Non-native Red-billed Blue Magpie Urocissa erythrorhyncha expanded in lowlands with moderate forest cover, with no significant impact on native common bird occupancy, in Shikoku, southern Japan

doi:10.21203/rs.3.rs-4746306/v1

Download PDF

Research Article

Non-native Red-billed Blue Magpie Urocissa erythrorhyncha expanded in lowlands with moderate forest cover, with no significant impact on native common bird occupancy, in Shikoku, southern Japan

https://doi.org/10.21203/rs.3.rs-4746306/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Non-native bird species have colonized and negatively affected natural ecosystems and social economics globally; however, most cases have been understudied. We evaluated the effectiveness of playback surveys for enhancing magpie detectability of the non-native Red-billed Blue Magpie (Urocissa erythrorhyncha), and revealed the drivers of the magpie distribution using an occupancy model that considers the detection process and effects of survey conditions in Shikoku, southern Japan. Using this model, we mapped the potential distribution of suitable magpie habitats across Shikoku. Furthermore, we obtained detection/non-detection data for native bird species [Narcissus Flycatcher (Ficedula narcissina), Varied Tit (Poecile varius), Japanese Tit (Parus minor), and Japanese Bush Warbler (Cettia diphone)], and evaluated the impacts of the magpie on occupancy of these native bird species using a multispecies occupancy model that considered interspecific interactions (i.e., co-occurrence or mutually exclusive occurrence patterns). The results showed that detection probability was enhanced by broadcasting a specific series of magpie calls in the early morning from late May to early July. Magpie occupancy was higher in areas of lower elevation and peaked in areas with moderate forest cover (76%). However, magpie presence did not significantly affect the occupancy of four native bird species. Mapping the distribution of magpie occupancy demonstrated that potentially suitable habitats are widely distributed in near-coast areas between lowlands and mountains, even in eastern Shikoku, which is not yet colonized. Therefore, before the magpie expands over Shikoku and becomes abundant, it will be necessary to further assess potential magpie impacts on local native species, develop efficient methods to capture the magpie, and establish a monitoring scheme in priority areas to block magpie expansion. Our approach using a combination of playback surveys and models considering detectability has the potential for application in studies of other non-native bird species, as well as to support their appropriate management.

Terrestrial Ecology

Conservation Biology

Biological invasion

Exotic species

Hierarchical modeling

Imperfect detection

Interspecific interaction

Globally, over 200 species of birds breed outside their native ranges, negatively affecting local ecosystems, agriculture, and human health (Shirley & Kark 2009; Martin-Albarracin et al. 2015). However, research on the effects of non-native birds on ecosystems has been insufficient to date, and the impact differs depending on the ecological traits of the non-native bird species (Shirley & Kark 2009; Martin-Albarracin et al. 2015). Moreover, eradication of non-native birds becomes difficult when a long time has elapsed since colonization. Therefore, an urgent need exists to identify the factors that determine the distribution of each non-native bird species, make plans for preventing their expansion, and assess their impacts on native species (Evans et al. 2014). Establishing efficient survey and analytical methods that are applicable to a wide range of non-native bird species is necessary to meet those goals.

Accounting for imperfect detection in bird surveys is critical, as surveyors miss some individuals (Kéry 2010; Zuberogoitia et al. 2020). Playback surveys, in which the call of the focal species is broadcast and responses are noted, have been used to increase detectability (Bibby et al. 2000). As an analytical method that considers the detection process, the occupancy model has been commonly applied to detection/non-detection data obtained from multiple surveys at a given site (MacKenzie et al. 2002; Kéry 2010). This model enables estimation of species occupancy at each site, the detectability in each survey, and the effects of environmental and survey conditions on these probabilities. For nocturnal rare bird species with low detectability, some studies have used a combination of playback surveys and occupancy models to predict large-scale distributions while avoiding underestimation (Kawamura et al. 2016; Zuberogoitia et al. 2020). This approach may also be useful for assessing non-native bird distributions, especially in the early stage after colonization.

