Why Taxonomies are Important to the Climate Change Resource Center
This site is all about using taxonomies for the USDA Forest Service's Climate Change Resource Center, to facilitate research of topics that span all disciplines, a way to navigate across disciplines is needed.
Domain-specific taxonomies offer one way to clearly map concepts and terms, and to see how they relate from different scientific and management perspectives.[i] A “climate change taxonomy” would make content more findable. To effectively organize results, and to facilitate browsing and searching, an extended taxonomy of key concepts related to climate change should be used.
Collections of resources can be associated (or "tagged") with each key concept. For advanced users, concepts may be used to filter other concepts: “water temperature” AND “salmon”, for example, combine to form a more limited collection.
Available climate-related taxonomies may be adapted and extended with concepts specific to the Forest Service, specifically to describe the interface between scientific research and natural resource management. To build and maintain such a collection of taxonomies, a development system must be available to collaborators inside and outside the agency, to enable editing and annotation.
A “boundary spanning” taxonomy for the CCRC need not hew to the traditional categorizations of disciplines or organizational bureaus. Rather it should elevate those “boundary objects” that cross disciplines and interests. Boundary objects are those concepts or views of the same thing that are different across communities, where the categorical edges of each community overlap, often accompanied by tension and debate. These are the places where the opportunity for exchange and growth of knowledge across communities is the greatest. Boundary objects that serve this purpose might be key concepts or presentations that bridge multiple disciplines and management layers, or touchstone images or media that similarly offer meaning to a wide spectrum of managers, such as maps or climate-model visualizations. [ii]
A "watershed" may be a boundary object for fisheries biologists and hydrologists. A model may be a boundary object if it is used in different ways by different disciplines. A map could be a boundary object, if its interpretation means different things to different communities. The classic definition comes from Starr and Bowker: "Boundary objects are those objects that both inhabit several communities of practice and satisfy the informational requirements of each of them. Boundary objects are thus plastic enough to adapt to local needs and constraints of the several parties employing them, yet robust enough to maintain a common identity across sites." [iii]
Spatial associations represented on maps, such as habitat or ecologic boundaries, span different disciplines and provide a common reference. Effects of changes in climate, listed as climatic and ecological impacts, are another common set of conceptual objects used to span boundaries in the climate-change dialog:
- Less snow
- Warmer water
- More wildfire
- Longer droughts
- and so on
The Nature Conservancy uses this lens of climate change impacts and consequences to organize its perspective on the issues: http://www.nature.org/initiatives/climatechange/issues/.
To stay abreast of change in the science and management practices responding to climate change, the CCRC must continually look at how the communities it serves self-define, and try to fairly represent the range of perspectives. Asking each community about its needs for climate change knowledge (broadly, including access to human expertise and experience as well as to information) is important in assessing how effective the CCRC is at meeting those needs.
Tools for semantic analysis
One of the tasks CCRC must grapple with is Identifying the key concepts and terminology used to describe climate change from the perspective of each discipline, then see where these key concepts and terms are shared across disciplines (e.g., are boundary objects). This will help identify where there are areas of common interest and attention, and where there may be gaps in understanding key issues across disciplines. To find the "language" of a discipline and organize it, there are several types of tools available.
Lexical analysis tools treat text as data. They count words and sometimes phrases, show the frequency with which these appear, and provide other summary statistics. Textalyser.net provides a free analysis service. See this example for analyzing one of the CCRC topic papers: http://textalyser.net?q=www.fs.fed.us/ccrc/topics/amphibians-reptiles.shtml. This kind of analysis can map terms across disciplines.
Text-mining or data-mining tools do lexical analysis across potentially large bodies of documents and data, identifying potential associations between and among them. CCRC does not presently have any text-mining tools available, but these tools are becoming more numerous and dropping in price, so this may be an option in the not-too-distant future.
Tagging tools match textual content with databases, typically to identify persons, placenames, and categories. OpenCalais is a free tagging service provided by Reuters. It does a pretty good job of finding people's names, place names, but is not adept at categorizing realms of scientific knowledge. You can install an OpenCalais plugin for IE or Firefox (Windows only, not Mac) called Gnosis to see how it works: <http://opencalais.com/documentation/gnosis>. OpenCalais tagging can be performed programmatically, meaning code can be written to add OpenCalais tags to any pages.
GeoMaker is another free tool that finds locations referenced in web pages or text, and from them produces map code or a geo-microformatted list of locations. Try pasting a URL from the CCRC into GeoMaker at <http://icant.co.uk/geomaker/>. GeoMaker can be run locally... it is free software.
Specialized, Restricted Vocabularies
The CCRC initially is using several restricted vocabularies to organize parts of the site.:
- Top-level topics, on which papers have been that focus on climate change as seen by disciplinary domains important to USDA Forest Service management activities.
- A different topic structure is used to organize the CCRC annotated bibliography.
- CCRC’s collection of links and documents is organized by its own functional headings.
