Introduction

The Canary Islands is a volcanic archipelago of Macaronesia located in the Northeast Atlantic Ocean between 14.8°N and 39.7°N and 96 km from northwest Africa (Fig. 1). The number of endemic species is high in many taxonomic groups, particularly those belonging to terrestrial organisms, such as vascular plants (573), arthropods (3202), or vertebrates (24) (Florencio et al. 2021). This archipelago is part of the Mediterranean biodiversity hotspot due to its high share of threatened endemics (Fernández-Palacios and Whittaker 2008). Tenerife is the largest and highest island of the archipelago, so it hosts a great diversity of habitats, from desert-like coastal areas, through cloud forest and pine forest, to alpine scrub at more than 2000 m a.s.l. Alpine scrub is one of the most isolated ecosystems with extreme climate conditions and high diversity of endemic species (Fernández-Palacios et al. 2014). This unique environment is present in only two of the Canary Islands, La Palma and Tenerife. The small populations and ranges of taxa in this ecosystem make them very vulnerable to human alterations such as climate change or the introduction of allochthonous species (Cubas et al. 2018, 2022). Among the invasive mammals found in the Canary Islands, the cat (Felis catus) and the European rabbit (Oryctolagus cuniculus) stand out, both present in the main ecosystems of the archipelago (Cubas et al. 2019; Medina and Nogales 2009).

Fig. 1
figure 1

Map showing a the location of the Canary Islands (red square), b the island of Tenerife and the Teide National Park (grey shape), c and the location of the samples in the National Park (white dots)

Full size image

Free-ranging cats are one of the greatest threats to wildlife worldwide (Doherty et al. 2016). This is exacerbated on islands, where cats are linked to 14% of historic global extinctions of insular birds, mammals, and reptiles (Medina et al. 2011). The impact of cats has been studied in at least 120 archipelagos. Studies carried out in the Canary Islands have shown that allochthonous mammals (mice, rats, and rabbits) are the major component in the diet of feral cats, followed by reptiles and native birds (Nogales and Medina 2009). This pattern coincides with that found in other archipelagos (Bonnaud et al. 2011). However, even though the major components of the diet are introduced species, the impact of cats on native fauna is severe. According to Bonnaud et al. (2011), cats on islands prey upon 248 species (179 vertebrates), of which at least 36 are threatened according to IUCN criteria. In the Canary Islands, feral cats act as opportunistic predators with a broad dietary spectrum, including at least 68 types of prey. Part of its diet is composed of native species (where 47% of the invertebrates, 44% of the birds, and all reptiles are endemic to the archipelago), while the 5 mammal species that are consumed are introduced (Medina and Nogales 2009). In the Canary Islands, rabbits are the most important prey in all the ecosystems, accounting for more than 55% of the total biomass of the diet, except in the laurel forest, where rabbits are scarce (Nogales and Medina 2009).

Since trophic webs in the Canary Islands currently integrate allochthonous and native species, the reduction of the density of one species (i.e. allochthonous) could have a negative impact on the other (i.e. native), caused by an increase of predation on the latter (Zavaleta et al. 2001; Courchamp et al. 2003). Recently, it has been observed that the reduction of rabbit population has resulted in changes in the diet of feral cats in a locality south of Tenerife Island, where native lizards and birds were found to be consumed more frequently than 15 years ago (Flores and Rando 2021). In relation to biomass, the consumption of rabbits decreased from 62.4 to 30.3%, while that of reptiles and native birds increased from 3.6 to 11.5% and from 2.4 to 39%, respectively, with a consequent increase in their impact on native species (Flores and Rando 2021). This is particularly serious for species with small populations and conservation problems (Flores and Rando 2021). Given that the reduction in rabbit densities appears to be a direct consequence of rabbit haemorrhagic fever virus (RHDV), a disease present in all the habitats in the island (Foronda et al. 2005), similar changes in the diet of feral cats might have occurred in other ecosystems throughout the island.

