BMSC Bamfield Marine Sciences Centre

Instructor(s): Dr. Kyle Wilson

Overview: Ecology is often applied and quantitative in practice. We observe nature, taking data across time and space, and then want to infer whether these observations are consistent with hypothesized patterns and process. Unfortunately, variation is a pervasive feature of nature and it can confound our ability to make meaningful inferences of the world around us. Enter statistics – which provides important tools that help us to surmount these challenges and tease apart signal from noise.

This course provides a bridge between introductory statistics and ecological practice by learning to confront models with real-world data. Students will learn basic and applied skills in this course including: conceptualizing models, translating concepts into code in R or Stan, fitting models to data via probability distributions, generating predictions, and comparing models. This course will go over likelihood and Bayesian approaches throughout the course and compare these two paradigms. Students will learn (or relearn) statistical approaches including linear, generalized linear, nonlinear, hierarchical, and integrated models. In each case, we will try to connect these lessons to local ecosystems and real-world ecological data. This course will allow students to address their own statistical questions through a course project offering opportunities to collect their own field data in local ecosystems, tackle their own pre-existing datasets, or using model simulation to generate data and understanding key concepts in statistics.

Research Skills: Students will improve on their abilities to translate ecological questions into quantitative models and confront models with data – each of which are key to modern ecological practice.

Practical Skills:  Ecology has increasingly become a quantitative, data-rich, and applied science. Students will become comfortable handling and manipulating large datasets to answer questions of interest and generate quantitative information through reproducible analyses, which are important skills for careers in academic, government, Indigenous, environmental non-government organizations, or private organizations (even outside of ecology). This course also develops skills in science communication through presentation of data and models in written, visual, and oral form.

Boat Use: You will be given the opportunity to drive boats if you choose to do so. Boat driving is recommended but not required for this course. Students may wish to drive boats so they can collect data by boat. Students who wish to drive boats at BMSC must hold a PCOC and valid first aid certificate and will participate in an introductory boat check-out on the first day of orientation.

Prerequisites: Students should have some comfort with quantitative skills relevant to ecology, with at least one of the following courses: (1) statistics, (2) calculus, and/or (3) quantitative ecology. Interest and experience in coding or programming relevant to data sciences would be beneficial. Students lacking any of these pre-requisites should email the instructor to get permission to take this course but should be prepared to do some additional legwork. All students should be willing to learn some statistical theory and apply it with programming and coding, primarily in R and Stan.

Physical Requirements: All students should be comfortable working long hours at their computer. Additionally, we will have opportunities to be outdoors for short lessons or field work on shoreline or from boats. For those interested in this, students should be comfortable on open boats, walking rugged shorelines in all types of weather. Field work is not mandatory, however, and accommodations will be made for students with any disabilities or other accessibility concerns. Students should note that Bamfield Marine Science Center is a remote and rural area and, as such, access to medical services can be somewhat limited (for physical or mental well-being).

Textbook (required): Statistical Rethinking: A Bayesian Course with Examples in R and Stan, 2^nd edition. Richard McElreath (2020). I recommend getting a hard copy of the 2^nd edition (should be available from university libraries or available for purchase). Alternatively, however, there are online .pdfs available of the slides and lectures and a less expensive 1^st edition.