Drought in south-west Australia links to urban immigration across multiple avian taxa

Introduction

With the continued growth of the human population, urban areas are expanding more rapidly than any other habitat type, leading to significant changes in the global landscape (Cazalis et al. 2023). This expansion has been associated with the decline or local extinction of many indigenous animal species from urban areas (McDonald et al. 2013). However, some taxa have successfully adapted to and even thrived in urban environments (Bonier et al. 2007; Spelt et al. 2019). Urban expansion occurs alongside climate change and overall declines in biodiversity, and there is a critical need to understand how these combined stressors interact to influence animal populations in and around urban areas.

Urban areas can serve as unexpected refuges for wildlife during environmental changes like droughts (Lowry et al. 2013). Although urban expansion typically leads to habitat loss, these areas can provide alternative resources such as water and food through gardens, artificial water sources such as ponds and lakes, or food waste that may be scavenged. These resources can attract various bird species when natural supplies are scarce (Magle et al. 2012). Additionally, urban structures offer unique nesting sites not available in drought-impacted rural areas, potentially increasing bird concentrations in cities (Bateman and Fleming 2012). Drought affects bird populations by limiting water and food, altering habitats, and increasing competition (Albright et al. 2010). As resources dwindle, these conditions often compel birds to move to areas where water and food are more readily available, including urban environments. In Australia, instances of bird immigration into urban areas in response to drought have historically been recorded for a number of species, including the galah (Eolophus roseicapilla), sulphur-crested cockatoo (Cacatua galerita), little corella (Cacatua sanguinea), and rainbow lorikeet (Trichoglossus moluccanus) (Davis et al. 2011). A recent climate anomaly in south-west Western Australia (SW WA) resulted in some of the lowest rainfall and highest temperatures on record (BOM 2024). During this event, anecdotal reports indicated an unusually high presence of certain bird species that are typically rare in the Perth Metropolitan Area (the major city in SW WA). In this study, we used the extensive citizen science eBird dataset to quantitatively assess changes in avian taxa in the greater Perth Metropolitan Area during the climate anomaly. We investigated correlations between reporting rates and cumulative rainfall, including both short-term (6 months) and longer-term rainfall trends (12 and 18 months) to better understand the relationship between bird movements and varying rainfall patterns. Our findings give new insight on how species from diverse bird groups – raptors, honeyeaters, and rails – can respond to extreme climatic events, and the potential role of urban habitats as the climate changes.

Materials and methods

Study context

South-west Western Australia (SW WA) is characterized by a Mediterranean climate with hot, dry summers and mild, wet winters. It was recently identified as a global drying hotspot (Masson-Delmotte et al. 2021). The study area was defined as a 25-km radius from Perth’s metropolitan center in SW WA, where approximately 80% of Western Australia’s population (2.3 million) live. Average annual rainfall in Perth has decreased by around 130 mm (15%), from ~860 mm to ~730 mm, over the past 30 years (1993–2023) when compared to the 30 years previous (1959–1988) (BOM 2024). The Perth metropolitan area, situated in the Swan Coastal Plain bioregion, was once a complex network of wetlands of which remnants remain (Davis and Froend 1999). The remnant areas comprise several significant water bodies (now artificial to varying extents) including Lake Joondalup, Herdsman’s Lake, Bibra Lake, and the Swan River and its tributaries, which are water bird hotspots (Fig. 1).

Fig. 1.

Satellite map with red circular shading indicating the urban area within a 25-km radius from Perth city, Western Australia. The inset map of Australia shows the approximate location of the study area (pink square) and bioregions.

Between October 2023 and April 2024, the Perth metropolitan area experienced record low rainfall and above average temperatures (BOM 2024, see Supplementary Fig. S1). During this time, the area received only 23 mm of rain, representing the city’s driest period since rainfall records began in the 1870s. Summer temperatures were 1.7°C above the long-term average, making it one of the warmest summers since observations began in 1910 (BOM 2024, Fig. S1b). In early 2024, large areas of vegetation began to turn brown and die across the south-west (Fontaine et al. 2024).

