October 16, 2020

PART 2/4 – Cornerstones of the first two decades

Early DFD logo, around 1990

Any look back can easily be influenced by the vantage point of one’s own activities and perceptions. For that reason I want to first mention in Part 2 some additional fundamental decisions about the future course of DFD that continue to have an effect up to the present:

One of these decisions about the future course of DFD was, without doubt, the integration of the “Neustrelitz Remote Sensing Station” as a branch of DFD in 1992. It was also a reunification into DLR at the initiative of DFD. In many respects the tasks, facilities and expertise in Neustrelitz were similar to those in Oberpfaffenhofen, so during the first few years numerous processes of coordination and shifts in priorities had to be gone through until step by step an overall DFD structure was established. In order to acquire the data coming from the PRIRODA/MOMS-2P mission on Russia’s MIR space station we erected the first X-band station in Neustrelitz. Five others were to follow.

The Neustrelitz facilities in 1990

Today, DFD in Neustrelitz is one of the “first ports of call” in the European ground station network. As a natural consequence, this year we were commissioned (in a work order) for the first time also by ESA to provide acquisition services relating to the operation of Copernicus satellites. In general, the Neustrelitz facilities have become a visible advertisement for DLR in Mecklenburg-Western Pomerania. Today, no fewer than four DLR institutes are pursuing activities in Neustrelitz that all developed from the original range of responsibilities of this DFD branch. They include, in addition to DFD with its large ground station facilities visible from afar, also the Institute of Communication and Navigation, the Remote Sensing Technology Institute, and since 2019 also the Institute of Solar-Terrestrial Physics.

New DFD building in Neustrelitz 2014
The TriBand-I antenna at DFD Neustrelitz, 2019

Another key development was the continuous involvement of DFD since 1991 in functions relating to operational ESA satellites, first as a Processing and Archiving Facility (D-PAF) for ERS-1 and ERS-2, then as a Processing and Archiving Center (D-PAC) for Envisat, and finally since 2014 until now as part of the EU Copernicus Mission. Today we handle the payload data from the two Sentinel-1 satellites, from Sentinel-3/OLCI, and from the Sentinel-5p satellite. This European integration can without doubt be considered part of the backbone of DFD.

Without the foresight in the mid-1980s to firmly anchor DFD also in an ESA context (at that time DLR had no significant missions of its own), much of our development would not have been possible.

In this line-up we also have in GARS O’Higgins an exceptionally successful ground station in Antarctica, visible from afar and constructed in cooperation with BKG.  Since 1992 it has provided us on 365 days of the year with valuable services for TerraSAR-X and TanDEM-X as well as other missions, and BKG with important VLBI-data for geodetic reference measurements.

GARS O'Higgins Antarctic station near the red buildings of the neighbouring Chilean station
Erhard Diedrich and Robert Metzig during an inspection in November 2016
DFD staff Marcelo Morais in a helikopter on a return flight from GARS 2018

For over ten years we have operated a second polar ground station for satellite data acquisition in the Canadian Arctic, in Inuvik. Temporary locations in Bishkek, Cordoba, Libreville, Kitab and Ulaanbaatar, as well as our long-term cooperation with Mexico in Chetumal, were precursors of this development. DFD was always highly international.

More setting of priorities, like multimission capability in the payload ground segment, was also initiated and stimulated. And our software development line DIMS – later in cooperation with the Werum company – made us European trailblazers. (For a long time ground segments were customized for each satellite. Today, multimission capability is standard. For that reason our technology is recognizable today at many places in ESA’s ground segment architecture.)

Assuming the function of a national satellite data archive, D-SDA, was another of our institute’s setting of priorities that continues to have an effect. Originally conceived as an in-house solution for national mission data, we also have been using for some time the infrastructure for the Copernicus Sentinel satellites, as a backloading station for national Copernicus services (CODE-DE), and in the future we will be using it for our own research data platform “TERRA_Byte” together with LRZ.

The robot archive in 2006

We operate a diverse range of data acquisition and processing systems that are distributed over several localities.

