EarthChem Solid Earth Geochemistry in Geoinformatics

1
EarthChemSolid Earth Geochemistry in
Geoinformatics

• www.earthchem.org

2
Why Do We Need Data Management in Solid Earth
Geochemistry?

• Geochemical data are essential for answering
  fundamental questions about the composition,
  structure, evolution of the Earth, its oceans,
  continents, and climate
• Problems
• Data is dispersed in literature, often not in
  electronic form
• Compilations by investigators are time-consuming,
  redundant, often incomplete
• Missing links among related data
• Data is lost due to incomplete publication

3
Data Management in Solid Earth Geochemistry
SedDB
4
PetDB, NAVDAT,GEOROC

• Offer the only generally accessible compilations
  of large volumes of data on the compositional
  variation of igneous rocks.
• Provide desktop access to the entire published
  geochemical literature within minutes,
• allowing researchers to address questions that
  otherwise would be dropped due to the large
  effort required to find and compile the data.
• allowing students to explore the global dataset
  within a formerly unimaginable timeframe that can
  be accommodated in the course schedule.

5
PetDB, NAVDAT, GEOROC

• Compile and serve ALL raw geochemical data
• Share common relational data model (Lehnert et
  al. 2000)
• Data fully integrated
• Wide range of sample analytical metadata
• Generally applicable for sample-based
  petrological and chemical data for rocks
• Each value linked to original publication or
  producer

6
Interactive, Dynamic Web Interfaces

• Select, filter, view, download customized data
  sets
• Explore metadata

7
Other Features (database-specific)
Visualization tools (NAVDAT)
Interactive map interfaces (NAVDAT)
Disparate data for individual samples linked via
unique sample IDs (PetDB)
Interoperability (PetDB)
8
Data Quality Control
Comprehensive analytical metadata

• allow proper data quality assessment

Example PetDB interface

• can be used as data quality filters

9
Content of PetDB, NAVDAT, GEOROC

• gt 4 Million individual chemical values
• for gt ca. 230,000
• igneous rock samples
• from gt 6,300 publications

10
Benefits of Rigorous Scientific Data Management

• Maximized Utility of the Geochemical Dataset
• Enhanced Data Quality Control
• Data Integration Visualization across the
  Geosciences
• Impact on Science Education

11
Maximize Utility of the Geochemical Dataset
More than just a timesaver, these databases make
it possible to address both global and regional
questions that I would otherwise never bother to
attempt. The amount of time saved is such that
countless ideas cross from the realm of the
totally impractical for a busy working scientist
into the realm of easy to squeeze into a spare
half hour. Simply put, I can now test theoretical
ideas against all the world's data, and can
readily compare any specific region I am working
on to its global counterparts. This is a
monumental benefit. Paul Asimov, California
Institute of Technology EarthChem User Survey
January 2005
12
Scientific Return
gt120 papers that cite PetDB GEOROC

• Plank, T. Constraints from Thorium/Lanthanum on
  Sediment Recycling at Subduction Zones and the
  Evolution of the Continents, Journal of Petrology
  46, 921-944, 2005.
• Ballentine, C.J. et al. Neon isotopes constrain
  convection and volatile origin in the Earth's
  mantle, Nature, 433, 33 38, 2005
• V. Salters A. Stracke Composition of the
  depleted mantle. G3, 2004
• Cipriani, A. et al. Oceanic crust generated by
  elusive parents Sr and Nd isotopes in
  basalt-peridotite pairs from the Mid-Atlantic
  Ridge. Geology, 32 (8), 657660, 2004.
• Herzberg, C. Geodynamic Information in
  Peridotite Petrology, Journal of Petrology, 45,
  2507-2530, 2004
• M. Hirschmann et al. Alkalic magmas generated by
  partial melting of garnet pyroxenite. Geology 31,
  2003
• Kellogg, J. B., Jacobsen, S. B., OConnell, R.
  J. Modeling the distribution of isotopic ratios
  in geochemical reservoirs, Earth Planet. Sci.
  Letters 217, 2004.