We conducted the present study of the trophic ecology of cats to test whether the change detected in the diet of cats locally in the coastal areas of Tenerife over a 15-year period (Flores and Rando 2021) has occurred also in other ecosystems of the island. Specifically, it was carried out in El Teide National Park, in the high mountain ecosystem, where the diet of cats had already been studied 35 years ago (Nogales et al. 1990). The objectives of this study were to determine the durability and prevalence of changes taking place in the Tenerife island ecosystem resulting from the decline in the rabbit population and to update information on their impact on the native biodiversity.

Material and methods

Study area

This study was performed in the high mountain ecosystem of Tenerife in Teide National Park (18,990 ha and a maximum elevation of 3715 m a.s.l.) (Fig. 1). The climate is mainly determined by orientation and elevation; average precipitation varies between 350 and 500 mm, the lowest temperature being reached in winter (− 15 °C) and the highest in summer (30 °C), with frosts from October to May and snowfall in winter (Díaz-Armas et al. 2019). Sampling was carried out at altitudes between 2000 and 2400 m covered mainly by the summit legume shrubland. This is characterised by Spartocytisus supranubius, a broom shrub endemic to Tenerife and La Palma, which can reach a height of 3 m (Del Arco and Delgado 2018). Part of the sampling was carried out in places where there are specimens of Juniperus cedrus trees due to their importance for migratory birds in autumn and winter (Rumeu et al. 2009).

Scat sampling and identification of diet items

The diet of feral cats was determined from scat analysis. Between May and August (2021), surveys were conducted in several sectors of the National Park to detect cat drop**s on the ground. Drop**s were found in semi-buried groups. Their characteristic cylindrical shape, size 100–150 × 18–30 mm, compact, consisting of a number of fragments joined together (Blanco 1998) make them clearly distinguishable from other mammal species present in the National Park. To avoid overestimating the number of prey, scat segments that were close to each other and had a similar appearance were treated as a unit (sample). In the laboratory, scats were broken up to extract bones, hair, and/or feathers to identify them. Prey remain identifications were made using materials from the reference collection of the Departamento de Biología Animal, Edafología y Geología de la Universidad de La Laguna. The minimum number of individuals (NMI) per species in each sample was determined with repeated anatomical structures. Because the scats remained unaltered over a long period before their disintegration, the collected material corresponded to all seasons.

Data analysis

In order to describe the feral cat diet and compare it to that from 35 years ago (Nogales et al. 1990), the frequency of occurrence of each prey item (FO) and the percentage it represented (%Pr) were calculated; from the average weight in grams of each prey individual (according to Nogales et al. 1990; Blanco 1998; Dunning 2008; Medina et al. 2010; Flores and Rando 2021), the mass (M) of each prey and its percentage of biomass (%W) in the diet were calculated. For large prey, which constitute more than the daily food intake by cats, the total biomass applied was 170 g (Fitzgerald and Karl 1979; Medina et al. 2010). With these data, the index of relative importance (IRI) was calculated for each type of prey. Invertebrates were included in one category and were not considered in the calculation of biomass because of the slight weight they could contribute with to the diet (Medina and Nogales 2009).

To compare the consumption of the different types of prey, chi-square tests were performed following the methodology of previous works (Medina and Nogales 2009) in R Studio 3.6.2 (R Core Team 2019). Two indices were calculated using the percentage of biomass (%W) of each prey item. The standardised Levins niche breadth index (B) (values close to 0 indicate diet specialisation, while those close to 1 suggest a broad diet). The overlap between the diet determined in this study and that from 35 years ago was evaluated by the Morisita similarity index (C) (values close to 0 indicate a large difference between diets, while those close to 1 suggest similar diets) (Krebs 1989).

