The impact of gigafire on vegetation structure, terrestrial vertebrate abundance, and diel activity

Fire regimes are shifting around the world due to climate and land-use
change, resulting in an increased frequency of large and severe wildfires.
However, the impact of extreme wildfire events on animal species remains
poorly understood. Particularly lacking is an understanding of how fire
affects animal behaviour. By examining how distinct vertebrate groups
respond to wildfire, we capture variations in resilience mechanisms that
further our understanding of fire ecology and aid conservation strategies
in increasingly fire-prone landscapes. Across an area of 180,110 ha, we
investigated the effects of the 2019–2020 Australian ‘Black Summer’
wildfires on vegetation structure (a proxy for vertebrate habitat), the
number of independent detections (as an indicator of abundance), and diel
activity of a range of terrestrial animals in Eucalypt forests of
south-eastern Australia. We examined the influence of fire severity on
vegetation structure, using wildlife cameras, animal species abundance,
and how burnt or unburnt areas and fire severity affect species’ diel
activity patterns 20 months post-fire. Areas subject to high severity
fires (in comparison to low–moderate) had less canopy cover and leaf
litter, and higher vegetation cover at measured (0 to >2 m)
understory heights, as well as greater abundance of logs. Fire severity
had a limited effect on the abundance of animal species, with differences
in abundance observed in eight of 29 species. A greater impact occurred
following high severity fire, where 8/29 species were affected, compared
to 3/29 in areas burnt at lower severity. However, only 4/29 species (high
severity) and 1/29 (low severity) responded negatively, demonstrating a
general resilience among many species to fire. Areas that burnt at high
severity had higher introduced mammal richness and reduced native mammal
diversity, suggesting the potential for introduced species to establish in
severely burnt areas. The diel activity patterns in areas subject to fire
differed for seven of 17 species with sufficient data, with these species
concentrating their activity during specific times in burnt areas compared
to unburnt areas. Such behavioural plasticity may facilitate species
persistence in environments modified by fire by allowing species to
exploit different resources or minimise predation risk. Understanding how
fire affects animal species, including animal behaviour, will be critical
as the world’s fire regimes continue to change.

# The impact of gigafire on vegetation structure, terrestrial vertebrate
abundance, and diel activity Dataset DOI:
[10.5061/dryad.x0k6djhx5](10.5061/dryad.x0k6djhx5) ## Description of the
data and file structure ###
File: **data\_frame\_sites\_dryad.csv** **The impact of gigafire
on vegetation structure, terrestrial vertebrate abundance, and diel
activity** This dataset encompasses comprehensive surveys of animal
species across three wilderness areas affected by an unprecedented
gigafire. It includes data on species richness, species diversity, and
various environmental variables such as fire severity, vegetation types,
and historical fire events. The data were collected to analyse the
resilience of animal communities and ecosystem functions following
large-scale fire disturbances. **Description of the Data and File
Structure** **data_frame_sites_dryad**: This is the primary data frame
used for analysis, containing observational data across different sites.
