The National Vegetation Survey Databank A New Zealand Archive

1
The National Vegetation Survey DatabankA New
Zealand Archive
http//nvs.landcareresearch.co.nz Wiser,
Bellingham Burrows. 2001. Managing
biodiversity information development of New
Zealands National Vegetation Survey Databank.
New Zealand Journal of Ecology 25(2) 1-17.
2
Who is involved?
Plant ecologists Peter Bellingham Susan
Wiser Larry Burrows Rob Allen Ian Payton Claire
Newell
Data entry and administration Michelle Breach
Database integration and strategy Jerry
Cooper Grant Hunter
Dept. of Conservation Liason Elaine Wright
Database design and programming Nick Spencer
Web design Mark Fuglestad
3
What is NVS?

• FRST-recognised archive of national importance
• A repository for vegetation data collected over
  the past 50 years including
• physical archive of data sheets, maps, photos
  etc.
• electronic archive of vegetation plot data (from
  early 1970s)
• Built on standardised, data collection efforts of
  former NZ Forest Service
• NZ-wide coverage
• Data now contributed and used by Dept. of
  Conservation, Landcare Research, regional
  councils, university academics and others
• Supported by analysis packages

4
(No Transcript)
5
(No Transcript)
6
(No Transcript)
7
(No Transcript)
8
(No Transcript)
9
Types of data collected on plots
10
(No Transcript)
11
(No Transcript)
12
(No Transcript)
13
Temporal coveragepermanent plots
Forests
Grasslands
7000
7000
6000
6000
5000
5000
Number of plots
4000
4000
3000
3000
2000
2000
1000
1000
0
0
1950s 1960s 1970s 1980s 1990s
1950s 1960s 1970s 1980s 1990s
Established Remeasured
14
Examples of how NVS data is currently being used

• Traditional uses
• Remeasurement of set or subset of a survey to
  assess change in vegetation composition and
  structure
• Succession, natural disturbance impacts, exotic
  plant invasion, exotic animal impacts
• Variation in composition across a region for
  conservation planning
• Vegetation description and classification
• Syntheses
• Use of forest structure data for national level
  estimates of C storage
• Developing national-level models to predict
  occurrence of major tree species in relation to
  climate
• Examining patterns of stand turnover with latitude

15
Why is NVS relevant to DOC monitoring staff?

• Awareness of existing data for your patch
• Context for new monitoring
• Source of data when remeasuring permanent plots
• Safe, archival storage of data

16
Redevelopment of NVS

• Fully function system in 1980s
• Institutional restructuring resulted in decline
  in monitoring and infrastructure to support it
• Efforts to rebuild over past 5 years now gaining
  significant momentum
• Consequence -- state of the art, user friendly
  systems dont yet exist, but are in the pipeline

17
State of Play 1 ½ years ago

• Data stored in ASCII in condensed form that could
  not be readily exported to the range of formats
  desired
• Data entry relied on a DOS-based system (Reflex)
  that few people knew how to use
• It was difficult to link encoded data within NVS
  to other data sets (e.g. PLANT NAMES)
• Data from different surveys was stored separately
  and was difficult to combine for synthetic
  analyses
• Unnormalised, flat file system led to internal
  inconsistencies and difficulty linking different
  types of data collected on a plot at a point in
  time or over time
• ASCII file formats too rigid to store all the
  data being collected

18
Example of storage formats
19
Database Integration project NVS Interim
Platform

• To move NVS data to a more flexible
  client-server relational database (SQL Server)
  with an MS-Access front-end, which requires
• A new relational database design
• No remodeling but remove constraints
• New linkages between metadata and vegetation data
• Interfaces for viewing and editing NVS data
• Migration of historic data to new database
  structures
• Ability to export data to a range of formats,
  including ASCII formats used for analysis
  packages

20
Progress to Date

• Relational database has been designed and
  implemented for metadata, composition, site, and
  forest plot data
• Interfaces for viewing and editing data have been
  created
• Metadata has been migrated to the new database
• Vegetation data has been migrated to the database
• Interface in operation

21
Database Integrationoutstanding needs

• Adopt or design a more efficient and robust model
  than interim solution
• Conduct data integrity checks where redundancy
  currently exists
• Site data
• Tree diameter data in time series
• Decide on best mechanisms to deal with data
  collected using non-standard methods

22
Externally funded projects 2001-2002

• NVS survey metadata on the Web (DOC)
• GIS interface for NVS plot location data on the
  Web (Landcare)
• Why?
• to increase awareness and use of data in NVS
• To enable potential users to determine if data
  stored in NVS will meet their needs
• To facilitate requests for data
• Gaps in NVS data will be more easily identified
• To improve data quality
• Pilot study in NZ on interagency data sharing
• Develop experience in use of new ESRI product
  range (Arc IMS/SDE)

23
Funded projects 2002-03

• Database Integration
• Access to interim NVS system for data entry
  uploading
• Design best-practice data models for NVS
• CMS demonstration module (Landcare)
• Internet capacity to enter and validate field
  data
• Generate key statistics required by policy makers
  
