Data Management: A Case Study

1
Data ManagementA Case Study

• Prairie Vegetation Monitoring at Herbert Hoover
  NHS

2
Whats been done?Data Management SOP

• Data acquisition - complete
• Verification and validation - complete
• Analysis in progress
• Dissemination in progress

3
Background

• Herbert Hoover NHS 186 acres in east-central Iowa
• Historically tallgrass prairie, savanna, and
  forest
• Prairie reconstruction began in 1971
• Ten years management
• Inventories began in 1982 Dr. Paul Christiansen
• Fire as a management tool
• Result an 81 acre tallgrass prairie oasis
• Retirement in 2005 with 20 years data legacy
• Needs data management
• HTLN to continue as Vital Signs monitoring
• Comparison of results
• HEHO also interested in fire effects monitoring

4
Data Acquisition

• Reports obtained from HEHO, MWR, HTLN
• Hardcopy and digital format
• Hardcopy scanned to pdf
• Vegetation data in Excel, otherwise data entry
• Spatial data in Excel, otherwise data collection
  

5
Data Acquisition
Issue
Action

• Reports scattered
• Poor scanning
• Data table images
• Data formats varied
• Some data lost
• Incomplete metadata

• Locate
• Scan again
• Data entry
• Modify to VegMon
• Addl. data mining
• Contact park, etc...

Data were managed in separate databases per year
of data then combined into a single VegMon
geodatabase.
6

• Access NRDT

7

• ArcCatalog data
• Feature classes
• Tables
• Relationship classes

8

• Tallgrass prairie oasis
• 5 methods
• 5 layers

9
Verification and Validation

• Reports reviewed for completeness
• Scientific and common names used in data tables
• Spatial data compared to location notes
• 100 of the records compared to their original
  source
• 10 of the records were reviewed a second time
• Validated for range and logic errors

10
Verification and Validation
Issue
Action

• Reports mixed up
• Taxonomy changes
• Multiple common names
• Spatial data incorrect
• Lookup tables inaccurate

• Rearrange
• Update to HTLN
• Update to HTLN
• Reviewed and corrected
• Modified VegMon tables

11
Analysis

• Used built in VegMon queries
• Exotic species summaries (exotic/native ratio,
  relative frequency, relative cover)
• Plant diversity, species evenness, guilds
• Compare legacy data with HTLN data

12
Document in NatureBib
Dissemination

• Document in NatureBib
• Citation data for annual reports

13
Dissemination

• Generate a new Bibkey ID
• The annual reports

14
Dissemination

• Upload reports
• Connection must be good
• Larger files transferred via ftp or email

a a
15
Dissemination

• Link names to report in NPSpecies

16
Dissemination

• Upload geodatabase
• Include metadata and browse graphic
• Example GWCA mammal inventory

17
Dissemination

• Enter info

18
Dissemination

• Search and find
• Data uploaded

19
Dissemination

• Data components

20
Dissemination

• View before download

21
Dissemination

• The metadata

22
Data Integration

• Christiansen retirement
• HTLN to continue Vital Sign monitoring
• HEHO interested in fire effects monitoring
• Incorporate VegMon and FEAT similar model
• Import/export utility?
• Continue analysis

23
Remaining Tasks

• Complete the analyses and finalize the report
• Complete the spatial metadata
• Continue investigating FEAT import/export utility
• Document and post

24

• Keynote

• The integration of Network and park monitoring
  programs has provided the basis for continued
  prairie monitoring at Herbert Hoover NHS. Such
  actions support the immediate goals of the IM
  program and mission statement of the National
  Park Service and serves to honor Dr.
  Christiansens life long dedication to prairie
  plant ecology.

25
Questions/Comments

• Michael H. Williams
• Data Manager
• Heartland IM Network and
• Prairie Cluster Prototype Monitoring Program
• Department of Biology
• Missouri State University
• 901 S. National Avenue
• Springfield, MO 65897
• Phone [phone]
• Email Michael_H_Williams_at_nps.gov
• http//www1.nature.nps.gov/im/units/htln/index.htm