Minimize TanDEM-X

TDX (TanDEM-X: TerraSAR-X add-on for Digital Elevation Measurement)

Spacecraft     Launch    Mission Status     Sensor Complement    Ground Segment    References

TSX/TanDEM-X is a high-resolution interferometric SAR mission of DLR (German Aerospace Center), together with the partners EADS Astrium GmbH and Infoterra GmbH in a PPP (Public Private Partnership) consortium. The mission concept is based on a second TerraSAR-X (TSX) radar satellite flying in close formation to achieve the desired interferometric baselines in a highly reconfigurable constellation. A contract to build the TanDEM-X spacecraft was signed in September 2006 between DLR and EADS Astrium.

The primary goal of the innovative TanDEM-X/TerraSAR-X constellation is the generation of a global, consistent, timely and high-precision DEM (Digital Elevation Model), corresponding to the HRTE-3 (High Resolution Terrain Elevation, level-3) model specifications (12 m posting, 2 m relative height accuracy for flat terrain). The HRTE-3/HRTI-3 models were defined by NGA (National Geospatial-Intelligence Agency), Washington, D. C. 1) 2) 3) 4) 5) 6) 7) 8) 9)

The achievable DEM height accuracy has been confirmed in Phase A by a detailed performance analysis taking into account all major system and scene parameters like the finite radiometric sensitivity of the individual radar sensors, co-registration and processing errors, range and azimuth ambiguities, baseline and Doppler decorrelation, the strength and orientation of surface and vegetation scattering, quantization errors, temporal and volume decorrelation, baseline estimation errors and the chosen independent post-spacing (horizontal resolution). 10) 11)

For generating the global DEM, roughly 300 TByte of raw data will be acquired using a network of ground receiving stations. Processing to DEM products requires advanced multi-baseline techniques and involves mosaicking and a sophisticated calibration scheme on a continental scale.

Beyond its primary mission objective of generating a global HRTI-3 DEM, TanDEM-X provides a configurable SAR interferometry platform for demonstrating new SAR techniques and applications, such as digital beamforming, single-pass polarimetric SAR interferometry, ATI (Along-Track Interferometry) with varying baseline, or super resolution. Close formation flight collision avoidance becomes a major issue and a new orbit concept based on a double helix formation has been developed to ensure a safe orbit separation.

Background: In the time frame 2007, the global coverage with topographic data at sufficiently high spatial resolution is inadequate or simply not available for scientific and governmental use. The first step to meet the requirements of the scientific community for a homogenous, highly reliable DEM with DTED-2 specifications was SRTM (Shuttle Radar Topography Mission), launch Feb. 11, 2000. SRTM, representing the first spaceborne single-pass interferometer, was built by supplementing the Shuttle Imaging Radar-C/X-Synthetic Aperture Radar system by second receive antennas mounted at the tip of a 60 m deployable mast structure. Within a ten day mission, SRTM collected interferometric data for a near global DTED-2 (Digital Terrain Elevation Data Level 2) land surface coverage. DTED-2 is the current basic high resolution elevation data source for all military activities and civil systems that require landform, slope, elevation, and/or terrain roughness in a digital format. DTED-2 is a uniform gridded matrix of terrain elevation values with post spacing of one arc second (approximately 30 m). SRTM mapped the Earth between 60 N and 56 S; however, there are still wide gaps, in particular at the lower latitudes.

The TanDEM-X/TerraSAR-X (TDX/TSX) constellation has the potential to close these gaps, to fulfil the requirements of a global homogeneous and high-resolution coverage of all land areas thereby providing the vital information for a variety of applications. The high-precision DEM models are of utmost interest for the civil and military communities, representing the basis for all modern navigation applications.



HRTI-3 definition


Relative vertical accuracy

90% linear point-to-point error over a 1º x 1º cell

2 m (slope ≤ 20%)
4 m (slope ≥ 20%)

12 m (slope < 20%)
15 m (slope > 20%)

Absolute vertical accuracy

90% linear error

10 m

18 m

Relative horizontal accuracy

90% circular error

3 m

15 m

Horizontal accuracy

90% circular error

10 m

23 m

Spatial resolution

Independent pixels

12 m (1 arcsec)

30 m (1 arcsec)

Table 1: DEM specification for HRTE/HRTI level 3 standard - and comparison with DTED-2 model

Figure 1 gives an overview of DEM-level coverage estimates of various observation technologies in the different HRTI classes. It should be noted that a surface area of 150 x 106 km2 represents a global coverage of Terra Firma (i.e., all land areas).


Figure 1: DEM-level versus coverage indicating the uniqueness of the global TanDEM-X HRTI-3 product (image credit: DLR)

Mission concept:

The TanDEM-X mission concept is based on an extension the TerraSAR-X mission by a second almost identical satellite, namely TanDEM-X. Flying the two satellites in a close formation with typical cross-track distances of 300-500 m provide a flexible single-pass SAR interferometer configuration, where the baseline can be selected according to the specific needs of the application. 12) 13) 14) 15) 16) 17) 18) 19) 20) 21) 22) 23)

The SAR (Synthetic Aperture Radar) instruments of TerraSAR-X and TanDEM-X are fully compatible, both offer transmit and receive capabilities along with polarimetry. These features provide a maximum of flexibility in supporting operational services (acquisition of highly accurate cross-track and along-track interferograms without the inherent accuracy limitations imposed by repeat-pass interferometry) and in data product quality. The following basic interferometric SAR (InSAR) observational modes are available (Figures 2 and -3):

1) Bistatic mode where the SAR instruments of both spacecraft look into a common footprint thus providing different views of the observed target area (Note: bistatic InSAR is characterized by the simultaneous measurement of the same scene and overlapping Doppler spectra with 2 receivers, avoiding temporal decorrelation; PRF synchronization and relative phase referencing between the satellites are mandatory). - One satellite serves as a transmitter and both satellites record the scattered signal simultaneously. In this tandem configuration, both spacecraft fly in a close orbit formation. The baseline of this configuration can be selected according to the specific needs of the application. This enables the acquisition of highly accurate single-pass cross-track and/or along-track interferograms without the inherent accuracy limitations imposed by repeat-pass interferometry due to temporal decorrelation and atmospheric disturbances.

2) Pursuit monostatic mode where both satellites are operated independently avoiding the need for synchronization; hence, both SAR instruments look acquire data from the same swath with a short time difference of a few seconds corresponding to an along-track distance of 30-50 km. Different to conventional repeat-pass (i.e., two‐pass or multi‐pass) InSAR observations, the temporal decorrelation is still small for most terrain types with the exception of ocean surfaces and vegetation in the case of moderate to high wind speeds.

3) Alternating bistatic mode is similar to bistatic mode, but the transmitter is switched from pulse to pulse between the two satellites.

The baseline for operational DEM generation is the bistatic mode which minimizes temporal decorrelation and uses efficiently the transmit power. This mode uses either TSX or TDX as a transmitter to illuminate a common radar footprint on the Earth's surface. The scattered signal is then recorded by both satellites simultaneously. This simultaneous data acquisition makes dual use of the available transmit power and is mandatory to avoid possible errors from temporal decorrelation and atmospheric disturbances.

The alternating bistatic mode can be used for phase synchronization, system calibration, and to acquire interferograms with two different phase to height sensitivities; the simultaneously acquired monostatic interferogram has a higher susceptibility to ambiguities especially at high incident angles.

A mission concept has been developed which enables the acquisition/generation of a global DEM within three years. This concept includes multiple data takes with different baselines, different incidence angles, and data takes from ascending and descending orbits to deal with difficult terrain like mountains, valleys, tall vegetation, etc.


Figure 2: Concept of TanDEM-X InSAR observations in bistatic (left) and monostatic (right) modes (image credit: DLR)


Figure 3: Schematic view of the alternating bistatic mode (image credit: DLR)

The TanDEM-X mission concept allocates also sufficient acquisition time and satellite resources to secondary mission objectives which cover the following application spectrum:

• Moving target indication with a distributed four aperture displaced phase centre system

• The measurement of ocean currents and the detection of ice drift by along-track interferometry

• High resolution SAR imaging based on a baseline-induced shift of the Doppler and range spectra (super-resolution)

• The derivation of vegetation parameters with polarimetric SAR interferometry

• Large baseline bistatic SAR imaging for improved scene classification, as well as localized very high-resolution DEM generation based on spotlight interferometry.

• Demonstration of high resolution wide-swath SAR imaging with four-phase-center digital beamforming.

In short, the TanDEM-X mission concept encompasses enabling technologies in a number of ways, including the first demonstration of a bistatic interferometric satellite formation in space, as well as the first close formation flight in operational mode. Several new SAR techniques will also be demonstrated for the first time, such as digital beamforming (DBF) with two satellites, single-pass polarimetric SAR interferometry, as well as single-pass along-track interferometry with varying baseline. 24)


Figure 4: Artist's view of bistatic observation by the TanDEM-X configuration (image credit: EADS Astrium)

TanDEM orbits:

Close formation flight of TerraSAR-X and TanDEM-X. The TerraSAR-X spacecraft remains its sun-synchronous dawn-dusk orbit with the following parameters: mean altitude of 515 km, inclination = 97.44º, local equatorial crossing time at 18 hours on the ascending node, nominal revisit period of 11 days (167 orbits in the repeat, 15 2/11 orbits/day. 25) 26) 27) 28) 29) 30) 31)

For setting up the effective baseline, TanDEM-X is separated from TerraSAR-X in the right ascension of the ascending node. This will span a horizontal baseline, which will be adjusted between 200 m and 3000 m to achieve the effective baselines required for DEM-acquisition at different latitudes. An additional vertical separation at the northern and southern turns is achieved by a relative shift of the eccentricity vectors of the satellites. The result is a complete separation of the two satellite orbits called Helix-formation, which enables a safe operation of close formations with minimum collision risk. Such a Helix formation with an offset in eccentricity vectors and a separation in the right ascension of the ascending node is shown in Figure 2.

The TanDEM-X operational scenario requires a coordinated operation of two satellites flying in close formation. Several options have been investigated and the Helix satellite formation has finally been selected. The helix configuration allows maintaining a relatively small distance between both satellites while at the same time avoiding the collision risk at the poles. This formation combines an out-of-plane orbital displacement (e.g. by different ascending nodes) with a radial (vertical) separation (e.g. by different eccentricity vectors) resulting in a helix-like relative movement of the satellites along the orbit. Since there exists no crossing of the satellite orbits, it is now possible to arbitrarily shift the satellites along their orbits, e.g. to adjust very small along-track baselines at predefined latitudes and to allow safe spacecraft operation without autonomous control.

