Data and code from: Spatial storage effect facilitates evolutionary rescue in rapidly changing environments

Published Oct 04, 2025 on Dryad. https://doi.org/10.5061/dryad.fqz612k51

Data files

Oct 04, 2025 version files 2.13 MB

distr

316 B
outst

308 B
pars1

7 B
parst

12 B
PRE.c

14.59 KB
README.md

6.69 KB
README.txt

2.88 KB
tra.c

2.75 KB
trajectory

2.10 MB

Abstract

The storage effect is a plausible natural mechanism that generates balanced genetic polymorphism in temporally varying environments. Balanced polymorphism may facilitate evolutionary rescue, promoting the persistence of populations otherwise destined for extinction. However, it is unknown whether the storage effect can be established in small populations whose size is allowed to vary, and if so, whether it will lead to evolutionary rescue. In this study, we investigate whether the spatial storage effect emerges and facilitates evolutionary rescue across small populations of variable sizes that inhabit heterogeneous, temporally varying environments and exchange migrants. We use an eco-evolutionary model to examine the phenomenon under a wide set of conditions, including the magnitudes and periods of temporal variation, habitat harshness, migration rates, the degrees of spatial heterogeneity, and increasing fitness oscillations over time, all within the framework of the logistic population growth model. We find that the storage effect emerges and that it increases the persistence of populations in harsh, temporally varying habitats beyond levels expected in the absence of the mechanism. This mechanism demonstrates how rapid evolution broadens the known conditions for population persistence in the face of rapid and continuous environmental changes.

Dataset DOI: 10.5061/dryad.fqz612k51

Description of the data and file structure

This package contains C programs used to generate the selection coefficients to mimic varying selection, tra.c, (developed by Gulisija and Kim 2015) and forward-in-time eco-evolutionary dynamics involving the spatial storage effect in temporally varying habitats (PRE.c), used in Rowland and Gulisija. 2025. Spatial Storage Effect Facilitates Evolutionary Rescue in Rapidly Changing Environments.

To run, please compile using:

gcc -Wall 'exe_file' -o 'program.c' -lm , or

cc 'program.c' -lm -o 'exe_file'

Brief descriptions of the programs and their input files follow:

tra.c

This program generates time-dependent selection coefficients (st). The input file 'parst' provides the total number of coefficients to be produced (TIME = 3 periods), period = P, and the smax. The output file is a column of values of st. The output file name is given as trajectory.PERIOD.WIDTH but the name should be changed to trajectory to be read by PRE.c.

PRE.c

This program simulates the eco-evolutionary dynamics. It requires three input files:

"trajectory" - contains three cycles of the "st", generated by tra.c.

"pars1" - contains a single number corresponding to the migration rate. In the sample 'pars1' the migration rate is 0.0125, which corresponds to 50 migrants at the carrying capacity of 4000.

"distr" - contains the counts of frequencies of derived allele out of 16000 simulation runs. From this distribution the populations are initiated, randomly, at the onset of selection. The sample distribution is generated under parameters and drift as reported in the manuscript (see Figure S1). This distribution can be preset (with total counts summing up to 16000) or simulated using PRE.c until equilibrium is confirmed with 16000 replicates and tallying up the occurrences (frequencies) for allele frequency.

Based on the parameters provided by the input files and those included in PRE.c, the system will eco-evolve until either the maximum simulation time of 100K (400000 generations) is reached or the population goes extinct (population size in V falls below 100 individuals, critical value based on Gomulkiewicz & Holt, 1995).

The output file (outst) reports: P (number of generations per cycle of selection), smax (maximum numerator fitness deviation due to varying selection), Migrants, mu (recurrent mutation rate), G RATE V (growth rate in V), G Rate R (growth rate in R), REPS (number of replicate populations), MaxTime (maximum simulation time), cumuHL (cumulative heterozygosity for metapopulation), HLpB (cumulative heterozygosity for V), timeEqu (number of replicate populations that persist to max time), polymorphs (number of polymorphic populations at max time), aveNtotal (average metapopulation size), aveNV (average population size in V), minExt (average survival time), extincts (proportion of replicate populations that went extinct)

Files and variables

File: distr

Description: Input file for PRE.c that contains the counts of frequencies of derived allele out of 16000 simulation runs. From this distribution the populations are initiated, randomly, at the onset of selection. The sample distribution is generated under parameters and drift as reported in the manuscript (see Figure S1). This distribution can be preset (with total counts summing up to 16000) or simulated using PRE.c until equilibrium is confirmed with 16000 replicates and tallying up the occurrences (frequencies) for allele frequency.

File: outst

Description: Output file for PRE.c that reports P (number of generations per cycle of selection), smax (maximum numerator fitness deviation due to varying selection), Migrants, mu (recurrent mutation rate), G RATE V (growth rate in V), G Rate R (growth rate in R), REPS (number of replicate populations), MaxTime (maximum simulation time), cumuHL (cumulative heterozygosity for metapopulation), HLpB (cumulative heterozygosity for V), timeEqu (number of replicate populations that persist to max time), polymorphs (number of polymorphic populations at max time), aveNtotal (average metapopulation size), aveNV (average population size in V), minExt (average survival time), extincts (proportion of replicate populations that went extinct)

File: parst

Description: Input file for tra.c that provides the total number of coefficients to be produced (TIME = 3 periods), period = P, and the smax. The output file is a column of values of st. The output file name is given as trajectory.PERIOD.WIDTH but the name should be changed to trajectory to be read by PRE.c.

File: pars1

Description: Input file for PRE.c that contains a single number corresponding to the migration rate. In the sample 'pars1' the migration rate is 0.0125, which corresponds to 50 migrants at the carrying capacity of 4000.

File: PRE.c

Description: This program simulates the eco-evolutionary dynamics and is the main code used in the study.

File: README.txt

Description: describes the package of programs used to generate the selection coefficients to mimic varying selection, tra.c, (developed by Gulisija and Kim 2015) and forward-in-time eco-evolutionary dynamics involving the spatial storage effect in temporally varying habitats (PRE.c), used in Rowland and Gulisija. 2025. Spatial Storage Effect Facilitates Evolutionary Rescue in Rapidly Changing Environments.

File: tra.c

Description: This program generates time-dependent selection coefficients (st). The input file 'parst' provides the total number of coefficients to be produced (TIME = 3 periods), period = P, and the smax. The output file is a column of values of st. The output file name is given as trajectory.PERIOD.WIDTH but the name should be changed to trajectory to be read by PRE.c.

File: trajectory

Description: Input file for PRE.c that contains three cycles of the "st", generated by tra.c.

Code/software