Download presentation
Presentation is loading. Please wait.
Published byClarence Ramsey Modified over 8 years ago
1
Workshop on Applied Hierarchical Modeling in BUGS and unmarked Patuxent Wildlife Research Center November 2015
2
AHM Book
3
Overview of unmarked Patuxent Wildlife Research Center November 2015
4
unmarked
5
Overview Models have a common data structure “metapopulation” sampling structure: multiple sites (spatially structured) and replicate samples (repeated measurement)!
6
unmarked
7
Origins Originally created by Ian Fiske abt. 2007/2008 Richard Chandler https://chandlerlab.wordpress.com/ Fiske, I., & Chandler, R. (2011). unmarked: An R package for fitting hierarchical models of wildlife occurrence and abundance. Journal of Statistical Software, 43(10), 1-23.
8
unmarked Unified framework for data analysis – Data manipulation – Data exploration – Model fitting (MLE) – Model selection – Model averaging – Goodness-of-fit tests – Prediction – Empirical Bayes estimators (BUP) – Publication quality graphics – Species distribution maps – Power analysis – Simulation studies – Etc… Also get this package: Mazerolle, M. J. (2011). AICcmodavg : model selection and multimodel inference based on (Q) AIC (c). R package version, 1(1.15).
9
What models are available in unmarked ? Available models Single-season occupancy models (MacKenzie et al. 2002) Royle-Nichols model (Royle and Nichols 2003) False-positive occupancy model (Royle and Link 2006; Miller et al. 2011) Binomial N-mixture models (Royle 2004b) Multinomial N-mixture models (Royle 2004a) Hierarchical distance sampling (HDS) models (Royle et al. 2004) “Open population” versions of the above: MacKenzie et al. (2003), Chandler et al. (2011), Dail andMadsen (2011).
10
What models are available in unmarked ? Models we would like to add Dynamic HDS models (Sollmann et al. 2015) Multi-state occupancy models (Nichols et al. 2007) Multi-species occupancy models for community data (Dorazio and Royle 2005)
11
What model should I use? Three questions: Abundance or occurrence? Static or dynamic? Sampling method?
12
What model should I use?
13
Data structure Sample units are called “sites” We record multiple “observations” at each site – In occupancy studies we make multiple “visits” – Could be observer or removal counts
14
Data structure All data must be put into an object called an unmarkedFrame umf <-unmarkedFrame(y=detectionData, siteCovs=siteData, obsCovs=list(wind=windData, date=dateData)) Why Bother? The data structure is not amenable to simple data.frames and lists Metadata such as distance interval cutpoints or measurement units can be bundled with the data Makes it easy to detect errors when created the unmarkedFrame Allows for custom tools to summarize and visualize data
15
Data structure
17
Typical (standardized) work flow
18
Typical work flow
21
Looking Ahead – Covariate Relationships
23
Looking Ahead – Empirical Bayes
24
Looking Ahead – Population Trends
25
Looking Ahead – Species Distribution Maps
26
Online Resources
27
Get involved! Developers wanted! Write a vignette or improve a help file Answer a question on the email list
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.