The Helix orbit for close formation flight, involving the maintenance of baselines of a cluster of spacecraft in orbit for cross-track and along-track interferometric observations, has been patented by DLR. The inventors are: Alberto Moreira, Gerhard Krieger, and Josef Mittermayer.

1) European Patent Office, Patent No: EP 1 273 518 A2 of Jan. 8, 2003. Title: “Satellitenkonfiguration zur interferometrischen und/oder tomographischen Abbildung der Erdoberfläche mittels Radar mit synthetischer Apertur.”

2) US Patent No: US 6,677,884 B2 of Jan. 13, 2004. Title: “Satellite Configuration for Interferometric and/or Tomographic Remote Sensing by Means of Synthetic Aperture Radar (SAR).”


Figure 5: Illustration of the Helix orbit configuration of both spacecraft (image credit: DLR)


Figure 6: Helical shape of interferometric baseline during one orbit (image credit: DLR)

The HELIX formation enables a complete coverage of the Earth with a stable height of ambiguity by using a small number of formations (e.g. ΔΩ ={300 m, 400 m, 500 m} and Δe ={300 m, 500 m}, where `Ω' is the right ascension of the ascending node, and `e' is the eccentricity. Baseline fine tuning can be achieved by taking advantage of the natural rotation of the eccentricity vectors due to secular disturbances and fixing the eccentricity vectors at different relative phasings. Since there exists no crossing of the satellite orbits, it is possible to arbitrarily shift the satellites along their orbits, e.g. to adjust very small along-track baselines at predefined latitudes and to allow safe spacecraft operation without autonomous control.

An appropriate reference scenario has been derived which enables one complete coverage of the Earth with baselines corresponding to a height of ambiguity of ca. 35 m within 1 year assuming a bistatic acquisition in stripmap mode with an average acquisition time of 140 s per orbit.

Both high precision orbit determination (POD) and interferometric baseline vector determination of the tandem configuration will be accomplished by means of the GPS-based TOR (Tracking, Occultation and Ranging) device, a dual-frequency receiver, which will be provided by GFZ as for TerraSAR-X.

Coarse orbit control and maintenance of the tandem configuration will be done as part of the regular maintenance maneuvers using thrusters. Fine-tuning of the Helix of the TanDEM-X satellite will be performed using additional cold gas thrusters.

TanDEM-X formation flight: The Helix formation geometry implies maximum out-of-plane (cross-track) orbit separation at the equator crossings and maximum radial separation at the poles. This is realized by small ascending node differences and by slightly different eccentricity vectors, respectively, as depicted in Figure 7. This concept of relative eccentricity / inclination vector separation results in a Helix-like relative motion of the satellites along the orbit and provides a maximum level of passive safety in case of a vanishing along-track separation. 32)


Figure 7: Formation building with relative eccentricity / inclination vector separation (image credit: DLR)

Legend to Figure 7: From left to right: (1) identical orbits, (2) maximum horizontal separation at equator crossings by a small offset in the ascending node (green arrow), (3) a small eccentricity offset causes different heights of perigee / apogee and hence yields a maximum radial separation at the poles. (4) Optional rotation of the argument of perigee to achieve larger baselines at high latitude regions.


During the development phase of the TerraSAR-X spacecraft, the TanDEM-X mission concept became a vision. However, a realization of the vision of two SAR missions in orbit could only have a chance with a necessary minimum extension of the SAR design on TerraSAR-X to support the synchronized operation of both radars. 'Minimum' meant that the TerraSAR-X schedule was not endangered and was further constrained to allow a cost-effective 1:1 rebuild approach for the SAR on TanDEM-X. 33)

For the spacecraft bus, the approach was constrained by only allowing software changes on TerraSAR-X. The bus design on TanDEM-X was extended to allow formation flight of both satellites - with TanDEM-X as the 'Master of the Constellation.' Particularly the bus hardware extensions were constrained by the tight schedule leading to strong orientation on existing hardware designs. The software changes are being verified during the TanDEM-X on-ground tests and will be uplinked to TerraSAR-X in preparation for the constellation flight.

Like the TerraSAR-X (TSX) satellite, the TanDEM-X (TDX) satellite is based on a mission-tailored AstroBus service module and a radar instrument developed according to the AstroSAR concept. Main differences to the TerraSAR-X satellite are the more sophisticated cold gas propulsion system to allow for constellation control, the additional S-band receiver to enable for reception of status and GPS position information broadcasted by TerraSAR-X, and the X-band intersatellite link for phase referencing between the TSX and TDX radars (the required modifications on the TSX spacecraft have already been implemented).


Figure 8: Artist's view of the TanDEM-X spacecraft (image credit: DLR)

The outer shape of the spacecraft is mainly driven by the accommodation of the X-band radar instrument, the body mounted solar array and the geometrical limitations given by the Dnepr-1 launcher fairing. A standard S-band TT&C system with full spherical coverage in uplink and downlink is used for satellite command reception and telemetry transmission.

An additional intersatellite S-band receiver, operating at the TerraSAR-X downlink frequency, will allow for the reception of status and GPS position information broadcasted by TerraSAR-X. It provides a 1-way link with which TanDEM-X can receive real-time position and velocity data from TerraSAR-X from its nominal 1-frequency GPS receiver. The TanDEM-X OBC (On-Board Computer) software uses such data from both satellites to generate a collision warning flag. Furthermore, this data is used by the TAFF (TanDEM-X Autonomous Formation Flying) algorithm running on the OBC (see Ref. 37).

Nominally, formation flying will be under ground control. The TAFF algorithm will be tested open-loop during the commissioning phase and could then become the standard approach for the constellation phases. The ISLR (Intersatellite Link Receiver) is laid out to receive the TerraSAR-X S-band transmissions in low power mode. It is cold redundant with each receiver/decoder cross-strapped to two patch antennas. This layout keeps contact gaps to less than 15 minutes in addition to an interruption imposed by the nominal high rate S-band contact with the ground station.

The OBC is a fully redundant unit that aims at performing the onboard data handling and the attitude and control functions on the satellites. The processor module in based on the ERC32, clocked at 40 MHz, and ensures an execution of software with a processing capability of more than 10 MIPS. The internal RAM memory comprises 6 MByte, with 4 MByte used nominally and 2 MByte reserved for the implementation of a cold redundancy.

The TanDEM-X attitude control system is based on reaction wheels for fine-pointing with magnet torquers for wheel de-saturation. A combined hydrazine/cold-gas propulsion system allows for orbit maintenance and rapid rate damping during initial acquisition. Attitude and orbit measurement is performed with a GPS/Star Tracker system during nominal operation and a CESS (Coarse Earth and Sun Sensor) in safe mode situations and during the initial acquisition. A combination of laser gyro and magnetometer allows for rate measurements in all mission phases.

CGS (Cold Gas Propulsion System): The CGS on TanDEM-X is of CryoSat-2 heritage and uses a high pressure tank of nitrogen gas. This provides small thruster impulses fitting the needs for constellation flight. There are 2 redundant branches each culminating in 2 redundant pairs of thrusters mounted on the satellite in each of the ± flight directions. A formation flight maneuver involves operation of a pair of thrusters in one of these directions.

The TanDEM-X spacecraft has a launch mass of about 1340 kg (payload mass of 400 kg); the nominal design life is five years after the end of the commissioning phase (estimated to be 3 months); the satellite consumables will last for 6.5 years after commissioning.

The Public-Private Partnership (PPP) between DLR and EADS Astrium has been extended to cover the design, build, launch, commissioning and operation of the TanDEM-X spacecraft. Like TerraSAR-X, TanDEM-X is a dual-purpose (scientific and commercial) Earth observation mission, providing its data services to the science (DLR) and to the non-science communities (Infoterra). This shared approach makes the program affordable to all parties of interest.


Figure 9: TanDEM-X in the satellite integration center at IABG (image credit: DLR, EADS Astrium)

Launch: The TanDEM-X spacecraft was launched successfully on June 21, 2010 on a Dnepr-1 launch vehicle with a 1.5 m long fairing extension. The launch provider is ISC Kosmotras, the launch site is the Baikonur Cosmodrome, Kazakhstan. 34) 35)

RF communications: A standard S-band TT&C system with 360º coverage in uplink and downlink is used for satellite command reception and housekeeping telemetry transmission. The uplink path is encrypted. Generated payload (SAR) data are stored onboard in a SSMM (Solid State Mass Memory) unit of 768 Gbit EOL capacity prior to transmission via the XDA (X-band Downlink Assembly) at a data rate of 300 Mbit/s. The X-band downlink is encrypted.

The on-board SAR raw data are compressed using the BAQ (Block Adaptive Quantization) algorithm, a standard SAR procedure. The compression factor is selectable between 8/6, 8/4, 8/3 or 8/2 (more efficient techniques can only be applied to processed SAR imagery). Both communication links are designed according to the ESA CCSDS Packet Telemetry Standard.


Rebuild of the TerraSAR-X satellite which was based on the Astrium Flexbus concept and extensive heritage from the CHAMP and GRACE missions


- X-band downlink horn antenna is mounted at the tip of a 3.3 m long boom
- SSMM (Solid State Mass Memory) data storage with a capacity of 768 Gbit (EOL)
- High-pressure nitrogen gas propulsion system for formation flying

Spacecraft launch mass

1340 kg (spacecraft: 1220 kg, fuel: 120 kg)

Spacecraft size

5 m length, 2.4 m diameter (hexagonal cross section)

Spacecraft design life

5 years nominal (after the end of the commissioning phase)

RF communications

- X-band of 300 Mbit/s link of payload data downlink with DQPSK modulation;
- S-band uplink of 4 kbit/s (2025-2110 MHz), BPSK modulation; S-band downlink of 32 kbit/s to 1 Mbit/s (2200-2400 MHz), BPSK modulation

Primary payload

Secondary payloads

- TDX-SAR instrument is identical to the TSX-SAR (TerraSAR-X SAR instrument)
in layout, operational performance and support modes.
- TOR (Tracking, Occultation and Ranging)
- LCT (Laser Communication Terminal)
- LRR (Laser Retroreflector)

Table 2: Overview of the TanDEM-X spacecraft parameters 36)

TAFF (TanDEM-X Autonomous Formation Flying)

TAFF is navigation and formation flying software package developed at DLR/GSOC. The overall objective of TAFF is to ease the ground and space operations. Its accurate orbit control performance facilitates the synchronization of the two SAR systems via dedicated horns. In fact the positions of the satellites will be known with a good precision well in advance of real operations. TAFF will enable a safe and robust formation control with minimum collision risk. 37) 38) 39)

On top of ensuring a stable and more precise baseline for SAR interferometry, TAFF will enhance the exploitation of along-track interferometry techniques. Along-track interferometry is enabled by a special configuration of the formation which provides dedicated osculating along-track separations at desired locations along the orbit. This method improves the detection, localization and the signal ambiguity resolution for ground moving targets and can be used for traffic monitoring applications. Furthermore real-time collision risk assessments will be performed by TAFF on a routine basis in order to support automated FDIR (Fault Detection Isolation and Recovery) tasks.

