Data Analysis

1
Data Analysis Visualization for the Census of
Marine Life (CoML)
Second SCOR Meeting on Coordination of
International Marine Research Projects December
8th, 2006
Patrick N. Halpin   Gabel Associate Professor of
the Practice of Marine Geospatial Ecology,
Marine Geospatial Ecology Lab Nicholas School of
the Environment and Earth Sciences Duke University
2
Data Analysis Visualization for the Census of
Marine Life (CoML)

• Overview data types
• Information dissemination interoperability
• Data visualization
• Workflow visualization example workflow

CoML 2010 Framework Committee charge to hold a
marine data visualization technical workshop
during 2007 (the Marine Geospatial Ecology Lab,
Duke University the Sea Mammal Research Unit,
St. Andrews)
3
Data is not information
Only some information is useful for decision
making
We convert data to information to decision
support through analytical workflows
A data warehouse is not a library
If we want people to use our data we need to
convert data warehouses into libraries of useful
information and published workflow processes
4
TOPP, POST, ArcOD, CHEss, MAR-ECO, GoMA
FMAP
HMAP
History of marine animal populations
Future of marine animal populations
CoML Field Projects
OBIS
Ocean Biogeographic Information System
5
regional nodes
specialized nodes
central data portal
CoML Field Projects
6
Marine Animal Data
Global Information Systems?
Ocean Biogeographic Information System
Hierarchical Information networks.
trophic information networks
Portals
fisheries / mgmt
specialist portals?
top predators
prey
producers
Nodes
oceanography
Regional nodes?
7
Marine Data Types
Biogeographic data X,Y location, species,
ancillary information
8
Common Marine Data Types
Purpose to create common data models for marine
GIS end users
Coming in 2007
9
(No Transcript)
10
Instantaneous point observation effort data
Observation (points) Effort (lines)
Cruise transects
Observations
11
Fixed location time-series data
FWS Beringian Seabird Colony Catalog 2004 (The
chart is sent from OBIS-SEAMAP)
Species code from the OBIS-SEAMAP web site
12
2D time telemetry tracking
(Location Series)
Sea Turtle Tracks (Caretta caretta)
Sea Surface Temperature (WCR)
Source http//www.po.gso.uri.edu/SST/
Source http//obis.env.duke.edu/datasets/ (Read
McClellan2004)
13
3D time telemetry tracking
(Location Series)
Turtles Cayman Islands
Dive Profiles 4D Data (X,Y,Z,T mm)
14
3D time environmental sensors telemetry
tracking
(Location Series)
Biologging data
15
Photo-ID tracking of individual animals
16
3D passive acoustic location / call data
Hydrophone array
3D Sperm whale locations
17
3D animal behavior models
DTag animal
GeoZui4D Ribbon-plots
Humpback whale bubble-net feeding
18
3D animal behavior models
Workflow
Humpback
Minke
Friedlaender et al 2007 (in press)
19
Marine animal habitat models
20
Animal behavior models
Bayesian state-space models of movement, speed
and environmental dynamics
21
Data Analysis Visualization for the Census of
Marine Life (CoML)

• Overview data types
• Information dissemination interoperability
• Data visualization
• Workflow visualization example workflow

22
OBIS-SEAMAP information system
The world data center for marine mammal, seabird
and seaturtle information.
23
(No Transcript)
24
Species information

• Species Profiles
• Natural history information
• Taxonomic classification
• Bibliography / Web Links
• Links to prey profiles

25
Internet map service standards
Emerging open-source, open-standards tools
provide a functional, open working environment
for ocean biogeographic information community.
physical oceanography
marine spatial ecology
26
Data Discovery Metadata
27
Data Analysis Visualization for the Census of
Marine Life (CoML)

• Overview data types
• Information dissemination interoperability
• Data visualization
• Workflow visualization example workflow

28
Data viewing (1)

• Dataset locations on Google Map
• Quick browsing
• Easy navigation
• Direct access to dataset page

29
Data viewing (2)

• Data on Google Earth
• Any types of data
• Effort/telemetry lines
• Informative popup
• Link to OBIS-SEAMAP

30
Data viewing (3)

• Integration with OBIS-SEAMAP application
• Turtle nesting trends
• Seabirds colonies abundance time series

Example of turtle nesting trend
Example of seabirds colony abundance
31
Data viewing (4)

• Overlay environmental data
• Monthly sea surface temperature over the period
  of survey

32
Data viewing (5)

• Time series animation
• Telemetry data
• Sea surface temperature
• Create your own animation

Example of animation setting window
33
Real-time telemetry tracking tools
34
Data Analysis Visualization for the Census of
Marine Life (CoML)

• Overview data types
• Information dissemination interoperability
• Data visualization
• Workflow visualization example workflow

35
Marine mammal habitat modeling
Data
Information
36
Motivation To translate data on marine mammal
distributions into useful information for ocean
scientists managers
Practical tools for marine mammal avoidance
Reducing shipping lane interactions
Reducing sonar interactions
NOAA Right Whale Surveys
37
Two linked projects
38
Available Data Sets
Survey Effort
45 SEAMAP data sets 1991 2002
Model individual species guilds of
species
Data Providers NOAA/NMFS -- NEFSC NOAA/NMFS
SEFSC UNC-W
39
Environmental variables
Distance to Shore
Distance to Shelf
SST
Depth
Distance to Fronts
40
Environmental sampling through time
automated processing scripts
T5
T4
Time
SST Chl Depth Shelf Fronts
environmental sampling
T3
Absences
T2
Example 11 years _at_ 8 day time step
T1
41
Habitat Modeling Workflow System
42
ROC evaluation
Receiver-Operator Characteristics curves
79 of the observations are correctly predicted
43
Area of Elevated Beaked Whale Encounter Risk
(Summer)
44
GIS Web Services for Environmental Planning

• Observations Effort
• By Dataset
• By Species
• Range Maps
• Predictive Models
• Expert-Reviewed

Selected area of interest for model query
45
Feedback to data providers
Spatio-temporal data Gap-Analysis
46
(No Transcript)
47
Data is not information
Only some information is useful for decision
making
We convert data to information to decision
support through analytical workflows
A data warehouse is not a library
If we want people to use our data we need to
convert data warehouse into libraries of useful
information and published workflow processes
48
Questions?