Furthermore, a recently developed multispecies occupancy model can estimate occupancy and detection probabilities, and their determinants, for individual species as well as interspecific interactions (i.e., co-occurrence or mutually exclusive occurrence patterns among species; Rota et al. 2016). The multispecies occupancy model has been applied to mammals to examine species interactions among non-native and native species (Kass et al. 2020; Twining et al. 2021). However, to our knowledge, no studies to date have applied the multispecies occupancy model to research on non-native bird species. In common field surveys of bird communities, distribution data for multiple species are obtained simultaneously, in a similar manner to the camera trap for mammals. Thus, the multispecies occupancy model may be a useful tool for assessing potential impacts of non-native bird species on multiple native species through predation or competition.

In Japan, more than 40 non-native bird species are naturalized (Eguchi and Amano 2004). The non-native Red-billed Blue Magpie (Urocissa erythrorhyncha; hereafter, magpie) has colonized Shikoku, southern Japan since 2000, and has expanded its range in forests (Sato et al. 2018). A magpie flock escaped from Tsushima Playland, a leisure park in Uwajima, Ehime (Matsuda 2023). Magpie distribution has likely expanded from that location, and one juvenile and two nestlings were detected in June 2015 and April 2016, respectively (Sato et al. 2018). The magpie is a species of the family Corvidae; its total body length is approximately 70 cm (Fig. S1; Brazil 2009; del Hoyo et al. 2009). It is naturally distributed from central China to Southeast Asia and preys on various organisms, including insects, frogs, bird eggs, and fledglings; it also consumes fruits (Brazil 2009; del Hoyo et al. 2009). Therefore, concern has arisen over negative effects of predation by the magpie on native species and impacts on agriculture (Sato et al. 2018).

We assessed the efficacy of playback surveys for detecting the magpie and revealed the distribution of this non-native species and its determinants using an occupancy model. We also examined its impacts on native bird species using a multispecies model. We predicted that playback of a specific series of magpie calls would enhance magpie detection probability; land use, topography, and distance from the escape point would explain the spatial distribution of magpie occupancy; and magpie presence might decrease the occupancy of some native bird species.

Study area and sites

This study was conducted in the Kochi and Ehime prefectures of western Shikoku. In our study area, magpies had been detected previously, and we anticipated that their populations were growing (Sato et al. 2018). Watari (2019) defined an ‘early stage of colonization’ as the period from the introduction to the establishment of non-native species, and we expected that magpie occupancy rate was lower in the far areas from the escape point because the magpie had already reached the fringe of the study area but did not seem to be common in the study area. In this area, the magpie has been reported to occur in evergreen broadleaved secondary forests, conifer plantations, and farmlands (Sato et al. 2018). In their native range, magpies occur in green spaces with anthropogenic openings and forests near human settlements (del Hoyo et al. 2009). The magpie may prefer landscapes containing both forests and open areas or warmer lowlands as habitats. Therefore, we selected survey points with widely varying land use (e.g., forest cover and proportion of natural forests among surrounding forests [index of local forest composition]) and topographic factors (elevation and Topographic Position Index [TPI]) using QGIS (https://qgis.org/ja/site/index.html), to cover most observation records over the past 20 years. We considered that elevation as a surrogate for temperature at the local scale (Kawamura et al. 2023).

For land use data, we used the High-Resolution Land Use and Land Cover Map (ver. 21.11) of the Japan Aerospace Exploration Agency (grid size: 10 m; period: 2018–2020; https://www.eorc.jaxa.jp/ALOS/jp/dataset/lulc_j.htm). Among 12 land use types, we treated two types of forests (deciduous and evergreen broadleaved forests) as natural forests. Evergreen conifer forests were treated as plantations. As the home range of the magpie was unknown, we calculated forest cover as a proportion of total land area within various radii from the sampling points (100, 200, 400, 600, 800, and 1,000 m), as an indicator of habitat extent. To examine the effects of local forest composition, the proportion of natural forest area to total forest area was calculated using 100 m as the detection radius for magpie.

We used elevation data provided by CGIS Japan and derived from a 10-m digital elevation model (http://cgisj.jp/download_type_list.php). We calculated the mean elevation within 100 m of each survey point. TPI was calculated at a scale of 250 m × 250 m as the difference between the elevation of a survey point and the mean elevation of its surroundings, with a negative value indicating a valley floor or concave area and a positive value indicating a ridge or convex landform (Zellweger et al. 2013). The scale of the TPI was suggested by Yonekura et al. (2001) to clearly represent a wide range of topographic features in Japan. We aimed to make each point spatially independent by setting a minimum distance between points of 2 km. Finally, we selected 42 survey points (six survey points on roads in each of seven municipalities: Fig. 1).

