K’gari (Fraser Island) World Heritage Area: evaluating change in plant species richness and composition over two decades

Introduction

Natural World Heritage listed conservation sites contain some of the Earth’s most valuable natural assets of outstanding universal value (UNESCO 2022a). However, in the context of evolving threats associated with human activity, the challenge of ensuring that natural areas sustain their outstanding universal value in the protection and conservation of biodiversity is an increasingly difficult one. K’gari (formerly known as Fraser Island), in the subtropical eastern coast of Australia, is an area of outstanding universal value and was inscribed on the World Heritage Area (WHA) List in 1992 (UNESCO 1992). At 122 km long and 181 851 ha, K’gari is the largest sand island in the world (UNESCO 1992). It features a diversity of ecosystem types and cultural values and is an outstanding example of ongoing biological, hydrological and geomorphological processes (UNESCO 1992; Walker et al. 2022). During the 2020 World Heritage Conservation Outlook Assessment (IUCN 2020), K’gari was rated as ‘Good with some concern’ but its vulnerability to threats was rated as ‘High’. The major threats to conservation values in K’gari include predicted climate change, unsuitable fire regimes, increased tourism and associated impacts including pollution, erosion and siltation, disturbance, and the introduction of invasive species (IUCN 2020; UNESCO 2022b). These threats are consistent with a recent review of the ecological and cultural environmental values of K’gari, which suggested that K’gari is a contested space between tourist visitation and associated environmental impacts (Walker et al. 2022).

Since European settlement, human impacts on K’gari ecosystems have been significant (FIDO 2022). Cutting of Kauri pine (Agathis robusta) and white beech (Gmelina leichhardtii) started in 1863, followed by blackbutt (Eucalyptus pilularis) and satinay (Syncarpia hillii) and subsequent clearing of some native vegetation for establishment of plantations (FIDO 2022). Sand mining on the island also started in 1971 and involved clearing of vegetation in strips. It resulted in less than 1000 ha affected before the industry was closed. However, timber harvesting continued until 1991. K’gari was listed as a World Heritage Area in 1992 (Walker et al. 2022).

Vegetation associations occurring on K’gari display an unusual level of complexity, with major changes in floristic and structural composition occurring over very short distances (IUCN 2020; Neldner et al. 2021). The vegetation comprises over 830 vascular plants and 57 pteridophytes. Together these cover 407 genera in 173 families, which are found in 16 broad vegetation groups (Applegate 2020; Neldner et al. 2021). These vegetation complexes reflect the long successional processes of sand-dune development and are characterised by abrupt transitions from younger coastal dune systems supporting rainforest and wet sclerophyll vegetation communities, with trees up to 50 m tall, to dry sclerophyll vegetation such as heathlands on older dunes (Walker et al. 1981, 2022; Applegate 2020; Neldner et al. 2021). Vegetation associations on K’gari also depict specialised adaptations to low fertility, fire, waterlogging and aridity (IUCN 1992). However, some of these ecosystems are currently at threat from inappropriate fire regimes and visitation pressures from human activity such as increased erosion and soil compaction linked to significant vehicle activity and overuse of some sites, and risk of incursion of weed species (Walker et al. 2022).

Significant human and financial resources have been directed towards the management of these threats including regulation of visitation, pest control, planned burning and bushfire response, as well as the protection and monitoring of biodiversity in K’gari (Hockings and Hobson 2000; Environmental Protection Agency 2005; Behrendorff et al. 2019; Neldner and Ngugi 2021). Among these initiatives is a long-term vegetation assessment program consisting of 56 permanently marked biodiversity monitoring sites that were set up by the University of Queensland (UQ) and Queensland Parks and Wildlife Service (QPWS) in 1995 (Hockings and Hobson 2000) and re-measured annually, with some sites re-measured until 2003. This time series dataset provides an opportunity to investigate the cumulative impacts of drivers such as bushfire.

