Experiences Developing a Semantic Representation of Product Quality, Bias, and Uncertainty for a Satellite Data Product Patrick West 1, Gregory Leptoukh 2, Stephan Zednik 1, Chris Lynnes 2, Suraiya Ahmad 3, Jianfu Pan 4, Peter Fox 1 (1)Tetherless World Constellation (2)NASA Goddard Space Flight Center (3)NASA Goddard Space Flight Center/Innovim (4)NASA Goddard Space Flight Cetner/Adnet Systems, Inc. EGU

Outline of Presentation Current Issues and Prior Work Definitions Our Approach to resolving these issues Our Focus Area –Multi-Sensor Data Synergy Advisor (MDSA) –Aerostat Applying our approach in the focus area Conclusion Questions 1

Issue climate model and various environmental monitoring and protection applications have begun to increasingly rely on satellite measurements. research application users seek good quality satellite data, with uncertainties and biases provided for each data point remote-sensing quality issues are addressed rather inconsistently and differently by different communities. 2

Problem Space Graphics, information here on how this relates to MDSA, DQSS, and AeroStat. 3

Definitions Product Quality: is a measure of how well we believe a dataset represents the physical quantity that it purports to. As such, it is closely related to (though not identical to) the level of validation of the dataset. It often varies within the dataset, with dependencies on such factors as viewing geometry, surface type (land, ocean, desert, etc.) and cloud fraction. Cf. Data Quality: Data Quality is typically applied to a particular instance of data (pixel, scan or granule). It describes how well the instrument and retrieval algorithm were able to resolve a result for that instance. 4

Definitions Uncertainty: has aspects of accuracy (how accurately the real world situation is assessed, it also includes bias) and precision (down to how many digits). Bias: has two aspects: –(1) Systematic error resulting in the distortion of measurement data caused by prejudice or faulty measurement technique (GL: modified from IAIDQ site) –(2) A vested interest, or strongly held paradigm or condition that may skew the results of sampling, measuring, or reporting the findings of a quality assessment: Psychological: for example, when data providers audit their own data, they usually have a bias to overstate its quality. Sampling: Sampling procedures that result in a sample that is not truly representative of the population sampled. (Larry English) 5

Focus Area – AeroStat Project 6

Approach semantic differences in quality/bias/uncertainty at the pixel, granule, product, and record levels outline various factors contributing to uncertainty or error budget; errors introduced by Level 2 to Level 3 and Level 3 to Level 4 processing steps, including gridding, aggregation, merging and analysis algorithm errors (e.g., representation, bias correction, and gap interpolation) assess needs for quality in different communities, e.g., to understand fitness-for-purpose quality or value of data vs. quality as provided by data providers 7

Approach Good Quality Documentation (based on standards and controlled vocabularies) is a necessary step to enabling semi-autonomous resource assessment. –Existing standards are ambiguous and not consistently implemented. (STRONG WORDS, NEED MORE DOCUMENTATION HERE, REFERENCES) 8

IQ Curator Model Introduction to it 9

IQ Curator Model Application to our Project 10

Application to Focus Area 11

Conclusion Quality is very hard to characterize, different groups will focus on different and inconsistent measures of quality. Products with known Quality (whether good or bad quality) are more valuable than products with unknown Quality. –Known quality helps you correctly assess fitness-for-use Quality Documentation (Metadata) is a key factor in determining Fitness-for-Purpose 12

References Levy, R. C., Leptoukh, G. G., Zubko, V., Gopalan, A., Kahn, R., & Remer, L. A. (2009). A critical look at deriving monthly aerosol optical depth from satellite data. IEEE Trans. Geosci. Remote Sens., 47, Zednik, S., Fox, P., & McGuinness, D. (2010). System Transparency, or How I Learned to Worry about Meaning and Love Provenance! 3rd International Provenance and Annotation Workshop, Troy, NY. P. Missier, S. Embury, M.Greenwood, A. Preece, and B. Jin. Quality views: capturing and exploiting the user perspective on data quality. Procs VLDB, (PDF) _vldb2006.pdfhttp://users.cs.cf.ac.uk/A.D.Preece/qurator/resources/qurator _vldb2006.pdf 13

Thank You Questions? Contact Information: –AeroStat Project Pages: –MDSA Project Pages: 14