Two GPS receivers are installed on each spacecraft. The dual-frequency IGOR GPS receiver of BroadReach Inc., which serves exclusively scientific purposes, and the single frequency MosaicGNSS receiver of EADS Astrium, whose navigation data are used by TAFF.

A one-way intersatellite link (ISL) is being implemented between the two satellites, using the existing S-band downlink system on TSX and an additional receiver on TDX. The link is designed to function properly up to distances of a few km (ca. 2-5 km).


Figure 10: Overview of the ground and space segments and their interface to TAFF (image credit: DLR)

The TAFF software package resides in the OBC of the TDX spacecraft. TAFF gets as inputs the GPS data provided by the GPS receiver onboard TDX and, through the ISL, also from the GPS receiver data onboard the TSX. TAFF uses the CGS (Cold Gas Propulsion System) to control the formation and performs in-plane control maneuvers in the flight and anti-flight directions only.

The in-flight performance validation of the experimental autonomous formation keeping system embarked by the German TanDEM-X formation has been performed during a 12-day-long closed-loop campaign conducted in June 2012. Relative control performance better than 10 m was achieved, demonstrating that a significant gain of performance can be achieved when the control of the formation is done autonomously on-board instead of on-ground. Furthermore, the formation keeping system was shown to be operationally robust, easy to operate and fully predictable, i.e. fully suited for routine mission operations. This campaign concludes successfully a series of validation activities, opening new doors to future innovative scientific TanDEM-X experiments for which enhanced formation control is required.

TAFF is the first onboard autonomous formation keeping system ever employed on a high-cost scientific formation flying mission with routine data acquisition. As such, it has to face inherent natural fears and reluctance to rely on onboard autonomy for critical activities like formation maintenance. TAFF aims at making evolving the minds by proving that a proper design of the formation (passively safe) as well as a smart implementation of the onboard navigation software (robust navigation and control, internal safety mechanisms) can guarantee simple, accurate and safe formation keeping.

Table 3: Inflight performance test of TAFF 40)


Figure 11: Illustration of the spaceborne DGPS tracking scheme (image credit: DLR)


Figure 12: Photo of the MosaicGNSS (left) and IGOR (right) devices (image credit: DLR)

Parameter / Instrument

MosaicGNSS (EADS Astrium)

IGOR (BroadReach Inc.)

GPS tracking capability

8 channels L1

16 x 3 channels L1/L2

Raw data

C/A: 5 m
L1: 3 mm

C/A, P(Y) 0.2 m
L1, L2: 1 mm

Power consumption

10 W

15 W

Radiation tolerance

35 krad

12 krad

Table 4: Key parameters of the onboard GPS receivers

Formation Flight and Safety measures:

The requirement of a configurable close formation between TSX and TDX arises from the need for a SAR interferometer in space. The satellites fly in almost identical orbits whereby the position of TDX describes a helix around the trajectory of TSX. This is achieved by separation of the relative eccentricity and inclination vector. The maximal radial separation is reached over the poles (vertical baseline typically between 200 - 500 m) and the maximum separation in normal direction occurs at the equator (horizontal baseline typically 200 – 500 m; see Figure 5). In this way, it can be assured that the radial and normal separation never become zero at the same time. The shape of the helix depends upon the mission phase. The formation with the smallest baseline had a minimum separation of 150 m. Orbit correction maneuvers are carried out with the hydrazine propulsion system simultaneously on both spacecraft with exactly the same ΔV. Additionally formation keeping maneuvers are needed to compensate the drift of the relative e-vector that arises from the J2-perturbation (Ref. 31). These maneuvers are made only on TDX with the cold gas system. 41) 42)

Thrusters were originally planned to be the prime actuators during non-nominal situations in AOCS safe mode. The experience with TSX showed, however, that the design with the thrusters mounted at the back of the satellite is far from ideal for flight in close formation. Analyses showed a collision risk of 1/500 due to orbit changes in case of a drop to the thruster based safe mode. 43)The reason is that just a minor part of the thrust is available for attitude control, whereas the major part is changing the orbit in an unpredictable way. - Hence, a second type of safe mode was implemented with the intention to control the attitude without changing the orbit. The so-called ASM-MTQ (Acquisition and Safe Mode-Magnetorquer) only uses the magnetic torque rods as actuators, whereas it still relies on CESS, magnetometer and IMU as sensors, just like the original ASM-RCS (Acquisition and Safe Mode-Reaction Control System).

However, the damping of the rotation rates and the recovery of the attitude takes longer in ASM-MTQ than in ASM-RCS due to the weakness of the magnetic field at 514 km altitude. The maximum overall body rate that can be handled are 0.5º/s due to the concept that the torque rods and the magnetometers are operated in alternation to allow disturbance free measurements of the Earth’s magnetic field.

The new FDIR (Fault detection, Isolation and Recovery) design intends to always use the magnetorquer based safe mode first when a severe anomaly has been detected. There are performance limitations in ASM-MTQ as mentioned above, and it might still become necessary to make use of the conventional but more powerful ASM-RCS. The latter will only be used if the continuation of the mission is seriously endangered. A possible scenario would be the battery voltage dropping below a certain value, a star tracker getting too hot or non-convergence of the attitude after three orbits. The thruster on-time is limited at first instance to make sure that the generated ΔV cannot lead to a collision of the satellites. A reboot of the on-board computer will follow in the worst case scenario when despite of limited use of the thrusters, no convergence was reached. The spacecraft will come up after the reboot in ASM-MTQ again, but this time with wider power/thermal limits. However, the described sequence will be tried only once. If there is still no convergence or the power/thermal limits are yet violated, the spacecraft will be sent by FDIR to ASM-RCS once more, but this time without limitations to the thruster on-time. 44)

The ISL (Inter-Satellite Link) is also used for surveillance, but is subject to some limitations. In the first place, the link only works in one direction and in the second, the connection is interrupted anytime the transmitter of TSX or TDX is switched to high-rate for ground station contacts. Therefore it is seen more as an extra safety rather than the part to rely on completely. The ISL is used to transmit some essential parameters of TSX (including GPS position and velocity) to TDX in order to feed TAFF algorithms (Tandem Autonomous Formation Flight).

AOCS surveillance: The most vital AOCS parameters, such as sensor performance, attitude errors, actuator commands, etc. are monitored on-board. In case of severe anomalies FDIR can react immediately and switch to the redundant hardware. During ground station contacts, a large number of parameters are checked in the mission control system against pre-defined limit settings and violations are indicated by yellow or red flags. The dump files (data covering also the time span in between ground station contacts) are screened with the same limit settings, and violations are reported by email. The events will subsequently be analyzed and it is then decided if they can be disregarded or if a threat to the satellite is developing.

Mission status:

• July 8, 2020: A research team from Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU) conducted the first study of area and elevation changes for all Alpine glaciers over a period of 14 years. This involved comparing three-dimensional terrain models obtained from the German radar satellite mission TanDEM-X and the German-US Shuttle Radar Topography Mission (SRTM) between 2000 and 2014. The team combined the elevation models with optical images from NASA’s Landsat satellites. They found that the Alps have lost approximately 17 percent of their total ice volume since the turn of the millennium. The team recently published the results of their study in the journal Nature Communications. 45)


Figure 13: Crevasses on the Aletsch Glacier (image credit: Christian Sommer)


• The Alps have lost approximately 17 percent of their total ice volume – more than 22 km3 – since the turn of the millennium.

• The FAU research team combined data from the three Earth observation missions TanDEM-X, SRTM and Landsat.

• Similar glacier studies generally assume that the size of glaciated areas will remain constant during an observation period. This can lead to a significant underestimate of the actual mass balance, particularly in highly dynamic regions.

• Focus: Space, Earth observation, global change.

- A 17-percent loss in ice volume is equivalent to more than 22 km3. With the exception of the highest elevations in the Central Alps, the melting of ice is now affecting higher glacier regions, and the trend is continuing.

- The most significant losses were recorded in the mountain massifs of the Swiss Alps. The large valley glaciers of the Bernese Alps alone lost approximately 4.8 gigatons of ice mass between 2000 and 2014. On average, the ice thickness decreased by 0.72 m each year, which corresponds to a volume of almost five km3. Local melting rates were several times higher in the lower reaches of the glaciers. One example is the Great Aletsch Glacier, the largest in the Alps. The surface near the glacier terminus contracted by five meters or more each year due to melting.

- The team from the FAU Institute of Geography obtained its findings by combining data from the three Earth observation missions TanDEM-X, SRTM and Landsat. The key benefit of this method was that it enabled an almost simultaneous comparison of area and elevation measurements. Similar studies from other mountainous regions around the world generally assume that the glaciated surface remains constant throughout the observation period. This can lead to a significant underestimation of the actual mass balance, particularly in highly dynamic glacier regions such as the Alps.


Figure 14: Glacial elevation changes in the Swiss Alps (image credit: Christian Sommer, background imagery of Landsat-8 and SRTM, USGS)


Figure 15: Photo of the upper Grindelwald Glacier, Bernese Alps (image credit: Christian Sommer)

• June 25, 2020: A new era in radar remote sensing began 10 years ago, on 21 June 2010, when the radar satellite TanDEM-X was launched. Since then, it has been orbiting Earth in close formation flight with TerraSAR-X, its three-year-older 'twin'. The distance between the satellites varies between several kilometers and sometimes only 120 meters. This enables the radar sensors to obtain a 3D view of Earth. This is referred to as a bistatic interferometer in space, which allows the terrain structure to be recorded in three dimensions in just one pass. This space mission continues to be globally unique. 46)

- The primary mission objective – the creation of a highly accurate global elevation model of Earth’s entire landmass – was achieved as early as mid-2016 with the completion of the TanDEM-X DEM (Digital Elevation Model). The digital elevation model provides precise topographic information and sets a new standard due to its high accuracy and global homogeneity. The DEM product is available in three different resolution variants. Depending on the quality requirements, elevation measurements were calculated for a grid of 12, 30 or 90 meters. The absolute height error, the inaccuracy of each measurement – only 1.3 meters – is extremely small and far exceeds the original requirement of 10 meters.