Fire is an important factor influencing vegetation community structure and composition in Australia. Between 1992 and 2020 there were 13 bushfires reported on K’gari occurring in the surveyed areas and with different causes (QPWS 2021). Most vegetation communities on K’gari are highly fire-adapted, having a range of survival and regenerative mechanisms including epicormic shoots and basal resprouts (Meiklejohn et al. 2021; Neldner and Ngugi 2021). Some, however, are fire sensitive including the foredune communities of Casuarina equisetifolia subsp. incana (State of Queensland (DES) 2022). These are killed by fire and show very poor regenerative capacity post-fire (Neldner and Ngugi 2021). Although this highlights the importance of managing bushfires and implementing appropriate fire regimes to maintain the integrity of the World Heritage values (Environmental Protection Agency 2005; Neldner and Ngugi 2021; State of Queensland (DES) 2022; UNESCO 2022b), it also raises the question of how these bushfire events have been affecting the vegetation communities.

In this study, we examine change in plant communities on K’gari World Heritage Area over two decades. The specific objectives were to determine: (1) whether any changes in plant species richness (number of species) and/or species composition had occurred at the regional ecosystem level over the monitoring period; and (2) the extent of infestation of invasive non-native plant species (weeds) over the monitoring period. Data from 22 permanently marked biodiversity monitoring sites established in 1995 were used in the study.

Materials and methods

Study area

K’gari (Fraser Island) lies off the south-east coast of Queensland, Australia between latitude −24.6976° and −25.7982°, and longitude 152.9369° and 153.3546° (Fig. 1). The First Nations People, the Butchulla, have had a long and continuous connection with K’gari island and the surrounding waters (FIDO 2015; IUCN 2020; Walker et al. 2022). The climate for the region is characterised by warm humid summers and mild dry winters. The annual rainfall ranges from 879 mm to 2484 mm, with a mean of 1551 mm. Temperatures range from a mean minimum temperature of 13.0°C to a mean maximum temperature of 28.7°C (SILO climate 2022).

Fig. 1.

Location of permanently marked vegetation monitoring sites established in 1995 and resampled in 2021 in K’gari (Fraser Island) World Heritage Area.

Results

Species richness between 1997 and 2021

The distribution of species richness values among sites sampled within each RE, and among all the sampled REs, is presented using a boxplot in Fig. 2a. The highest species richness was recorded in RE 12.2.9 (Banksia aemula low open woodland), ranging from 23 to 42 species and a mean of 32 species (Table 1). Generally, there was greater variability in the number of species observed during surveys within each RE than between regional ecosystems, except for RE 12.2.15g (Fig. 2a). The median richness for most of the monitored REs was similar and ranged from 24 to 28 species. The lowest species richness (nine species) was recorded in RE 12.2.15g, which was characterised as closed sedgeland.

Fig. 2.

Plant species richness and composition for nine vegetation communities monitored over time in K’gari (Fraser Island) World Heritage Area. (a) boxplot of species richness. (b) boxplot of Jaccard similarity indices. (c) boxplot of Bray–Curtis dissimilarity indices comparing species composition in sites measured in 1999 and 2021. (d) is a comparison between native species richness derived from sites sampled in 2021 (black) and from technical descriptions of each regional ecosystem (hatched). Error bars show minimum and maximum values, X is the mean, and the middle bar is the median of richness values.

The results of the analysis of variance of species richness among all the REs showed highly significant differences (P < 0.001) but within each of the four regional ecosystems that had more than three sites, the differences were not statistically significant (Table S2). A paired samples t-test performed to compare species richness for all sites observed in 1999 and 2021, showed no significant difference in species richness between the two survey dates (t₂₁ = 0.575, P = 0.286). Similar comparisons of species richness done among sites that were burnt and unburnt in 2021 against the 1999 observations showed no statistically significant differences among the burnt sites (t₁₁ = 1.182, P = 0.131), nor among the unburnt sites (t₉ = −0.569, P = 0.292).

