Data conversion from ADD to Global Map

1
Data conversion from ADD to Global Map

• FUJIMURA Hidenori (ISCGM / GSI)
• Peter PULSIFER (Carleton University)

2
Contents

• A short description on Global Map (GM)
• Problems in data conversion from Antarctic
  Digital Database (ADD) to Global Map
• Inverse-projection from ADDs polar stereographic
  projection
• Data model matching / mapping between ADD and GM

3
Global Mapping project

• National mapping organization of each country
  develops Global Map contribution (GM)
• conforms with standardized Global Map (GM)
  specifications
• Intended for environmental analysis and global
  change monitoring
• 129 countries or regions participating
• Global map for 12 countries have been released
  see http//www.iscgm.org/

GM sample
4
Progress of GM project

• Many organization are now developing GM.

5
ADD data to GM data

• Global Map for Antarctica will be derived from
  Antarctic Digital Database (ADD)
• ADD project provided GM project with the released
  vector data and DEM data.

6
Conversion software problem

• ADD data is projected on the Polar Stereographic
  projection.
• Global Map uses geographic coordinate.
• Requires inverse-projection
• Could not find inverse projection function in
  commercial GIS software that worked effectively
  with raster data
• Tested ArcView3, Arc/INFO7, ArcGIS8 and ERDAS

7
Use of Open Source Software

• Solution used was PROJ.4, an open source
  projection tool
• Open source software can be externally validated
• The program (Ruby script) will be developed
  within testing/validation framework

8
Elevation data conversion

• We have to get elevation data for 155520000
  points, which cannot easily be stored in RAM at
  once.
• We use XML file as storage stream data to script

9
(Plan of) Implementation

• Ruby script for each step will be tested.
• Use of (XML) file will facilitate testing of
  intermediate data

10
Mapping of the data model

• Which data in ADD to which data in GM?
• We have to analyze each data model at first.
• Draw UML-like class diagrams for the data model.
• After data model analysis, in principle
• We have to conserve the consistency of the Global
  Map specifications.
• We want to map as much information as possible.

11
Global Map point data
12
Global Map line data
13
Global Map polygon data
14
Global Map package
Transportation
Hydrography
Boundary
Population
15
ADD point data
16
ADD line data
17
ADD polygon data
18
ADD pack-age
19
Attribute mapping

• By analyzing the diagrams and specifications, we
  found some difficulty for mapping the data (which
  we call semantic heterogeneity)
• We have to confirm the definition of
  nationality (ADD) and countryCode (GM) in
  detail
• source (ADD) and sourceCountry are nearly the
  same concept, but hard to map. Because source
  contains not only countries.

20
Attribute mapping cont.

• Antarctica has many kinds of coastlines
  identified
• A number of conversion options
• Map all the coastlines without information on
  kind?
• Map only one kind of coastline?
• Select most relevant coastline classes and
  aggregate to a composite coastline for GM
• etc.

21
Conclusion

• Inverse projection
• We are developing the inverse projection scripts
• Considering testability to ensure reliability
• Data model matching
• We are now analyzing the two data models
• Some semantic heterogeity exists
• More vector to raster conversion (land cover,
  land use and vegetation.)
• Global Map Antarctica will be derived from this
  work.