13
Application to Education
14
Challenges for Database Providers

• Optimize interaction with the data for a broad
  audience ranging from the casual to the expert
  user
• Efficiently populate databases with legacy and
  new data
• Integrate data with the larger Earth Science
  dataset
• Ensure longevity of data systems

15
The Problem of Distributed Datasets
A typical science question What is the
relationship between what is being subducted at
the Aleutian trench and what is being erupted in
Aleutian volcanoes?

• Need Nd, Sr, Pb, Hf isotope ratios, and
  incompatible trace element compositions

Aleutian Volcanics
North Pacific (Juan de Fuca Ridge) MORB
16
The EarthChem Consortia
Founded in 2003 by R. Carlson, A. Hofmann, K.
Lehnert D. Walker

• Build an integrated data management and
  information system for solid earth geochemistry,
• based on and expanding the collaboration of
  PetDB, GEOROC, and NAVDAT.

• Nurture synergies among projects
• Minimize duplication of efforts
• Share tools and approaches

17
EarthChem Activities

• Community Workshop (October 2003, Carnegie
  Institution Washington)
• Reviewed the current status of data management
  efforts in Solid Earth Geochemistry
• Discussed ways in which these activities can grow
  and collaborate to best participate in and
  contribute to the Cyber Infrastructure revolution
  in the Geosciences

• Exhibits demos at AGU 2003 2004 and GSA 2004
• Presentations at GSA2003, AGU2004, various
  workshops
• Session on Geoinformatics for Geochemistry at
  AGU 2004, co-chaired with GERM
• Web site at www.earthchem.org

18
EarthChem Priorities

• Build the EarthChem portal as a central access
  point to a system of federated geochemistry
  databases (One-Stop Shop for Geochemical Data)
• Ensure efficient and continuing update and
  expansion of data holdings

Proposal submitted to NSF (EAR IF) January
2005 K. Lehnert, D. Walker
19
One-Stop-Shop for Geochemical Data
Users
Geoscience CI
Interoperability
EARTHCHEM PORTAL Uniform data submission Search
capability across federated databases Standardized
integrated data output Generally applicable
tools for DQ assessment data analysis/visualizat
ion
SedDB
and more..
20
Building the One-Stop Shop

• Interface federated databases
• Implement web services SOAP/XML/WSDL, OAI, OGC
• Standardize metadata (ISO19115, OGC-GML)
• Systematize nomenclature vocabulary
  (ontologies)
• Register database schemas with GEON?
• Implement unique sample identification through
  use of the International Geo Sample Number
• Build user interfaces with flexible data
  selection and extraction, tiered for different
  levels of expertise
• Use customized GEON Portal technology?
• Use EarthChem map viewer, GeoMapApp browser, or
  other tools to integrate with other data types
  such as seismic tomography, gravity, structural
  features, etc.
• Provide tools for data evaluation such as
• interactive discriminant plots, P/T calculators,
  data quality filters

21
The Bottleneck Data Entry

• Difficult to find knowledgeable data managers
• Missing metadata (e.g. locations, analytical
  info)
• No unique sample identification
• Missing standards for data presentation (e.g.
  units)
• Unavailable data files
• Errors in original data tables
• Missing cooperation from authors

EXPENSIVE!
22
Efficient Update Expansion of Data Holdings

• Encourage direct data contributions from the
  community
• Build on-line data submission capability for
  future data (compliance with data policies for
  science programs!)
• Provide services for on-line storage of routine
  data about analytical procedures (MyEarthChem)
• Facilitate incorporation of existing large data
  compilations
• Provide technical assistance to investigators who
  want to compile new datasets

23
Facilitate Community Contributions

• Assist contributors with design, implementation,
  population of databases.
• Serve databases via the EarthChem portal.
• Contributed datasets will retain their identity
  within the EarthChem system.
• PILOT PROJECT
• A relational database of the Mexican Volcanic
  Belt
• Straub, Ferrari, Langmuir

24
Expansion of Data Holdings

• Generate additional datasets
• Identify and prioritize new target datasets
  through community outreach and the EarthChem
  Advisory Committee
• Data entry by dedicated EarthChem personnel

25
Integration with Science GeoInformatics
26
A Users Vision
in theory the best thing would be one big
Geo-database where all different types of
geochemical reservoirs are included and all
analytical tools as well and where you can search
for either regions or reservoir type or method...
ok thats a big goal.