Fighting the flames: site-specific effects determine species richness of Australian frogs after fire

Introduction

Fire has regulated patterns of biodiversity for millions of years, but recent anthropogenic disturbances are altering fire impacts on the persistence of biodiversity (Kelly et al. 2020). Anthropogenically mediated climate change has increased fire severity and duration and geographic scale across the globe (Moritz et al. 2012; Jolly et al. 2015), exemplified by recent fires in western USA and south-eastern Australia (Nolan et al. 2020; Ward et al. 2020). Changes in fire regime are driven in part by: climatic variables, extending periods of dry, hot and windy weather; vegetation changes due to unprecedented precipitation periods; and increased likelihood of ignitions (Williams et al. 2009; Moritz et al. 2012; Kelly et al. 2020).

Many species have adapted to specific fire regimes, so disruption can impact both populations and entire ecosystems (Kelly et al. 2020). As fire increases in severity, frequency and scale under climate change, it can directly increase mortality rates, reduce available resources and habitat and fragment populations (Ward et al. 2020), increasing extinction rates of biodiversity (Gill and Bradstock 1995). Despite this potentially deleterious outcome, responses to fire are poorly documented, particularly for understudied taxa like amphibians, invertebrates and reptiles (Dale et al. 2001; Pastro et al. 2014; dos Anjos et al. 2021).

Frogs are already experiencing population declines globally due to various anthropogenic threats (Houlahan et al. 2000; Campbell Grant et al. 2020). They are likely to be particularly susceptible to high mortality rates or population decline with fire, given their low vagility and reduced ability to flee fire fronts, reliance upon water sources and specific habitat requirements that may be degraded post-fire (Rowley et al. 2020). They may also be increasingly vulnerable to extinction due to increased inbreeding, lower population numbers and low immigration rates with fires (Potvin et al. 2017). The potential vulnerability of frogs to fire, and other key threats such as chytridiomycosis (Berger et al. 2016), make studying the response of frogs to fire critically important.

Most of our understanding of how frogs respond to fire is biased towards North American species (dos Anjos et al. 2021), with few investigations in Australia by comparison. However, effects on Australian species vary. Some studies show frog species are not obviously affected by fire (Bamford 1992; Lowe et al. 2013; Rowley et al. 2020), whereas others show some species experiencing population declines. For example, Litoria littlejohni (Daly and Craven 2007) and Geocrinia lutea (Driscoll and Roberts 1997) experienced declines, and one study even documented a local population extinction (Lemckert 2000). Despite the contributions of finer-scale or localised studies elucidating the effects of fire on Australian frogs, we lack evidence across larger spatial scales or for mega-fires, such as those experienced in 2019/2020, and those that are predicted to increase under future climate change (Lee et al. 2022). Further, such conflicting evidence for a few species limits our ability to extrapolate the impact of fires from previous research.

Fires can be unpredictable in space and time and consequently, assembling resources to study their effects on short notice can be difficult. With the rise of citizen science, it is possible to investigate impacts of fire over large spatial scales and on a time-sensitive basis (Kirchhoff et al. 2021). Indeed, citizen science was instrumental in providing preliminary data on short-term persistence of frogs across large spatial scales after the ‘Black Summer’ bushfires in Australia, with at least 45 out of 66 species detected, and all 33 species with more than five calling records pre-fire recorded calling post-fire (Rowley et al. 2020). However, this dataset was only short term, failed to sample range-restricted and rainforest species and focused on summer-breeding. This is particularly relevant because although some individuals may survive the immediate impact of a fire, they are persisting in a burnt landscape with fewer resources and potentially experiencing starvation, as well as an increased risk of predation (Nimmo et al. 2019; Dickman and McDonald 2020). Additionally, the impact of varied fire severity on frogs was not assessed.

In this study, we used large-scale citizen science data, coupled with two remotely sensed datasets from the extensive and high-severity 2019/20 Australian bushfire season to examine: (1) persistence of frog species, measured by species richness, 18 months post-fire across Australia; and (2) effects of varying fire severity on frog species richness and recovery in eastern Australia, where megafires were concentrated. More specifically, we hypothesised that fire would have a negative effect on species richness, particularly at higher severity levels.

