The first 3D experiment
It was during a sleepless night, with hardly any drop in the midsummer temperatures, when the idea for a radar experiment with both satellites, TerraSAR-X and TanDEM-X occurred: “What if we could prove that interferometry is possible with two satellites even before the final formation is reached?”
Since the textbook launch of TanDEM-X on 21 June, it has been approaching TerraSAR-X – which is flying ahead of it – at a speed of about 630 kilometres per day. Both satellites travel around Earth at altitudes between 515 and 510 kilometres in polar orbits – with every orbit they fly over the North Pole, and 48 minutes later, the South Pole. This type of orbit is common for Earth observation satellites, as Earth rotates beneath the satellite's orbit and the whole of its surface can be observed in a fixed period of time – eleven days in this case.
On 12 July, the first deceleration manoeuvres began, to prevent TanDEM-X from overtaking TerraSAR-X. On 16 July, the distance between the satellites had shrunk to 370 kilometres. This distance corresponds to a time gap of 48 seconds, during which Earth continues to turn, so that the trailing satellite does not cover the same territory. Although the orbits are nearly identical for both satellites, Earth’s rotation causes the ground tracks at the equator to be separated from one another by more than 20 kilometres. This distance is too great for interferometric observations. At the poles, the ground tracks approach one another and eventually intersect. This provides a unique opportunity to observe the same area with two satellites, one after another, from similar positions and with a reduced distance between them. However, because of the 48-second time delay and Earth's rotation, adjustments must be made for each different imaging location.
First TanDEM-X interferogram (left) and the digital elevation model derived from it. These show October Revolution Island, the largest island of the Severnaya Zemlya group in the Arctic Ocean. Credit: DLR.
Now it is necessary to delve deep into the engineer’s bag of tricks and make full use of the possibilities offered by the radar instruments. Here, the heart of the radar – the antenna array with its 384 transmit and receive modules – comes into play. Through skilful command and control of the antenna modules, the ‘viewing’ direction of the radar can be changed without the need for the satellite to be rotated. This targeted ‘squint’ is the basis for acquiring an interferogram with the satellites in their unusual spatial configuration, and then deriving a digital elevation model.
Since the data generated by this technique differs from that produced during normal operation, it is transformed to obtain images using special software for the processing of experimental radar data – known as TAXI, or Experimental TanDEM-X Interferometric Processor. After a further step known as coherent superposition, the interferogram was derived and from that the final product, a digital terrain model, was produced. A special feature of this image is the good height resolution – about 10–20 centimetres – at which even the smallest elevations on the surface of the ice are visible clearly.
The enlarged detail of the interferogram (left) and the corresponding elevation model (right) show a part of October Revolution Island and icebergs. Credit: DLR.
The fact that it has been already proved possible, under these difficult conditions, to produce an interferogram and a first elevation model is an enormous relief. It is clear to us that the conditions for achieving the mission objective – the creation of a digital terrain model of the entire land surface of Earth – have been achieved.
I would like to sincerely thank my colleagues at the DLR Microwaves and Radar Institute who prepared the experiment despite other commitments, commanded the satellites and successfully processed the data. Without their hard work, it would not have been possible to carry out this unique experiment in such a short time.