- Key words and phrases from presentations made at the two-day workshop entitled 'Adapting to Climate Change in National Forests' on April 19-21, 2010 at the Skamania Lodge in Stevenson, WA.
If we think of the CCRC taxonomy as one big list/glossary, this list can be rearranged in different ways (topic papers, bibliography, links). It can be collapsed to show only high-level concepts, or expanded to show specialized conceptual hierarchies.
CCRC taxonomy terms may (or may not, but probably do) exist in other taxonomies. That’s a good thing: it links CCRC into other, domain-specific taxonomies. We can also take advantage of taxonomies like subject headings and place name lists used in automatically tagging pages and documents with OpenCalais.
We may make selected taxonomies easily available for tagging items in the CCRC. We could do this by providing suggestions as visitors start typing in the search box, or in forms for adding metadata as we add new pages or documents.
Several formal ontologies (conceptual taxonomies) already have been developed for climate change, but key among these in the United States are the Semantic Web for Earth and Environmental Terminology (SWEET) ontologies developed at the NASA Jet Propulsion Laboratory at CalTech, intended as a common semantic framework for various Earth science initiatives. The SWEET 2.0 ontologies cover astronomy, atmosphere, biology, chemistry, climate, earth science, geography, geology, human activity, hydrology, math, oceanography, physics, science, space, and time. SWEET includes more than 6,000 concepts in more than 200 individual taxonomies.
EcoResearch uses the SWEET ontologies for its "Media Watch on Climate Change" service <http://www.ecoresearch.net/climate/#3>. This applies the terms found in the ontology with full-text indexing of news items posted online by Fortune1000 companies, environmental NGOs, and blogs. It uses the terms in the ontology to visually map the news stories as they are conceptually related to one another.
Formal geospatial ontologies are well established, and even required as a standard for federal projects: the Content Standard for Digital Geospatial Metadata (CSDGM) <http://www.fgdc.gov/standards/projects/FGDC-standards-projects/metadata/base-metadata/index_html>. Of course, adhering to the full CSDGM, which provides for more than 300 metadata elements, could be a full-time job. Instead, CCRC needs to identify a subset of elements that adhere to the standard.
A taxonomy for CCRC could incorporate all or part of SWEET and CSDGM, and add specialized vocabularies to extend them.
Open Vocabularies and Annotation
People have their own unique ways of seeing and describing the world, often influenced by culture. By allowing users who benefit from CCRC resources to annotate and tag those resources in their own words that are not necessarily included in restricted vocabularies or standard ontologies, we provide new opportunities to span boundaries and connect climate change science and management in ways we may not otherwise consider. Collections of keyword tags generated in this way are called “folksonomies,” and can be made available alongside other taxonomies.
How Will CCRC Use Taxonomies?
A CCRC taxonomy could be used to populate auto-suggestion lists in a search form. As each letter is typed into the form, a request goes to the server, which suggests matching words or phrases pulled from the CCRC taxonomy.
Because a taxonomy associates terms with one another, based on their relationships, it can be used to generate visual images of how concepts relate to one another, in concept maps or tag clouds, which in turn can be used for content navigation.
The CCRC top-level topics selected as featured perspectives on climate change offer different lenses through which to view climate change. Together, they describe various perspectives on climate change that are of the greatest relevance to the Forest Service mission. Sub-topics of these top-level topics provide richer, focused detail. Think of them as sidebars to a bigger picture. Sub-topics may appear under multiple topics, or even under other sub-topics.
All the topic and sub-topic categories are part of the CCRC taxonomy. The CCRC taxonomy also includes subject-area headings for bibliographic references and links. These headings are organizing categories, arranged to make references and links browseable.
The process of maintaining taxonomies is necessarily one of continual improvement, each revision adding to knowledge organization and the improvement of search results. A collaborative editing space for public taxonomies is freely provided by the Florida Institute for Human and Machine Cognition.
The CCRC taxonomy will be a glossary against which every resource in the CCRC collection will be indexed. The full text of each resource, and the text of the resource's metadata (the data describing the resource), is compared to the CCRC taxonomy, and any occurences of terms from the taxonomy in the resources is recorded as a location index entry. Think of this like the index at the back of a reference book: it records the exact location of every place the taxonomic terms appear in the resources. The software used to provide this search indexing depends on the content management system used for the CCRC. The CCRC is currently experimenting with using the Drupal content management system for this purpose.
The index is used to suggest search terms, using an "autosuggest" script that presents words and phrases that include the letters the user has typed in a search box. Software like TagDragon or gsuggest produce search widgets such as recommender search and visual taxonomic trees, to assist in finding CCRC resources.
[ii] Susan Leigh Star and James R. Griesemer. Institutional Ecology, 'Translations' and Boundary Objects: Amateurs and Professionals in Berkeley's Museum of Vertebrate Zoology, 1907-39. Social Studies of Science 1989; 19; 387.