Materials and methods

Results

Overall, 36 well sampled species (≥5 records before fire) called in the short-term after fire, and all 49 species previously recorded with FrogID in FESM burnt areas called long-term post-fire. These species were taxonomically diverse, representing three of five native frog families in Australia (Fig. 1). Frogs responded similarly in the NIAFI/DEA Hotspot data (Fig. S1). About 58% of the frog species ever detected in FESM burnt areas by FrogID (including those with <5 records) before the fire were calling after 3 months, increasing to 72% after 6 months, 92% after 12 months and 98% after 18 months, until all were calling by around 2 years post-fire (Fig. S2). Frogs with the highest number of records post-fire were of low conservation concern (categorised as ‘least concern’ by the IUCN). Seven threatened species, Crinia tinnula (VU), Litoria subglandulosa (VU), Litoria olongburensis (VU), Philoria pughi (EN), Philoria sphagnicola (EN), Mixophys iteratus (EN) and Litoria watsoni (EN), and one near-threatened species, Adelotus brevis, called long-term after fire but had low numbers of records (Fig. 1).

Fig. 1.

Well sampled frog species with high numbers of FrogID submissions (n ≥ 5) calling before fire (blue), short-term after fire (yellow) and longer-term after fire (orange) using New South Wales Fire Extent and Severity Mapping (FESM) data. All species were found to be calling in the long-term. Red dots indicate threatened and near-threatened species. Horizontal axis is truncated due to high numbers of records for some species. N = unique number of sites.

Frogs from all ecological groups (terrestrial, ephemeral pond, permanent pond, stream associated, terrestrial and bog/soak) were present after fire, with comparably proportional records between groups for before, short- and longer-term after fire (Fig. S3). However, frog species associated with permanent waterbodies were the best represented in the dataset, and there were low number of records for frogs associated with bogs/soaks and those that relied upon both terrestrial and ephemeral ponds (Fig. S3). Frogs that had a low (0–20%) dependence on fire-sensitive vegetation were best represented in our data, and those with a high dependence were present post-fire but had a low number of records (Fig. S4). The most common frog species in our dataset were those that found refuge under debris during fire; however, frogs from all refugia groupings were present post-fire, albeit in lower numbers (Fig. S5). Mean species richness did not significantly change post-fire compared with pre-fire (estimate = −0.130, z = −0.986, P = 0.324; Fig. 2). Instead, changes in species richness were site specific, increasing for nine sites, remaining unchanged for four sites and decreasing for six sites post-fire (Fig. 3; Table S1). Although species richness was, on average, higher in low-severity burn areas, species richness did not significantly differ with varying fire severity level, even after accounting for vegetation type and sample size (P > 0.05; Table 1, Fig. 4). Similarly for DEA Hotspot/NIAFI data, changes in species richness were site-specific (Table S2). Specific changes to species and certain ecological groups varied across severity levels (Tables S3–S7).

Fig. 2.

(a) Overall species richness for bushfire-affected areas before and after fire for DEA Hotspot/NIAFI datasets encompassing New South Wales and Victoria. (b) FESM fire severity dataset encompassing only New South Wales. Slope of lines represent direction of change in species richness for each sampling site (either increase, decrease, or no change).

Fig. 3.

Change in species richness per DEA/NIAFI site in southeast Australia. Species richness increased in some sites and decreased in others post-fire. Some sites did not change species richness. Shaded areas indicate fire zones.

Fig. 4.

Species richness per site under four different fire severity categories (low, moderate, high and extreme) in New South Wales. Species richness overall did not significantly change before and after fire between the different levels of fire severity. Shaded areas indicate fire zones. N represents the number of sites within each fire severity category.

Discussion

We present a geographically and taxonomically broad analysis examining post-fire responses for frog species across south-eastern Australia – a continent increasingly affected by climate change and consequently increasing fire severity and extent (Bureau of Meteorology, CSIRO 2020; Lawrence et al. 2022). Encouragingly, Australian frog species richness did not decline overall after fire even with increasing fire severity, against our original hypothesis. Species richness increased in some survey areas but decreased in other areas, suggesting site-specific effects.

