Geoscience Reference
In-Depth Information
and commissioned industrial work dating from the 1960s, when the TFRI's annual
reports fi rst included extensive bibliographies.
In total, over 30,000 research outputs and over 30,000 meta-data 2 outputs were
collected from over 30 of the approximately 40 ILTER networks, spanning approxi-
mately 40 years of research. The networks from which no bibliographic information
was obtained consist of networks that are inactive (e.g., Canada and some networks
in Eastern Europe and Africa) or recently established (e.g., Philippines and
Malaysia). While some regional and national networks actively maintained compre-
hensive bibliographies of their own research outputs, others maintained bibliogra-
phies at the sub-regional or site levels.
The set of ILTER research sites is not identical to the combined sets of research
sites under each of the regional or national networks. Furthermore, non-ILTER
research is conducted at many ILTER research sites. Consequently, the 30,000
research outputs collected include research outputs produced at research sites and
by individuals affi liated with national LTER networks, but which may not be for-
mally part of the ILTER Network. Inclusion of such research outputs from outside
the formal core of the ILTER network is consistent with the inclusion of networks
that have in their bibliographies work initiated or published before the formal estab-
lishment of the ILTER network in 1994.
All available abstracts from meta-data outputs and over 5,400 abstracts from
other research outputs also were collected into the same database, containing among
others the following columns:
()
(()
Author s
,
Title Year Publication Name Keywords LTER Network s
,
,
,
,
, Abstract
In cases of multiple authorship, a single publication may appear in the bibliogra-
phy of more than one LTER network. However, each publication only is counted
once. Texts of the research output and meta-data titles and abstracts were automati-
cally deconstructed into one-to-three word long alphabetized N-grams (Cavnar and
Trenkle 1994 ) of Porter Stemmed ( 1980 ) words, excluding stop word such as “of”,
“is”, and “the”. 3 Plausible place-names were initially identifi ed as those containing
2 Meta-data are searchable data about data. In LTER, a meta-data record about a data set might
include time and location of data collection, methods used, species and geographies involved, etc.
Many LTER networks (also) publish their data and meta-data in a Global Biodiversity Information
Facility repository or other repositories.
3 For example, the title “Geographical and Thematic Analysis of Publications Generated at ILTER
Sites” would be processed fi rst into: “geograph*”, “themat*”, “analysi*”, “public*”, “gener*”,
“ILTER*”, “site*”, “geograph* themat*”, “analysi* themat*”, “analysi* public*”, “gener*, pub-
lic*”, “analysi* geograph* themat*”, “analysi* themat* public*”, “analysi* gener* public*”,
“gener* ILTER* site*”. Each N-gram was considered to be a plausible concept discussed in the
research outputs. Other concepts included “disturb*” (capturing “disturbance”, “disturbed”, etc.),
“chang* environment*” (capturing “changing environments”, “environmental change”, etc.), and
“chang* impact*” (capturing “impact of change”, “changes impact”, etc.).
Search WWH ::




Custom Search