Figure 16: Global TanDEM-X DEM (Digital Elevation Model), image credit: DLR

- All elevation data are accessible to the scientific community through DLR’s application process and are used by more than 4000 scientists from 97 countries. The focus of scientific interest is naturally on Earth sciences such as geology, glaciology, oceanography, and hydrology. However, applications for observing vegetation, environmental protection, land use, urban and infrastructure planning, cartography and crisis management also access the extensive data sets and evaluate them according to their needs.

- The elevation data of the 90-meter product variant are freely available for scientific purposes and can be downloaded in the TanDEM-X Science Service System after a simple registration without a an application process.

- TanDEM-X also offers the unique possibility of specifically investigating changes in Earth’s surface. For this purpose, repeated images of the same area are acquired. This process can, for example, make the melting of ice masses at the poles visible and measurable, and allow the monitoring of glaciers worldwide.

- The interferometric images not only enable the generation of highly accurate elevation models, but also contain information that allows a detailed differentiation between forested and open areas. The global TanDEM-X forest map was created for this purpose. Repeated images can be used to monitor rainforests.

- The examples of glacier melt and deforestation (Figures 17, 18 and 19) show how well dynamic changes in different regions of Earth's surface can be observed with the radar satellites TerraSAR-X and TanDEM-X.


Figure 17: Patagonian ice fields (image credit: DLR/EOC)


Figure 18: TanDEM-X elevation model -brittle ice shelf of the Thwaites Glacier (image credit: DLR, CC-BY 3.0)


Figure 19: TanDEM-X forest map (image credit: DLR)

- An interesting example of the change in topography can be seen at the Aurora North oil sand mine in Alberta, Canada, between 2012 and 2016 (Figure 20). The expansion of the mining area by several hectares (green area in the center) as well as the spoil heaps (brownish structure at the lower right) are a measure of the mine's productivity – and the environmental changes.

- Therefore, after the completion of the global elevation model in 2016, work began on a global 3D change map, which will be completed shortly. With an elevation accuracy in the meter range, the new global 3D change map will be available for scientific and commercial applications in 2022, following extensive, highly complex data processing and calibration.

- TanDEM-X and TerraSAR-X have now been in service for more than twice their expected lifetime. They were designed with a service life of five-and-a-half years. Both satellites remain fully operational and still have sufficient propellant for several years of operation in orbit.

- In the future, further surveys are planned that will concentrate on monitoring ice sheets and permafrost areas, large-scale forest surveys, especially for tracking deforestation, and the observation of 2000 cities worldwide for continuous mapping of urban settlement areas.


Figure 20: TanDEM-X images of the Aurora North mine in Canada (image credit: DLR)

What will happen after TanDEM-X?

- TanDEM-X has impressively demonstrated the unique possibilities of bistatic interferometry with radar sensors. The future mission proposal High Resolution Wide Swath (HRWS) is based on four satellites – one main satellite that transmits and receives radar signals, and three small companion satellites that just receive the signals reflected by Earth. This type of formation flight offers various perspectives of Earth due to the different distances between the satellites, thus enabling simultaneous height measurements with varying degrees of accuracy. The generation of elevation models is considerably facilitated and accelerated by combining the measurements, so that digital terrain models of any region on Earth can be delivered on request after a short waiting period.

- As a follow-on mission to TanDEM-X, DLR has designed the bistatic L-band mission Tandem-L in recent years. This mission will be able to map Earth's entire landmass on a weekly basis. In addition to innovative, high-performance imaging technology, the different wavelengths of the radar signals play a decisive role here. The radar signals transmitted by TanDEM-X are in the X-band – with a wavelength of approximately three centimeters – and are essentially reflected at the surface of vegetation. L-band signals – with a wavelength of approximately 25 cm – penetrate the entire volume of vegetation down to the solid ground below.

- A bistatic L-band system thus enables a tomographic recording of forest areas and the imaging of the 3D structure, which is a prerequisite for making a precise determination of global biomass and the changes it undergoes. This is the central and so far, insufficiently known quantity in the carbon cycle. The mission will set new standards in Earth observation, observe global change with unprecedented quality and enable important recommendations for action. Efforts are currently underway to implement this mission within the EU's Copernicus program.

• September 13, 2019: The Kangerlussuaq Glacier is the largest glacier on the southeast coast of Greenland and flows into the Fjord of the same name. The glacier front, which in the past was protected by an ice melange – a mixture of sea ice and calved icebergs – is retreating at an increasing rate. The glacier calves approximately 24 km3 of ice into the ocean every year. This corresponds to about five percent of the amount of ice lost annually by the entire Greenland ice sheet. Using a time series of 150 TanDEM-X elevation models of the Kangerlussuaq Glacier, scientists from Swansea University in the United Kingdom have measured the decrease in the glacier’s surface height. 47)


Figure 21: Landsat image of Greenland's Kangerlussuaq Glacier. Focus: space, Earth observation, climate change (image credit: USGS, DLR)

Glacier physics in the Arctic Ocean

- Glaciers are very sensitive to temperature changes because they are in direct contact with the sea and the atmosphere. Normally, in winter, an ice melange forms a natural protective shield that restricts or even completely prevents the calving of ice masses. During the summer months, the protective shield is not present, and the calving rate of the glacier increases.

- In 2017 and 2018 the ice melange was weakened. This was probably due to a combination of atmospheric and oceanic warming, so the glacier lost large ice masses even during the winter months. By late summer 2018 – after two years of calving all year round – the Kangerlussuaq Glacier retreated further inland than it has ever done in 80 years of observations. The analysis is based on high-resolution radar data from the TanDEM-X satellites acquired during 2017 and 2018, and digital elevation models created especially for the research team in Wales by the German Aerospace Center (Deutsches Zentrum für Luft- und Raumfahrt; DLR).

- Glaciers lose thickness faster when they retreat horizontally. Glaciologist Suzanne Bevan and her colleagues mapped the depth of the glacier bed and fjord floor – and thus determined exactly where the glacier is floating in the water and where it is resting on the land. A comparison of the maps over the two years shows a seasonal advance as well as a clear retreat of the glacier ice front – in addition to the usual slight thickness fluctuations over the seasonal cycle. This has resulted in an elongation, referred to as dynamic thinning and has led to a lowering of the glacier surface. The maps show that between 2016 and 2018 the glacier lost most of its 5 km long floating tongue. In the same period, its thickness decreased by a total of 35 m, which corresponds to approximately 35%.


Figure 22: Surface height profiles of the Kangerlussuaq Glacier as derived with TanDEM-X data. Focus: space, Earth observation, climate change (image credit: DLR)


Figure 23: Retreat of the glacier tongue of the Kangerlussuaq Glacier as derived with TanDEM-X data. Focus: space, Earth observation, climate change (image credit: DLR)

- The TanDEM-X radar satellite mission: TanDEM-X was initiated on behalf of DLR with funding from the German Federal Ministry for Economic Affairs and Energy (BMWi) in public-private partnership with Airbus Defence and Space. DLR is responsible for the scientific use of the TanDEM-X data, the planning and execution of the mission, the control of the two satellites and the generation of the digital elevation models.

• June 12, 2019: A precise understanding of glacier evolution requires knowledge of a glacier's exact mass. This is important in South America, in the tropical regions between Bolivia and Venezuela, where meltwater from glaciers provides drinking water during the dry season. However, basic data about mass changes in glaciers are not easy to obtain. Mass losses are also contributing to rising sea levels on a global scale. As a new analysis method using TanDEM-X data shows, this is particularly true for Patagonia. 48)

- Prior to the availability of this method, scientists had to measure changes in glacier mass on site, which is difficult for large and inaccessible areas. An example is the Patagonian Ice Fields, which lie in the Andes – on the border between Chile and Argentina – and cover an area of almost 18,000 km2. Alternatively, satellite-based gravitational field measurements can provide information about the mass balance. However, this method is not suitable for glaciers in tropical regions with low ice cover. TanDEM-X now offers the possibility of determining the mass balance of glaciers using radar remote sensing. This has the advantages of being a uniform measuring method and offering a higher degree of precision than ever before. Scientists from the Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU) have developed a special processing method, which they have used to obtain a detailed picture of the mass changes in all of South America's glaciers from radar data for the first time.


Figure 24: TerraSAR-X image of the Upsala Glacier in Patagonia, Argentina. Artificially colored TerraSAR-X image (strip mode) of the Upsala Glacier, created using data acquired on 7 January 2008. The colors provide information about the roughness of the terrain. Areas that appear predominantly smooth to the radar are tinted in darker shades of blue and gray. Areas with a coarser surface texture are shown in yellow [image credit: DLR (CC-BY 3.0)]

Entire glaciers have disappeared

- The FAU study shows that the Patagonian Ice Fields have suffered the biggest losses – in addition to the mass losses in the large ice sheets, entire glaciers have already disappeared. Between 2000 and 2015, Patagonia's Ice Fields shrank by around 17.4 gigatons. This equates to a decrease of 19.3 km3 per year and exceeds even the mass losses of glaciers located in the tropics. Analysis of the TanDEM-X data confirms previous investigations and reveals a dramatic change, which was previously only confirmed for areas in Bolivia and Peru.