Discussion

Understanding how threats, and the management actions used to mitigate them, affect species richness and composition over time is important for the management of World Heritage Areas. In the K’gari World Heritage Area, visitor impacts, inappropriate fire regimes including lack of fire, invasive plants and animals, and climate change are known key threats (UNESCO 2022b; Walker et al. 2022). In this study, two primary components of species diversity were investigated: (1) species richness and (2) species composition. These were examined using data collected over 24 years (between 1997 and 2021) within the K’gari World Heritage Area. Despite observed fluctuations in species richness during site revisits, the overall differences within regional ecosystems were not statistically significant. There were, however, significant differences in species richness among regional ecosystems, which is consistent with differences in species assemblages that characterise regional ecosystem communities. This finding provided support for the use of regional ecosystems as a stratification criterion when assessing changes in the vegetation communities (Sattler and Williams 1999).

Fire is a major ‘tool’ in the management of biodiversity values of K’gari WHA (State of Queensland (DES) 2022). Although fire regimes that are not appropriate to the ecosystems are recognised as a ‘High Threat’ to the conservation values of the heritage listed area (IUCN 2020), fire is a critical and natural component of the ecology of most ecosystems on K’gari (Queensland Herbarium 2021c; State of Queensland (DES) 2022). Plant species in most of these ecosystems are highly adapted to fire, having one or more reproductive and or vegetative traits that enable them to survive and regenerate. These include seed banks in the soil (e.g. Acacia spp.) or canopy (e.g. Banksia spp. but not Banksia integrifolia), resprouting from epicormic shoots (e.g. eucalypts, B. aemula, Melaleuca quinquenervia) or an unharmed apical bud (e.g. Livistona decora) or a lignotuber (e.g. B. aemula, Banksia robur; most eucalypts but not E. pilularis) or other subterranean organs such as rhizomes, tubers and bulbs (Meiklejohn et al. 2021; Neldner and Ngugi 2021; State of Queensland (DES) 2022). There are relatively few obligate seed regenerators in these ecosystems. The reproduction method and the extent of recovery is influenced by the intensity and timing of the fire. In contrast the C. equisetifolia subsp. incana communities on frontal dunes are not adapted to fire and the species is readily killed by fire and has no canopy or soil stored seed (Meiklejohn et al. 2021).

All the sites in this study had experienced planned burn and/or bushfire events within the timeframe of this study, with most of the sites having had at least one fire event. Widespread intense bushfires tend to have the greatest impact on biodiversity as reported in K’gari WHA (Meiklejohn et al. 2021; Neldner and Ngugi 2021), New South Wales (Deb et al. 2020) and Victoria (Lindenmayer and Taylor 2020) in the 2019–2020 fires. This impact is exacerbated when they become too frequent.

Forest and woodland bushfires in south-eastern Australia have been relatively infrequent in the past and characterised by bushfire return intervals of >20 years (Williams et al. 2012; Murphy et al. 2013; Lindenmayer and Taylor 2020; Canadell et al. 2021). On North Stradbroke Island (Minjerribah), south-eastern Queensland and prior to 1996, a bushfire return interval of 30 years was recorded, but between 1996 and 2014 two bushfires were recorded (Kington et al. 2016; Williams et al. 2020). In south-eastern Queensland forests, a single bushfire in dry sclerophyll eucalypt forest was recorded between 1946 and 1996, a bushfire return interval of 50 years (Lewis 2020). In this study, the average bushfire return interval in our plots (Table 3) since 1994 was estimated at 13 years. This suggests that bushfire occurrence in the last two decades is more frequent than broader averages reported in south-eastern Queensland in the past.