Nevertheless, much would have developed differently without the success story of “radar” that began in the mid-1980s with our first responsibilities for Seasat-1 and later for SIR-B. After carrying out several studies for ESA, a new main emphasis, SAR data processing, was established through participation in ERS-1 starting in 1991 and in the two 1994 SIR-C/X-SAR missions. This enabled us to build up comprehensive expertise in processor development that continues to earn us international recognition. A team in the department established by Wolfgang Noack, SAR data processing (DFD-SD), began what was at that time a most courageous attempt to develop the processor for the C-band SAR on ERS-1 using an AI approach. Unfortunately, it was not ready in time so almost at the last minute we had to rely on a commercial product from MDA, a Canadian company. The systems were then created, integrated and tested in a tour de force in Klaus Reiniger‘s department, and they were up and running in time for the launch of the mission. What a relief it was when ERS-1 project leader Jörg Gredel could proudly present Federal Research Minister Heinz Riesenhuber with the first image data produced by ERS-1! Geocoding the complex SAR images as an element of the total processing chain was another new DFD methodological priority, developed and realized under the leadership of Gunter Schreier at the end of the 1980s. It lives on in Achim Roth’s team, where it was elaborated step by step using sophisticated mosaicking techniques.

Jörg Gredel presents Federal Minister of Research Heinz Riesenhuber with the first ERS-1 image in 1991

Parallel to this development, a dynamic young engineer who had just completed his post-doctoral thesis came to DFD in 1989: Dr Richard Bamler. He had the task of developing the technically entirely different X-band processor for the X-SAR instrument that was provided by DLR and flew on the Space Shuttle Endeavour in two very successful missions in April and October1994 Together with Hartmut Runge his team was able to master all the challenges connected with processing X-band SAR data, and our processor decisively contributed to the great success of the two Shuttle missions. Under Richard Bamler’s leadership – he soon had his own department for developing algorithms and methodologies (DFD-AV) – SAR processor development received an extraordinary impulse. Shortly thereafter in 2000 we could also process the complex SRTM data with our own interferometric SAR processor to generate elevation models. And they were of a quality never achieved before. Rumours went through the international scene that could not be overheard – and they reached JPL in Pasadena. That was the breakthrough. Just a few years later with TerraSAR-X starting in 2007 and TanDEM-X in 2010 we could again provide evidence of our knowledge and technology for all to see. It was evident in all the conceivable processing modes and in data products of outstanding quality – available already two days after satellite launch – made possible with our SAR processors. Both satellites are still today our DLR flagships.

A photo from the early years of SAR data processing taken at a meeting with ASI colleagues in Matera:
DFD colleagues Hartmut Runge (2nd from left), Wolfgang Noack (6th from right), Gunter Schreier (4th from right.) Richard Bamler (3rd from right).

At the same time another area of processor development emerged connected with atmosphere spectrometers. The initial spark came from GOME  on ERS-2. Starting around 1993 Wolfgang Balzer had the difficult task of developing the standard ozone processor for the GOME instrument. (It was later also used for the SCIAMACHY instrument on Envisat.) Nobody had the least idea of how to go about it. However, these colleagues learned fast; they networked efficiently with the stars of the scene in England (like Dr Robert Spurr), with the group around the instrument’s principle investigator, Prof. John Burrows in Bremen, and others. A team of high-performance colleagues was assembled that energetically tackled the new topic and had to endure many an unpleasant discussion about themselves. Dr Diego Loyola was one of the colleagues involved from the very beginning.

In my own value-adding group we stumbled around for a while trying to develop well-made, integrated global ozone products on time. A young physicist from Wuppertal, who knew what approach was required. boosted our efforts starting in 1996: Dr Michael Bittner. From the first solutions a whole series of highly complex atmosphere products came into being that also took models into account, and that work continues to the present day. This later brought us a mandate as World Data Center for Remote Sensing of the Atmosphere (WDC-RSAT) after a rigorous assessment by the DFG in 2000.

All these developments were ingredients for a promising mixture that laid the groundwork for something innovative in content and structure. All the same, the 1990s were, on the whole, a difficult time. We had financial worries and challenges to overcome because institutional financing was significantly reduced and increasingly more effort had to be put into finding third-party funding. And just at this time our director Winfried Markwitz had to slow down and, in the end, even prematurely retire from DLR for health reason. A new director had to be found – and was soon chosen, Prof. Christoph Reigber, a prestigious researcher and successful department head at GFZ-Potsdam who had gained a reputation working with Champ satellite data (“Potsdam Potato”) and on other topics. There had long been professional and personal links to DFD. Mr Reigber even introduced himself as “DFD director in spe” when he welcomed then federal president Roman Herzog on his visit to DFD. But it didn’t happen – and things got complicated. The German states of Brandenburg and Bavaria could not come to an agreement about the conditions for his transfer, so we began 1996 without a director. Dr Wolfgang Mett, who had come to us from BMFT a few years previously to work on strengthening the area of remote sensing of the atmosphere at DFD, was appointed acting director. He did not have an easy time holding the different strands together. All the department heads had strong personalities and long experience at DLR (except for the two greenhorns Bamler and Dech), and each of them had his own quite convincing ideas and pursued reasonable goals: Dr Hans-Diedrich “Dieter” Bettac (†) in Neustrelitz, Jörg Gredel, Hermann Rister, Klaus Reiniger (†), Kurt Schmidt and Hans Walek in Oberpfaffenhofen. I often met with Klaus Reiniger at that time and we shared our concern about how things were going with DFD and its cohesion. Then one day Klaus said, “you and Richard should do it”. That startled me and at first I quickly brushed aside the idea.