Bird survey

To examine the responsiveness of the magpie to playback of their calls, we first conducted a preliminary indoor survey at the Nature Center of Shimanto Fairy Pitta Forest (Shimanto town, Kochi Prefecture). We selected five types of calls downloaded from the xeno-canto database and a call recorded in our study area, to cover a wide range of magpie call types (see sonograms in Fig. S2) and areas because the function of each call type remain unknown and the subspecies was not specified for the population in Shikoku (Sato et al. 2018). Both two captive individuals responded to playback of these calls by jumping or calling back. At six sites where magpies were detected at least once before the survey, we broadcast these six call types as a preliminary field survey in early April 2022. At two sites, magpies called back and approached the speaker.

We created a 3-min sound file for each of the six types of calls and rearranged them to create six 18-min playback files (ABCDEF, ABEFCD, CDABEF, CDEFAB, EFABCD, and EFCDAB). A single surveyor (HM) played back these files using a speaker (SoundLink Revolve +; Bose, Framingham, MA, USA) connected to a player (PCM-A10; Sony, Tokyo, Japan) with the volume set to maximum, and a magpie decoy was hung from a tripod at each point. The number of magpie individuals calling back within 100 m of the speaker was recorded at 3-min intervals (i.e., we obtained six observations in a single visit).

We aimed to assess the effects of magpie occupancy on as many native bird species as possible, including both rare and common species. We presumed that smaller bird species than the magpie can be prey for the magpie and that species with unique and easily identifiable songs, mainly songbird species, were suitable for this assessment. Therefore, when we conducted the playback survey for the magpie, we also recorded the detection/non-detection of eight target species: Narcissus Flycatcher (Ficedula narcissina), Japanese Bush Warbler (Cettia diphone), Japanese Tit (Parus minor), and Varied Tit (Poecile varius) as common species, and Fairy Pitta (Pitta nympha), Japanese Paradise Flycatcher (Terpsiphone atrocaudata), Eastern Crowned Warbler (Phylloscopus coronatus), and Ruddy Kingfisher (Halcyon coromanda) as rare species in this area (Ueta et al. 2011). Although we did not record detection/non-detection data during each 3-min interval for these small birds, we considered all species detected during an 18-min recording period within a single visit to be detected during all six 3-min sampling intervals, as only actively singing individuals were detected. We performed surveys once per month from April to August 2022 (four to five visits to each point). Surveys were conducted between dawn and 9:00 and between 16:00 and dusk.

Statistical analysis

Occupancy model for the magpie

We used the unmarked package (v1.2.5; Fiske & Chandler 2011) and ubms package (v1.2.2; Kellner et al. 2021) in R (v4.2.2; R Core Team 2022). To reveal the factors determining the magpie distribution and predict habitat suitability across Shikoku, we used an occupancy model (link function = logit) with magpie detection/non-detection in each 3-min sampling period as the response variable (MacKenzie et al. 2002). Application of the occupancy model to this data arrangement enabled us to examine the playback effects of each call type on magpie detection probability. Implementing the occupancy model increases model computation time because the relationships between occupancy of the target species and environmental factors (the ecological process), and between detectability and survey conditions (the observation process), are estimated simultaneously (Goldstein and de Valpine 2022). To select the most parsimonious model for prediction (final model) with less computation time, we used a two-step model selection process in which explanatory variables for the ecological process were selected first, followed by those for the observation process, using the “occu” function in the unmarked package (based on a maximum likelihood estimation).

First, to select explanatory variables representing ecological processes, we considered a quadratic term or logarithmic transformation of each environmental factor (forest cover, natural forest proportion, elevation, TPI, and distance from the escape point) after standardization. Correlations among these factors were not high (|r| < 0.62). In this procedure, we included and fixed explanatory variables, i.e., survey date, its quadratic term, survey time and six types of playback calls for the observation process (see details in the next paragraph). For forest cover and natural forest proportion, we constructed three models including only the linear term, only the quadratic term, and both the linear and quadratic terms as explanatory variables. For the elevation and distance from the escape point, we constructed two models using the linear terms or logarithmically transformed values. The significance of explanatory variables in the model with the lowest Akaike information criterion value was determined at the 5% level for each environmental factor, and all significant factors were used as explanatory variables for ecological process evaluation, as described in the following steps.

