GEOSS Future Products: Recommendations from a Workshop

The Global Earth Observing System of Systems (GEOSS) is a platform to make varied sensor and model data available in an interoperable manner. Data streaming from in-situ and remote sensing sensors (Sensor Web) and models (Model Web) offers great potential to generate a wide portfolio of on-demand and near real-time products.


To meet the GEOSS aim of achieving interoperability of existing and new systems that provide essential environmental observations and information, the “GEOSS Future Products Workshop” was held March 26-28, 2013, at the NOAA Science Center in Silver Spring, Maryland, with participation from more than 50 people from different sectors in private industry, academia, and government agencies coming from U.S., Europe, Africa and Brazil.


The full report is posted online at https://portal.opengeospatial.org/files/?artifact_id=53993


All presentations are posted online at http://www.ogcnetwork.net/node/1872


The scope and focus of GEOSS (Figure 1) provided a basic structure for the Workshop. Sensor Webs are a method to increase accessibility to products from Earth Observation Systems. Model Web is a concept for improving the access to, and interoperable use of Earth System Models. The observations and models support decisions related to societal benefits.  Collectively, the contributed systems of GEOSS work together to: address identified common user requirements; acquire observational data; process data into useful products; exchange, disseminate, and archive shared data, metadata and products; and monitor performance against the defined requirements and intended benefits.


The remainder of this article presents and discusses the recommendations coming from the GEOSS Future Products Workshop.


image 1

Figure 1. GEOSS Observations and Models linked to Societal Benefits. Source: GEOSS 10-Year Implementation Plan: Reference Document, GEO Document 1000R, February 2005.



 


Using the methods of science to guide use of observations and predictions.


GEOSS connects observations, models and decision support for policy decisions and societal benefits.  To design such information systems, a science-driven approach is needed to connect observations and predictions. A presentation during the Workshop by Peter Colohan of the Office of Science Technology and Policy (OSTP) emphasized three principles for GEOSS success: science-based decision-making, open access to data, and international activity.


Professor David Maidment’s presentation to the workshop developed the science-based approach to defining GEOSS by considering the main epistemological methods of science: deduction, experiments and observation. Maidment identified, as the heart of GEOSS, the Model Web as deduction and Sensor Web as observation.  Developing the “Model as a Service” approach will advance how observations and predictions can be used together. An example is presented in Figure 2 for studying river flows.


The Utility of Sensor Web


image 2

Figure 2. Implementing River Flow Models as Services. Source: D. Maidment.



The Sensor Web approach, as reviewed in the Workshop, is delivering results for operational systems but requires additional areas of development in order to meet remaining challenges and to achieve its full vision. Additional developments of Sensor Web should be undertaken in the areas of accuracy and timeliness in providing data products, distributed computing, improving robust services, diverse data models and metadata using standards.


To make products directly usable, Sensor Web services need to be made more accessible to end users, perhaps by defining a “GeoSocial API” that provides clients a layer over the low-level services.


image 3

Figure 3. “What if?” model for Dengue Fever prediction. Source: Geller.



The Model Web Vision


Models codify the best understanding we have about physical phenomena and process. The Model Web vision is: A dynamic web of models, integrated with databases and websites, to form a consultative infrastructure where researchers, managers, policymakers, and the general public can go to gain insight into “what if” questions. Figure 3 provides an example of Model Web for Dengue Fever.  The Model Web vision is seen as a great support for Integrated Environmental Modeling.


The Model Web vision and a roadmap initiated in the Workshop addresses the potentiality of the Model Web in supporting applications by non-scientists/decision-makers; defines a user-driven approach to the Model Web; and puts forward a possible roadmap for the technological development and social developments of the Model Web.. A key result announced at the workshop was an experiment that used the Open Modelling Interface (OpenMI) and Earth System Modeling Framework (ESMF) together. Such testing is a major step to bridging the interoperability in Model Webs. The Workshop results are the basis of a manuscript to be published in the Environmental Modelling & Software journal under the lead authorship of Stefano Nativi.


Interaction between Sensor Web and Model Web


Improving Sensor Web- Model Web interactions will lead to improved accuracy of the observations, as models help sensor systems target most relevant observations, and the predictions, where the most current and relevant observations provide the initial inputs to the models. Currently, there is little integration or coupling between Sensor Web data and modeling.


An element of this interaction is to continue development of methods for workflow. In science, the workflow may not be clear from the start. An issue to be addressed is how to facilitate exploration and provenance, and enable repeatability, portability, and domain specific tool-sets. With time, scientific workflows can become the basis for hypothesis development and peer review of new theories.  Professor Liping Di of George Mason University presented results of experiments in Sensor-Model Integration themes through a Self-Adaptive Earth Prediction System (SEPS).