Annual prey intake of feral cats in Teide National Park

The annual prey intake ingested by cats in the National Park was estimated assuming an intake of 170 g by day (Fitzgerald and Karl 1979; Medina et al. 2010). Moreover, following the methodology of previous work, a year was considered to be 365.25 days long (Blancher 2013; Loss et al. 2013; Woinarski et al. 2017; 2018). Thus, the biomass ingested by a single cat in 1 year was calculated by multiplying 365.25 days by the biomass ingested in 1 day. The number of prey ingested was calculated from the percentage of biomass (%W) contributed by each prey item to the diet, i.e. by multiplying (%W/100) by the biomass ingested by a cat over a year and dividing this number by the weight of each prey item.

Results

Contemporary cat diet

A total of 727 prey items were identified in the 301 analysed drop**s, of which 648 belonged to vertebrates (Table 1, Fig. 1, Supplementary Information). A minimum of eight species of vertebrates were identified: 2 endemic reptiles (Tenerife lizard Gallotia galloti, Tenerife skink Chalcides viridanus), 4 allochthonous mammals (European rabbit Oryctolagus cuniculus, black rat Rattus rattus, house mouse Mus musculus, and Etruscan shrew Suncus etruscus), and two birds (Berthelot’s pipit Anthus berthelotii) and rock pigeon (Columba livia) (Table 1). Invertebrates showed a frequency of occurrence of 26.2% and a minimum number of individuals of 79.

Table 1 Results of the analysis of cat drop**s collected in Teide National Park in 1986 (n = 221) (Nogales et al. 1990) and in 2021 (n = 301) for this work
Full size table

Endemic reptiles with an NMI of 418 were the main group preyed upon (χ2 = 381; df = 2; p < 0.001), representing 64.5% of all prey consumed and 29.9% of the biomass (Table 1). The lizard was the most frequent prey (84.7%), providing most of the biomass of reptiles to the diet (29.8%). The set of mammals contributed 67.8% of the total biomass. Rabbit remains were present in 31.2% of the samples, representing a minimum of 94 specimens. Additionally, rabbits made the highest contribution of biomass at 53.9%, even though it only represented 14.5% of all prey consumed. Interestingly, two Etruscan shrews were identified, this being the first time that this species has been detected among cat prey in the Canary Islands. Bird remains were identified in less than 5% of the samples, with 2 passerines and 12 nonpasserines. These represent only 1.8% of the biomass in the diet.

Contemporary vs. past cat diet

Nogales et al. (1990) processed 221 samples collected in 1986 and identified 523 prey items, 451 of which were vertebrates (Table 1). Among the vertebrates, they identified a minimum of 8 groups: 3 endemic reptiles (Tenerife lizard, Tenerife skink, and Tenerife gecko Tarentola delalandii), 3 exotic mammals (rabbit, black rat, and house mouse), and 2 groups of birds (passerines, nonpasserines).

Consumption of reptiles found in this study appeared significantly higher as compared to the data collected by Nogales et al. (1990) 35 years ago (χ2 = 10.2, df = 1; p = 0.001). Although contribution of rats and mice to the cat diet increased 1.1 and 1.2 times, from 1986 to 2021, respectively, the difference was not significant (χ2 = 0.015; df = 1; p = 0.900 and χ2 = 1.199; df = 1; p = 0.274). In contrast, consumption of rabbits as well as mammals, in general, decreased highly significantly in 2021 as compared to 1986 (χ2 = 31.807; df = 1; p < 0.001 and χ2 = 13.245; df = 1; p < 0.001, respectively). The amount of birds in the cats’ diet in 2021 was 2.75 times higher than in 1986, however, no significant differences were found (χ2 = 1.945; df = 1; p = 0.163). Levins index calculated for the cats diet in this study was slightly higher (B = 0.22) than that for 1986 (B = 0.11), suggesting more specialised diet 35 years ago. On the other hand, the Morisita index indicates a high degree of overlap in the diet (C = 0.94) between both studies.