- The TanDEM-X terrain models record height differences with an accuracy of one meter, allowing even individual glaciers to be accurately measured. Geographers working with Matthias Braun and Tobias Sauter in the fields of remote sensing, geoinformation and physical climatology at FAU are using these data from the 2011–2015 period and comparing them with data acquired by the Shuttle Radar Topography Mission, which took place in 2000. Using a complex procedure, they used the differences to calculate the height changes for the glaciers and hence the changes in their mass. By using the high-resolution TanDEM-X data and the new processing method, the FAU researchers were thus able to analyze the large Patagonian inland ice sheets separately from the surrounding glaciers for the first time. The results of the extensive study were published in the journal Nature Climate Change and may be included in the next report of the Intergovernmental Panel on Climate Change. 49)

Continuation of TanDEM-X and Tandem-L

- The global terrain model derived from data acquired by the TanDEM-X mission is currently being updated, with the researchers from the FAU hoping to benefit even more from these data in the future. They are looking to extend their analysis to other regions and update it over time. The geographers are also excited about Tandem-L, the follow-up mission to TanDEM-X. Tandem-L aims to map Earth's landmass on a weekly basis. In the L-band frequency range, at a wavelength of 23.6 cm, the radar signals of the new satellites would be able to penetrate through the vegetation and into the ground. Radar tomographic images will form part of the Tandem-L mission, thus ensuring even more accurate coverage of glacier masses. In future, FAU scientists will be able to observe the glacier regions of South America with high temporal and spatial precision and thus gain further valuable insights.


Figure 25: Mass changes in the Southern Patagonian Ice Field revealed by radar. Detail from a map showing the changes in height of the Southern Patagonian Ice Field. The names on the map indicate the most important outlet glaciers; a) Pixel-base estimation of the mass change over the period 2000-2015; b) estimated average mass change for the area of the glacier catchment area. Elevation models from the SRTM (Shuttle Radar Topography Mission) in 2000 and from the TanDEM-X mission conducted from 2011-2015 were used in the analysis (image credit: Malz et al., 2018)


Figure 26: On-site glacier measurements. Glaciological ground measurements on the Gray Glacier in Argentina. Poles are drilled into the glacier to measure the rate of melting. Such measurements serve as references for analysis using satellite data (image credit: FAU, Matthias Braun)

• August 2019: TanDEM-X (TerraSAR-X add-on for Digital Elevation Measurement) is successfully operating now already since 2010 and has opened a new era in spaceborne radar remote sensing. A single-pass SAR-interferometer with adjustable baselines in across-and in along-track directions is formed by adding a second (TDX), almost identical spacecraft to TerraSAR-X (TSX) and flying the two satellites in a closely controlled formation. TDX has SAR system parameters which are fully compatible with TSX, allowing not only independent operation from TSX in a mono-static mode, but also synchronized operation (e.g. in a bi-static mode). With typical across-track baselines of 200-600 m DEMs with a spatial resolution of 12 m and relative vertical accuracy of 2 m has been generated. The Helix concept provides a save solution for the close formation flight by combining a vertical separation of the two satellites over the poles with adjustable horizontal baselines at the ascending/descending node crossings. 50)

- Beyond the generation of a global TanDEM-X DEM as the primary mission goal, applications based on cross-track as well as along-track interferometry (ATI) are important secondary mission objectives. Furthermore, TanDEM-X supports the demonstration and application of new SAR techniques, with focus on multi-static SAR, polarimetric SAR interferometry, digital beamforming and super resolution.

- TanDEM-X has successfully achieved its primary mission objective, the generation of a global digital elevation model (DEM) with unprecedented accuracy. Despite being well beyond their design lifetime, both satellites are still fully functional and have enough consumables for operation into the 2020s. Besides the generation of a global change layer until 2020, the bistatic operation of TanDEM-X offers unique opportunities for highly innovative scientific applications as well as for the demonstration of new imaging techniques.

• May 06, 2019: Forests are Earth's lungs; they help to reduce greenhouse gas concentrations in the atmosphere and thus counteract global warming, while also providing protection and resources for humans, animals and plants – and they are being lost at an alarming rate. As the view from space reveals, forests cover about one third of Earth’s landmass today. More than half of the world’s forests, which have fallen victim to deforestation since the middle of the 20th century in particular, have already been lost. The German Aerospace Center (Deutsches Zentrum für Luft- und Raumfahrt; DLR) has created a special dataset to monitor, assess, and protect the current state and development of this green organ with precision – the global TanDEM-X Forest/Non-Forest Map. Interferometric data acquired by the German TanDEM-X radar satellite mission for the creation of a global elevation model were used for this purpose; algorithms from the field of Artificial Intelligence were developed for global data processing. These have been optimized for different types of forests based on tree height, density, and structure. This has resulted in a global map that shows the extent of forested areas at a resolution of 50 meters. DLR’s global TanDEM-X Forest/Non-Forest Map is now available free of charge to scientific users. 51)


Figure 27: Global TanDEM-X forest map (image credit: DLR)

- Radar satellites can acquire image data regardless of the weather or time of day – a particular advantage when it comes to mapping tropical forests, which are usually covered by clouds. The TanDEM-X Forest/Non-Forest Map closes the gaps that previously existed in the data and, for the first time, provides a uniform overview of the rainforests in South America, Southeast Asia, and Africa. The findings are important for authorities and scientists alike, as these areas must be protected from illegal logging and preserved as important stores of carbon.

- The new map can also help scientists to more precisely determine the forest biomass – a key factor when studying the global carbon cycle. The TanDEM-X Forest/Non-Forest Map thus provides an important dataset for research into global change and makes a variety of applications in agriculture, forestry, regional development, and land-use planning possible. In addition, it also allows more precise predictions to be made and appropriate measures to be taken to address the societal challenges arising from global change.

- The DLR Microwaves and Radar Institute has processed more than 400,000 datasets for the project. The datasets were acquired between 2011 and 2015 as part of the TanDEM-X mission. The radar experts have developed special algorithms that first evaluate each image individually and then combines them to form a global map with the goal of extracting and classifying forest-related information from the vast quantities of data. These algorithms are based on machine learning in the field of Artificial Intelligence. More information is available in an article published in the journal 'Remote Sensing of Environment' (Volume 205, February 2018). In the future, it will be possible to evaluate new satellite data and compare it with the global TanDEM-X map, for instance using time series analyses.

- The developers used additional remote sensing data to validate the calculated results and differentiate forest areas from non-forested regions with a greater degree of accuracy. In particular, this includes the ‘global urban footprint’, a global map of settlements created at the DLR Earth Observation Center (EOC), as well as the mapping of water bodies by ESA’s Climate Change Initiative. The distribution of the global TanDEM-X Forest/Non-Forest Map is managed by the German Satellite Data Archive at the EOC and made available to users. The German Space Operations Center (GSOC) is responsible for the operation of the TanDEM-X radar satellite mission.

- Tandem-L – forest monitoring in the future: Assessing and monitoring forest resources is a key task for current and future radar satellite missions. In particular, Tandem-L – a proposal for a highly innovative satellite mission – could in the future generate forest maps on a weekly basis and derive forest height, structure, and biomass accordingly. With its innovative imaging technology and the resulting enormous recording capacity, Tandem-L is also designed to observe other dynamic environmental processes on the Earth’s surface. The mission will set new standards in Earth observation and thus significantly contribute towards addressing global societal challenges.

• February 1, 2019: The Thwaites Glacier, one of the most fragile glaciers in western Antarctica, is melting inexorably into the Amundsen Sea at an ever-increasing rate. Until now, it has been responsible for approximately four percent of the global rise in sea level and will cause the oceans to rise by over 65 centimeters in future as its remaining ice melts. With the German radar satellites TerraSAR-X and TanDEM-X, it is now possible, for the very first time, to observe Thwaites Glacier and other polar regions at regular intervals, with high resolution and in three dimensions. Scientists from DLR ( German Aerospace Center) have generated special TanDEM-X elevation models to better understand and predict the melting processes and changes occurring on Thwaites Glacier. The results of the NASA-led study have now been published in the scientific journal Science Advances. 52) 53)


Figure 28: TanDEM-X elevation model -brittle ice shelf of the Thwaites Glacier. For the first time, TanDEM-X elevation models and data from the latest generation of radar satellites enable detailed observation of glacier changes (image credit: DLR, NASA)

- There is a gigantic, 350-meter cavity in the floor of the Antarctic glacier, with the penetrating seawater continuously eating further into the ice. Experts have long suspected that Thwaites is not firmly attached to the bedrock beneath it, but the size of the cavity and the formation of subglacial channels was as surprising as it was alarming. Satellite data acquired by the partners from the United States, Germany and Italy revealed that a total of 14 billion tons of ice have already been washed out, mainly in the last three years. The melt rate was calculated based on TanDEM-X images.

- In addition, the TanDEM-X elevation models reveal the glacier's special dynamics. The changes in the ice surface elevation were measured with millimeter accuracy, allowing important conclusions to be drawn about the underlying melting processes. With images from the Italian Cosmo-Skymed satellites, it was possible to closely monitor the glacier's 'grounding line', which marks the threshold at which the ice mass no longer has bedrock beneath it and begins to float in the sea. Scientists thus discovered that although the glacier surface is rising, the overall thickness of the ice is decreasing. The consequences of interactions between ice masses and penetrating seawater are far greater than previously thought. These and other such insights are essential to predict the effects of glacier melt on global sea levels more accurately. The current study shows the decisive role played by innovative radar satellite technologies.

- For the detailed time series analyses, the DLR experts ordered a total of 120 TanDEM-X images over the period from 2010 to 2017. A time series of elevation models was created from these using the global TanDEM-X elevation model. "This unique capability of TanDEM-X makes it possible to accurately observe changes in surface topography and thus provide in-depth analyses of melt processes in the polar ice caps," says co-author Paola Rizzoli from the DLR Microwaves and Radar Institute.

- The highly accurate determination of the glacier's structure is achieved thanks to high-precision interferometric processing, geocoding and calibration of TanDEM-X images, which was implemented at the DLR Microwaves and Radar Institute. The input data is provided by the automated TanDEM-X processing chain of the DLR Remote Sensing Technology Institute. The data from TerraSAR-X and TanDEM-X are received by the German Remote Sensing Data Center at its stations in Neustrelitz, Inuvik (Canadian Arctic) and GARS O'Higgins (Antarctic). The satellites are operated by the German Space Operations Center at the DLR site in Oberpfaffenhofen.

- New radar remote sensing technologies and methods make it possible for scientists to conduct more targeted research into critical climate processes and further improve predictive models. The latest findings on the development of Thwaites Glacier provide a valuable guide for climate and environmental research. The study 'Heterogeneous retreat and ice melt of Thwaites Glacier, West Antarctica' was written by Pietro Milillo of the NASA Jet Propulsion Laboratory with co-authors from the University of California, the German Aerospace Center (DLR) and the Université Grenoble Alpes, and is available here on the online portal of the journal Science Advances.