In Victoria, Australia, of the 1.5 million ha burned during the 2019–2020 fire season, 600 109 ha have burned twice, and 112 957 ha have burned three times over the past 25 years; indicating a recurrence that is far shorter than the expected natural interval for these forest types of 75–150 years (Lindenmayer and Taylor 2020). These results are consistent with the finding of Dutta et al. (2016) that the frequency of bushfire is increasing significantly, and the finding of Canadell et al. (2021) that the mean number of years since the last bushfire in Australia has been decreasing consecutively in each of the past four decades. Further, Canadell et al. (2021) found that the frequency of very large bushfires has markedly increased since 2000 and is correlated with anthropogenic climate change.

In the 2021 K’gari post-fire recovery study (Neldner and Ngugi 2021), the extent of negative effects was dependent on vegetation type and the severity of the fire at the site and appeared to be short lived. Much of the vegetation was recovering by vegetative resprouting across the whole plant, or from the base of the stem for those that had been severely burnt (Meiklejohn et al. 2021; Neldner and Ngugi 2021). The regenerative adaptations are likely to explain why there is no statistical difference between the species richness reported on sites within the same regional ecosystem that had several bushfires versus those with few or no bushfires, in the last 26 years.

The Jaccard similarity indices enabled more detailed examination of the species composition. Although the differences in the Jaccard similarity indices within and among regional ecosystems were not statistically significant, the values were generally low as shown by an average value of 42%. A detailed examination of the species data used for the Jaccard analysis revealed that there were large differences in the list of unique species recorded at the two different primary sampling occasions (1999 and 2021). About 28% of the species observed during the 1999 visit were not recorded in the 2021 visit and vice versa (Appendix 2 (Ngugi et al. 2022)), and annuals contributed less than 10% of the total species richness. Although misidentification of species may have contributed to the discrepancy, differences originating from bushfire exposure cannot be excluded as a possible cause for the differences in observed species – especially as 57% of the sites in this study were burnt in the 2020 bushfire. The sites burnt in 2020 were in different stages of ecosystem recovery depending on their fire severity and vegetation type. Some sites were characterised by massive germination of seedlings from the soil seedbank for species such as Acacia (Meiklejohn et al. 2021; Neldner and Ngugi 2021).

Species composition results obtained using Bray–Curtis dissimilarity indices indicated high dissimilarity between 1999 and 2021 (average 65.9%). This is partly because plant cover was used as a surrogate for species abundance for this study and vegetation cover was adversely affected by the fire severity of the 2020 fire event in K’gari WHA. Species differ from one another in their environmental tolerances, resource use, and interactions with other species, and hence their abundance has a major influence on ecosystem functioning and stability (Cleland 2011). In K’gari, the differences in species responses to fire were evident in the recovery of different regional ecosystems following the 2020 fires (Meiklejohn et al. 2021; Neldner and Ngugi 2021).

Species richness across the eight regional ecosystems sampled in 2021 on K’gari was very similar to the overall species richness recorded for a larger pool of sites upon which the regional ecosystem technical descriptions are based. Mean native species richness of 26 species observed in 2021 was also very similar to 25 species obtained from the technical description for the regional ecosystems (Queensland Herbarium 2022). These results suggested that species richness rebounded over time as sites recovered from fire events.

The use of historical monitoring data to assess vegetation change over time is not without its challenges. Firstly, the two monitoring programs used different plot sizes. The 1995 monitoring program used variable plot sizes ranging from 25 to 400 m² (Hockings and Hobson 2000) which were considerably smaller than the 500 m² standard plot used by the Queensland Herbarium in 2021. Despite the differences in plot size, the analysis of variance results comparing species richness data collected using the two methods did not show statistically significant differences. It is recommended that the Queensland Herbarium methodology be used for future monitoring.

