Using theory-driven Integrated Population Models to evaluate competitive outcomes in stage-structured systems

Matthieu Paquet, Frédéric Barraquand · 2025

Read Using theory-driven Integrated Population Models to evaluate competitive outcomes in stage-structured systems by Matthieu Paquet, Frédéric Barraquand (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-06725.txt.

Introduction

Matthieu Paquet and Frédéric Barraquand built a merged model in 2025 to call who wins when two species vie for the same ground. Each species has life stages, and young and old may vie in unlike ways. Field data on counts, marked returns, and births are each thin on their own. Their model pours all three into one frame, with a math core that gives clean rules for when each species can invade the other's stand.

A Bayesian fit carries all doubt through to those calls. When true coexistence was built in, the model called it right. Strengths of the clash were not always pinned down, though. A forced wipe-out was hard to spot, yet the model at least flagged that case as unsure.

Catalog record

CREATORMatthieu Paquet, Frédéric Barraquand
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE47 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-06725

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Using theory-driven Integrated Population Models to evaluate competitive outcomes in stage-structured systems Matthieu Paquet, Frédéric Barraquand Abstract Predicting competitive outcomes typically requires fitting dynamical models to data, from which interaction strengths and coexistence indicators such as invasion criteria can be produced. Methods that allow to propagate parameter uncertainty are particularly indicated. These should ideally allow for competition between and within species at various life-stages, and make the best out of multiple data sources, each of which can be relatively scarce by statistical standards. Here, we embed a mathematical model of stage-structured competition between two species, producing analytical invasion criteria, into a two-species Integrated Population Model. The community-level IPM allows to combine counts, capture-recapture, and fecundity data into a single statistical framework, and the Bayesian formulation of the IPM fully propagates parameter uncertainty into invasion criteria. Model fitting demonstrates that we can correctly predict coexistence through reciprocal invasion when present, but that interaction strengths are not always estimable, depending on the prior chosen. Our competitive exclusion scenario is shown to be harder to identify, although our model allows to at least flag this scenario as uncertain rather than mistakenly present it as coexistence. Our results confirm the importance of accounting for uncertainty in the prediction of competitive outcomes. Keywords : competition; coexistence; invasion analysis; Integrated Population Model Correspondence to [email protected] 1 Introduction Two-species coexistence is typically predicted by Lotka–Volterra models when intraspecific competition is stronger than interspecific competition ( Chesson, 2000 ; Letten et al., 2017 ) . However, any deviation from the unstructured two-species Lotka–Volterra framework, such as adding more species ( Barabás et al., 2016 ) or more ages / stages ( Kohyama, 1992 ; Moll & Brown, 2008 ; Fujiwara et al., 2011 ) can make the mapping between species interaction strengths and competitive outcomes (coexistence, exclusion, priority effects) much less straightforward. In stage- or age-structured populations, inferring coexistence therefore requires further investigations by calculating invasion criteria, or running simulations ( Moll & Brown, 2008 ; Fujiwara et al., 2011 ) . Yet, it is of particular interest to investigate coexistence in such stage-structured systems, as in most communities some age- or stage-structure will be visible and potentially mediate the susceptibility of individuals to competition, be those communities made of trees, birds, fishes or insects. Theoretical models have actually shown that the presence of age structure itself could contribute to species coexistence ( Moll & Brown, 2008 ; Fujiwara et al., 2011 ) , at least in systems with few, similar, strongly competing species ( Bardon & Barraquand, 2023 ) . While these advances in modelling competing stage structured populations are shedding light on our understanding of conditions promoting species coexistence, whether and under which circumstances these models’ parameters could be estimated using real data remains unknown. As dynamic population models with interactions are parameter-rich, there are expected challenges associated with fitting such models to data and estimate species interactions and coexistence ( Paniw et al., 2023 ) . One of them being that some parameters such as species interaction coefficients can be hard to estimate from population count