Review of the Growing Modeling Toolkit

1
Review of the Growing Modeling Toolkit

• Bruce G. Marcot
• USDA Forest Service
• Portland, Oregon USA

2

• Marcot, B. G. 2006. Review of the growing
  modeling toolkit special session. Presented 5
  December 2006 at Habitat and Habitat Supply
  Modeling Practitioner's Workshop, 5-7 December
  2006. Ministry of Forests, Research Branch,
  British Columbia, Canada. Invited. Chase, B.C.
  Canada.

3
Models, Models, Models
4
Models, Models, Models
5
Models, Models, Models
6
Models, Models, Models
7
Models, Models, Models
8

• Lots of models to choose from !

9

• measurable surrogate

inference ?
biodiversity parameter
10

• Influence diagrams

11

• Influence diagrams
• Concept mapping
• Concept diagrams
• Cognitive Map
• Mental Map, Mind Map

http//intraspec.ca/cogmap.php http//www.cs.joen
suu.fi/marjomaa/Knowledge_Representation/doc/Know
ledge_Representation-56.htm
12
Influence diagrams
Source Marcot, B. G., et al. 2001. Forest
Ecology and Management 153(1-3)29-42.
13

• Influence diagrams
• Mindjet MindManager Pro
• Inspiration
• Personal Brain
• Netica

14
Mindjet MindManager Prohttp//www.mindjet.com
15
Inspirationhttp//www.inspiration.com/
16
Personal Brianhttp//www.thebrain.com/
17
Neticahttp//www.norsys.com
18
IHMC CmapToolshttp//cmap.ihmc.us/
19

• Building influence diagrams
• empirical data
• expert judgment / opinion
• knowledge engineering
• peer review
• expert paneling (e.g., Delphi)
• combination

20

• From influence diagram
• to models galore !

21
Path regressionQuality Deer Management (QDM)
Source Woods, G. R., D. C. Guynn, W. E. Hammitt,
and M. E. Patterson. 1996. Determinants of
participant satisfaction with quality deer
management. Wildl. Soc. Bull. 24(2)318-324.
22
Source Hudson, R. J. 1995. Paths to
conservation. Pp. 318-322 in J. A. Bissonette
and P. R. Krausman, ed. Integrating people and
wildlife for a sustainable future. The Wildlife
Society, Bethesda, Maryland. 715 pp.
Process model STELLA http//www.iseesystems.com/
23
Process model STELLA http//www.iseesystems.com/
24
neural network
25
Types of Models

• Analytic and numerical population models
• Leslie matrix life tables
• Genetic models of inbreeding, genetic drift
• Simulation models
• GIS-based models
• Spatially explicit, individual-based models
• Knowledge-based (expert) models
• Expert systems
• Other expert-based models

26
Types of Models

• Statistical empirical models
• Correlation, multivariate models
• Regression tree, classification tree

27
Regression tree 3 viability risk levels
28
Types of Models

• Statistical empirical models
• Correlation, multivariate models
• Structural equation models (SEMs) a
  modeling procedure

29
Structural Equation Models (SEMs)

• A way to formalize and construct relationships
  among variables.
• Observational data
• A generalization of many statistical techniques
• Regression, discriminant analysis, canonical
  correlation, factor analysis
• Differentiates among direct relationships,
  indirect causal relationships, spurious
  relationships, association without causation

30
Structural Equation Models (SEMs)

• Create the model structure as an influence
  diagram including unexplained variance.
• Expand the latent variables into their components
  e.g., habitat into measurable veg. variables.
• Compute regression weights for each variable.
• Partial correlation analysis
• Bayesian conditional probabilities
• Estimate measurement errors of each component
  variable.
• This depicts the amount of uncertainty in the
  habitat-species relations represented in the
  model.

31
Structural Equation Models (SEMs)

• The final SEM model depicts
• Specific variable relations
• Degree of uncertainty of those variables
• The relations among the variables
• SEM tests the hypothesized underlying causal
  relations among variables by analyzing their
  covariance structure.
• Goodness-of-fit tests of congruence between the
  variance-covariance matrix derived from
  observational data to that suggested by the
  hypothetical causal structure of the model (the
  predicted moment matrix).

32
Structural Equation Models (SEMs)

• Methods of estimation for the goodness-of-fit
  tests
• MLE (maximum likelihood estimation), for
  multivariate normal data Ngt200 samples
• WLS (weighted least squares asymptotically
  distribution free) methods, for continuous but
  nonnormal data
• Polychoric correlation analysis, for ordinal
  variables (computes correlation between
  unobserved normal variables then uses WLS
  methods)
• Software for doing SEM
• LISREL, EQS, AMOS, CALIS, SYSTAT

33

• Statistical empirical models
• Post-hoc pattern analysis
• Knowledge discovery
• Rule induction (problems w overfitting data)
• Data mining (association analysis)
• Text mining

34
Types of Models

• Statistical empirical models
• Post-hoc pattern analysis
• Knowledge discovery
• Rule induction
• Data mining
• Text mining

35
Text Mining

• Biodiversity
• biocomplexity
• ecological complexity
• ecological functions
• disturbance regimes
• ecosystem resilience
• stability, resistance
• ecological integrity
• ecosystem services
• sustainability . etc.

