Invasive Species Systems Task Group Status Update and Some Performance Analysis

1
Invasive Species Systems Task Group Status Update
and Some Performance Analysis

• Jim Graham
• PhD Candidate, Colorado State University
• Annie Simpson, Michael Brown, Greg Newman,
  Catherine Jarnevich, Tom Stohlgren

2
Invasive Species

• 120 Billion spent annually in the US
• Between 5000 and 50,000 invasive species in the
  US
• Impact agriculture, environment, and human health
• Reduce area for viable grazing
• 46 of species on threatened and endangered list
  are listed because of invasive species
• Human health concerns (e.g. West Nile Virus)

3
Resource Managers Need Help!
4
Invasive Species System
Users
End-Users
Web Pages
Portals
Search Engines
Modelers
Consumers (Clients)
Web Services
Herbariums
Museums
Invasive Databases
Inspection Databases
Providers (Server)
5
Customers

• End-Users
• Researchers
• Citizen scientists
• Decision makers
• Resource managers
• Data Consumers
• Modelers
• Map Generators
• Portals
• Data Providers
• Over 252 invasive species databases
• in the US alone (Crall et al. 2006)

6
Research and Production
7
Software Lifecycle
Investigation
Design
Implementation
Testing
Maintenance
Time
8
ISSTG Initial Schedule

• October, 2006
• Formation of TDWG task group
• December, 2006
• Investigation completed
• Draft design
• February, 2007
• Design completed
• May, 2007
• Test implementation
• September, 2007
• Toolkits available?

9
Current Status

• Investigation
• Initial use case analysis
• Draft requirement specification
• Starting alternatives analysis
• Starting feasibility studies
• Design
• Draft schema
• Terminology definition tools
• Implementation
• Prototypes

10
Use Case Analysis

• Example Cases
• Create map of species
• Research status of species
• Find environmental niche of species
• Research control methods
• Find native distribution of species
• Possible Actions
• Refer to existing solution
• Build web site
• Build portal
• Add to the protocol

11
Some Requirements

• Provider Requirements
• Common language support
• Platform independent
• Broad database support
• Easy to implement, use
• Fast access to large amounts of occurrence data

12
Invasive Species Data

• Existing Standards
• Locations
• Taxa identification
• Dates
• Depth, Elevation, Sex, LifeStage
• Extensions
• BioStatus (non-native, harmful)
• Additional ancillary data (percent cover, date
  introduced)
• Management information
• Life history

13
Additional data
14
Quality
15
Search Times on Resources
Times are in seconds
16
Performance Variables

• Internet connect speed
• Number of hops between client and server
• Server speed
• Web server software
• Protocol complexity (3)
• Provider software (3)
• Database software
• Database structure
• Number of records (1)
• Search type
• Fields requested
• Number of matches (2)

17
Search time vs. number of records
18
Search time by number of hits
19
Protocol and Provider Software
20
Typical Harvest Times?
- Only the Internet sample time includes Internet
latency
- Times are in seconds unless noted otherwise
21
Conclusions

• Protocol fit to draft requirements
• Removing embedded requests improves performance
  and robustness
• Current DiGIR SW has performance problems with
  SQL Server
• Will look at
• other provider software
• Simple request/response
• Schema fit to requirements
• Portions of DarwinCore and ABCD
• Additional extensions

22
Duplicate Standard Nightmare!

• TIFF
• Too complicated to use
• Toolkits were late
• JPEG, GIF, IMG, PNG, EPS, Targa
• Thousands of formats!
• TWAIN
• Simplest solution that met the requirements
• Delivered with cross-platform toolkits
• Only scanner interface standard
• Can we get all the folks working on occurrence
  data together?

23
Acknowledgements

• Funding
• NASA
• USGS, National Biological Information
  Infrastructure
• Tamarix Coalition
• Help
• Catherine Jarnevich
• Annie Simpson
• Michael Browne
• Bob Morris
• Jacob Asiedu
• Dave Vieglas
• James Macklin
• Everyone at the Natural Resource Ecology Lab at
  Colorado State University

24
----------------
25
WMS Performance
26
Harvest DiGIR Records

• Seven invasive species, one genus
• Harvested all available data from 637 registered
  resources (UDDI)
• 100 records maximum per request
• Fields
• DecimalLatitude, DecimalLongitude,
• ScientificName, Kingdom
• EarliestCollectionDate, LatestCollectionDate
• Removed embedded requests

27
Forecasting at Various Scales
Local
Regional
Global
National
28
(No Transcript)
29
Top 10 Times for S. vulgaris
30
Rasters from JPL

• Web Map Service (WMS) Protocol from OpenGIS
• Requested 1000 rasters
• Randomized
• from 100x100 to 1000x1000 pixels
• Within the continental US
• Time from request to entire image received
• Digital elevation rasters

31
www.NIISS.org
32
Invasive Species Cyberinfrastructure
33
3D Spatial Prototype
Hackberry Canyon Watershed, Utah
Relative Biomass (g/0.5m2)
Kumar, Sunil 2006