Tree List Generator TLG v1.0 - PowerPoint PPT Presentation

1 / 16
About This Presentation
Title:

Tree List Generator TLG v1.0

Description:

Initial tree list estimation for individual tree based growth models ... targeted use of the SMC-ORGANON modeling database to develop a methodology ... – PowerPoint PPT presentation

Number of Views:75
Avg rating:3.0/5.0
Slides: 17
Provided by: ericctu
Category:
Tags: tlg | generator | list | organon | tree

less

Transcript and Presenter's Notes

Title: Tree List Generator TLG v1.0


1
Tree List Generator (TLG)v1.0
  • Kevin R. Gehringer
  • Eric C. Turnblom

2
TLG - Topics
  • Introduction / background
  • Models / Methods
  • Results / Discussion
  • Operation / Output
  • Future work / conclude

3
TLG - Introduction
  • Initial tree list estimation for individual tree
    based growth models
  • Basically targeted use of the SMC-ORGANON
    modeling database to develop a methodology for
    generating initial tree lists
  • Tree List Generation Database software provides a
    means for producing reasonable tree lists for a
    forest stand based on a minimal amount of
    information, such as SI, age, TPA, etc.

4
TLG - Introduction (2)
  • Tree list is a compatible set of DBH and height
    measurements for each tree as well as species
  • TLG is designed for managed, mixed species stands
    of Douglas-fir and western hemlock as dominant
    species (with mixtures of other minor species)
  • Untreated stands and thinned stands throughout
    southwestern British Columbia, Washington and
    Oregon west of the Cascade Crest

5
TLG - Introduction (3)
  • Tree list generation requires four things
  • A method for classifying stands
  • A method for selecting an appropriate stand class
  • A method for generating a tree list from a stand
    class
  • A wide variety of stand measurement data sets
  • Useful to think under the framework of generating
    a random stand of trees using a small set of
    stand attributes as the seed

6
TLG - Methods
  • Weibull density distribution based methods
  • Weibull is strongly unimodal
  • Represents flat distributions poorly
  • Mixed species stands present challenges
  • Parameter prediction functions may change with
    new data
  • Non-parametric, density estimation methods (e.g.,
    B-splines, etc.)

7
TLG - Methods (2)
  • Tree List Generation Database method
  • Explicit classification based on a
    multidimensional histogram (data speaks for
    itself)
  • Histogram is a consistent estimator of a density
    function
  • Histogram allows refinement of its classification
    if sufficient data warrant higher resolution
  • Histogram does not require reevaluation of
    classification rules when new data is added
  • Histogram dimensions are stand attributes, e.g.,
    site index, QMD, stand type, etc

8
TLG - Methods (3)
  • Tree List Generation Database method (contd)
  • Stands with similar attributes belong to same
    hist. Bin
  • Given a set of desired stand attributes,
    similarity scores are calculated for stands
    defined by the tree list generation database
  • The k smallest similarity scores are used to
    select the collection of stands to form the
    canonical stand from which a tree list may be
    generated

9
TLG - Methods (4)
  • Two fundamental assumptions
  • Forest stands, or sample plots, having similar
    attributes, e.g., site index, QMD, species
    composition, etc., are similar
  • Attributes of forest inventory data collected
    from a representative set of sample plots from
    within a larger stand may be extended to the
    whole stand

10
TLG - Results
  • Simulated stands generally agreed with the actual
    stands they were generated to mimic at the
    individual tree level defined by DBH, height and
    species
  • Based on a single random draw
  • For DBH alone, misclassification rate was 2.6
  • For height, misclassification rate was 13.2
  • For species comp., misclassification was 8.3

11
TLG - Operation
  • Conventional DOS command line interface
  • Stand Description files specify input values for
    the set of measured or desired attributes used as
    the histogram bin dimensions for which a
    simulated tree list is desired
  • comments appear after a sign
  • KEYWORD VALUE format defines desired
    attributes

12
TLG - Operation
  • Example stand description file
  • TREATMENT_TYPE UNTREATED
  • SITE_INDEX 140.0
  • STAND_TOTAL_AGE 57.0
  • STAND_ORIGIN NATURAL
  • STAND_DENSITY 160.0
  • QMD 12.6
  • STAND_TYPE PURE_DF

13
TLG - Operation
  • Random Measurement files are produced by the tree
    list generation process and contain attributes
    for the stand and a simulated stand of trees
  • Three files are needed to set up or change the
    stand classification attributes
  • Schema file defines multidimensional histogram
    bin dimensions
  • Species mapping file maps 2-6 char. Code to
    integer
  • Stand measurement file added to tree database

14
TLG - Future work
  • Add new Type I and Type III data to the set of
    stand measurement files included in the tree list
    generation database distribution set
  • Add Graphical User Interface (GUI) such as is
    used with DFSIM, that is, DFSIMIN

15
TLG - Possible GUI
16
Tree List Generator
  • Please visit lthttp//www.standmgt.org/gt
  • Click on the SOFTWARE link on L.H.S.
  • Questions / Comments ?

Kevin R. Gehringer Eric C. Turnblom
Write a Comment
User Comments (0)
About PowerShow.com