Data and code from: Dispersing across habitat boundaries: uncovering the demographic fates of populations in unsuitable habitat
Data files
Feb 23, 2024 version files 18.63 MB
-
avena_fit.RData
-
avena.RData
-
datass.RData
-
new_ds.RData
-
new_ga.RData
-
README.md
-
Szojka_Ecosphere_2024.zip
Feb 23, 2024 version files 18.63 MB
-
avena_fit.RData
-
avena.RData
-
datass.RData
-
new_ds.RData
-
new_ga.RData
-
README.md
-
Szojka_Ecosphere_2024.zip
Abstract
Patchy landscapes are characterized by abrupt transitions among habitats, forcing species to cross habitat boundaries in order to spread. Although individuals that disperse into unsuitable habitat are often presumed dead in population models, their demographic fates may not be so simple. If survival within unsuitable habitat is possible (however transiently), then through time, individuals may be able to reach distant, suitable habitat, buffering species from extinction. In a fragmented Californian grassland, we explored the fates of individuals that crossed habitat boundaries, and if those fates differed among habitat specialists dispersing between two habitat types: serpentine habitat patches and the invaded non-serpentine matrix. We surveyed the diversity of seedbank and adult life stages along transects that crossed boundaries between patches and the matrix. First, we considered how patch specialists might transiently survive in the matrix via seed dormancy. We found that dormancy maintained populations of patch specialists deep into the matrix, as abundances of individuals in the seedbank and in the germinated adult communities increasingly differed with distance into the matrix; these patterns were not observed for matrix specialists in serpentine patches. Seeds of patch specialists that lacked morphologies for assisted dispersal accumulated downslope of patches, suggesting that even the most dispersal-limited species could eventually reach suitable patches even if they first land in the matrix. Second, we investigated dispersal of an invasive matrix specialist (Avena fatua) into patches to assess if sink populations existed across the habitat boundary. We found that A. fatua was largely absent deep in patches, where reproductive outputs plummeted and there was no evidence of a dormant seedbank (in contrast to a sizeable seedbank of patch specialists in the matrix). Our results not only reveal the demographic fates of individuals that land in unsuitable habitat, but also that their ecological consequences differ depending on the direction by which the boundary is crossed (patch ➝ matrix ≠ matrix ➝ patch). Dormancy is often understood as a mechanism for persisting in face of temporal variability, but it may serve as a means of traversing unsuitable habitat in patchy systems, warranting its consideration in estimates of habitat connectivity.
README: CODE AND DATA ASSOCIATED WITH MANUSCRIPT: "Dispersal across habitat boundaries: uncovering the demographic fates of populations in unsuitable habitat"
################################################################################
AUTHORS: Megan Szojka (ORCID ID: 0000-0002-8937-9074) and Rachel M. Germain (ORCID ID: 0000-0002-1270-6639)
################################################################################
OVERVIEW:
This project explored the fates of seeds that crossed habitat boundaries, and if those fates differed among habitat specialists dispersing between two habitat types: serpentine habitat patches and the invaded non-serpentine matrix. We surveyed the diversity of seedbank and adult life stages along transects that crossed boundaries between patches and the matrix. Our seedbank samples were collected in October 2018 from serpentine grasslands within McLaughlin Nature Reserve in Northern California (Dataset: new_ga.Rdata; Fig. 2a). Our surveys of the germinated community were done during peak biomass post-growing season (May 2019) in adjacent plots to those where we collected seedbank samples (Dataset: new_ds.Rdata; Fig. 2b). As such, the seedbank communities collected prior to the growing season can be directly compared to the germinated community post growing season (Dataset: datass.Rdata; Fig. 1).
The method of species dispersal may either promote or hinder species ability to cross habitat boundaries. As such we analysed whether species dispersal mode, either unassisted, animal-dispersed, or wind-dispersed, interacted with the direction of transects on a slope (either up slope, lateral, or down slope) to facilitate seed dispersal (Dataset: new_ga.Rdata and new_ds.Rdata; Fig. 3).
We were additionally interested in the effects of habitat boundaries on an invasive grass: Avena. We collected data from a subset of serpentine patches on Avena demography, including survival and seed production along transects extending from non-serpentine into serpentine habitat (Dataset: avena.Rdata and avena_fit.Rdata; Fig. 4).
CODE TO RUN PROJECT:
We ran all code in R version 3.6.1. All analyses and figures are housed in the script "Ecosphere_review - rmg.R, which loads in raw data in the form of R dataframes (.Rdata). This code requires R packages emmeans (version 1.8.5), tidyverse (version 2.0.0), car (version 3.1.2), lme4 (version 1.1.32), visreg (version 2.7.0), glmmTMB (version 1.1.7), scales (version 1.2.1), DHARMa (version 0.4.6), ggeffects (version 1.2.3), codyn (version 2.0.5), cowplot (version 1.1.1), grid (version 4.2.2), and magick (version 2.8.1).
