ESA 2013 IPM Workshop: An Introduction to Integral Projection Models

  1. Introductory slides
  2. Introductory exercises and data
  3. IPMpack introduction: slides, Metcalf et al MEE 2013
  4. IPMpack vignette
  5. Fecundity in IPMpack slides
  6. Fecundity in IPMpack exercises

Short Course: Modeling Biodiversity Patterns and Ecological Processes

This course is taught as a short module that introduces ecologists to Classical and Bayesian statistical models, maximum entropy models, and simulation models, originally offered in the Spring of 2011. It’s designed as a workshop using R, with a focus on Bayesian modeling and application to species distribution modeling. The original course webpage is here; this page represents an updated version.

PLEASE NOTE THAT THE LINKS BELOW WERE BROKEN WHILE REVAMPING MY WEBSITE AND WILL BE FIXED THE NEXT TIME I TEACH THE COURSE. IF YOU NEED ANYTHING BEFORE THEN, JUST EMAIL.

Course Summary

This module will explore methods for developing predictive models of biodiversity patterns and ecological processes. We will introduce the R statistical programming language and use it to demonstrate exploratory data analysis. We will cover statistical modeling (e.g. Hierarchical Bayes) as well as non-statistical modeling (e.g. Maxent), and simulation modeling (e.g. cellular automaton) approaches and demonstrate how to use these models with datasets provided by EEB faculty members and Intro IPM exercises.r. Participants will be provided with working code to help explore models on their own and under the guidance of instructors.

Course Objectives

The goal of this course is to provide students with experience using models that can be applied directly to their own research. As such we will focus on a developing a complete understanding of simpler models, as opposed to a peripheral understanding of more elaborate models. Each session will focus on one or two different modeling strategies or case studies. These meetings will be structured as a workshop with lecture interspersed with guided programming exercises. We will primarily use the statistical programming language R, and will assume that students do not have programming experience. The goal is to balance building confidence with simple models while considering models with sufficient complexity to be useful. We will provide homework assignments where participants make simple modifications to sample code from models discussed in class. Extra help will be available during regularly scheduled office hours for programming questions.

Prerequisites

As a prerequisite, students will be provided with a basic introductory packet on the statistical software package R that guides them through sample code. While the course will assume that students are unfamiliar with R, we will expect that students can work through some of the most elementary calculations on their own so that class time can be used efficiently. Students will not be expected to write their own code during the course, but will be asked to read prepared code prepared and make basic modifications.

Workshop Content

[ Open All | Close All ]

Day 1

Goal

Introduction to R

Content

Exploratory data analysis with for community abundance patterns
Slides

Excerises

Part 1, Part 2
Data for exercises. You don't need to download these, R will do it for you.
fynbos site abundance.csv
fynbos sites.csv

Homework

Suggested Further Reading

  • An Ecological Modeler’s Primer on R. A tutorial on basics, entering data, functions and plotting.
  • icebreakeR. A clearly written and complete medium-length tutorial.
  • Quick-R. A very succinct guide for a handful of standard analyses. Nicely demonstrates how simply some analyses can be coded.
  • My R resource page for ecologists.

Day 2

Goal

Introduce species distribution modeling using Generalized Linear Models (GLMs) and Generalized Additive Models (GAMs)

Content

Rough introduction to species distribution modeling, GLMS, GAMs.
Slides
Exercises
Data for exercises. You don't need to download these, R will do it for you.
IPANE site data
selected IPANE presence/absence data
In case you're interested, the IPANE website

Homework

Suggested Further Reading

  • Guisan, A., T. C. Edwards, Jr., and T. Hastie. 2002. Generalized linear and generalized additive models in studies of species distributions: setting the scene. Ecological Modelling 157:89-100
  • Yee., T. and N. Mitchell. 1991. Generalized additive models in plant ecology. Journal of Vegetation Science 2: 587-602.

Day 3

Goal

Introduce concepts behind Bayesian modeling

Content

Fundamentals of Bayesian modeling (basic probability rules, Bayes’ Rule, priors and posteriors, examining model output), Building nonhierarchical regression models from scratch.
Download MCMCRobot (scroll down to the bottom of the page). This is a nice little program that Paul Lewis wrote for teaching MCMC concepts. It runs on a PC by clicking on the .exe file. Mac users will have to follow along on this one unless you have experience running pc programs through wine.
Slides
Exercises will be handed out in class
A Bayesian Jargon List to keep track of new terminology

Homework

Suggested Further Reading

  • Clark, J. 2005. Why environmental scientists are becoming bayesian. Ecology Letters 8:2-14 Ellison, A. 2007. Bayesian inference in ecology. Ecology Letters 7: 509–520

Day 4

Goal

Provide experience with more Bayesian models.