Results

The final filtered dataset comprised a total of 749,477 species observations from 4460 locations within the study area, with 273 species recorded in total. The most extreme part of the drought was identified as the ~7-month period spanning October 2023–April 2024, where monthly rainfall was 70–100% lower than average and monthly maximum temperatures were up to 12.5% higher than average (Fig. S1). Four species exhibited very high reporting rates during the drought period, all with reporting rates surpassing 2: (1) the black-shouldered kite (Elanus axillaris); (2) the black-tailed native-hen (Tribonyx ventralis); (3) the tawny-crowned honeyeater (Gliciphila melanops); and (4) the western spinebill (Acanthorhynchus superciliosus) (Fig. 2). We identified the Australian hobby (Falco longipennis), dusky moorhen (Gallinula tenebrosa), New Holland honeyeater (Phylidonyris novaehollandiae), and singing honeyeater (Gavicalis virescens), as ecologically similar species that exhibited no major change in reporting rates during the drought event, for comparison (Fig. 2). Other species with high reporting rates included the bush-stone curlew (Burhinus grallarius) and the yellow-billed spoonbill (Platalea flavipes). However, these species had low average encounter rates, and therefore were not included in further analyses. The only species that showed a decrease in reporting rate <0.5 was the fan-tailed cuckoo (Cacomantis flabelliformis), but the encounter rate for this species was generally low and therefore the data was not deemed adequate for further investigation.

Fig. 2.

Trends in eBird reporting rates (left y-axis) and proportional differences in cumulative 12-month rainfall (right y-axis) from the long-term average for selected bird species from January 2019 to August 2024. A reporting rate of one indicates that the reporting rate in a given month was the same as the 36-month rolling average while a value greater or less than one means that it reporting rate was higher or lower than the 36-month rolling average. The recent drought event spans ~7 months as shown by vertical dashed red lines showing the beginning of the drought event (October 2023) and the time when the drought was broken by meaningful rainfall (76 mm) in May 2024. Left panels show species with increased reporting rates during the climate anomaly: black-shouldered kite (Elanus axillaris); black-tailed native-hen (Tribonyx ventralis); tawny-crowned honeyeater (Gliciphila melanops); and western spinebill (Acanthorhynchus superciliosus). Right panels show similar reference species that maintained stable reporting rates: Australian hobby (Falco longipennis); dusky moorhen (Gallinula tenebrosa); New Holland honeyeater (Phylidonyris novaehollandiae); and singing honeyeater (Gavicalis virescens).

The peak reporting rate during the drought period varied for the four species identified. Reporting rates for the black-shouldered kite increased to ~3 in July 2023 (3 months prior to the drought event), before declining back to near the baseline in September 2023, and then steadily increasing from October 2023 (the start of the drought event) to peak at 9.2-times higher than the average in January 2024, ~3 months after the onset of the drought period. The reporting rate for the black-shoulder kite remained high (~7) through to April 2024. In May 2024, when the drought broke, the reporting rate dropped to 3.5, and down to 1.1 by August 2024.

The tawny-crowned honeyeater peaked in January and February 2024, at 8–8.5. The western spinebill peaked a couple of months later, around 2.8–3.3 from February to April 2024. Conversely, the black-tailed native-hen reporting rate peaked at 10 in November 2023, early in the drought period, staying high in December and January at 8.8 and 5.8, before dropping to zero. This species also had a relatively high reporting rate in September 2023, which we defined as prior to the main drought, although noting that the rainfall was ~40% below average in September. Reporting rates for all four species returned to near the baseline (1) or to zero, within several months after the drought broke in May 2024 (Fig. 2).

The top models predicting species reporting rates indicated that the black-shouldered kite, tawny-crowned honeyeater, and western spinebill were best predicted by percentage changes in 6-month cumulative rainfall (Table 1, Fig. 3). Model performance, measured by Nagelkerke’s R², was moderate for the western spinebill (R² = 0.31), and higher for tawny-crowned honeyeater (R² = 0.48), and black-shouldered kite (R² = 0.55). The null model was the top model for the black-tailed native-hen, with all other models also within 2 AICc, however the rainfall predictors were not significant in these models. For the black-shouldered kite, tawny-crowned honeyeater, and western spinebill, 6-month cumulative rainfall was significant, with estimates of −2.77 (P = 0.00), −2.60 (P = 0.00), and −1.34 (P = 0.00), respectively.

Fig. 3.

Generalized linear model predictions of reporting rate (y-axis) as a function of six month cumulative rainfall for the western spinebill, black-shouldered kite, and tawny-crowned honeyeater in the Perth metropolitan area. The shaded areas represent the 95% confidence intervals. Black points are the raw data points for each month included in the analysis. Photos provided by Harry Moore, Gary Tate, and Kaylene Taylor.

Discussion

The impact of climate change and drought on bird populations has become increasingly evident as shifts in environmental conditions force wildlife to adapt to a hotter and drier world (Albright et al. 2010). We observed a sharp increase in reporting rates for four bird species, up to nine times higher than the average, within the Perth greater metropolitan area, indicating a significant rise in their presence during and around a period of record low rainfall and record-high temperatures. Reporting rates for three of these species, the black-shouldered kite; the tawny-crowned honeyeater; and the western spinebill were found to be strongly correlated with deviations in cumulative rainfall from the Perth greater metropolitan area average. The black-tailed native-hen’s reporting rate peaked early in the drought and was not found to be signficantly related to the cumulative rainfall predictors we examined. This species is known to be highly eruptive and is only occasionally reported in Perth, so it may have been responding to other factors as well as having patchier data making model fitting more difficult. Our findings contribute to the growing body of evidence highlighting the disruptive impacts of climate change on bird communities and add new insight into the potential role of urban environments during climate anomalies.

