Index: MacKenzie et al 2002 - Site Occupancy with Imperfect Detection

MacKenzie et al 2002 - Site Occupancy with Imperfect Detection

Routing Summary

Full capture of MacKenzie, Nichols, Lachman, Droege, Royle and Langtimm (2002), Ecology 83(8): 2248-2255: the occupancy likelihood with imperfect detection, simulations, the Maryland frog example, and extensions. 4 notes.

• Need the paper on one page, headline numbers, PyMC example link, or the section-to-note map? → MacKenzie 2002 - Site Occupancy with Imperfect Detection - Overview
• Need the likelihood (Eq. 1), covariates, missing visits, bootstrap SEs? → MacKenzie 2002 - Occupancy Model and Likelihood
• Need the 243-scenario simulation or the toad and peeper results (Table 1)? → MacKenzie 2002 - Simulation Study and Maryland Anuran Results
• Need design rules, caution rule, extensions, or a PyMC marginalised-occupancy sketch? → MacKenzie 2002 - Extensions, Design Guidance and PyMC Sketch

Concept Map

Concept	Note	Type	Depends On	Key Result
Sampling assumptions (closure, no false positives)	MacKenzie 2002 - Occupancy Model and Likelihood	definition	-	sites closed during surveys
Occupancy likelihood	MacKenzie 2002 - Occupancy Model and Likelihood	theorem	assumptions	all-zero history = $\psi \prod (1-p)+1-\psi$
Simulation design and findings	MacKenzie 2002 - Simulation Study and Maryland Anuran Results	example	likelihood	unbiased if $p\ge 0.3$, $T\ge 5$
Frog Table 1	MacKenzie 2002 - Simulation Study and Maryland Anuran Results	example	likelihood	toad 0.49 vs naive 0.34
Caution rule and design guidance	MacKenzie 2002 - Extensions, Design Guidance and PyMC Sketch	theorem	simulation	doubt $\hat{\psi }\approx 1$ with $p<0.15$, $T<7$

Notes

• MacKenzie 2002 - Site Occupancy with Imperfect Detection - Overview — CONTAINS: question, contribution, key results, PyMC E28 link, Source Map.
• MacKenzie 2002 - Occupancy Model and Likelihood — CONTAINS: assumptions, site-level and pooled likelihood, bootstrap, logistic covariates, missing visits, worked non-detection example, Python MLE.
• MacKenzie 2002 - Simulation Study and Maryland Anuran Results — CONTAINS: simulation factors and findings table, field data protocol, Table 1 reproduced with interpretation.
• MacKenzie 2002 - Extensions, Design Guidance and PyMC Sketch — CONTAINS: caution rule, design guidance, robust design, co-occurrence parameter growth, heterogeneity, PyMC sketch.

Sources

• MacKenzie 2002 - Estimating Site Occupancy Rates When Detection Probabilities Are Less Than One — USGS ARMI copy (8 pp). Ecology 83(8): 2248-2255.

See Also

• Barker 2018 - Reliability of N-Mixture Models - Overview — abundance counterpart.
• Concept - Marginalising Latent Discrete Counts for HMC, Hierarchical Models, Monsters and Mixtures.