AATSR History and Background

A short history of the ATSR programme compiled by David Llewellyn-Jones and Chris Mutlow:


The need for precise and systematic measurements, on a global basis, of geophysical parameters such as sea surface temperature was recognised almost 25 years ago by the scientists from the UK, Australia and France who proposed the Along Track Scanning Radiometer (ATSR) in response to an Announcement of Opportunity (AO) from the European Space Agency (ESA) for experiments to be flown on the ERS-1 satellite. ATSR was designed to tackle the issue of observing global sea surface temperature (SST) from space to the very high levels of accuracy and precision required for climate research as specified by the World Climate Research Programme (WCRP).

These requirements were considered and endorsed at a workshop held at the Rutherford Appleton Laboratory in 1980. The workshop was convened at the instigation of Professor Sir John Houghton, then Deputy Director of RAL, subsequently Director-General of the UK Met Office and the first Chairman of the Inter-governmental Panel on Climate Change (IPCC). The workshop was attended by representatives of major meteorological services and space agencies, as well as the wider scientific community. Invitees included: Francis Bretherton (NCAR), Carl Wunsch (MIT), Pierre Morel (CNES) Guy Duchossois (EO programme manager, ESA), Stan Wilson (head of Oceanographic Applications, NASA) and a number of representatives of the active UK oceanographic and meteorological research community.

Following the requirements identified and endorsed at this workshop, the continuing importance of global SST as a key variable in the climate change debate is reinforced by many programmes, for example the CLIVAR and IGOS Programmes, which give SST as one of the critical geophysical quantities for their climate studies. They also specify the levels of required accuracy as 0.1-0.3K. ATSR is one of the few sensors, and arguably the only current sensor, that can deliver the required SST data of this quality.

In response to these requirements and following a number of feasibility assessments, it was realised that these required levels of measurement accuracy could be achieved from space.

Up until the development of ATSR, previous sensors had relied on the use of multi-spectral methods to provide the atmospheric correction necessary to carry out accurate SST retrievals. However this method has significant limitations, especially in the presence of aerosols and other atmospheric anomalies, and thus became heavily dependant on the use of in situ data, of highly variable and sometimes questionable quality and sporadic coverage, to tune the algorithm, rather than rely on the physical basis of the measurement methodology. Theoretically, the technique of along-track scanning offered the capability to generate an improved atmospheric correction that increases the skill and independence of the SST retrievals compared to that possible with the previous sensors, namely NOAA’s Advanced Very High Resolution Radiometer (AVHRR) and successor instruments such as MODIS and VIRI.


The ERS-1 ATSR instrument was built as a new and experimental instrument concept. Its design added the novel method of along track scanning to the already established multi-spectral techniques used by AVHRR to produce new high-quality SST observations with an improved atmospheric correction that is more robust against the corrupting effects of the atmosphere. ATSR also introduced a substantially improved internal calibration system, at a time when very serious concerns were emerging about the overall stability and basic validity of the on-board calibration schemes of existing sensors. Also, ATSR-1 carried, for the first time in space, a new UK-developed closed-cycle mechanical cooler, which allowed the detectors to function at their optimum temperatures of around 80-100°K, without the accommodation problems usually posed by a passive cooler.

These technical features, embodied in an excellent optical design, resulted in the deployment of ATSR-1 as a near-ideal radiometer in space, of a quality that has not been equalled in any civilian thermal imaging space sensor other than ATSR-2 and AATSR.

Thus the design concept for the ATSR was developed and the instrument successfully proposed, in 1981, to ESA in response their Announcement of Opportunity for additional instrumentation on board the ERS-1 satellite.

Initially, the plans were for a single ATSR instrument to support process studies in climate research, but after the successful launch of ATSR-1 this developed into a multi-instrument programme covering a period of over a decade. This development allowed the main scientific objective of the ATSR Programme to broaden to include the monitoring and detection of climate change, as it would produce a long-term global sea surface temperature (SST) data set of sufficient accuracy and stability that it can be used for quantitative climate change detection and monitoring. This requires a series of instruments producing SST data with an accuracy and precision of better than ±0.3K (1?) over a 10-15 year period. Furthermore, in addition to needing high-quality instruments, it also requires that careful attention to detail is needed in the processing of the data sets, and in calibration and cross-calibration of the sensors.

At this point it is worth noting that, in addition to the scientific achievements which have emerged as a result of this programme and which are described elsewhere there have been some very important “spin-offs” from the ATSR programme. Firstly, when ATSR-2 was proposed in response to ESA’s AO, new UK research council funding of £2m per annum was obtained as a result of a subsequent bid madre by SERC to central government. This money now comprises part of the UK’s present EO programme which was subsequently transferred to NERC. Secondly, as it became clear that the ATSR series could provide input to the climate record, the UK Department of the Environment (now DECC) was sufficiently convinced to commit funding for the third ATSR (AATSR) which is now producing excellent data. This is the first time in theUK (and probably in the rest of Europe) that a new “user depertment” of government has committed funds to a space sensor procurment programme.