Figure 29: Ice thickness change of Thwaites Glacier. (A) Ice surface elevation from Airborne Topographic Mapper and ice bottom from MCoRDS radar depth sounder in 2011, 2014, and 2016, color-coded green, blue, and brown, respectively, along profiles T1-T2 and (B) T3-T4 with bed elevation (brown) from (16). Grounding line positions deduced from the MCoRDS data are marked with arrows, with the same color coding. (C) Change in TDX ice surface elevation, h, from June 2011 to 2017, with 50-m contour line in bed elevation and tick marks every 1 km (image credit: Thwaites Glacier Study Team)

• October 8, 2018: The 90 m TanDEM-X Digital Elevation Model has been released for scientific use and is now available as a global dataset. By providing this data, the German Aerospace Center (Deutsches Zentrum für Luft- und Raumfahrt; DLR) follows the EU data policy under the Copernicus Earth observation program, which encourages free and open access to satellite data. 54)

- When on 21 June 2010, the TanDEM-X radar satellite was launched to space, its 'twin' TerraSAR-X had already been in Earth orbit since 15 June 2007. Since then, the two German radar satellites have been recording Earth flying in close formation. As they fly over Earth, both satellites 'see' the same land area, but from slightly different perspectives. The signal reflected by the ground arrives at the satellites with a small time offset due to the different range. This range difference is recorded interferometrically with millimeter precision. To calculate exact heights, multiple images of the Earth’s entire land surface had to be taken between 2011 and late 2015. The distance between the twin satellites varied between 500 m and, on occasion, just 120 m. This made the creation of a Digital Elevation Model (DEM) of the Earth’s surface on DLR computers in Oberpfaffenhofen. The full-resolution data, with a horizontal sampling distance of 12 m, also allowed the creation of versions with reduced resolutions of 30 m and 90 m, respectively. While access to the 12 m and 30 m elevation models is subject to restrictions due to the potential for commercial exploitation, and thus requires a scientific proposal, the 90 m DEM is now available on a DLR server and can be downloaded free of charge for scientific data use.


Figure 30: Global TanDEM-X Digital Elevation Model (image credit: DLR)

- The TanDEM-X DEM covers all of Earth’s land surfaces, totalling over 148 million km2. The absolute height accuracy is 1 m. This 3D image of the Earth was completed in September 2016 and is approximately 30 times more accurate than any other global dataset. The elevation models generated with TanDEM-X and TerraSAR-X also have the advantage of being the first to capture the Earth with uniform accuracy and no gaps.

- "Given that we are offering free, straightforward access to the 90 m TanDEM-X DEM, we expect several hundreds of thousands of downloads in the coming months for applications in Earth, hydro and environmental sciences, as well as infrastructure planning and remote sensing," says Alberto Moreira, Director of the Microwaves and Radar Institute. The contact for commercial users continues to be Airbus Defence Space.

- In total, over 2400 scientists from 70 different countries are working with the radar data from TanDEM-X and TerraSAR-X. The digital elevation models can be used to create topographic maps, but also to monitor land use and vegetation, collect hydrological information such as drainage paths or soil moisture, and observe polar ice caps or glaciers.

- The two satellites continue to fly in close formation and acquire images of Earth to detect topographic changes that have occurred as a result of earthquakes or in glaciers, permafrost regions, agricultural areas or urban zones, to give but a few examples. TerraSAR-X and TanDEM-X are still functioning flawlessly after 11 and eight years in orbit, respectively, and have long surpassed their nominal lifetime of 5.5 years. "The quality of the data from TerraSAR-X and TanDEM-X is still excellent, with both radar instruments working just like they did at the start of their mission. Given their remaining fuel resources and good condition of the batteries, it seems likely that they will continue to operate beyond 2020," explains DLR Mission Manager Stefan Buckreuss.

• June 2018: The missions TSX (TerraSAR-X) and TDX (TanDEM-X jointly share the same space segment consisting of two almost identical satellites orbiting in close formation. They are operated using a common ground segment, that was originally developed for TSX and that has been extended for the TDX mission. A key issue in operating both missions jointly is the combination of the different acquisition scenarios: TSX requests are typically single scenes for individual scientific and commercial customers, whereas the global DEM as well as science products require a global mapping strategy. Thus the TSX mission goal is retained and served by both satellites. 55)

1) On-Board Resources Status: TSX and TDX have reached their nominal lifetime at the end of 2012 and 2015, respectively. Therefore it is worth to have a look at the status of the irretrievable on-board resources. Especially the propellant and the battery status are crucial factors determining the future progress and remaining duration of the mission.

- Propellant: The consumption of propellant (Hydrazine) is deter-mined by the number of maneuvers, respectively factors as aerodynamic drag, solar activity, tidal forces, space debris avoidance maneuvers, etc. In addition, the adjustment of the formation between TSX and TDX for bistatic operations consumes a large portion of propellant on the TDX satellite. Currently the propellant filling level of TSX is about 46% and 43% for TDX. This ensures an extension of the mission of three years at least.

Also the filling level of the cold gas used by the additional propulsion system on TDX (only required for close formation flight) still amounts to 10%. This allows a fine orbit adjustment for further six month approximately, whereby strategies have been developed to reduce the cold gas consumption by an increased use of hydrazine thrusters. For this reason almost no cold gas was consumed since mid of 2016. Thus the overall propellant status is currently considered as uncritical for the next years and will not impair the continuation of the mission.

- Battery: The retained TSX battery capacity is about 67% and TDX battery capacity is about 77%. According to analysis, the batteries are in excellent health – exceeding original degradation predictions. However, measures have been taken to conserve the batteries, as for example a limitation of the datatake length during the polar eclipse when the satellites are in the shadow of the Earth.

2) Radar Instrument Status: The usability of the SAR products depends strongly on the absolute calibration and stability of the processed products. The SAR instrument plays a major role in the stability of the whole chain. To ensure the stability of the instrument and to detect weak components at an ear-ly stage, special test datatakes are evaluated on a regular base comprising:

- Health checks for transmit/receive modules

- Repeated acquisitions over corner reflectors and rain forest

- Ultra Stable Oscillator frequency measurements.

Dedicated long-term system monitoring activities make use of these measurements and still confirm the high performance and the excellent stability and radiometric calibration of the SAR system.

3) Global DEM Production: After the launch in June 2010, the TDX SAR system was calibrated and thereafter a comprehensive testing of the various safety measures, the close formation was achieved mid October 2010. The operation at typical distances between 120 m and 500 m is running remarkably smooth and stable since then. Final phase, delay and baseline calibration have reached such an accuracy level, that more than 90% of all Raw DEMs (long data takes are processed to scene based DEMs of 50 by 30 km extension, called Raw DEMs) are within ±10 m compared to SRTM/ICESat data already before the final calibration step using ICESat data as reference heights. More than 500,000 Raw DEMs have been generated in a fully automated process employing multi-baseline interferometric techniques. The first and second global coverages (except Antarctica) were completed in January 2012 and March 2013, respectively. Difficult terrain (e.g. mountains, deserts) have been mapped up to 6 times under special viewing geometries. Antarctica was also mapped twice during local winter conditions. The primary data acquisition program was concluded in 2015.

The final calibration and mosaicking chain was fully operational since the end of 2013 and as of September 2016 the production of the global TanDEM-X DEM was finished. The final global DEM consisting of more than 19,000 1° by 1° tiles is well within specification. A comprehensive system has been established for continuous performance monitoring and verification, including feedback to the TDX acquisition planning for additional acquisitions. The cumulative absolute height error is with 0.9 m outstanding (excluding ice and forested areas) and one order of magnitude below the 10 m requirement.

Beyond the generation of a global TDX DEM as the primary mission goal, a dedicated science phase from 2014 to mid of 2016 aimed at demonstrating the generation of even more accurate DEMs on local scales and applications based on along-track interferometry and new SAR techniques, with focus on multistatic SAR, polarimetric SAR interferometry, digital beamforming and super resolution.

4) Global DEM Update: During the production of the DEM, it turned out that there are height differences from different acquisition periods. In particular, repeated acquisitions for scientific purposes clearly show that the Earth's surface is a very dynamic system when analyzed at this level of accuracy. Not only height changes in glaciers, permafrost regions and forests but also agricultural activities and changes in infrastructure leave clear signals in the X-band DEM.

Therefore, in 2017 the mission decided to acquire an additional complete coverage of the Earth’s landmass and to provide an independent unique DEM-dataset from a well-defined time span (September 2017 until the end of 2019) to be used specifically for the assessment of temporal changes in comparison to the TDX DEM on global scales. Hence the name of the resulting product is “Change DEM”.

The Change DEM will allow monitoring topographic changes on a global scale. In addition, data to provide updates for dedicated areas and for gap-filling in the global DEM will be collected as well. The data takes for this product are still conducted in bistatic operation in close formation, started in September 2017 and are expected to last until end of 2019.

The acquisition planning for the global Change DEM is based on the experience and lessons learned from the TanDEM-X global DEM acquisition. All landmasses of the Earth were separated in dedicated acquisition areas as shown in Figure 31. Each acquisition area is furthermore constrained by certain acquisition requirements in terms of season, number of coverages and desired baselines:


Figure 31: Areas to be acquired for the Change DEM with dedicated parameters (image credit: DLR)

- Glaciers (light blue) will be acquired twice during local winter in order to avoid low coherence of melted ice and snow.

- Mountains with forest (red) will be acquired twice in local summer time in order to acquire additional information for phase unwrapping where this information is too sparse in the present DEM.

- Temperate & boreal forest (dark green) will be acquired in local summer (”leaf on”).

- Tropical forest (light green) will be acquired all year round.

- Deserts (yellow) will be acquired with steep incidence angles to ensure a sufficiently high signal-to-noise ratio.

- Deserts with mountains (orange) will also be acquired twice.

- Most of the polar regions (grey) were already acquired in the respective local winter seasons of 2016/2017.

- The rest of the world (brown) will be acquired once independent of the season, permafrost areas (also brown, north of 60 degree latitude) in the local winter season.

Besides that, several constraints of operational nature are considered as well. The memory on board is shared between the TSX and the TDX mission. In addition, the data take length is limited due to the degradation of the battery after ten years operation. -Also, only a reduced number of ground stations com-pared to the first global DEM acquisition are available for data downlink.

