NPL is leading a consortium of European technical experts t review existing quality assurance/control practices and then to consider and propose strategies for improvement for the future as part of an ESA Strategy.

The establishment of a joint EU/ESA programme on Earth Observation was a corner stone within Europe. It brought together the principal key users of data (EU) with the data generators (ESA) to develop a coherent strategy to ensure that Europe fully exploits the potential of EO data both from a scientific and commercial perspective within the next decade.

The establishment of the joint EU, ESA GMES initiative is a measure of the importance governments and the Commission have given to issues that EO data can help to address.

Examples of applications for EO data include:

• Understanding the drivers for climate change:
• To inform European policy on necessary steps to protect the living conditions of citizens, whilst not putting at risk European competitiveness with ineffective or unnecessary emission controls.
• To remove any equivocation over the affects of human activity on the Earth’s climate, to strengthen the European negotiation position with countries, which are not yet convinced of the urgency to reduce industrial and transport emissions.
• Monitoring the effects of climate change:
• Monitoring European coastal zone to manage the effects of sea level rise.
• Monitoring biodiversity.
• Monitoring of food production, pests and disease across Europe.
• For the security of food production.
• To measure the impact and police subsidies provided by the EU to food producers.
• To measure the impact of the release of GMO on biodiversity.
• Monitoring air and water quality for the health of citizens.
• Monitoring and policing pollution so that those responsible for major pollution events can be challenged for compensation.
• Management of major calamities, e.g. cyclones, floods and storms.

These issues affect decisions on investments and regulations that will have significant economic impact in Europe and require data that is unequivocal, from which international protocols that tackle issues at a global level can be agreed, and ultimately will require data that is robust enough to be challenged in court.

NPL is leading a consortium of European technical experts to review existing QA/QC practises and then to consider and propose strategies for improvement for the future, as part of an ESA study.

The experts were selected to cover all sensor domains: Optical, SAR, Atmospheric Chemistry and Altimetry as well as ground segment activities, to ensure that the whole EO sector was included.

It should also be noted that in practise any long-term QA/QC strategy should also include pre-flight activities as well as post-launch since these can have significant impact in the ability of any mission to meet the data quality aspirations of its intended customers.  This aspect is not covered by this study.

One of the early findings of this study is that the EO community are keen to have improved QA and whilst not wanting necessarily to undergo formal accreditation to to quality systems such as those of ISO, they are keen to make use of many of the principles they contain e.g. written procedures, uncertainties, traceability and independent validation.  Thus there is clear recognition and broad acceptance that there is a need to establish a full QA plan for each data product.

The reliable supply of high quality EO data products is ultimately dependent on the establishment and operation of a full QA plan for all stages of an instrument’s development.  Such a QA plan when fully implemented not only helps to improve the efficiency of supply but also helps to trace and quickly identify sources of error, which will inevitably occur with such complex processes.  The establishment of such QA activities is one of the major steps towards the ultimate goals of formal QA systems like ISO 9001.

In many of its activities ESA, in common with other space agencies, can point with confidence to such QA activities and the efficient production of data products, which meet the users expectations.  For example, there are always QA/QC plans prepared for every EO mission and these will be the principle source of information for much of this study.   However, as the demands of the user increase, the level of detail at which these QA activities are carried out similarly needs to increase.  This is particularly important in the case of accuracy; a critical issue for many scientific studies, most notably, climate change.

Schematic diagram summarising the activities carried out to QA a data product.

Figure 1 summarises the basic steps involved in the production of a typical level 2 data product.  In this diagram the satellite instrument produces some data products on-board.  It does this using an algorithm that converts a primary measurement signal e.g. digital counts into a level 1 or level 2 product.  The algorithm functions through using a series of inputs some based on pre-flight calibrations others from other satellites or instruments.  These “instrument produced” products are then sent to ground where they may be distributed directly or undergo further processing or checks using another algorithm with other inputs.  When completed these products are released to users and archived.  To add confidence/validate this process, checks are made through a series of cal/val support activities.  These are designed to measure similar quantities to those of the satellite instrument geo-located in time and space.  Such comparisons can be used to verify the processing steps of the satellite instrument and in the event of differences provide new information to the “instrument calibration team” which can then decide if they should “revise” any of the inputs to the data processing algorithms.  In figure 1, the link between cal/val support activities and User is because often these have some degree of commonality, at least in the present implementation.  As we move towards the full implementation of GMES and the uptake of its products/services, the users are likely to become more remote and isolated from the process, simply specifying/buying products like a car with little interest in evaluating or refining its manufacture.

The cal/val activities whilst predominantly occurring during the commissioning phase of the mission continue to a lesser extent throughout mission life, both to monitor change, but also to take account of any new information which may be the result of a new technique/algorithm or to test an aspect of the mission performance in some extreme of its operational boundaries.  Any decision to change the calibration coefficients or any processing step for an operational mission will require significant justification given the cost and potential confusion to end users of the data products.  Consequently the QA/QC processes need to be extremely robust and their configuration managed to ensure that the value of any such change can be fully quantified.

Therefore in terms of QA it is essential that every aspect of such a process is well documented and where appropriate, demonstrable to an independent reviewer.  In particular, where any measurements, calibrations or input variables are required as part of the process each should have associated with it an uncertainty.  In addition, the value of the measurand (e.g. spectral radiance) or input variable (e.g. molecule absorption cross-section) should, where appropriate, be able to demonstrate full traceability back to the International System of units, (SI) in a manner acceptable to the international community.