There has also been significant benefit to UK industry through knowledge transfer as a result of the industrialisation of the ATSR programme. Although ATSR-1 was not designed for transfer to industry, rather as a “one-off” scientific experiment, successive ATSR’s were built with increasing involvement of industry, culminating with ASTRUM (UK), as prime contractor, buiding the AATSR sensor and coolers, while the on-board calibration targets were built by AEA Technology. In the software area, there has been similar knowledge transfer to European industry concerning the AATSR processer. ATSR-1 was developed by a consortium of science institutes in the UK, France and Australia. These were in the UK Rutherford Appleton Laboratory, Mullard Space Science Laboratory (part of University College London), the Atmospheric Physics Department of Oxford University, and the Meteorological Office. The project was largely funded by the UK Science and Engineering Research Council; however Australia through CSIRO’s Division of Atmospheric Physics and Research and its Office of Space Science and Applications made a substantial contribution to the programme, involving hardware as well as the collection of validation data. In France, an additional microwave part of the instrument, incorporated at ESA instigation to provide a total water vapour range correction for the ERS-1 radar altimeter, was developed, with CNES funding, by the Centre de Recherches en Physique de l’Environment Terreste et Planétaire (CRPE).

As already mentioned above, the ATSR-1 instrument was developed in response to an AO from ESA, and was specifically designed to provide the SST information urgently needed for the increasingly demanding data requirments of climate research, as well as to produce stable and well calibrated image data sets for use in a wide range of atmospheric and oceanographic studies, for example the detection of clouds and aerosols. The first instrument, ATSR-1, was launched in July 1991 as an experimental, infrared-only sensor, with channels at 1.6, 3.7. 10.8 and 12?m designed to make accurate and precise observations of SST for climate research applications, and it successfully proved that the ATSR concept was capable of delivering the required SST data. Apart from the unfortunate failure of the 3.7?m channel in May 1992, ATSR-1 performed well until the demsie of ERS-1, some time after the launch of ERS-2.

Robust and accurate calibration coupled with long-term stability are key attributes need in sensors producing data for climate change research, and other quantative remote sensing applications. Therefore, in-flight calibration and and long-term stability are major deisgn drivers for the ATSR sensors. To this end, both ATSR-1 and -2 carry a two blackbodies for infrared calibration, in addition to the reflected-wavelengths calibration system mentioned above.

The primary objectives of the ATSR-1 and -2 missions are:

  • The collection of a high-quality record of spatially-averaged SST data at 30 arc-min and 10 arc-min resolutions.
  • The collection of well calibrated infrared brightness temperature and visible reflectance images of the Earth's surface at the full instrument resolution (nominally 1km) for use by oceanographers and other scientists
  • To provide new information about clouds and aerosols in the atmosphere
  • The development of processing software for ATSR was more problemmatical. The instrument was funded on the clearly stated assumption that the task would be externally funded and implemented through the ESA system, which in turn was based on the Prodessing and Archiving Fcilities (PAFs). In the event, foa a variety of reasons, the PAF with responsibility for processing ATSR data was not able to develop an opertional processor of adequte performance and, as a fall-back, the prototype software available at RAL was hastilly integrsted into a continuously running processing chain , known as SADIST. It is important to recall this situation, as it has resulted in a rather slow developement of ATSR data-products and delayed their scientific utilisation, despite the very high quality of the sensor’s data record.

    ATSR-2 and Beyond

    ATSR-2 was a further development from of the ATSR sensor which that added a visble capability to the sensor by the addition of channels at 0.55, 0.67 and 0.85?m to extend the capability of the instrument for land remote sensing, particularly for odeveloping effective atmospheric correction schemes, and cloud property retrievals. ATSR-2 was launched in April 1995.

    Also, ATSR-2 carries a innovative on-board calibration system for the visible, or “reflected”, channels. A Russian opal diffuser plate is illuminated by the Sun once per orbit and allows the instrument to calibrate its visible detectors using Sunlight.

    The ATSR series was continued by the launch of the Advanced ATSR instrument on ESA’s Enivsat satellite in March 2002, this sensor was now the operational AATSR instrument and it continued the SST data continuity from ATSR-1/2 and the visible mission of the ATSR-2.

    Routine ATSR-1 data collection started on the 1st August 1991, and ended when the ERS-1 satellite was hibernated in June 1996. From then until the loss of the ERS-1 satellite on the 10th March 2000, ATSR-1 collected data regularly every 70 days during each of the 3-day reactivations of the ERS-1 satellite from hibernation. Thus, ATSR-1 has provided 9-years of SST observations, including a period of nearly 5-years continuous coverage from August 1991 until June 1996.

    Data collection from ATSR-2 started in May 1995 and, apart from a temporary shutdown due to a scan mechanism drive anomaly from the 22nd December 1995 until early July 1996, ATSR-2 remained in near-continuous operation until the 31st January 2008.

    From ATSR-1, -2 and AATSR, nearly 20 years of continuous high-quality SST data is available covering the period from August 1991 until the 8th April 2012. Work has still continued on the improving the algorithms to account for variations in volcanic aerosols and instrumental effects (e.g. the ATSR-1 detector warm-up), and to improve the cloud detection scheme, particularly in areas where marine stratus are prevalent, so that the “best” SST data set can be produced from the joint mission. This is painstaking work which has required attention to detail, and which involves statistical analysis of large quantities of spatially- averaged and image data by the project team.