Assuming more responsibility

At the same time the DLR Executive Board was also debating what should be done with DFD as to its organisation. Prof. Walter Kröll was the board chairman then, and Prof. Achim Bachem the board member responsible for space affairs, and thus our board liaison. Only later did I learn that Prof. Ulrich Schumann, long-time director of the Institute of Atmospheric Physics, had encouraged Prof. Kröll to refrain for the time being from going through another possibly inconclusive recruitment round, and to first place some confidence in “the young fellows” Richard Bamler and Stefan Dech (both with postdoctoral qualifications and employed in complementary roles). A short time later, in spring 1998, Prof. Bachem asked us to come for a discussion. Richard Bamler and I already had some premonition of what was on the horizon, and it was soon enough clear to us that we couldn’t get out of it and had to take over responsibility for DFD.

Leadership duo Stefan Dech and Richard Bamler in the early years of the institute cluster

We were requested to develop for presentation to the Executive Board a concept for the future including both scientific/technical and structural aspects, which kept us fully busy for several months. Especially as to structure we proposed a number of necessary changes to quickly give DFD a stable economic foundation again – some of them quite painful, and requiring a great deal of our energy and powers of persuasion. Entire departments were closed and existing ones modified and restructured. On May 26, 1998 (the year Gerhard Schröder was elected federal chancellor) Richard Bamler and I presented our concept at a meeting of the entire Executive Board, and on the same day we were appointed joint heads of DFD. Those whom we had proposed as department heads were likewise reappointed. We had already with heavy hearts said good-by to the idea of ourselves remaining in charge of ”our” departments DFD-FE (Remote Sensing Applications) and DFD-AV (Algorithms and Methodologies), into whose development we had put so much heart-blood, Dr Harald Mehl became my successor for Remote Sensing Applications and Dr Tom Rother from Neustrelitz became Richard Bamler’s successor to head DFD-AV. Starting in 1998 the crew comprised of the other department heads was as follows: Jörg Gredel: Strategic Planning and Resources, Gunter Schreier: Business Development and Marketing, Dr Klaus Reiniger: International Ground Segment, Dr Dieter Bettac: National Ground Segment, Dr Kurt Schmidt: Information Technology, and Dr Michael Bittner: Climate and Atmosphere Products. Dr Wolfgang Mett transferred to Cologne to head Mr Bachem’s staff of experts for his Executive Board role.

It was of course a milestone for our concept that from that moment on DFD would also be engaged in research activities and directed as an institute and no longer a “facility”. Until mid-1999 there followed for us a very intensive first working year in our new function, characterized by a collegial atmosphere of renewal despite more than enough in-house work to organize and establish effective processes – often enough down to the last detail. Collaboration with Richard Bamler in our shared leadership role was always close and trusting as well as pragmatic. That was already at that time a key to the success of DFD, and up to the present time also of the entire EOC.

At that time DFD had two other branch locations besides the one in Neustrelitz. In Cologne Dr Robert Backhaus was in charge of a group of scientists with recognized research and project experience. It remained part of the Remote Sensing Applications department. A new addition was a group under Dr Maria von Schönermark in Berlin, who strengthened methodologically the newly created department Climate and Atmosphere Products. DFD then had a total of about 260-270 staff members and an extremely wide range of expertise.

Our institute retreat in 1999 had on its agenda “The future of Data and Information Management in the Ground Segment”. Kurt Schmidt and his colleagues, led by a young team leader, Eberhard Mikusch, convinced us that the adopted path and the developments to date were worth every invested penny (we still had the D-Mark!) and would soon pay off. In due course we successfully continued these developments, and indeed they did pay off. DIMS is today not only a load-bearing pillar of DLR earth observation missions, but in many cases also of ESA’s.

Exhausted but pleased after the 1999 cluster retreat on the island of Frauenchiemsee.
From left to right: Frank Steiper, Erhard Diedrich, Willi Wildegger, Eberhard Mikusch, Stefan Dech, Klaus Reiniger (†), Tom Rother, Susan Giegerich, Kurt Schmidt