Second, we selected explanatory variables for the observation process using a method similar to that used for ecological process. We included survey date, survey time, and six playback call types in the analysis. Survey date was defined as the number of days elapsed (i.e., 1 April = 1), which was standardized prior to the analysis. Survey time was a categorical variable (morning or evening). Each of the six types of playback calls was included as a categorical explanatory variable consisting of a binary number, where 0 indicated before playback and 1 indicated during/after playback. For example, in a playback file containing six types of calls, in the order ABCDEF, we described explanatory variables for the six 3-min intervals as follows: A, 111111; B, 011111; C, 001111; D, 000111; E, 000011; F, 000001. One of the six types of calls was played back during each 3-min playback interval, and all six playback call types were played back in every 18-minute visit. The magpie detectability during each visit was expected to differ depending on the playback order of six types of calls. We therefore included six playback call types in every model and the intercept of the detection model represents the hypothetical detection probability prior to any playback; thus, magpie responses before and after playback of each call type could be compared using this modeling framework. Considering the low number of magpies detected, we assessed the significance of each explanatory variable at the 10% level and thus included marginally significant factors in the observation process.

Finally, to consider the effects of temporal pseudoreplication of multiple observations within each visit, we constructed the final model using the random intercept in the detection model for each visit in each site with the “stan_occu” function (chains = 3, iter = 10000) in the ubms package (Bayesian framework, computation time: 9 min). We used the default setting for prior distributions. We used this model to predict the distribution of suitable habitats across Shikoku. The MacKenzie–Bailey chi-square goodness-of-fit test was performed using the “gof” function of ubms.

Multispecies occupancy model

To evaluate the impact of magpie presence on occupancy of native small bird species, we used a multispecies occupancy model (“occuMulti” function of the unmarked R package: Fiske & Chandler 2011; Rota et al. 2016). Of the eight native species that we surveyed, four (Fairy Pitta, Japanese Paradise Flycatcher, Eastern Crowned Warbler, and Ruddy Kingfisher) were excluded from the analysis because they occurred at fewer than three points. We used the detection/non-detection data for the magpie and four native species (Narcissus Flycatcher, Japanese Tit, Varied Tit, and Japanese Bush Warbler) in each 3-min sampling interval as the response variables, although the detection/non-detection for native species were the same among six 3-min samplings in a single visit (see details in Bird survey). This modeling approach using repeated 3-min detection statuses of five species in a single model enabled us to estimate the interaction between magpie occupancy and that of multiple native bird species (i.e., magpie:[Narcissus Flycatcher], magpie:[Japanese Tit], magpie:[Varied Tit], and magpie:[Japanese Bush Warbler]), simultaneously, while considering playback effects on the detection probability of the magpie. For the magpie, the explanatory variables for the ecological and observation process were the same as in the final model. Similarly, for each native species, we used single-species occupancy models, with explanatory variables selected based on the quadratic term or logarithmic transformation of each environmental factor (forest cover, natural forest proportion, elevation, and TPI) after standardization. The explanatory variables for the observation process were fixed to be survey date, its quadratic term, and survey time. We assumed that magpie presence affects only the intercept for native species occupancy (i.e., mean occupancy).

Mapping the potential distribution of the Red-billed Blue Magpie

In order to predict the potential future expansion, we mapped the habitat suitability for the magpie at 200-m resolution across Shikoku using the final model and the environmental data that were used for survey point selection. First, we calculated the mean elevation within a 100-m radius from the center of each 10 m × 10 m cell (i.e., 317 cells were included in this range) using a 10-m digital elevation model. Then, we calculated forest cover within a 600-m radius from the center of each 10 m × 10 m cell (i.e., 11,289 cells were included in this range) using the 10-m land use map. To create data at a 200-m resolution from those at a 10-m resolution, we averaged the elevation and forest cover data across 400 10 m × 10 m cells within each 200-m grid. Finally, we mapped the expected occupancy probability by substituting these 200-m resolution environmental data into the final model, although with extrapolated ranges of the forest cover and elevation.