**Variables in data_frame_sites_dryad** Below is a detailed description of
the variables included in the data_frame_sites_dryad data frame: **Site
identifiers and observation labels**
·         **site**: Unique identifier for each camera‐trap deployment site. ·         **location**: Name of the wilderness area (one of the three study regions) where the site is located. ·         **landscape**: Numeric code for the landscape unit within each study area. ·         **obs**: Label used as the observation‐level random effect in hierarchical models. **Deployment dates and camera details** ·         **start_date**: Date on which the camera was deployed (YYYY-MM-DD). ·         **collection_date**: Date on which the camera’s memory card was retrieved (YYYY-MM-DD). ·         **date_of_last_picture**: Date of the most recent image captured during that deployment (YYYY-MM-DD). ·         **deployment_days**: Total number of days the camera was active. ·         **camera_type**: Model of camera trap used at the site (e.g., Enduro Swift, Reconyx). **Site location and Environmental variables** ·         **latitude**: Latitude of the camera site in decimal degrees. ·         **longitude**: Longitude of the camera site in decimal degrees. ·         **topography**: Broad landform category at the site (e.g., ridge, valley). ·         **elevation**: Elevation at the site in metres above sea level. ·         **rainfall**: Mean annual rainfall for the site’s locale (millimetres). ·         **vegetation**: Dominant vegetation type at the site (e.g., shrubland, open forest). ·         **ndvi**: Mean Normalized Difference Vegetation Index over the deployment period, indicating vegetation greenness. ·         **terrain_ruggedness**: Terrain Ruggedness Index calculated for the area surrounding the site. **Fire metrics** ·         **fire_severity**: Categorical fire‐severity class assigned to the site (unburnt; low/moderate; high). ·         **fire_count**: Number of historical fires recorded within the site containing the site. **Biodiversity metrics** ·         **Total_richness**: Total number of species recorded by that camera across all periods. ·         **Total_native_mammal_richness**: Number of native mammal species detected at the site. ·         **Total_introduced_mammal_richness**: Number of introduced (non-native) mammal species detected. ·         **Shannon_diversity**: Shannon diversity index for all species at the site. ·         **Shannon_native_mammal**: Shannon diversity index calculated only for native mammal species. ·         **Shannon_introduced_mammal**: Shannon diversity index calculated only for introduced mammal species. **Species data** Individual species data show the sum of the number of 30-minute events at each site in each period. ### File: data\_frame\_vegetation\_dryad.csv **The impact of gigafire on vegetation structure, terrestrial vertebrate abundance, and diel activity** This dataset encompasses vegetation surveys across three wilderness areas affected by an unprecedented gigafire. It includes data on species richness, functional diversity, predator-prey interactions, and various environmental variables such as fire severity, vegetation types, and historical fire events. The data were collected to analyse the resilience of animal communities and ecosystem functions following large-scale fire disturbances. **Description of the Data and File Structure** **data_frame_vegetation_dryad**: This is the primary data frame used for analysis, containing vegetation data across different sites. **Variables in data_frame_vegetation_dryad** Below is a detailed description of the variables included in the data_frame_vegetation_dryad data frame: **Survey and site identifiers** * **location**: Study region (one of the three wilderness areas). * **survey number**: Sequential identifier for each survey event. * **landscape**: Numeric code for the landscape unit. * **site**: Unique code for each camera‐trap deployment. * **obs**: Observation‐level random‐effect label. **Temporal variables** * **date**: Date of the survey or detection (YYYY-MM-DD). * **time**: Time of the survey or detection (HH:MM:SS). **Spatial coordinates and positioning** * **landscape_position**: Position within the landscape (e.g., mid slope, flat, upper slope). * **latitude**: Site latitude (decimal degrees). * **longitude**: Site longitude (decimal degrees). **Environmental covariates** * **elevation**: Metres above sea level. * **rainfall**: Mean annual rainfall (mm). * **vegetation**: Dominant vegetation community at the site. * **ndvi**: Mean Normalized Difference Vegetation Index during the survey. * **terrain_ruggedness**: Terrain Ruggedness Index around the site. **Fire metrics** * **fire_severity**: Categorical class (unburnt; low/moderate; high). * **severity**: Continuous index of local fire impact. * **fire_count**: Number of historical fires in that site. **Habitat-structure metrics** * **canopy_species**: Dominant canopy species. * **Canopy cover average**: Mean percent canopy cover taken at three points (0m, 25m, 50m) along the vegetation transect. * **Canopy height average**: Mean canopy height (m) taken at three points (0m, 25m, 