thermospheric densities for all Swarm satellites, and to correct and augment the accelerometer observations available for Swarm-C only, to generate higher resolution densities.

This presentation describes the strategy that is used to estimate thermospheric densities from the Swarm GPS observations. Several modifications which have been implemented recently will be explained in detail. These modifications include a more refined handling of the solar radiation pressure modelling, which is especially important for the current solar minimum conditions, as well as an improved modelling of the satellite geometry and the gas-surface interactions. The resulting GPS-derived densities for the entire mission are presented and the impact of the different modifications on their accuracy is assessed.