• TFBIS
• NVS prototype - model for Internet-based user
  access to data
• Digitise backlog ecological field survey data for
  vegetation surveys
• PGSF
• Complete migration of NVS data
• Ongoing maintenance and requests

24
How do I enter vegetation plot data?

• Landcare Research can enter data for a fee
• NVS prototype interface will be available early
  in 2003
• Data entry yourself in EXCEL

25
Data checking and correction

• Validation in interface
• Checks as data is entered
• Data types correct
• Certain fields compulsory
• Pick lists for coded values (Species, tiers,
  cover classes, categorical site variables)
• Prohibits duplicate plots, subplots, species
  within tiers, tree tags
• Range checks
• Checking routines in PC-packages
• Plot by plot check after data entered
• Species specific range checks
• growth/shrinkage of individual trees
• date sequences
• misalignment of data

26
Errors not checked by current system

• Tagged trees
• Species names changing
• Previously tagged trees retagged with new tag
  number
• Temporal checks
• Consistency with earlier site measurements
• Consistency with species records (e.g.
  nomenclature, ids)
• Logic checks
• Species only recorded within known range limits
• Plot locations recorded are within limits of
  survey area
• Altitude and slope consistent with plot location

27
PC Analysis Packages

• PC-RECCE
• PC-DIAM
• PC-USTOREY
• PC-TRANSECT

28
PC packages

• Check data
• Classification of plots into groups based on
  composition (total, trees, seedlings or saplings)
• Summaries of all plots, individual plots, or
  groups
• Frequency abundance of species (by tiers, size
  classes etc.)
• Basal area, growth, mortality of trees
• Density of trees, seedlings, saplings
• Site factors
• Export of summarised data suitable for import
  into graphics or statistical packages

29
Pitfalls when analysing change

• Nomenclature
• Tagged trees

30
Flow diagram of PC-Recce options
Stats package
PC-Ord
31
Flow chart of PC-Diam options
Stats package
Graphics package
32
Advanced analyses

• Regression modelling
• Statistical software SAS, SYSTAT, STATISTIX
• Multivariate analysis of composition data
• CANOCO
• PC-ORD
• Requires advanced training and considerable
  expertise

33
Interpreting vegetation plot data

• Use of exclosures to examine impacts of fallow
  deer on kanuka forest
• Patterns of tree mortality and recruitment
  throughout New Zealand
• Invasion of mountain beech forest by an adventive
  herb
• Influences of white-tailed deer on forest
  regeneration
• Establish the distribution and quantity of forest
  species that are key takehe food plants

34
Description of forest in proposed wapiti area,
West Nelson (Rose 1985)

• 1000 Recce plots to define and describe
  communities and their relationships to
  environmental factors
• Cluster analysis to define communities at 3
  levels
• Type groups with similar phyiognomic dominants
• Association groups within a type grouped at a
  higher level of similarity
• Sub-association groups within associations
  that are highly related

35
(No Transcript)
36
(No Transcript)
37
Invasion of mountain beech forest by an
adventive herb (Wiser et al. 1998)

• Permanent plots in mountain beech forest measured
  in 1970, 1985 and 1993
• Used recce data, seedling subplot data to examine
  changes in abundance of Hieracium lepidulum
• Used site data, community structure data, and
  stand dynamics data to interpret patterns

38
Progress of Hieracium lepidulum invasion
1970 - 11
1985 - 43
Plots invaded Plots not invaded
1993 - 57
39
Progress of Hieracium lepidulum invasion
1970 - 11
1985 - 43
Plots invaded Plots not invaded
1993 - 57
40
Predictability over time
60
with soils
50
without soils
40
deviance explained
30
20
10
0
1985
1993
Year
1970
Forest-margin dispersal limited
More species rich sites on fertile soils
41
Effects of an earthquake on forest structure
(Allen et al 1999)

• 250 permanent plots established in 1970
• Remeasured multiple times
• Allowed examination of impacts of a range of
  disturbances heavy snow, pathogens, Arthurs
  Pass earthquake in 1994

42
Multiple disturbances
200
180
160
Biomass (Mg/ha)
140
120
100
0
1970
1980
1990
2000
Year
Snow Storm
Pathogen
Earthquake
43
Influences of white-tailed deer on forest
regeneration(Bellingham Allan 2002)

• Stewart Island
• 5 exclosures and controls established in 1979
• 84 permanent plots
• Mixture of objective vs. subjectively located
• Different measurement histories
• Pellet counts

44

• Seedling density increased in both exclosures
  and controls
• Proportionally higher density in exclosures

45
References provided

• Resource description
• Rose 1985
• Stewart et al. 1993
• Description of change
• Wiser et al. 1998
• Allen et al. 1999
• Change in response to management
• Allen et al. 1984
• Smale et al. 1995
• Bellingham et al. 2002
• Large-scale patterns in forest dynamics
• Bellingham et al. 1999

46
Contact information

• Web site http//nvs.landcareresearch.co.nz
• E-mail nvs_at_landcareresearch.co.nz
• Data requests and general info
• Michelle Breach / Larry Burrows 03-325-6700