Data from: Interacting livestock and fire may both threaten and increase viability of a fire-adapted Mediterranean carnivorous plant
Paniw, Maria, University of Cádiz
Quintana-Ascencio, Pedro F., University of Central Florida
Ojeda, Fernando, University of Sheffield
Salguero-Gomez, Roberto, University of Queensland, University of Sheffield
Published Jan 17, 2018 on Dryad.
Cite this dataset
Paniw, Maria; Quintana-Ascencio, Pedro F.; Ojeda, Fernando; Salguero-Gomez, Roberto (2018). Data from: Interacting livestock and fire may both threaten and increase viability of a fire-adapted Mediterranean carnivorous plant [Dataset]. Dryad. https://doi.org/10.5061/dryad.40qt2
1. Quantifying interactive effects of environmental drivers on population dynamics can be critical for a robust analysis of population viability. Fire regimes, among the most widespread disturbances driving population dynamics, are increasingly modified by and interact with human activities. However, viability of fire-adapted species is typically assessed overlooking disturbance interactions, potentially resulting in suboptimal management actions. 2. We investigated whether increasing human disturbances in fire-prone ecosystems may pose a threat or an opportunity to improve population viability, using demographic data of the carnivorous, post-fire recruiting plant Drosophyllum lusitanicum, endemic to heathlands in the southwestern Mediterranean Basin. We built integral projection models and simulated population dynamics under different combinations of two key disturbance types affecting populations: fire and livestock browsing and trampling. We used perturbation analyses to determine potential long-term consequences of maintaining fundamentally different disturbance types. 3. Despite most populations inhabiting browsed habitats, simulations showed a greater extinction risk in populations under high livestock pressure compared with ones under low or moderate pressures. Extinction risk decreased when fire return intervals shortened in populations under low or moderate livestock pressure; however, the opposite pattern emerged in heavily browsed populations, where short intervals between fires increased extinction.
4. Elasticity analyses showed that decreases in viability under frequent disturbance interactions (heavy browsing and frequent fire) may be explained by selection against seed dormancy in populations with frequent browsing and trampling. This may potentially cause populations to collapse when fires kill above-ground plants without populations being able to recover from a seed bank. 5. Synthesis and applications: Incorporating disturbance interactions can result in a different assessment of viability of a fire-adapted species than considering fire regimes alone. In Mediterranean ecosystems, fire management may be more effective when integrating moderate human activities. However, replacing fires by human disturbances, a currently widespread strategy in many fire-prone ecosystems, is not recommended since it may fundamentally alter population dynamics and selection pressures and decrease viability of fire-adapted species.
Overview of the R code provided in the manuscript
Here, we provide an overview of the R scripts and data files to accompany the main text, "Interacting livestock and fire may both threaten and increase viability of a fire-adapted Mediterranean carnivorous subshrub" and found in this depository. The .R files should be opened with an R editor (e.g., R Studio). The R code is fully commented.
Demographic transitions of Drosophyllum lusitanicum populations recorded in five annual censuses (from 2011to 2015) in eight populations differing in time-since-fire (TSF) and livestock browsing (LS) in their habitats. These data are used to quantify vital rates of above-ground individuals.
Seed fates (in a binary format) inferred from two experiments. These data are used to quantify the transitions related to the seedbank.
In case the reader wishes to forego the step of fitting the Bayesian models, which can be very time consuming (> 24 h on 3.40 GHz processor), we provided a mcmcOUT.csv file with 600 posterior parameter values for each of the parameters estimated with Bayesian models using uninformative priors.
Executes and saves the results of a Bayesian model quantifying all vital rates using uniformed priors; illustrates basic diagnostics that can be run on the results of an MCMC run (i.e., the posterior parameter distribution) to check for model convergence and autocorrelation of the posterior samples.
Demonstrates how to construct IPMs including continuous (above-ground) and discrete (seedbank) transitions for parameter means of the Bayesian models; saves IPMs for each combination of TSF (n =5) × LS (n = 2) × site (n = 8) (Part A). The code also saves IPMs for each TSF × LS averaged over sites (Part B). Site-specific and average kernel transitions for the vital rate growth (γ in main text) are also saved. The IPMs in Part A are used to calculate the stochastic population growth rate log λs, and extinction probability, Pq(t) by t = 300 years. The IPMs in parts A and B are used to calculate elasticities of log λs to changes in mean transitions across environmental states (TSF & LS) and in the standard deviation of transitions.
Runs and plots results of simulations of log λs using IPMs constructed for mean parameter values. The simulations consider fire as a stochastic process in which transitions between TSF states are based on fire return interval. Transitions between LS are not considered.
Runs and plots results of simulations of log λs using IPMs constructed for mean parameter values. Here, the growth kernel transitions are also included alongside the full IPMs. The simulations consider both fire and livestock browsing as an integrated stochastic process in which transitions between combinations of TSF × LS states are based on fire return interval and browsing management efforts.
Runs and plots results of simulations of log λs and elasticities using IPMs constructed for lower vital rate survival (σ). The simulations consider fire as a stochastic process in which transitions between TSF states are based on fire return interval. Transitions between LS are not considered.