5) Global Change DEM: This Change DEM benefits from improvements in the acquisition planning process and the data processing which enables to achieve reliable DEM data of high accuracy with fewer acquisitions. For this goal, the use of an edited TanDEM-X DEM as “starting point” for the processing is mandatory.

Since the limited satellite resources and time do not allow several coverages for the majority of the landmass, the Change DEM is processed on the basis of the final TanDEM-X DEM product by a newly developed so-called “delta-phase” approach instead of the Dual-(or Multi-)Baseline-Phase-Unwrapping algorithm developed for the mission. The phase unwrapping is now based on an edited version of the global DEM to reduce the density and number of the interferometric fringes. This approach has been tested with demanding acquisition data of very low height-of-ambiguity, yielding a nearly error-free data set. It is important to note, that - although the process starts with the first global DEM - the new phase (height) values are independent of the old ones.

Areas which show no significant changes are used to pre-calibrate the individual DEM scenes (Raw-DEMs) prior to geocoding. This further reduces possible offsets and horizontal shifts in the data – facilitating final calibration and mosaicking. The absolute height accuracy which is driving the use for temporal height change detection, will be in the same order as the first Global DEM, respectively, well below 10 meters.

The lack of several coverages will affect the relative height error performance. As detailed before, a sophisticated acquisition scenario has been developed to maximize the performance. Yet, the random errors will vary slightly over the swathes (in range) since a clapboard pattern as applied in the 1st global DEM is no longer available. Nevertheless, the Change DEM will be processed to same pixel spacing as the Global DEM but with slightly more filtering (more interferometric looks and different filters) applied at the benefit of a lower random height error. Most data will have approximately the height of ambiguity values as the first global coverage (locally even better), thus the relative height error performance is expected to be comparable to the intermediate TDX DEM product (IDEM) tiles which were generated from selected 1st global coverage data only. Unlike the IDEM, the coverage of the Change DEM will be nearly complete and unaffected by larger phase-unwrapping errors. The relative height error depends on the geographical regions outlined in the previous chapter. This means the expected value is in the order of 1 to 2 meters for the majority of the mapped area and increases to about 4 meters over difficult terrain as mountains or deserts.

In summary, in June 2018 TSX will have exceeded its nominal lifetime by 5.5 years and TDX by 2.5 years. Fuel and battery status are considerably better than predicted. The radar performance and calibration of the individual satellites is still within specification or better and no indication of any degradation is noticeable at the moment.

As both satellites are still working very well and have plenty of resources left, it is planned to continue the mission beyond 2020 with the focus on selectively updating and improving the global TanDEM-X DEM and generating a global Change DEM as a self-contained product.


Figure 32: TanDEM-X Raw-DEM of an open pit mining in Wyoming, USA (left) and a three-dimensional change map with 6 m x 6 m resolution from 2016 (right), image credit: DLR

• February 9, 2018: The satellite duo, TerraSAR-X and TanDEM-X, continue to orbit in close formation to make bistatic observations for scientific applications. In addition, observations are being made to fill small areas in the DEM and to improve the quality of the data. Furthermore, observations are being conducted to capture topographic changes. 56)

- Both satellites have used their consumables frugally; each spacecraft spent so far somewhat less than half a tank volume of hydrazine; the batteries are in good operational conditions, and the quality of the radar imagery remains excellent since the start of the missions. The project expects a continuation of operational services of the two missions to at least 2020, subject to unforeseen events.

- A very recent paper has been published providing forest maps on the basis of interferometric TanDEM-X data. 57)


Figure 33: Global TanDEM-X Forest/Non-forest Map (image credit: DLR)

- Project Forest/Non-Forest Map: 58) The TanDEM-X Forest/Non-Forest Map is a project developed by DLR/MRI (Microwaves and Radar Institute), within the activities of the TanDEM-X mission. The goal is the derivation of a global forest/non-forest classification mosaic from TanDEM-X (i.e. TerraSAR-X and TanDEM-X) bistatic InSAR (Interferometric Synthetic Aperture Radar) data, acquired for the generation of the global DEM (Digital Elevation Model) between 2011 and 2015 in stripmap single polarization (HH) mode.

- In this work, the global data set of quicklook images was used, characterized by a ground resolution of 50 m x 50 m, in order to limit the computational burden. For classification purposes several observables, systematically provided by the TanDEM-X system, can be exploited, such as the calibrated amplitude, the bistatic coherence, and the DEM height information. In particular, the volume correlation factor quantifies the amount of decorrelation due to multiple scattering within a volume, which typically occurs in presence of vegetation.

- This quantity is directly derived from the interferometric coherence and used as main indicator for the identification of vegetated areas. For this purpose, a fuzzy multi-clustering classification approach, which takes into account the geometric acquisition configuration for the definition of the cluster centers, is individually applied to each acquired scene.


Figure 34: TanDEM-X Forest/Non-Forest Map example over the Alps (image credit: DLR)


Figure 35: TanDEM-X Forest/Non-Forest Map example over the Amazon Rainforest (image credit: DLR)


Figure 36: TanDEM-X Forest/Non-Forest Map example, a zoom-in over the Amazon Rainforest in the state of Rondonia, Brazil (image credit: DLR)

• July 2017: After the launch in June 2010 and the subsequent commissioning phase, global DEM acquisitions started in December 2010. Parallel to the first month of operational data acquisition the team concentrated its efforts on the calibration of the bistatic interferometer. Correction of differential delays between TSX and TDX was necessary to facilitate the utilization of radargrammetry for resolving the 2π-ambiguity band. Phase, delay and baseline calibration have reached such an accuracy level, that more than 90% of all so-called Raw DEMs (long data takes are processed to scene based DEMs of 50 km by 30 km extension) are within ±10 m of DEMs derived from SRTM/ICESat data already before the final calibration step using ICESat data as reference heights. More than 500,000 Raw DEMs have been generated in a fully automated process employing multibaseline interferometric techniques. 59)


Figure 37: The global TanDEM-X DEM is a consistent data set covering all land surfaces at unprecedented absolute height accuracy of about 1m at a horizontal sampling of 12 m by 12 m. Between 2011 and 2014 at least two acquisitions have been collected by the bistatic TanDEM-X SAR interferometer, mountainous areas have covered up to six times (image credit: DLR)

- The first and second global coverages (except Antarctica) were completed in January 2012 and March 2013, respectively. After some gap-filling, Antarctica was mapped for the first time under local winter conditions. In early August 2013 the satellite helix formation was changed to allow imaging of mountainous areas from the opposite viewing geometry. Due to a low SNR, desert areas had to be re-acquired as well, but at steeper incidence angles. Afterwards the satellites were maneuvered back to the original formation and Antarctica was covered again at larger baselines. The primary data acquisition program was concluded mid-2014.

- A comprehensive system has been established for continuous performance monitoring and verification, including feedback to the TanDEM-X acquisition planning for additional acquisitions. Since the end of 2013 the final calibration and mosaicking chain has been fully operational and completed the global DEM consisting of more than 19,000 1º by 1º (lat/long) tiles in September 2016.

- Global DEM Performance: The quality of the final DEMs is well within the expected performance for the global DEM. Figure 38 shows as an example the absolute height accuracy (90% linear error) per tile derived from the comparison of the TanDEM-X heights against ICESat validation points. The cumulated absolute height error over the complete data set totals 3.5 m. If the forested and ice-covered areas are excluded, where the X-band reflective surface deviates from the laser surface due to different penetration, we end up with an outstanding 0.9 m global absolute height error that is one order of magnitude below the 10-meter requirement.


Figure 38: TanDEM-X DEM absolute height accuracy (90% linear error) per 1º by 1º DEM tile; the cumulated absolute height error for ice-free and non-forested areas is with 0.9 m one order of magnitude below the 10-m requirement (image credit: DLR)

- As the system is very well calibrated and tilts and trends are negligible, the relative height accuracy is well described solely by the random errors in the system. It can be calculated from the interferometric coherence and the resulting phase error. It is specified as the point-to-point error within a 1° by 1° tile. Again, excluding ice and forest areas, where additional volume decorrelation deteriorates the coherence and in consequence the relative height error, 97.8 % of all DEM tiles fulfil the relative height error specification of 2 m (4 m) for flat (steep) terrain.

- Finally, compared to SRTM the TanDEM-X DEM features a much lower percentage of void areas (global count of 0.1 %), especially in desert areas, a result of the reacquisition at steeper incidence angles and hence better SNR. Further details on the TanDEM-X DEM quality can be found in Ref. 60)

- The above mentioned performance monitoring used the interferometric coherence as the key quantity to control the data acquisition, processing and mosaicking. In case of volume scatterers the coherence is mainly determined by the volume decorrelation effect, which in turn can be used to discriminate forested and non-forested areas as shown in Figure 39. As a byproduct of the global DEM a global forest map at a resolution of 50 m x 50 m is currently being generated and is planned to be made freely available for scientific users. 61)

- As both satellites are still working very well and have plenty of resources left, an agreement to continue the mission was concluded between DLR and AIRBUS Defence & Space. Key objectives for this extra mission phase are the acquisition of interferometric data for improvements of the global DEM and the generation of a global change layer, that can be considered as a demonstration for the future climate research and environmental monitoring mission Tandem-L. If the baseline geometries are suitable, further scientific experiments will be included in the timeline as well.


Figure 39: Forest map of an area in Rondonia/Brazil derived from TanDEM-X coherence quick looks (50 m x 50 m resolution), image credit: DLR

- In summary, the generation of the global DEM, the primary mission objective, has been successfully completed. Quality and coverage of the data are outstanding. A science phase dedicated to demonstrating applications based on along-track interferometry and new SAR techniques, has been finished last year. A continuation of the mission was approved with the main objective to use bistatic interferometry in close formation flight to generate a global 3D information change layer.

- TanDEM-X has demonstrated the feasibility of an interferometric radar mission with close formation flight and delivers an important contribution for the conception and design of future SAR missions. One example is Tandem-L, a mission for monitoring dynamic processes on the Earth surface with unprecedented accuracy

• July 6, 2017: Airbus DS has again expanded its WorldDEM portfolio with the launch of WorldDEM4Ortho. Tailored for orthorectification of high and very high resolution optical and radar satellite data, WorldDEM4Ortho will enable corrections of all distortions induced by the topographical variations of the Earth’s surface and satellite orientation when acquiring an image. Covering the Earth’s entire land surface, WorldDEM4Ortho is the most consistent and accurate elevation model for orthorectification on a global scale. 62)

- Without these geometrical corrections, satellite images cannot be used in GIS (Geographical Information Systems) or for any mapping related applications. With the huge development of new geolocated applications like business analytics, location-based services or tourism, the needs for such a consistent and precise elevation model are exploding.

