Expertise for the nonspecialist: delivering forestrelated information to nonforesters

1
Expertise for the non-specialist delivering
forest-related information to non-foresters

• Chair and organiser
• Roger Mills, Oxford University Library Services
• Co-ordinator, IUFRO Research Group 6.03.00
  Information Services and Knowledge Organization

2
Forest Research

• Projects over many decades have produced a wealth
  of data, published and unpublished
• Now finding uses in other disciplines
• environmental management
• climate change assessment
• biodiversity conservation
• economic planning
• economics
• politics
• social science
• law
• Easy to access with modern technologies
• data frequently needs processing or harmonisation
  to make it usable
• Raises many issues of intervention, explanation
  and training which fall partly or wholly on the
  library and information sector

3
Todays workshop

• Highlight some of the issues
• Present case studies
• Discuss what we can do to ensure that users
  unfamiliar with the forestry subject area can
  make best use of available data
• Make a wish list for future action in IUFRO,
  IAALD, other fora

4
Trees grow slowly

• Not like cabbages generations needed for
  controlled study
• No equivalent to Rothamsted experiments started
  in 1843 and still going
• Majority of forest studies carried out for a
  particular end and data collection not primary
  purpose

5
Data gathering

• Traditionally
• Field trials
• Gather data
• Analyse on paper
• Publish conclusions
• Data stays in a drawer

6
Early computing

• Data on tapes, punched cards etc
• Physically managed by central computing units
• Data preserved though may not be fully catalogued
  or readable long term

7
Modern computing

• Gathered on portable devices
• Analysed on PC
• Stored on removable media
• No central responsibility, existence known only
  to researcher
• Unknown, unreachable, unreadable
• So data is recompiled

8
Forest data

• Time dependent, not repeatable
• Time series important significant variations may
  occur over relatively short periods
• Essential to preserve all historical data we can

9
Impact of web

• Preserving data in a mediated library allows
  delivery with health warnings
• Make it web-accessible leaves open to
  misinterpretation
• But harmonised data useful in many non-forestry
  contexts
• Problem lies in the harmonisation

10
DBH

• Diameter at Breast Height
• How high is your breast?
• 1.3m (43) (USA etc)
• 1.4m (46) (UK etc)
• 1.5m (for ornamental trees).
• Decimal conversions also introduce variations
  46 is more accurately 1.37m.

11
A little knowledge is a dangerous thing

• Adding stats for DBH from different areas without
  conversion will be misleading
• Can lead to bad decision making
• Eg in climatology, basing estimates of carbon
  incorporation on forest volume

12
Whats that got to do with librarians?

• Aim to make data readily available to all who can
  use it, without restriction or censorship
• Internet helps, but aids unintentional or
  intentional misuse
• Answer better metadata and user education

13
GFIS

• Data harmonization originally an aim of Global
  Forest Information Service
• Not achieved because of manpower required to
  generate extra metadata defining conversion
  requirements, or just warning of
  incompatibilities
• Most data not compiled for international use, no
  funding to provide metadata at source

14
EU to the rescue

• 1989 regulations to set up European forest and
  Communication System
• well-structured and relaiable forest information
  at European level
• NEFIS Network for a European Forest Information
  Service 2003-5
• http//www.efi.int/portal/project/nefis

15
Into operation

• European Forest Information and Communication
  Platform (EFICP)
• http//eficp-info.jrc.it/
• Long gestation common
• Political requirement
• Development of prototype
• Study problems
• Development of production system
• Now 19 years since original Regulation

16
Use it or lose it

• Communicate existence of system
• Make it easy to use and reliable
• Must save users time
• NEFIS project illuminates problems
• Many relate to librarians traditionbal expertise
• Terminology
• Classification
• Quality assessment
• Searchability
• Interoperability
• High-quality metadata

17
Iterative development

• Distribute technology favours new uses/users for
  existing data
• Infrastructure needs
• Advanced spatio-temporal data collection and
  information management
• Dissemination and fusion of heterogeneous
  distributed information
• Sophisticated analysis, modeling and
  visualization of information
• Designed to outlive current software

18
Cf Bioinformatics

• Single information system holds
• Sequencing data
• Tools for annotation
• Tools for analysis
• Publications resulting from analysis
• E.g. NCBI http//www.ncbi.nlm.nih.gov/

19
(No Transcript)
20
An integrated system for forestry?

• Much wider variety of data types
• Much wider community of users
• And of technical infrastructure
• NCBI model bridges data acquisition, analysis and
  curation
• Publishing models increasingly incorporate raw
  data source with peer-reviewed research

21
Publishing data

• Author complies dataset containing forest cover
  statistics spanning multiple jurisdictions and
  century-long time series
• Data acquisition and harmonisation methods
  recorded in metadata
• Publishes package so data remains available
  long-term for use or further analysis by others,
  retrievable alongside journal articels

22
Open Access

• Non-subscription environment to ensure wide
  availability
• Requires new approach to resaerch funding
• And long-term funding for data curation
• That role likely to fall on library community
• Business and technical expertise in archiving
• Developing and supporting integration and
  interoperability tools
• Online repositories

23
Developing standards

• NEFIS datasets too different to achieve
  interoperability
• Demonstrated need
• EU European Interoperability Framework 2004
• Technical
• Semantic precise meaning
• Organizational
• Last two most challenging

24
Semantic interoperability

• Descriptive metadata
• Controlled vocabularies
• Ontologies
• User-nominated terms requires editor
• Tagging
• Quality
• Accuracy
• Logical consistency
• Completeness
• Positional accuracy
• Lineage
• Non-censorious indication quality report

25
Data location

• Providers server
• Or central?
• If local, owner responsible for metadata
  management
• Interoperability requires metadata on
• Protocols for query translation
• Mapping of filed labels
• Field contents
• Backround information
• Associated files
• Realed IPR
• Required executables
• Language and character set
• Access control mechanisms
• Standards to be agreed so all new compilations
  and reloaded legacy data have this information

26
NEFIS Demonstrator

• No data harmonization
• Showed feasability of retrieving and analysing
  data for a single request to multiple servers in
  multiple countries
• Comprises
• Resource discovery toolkit searches metadata
• Remote search demonstrator managing data
  retrieval form multiple sources
• Visualisation toolkit (VTK) naïve and expert
  modelling of retrieved data

27
EDA

• Exploratory Data Analysis
• Unbiased examination of data to detect patterns,
  trends, relationships rather than answer
  preconceived question
• Mirrors bioinformatics approach
• NEFIS data specially prepared
• Adoption of common standards could allow
  development of VTK with no need for human
  intervention in preparing data

28
Librarians are key

• In
• Curating data
• Developing and supporting implementation of
  standards
• Ensuring ready access to data
• Promoting use
• Universal Data Control UDC
• Its classification, Captain, but not as we know
  it or maybe it is!
• So lets do it.