50m) along the vegetation transect. * **Log volume sum**: Total fallen-log volume (m³) taken from their height and width within 5m on either side of the vegetation transect. * **Log count**: Number of fallen logs within 5m on either side of the vegetation transect. * **large_rocks**: Count of large rocks within 5m on either side of the vegetation transect. * **burrow_count**: Count of wombat burrows within 5m on either side of the vegetation transect. **Vegetation and substrate metrics** **Ground cover** ·         **Bare ground**: Cover of exposed soil along the vegetation transect. **Understorey structure cover** ·         **Ground, Low**, **Medium**, **High**, **Very high vegetation cover**: Cover of understory vegetation within each height class along the vegetation transect. **Leaf-litter cover** ·         **Leaf litter**, **Low/Medium/High/Very high leaf litter**: Cover of litter layer and its density classes along the vegetation transect. **Substrate feature** ·         **Rock**: Cover of rocks along the vegetation transect. **Plant functional groups** ·         **Eucalyptus adult, Eucalyptus basal, Eucalyptus dead, Eucalyptus epicormic, Eucalyptus juvenile**: Cover of individuals in each life stage along the vegetation transect. ·         **Forb, Grass, Other tree, Rush, Sedge, Shrub**: Cover of each functional group along the vegetation transect. **Tree-size metrics (DBH classes)** * **DBH alive 5–20**, **20–50**, **51–80**, **81–120**, **121+**: Count of the live stems in each diameter‐at‐breast‐height (cm) class within 5m either side of the vegetation transect. * **DBH dead 5–20**, **20–50**, **51–80**, **81–120**, **121+**: Count of the standing dead stems in each DBH class within 5m on either side of the vegetation transect. **Note:** Variables ending in **A** (e.g., **Bare groundA**, **ForbA**) record the number of sampled grid cells in which that category was present; the corresponding variable without **A** gives the total count, area, or measurement for that category. ### **site\_images\_30min\_dryad.csv** **The impact of gigafire on vegetation structure, terrestrial vertebrate abundance, and diel activity** This dataset encompasses 30-minute event data of animal species across three wilderness areas affected by an unprecedented gigafire. It includes data on species activity and various environmental variables such as fire severity, vegetation types, and historical fire events. The data were collected to analyse the resilience of animal communities and ecosystem functions following large-scale fire disturbances. **Description of the Data and File Structure** **site_images_30min_dryad**: This is the primary data frame used for analysis, containing observational data across different sites. **Variables in site_images_30min_dryad** Below is a detailed description of the variables included in the site_images_30min_dryad data frame: **Event identifiers** * **event number**: Unique identifier for each detection event. **Site identifiers and location** * **site**: Unique code for each camera‐trap deployment site. * **location**: Name of the wilderness area (one of the three study regions) in which the site is situated. * **landscape**: Numeric code for the landscape unit within each study area. **Deployment dates and camera metadata** * **start_date**: Date on which the camera was deployed (YYYY-MM-DD). * **collection_date**: Date on which the camera’s memory card was retrieved (YYYY-MM-DD). * **data_of_last_picture**: Date of the most recent image captured during that deployment (YYYY-MM-DD). * **deployment_days**: Total number of days the camera was actively recording. * **camera_type**: Model of camera trap used at the site (e.g., Enduro Swift, Reconyx). **Site location and Environmental covariates** * **latitude**: Latitude of the camera site in decimal degrees. * **longitude**: Longitude of the camera site in decimal degrees. * **topography**: Broad landform category at the site (e.g., ridge, valley). - **elevation**: Elevation of the site in metres above sea level. - **rainfall**: Mean annual rainfall for the site’s locale (millimetres). - **vegetation**: Dominant vegetation type at the site (e.g., shrubland, open forest). - **ndvi**: Mean Normalised Difference Vegetation Index over the deployment period, indicating vegetation greenness. - **terrain_ruggedness**: Terrain Ruggedness Index calculated around the site. **Fire metrics** * **fire_severity**: Categorical fire‐severity class at the camera location (unburnt; low/moderate; high). * **fire_count**: Number of historical fires recorded within the site. **Detection details** * **common_name**: Common name of the species detected in the event. * **DateTimeOriginal**: Original timestamp of the image capture (YYYY-MM-DD HH:MM:SS). * **date**: Date of the detection event (YYYY-MM-DD). * **time**: Time of the detection event (HH:MM:SS). * **number_of_objects**: Number of objects (individuals) recorded in the detection event (e.g., number of animals in the frame). ## **Sharing/Access Information** The dataset can be accessed through the Dryad digital repository: [https://doi.org/10.5061/dryad.x0k6djhx5](https://doi.org/10.5061/dryad.x0k6djhx5)