data alone ( Barraquand & Gimenez, 2019 ) . Combining several datasets into integrated models can help estimating parameters that would not be identifiable if datasets were analysed separately ( Cole & McCrea, 2016 ) , including for estimating species interactions ( Barraquand & Gimenez, 2019 ) . Stage-structured multispecies integrated population models (IPMs) have been developed to estimate species interactions ( Péron & Koons, 2012 ; Barraquand & Gimenez, 2019 ; Quéroué et al., 2021 ; Paquet & Barraquand, 2023 ; Viollat et al., 2024 ) . However, simulation-based evaluations have been performed in a trophic interaction set-up ( Barraquand & Gimenez, 2019 ; Paquet & Barraquand, 2023 ) and whether these models provide reliable estimates of competitive interactions is still poorly known, particularly in cases where both intra- and interspecific interactions act on the same vital rates. Moreover, they rely on phenomenological functions (using log and logit links), which is very convenient from a statistical perspective but has some drawbacks for ecological interpretation. Indeed, stage-structured models with log and logit links in vital rates ( Moll & Brown, 2008 ; Péron & Koons, 2012 ) do not allow for immediate assessment of competitive outcomes from interaction strengths, and analytically tractable invasion criteria are not available either, which means that coexistence can only be assessed by simulating the fitted models. More theory-driven competitive functional forms, and resulting population models, could be used to infer populations’ coexistence. This should be particularly interesting in models for which invasion criteria have been obtained analytically ( Fujiwara et al., 2011 ; Bardon & Barraquand, 2023 ) . Finally, when modelled in a Bayesian framework, such two-species theory-driven IPMs could allow to estimate invasion criteria (hence species coexistence or exclusion) together with their uncertainty, potentially even when all model parameters are not identifiable individually ( Kao & Eisenberg, 2018 , by analogy to epidemiological models,) . This need to correctly propagate uncertainty has been recently reflected in the plant coexistence literature ( Bowler et al., 2022 ; Armitage, 2024 ) . We therefore draw here on the development of such stochastic multispecies IPMs in animal ecology to improve the evaluation of competitive outcomes, such as coexistence or exclusion in plant and animal ecology alike. In the following, we propose a two-species Integrated Population Model—a dynamic, demographically structured model relying on multiple data sources—to estimate competitive interaction coefficients and evaluate coexistence or exclusion of two (potentially) competing stage-structured populations. In this model, adults can affect both adults and juveniles of either species, so that both fecundities and juvenile survival rates can depend upon adult densities of both species. Unlike previous multispecies IPMs relying on Ricker-based functional forms (log link functions), this model encapsulates a Beverton-Holt two-species juvenile-adult population model, for which analytical expressions of invasion criteria have been obtained ( Bardon & Barraquand, 2023 ) . Using four sets of parameters (i.e., three contrasting sets leading to coexistence through different pathways, and one leading to one species excluding the other), we assess the identifiability and accuracy of model parameters estimates, including species interactions, as well as our ability to estimate the mathematically-derived invasion criteria. 2 Methods 2.1 Description of the two-species two-stages stochastic model The model is based on Bardon & Barraquand (2023) and their stage-structured difference equation is then embedded in a stochastic framework to account for demographic stochasticity (similarly to Barraquand & Gimenez 2019 ). For each species we used a post-breeding census model where we have the expected number of breeders at each time step: n_{a}(t+1)=n_{a}^{\text{old}}(t+1)+n_{a}^{\text{new}}(t+1) (1) where n_{a}^{\text{old}}(t+1) represents the number of adults (breeders) that were already adults at the previous time step, and n_{a}^{\text{new}}(t+1) represent the number adults that were juveniles at the previous time step. We model demographic stochasticity using Binomial and Poisson distributions, so that n_{a}^{\text{old}}(t+1)\sim\text{Binomial}\left(s_{a},n_{a}(t)\right), (2) with s_{a} the adult yearly survival, and n_{a}^{\text{new}}(t+1)\sim\text{Binomial}\left(s_{j}(t),n_{j}(t)\right) (3) with s_{j}(t) the juvenile yearly survival (that is time dependent, see below), and n_{j}(t) the number of juveniles at time t. Finally for the number of juveniles we have n_{j}(t+1)\sim\text{Poisson}\left(f(t)n_{a}(t)\right), (4) with f(t) representing fecundity (the number of offspring of the modelled sex at each time step). The population model used

Read the complete text