Analyzing environmental‐trait interactions in ecological communities with fourth‐corner latent variable models

Abstract

In ecological community studies it is often of interest to study the effect of species related trait variables on abundances or presence‐absences. Specifically, the interest may lay in the interactions between environmental and trait variables. An increasingly popular approach for studying such interactions is to use the so‐called fourth‐corner model, which explicitly posits a regression model where the mean response of each species is a function of interactions between covariate and trait predictors (among other terms). On the other hand, many of the fourth‐corner models currently applied in the literature are too simplistic to properly account for variation in environmental and trait response and any residual covariation between species. To overcome this problem, we propose a fourth‐corner latent variable model which combines the following three features: latent variables to capture the correlation between species, fourth‐corner terms to account for environment‐trait interactions, and species‐specific random slopes for modeling excess heterogeneity between species in their environmental response. We perform an extensive numerical study comparing a variety of fourth‐corner models available in the literature which account for the aforementioned sources of variation to varying degrees. Simulation results demonstrate that the proposed fourth‐corner latent variable models performed well when testing for the fourth‐corner (interaction) coefficients, across both Type I error and power. By comparison, some models that do not full account for all relevant sources of variation suffer from inflated Type I error leading to potentially misleading inference. The proposed method is illustrated by an example on ground beetle data.

Estimating interaction matrices from performance data for diverse systems

“…Our github repository also contains the data _prep.R file, which will show how to calculate these vectors from the input data. Secondly, we impose the following constraints to improve convergence, avoid overparameterisation and maintain identifiability of our parameters r d k and e j (Huber et al 2004; Kidziński et al 2021; Niku et al 2021). We define the effect parameters as a unit vector, which means we only require K-1 degrees of freedom (where K is the total number of neighbour elements) to estimate all effect values.…”

Hay meadows' overriding effect shapes ground beetle functional diversity in mountainous landscapes

“…ter Braak (2019) showed, however, that Jamil's model is not robust against unobserved environmental variables or other characteristics of sites (plots or pitfalls) that are uncorrelated with the environmental variable(s) in the model but do interact with the trait(s) in the model. For more robustness, ter Braak (2019) included a random term for site‐specific trait effects in the model, which performed well in the simulation study by Niku et al (2021) in comparison with their even more sophisticated generalized linear latent variable models, a Bayesian version of which has been developed by Ovaskainen et al (2017). It is not completely clear what adaptations of these models for hierarchical designs retain the robustness.…”

Trophic strategies explain the ocean niches of small eukaryotic phytoplankton

“…In summary, this model quantifies whether the relationship between relative abundance and an environmental variable for an OTU is predicted by that OTU’s clearance rate. The GLMM approach is appropriate because it models # of reads while accounting for variation in total reads, and allows for uncertainty in relative abundances and environmental relationships while quantifying CReff ( 45, 46 ). The assumption of logit-linear environmental responses was appropriate based on visual inspection of the data (Fig.…”