The Oregon Chapter of The Wildlife Society is hosting a two and a half day RSF workshop, taught by Dr. Ryan Long from the University of Idaho. The workshop will be held October 22-24 2019, in Adair Village (Near Corvallis), at the ODFW Office conference room. Address: 7118 NE Vandenberg Ave, Adair Village, OR 97330. Registration is $200; bring your own food and snacks.
Topics will include:
1) Attendees need to provide their own laptop computers for completing workshop exercises.
2) Computers will need to have the following software installed:
- The most recent (or at least a relatively recent) version of ArcGIS (ArcMap with spatial analyst extension enabled: NOT ArcGIS Pro).
- The most recent 64-bit version of R (https://www.r-project.org/). Attendees are welcome to use R Studio or a similar interface if they want, but I do everything in base R, so that’s how the exercises and demonstrations are set up.
- Make sure you install the lme4 and MASS packages prior to the workshop.
- The 32-bit version of R 3.2.0. The package for fitting resource utilization functions only runs in slightly older versions of R.
- In the 32-bit version of R 3.2.0, the RUF package needs to be installed by running the following line of code: install.packages(“ruf”,repos=”http://www.stat.ucla.edu/~handcock”)
DETAILS: Analysis of Resource Selection by Animals
Course Description
Space-use decisions made by animals in heterogeneous environments can reflect a variety of important processes, including the acquisition and investment of energy, avoidance of mortality from predation or other sources, intra- and interspecific competition, and interactions with both natural and anthropogenic features of the landscape. Consequently, quantifying patterns of resource selection by animals can provide key insights into relationships among the environment, individual fitness, and population dynamics that are critical for making effective management and conservation decisions. Although powerful model-based approaches to quantifying resource selection have been developed in recent years, many managers and researchers continue to use outdated techniques that provide limited insight into complex wildlife-habitat relationships.
The objective of this course is to provide participants with the skills and confidence necessary to proceed from a raw dataset of animal locations and habitat characteristics to a final resource selection function using modern modeling techniques. Course structure will consist of lecture modules in the mornings (roughly 30% of the course) focused on key elements of the background and theory of resource selection analysis, and hands-on computer labs in the afternoons (roughly 70% of the course). Some previous experience with ArcGIS and/or R statistical software will be helpful.
Course Topics
Lecture: Introduction to resource selection analysis
- Central definitions and concepts (use, availability, selection, preference, etc.)
- Spatial and temporal scale (1st through 4th order selection and the importance of daily and seasonal patterns of selection)
- Sampling and study design (the various sampling schemes and units typically associated with resource selection studies)
- Categorical data and selection ratios (2D vs. 3D selection ratios, selection ratios as the response variable in a modeling framework)
- Modeling resource selection (advantages, disadvantages, goals, and steps)
Lecture: Logistic regression
- The logistic model and classic logistic design
- Difficulties of the classic approach
- Mixed-effects logistic regression (with a discussion of conditional logistic regression) Hands-on computer lab: Modeling resource selection using mixed-effects logistic regression
Lecture: Modeling use as a continuous variable
- Resource utilization functions (RUFs; Marzluff et al. 2004, Millspaugh et al. 2006)
- Negative binomial regression (Sawyer et al. 2006, 2007, 2009)
- Hands-on computer lab: Modeling resource selection using the RUF approach
Hands-on computer lab: Modeling resource selection using negative binomial regression Interactive presentation: Mapping predicted probability of use from an RSF across a landscape Interactive presentation: K-fold cross validation
Instructor Contact Information
Ryan Long
Department of Fish and Wildlife Sciences
University of Idaho, Moscow, ID 83844
E-mail: ralong@uidaho.edu
Phone (office): 208-885-7225
For questions about registration or the course in general, please contact barbaragarcia@ortws.org. For technical questions including those related to software, please contact the course instructor Ryan Long at ralong@uidaho.edu.