Skip to main content

Data from: Do Tasmanian devil declines impact ecosystem function?

Cite this dataset

Stephenson, Torrey (2024). Data from: Do Tasmanian devil declines impact ecosystem function? [Dataset]. Dryad.


Tasmanian eucalypt forests are among the most carbon-dense in the world, but projected changes in climate could destabilize this critical carbon sink. While the impact of abiotic factors on forest ecosystem carbon dynamics have received considerable attention, biotic factors, such as the input of animal scat, are less understood. Tasmanian devils (Sarcophilus harrisii)—an osteophageous scavenger that can ingest and solubilize nutrients locked in bone material—may subsidize plant and microbial productivity by concentrating bioavailable nutrients (e.g., nitrogen and phosphorus) in scat latrines. Dramatic declines in devil population densities are driven by the spread of a transmissible cancer and may have underappreciated consequences for soil organic carbon (SOC) storage and forest productivity by altering nutrient cycling. Here, we fuse experimental data and modeling to quantify and predict future changes to forest productivity and SOC under various climate and scat-quality futures. We find that devil scat significantly increases concentrations of nitrogen, ammonium, phosphorus, and phosphate in the soil, and shifts soil microbial communities towards those dominated by r-selected (e.g., fast-growing) phyla. Further, under simulated increases in temperature and precipitation, devil scat inputs are projected to increase above- and belowground net primary productivity and microbial biomass carbon through 2100. In contrast, when devil scat is replaced by lower-quality scat (e.g., from non-osteophageous scavengers and herbivores), forest carbon pools either increase more slowly or decline. Together, our results suggest biotic factors will interact with climate change to drive current and future carbon pool dynamics in Tasmanian forests.

README: Do Tasmanian devil declines impact ecosystem function?

This dataset contains .lis DAYCENT output files for all present day to 2100 (future) model runs (3 sites * 3 climate scenarios * 2 scat input qualities = 18 output files), which were used to generate figure 4. Additionally, there are 3 .lis DAYCENT output files which represent historical land use ("hlu"), which were used to spin up DAYCENT models to the present day.
This dataset also contains soil nutrient and microbial community data used to generate figures 2 and 3.

Description of the data and file structure

Soil chemistry data columns are as follows:

Columns A:E = sample ID information including treatment, replicate, and site. For this manuscript, site = "Goulds Country" corresponds to "Eastern", site = "Arthur River" corresponds to "Western", and site = "Takone" corresponds to "Central."

Columns F:M = soil chemical properties. units, where applicable, are provided in column headers.

Columns N:AD = soil biological properties. Relative abundance of bacterial phyla/fungal classes as well as total DNA biomass and r:K ratio.
Note: "NA" values in columns W:AD indicate that the sample in question was not used for 16S community analysis. In this dataset, Arthur River replicate 6 (both treatment and paired control) was excluded from analysis due to low read quality of 16S sequencing data. 

Columns AE:AI = soil chemical properties from ICP analysis--total P, Ca, Fe, Mg, and Al
Note: "NA" values in columns AE:AI indicate samples that were not extracted for ICP analysis due to experimental constraints. ICP data were collected for 3 of the 6 total replicates collected at each site.


Additional code and data can be found at:


National Science Foundation, Award: DEB-2054716