Dr. Frédéric Mélin
EU Joint Research Centre (JRC)
Dr. Roland Doerffer
Helmholtz Zentrum Geesthacht
Institute for Coastal Research, Max-Plank- Str. 1
Scientific and Programmatic Background and Rationale
Ocean colour remote sensing data are affected by errors and uncertainties as with any other observational data. The main contributing factors are measurement and calibration errors as well definition issues concerning the nature of the object derived from top-of-atmosphere reflectance spectra and the volume of water from which the measured quantity (concentration, inherent optical property) is derived. Factors which have a significant influence on the accuracy of remote sensing data are, for example:
- correction of the influence of the atmosphere,
- bio-optical model, which is the basis of the retrieval algorithm,
- the vertical distribution of the quantity to be derived along with the vertical penetration depth of the radiation in the spectral bands, which are used in the algorithm,
- sub-pixel patchiness of the water constituents,
- reflection by sea bottom,
- uncertainties in in situ data, which are used for calibration and as “true values”,
- different water bodies represented in in situ and remotely sensed data.
All of these and further factors determine the deviation between the value of the variable, which has been derived from the top-of-atmosphere reflectance, and the unknown true value. For determining the uncertainty, the following procedures are possible:
- the uncertainty can be determined from statistical comparisons between in situ and remote sensing observations on global or regional scale or based on different water types,
- the uncertainty can be determined from round-robin comparisons of different algorithms and in situ or synthetic data,
- the uncertainty can be determined by sensitivity studies using radiative transfer models,
- the uncertainty can be determined on a pixel-by-pixel basis directly by fitting the reflectance spectra of a forward model radiative transfer model or its proxy to the measured one.
All these methods complement one another. For the user of remote sensing data the knowledge of the uncertainty or confidence range is of high importance. For the assimilation of remote sensing data into numerical oceanographic models, the confidence range determines how well the model results have to meet the observations. Up to now, most of the remote sensing data are provided without uncertainties. In most cases only the general knowledge from validation of remote sensing data is provided, or flags indicate pixels which are out of scope or which should be used with caution.
Due to the demand for improving the information about uncertainties, an IOCCG working group was established with the objectives to analyze and discuss the issue of uncertainties in ocean colour remote sensing, and to propose procedures to determine uncertainties and ways to distribute the information as part of the remote sensing data products. Since this is a general problem affecting all providers of ocean colour data, the IOCCG, as an international organization, has chosen to address this issue.
Terms of Reference
- Analyse factors which determine uncertainties for various water types.
- Analyse the problem of uncertainties for different product types, such as IOPs, k-values, z90 depth, and Chl concentrations.
- Analyse the problem of the propagation of uncertainties in L3 and merged products.
- Review, analyze and document methods which allow us to determine out of scope conditions and uncertainties including flagging.
- Develop and compile a data set to test uncertainty and out-of-scope algorithms.
- Recommend procedures to be implemented in ground processors and formats to present errors/uncertainties in data products.
- Summarize the results and recommendations in the form of an IOCCG report.
Proposed Working Group Membership
- Marc Bouvet, European Space Agency (ESA)
- Prakash Chauhan, ISRO, India
- Roland Doerffer (Co-Chair), HZG/Brockman Consultants, Germany
- Stephanie Dutkiewicz, Massachusetts Institute of Technology (MIT), USA
- Hiroshi Kobayashi, University of Yamanashi, Japan
- Frédéric Mélin (Co-Chair), EU Joint Research Centre (JRC), Italy
- Menghuan Wang, NOAA\NESDIS, USA
- Jeremy Werdell, GSFC, NASA, USA