Data from: Tropical biome switching: Ant communities transition from savanna to rainforest following cessation of burning
Data files
Oct 09, 2025 version files 114.72 KB
-
IR_ants_2022.csv
111.12 KB
-
README.md
3.60 KB
Abstract
Savannas and rainforests co-occur as mosaics across large areas of the tropics. These mosaics are dynamic, with savanna and rainforest switching in relatively short-time scales, largely through the effects of fire. Most research to date focuses on assessing how this biome switching affects vegetation, with little attention given to faunal responses to transitions between savannas and rainforests.
Aim: To assess faunal responses to biome switching from savanna to rainforest following 50 years of fire cessation.
Location: Iron Range (Kutini-Payamu) on Australia's Cape York Peninsula.
Taxon: Subterranean, leaf litter, epigaeic, and arboreal ant assemblages.
Methods: We sampled ant assemblages using pitfall, subterranean and arboreal traps, and leaf litter extractions across different vegetation types: savanna, secondary forest, primary dry forest and primary wet forest.
Results: We find a clear switch from savanna to rainforest ant communities and that the ant communities in secondary forest are very similar to that in primary forest. This is despite low similarity in floristic composition due to a poor representation of primary rainforest trees. However, the secondary forest is still missing a range of specialist forest taxa, especially cryptobiotic species from the soil and leaf litter, and possibly also specialist canopy species.
Main conclusions: The current 50 yr-old secondary forest has much of the ant biodiversity value of primary rainforest, including for regionally endemic species that are rainforest specialists. This indicates that ant communities respond more strongly to vegetation structure than to plant species composition. However, some specialist taxa are not in secondary forests, and it is unclear under which circumstances they may reappear.
https://doi.org/10.5061/dryad.mcvdnck9r
Description of the data and file structure
These data (IR_ants_2022.csv) result from sampling at the Iron Range Research Station (IRRS) in northern Cape York Peninsula, Queensland, Australia, done in June 2022. The sampling targeted subterranean, leaf litter, epigaeic, and arboreal ant assemblages in different vegetation types: savanna, secondary regrowth forest, primary dry forest, and primary wet forest. In total, we sampled 9 plots, deploying 180 pitfall traps, 180 arboreal traps, 360 subterranean traps, and undertook 28 Winkler leaf litter extractions.
Files and variables
File: IR_ants_2022.csv
Description: Raw data used for all analyses.
Variables and description
| IRRS_name | Site name as used at the Iron Range Research Station (IRRS) |
|---|---|
| site_name | Site name used in the study |
| habitat | vegetation type as either savanna, secondary regrowth forest, primary dry forest, or primary wet forest |
| forest_savanna | vegetation type as either savanna or forest (secondary regrowth, primary dry or primary wet) |
| ll_vol | The volume (ml) of leaf litter collected in the winkler extraction for that transect |
| sampling_date | The date of sampling: day (numerical), month (Roman), year (numerical) |
| method | The sampling method, which is also equivalent to the assemblage stratum (i.e., pitfall = epigaeic, arboreal trap = arboreal, subterranean trap = subterranean, winkler = leaf litter) |
| transect_nb | The number associated with a transect within a plot. |
| trap_nb | The number associated with a trap within a plot for one method |
| species | Name of species |
| abundance | The number of ants of one species in a particular trap, capped at 50 |
Note: 'NA' means not available or missing data. For instance, if a trap was missing, we entered the value 'NA' in the species and abundance columns.
Code/software
The only software needed to analyze our data and run our code is R.
This dataset consists of data on subterranean, leaf litter, epigaeic and arboreal ant assemblages. It was collected using pitfall, subterranean and arboreal traps, and leaf litter extractions. These were placed in 9 sites in the Iron Range of the northern Cape York peninsula of Australia. The vegetation of these sites consisted of savanna, secondary forest, dry primary forest and wet primary forest.