Annual prey intake of feral cats in Teide National Park

The annual predation of vertebrates preyed upon by a single cat has been estimated to be approximately on average 1356 annually (Table 2): approximately 860 lizards, 197 rabbits, and 29 birds. Assuming, a density of 1 cat/km2 and an area of 190 km2 for the National Park, the feral cat population consumes approximately 257,737 vertebrates annually, of which 163,465 are lizards, 37,384 are rabbits, and 5587 are birds (Table 2).

Table 2 Annual numbers of vertebrate intake of both a single cat and the cat population assuming a density of 1 cat/km2 in Teide National Park and considering that 170 g of prey equals one day’s consumption
Full size table

Discussion

Shift in the trophic ecology

Previous studies conducted in different ecosystems of the Canary Islands indicated that feral cats mainly fed on introduced rabbits and rodents (Nogales and Medina 2009; Piquet et al. 2019). In the high mountains of Tenerife, introduced mammals accounted for 81.8% of the cat diet biomass (Nogales et al. 1990). The present work shows a significant decrease in mammal consumption in the alpine ecosystem of Tenerife, both in terms of percentage and biomass. The differences detected when comparing the total consumed mammals can possibly be explained by the notable difference in rabbit consumption, as 35 years ago, cats consumed twice as many rabbits as today (28.6% in 1986 and 14.5% in 2021), which represents a 19.1% decrease in the biomass (73% in 1986 and 53.9% in 2021) contributed by this mammal to the cat diet in the National Park. The values obtained in this study are the lowest ever recorded for cat diets in high mountain scrubland in all the Macaronesian islands studied thus far (Medina et al. 2010) and are among the lowest of all those recorded for the Canary Islands (Nogales and Medina 2009; Piquet et al. 2019; Flores and Rando 2021). The decrease in rabbit density in different habitats of Tenerife has been recorded for over a decade (Nadal 2020) and in censuses in Teide National Park, indicating a low density since 2014 (MAAPNT 2020). A plausible explanation for this tendency is the impact of pathogens, such as rabbit haemorrhagic disease virus (RHDV) and specifically its new variant RHDV2. The RHDV2 variant was detected on the island for the first time in 2016 (Martín-Alonso et al. 2016) and has been suggested to be replacing RHDV, with a similar prevalence across Tenerife’s habitats, including high mountain areas (Martín-Carrillo 2018). The RHDV2 variant has also been responsible for the drastic decrease in rabbit populations in the Iberian Peninsula in some areas, with reductions of up to 75% (Delibes-Mateos et al. 2014).

Additionally, this trend is directly linked to an increase in reptile intake, specifically lizards. The frequency of occurrence of reptiles found in this study is one of the highest detected in islands worldwide (Bonnaud et al. 2011) and similar to those recorded in Santa Luzia (Cape Verde), where the frequency of occurrence and biomass of reptiles in cats’ diet was 91.7 and 72.8%, respectively (Medina et al. 2020). In this case, high frequency of reptiles was consequence of the absence of rabbits and rats, the house mouse being the only allochthonous mammal present on the island (Medina et al. 2020).

The absence of significant differences in the consumption of birds between the diet today and that 35 years ago can be explained by the fact that birds constituted a minor component in the diet of cats in the National Park in both periods. The slight difference observed between Levins index of both periods is probably affected by different contributions of rabbit in the cat diet. On the other hand, the high degree of overlap between both diets could be explained in terms of the prey found which are essentially the same. In addition, the two most frequent and important prey (rabbits and lizards) also coincide in each study.