The Diversity of GEOSS Users


An emergent theme in the Workshop was the need to better understand the diversity of GEOSS users. Figure 4 shows a perspective on the range of GEOSS users. GEOSS cannot be fully defined without identifying the different types of users that build and interact with it. Users are involved in so many steps of data generation (from upload, modification and manipulation to data consumption and development) that it is not possible to talk about the GEOSS user per se. Several presentations in the Workshop addressed approaches to meeting the needs of the diverse groups of GEOSS users. The “GeoSocial API” introduced for Sensor Webs can support the use of simpler client tools with GEOSS.


image 4

Figure 4. Range of GEOSS Users. Source: GEOWOW.



Authentication and other Interoperability Issues


Data resources being provided by organizations requiring registration and login are impediments to seamless interoperability of data. As presented in the workshop by Steven F. Browdy, this is being addressed during the Architecture and Implementation Pilot (AIP) activities of GEO. The desired goals include Single Sign-On (SSO) capabilities within a GEOSS-wide authentication federation.


Several recommendations regarding authentication and other interoperability issues were presented.  These recommendations should be considered by the GEO Standards and Interoperability Forum (SIF) and the AIP activities of GEO:


• Authentication and SSO should be initially focused on OpenID and SAML-2 (Figure 5).
• Authorization should be added to the authentication and SSO federation as soon as possible.
• Use metrics for GEOSS should include some location information to gauge how widespread the reach of GEOSS is.
• The Semantics Registry/Component needs to be deployed into the GEOSS Common Infrastructure (GCI)as soon as possible.
• Linked data and resource crawling should be investigated as a means to solve GEOSS registration issues.
• Broker interoperability should be investigated as a way to serve communities better, and to handle the existence of different types of brokers.
• A tutorial for the Best Practices Wiki focusing on the adoption and use of a consistent schema for unique and persistent identifiers for GEOSS resources should be created.


image 5

Figure 5. Authentication Framework. Source: S. Browdy.



Further Development of the GCI


The GEOSS Common Infrastructure (GCI) allows the user of Earth observations to access, search and use the data, information, tools and services available through GEOSS. Overall, the GEOSS Future Products Workshop generated a healthy and important debate that reflected that data user needs are not fully supported by data providers. There was agreement that further development of GCI is needed, in particular regarding diversity of communities (Figure 6). In order to consider the future products of GEOSS, several recommendations about the current state of GCI and what is needed to achieve expectations were developed:


• Refine the GCI, focusing on the core need of seamless discovery and access to data that is not currently being met. Automated methods, and utilizing the broker approach, should be emphasized.
• Use methods of Linked Open Data to continue increasing incorporation of semantics and linkages of data using a CEOS vocabulary.
• Use of Digital object Identifiers (DoIs) should be considered for GEOSS in part based on NOAA and NASA experiences with the use of DoIs that offer best practices and lessons learned for handling Identifiers.
• Address the anticipated challenges ahead related to general coordination with licenses and user management, types of semantic information, independent community portals from SBAs, mobility, tracking of content quality, etc.
• Perform an introspective assessment to help facilitate the realization of the GCI vision.  The main goal of this assessment will include the formulation of a plan to inform GEO on how to move the GCI forward.


GEOSS Capacity Building


GEOSS Capacity Building recommendations based on the workshop discussions include:


• GEO should continue efforts to help GEOSS users better understand what GEOSS is and what it can do to improve their lives and professional efforts.GEO should continue to develop technical documentation of the GCI and the greater GEOSS architecture, and an expanded set of tutorials to meet the needs of GEOSS providers and GEOSS users. The GEOWOW project should continue tasks that support capacity building, including improved data discovery, easier data access, new data registration mechanisms, and support for the GEOSS Data CORE.
• GEO should continue efforts to help GEOSS users better understand what GEOSS is and what it can do to improve their lives and professional efforts.
• A greater and more effective distribution of information regarding what GEOSS is should be disseminated through social media.


image 6

Figure 6. GEOSS Community Components. Source GEOWOW.



George Percivall serves as chief engineer of the Open Geospatial Consortium (OGC), responsible for OGC’s Interoperability Program and Compliance Program. His roles include articulating OGC standards as a coherent architecture, as well as addressing implications of technology and market trends on the OGC baseline. He is the OGC principal representative to the Group on Earth Observations and is the Coordinator for GEO Task IN-05, GEOSS Design and Interoperability.


Steve Browdy serves as president of OMS Tech Inc. and as co-lead for the Standards group of the IEEE Ocean Engineering Society GEOSS Technical Committee.  He is the component lead for the Design and Interoperability component of the GEO Task IN-05.  He also serves as co-chair of the GEO Standards and Interoperability Forum and as co-chair of the GEOSS Common Infrastructure subgroup of the GEO Data Sharing Working Group.


For more on this subject, please check our previous article on Directing the Evolution of GEOSS Technical Architecture.


 


Topic:



Tags: