Initial Ideas on Geo-Science Use Cases    (2JX9)

http://ontolog.cim3.net/file/work/SOCoP/Pictures/GEOSS%20Data%20Semantic%20View.jpg    (2RGR)

Use Cases capture Geoscience interests and requirements. Ideally a Use Case describes:    (2RUF)

Version 2 of GPDD, released in July 2010, is the largest collection of animal and plant population data in the world, it brings together ~ 5,000 time series in one integrated database which can be searched by:    (2RQ2)

This latest version has:    (2RPX)

Both factors make GPDD a good candidate to expand the use of such data beyond fixed queries and allow integrated use with other data sets. One might look, for example, how animal and plant populations are being affected as part of changed land use.    (2RPZ)

Indeed a primary use case being considered for SOCoP INTEROP is Land Use and Land Cover.    (2RQ0)

One issue is the type of classification that support the dynamics of geo-science semantic uncertainty of geographic data and how this affects studies of land use and land cover. This Use case was discussed at some length at the Dec. Workshop and notes are available. For example there is an interest in land use dynamics and how land cover changes when we move from one state or region to another (e.g. Arkansas Vs. Pennsylvania). The climate is vastly different and in such moves and this complicates communication between researchers.    (2RGU)

Another example discussed for use concerns a land cover map on Somalia.    (2RGV)

Such examples call out the problem of top-down definitions of geo-features. One would like bottom up generated feature categories which may provide better semantics or land use.    (2RGW)

Currently several SOCoP members are working on a Use Case around such examples; along with the vocabulary, data and ontology issues. Progress is expected in April and May, 2011 so that the issues can be discussed at a June workshop. A example is work on landscapes and landforms. See Toward Foundational Ontology of the Landscape by Gaurav Sinha with David Mark http://oak.cats.ohiou.edu/~sinhag/docs/papers/SinhaMark_GIScience2010.pdf    (2RGX)

Other Use Cases (many interrelated) under consideration include:    (2RGY)

Environment    (2ROD)

Environmental data is key to many topics of interests including the spread of disease, land use, marine science, ecology (such as researcher in the NSF Long Term Ecological Research program), Biodiversity (such as testable using GPDD) etc.    (2ROE)

There is a relevant ontology for environment called Envo (http://www.environmentontology.org/) but there are issues with the conceptual distinctions it makes and has these might be enhanced for wider use.    (2ROF)

Some of this ontology builds on older Science Environment for Ecological Knowledge (SEEK) Project - a five year initiative that worked on creating a cyberinfrastructure for:    (2RPV)

Climate Change    (2ROP)

There are many research issues involving Big Data integrations and use of multiple types of data such from instrumental, paleoclimatic, satellite, and model-based sources. The USGS, for example is interested in' long-term alteration in the characteristic weather conditions of a region, such as changes in precipitation and temperature. A broad view of that research is at: http://www.usgs.gov/global_change/    (2ROQ)

Research sub-topics within this include:    (2ROR)

Assessing Historic Climate Variability and Change by Regions and Land Types    (2ROS)

There is a general problem with adequate integration of more recent,“abundant & precise” spatial data data from traditional means along with historical datasets. Such integrations are need to make progress on and "solve" complex problems such as climate variability. A challenging problem is assessing relatively recent climate in relation to forest composition, structure, as well as animal, plant (GPDD) and human health. Historic analysis may involve model-based “re-analyses” using existing data in a new way. For example, hybrid model-observational data sets can be created by assimilating local observations into a regional (or even global) forecast model for a given time period (such as a 30 year perspective 1980 to 2010). Such broader models are harmonized to provide more physically consistent/expanded depictions of the observed time-evolving climate system. Such harmonized models have become very useful for climate system research. A challenge for such reanalysis requires sustained efforts that include versions of models that are fixed in a usable way. As noted in Climate Data Challenges in the 21st Century, Overpeck et al, 2011:    (2RX8)

"Future reanalysis methods will include more diverse observational data types (such as atmospheric chemistry, biospheric, oceanographic, and cryospheric data) and longer time scales (including paleoclimatic time scales)."    (2ROT)

Evaluating Ecosystem Effects    (2ROU)

A challenge is to evaluate natural process and disturbance effects across locations and their diverse physical and ecological attributes of natural resources. Where for example are animal and plant population fluctuating rapidly over time? For marine species how is this influenced by ocean and air temperatures?    (2ROV)

Marine and Oceans Observations    (2RGZ)

The OCEAN OBSERVATORIES INITIATIVE or OCI and NSF Ocean Observing Initiative (OOI) provide a basis for this use case. Phenomena and sub-topic areas include:    (2RH0)

See https://confluence.oceanobservatories.org/dashboard.action for a discussion.    (2RH7)

Epidemiology and Disease    (2RHF)

Examples include Dynamic Continuous-Area Space-Time models to track disease spread. This use case would be collaboratively developed with BoK2 member Sean Ahearn who has studied the issues and who has a preliminary ontology. A sewer overflow is discussed below which has health implications.    (2RHG)

Ecological Synthesis, Disasters and Hydrology    (2RH8)

Relevant data and even ontologies may be available from USGS and CUAHSI Hydrologic Information System    (2RHH)

Disaster situations include flood events along with:    (2RH9)

GEON Project Use Case    (2RHB)

Some work can build on the GEON effort which used 3D structural models with accompanying physical parameters of density, seismic velocities, geochemistry, and geologic ages    (2RHC)

The site www.geongrid.org describes some of this work which provides a start on Geologic Ontologies, Rock Classification Ontology and Spatial ontologies.    (2RHD)

Applied eSCIENCE USE CASE (combined sewage overflows)    (2RHE)

An INTEROP Use case may be build on the eScience use case (see http://www.ncsa.illinois.edu/~yongliu/documents/MSeScience2008-EventOntologyPaper-final.pdf) which centers on the optimal control of combined sewage overflows (CSO) events that are common in aged metropolitan areas. They use the city of Chicago, where storm water drainage and sewage pipelines share the same physical infrastructure. The material below is taken from this article on Event Ontologies.    (2RHK)

When storms hit such areas the capacity of sewage pipelines can be exceeded. If so the system gates of the pipes have to be opened and a combined discharge of sewer and stormwater will go to Chicago's available spaces (e.g., streets) or even residential basements. The result is a combination of unwanted public health problems and flooding issues. There are many types of events happening when CSO events take place, such as storm events (example data from NCDC can be found at http://data.nicar.org/node/40), operational control events (open/close gates). All these events are inherently spatiotemporal and have complex causal relationships and structures. To effectively respond to CSO events, a knowledge base needs to be developed to facilitate event-to-event understanding and move towards event-driven adaptive observation, modeling and control. This use case is similar to many coupled engineering and natural systems based on sensors. For example, ubiquitous sensors are leading toward an Event_Web, merging data from many sources. In a wider context, the architecture of a event-oriented system is similar to other systems, including interactive multimedia systems. The existence of ubiquitous sensor-data will transform , human-computer interactions potentially enabling interactions more akin to musical and dance performance than pointing and typing.    (2RX7)

To develop these systems an event-oriented ontology is necessary to enable open inter-operation of data and metadata from many sources, which will aid the coupling of computational models and the creation of persistent knowledge across multiple disciplines.    (2JXA)