This shift in the trophic behaviour of cats, caused by variations in the prey availability (Fitzgerald and Turner 2000; Zavaleta et al. 2001; Courchamp et al. 2003), had already been described for the coastal scrubland in Montaña de Guaza (Tenerife) where changes in the proportions of prey consumed were observed (Flores and Rando 2021). That study showed a significant reduction in the contribution of mammal biomass (mainly rabbits) from 93.9% in 2004 to 40.5% in 2020 and a consequent increase in the contribution of birds and reptiles to the cat’s diet from 2.4 to 39% and from 3.6 to 11.5%, respectively. In the same way as in the case of the Teide National Park (this study), this shift was suggested to be driven by the general decline in rabbit populations in Tenerife (Nadal 2020) because of the high prevalence of rabbit haemorrhagic disease virus (RHDV2) in the island (Foronda et al. 2005; Martín-Alonso et al. 2016).

Conservation concerns

The data of this study showed the enormous predation pressure exerted by cats on native species in the high mountains of Tenerife. Of the estimated 257,739 vertebrate specimens preyed upon by cats in Teide National Park annually, two-thirds, 171,836, corresponded to native species (reptiles and birds). These figures reflect a serious conservation problem in the National Park with very negative consequences for native species and the ecological functions they perform. Therefore, this situation should be a priority in the management and conservation actions to be carried out in these places.

Since trophic networks integrate both invasive and native species, the decrease in the density of an exotic prey can affect the native species negatively due to an increase of predation (Zavaleta et al. 2001; Courchamp et al. 2003). This has been observed in the Montaña de Guaza, where a decrease in rabbit consumption by cats has a negative impact on the population of the critically endangered endemic giant lizards of Tenerife, Gallotia intermedia, decreasing their range and abundance (Padilla et al. 2019; Flores and Rando 2021). In the high mountains of Tenerife, the fall in rabbit density has increased the importance of reptiles in the cat’s diet and, to a lesser extent, that of birds. More than 60% of the vertebrate prey found in cat faeces correspond to lizards.

Gallotia galloti has been identified as the primary short-distance disperser of cedar seeds in the high mountains of Tenerife (Rumeu et al. 2014). Even though lizards show a high density in the National Park (Rumeu et al. 2014), the high predation on lizards by feral cats could decrease their densities and attenuate the ecological role they play within the ecosystem. On the other hand, the high predation on lizards may also have negative consequences for their native predators by direct competition. Specifically, in the diet of the kestrel Falco tinnunculus, lizards represent 50% of the biomass (Carrillo et al. 2017), while in the diet of the southern grey shrike Lanius meridionalis, they represent between 71 and 93% of the biomass (Padilla et al. 2009).

During the coldest months, birds may be more heavily predated upon. Our data show that the cat population of the National Park could be consuming 5587 birds per year. Moreover, the highest intake has been detected in sectors hosting the most important cedar populations, which are very important places for birds during the winter (Rumeu et al. 2009). In this context, the ring ouzel Turdus torquatus that plays a key role as a long-distance disperser of the Canary cedar, especially considering the scarcity of the common raven Corvus corax in the National Park (Rumeu et al. 2014), is a species of serious concern. This is because it spends the winter and part of the spring in the high mountain of the island (Rumeu et al. 2009), a period when the reptiles show less activity due to the low temperatures. During the coldest season, T. torquatus and other bird species frequent areas with cedars, seeking both food and shelter, making them very vulnerable to cat predation.

The fencing of cedar populations to prevent the cat access (Dickman 2012) or selective trap** of cats may be some of the actions needed to try to reduce the impact that cats are currently having on native fauna in the Teide National Park. On the other hand, any conservation action that involves a reduction in rabbit density must also be associated with cat control. This is the only way to prevent increasing impacts on native species, as we have described in this study.

Conclusion

The data obtained in this work showed not only huge impact of cats on the native fauna of a National Park but they also indicate the need for updated studies on the trophic ecology of invasive predators on oceanic islands since their diet can shift over time. This work and that of Flores and Rando (2021) showed that this has happened in Tenerife after a disease affecting rabbits arrived on the island, resulting in a greater impact of cats on native species. For this reason, there is a great need to update and improve cat management and control strategies, especially in those areas where the rarest native species may be providing essential ecological functions for ecosystems.