Below is a description of variables within and purpose of each data file.
DESCRIPTION OF DATA FILES:
Variables that are shared between multiple data files are summarized here for brievity:
'Species' are coded by the first three letters of the genus and the first three letters of the specific epithet, based on species names found on https://www.calflora.org/.
'Status' summarized whether species were 'native', 'invasive', or 'naturalized'.
'Distance' was measured in meters, where 0 is edge/boundary between matrix and patch, negative numbers reflect distance towards center of patch, and positive numbers reflect distance into non-serpentine matrix.
For model fits, distance was transformed exponentially ('Dist.exp'), or logged ('Dist.log'),
'Treatment' included two levels: 'ds' refers to the surveys of germinated adults, or 'ga' refers to the surveys of seedbank individuals.
'Transect' was categorized by position on the landscape into four levels: 'lateral' means transect was horizontal with respect to the direction of landscape slope, 'up slope', 'down slope', or along an 'animal path'.
'Position' was categorized by plot position within a transect into five levels: 'c' = center of serpentine patch, 'p'= 1 meter into serpentine patch from edge, 'e' = edge/boundary between matrix and serpentine, 'm1' = 1 meter into non-serpentine matrix from edge, 'm5' = 5 meters into non-serpentine matrix from edge.
'Patch', 'T_ID', and 'Plot' are variables that assign a unique id to serpentine patches, transects, and survey plots, respectively.
'Disp.mode' reflects the dispersal mode that characterizes each species' seeds, and could be three options: 'U' = unassisted dispersal, 'V' = vertebrate / animal dispersed, or 'W' = wind dispersed.
Productivity represents the mean ndvi extracted for each serpentine patch, but was not used in analyses.
'slope.m' or 'Slope' was also extracted as the average slope each serpentine patch was on the landscape (change in elevation over the diameter of patch in meters), but was not used in analyses.
'Area' represents the meters squared of each serpentine patch.
'Bare' is a measure for each plot in the adult community surveys of percent cover that is bare ground.
avena
Includes the estimated 'Abundance' of Avena (avefat) within each plot. We estimated these numbers by multiplying the % cover by the average size of one avena individual.
avena_fit
From a subset of the patches we used to collect surveys on seedbank and germinated communities, we assessed the fitness of Avena via seed production. 'seed' is the recorded seed production of each Avena individual, and 'repl' is a measure of reproductive replacement, where 1 is subtracted from the seeds produced by each individual.
datass
This dataset combines the adult community and seedbank survey data to compare metrics between them.
'Richness' reflects the number of species found in each plot and each treatment.
'Abundance.plot' is a rescaled version of raw abundance (used in seedbank surveys) and cover estimates (used in adult surveys) ranging from 0 to 100. This metric is not used in our analyses.
'Intercept_rich' was calculated to estimate the richness between the seedbank and adult communities, considering that the we germinated only a subset of the seedbank sample collected. Specifically, we compared the mean difference of richness between the seedbank and adult communities within center plots, then added the difference (4.5 spp) to each seedbank plot. We did not use this metric in analyses.
'J' represents Pielous' J which is a measure of evenness within each community, and can range from 0 to 1.
We calculated Bray-Curtis dissimilarity, summarized by axis one ('Bray1'), and axis two ('Bray2').
new_ds
Survey of species in the field, representing the diversity and abundance of species that germinated from the seedbank. These surveys took place in plots adjacent to those that we used to collect seedbank material.
'Profile_ds' is the proportion of cover a species takes up within a plot, calculated by taking the percent cover of each species and dividing it by the total percent cover within a given plot.\
'Richness_ds' is the number of species found in each plot.\
'Cover_OG' is the percent cover each species takes up within each plot.\
'Rescaled_abundance_ds' is 'Cover_OG' rescaled between 0 and 100.
new_ga
Survey of species from germination assay, representing the diversity and abundance of species in the seedbank.
'Profile_ga' is the proportion of individuals that each species takes up, calculated by taking the number of individuals of each species within a plot divided by the total number of individuals within that plot.\
'Richness_ga' is the number of species found within each germination assay.\
'Abundance_OG' us the number of individuals found in each germination assay for each species.\
'Rescaled_abundance_ga' is 'Abundance_OG' rescaled between 0 and 100.
SHARING INFORMATION AND CONTACT:
contact author Megan Szojka megan.szojka@gmail.com with any questions regarding data usage.
dryad DOI: 10.5061/dryad.jsxksn0hj
##################################################################
Methods
See Methods of manuscript