Secondly, two different methodologies were used for estimating species cover: the Braun-Blanquet method that used cover class to estimate the cover of each species, and the Queensland Herbarium methodology that estimated the cover of each species to the nearest one percent. It is therefore likely that species cover estimates based on the conversion of Braun-Blanquet scale to a single value (Matteodo et al. 2016) may have underestimated or overestimated cover values, and hence contributed to the greater dissimilarity obtained between survey events. Thirdly, it is possible there were discrepancies in species identification between the two survey events. However, in both programs, where a species could not be positively identified in the field, specimens were collected and taken to the Queensland Herbarium for identification. In a few cases, in the absence of flowers or fruit, identification to species level was not possible.

Increased tourism in K’gari has been recognised as acting as a driver for the introduction of invasive species (IUCN 2020; Walker et al. 2022). This study recorded only six invasive plant species in the 22 sites sampled. Most of these invasive species are not widespread over K’gari except lantana and the corky stem passionflower, which is becoming widespread in human-use areas. A comprehensive survey of invasive flora was not conducted as part of this study, but the Queensland Herbarium database (HERBRECS) has specimens of 109 naturalised non-native (invasive) plant species that have been collected on K’gari (Queensland Herbarium 2021b). Hence the results are encouraging compared to those from other frequently visited Parks in the south-east Queensland region (Ngugi et al. 2014) and more remote areas such as protected areas of the Cape York Peninsula (Ngugi and Neldner 2017), which have much greater numbers of recorded invasive species. Once invasive species become widespread, their control can be very protracted. The successful removal of bitou bush (Chrysanthemoides monilifera subsp. rotundata) from K’gari required a 35-year period of collaborative annual surveys and control efforts (Behrendorff et al. 2019). Vigilance will always be required on K’gari because of the continuous potential source of new propagules associated with the dispersal capacity of birds and wind, the high volume of vehicles, campers and bushwalkers to the island and the residents and visitors of the permanent townships.

Conclusion

This study investigated changes in species richness and composition among vegetation communities on K’gari (Fraser Island) over two decades. In addition to the ongoing threats of climate change, increasing tourism is acting as a driver for several threats to the WHA, which include fire ignition, pollution, erosion, siltation, disturbance, and the introduction of invasive species (IUCN 2020). Widespread, frequent, and severe bushfire is also a threat to the conservation of biodiversity on the Island.

Vegetation changes were investigated using time series data from monitoring sites established in 1995 soon after the declaration of the World Heritage status of the island. Species richness and composition recorded from these monitoring sites were compared with data collected from the same sites in 2021. Although there were fluctuations in species richness within regional ecosystems over time, there were no statistically significant differences in species richness over the monitoring period. Species canopy cover was adversely affected by the 2021 fire events and resulted in statistically significant differences (P = 0.0298) between dissimilarity indices of burnt and unburnt sites, with recently burnt sites showing greater dissimilarity. There was a large discrepancy in the list of species that were observed in 1999 and those observed in 2021. Although the overall richness was similar, only 42% of the species were common between these survey events. The extent to which this difference in unique species was related to exposure to recent bushfires could not be ascertained. Hence the possible impacts of fires on the species that were recorded in 1999 but not recorded in 2021 and vice versa should be investigated. The results suggest that most vegetation communities on K’gari, although adversely affected by intense bushfire in the short-term, were resilient and demonstrated adaptations to survival and recovery. An exception was C. equisetifolia subsp. incana woodland to low open forest on exposed frontal areas, which were killed by fire and not regenerating. Targeted planned burning is a key management tool aimed to achieve fire regimes that are appropriate to the ecosystems and to mitigate risk of severe bushfires. We recommend continued implementation of the existing planned burn guidelines for the south-east Queensland Bioregion (State of Queensland (DES) 2022).

Supplementary material

Supplementary material is available online.

Data availability

The data used for this analysis can be made available by contacting the authors.

Conflicts of interest

The authors declare that they have no conflicts of interest.

Declaration of funding

This study was jointly funded by the Queensland Herbarium and Biodiversity Science and the Queensland Parks and Wildlife Services & Partnerships (QPWS&P) Technical Services.