Data from: The effectiveness of seed rain in recolonising an ecotonal mesic forest following extreme severity fire
Data files
Mar 03, 2026 version files 197.05 KB
-
2024_distance_dispersal_surveys.csv
35.12 KB
-
README.md
4.65 KB
-
SeedRain.csv
157.29 KB
Abstract
As fire severity and extent increase, recovery of fire-sensitive species in plant communities may rely increasingly on seed rain and dispersal from adjacent unburnt areas. The role dispersal plays in driving recovery trajectories is surprisingly understudied. Additionally, how traits drive dispersal can contextualise compositional differences and indicate recovery trajectories. This paper investigated seed rain in recently burned mesic forest and the role of traits in driving post-fire recovery across a gradient of fire severities. We placed seed traps at two locations in the Blue Mountains, located in south-eastern Australia. We collected and identified all seeds that fell into seed traps each month for a year, three years following fire of differing severities. We compared species diversity of seed collections with extant vegetation and measured the minimum dispersal distance each species likely travelled. We used variable selection to identify how traits impacted dispersal distances. Rainforest species were generally lacking from recently burned sites in both the extant vegetation and the seed rain. Composition of the seed rain indicated dispersal occurred primarily from local sources, with minimum distance travelled largely determined by life form, dispersal type and dispersal height. Differences in species richness and composition between unburnt and sites burnt at differing severities was observed in both the extant vegetation and in the seed rain. Increasing fire severity suggests lengthening recovery times, with recolonisation from unburnt areas minimal and slow. The commonality of local dispersal highlights the importance of surviving individuals for maintaining diversity and replenishing populations.
Dataset DOI: 10.5061/dryad.cjsxksnjw
Preface
The following README provides information regarding two csv files containing the raw data for the manuscript ‘The effectiveness of seed rain in recolonising an ecotonal mesic forest following extreme severity fire’.
Files and variables
Dataset Title
SeedRain.csv
Overview
This data set contains all raw data for seeds recovered from seed traps and related trait groupings for the manuscript ‘The effectiveness of seed rain in recolonising an ecotonal mesic forest following extreme severity fire’. The data is in the format of a csv.
Table Columns
Site – Mt Wilson or Mt Tomah
SeverityGroup – Severity as described in the methods of the paper, (Unburnt, Low/Moderate, High/Extreme)
SeedTrapline – Row of seed traps, unique values based on site, severity, and distance from fire edge.
DistanceEdge – Distance from fire edge, numeric value, positive values indicate distance into burnt area, negative values indicate distance outside of burnt area (unburnt).
TrapCount – Unique identifier for trap per site, severity and distance.
DateInitial – Indicates earliest date seed can have dispersed into trap
DateCollected – indicates date seed collected
Family – Species family. NA indicates no associated species with seed trap for time point.
Species – Scientific species name. NA indicates no associated species with seed trap for time point.
DispersalType – Common dispersal type of species. Categorical: Adhesion, Ant, Ballistic, Bird or wind, Gravity, Vertebrate, Wind. NA indicates no species. Vertebrate indicates seed types likely dispersed by animals, including mammals and birds, usually possessing a fleshy fruit. Bird or wind referred to seed types likely dispersed by either, usually not possessing a fleshy fruit. NA indicates no associated species with seed trap for time point.
HeightRange – Average plant height in metres grouped into categories: 0.5 to 1.5, 0.1 to 0.5, 0.5 to 3, 0.5 to 10, 10 to 25. NA indicates no associated species with seed trap for time point. ‘Unknown’ indicates species was unknown, and value could not be determined.
Form – life form of species. Categories: Climber, Grass, Herb, Sedge, Shrub, Tree. NA indicates no associated species with seed trap for time point.
SeedSize – Average weight of seed per species as described in the methods of the paper. Numeric value in mg. NA indicates no associated species with seed trap for time point.
Count – number of seeds per species for time period in trap. NA indicates no associated species with seed trap for time point.
MinimumDist - Distance of closest mature adult to trap for each species. NA indicates no associated species with seed trap for time point. ‘Unknown’ indicates species was unknown, and value could not be determined, and ‘Not found’ indicates closest mature adult could not be located.
SeedTrapCoverT1 – cover over traps as described in Supplementary material 2 methods measured 3 years post fire.
SeedTrapCoverT2 - cover over traps as described in Supplementary material 2 methods measured 3 years 8 months post fire.
Dataset Title
2024_distance_dispersal_surveys.csv
Overview
This data set contains all raw data for vegetation surveys conducted in 2024 to determine dispersal distances of species found in seed traps. This data is associated with the manuscript ‘The effectiveness of seed rain in recolonising an ecotonal mesic forest following extreme severity fire’. The data is in the format of a csv.
Table Columns
Date – Date plant surveys conducted
Site – Mt Wilson or Mt Tomah
Species – Scientific species name
SeedTrapline – Row of seed traps, unique values based on site, severity, and distance from fire edge. NA values indicate no related line of seed traps.
SurveySiteCode - unique values for surveys based on site, severity, and distance from fire edge or transect number.
DistanceEdge – Distance from fire edge, numeric value, positive values indicate distance into burnt area, negative values indicate distance outside of burnt area (unburnt).
SeverityGroup - Severity as described in the methods of the paper, (Unburnt, Low/Moderate, High/Extreme)
Code/software
Software
RStudio version 4.3.0 was used for all data analysis (RStudio Team, 2023).
References
NSW Government 2023. BioNet Atlas. NSW Government.
RStudio Team. 2023. *RStudio: Integrated development for R *[Online]. Boston, MA: RStudio, Inc. Available: http://www.rstudio.com/ [Accessed 22 March 2020].