2018 ESIP Summer Meeting

During the 2017 ESIP Summer Meeting, the Information Quality Cluster (IQC) sponsored a plenary panel session and a breakout session focused on Earth science data uncertainty. Expert panellists presented to the ESIP audience key aspects of scientific quality and addressed questions such as "How is uncertainty determined and characterized in the products of their research or application? What are the major side effects and limitations of common statistical techniques used to quantify and characterize uncertainty? What is the impact of uncertainty on the quality of their data products? How is data uncertainty accounted for when multiple sources of data are spliced and woven into a single product? How do they document and convey the information about uncertainty to other scientific users? What is the best way of conveying uncertainty to the (possibly skeptical) public?" Following considerable discussion during the breakout session, a one of the key action items recommended was that a clear understanding of the concept of uncertainty, and its communication to users was essential, and that the IQC should develop a white paper to satisfy this objective. During the 2017 ESIP Winter Meeting, the IQC held a session titled "Formulation of a White Paper on Earth Science Data Uncertainty" where further presentations by experts was held about the mathematical basis for uncertainty as well as uncertainty from the point of view of scientific data producers and applications' users, followed by subgroup discussions for formulating the white paper. An outline of the white paper has been developed and reviewed by IQC members and a number of individuals have signed up to be co-authors and/or reviewers.

The purpose of this breakout session is to continue this progress by acquiring a final set considerations and recommendations from additional domain experts spanning multiple disciplines of Earth science and data science. This will be followed by a working session to finalize the outline and establish the final points and issues that should be addressed by the paper.

The following is a list of speakers, titles and brief descriptions of the talks to be delivered by the panelists:

Michael Little – NASA AIST Program
Presentation Title: “An AIST Program View of the Significance of Uncertainty in Data Exploitation”
Presentation Summary: While uncertainty is important, everyone has a different idea about what it means, how to characterize it and what it should be used for. I hope to provide a brief overview of how I think it’s used in programmatic decisions.

Jeff Privette – NOAA/NCEI
Presentation Title: “NOAA/NCEI’s Approaches to Informing Users on Uncertainty”
Presentation Description: Determining uncertainty in environmental data is critical but typically challenging and expensive. Therefore, many scientists either ignore, guess at, or use validation studies to estimate it. This can lead to misleading results in climate monitoring, especially in derivative or time-series results such as climate anomalies or rankings. NCEI is adopting a policy to inform data users of what uncertainty, validation, and quality assurance has been applied so that users are informed prior to ordering the data.

Faozi Said – NOAA/NESDIS
Presentation Title: “Leveraging Cal/Val to Effectively Communicate the Physical Data Limitations”
Presentation Summary: Using a real world example, we explore on a high level the calibration and validation (Cal/Val) challenges we face and how the possible issues and uncertainty in the data are communicated to the end user.

Jonathan Hobbs – Jet Propulsion Laboratory
Presentation Title: “Probability as a Foundation for Data Uncertainty: Applications in Remote Sensing”
Presentation Summary: Earth science data records often include products that combine models with indirect observations of their quantity of interest. Probability serves as a foundational tool for characterizing uncertainty for these complex methods. This presentation will illustrate these ideas in the context of remote sensing retrievals.