These variations in frog species richness at site level is consistent with other studies, identifying some increases (Brown et al. 2011; Allingham and Harvey 2013; Grundel et al. 2015; Mester et al. 2015) but also decreases after fire (Cano and Leynaud 2010; Constible et al. 2016). Such dichotomous responses occur in Australian frogs, with no discernible effects of fire detected for Heleioporus eyrei, Litoria olongburensis, L. freycineti or Crinia tinnula, but a negative impact was documented for both Litoria littlejohni (Daly and Craven 2007) and Geocrinia lutea (Driscoll and Roberts 1997). Furthermore, Beranek et al. (2023) found that amphibian metacommunities in the southern regions of NSW were negatively impacted by high fire severity, whereas an effect was not detected in the northern region. Generalist species may respond positively to fire (e.g. Anaxyrus spp.; dos Anjos et al. (2021), compared with specialist species (e.g. Phyllodytes luteolus; (Papp and Papp 2000; dos Anjos et al. 2021). Our generalist/common species (Crinia signifera, Litoria peronii, Limnodynastes peronii) had the most records compared with more specialist species with few records (for example, our threatened species). Persistence in post-fire landscapes reflects evolutionary history (see Beranek et al. 2023), ecological interactions, time since last burn, availability of local waterbodies and available shelter (dos Anjos et al. 2021). Some frogs can hide in burrows, ant-mounds and even bromeliads during fires (Moreira et al. 2009; Nomura et al. 2009; dos Anjos et al. 2021). Australian frogs take shelter in hollows, under bark, in dense vegetation, within deep soil cracks or within reeds such as Lomandra spp. (Mahony et al. 2023). Changes in diversity are likely to be considerably less sensitive than changes in abundance, and fire may impact frog population size, which may not be captured in our dataset (Hillebrand et al. 2018; Antão et al. 2020). It would also be useful to examine what factors drive adaptiveness to fire regimes in different frog species and what makes some species more susceptible. Our analysis was not able to necessarily differentiate between persistence and recolonisation, because some species presence post-fire could be due to quickly recolonising burnt areas rather than persistence through the fire period. Fine-scaled in-situ research, however, could contribute to better understanding these differences after future fire events.

Although our surveys were widespread, they are unlikely to include all frog species present. Time of year and weather, in particular, influence frog calling activity (Thompson et al. 2022), with some species only detectable under specific conditions (e.g. Litoria brevipalmata are rarely detected except after recent heavy rainfall). In addition, range-restricted species that occur in remote areas are less well sampled via the FrogID project and other citizen science projects. Other survey methods, such as targeted scientific surveys or passive acoustic monitoring (Beranek et al. 2023), may be necessary to elucidate the effects of fire on these species. This may explain why, despite considerable data collection, there were still relatively few sites with high fire impact (n = 11; Fig. 4). Other species may remain undetected if surveys were not conducted at the right time of year (e.g. Pseudophryne bibronii are most active in autumn). Moreover, we had relatively few records for many threatened species (in some cases none), limiting our ability to test their responses after the 2019/20 bushfires. This may be due to low survey effort in the habitat of many of the threatened frog species in the fire zone (i.e. relatively remote rainforest habitats). Thus, our analysis is complementary to targeted, ‘on the ground’ surveys for threatened species to quantify their responses post-fire (Beranek et al. 2023). Nonetheless, the high levels of sampling overall by citizen scientists post-fire increases confidence in areas where we see sites with decreased species richness. For example, less than half of the species were detected at a site in the Blue Mountains (NSW) (seven before to only three species post-fire), despite over three times more sampling effort post-fire. The site-specific changes our data reveal will inform identification of priority zones for future surveys to determine the full effects of fire on frog populations. For instance, these field-based surveys will be necessary to detect breeding success and to assess less readily detectable impacts from ecological data, such as genetic erosion (Potvin et al. 2017).

Moreover, unlike most Australian bushfire seasons, the subsequent onset of La Niña (September 2020) provided additional rainfall to the burnt landscape (Bureau of Meteorology 2020) – offering conditions likely more conducive for frog breeding activity and persistence. Thus, we emphasise that our results should be interpreted with caution: future results may differ if we see similar bushfire activity under more erratic El Niño Southern Oscillation (ENSO)/Indian Ocean Dipole (IOD) cycles due to climate change (Cai et al. 2009, 2013; Fasullo et al. 2018). For instance, an overall decrease in species richness may occur under El Niño conditions because necessary requirements for breeding recovery (such as water availability) may not be met. Furthermore, this study does not consider the interactions with other stressors and extreme events predicted under climate change, such as drought (King et al. 2017) or disease (Hossack et al. 2013). Thus, future research integrating these additional stressors with fire would be of great interest.

In conservation biology, we require extensive yet timely data to make informed decisions (Kirchhoff et al. 2021). After the severe bushfires, the global pandemic made collection by professional biologists problematic (Corlett et al. 2020). As a result, species priority lists for post-fire monitoring were estimated based on expert advice in the absence of the most up-to-date data. However, our study identifies the importance of citizen science in providing critical data for conservation, especially in response to unprecedented fires, and highlights how citizen science may also be relevant for other disturbances. It provided useful data on the effects of fire severity and extent, which did not significantly affect species richness of frogs overall in either the short- or long-term after fire. We emphasise the need for ongoing citizen science data collection and targeted scientific monitoring, especially for species less well-sampled, via citizen science and for threatened species.

Supplementary material

Supplementary material is available online.