To evaluate the performance of the final model, we used 52 observation records collected over the past 20 years, separately from our surveys (hereafter, observational data: Tanioka and Hamada, unpublished data). The observational data were collected by bird watchers with bird identification skills in a manner differing from that of our surveys (i.e., occasional observation without playback). All but five records were collected before our surveys. Although observation times and detection radii were not recorded, it is possible that magpies have continued to inhabit the area because magpies were observed multiple times in areas within 10 km of each observation site. We compared the predicted occupancy probability, elevation, and forest cover of these observation locations (i.e., presence only data) with those of “detected” and “non-detected” points of our field survey, and examined the features of observation sites despite low predicted occupancy probability.

Factors affecting magpie occupancy and detectability

We detected magpies during 17 of 183 visits and at 13 of 42 points (a maximum of three magpies were detected at individual points). All individuals were detected with calls. Multiple visits with detections occurred at four points. Forest cover within 600 m, its quadratic term, and elevation were selected as explanatory variables for the ecological process, while survey date, its quadratic term, survey time, and the playback of six types of calls were selected as explanatory variables for the detection process (Table 1, Table S1). For the ecological process of occupancy model, parameter estimates were similar between models that did not consider and considered pseudoreplication derived from multiple observations at each visit (Table 1, Fig. 2). The magpie occupancy probability was higher in areas of lower elevation and tended to be high in areas with moderate forest cover, peaking where forest cover was 76% (Fig. 2b). Occupancy was predicted to approach 0 in areas with elevation > 400 m (Fig. 2d).

By contrast, the expected detection probability (i.e., the intercept estimate) was lower in the model that considered temporal pseudoreplication than in the model that did not (Table 1, Fig. 3), although the direction of the estimated coefficients for the explanatory variables (i.e., positive or negative) and the magnitude relation of the playback effect of each call type did not differ between these two models (Table 1, Fig. 3). The final model with random visit effects predicted that the detection probability peaked on 18 June and decreased late in the season (Fig. 3b). The detection probability tended to be lower in the evening than in the morning (Fig. 3d). Playback effects were estimated to be highest for the D-type call, followed by C- and F-type calls, because detection probability tended to increase after these playback types (Fig. 3f). On the morning of 18 June, the predicted detection probability during each 3-min sampling interval was initially very low (~ 0), and the cumulative detection probability was 17–29% over the 18-min visit (Fig. 3h). The goodness-of-fit test for the final model with random effects indicated no significant difference between the prediction and collected data (posterior predictive p = 0.95).

Effects of environmental factors and magpie presence on occupancy of native bird species

Detection probability decreased in the late season for all native species and that of the Japanese Tit was high in the morning (Fig. S3, Table S2). Occupancy of the Narcissus Flycatcher tended to be higher in areas with moderate forest cover (72%), while that of the Japanese Tit increased with increasing TPI (i.e., at ridges) and natural forest proportion (Fig. 4, Tables S3 and S4). We detected no effects of environmental factors on the Varied Tit or Japanese Bush Warbler (Fig. 4, Tables S3 and S4). No significant impact of magpie presence was detected on the occupancy of these native species (Fig. 4, Table S3).

Mapping predicted suitable magpie habitats across Shikoku Island

Areas of high predicted occupancy probability were located near the coast, between lowlands with abundant farmlands and forested mountains, indicating that potentially suitable habitats are abundant even in eastern Shikoku, where magpies do not currently occur (Fig. 5a). At the locations of previous observations, the predicted occupancy was < 0.10 at 20 of 52 sites. Its median value for previous observations from April to August (to match our survey season) was approximately 0.13, similar to our non-detected points (0.14: Fig. S4a). These previous observations in areas of low predicted occupancy occurred mainly in mountainous areas at high elevations (250–600 m) with high forest cover or in lowlands with low forest cover (Fig. 5b, Fig. S4b, c).

Efficient survey methods for the magpie

Our results showed that surveying with playback of the D-type call in the early morning from late May to early July was most effective. The detection probability over 18 min (per visit) was predicted to reach 17–29% when each call type was played back during each 3-min sampling interval. Although the detection probability may be enhanced by a playback survey in the morning during the appropriate season, the expected detection probability was lower when we considered temporal pseudoreplication, suggesting that detectability may vary greatly depending on individual-level factors, such as location of magpie individuals on the survey visit and breeding stage on each visit. Thus, each site would be surveyed multiple times to compensate for imperfect detection. The breeding season of the magpie has been reported as between March and August in China (Guo et al. 2022) and may be similar in our study area.