- WorldDEM4Ortho is based on the global WorldDEM dataset. It is produced via a fully automated process and features a vertical accuracy of four meters in a 24 m raster. Identified disturbing terrain artefacts are removed. Bodies of water like lakes or sea are flattened. Rivers are stepped with a flow that follows the surrounding shorelines. Adaptive smoothing processes are also applied to different landscapes and land-use such as urban areas to avoid distortions in the orthorectified image.

• May 19, 2017: Look at a mangrove and it’s clear that it is unlike any other forest type. Roots rise above ground in looping, arching shapes—an adaptation to the low-oxygen soils of subtropical coastal areas. But other aspects of these forests are less obvious, starting with their importance for storing carbon dioxide and keeping it out of the atmosphere.

- Research has shown that mangroves account for only 3% of global forest cover. However, they happen to be the most carbon-rich type of forest in the tropics. This means that mangrove loss can have a large effect; up to 10 % of global carbon emissions from deforestation has been attributed to mangroves. (When trees are harvested and die, whether they are burned or eventually rot, their stored carbon is released to the atmosphere.) Because of their importance to the carbon cycle and climate, researchers have been investigating the structure of these forests from the ground, air, and space. 63) 64) 65) 66)

- The maps of Figures 40 and 41 are the result of one such effort, which uses satellite radar data to model the height of mangrove canopies in Africa. This map, based on a model developed by SeungKuk Lee of NASA’s Goddard Space Flight Center, shows tree canopy heights for 2015 in the vicinity of the Akanda and Pongara national parks in Gabon. The parks span 540 and 929 km2, respectively, and together account for 25 percent of Africa’s protected mangrove area.

- Dark greens represent areas where mangrove trees are the tallest. The darkest greens appear in Pongara National Park, where trees tower up to 60 meters in places—some of the tallest mangroves in the world.

- Information on the height of mangrove canopies can help scientists estimate things like the total amount of biomass in a forest. That information, in turn, can be used to get more precise estimates of how much carbon is locked up in a mangrove, and how land cover changes are affecting where that carbon ends up.

- But mangrove forests are not the same everywhere. Forests across Africa and around the world contain mangroves of various species and structures, with various capacities for storing carbon. Air- and ground-based data are important for helping researchers estimate total mangrove ecosystem carbon stocks in Gabon and around the planet.


Figure 40: Overview of the mangrove parks in Gabon, Africa, acquired in 2015 by TanDEM-X of DLR (image credit: NASA Earth Observatory, maps by Joshua Stevens, using canopy height data courtesy of SeungKuk Lee/NASA GSFC/NASA Carbon Monitoring Systems. Story by Kathryn Hansen)


Figure 41: Detail map of the mangrove parks in Gabon, acquired in 2015 by TanDEM-X of DLR (image credit: NASA Earth Observatory, maps by Joshua Stevens, using canopy height data courtesy of SeungKuk Lee/NASA GSFC/NASA Carbon Monitoring Systems. Story by Kathryn Hansen)

• January 13, 2017: The satellite duo, TerraSAR-X and TanDEM-X, continue to orbit in close formation to make bistatic observations for scientific applications. Both satellites have used their consumables frugally; each spacecraft spent so far about half a tank volume of hydrazine; the batteries are in good operational conditions, and the quality of the radar imagery remains excellent since the start of the missions. The project expects a continuation of operational services of the two missions to about 2020 — these predictions depend of course on the assumption that no unanticipated events occur. After all, in June 2017, TerraSAR-X will be 10 years on orbit. 67)

• October 17, 2016: The German satellite duo TerraSAR-X and TanDEM-X have consistently delivered one-of-a-kind Earth observation data since 2007 and 2010, hence shaping the international research landscape. Now, scientific users from across the globe have gathered for the TerraSAR-X and TanDEM-X Science Meeting at DLR (German Aerospace Center) in Oberpfaffenhofen, where they will discuss the results obtained from the data and define requirements for future remote sensing technology. 68)


Figure 42: TSX/TDX terrain image of Mount Erebus in Antarctica (image credit: DLR)

Legend to Figure 42: Within the framework of the German TSX/TDX radar mission, the polar regions were surveyed for the first time in a comprehensive and highly accurate manner, which is of vital interest to climate research. The terrain model of Figure 42 shows a region of the Antacrtic around the 3794 m Mount Erebus (upper left), an active volcano covered by glacier ice.

• October 4, 2016: The new three-dimensional map of Earth has been completed. Mountain peaks and valley floors across the globe can now be seen with an accuracy of just one meter. The global elevation model was created as part of the TanDEM-X satellite mission; it offers unprecedented accuracy compared with other global datasets and is based on a uniform database. The approximately 150 million km2 of land surface were scanned from space by radar sensors. 69)

- The quality of the global elevation model has surpassed all expectations. Exceeding the required 10 meter accuracy, the topographic map has an elevation accuracy of a single meter. This is a result of excellent system calibration. The distance between the two satellites in formation flight, for example, is determined with millimeter precision. The global coverage achieved by TanDEM-X is also unparalleled – all land surfaces were scanned multiple times and the data was then processed to create elevation models. In this process, DLR's remote sensing specialists created a digital world map consisting of more than 450,000 individual models with pixel by pixel height detail – creating a special kind of three-dimensional mosaic.

- This mission broke new ground in many areas. The close formation flight of the two satellites at a minimum distance of 120 m has become as routine as the various maneuvers required to continuously change the formation and adapt it to the requirements of the imaging geometry. A similar situation applies to bistatic radar operation; simultaneous data acquisition using two radar satellites was initially a major challenge, but was a necessity to ensure the high accuracy of the elevation models. DLR is now a world leader for this pioneering technology.

- TerraSAR-X and TanDEM-X have long exceeded their specified service lives and continue operating faultlessly and in such an efficient way that they still have enough propellant for several more years. Completion of the 3D world map does not signify the end of the mission. Due to the special nature of the formation flight, further scientific experiments are scheduled. Alberto Moreira points out: “Earth as a system is highly dynamic, which is also reflected in its topography. Through frequent updates, we could capture such dynamic processes systematically in the future. This is the primary goal of the Tandem-L mission that we have proposed.”


Figure 43: Crater landscape of the Nevada Test Site (image credit: DLR)

Legend to Figure 43: The 'Nevada Test Site' was, from 1951, the area for numerous nuclear tests. The desert area, 100 km northwest of Las Vegas, is dotted with explosion craters.

• July 2016: The TSX and TDX satellites were designed for a nominal lifetime of 5.5 years. Predictions based on the current status of system resources indicate a lifetime for both satellites and a joint operation until at least 2020 and 2018, respectively. 70)

• June 2016: HDEMs (High-resolution DEMs): After the great achievement in terms of DEM performance of the TanDEM-X mission and the demonstration of the high potentiality of the different experimental modes, the plan of the mission is to continue the acquisitions. In particular, acquisitions for HDEMs and acquisitions for further improvement of the DEM will be carried on, together with the always ongoing science activities. 71)

- As a next milestone, the TanDEM-X mission will carry on a new goal: so-called "high-resolution" DEMs (HDEMs) will be produced for selected areas with an independent posting of 0.2 arcsec (< 6m) by maintaining and even reducing the relative height error (for TanDEM-X DEM 2 m, for HDEM goal 0.8 m for specific areas). The data set will be available on special user-request. Next to the acquisition strategy the following paper highlights the proposed HDEM processing chain. 72)

- The performance on the first HDEM acquisitions performed during the science phase in 2015 is currently under analysis. To support further acquisitions of HDEMs and in order to optimize the SAR acquisition parameters, studies on the relative height accuracy have been performed (Ref. 71).

- Figure 44 shows the relative height accuracy as a function of the range bandwidth for a flat soil-rock terrain acquired with an incidence angle of about 41º, for different backscattering coefficients σ0. Different colors represent the relative height accuracy resulting from different combinations of height of ambiguity (HoA) values. The HoA is defined as the height difference equivalent to a complete 2π cycle of the interferometric phase. It depends on the imaging incidence angle and is inversely proportional to the baseline length. It is a direct scaling factor that relates the interferometric phase error to the relative height accuracy.

- From the figure, one can see the improvement in the height accuracy when increasing the bandwidth for σ0 larger than -15 dB. From the plot, one can also notice the need to combine together several acquisitions with smaller HoA in order to achieve the goal of 0.8m in the relative height accuracy (the black line in Figure 44).


Figure 44: Relative height accuracy versus range bandwidth for different combinations of acquisitions and σ0 values. The black horizontal line indicates the required height accuracy of 0.8 m (image credit: DLR)

- The HDEM regions of interest are shown in Figure 45. These regions include "preparation-areas" (in red, excluding Antarctica) which have good coherence and relatively flat terrains, but exclude forested areas where volume decorrelation effects are strong. In addition, some smaller "demo-areas" (the ones in blue in Figure 45) will be acquired three to four times in order to improve the relative height accuracy, to demonstrate the high resolution DEM performance.

- The new acquisitions are planned with a higher bandwidth of 150 MHz, compared to 100MHz of the nominal DEM acquisitions. This is required to ensure sufficient multi-looking even with the higher resolution of 6m x 6 m. In addition, higher BAQ (Block Adaptive Quantization) of 4 bits/sample (compared to 3 bits/sample of the DEM) is applied where possible to further improve the height accuracy.


Figure 45: Regions of interest for high resolution acquisitions from 2016. In red the "preparation-areas" which are planned only once, in blue the "demo-areas" which will be acquired three to four times (image credit: DLR)

- The Helix formation has been optimized as well in order to reach the required height accuracy (with HoAs ranging from 15 m to 25 m.) at all the latitudes and for all the incidence angles. In particular, as it is shown in Figure 46, the horizontal distance between the two satellites increases from 600 m to 1200 m over 8 months, while the vertical distance remain constant for most of the time, with a small increase in the last one and a half month in order to allow acquisitions at higher latitudes.

- Additional acquisitions are also planned over the Antarctica coast (also shown in Figure 45) in order to improve performance for future versions of the DEM (Ref. 71).


Figure 46: Formation parameters evolution over time for the HDEM acquisitions (image credit: DLR)