Drought has been correlated with bird immigration in other Australian metropolitan areas. For example, in the urban centers of Sydney and Melbourne on Australia’s east coast, parrot immigration across various species has correlated strongly with decreases in inland rainfall (Davis et al. 2011; Loyn and Menkhorst 2011). We did not detect a strong response to the current drought for any of the parrot species, however, similar trends have been observed historically within the study area for little corellas and galahs, with urban reporting rates increasing substantially in the late 1990s, corresponding with the millennium drought (Recher 2023). While anecdotal evidence suggests that other bird groups, such as honeyeaters and birds of prey, can also be reported more frequently in urban areas during droughts, evidence within the literature remains scarce.

The black-shouldered kite, a small falcon-shaped hawk, was the only bird of prey that exhibited a significant increase in reporting in the eBird dataset at the time of the climate anomaly. This species typically occurs in open country, and has been characterized as partially nomadic, with its presence often reflecting prey availability (mostly rodents, lizards, and invertebrates (Menkhorst et al. 2017) and rainfall (Chambers 2008). For example, Pavey and Nano (2013) observed the black-shouldered kite in the Simpson Desert only after significant rainfall events led to a surge in rodent populations. We identified cumulative rainfall in the previous 6 months as the strongest predictor of reporting rate for this species. It is likely that the black-shouldered kite has expanded its range since European arrival (Menkhorst et al. 2017), benefitting from land-clearing and irrigation practices that create suitable hunting areas. Given these traits, and past adaptation to changing habitat, it is perhaps not surprising that the black-shouldered kite may have capitalized on greater food resources in metropolitan area in response to poor conditions and assumed resource depletion in surrounding areas.

The two honeyeaters (western spinebill and tawny-crowned honeyeater) with the strongest response to the drought are both specialist nectar feeders (Menkhorst et al. 2017). Honeyeaters are known make short-moderate landscape scale movement in search of flowering plants. For example, Morris and Wooller (2001) observed that honeyeater movements in a semi-arid eucalypt woodland near Kambalda, south-west Australia, were linked to fluctuations in nectar availability following rainfall, with species entering the area opportunistically in response to increased flowering. The western spinebill is a resident species in Perth, typically inhabiting heath, woodlands, and forest, although it will move in search of blossoming vegetation. Long term data from King’s Park, a large urban bush reserve, suggests the species is likely declining in Perth (Recher 2023). The tawny-crowned honeyeater is somewhat ecologically similar to the western spinebill but is a heathland specialist (Menkhorst et al. 2017). While both species are among the least commonly reported honeyeaters in Perth, most observations typically occur between February and April, coinciding with the driest months in Perth.

We have made the interpretation that the increases in reporting rates are most likely explained by urban immigration, with birds having moved into the metropolitan area from surrounding landscapes. Another explanation would be an increase in reproductive output causing a local population increase. However, the timeframe when the sharp increase in reporting rates occurred makes this explanation unlikely. We limited our analysis to the Perth metropolitan area, where a large number of eBird surveys are conducted and where the data is of high quality. Identifying if corresponding declines in reporting rates occurred outside of the study while reporting rates increased inside the study area would shed more insight on the timing and direction of bird movement. Our study correlated reporting rates to rainfall for some species, but it was beyond the scope to identify other mechanisms that may be difficult to measure but are nonetheless important. Further research is needed to understand the full range of factors influencing the patterns we have identified here. Various factors may limit successful bird immigration to urban areas, even in the case when a move offers potential benefit to a species. For example, some birds are highly site-attached and unlikely to move outside of their territory; or urban residential birds may protect resources and compete with immigrating birds.

Climate change and the continuing spread of urbanization are altering the availability of natural habitats for many species. Documenting and understanding the differential responses of taxa is essential for developing suitable conservation measures (Bateman and Fleming 2012).

We identified strong drought responses in bird species from distinct groups (raptors, honeyeaters, and a rail). However, other similar species showed no significant change in reporting rate, suggesting the effect is highly species dependent. Our findings, though limited in scope, provide important considerations for urban ecology, suggesting that when conditions are unfavorable in surrounding areas, fauna may move closer to cities in search of alternative resources (Johnson and Munshi-South 2017).

Supplementary material

Supplementary material is available online.