Effects of environmental factors on magpie occupancy

Magpie occupancy was higher in areas of low elevation with moderate forest cover (76%). In Southeast Asia, the magpie has been reported to occur in environments containing both forests and open lands, such as forest edges, green spaces, and forests near human settlements (McKinnon & Phillipps 2000, del Hoyo et al. 2009). The main non-forest land use in our study area was farmland, and magpies likely preferred forested landscapes with such anthropogenic openings. Among several candidates, 600 m was identified as the most effective scale for forest cover. Although no data for the home range or territory size of the magpie are available, given the relatively large body size of the magpie, daily foraging movement may be conducted at this scale (Jetz et al. 2004). In contrast, natural forest proportion within 100 m did not significantly affect magpie occupancy. In their native range, broadleaved forests are considered the main magpie habitats, but conifer or Lophostemon confertus plantations are also used (Kwok & Corlett 2000, del Hoyo et al. 2009, Robson 2015). Thus, the tree species preference of the magpie may be low at this scale. In central China, most nests were built on a bamboo species (Phyllostachys sulphurea) and an oak species (Quercus acutissima) (Guo et al. 2022). Phyllostachys spp. have been expanding and Quercus spp. are widely distributed in lowlands of southern Japan (Tanaka & Matsui 2007, Someya et al. 2010). In our area, the nesting behavior of the magpie was observed at a broadleaved shrub (Tanioka & Fukuda 2023). Thus, nesting resources may be abundant for magpies.

The negative effect of elevation may reflect the difference in temperature. Magpies are distributed mainly below 1,500 m in eastern China and Southeast Asia, and no stable populations were observed in highlands (McKinnon & Phillipps 2000, Robson 2015). Similarly, in Shikoku, highlands with lower temperatures and larger temperature fluctuations may be unsuitable for breeding in the magpie, which is naturally distributed in warmer areas. Moreover, considering the non-significant effect of distance from the escape point, we concluded that the early stage of magpie colonization has already passed for our study area.

Effects of magpie presence on native species occupancy

We did not find that occupancy of the four native species decreased due to the presence of magpies. This result suggests that each magpie individual has not had devastating negative impacts on habitat occupancy of these native species. However, negative impacts on native bird species may be detected when magpies are abundant. Therefore, before the magpie becomes prevalent, it is necessary to assess magpie impacts on native species using other measures. Candidate measures include the abundance, reproductive success, or multi-year occupancy status data for native species. Moreover, we found no survey points occupied by the Fairy Pitta, which is classified as Vulnerable in the IUCN Red List (BirdLife International 2023) and is thought to be negatively affected by magpie predation, despite the lack of direct evidence (Japan Broadcasting Corporation 2023). Although it would be difficult to ensure the number of detection sites required to obtain reasonable occurrence estimates of the magpie and rare species, efforts to estimate magpie impacts on rare species are needed. Applying playback surveys to both the magpie and rare species, and using the multispecies occupancy model, may enable us to overcome this challenge.

Suitable magpie habitat distribution across Shikoku and management implications

Mapping of predicted magpie occupancy demonstrates that potentially suitable habitats are widely distributed in near-coast areas located between lowlands and mountains in western Shikoku, which has been colonized by the magpie, and in eastern parts of this area, which is not yet colonized. The predicted occupancy was low in some locations where magpies have been observed in the past 20 years. The magpie may have high dispersal ability because it has long wings for its light weight (Tan et al. 2021), and may also occur in fully forested areas at high elevations or in lowlands with low forest cover when they perform natal dispersal or occasional seasonal movement (del Hoyo et al. 2009), or where local environmental conditions are particularly suitable. Surveys at high-elevation sites or in the dispersal/wintering seasons would be useful to explain the low predicted occupancy in areas where magpies were previously observed.

