Joint species distribution modelling with ther‐package Hmsc

Abstract

Joint Species Distribution Modelling (JSDM) is becoming an increasingly popular statistical method for analysing data in community ecology. Hierarchical Modelling of Species Communities (HMSC) is a general and flexible framework for fitting JSDMs. HMSC allows the integration of community ecology data with data on environmental covariates, species traits, phylogenetic relationships and the spatio‐temporal context of the study, providing predictive insights into community assembly processes from non‐manipulative observational data of species communities. The full range of functionality of HMSC has remained restricted to Matlab users only. To make HMSC accessible to the wider community of ecologists, we introduce Hmsc 3.0, a user‐friendly r implementation. We illustrate the use of the package by applying Hmsc 3.0 to a range of case studies on real and simulated data. The real data consist of bird counts in a spatio‐temporally structured dataset, environmental covariates, species traits and phylogenetic relationships. Vignettes on simulated data involve single‐species models, models of small communities, models of large species communities and models for large spatial data. We demonstrate the estimation of species responses to environmental covariates and how these depend on species traits, as well as the estimation of residual species associations. We demonstrate how to construct and fit models with different types of random effects, how to examine MCMC convergence, how to examine the explanatory and predictive powers of the models, how to assess parameter estimates and how to make predictions. We further demonstrate how Hmsc 3.0 can be applied to normally distributed data, count data and presence–absence data. The package, along with the extended vignettes, makes JSDM fitting and post‐processing easily accessible to ecologists familiar with r.

Fungal communities decline with urbanization—more in air than in soil

“…To examine in more detail the responses of the most prevalent species to the environmental and spatial predictors, we fitted the joint species distribution model of Hierarchical Modeling of Species Communities (HMSC, [ 35 ]) separately to the air (model HMSC-air) and to the soil (model HMSC-soil) data with the R-package Hmsc [ 36 ]. For these analyses, we selected species which occurred in at least 20 samples of each sample type.…”

A database and synthesis of euglossine bee assemblages collected at fragrance baits

“…We analysed several subsets of the data by fitting latent variable joint species distribut i o n m o d e l s u s i n g t h e H i e r a r c h i c a l Modelling of Species Communities (HMSC) framework of Ovaskainen et al (2017), implemented in the Hmsc 3.0 R package (Tikhonov et al 2020).…”

“…We initially attempted to fit the model with Poisson log-normal errors but experienced poor mixing properties of the MCMC sampling scheme, a known problem in MCMC-based joint species distribution models (Tikhonov et al 2020). We therefore chose to analyse the data using a 'hurdle' approach, where we fitted one model with binomial errors (probit link) to data truncated to presence-absence and a second model with Gaussian errors to log-transformed species abundances conditional on presence (i.e.…”

Species co‐occurrences in ectoparasite infracommunities: Accounting for confounding factors associated with space, time, and host community composition

“…speciations, extinctions, and colonisations), and (c) neutral processes (Weiher et al, 2011;Ovaskainen et al, 2017). Furthermore, the species that compose a community respond to the environment and to each other as an ensemble rather than separately (Tikhonov et al, 2020). Recognising these factors, JSDMs aim to consider information on the abundance or presence/absence of many species simultaneously and, at the community level, to incorporate the effects of environmental factors and interspecific interactions on species abundance or incidence (e.g.…”

“…The models were fitted in a Bayesian framework by running the Markov chain Monte Carlo (MCMC) estimation scheme and sampling from the posterior distribution of the model parameters (Dallas et al, 2019;Tikhonov et al, 2020). The posterior distribution was sampled with two MCMC chains with 100*thin iterations removed as burn-in, and 200*thin iterations remained for sampling with the thin equal to either 10 or 100.…”