Date: 2/22/22

Location: Zoom

Speaker: Adam Mahood

Abstract: Multivariate analysis can be confusing. In plant community ecology, for example, a single site can have tens or even hundreds of species. It can be hard to understand how external and internal factors interact to affect species communities. Often ecologists use methods like principal components analysis or non-metric multidimensional scaling that simplify the variance to try and understand effects on species, but these analyses can be kind of vague and hard to interpret. Recently, with the advancement of both computational capacity and statistical techniques, several new methods have been developed that allow researchers to simply model every species in the community at once. This is called joint species distribution modelling (JSDM). In this talk we will use some data from a 2019 study to illustrate how to create a JSDM, using the R package Hmsc. Attendees have the option to clone the github repository in advance, download the appropriate R packages, and follow along.

https://github.com/admahood/ff_study

Bio: Adam Mahood is a post-doc at Earth Lab. He studies the ecology of fires, plants and plant invasions. He is not by any means an expert in joint species distribution modelling, but he has used it a few times and thinks it's really cool.