Transient Models

1
Transient Models
2
Transient Models

• Extension of the CJS
• Attempt to separate mortality and permanent
  emigration
• Equivalent to Brownie and Robson Model, but
  developed for different reasons
• B and R, Model 2
• Initial capture/handling may affect the survival
  of marked individuals
• Separate estimate for newly (f'i) and previously
  marked (fi) individuals
• Change diagonal in PIM

3
Transients

• Survival of newly and previously marked
  individuals may also differ because
• Initial sample includes non-residents
• Transients individuals with a future capture
  probability of 0.0
• Permanent emigration after initial capture
• Structure equivalent to Brownie and Robson

4
Transient Structure

• Survival of newly marked individual
• f'i (1 - ?i)fti   ?ifri
• Where, ?i is the probability that a newly marked
  individual is a resident
• fti and fri are the survival probabilities for
  transients and residents, respectively.
• fti 0.0, thus f'i ?ifri and
• ?i f'i / fi
• Identifiably with first cohort in t-specific
  model
• Implemented by conditioning on second capture

5
Age Structured Models

• Another reason for differences in initial
  survival
• f'i (1 - ?'i)fyi   ?'ifai
• where, ?i is the probability that a newly marked
  individual is an adult and
• fyi and fai are the survival probabilities for
  young and adult animals, respectively.
• First interval survival probability is the
  weighted average survival probability of young
  and adult animals. 
• fyi is unknown and neither fyi nor  ?'i can be
  estimated, unless the model is parameterized as a
  transient model
• . Need ancillary data to determine which source
  of heterogeneity is operating

6
How to detect heterogeneity?

• Difference in survival of newly and previously
  marked
• LRT between model 2 and next simplest model w/o
  this structure
• Interpret components of GOF (TEST3.SR release)

7
Transients - Mechanics

• Condition on second capture
• Brute force
• No estimate of proportion residents
• Model 2 in program MARK
• Adhoc estimate of ?i f'i / fi
• Program TMSURVIV
• Computes estimates of survival and capture
  probability and the proportion of transients in
  open population capture-recapture data
• http//www.mbr-pwrc.usgs.gov/software.htmla

8
Transients Problems?

• Heterogeneity
• How to separate low capture or survival from true
  transients
• Effects of temporary emigration and permanent
  emigration following initial capture
• What is the definition of a population?
• Should it include transients?
• Spatial and temporal variation in transient
  behavior when used for monitoring