NADM , Links, GeoMatter and friends

NADM , Links, GeoMatter and friends

An overview of NADM Eric Boisvert & Vincent Desjardins

Centre géoscientifique de Québec

Schedule

What is NADM ?

What is <geo>Link (Cordlink, Hydrolink, GASLLink,etc...)

What is GeoMatter ?

How does it work ?

A down to earth example

NADM North American Data Model

Result of an international effort (USA and Canada, Mexico could eventually jump in)Lead by a « Steering Commitee » composed of USGS, AASG, GSC and OGS.SC created a set of 6 Technical Teams addressing specific issues of the Data Model

Requirements Data Model Design Documentation Data Interchange Science Language Tool Development

NADM North American Data Model

GSC has 2 seats, 1 occupied by Boyan Brodaric and 1 rotating seat.Provincial also have 2 seats, one occupied by Brian Berdusco (OGS) and another to be filled (Eric Grunsky, Alb. ?)Remaining seats occupied by USGS and AASG (American Association of State Geologists)

NADM North American Data Model

They produced several documents but the most important for us so far is the Version 4.3 of the database structure.http://geology.usgs.gov/dm/model/Model43a.pdf

This is the only true NADM data structureA NADM inspired version has been developed for the Cordlink ResSources project. It is called V 5.2 or simply the  « Cordlink version »http://cordlink.gsc.nrcan.gc.ca/cordlink1/info_pages/English/dm52.pdf

NADM 4.3 versus 5.2 at a glance

5.2 is a generalization of 4.35.2 is enhanced to serve as a Virtual Library core (handles images and texts)5.2 has « shared attributes »5.2 embeds higher level of information description and relationships5.2 embeds tool specific metadata (system tables)

NADM logic (5.2)

Source

Legend schemeLegend

items

Spatial objects

references

Compound

Object

Archive

Singular

Object

Archive

Attributes

NADM logic (5.2)

Source: Map related information- Author, Edition, Publication, etc.- Scale, projection, resolution

Legend : Scheme and Classification-Labels, symbolisation, Legend description

Spatial References: Geographical objects- References to points, lines, polygons, volumes, etc..

COA, SOA, Attributes: actual data-Rock description, structural measurements, sample descriptions, ages, stratigraphic positions

Glossary SourceScheme

Glossary Classification ObjectSymbol

Glossary Spatial objects

Glossary

striae

Classification

Scheme

SOA:

Type: striae

Azimuth: 210

Direction: Unknown

Chronology: 1

•Glacial flow indicator

•Striae

Faint marks on rock

Except on volcanic outcrops

Why is it so interesting ?

Data deep : it incorporates both factual and interpretative information

v 5.2 integrates images and texts

Incorporates « conceptual » integration through COA hierarchy

Say what ?

Incorporates « conceptual » integration through COA hierarchy

Universe

Atmosphere

Lithosphere

Hydrosphere

Biosphere

Continental

Oceanic

Plateform

Sea floor

COA

Most difficult concept to understand

It’s a hierarchical list of « things » that will appear on a map

There is a fuzzy frontier between COA and legend

COA

Saint Lawrence LowlandPotsdam Group

• Cairnside Formation

• …

Chazy Group

IgneousPlutonic

• Mafic Gabbro …

• Intermediate …

• Felsic Syenite Granite …

Volcanic• Mafic

ContinentalAlluvial

• Alluvial terrasses

• …

Glacial

Lithostratigraphic

Morphogenetic

Lithologic

COA

NADM allow multiple types of COA arborescences to coexists

UniverseRock Unit

• …

MorphoGenetic• ...

Rock Unit

specific attribute

table

Morphogenetic

specific attribute

table

COA

COA

Emphasize the fact that a map is always create with a conceptual frameworkGeologists are not interoperable between themselves (specially if you include time factor) (D. Lebel, March 21, 2000)

So this is not a limitation of NADM itself

Geologists were

not using the same COA

tree

Interoperability & integration

A B

A1 A2 B1

Corporate wide concepts

Local

Why do I annoy you with this COA tree stuff ?

Example of different concept levels

Corporate levelTholeiite

Local LevelHigh Mg Tholeiite with large amount of feldspar phenocrists.

Someone working in volcanic petrology could

have 24 differents classes of Tholeiites, but unless the client

is looking something very specific, chances are he would stop

at the tholeiite level

Interoperability

A B

A1 A2 B1

A B

A1 B2 B1

Ba Bb BcSite 1 Site 2

Do you have any A ?

A1 -> A

A2 -> AA1 -> A

Benefit of the COA approach

Can create a global integration of geosciences « things »

Corporate don ’t have to update the tree everytimes to accomodate new concepts

More specific concepts can be created at local level without affecting the rest

Helps search and retrieval of information in multiple (distributed) databases

NADM Important consideration regarding GIS

NADM data

structure

DatasetID

SpatialObjId

GIS

NADM only handles references to

Spatial objects

DatasetId = Coverage

SpatialObjId = item

NADM Where does it fit ?

NADM structured

System

Cordlink

GeoServ

Spatial Objects pool

NADM Where does it fit ?

NADM structured

System

Spatial Objects pool

ArcView

Shape files, ArcInfo Coverages

DatasetIdSpatialObjId

Cordlink, HydroLink and others

Application that taps into NADM structure to deliver information

5.2 version has been developed to respond to Cordlink needs

Cordlink is built around the « Virtual Library » paradigm

Hydrolink is a « port » of Cordlink to address hydrogeological purposes

Cordlink and Hydrolink

Cordlink, HydroLink and others

NADM structured

System

Spatial Objects pool

SDL and WFM

Browser

MapGuide

Server

ColdFusion

Server

MapGuide

Viewer

.cfm files

HTML

request

DatasetIdSpatialObjId

DatasetIdSpatialObjId

What’s next ?

NADM is getting more and more attention as a data integration approach

Is this truly perfect ???

Nope, there are a couple of flawsSymbolisation has a couple of conceptual problems, for instance, some symbolisation is SOA basedFormally, a single spatial object can be related to more than one COA in the same map (leading to another symbolisation problem)

I personally think symbolisation should have been left out of the model

SOA symbolisation

Gabbro X

COA

Therefore, these blobs are related to this COA by the mean

Of this legend item and given this symbol

Specific attributes of blobs

Are not required

SOA symbolisation

Schistosity

COA

1- Multiple symbol for 1 point

2- Rotation from SOA

3- Dip label also from SOA

45

30

44

SOA

SOA

SOA

Multiple COA

Gabbro X

COA

COAAquitar

WARNING – Personal opinion

Keep the database at the data content level

Implement external symbolisation processes (such as Coleman-Sadd’s GeoLegend) to create map out of database content

This is an open debate

What is Geomatter ?

We have NADM in the core, <geo>link to diffusion information, now how do we populate it ?

GeoMatter is essentially a tool to populate the database

NADM version 5.2 - scary version

Friendly version

GeoMatter philosophy

COA Generalization

Hide data structure complexity

What ’s missing ?

Geomatter

Curly

« scripts »

Entry tools

Cordlink and other

<geo>links

Cubewerks stuff

Diffusion tools

Mostly none

Geomatter (?)

Management tools

None

Import/Export tools

What ’s next ?

servlet

OGDI

Abstraction

layer

NADM

OpenGIS

Java

clientNADM API

OGDI API