36
Text Mining

• Biodiversity
• biocomplexity
• ecological complexity
• ecological functions
• disturbance regimes
• ecosystem resilience
• stability, resistance
• ecological integrity
• ecosystem services
• sustainability . etc.

• gt13,000 references
• EndNote biblio. database
• concept proximity analysis

37
Marcot, B. G. In revision. Biodiversity and the
lexicon zoo. Forest Ecology and Management

• Text mining Concept map

38
http//www.kartoo.com
39
http//www.kartoo.com
40
Data mining Information mapping - topographical
maps - closeness maps - interactive trees -
concept clustering - (many others)
41
Decision-Support Models
42
Decision-Support Models

• Many tools
• Bayesian statistics, Bayesian belief networks
• Data and text mining
• Decision tree analysis
• Expert systems
• Fuzzy logic, fuzzy set theory
• Genetic algorithms
• Rule and network induction
• Neural networks
• Reliability analyses
• Landscape simulators

43
Bayesian Belief Network Model
44
Influence Diagrams as Bayesian Belief Network
Model
45
Influence Diagrams as Bayesian Belief Network
Model
46

• sensitivity analysis
• identifies most influential factors
• identifies degree of influence

47
Influence Diagrams as Bayesian Belief Network
Model
48
Influence Diagrams as Bayesian Belief Network
Model
Node Mutual Variance of
---- Info Beliefs
Caves or mines 0.02902 0.0069284
Lg snags or trees 0.00953 0.0023053
Cliffs 0.00599 0.0014514
Forest edges 0.00599 0.0014514
Bridges, buildings 0.00063 0.0001543
Boulders 0.00002 0.0000038
49
Fuzzy logic model NetWeaverPenn St. Univ.
fuzzy logic (NetWeaver, Penn St. Univ.)
50
fzcalc.exe http//www.vspdecision.uni-hannover.de/
Fuzzy arithmetic
51
fzcalc.exe http//www.vspdecision.uni-hannover.de/
Fuzzy arithmetic
52
EMDS Ecosystem Management Decision
Supporthttp//www.institute.redlands.edu/emds/
53
EMDS Ecosystem Management Decision
Supporthttp//www.institute.redlands.edu/emds/
Stream condition depends on reach
condition riparian vegetation bank
stability w/d ratio pool frequency large
wood spawning fines water temperature
54
Decision-Support Models

• inputs expressed as management activities ...
  or environmental variables potentially affected
  by management
• outcomes expressed as probabilities ... risk
  management
• explicitly show management decisions and values
  of outcomes

55
decision tree (DecisionPro, Vanguard Software
Corp.)
56
Influence Diagrams as Bayesian Belief Network
Model
57
Influence Diagrams as Bayesian Belief Network
Model
58
Decision-Support Models

• Many decision-analysis techniques are available
• Value of perfect information
• Value of additional or sample information
• Credibility of information
• Quantitative measures of the state of knowledge

59
Decision-Support Models

• Many analysis methods
• MAUT (Multi-Attribute Utility Theory)
• Goal Hierarchy
• AHP (Analytic Hierarchy Process)
• MCDM (Multiple Criteria Decision Making)
• EPAs Quantitative Risk Analysis

60
Decision-Support Models --Useful Model Attributes

• probability-based
• accounts for missing data
• provides for sensitivity testing
• provides management hypothesis (adaptive
  management)
• incorporates new data to update model functions,
  probabilities, structure
• allows rapid prototyping
• combines expert judgment w/ empirical data
  multiple experts

61
Population Models
62
Population Models

• Populus general popn dynamics(www.cbs.umn.edu/p
  opulus/)
• Vortex (www.vortex9.org/vortex.html), Nemesis
  PVA, viability
• RAMAS (www.ramas.com)
• RAMAS Red List, RAMAS GIS, RAMAS MetapopRAMAS
  Landscape, RAMAS Multispecies AssessmentRAMAS
  Ecotoxicology, RAMAS Ecosystem, RAMAS Stage
• Outbreak disease (www.vortex9.org/outbreak.html)
  
• SELES spatially explicit landscape event
  simulator (http//www.ncgia.ucsb.edu/conf/SANTA_F
  E_CD-ROM/sf_papers/fall_andrew/fall.html)
• Biomapper habitat suitability mapping, GIS
  (www2.unil.ch/biomapper/)
• PATCH individual, spatially-explicit
  simulator(Nathan Schumaker, U.S. EPA)

63
Models of Vegetation Processes

• Vegetation growth, ecological succession,
  ecosystem disturbance, other processes
• Techniques
• Markov chain
• Transition matrix analysis
• Loop analysis, graph theory
• Models
• VDDT, ESSA, MARXAN, ECOSIM ECOPATH, PATH

64
Models of Optimizing Land Allocations

• Many techniques
• e.g., genetic algorithms, neural networks,
  adaptive kernel
• Multi-objective integer programming (MOIP)
• Nonlinear integer programming
• Models SITES, BioMapper
• Approaches reserve complementarity, redundancy,
  representativeness

65
Managing Under Uncertainty

• many tools in the toolkit
• make explicit the inference from habitat to
  biodiversity parameter
• display uncertainty of parameters and
  relationships
• useful for risk management
• clearly show risk attitude
• clearly articulate decision criteria
• very helpful for monitoring, adaptive research