Content

Bayesian regression models.
Slides
Exercises will be handed out in class

Homework

- Finish any exercises not completed in class

Suggested Further Reading

  • Clark, J. 2007. Models for Ecological Data. Princeton University Press.
  • Wikle, Christopher K. and L. Mark Berliner. 2007. Bayesian tutorial for data assimilation. Physica D: Nonlinear Phenomena 230: 1-16.

Day 5

Goal

Develop basic skills for hierarchical Bayesian models

Content

General hierarchical modeling discussion, Hierarchical regression, Use canned hierarchical models, Hierarchical Bayesian spatial models for species' distributions.
Slides
Exercises
Data for exercises. You don't need to download these, R will do it for you.
Live Aloe Data
Dead Aloe Data
New England 5' x '5 Environmental Data
IPANE Occurence Data
Preliminary spGLM run for Celastrus

Homework

- None

Suggested Further Reading

  • Ibáńez, I., Silander, J.A, Jr., Wilson, A., LaFleur, N., Tanaka, N., and Tsuyama, I. 2009 Multi-variate Forecasts of Potential Distribution of Invasive Plant Species Ecological Applications 19(2) 359-375.
  • Additional tutorial info for spBayes
  • Gelman and Hill. 2008. Data Analysis Using Regression and Multilevel/Hierarchical Models

Day 6

Goal

Build hierarchical Bayesian models

Content

Use JAGS to develop custom models.
Slides
Exercises

Homework

Suggested Further Reading

  • Simple WinBUGS models
  • WinBUGS Examples Vol. 1, Vol. 2
  • Latimer, A.M., S. Banerjee, H. Sang, E. Mosher and J.A. Silander (2009). Hierarchical models facilitate spatial analysis of large data sets: A case study on invasive plant species in the northeastern United States. Ecology Letters 12:144-154.
  • Agarwal, D. K., J. A. Silander, Jr., A.E. Gelfand, R.E. Dewar, and J.G. Mickelson, Jr. 2005. Tropical Deforestation in Madagascar: Analyses using hierarchical, spatially explicit, Bayesian regression models. Ecological Modelling 185:105-131
  • Hooten, M.B., C.K. Wikle, R.M. Dorazio, and J.A. Royle. 2007. Hierarchical spatio-temporal matrix models for characterizing invasions. Biometrics 63: 558-567

Day 7

Goal

Use Maxent to model species distributions

Content

We will develop the general theory of Maxent, demonstrate how to interpret results and discuss important considerations about settings.
Download Maxent. This is free runs on any operating system. Make sure you've got it up and running before class.
Slides

Excerises

Maxent tutorial Data: Maxent Tutorial Data, IPANE Maxent Data

Homework

Suggested Further Reading

  • Phillips, S., Anderson, R. and R. Schapire. 2006. Maximum entropy modeling of species geographic distributions. Ecological Modelling, 190:231-259.
  • Phillips, S. and M. Dudík.2008 Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation.Ecography 31: 161-175.

Day 8

Goal

Develop in-depth understanding of Maxent

Content

Customizing Maxent models, New Features, Projections under climate change
Required pre-class reading: Elith, J.,S. Phillips, T. Hastie, M. Dudík, Y. Chee and C. Yates. A statistical explanation of MaxEnt for ecologists. Diversity and Distributions, 17:43-57, 2011.
Slides
Exercises

Homework

- Finish Exercises

Suggested Further Reading

  • Phillips, S. M. Dudík, J. Elith, C.H. Graham, A. Lehmann, J. Leathwich and S. Ferrier. 2009. Sample Selection Bias and Presence-Only Species Distribution Models: Implications for Background and Pseudo-absence data. Ecological Applications 19:181-197
  • Elith, J. et al. Novel methods improve prediction of species' distributions from occurrence data. 2006 Ecography 29: 129-151.
  • Elith,J. M. Kearney and S. J. Phillips. 2010.The art of modelling range-shifting species,Methods in Ecology & Evolution 1:330-342

Day 9

Goal

Discuss rule-based mechanistic Cellular Automaton (CA) models

Content

General background on CA models, Spread models, Case study of invasive plant spread in New England. Slides
Exercises
Data for exercises. You don't need to download these, R will do it for you.
New England 5' x '5 Environmental Data
IPANE Occurrence Data

Homework

- None

Suggested Further Reading

  • Merow, C., N. LaFleur, J. Silander, A. Wilson & M. Rubega. 2011. Developing dynamic, mechanistic species distribution models: predicting bird-mediated spread of invasive plants across northeastern North America. American Naturalist 178: 30-43. Supplement
  • Jongejans, E., O. Skarpaas, and K. Shea. 2008b. Dispersal, demography and spatial population models for conservation and control management. Perspectives in Plant Ecology, Evolution, and Sys- tematics 9:153–170
  • Higgins, S. I., and D. M. Richardson. 1996. A review of models of alien plant spread. Ecological Modelling 87:249–265