The Eurasian Magpie (Pica pica) has colonized Kyushu, southern Japan, as a non-native species of Corvidae; however, its distribution has not expanded widely. This lack of expansion has been explained as arising from forested mountains forming a barrier to the dispersal of this species, which prefers open lands, including farmlands (Eguchi 2016). Although the Red-billed Blue Magpie distribution in inland Shikoku may be restricted by areas of high elevation and forest cover, further expansion is a cause for concern because this species prefers more forested environments than the Eurasian Magpie and its activities have been observed in areas with low predicted occupancy. In Shikoku, monitoring of inland areas and southeastern areas adjacent to areas with high predicted occupancy is necessary. To prevent the expansion of this species across Japan, monitoring of the Sadamisaki Peninsula near Kyusyu and the Takanawa Peninsula near Honshu is critical (Fig. 5a), and rapid management including capture and removal is required when the magpie is detected. Therefore, we should develop a strategy for controlling further expansion of the magpie through establishment of an efficient capture method before the magpie becomes more prevalent.

ACKOWLEGMENTS

We greatly thank the following people for their help: H. Kuroda for providing sound sources of the Red-billed Blue Magpie call; K. Yamaura for providing the magpie decoy; A. Tomiyama for providing environmental data and survey assistance; members of the Plant Ecology Laboratory, Faculty of Science and Technology, Kochi University for data confirmation; and T. Nakamura, K. Kawakami, Y. Watari, M. Senzaki and anonymous reviewers for useful comments. This study was supported by the Environment Research and Technology Development Fund provided by Ministry of the Environment of Japan [JPMEERF20234002].

CONFLICT OF INTEREST

No potential conflict of interest was reported by the authors.

Bibby CJ, Burgess ND, Hill DA, Mustoe S (2000) Bird Census Techniques. Academic, London
BirdLife International (2023) Species factsheet: Pitta nympha. Downloaded from http://www.birdlife.orgon02/05/2023
Brazil M (2009) Birds of East Asia: China, Taiwan, Korea, Japan, and Russia. Christopher Helm, London, pp 306–308
del Hoyo J, Elliott A, Christie D (2009) Handbook of the Birds of the World, vol 14. Lynx Edicions, Barcelona, Spain, pp 594–596
Eguchi K (2016) The Eurasian Magpie. Jpn J Ornithol 65:5–30 (in Japanese with English abstract)
Eguchi K, Amano HE (2004) Spread of exotic birds in Japan. Ornithol Sci 3:3–11
Evans T, Kumschick S, Dyer E, Blackburn T (2014) Comparing determinants of alien bird impacts across two continents: implications for risk assessment and management. Ecol Evol 4:2957–2967
Fiske I, Chandler R (2011) Unmarked: an R package for fitting hierarchical models of wildlife occurrence and abundance. J Stat Softw 43:1–23
Goldstein BR, de Valpine P (2022) Comparing N-mixture models and GLMMs for relative abundance estimation in a citizen science dataset. Sci Rep 12:12276
Guo W, Hu Z, Lin B, Kuang Y, Cao H, Wang C (2022) Nest site selection and breeding ecology of the Red-billed Blue Magpie Urocissa erythrorhyncha in central China. Anim Biol 72:153–164
Japan Broadcasting Corporation (2023) Kochi News Web Sanjaku-taisaku ni gyousei no kyouryoku wo Yairocyou no hogo-dantai (Fairy Pitta Conservation Group requests governmental supports for managing the Red-billed Blue Magpie). (in Japanese). https://www3.nhk.or.jp/lnews/kochi/20230110/8010016758.html
Jetz W, Carbone C, Fulford J, Brown JH (2004) The scaling of animal space use. Science 306:266–268
Kass JM, Tingley MW, Tetsuya T, Koike F (2020) Co-occurrence of invasive and native carnivorans affects occupancy patterns across environmental gradients. Biol Invasions 22:2251–2266
Kawamura K, Yamaura Y, Senzaki M, Yabuhara Y, Akasaka T, Nakamura F (2016) Effects of land use and climate on the distribution of the Jungle Nightjar Caprimulgus indicus in Hokkaido, northern Japan. Ornithol Sci 15:203–212
Kawamura K, Yamaura Y, Nakamura F (2023) Early successional habitats created through plantation harvesting benefit the Gray Nightjar (Caprimulgus jotaka): An 8-year survey in central Hokkaido, northern Japan. J Res 28:289–296
Kellner KF, Fowler NL, Petroelje TR, Kautz TM, Beyer DE Jr, Belant JL (2021) ubms: An R package for fitting hierarchical occupancy and N-mixture abundance models in a Bayesian framework. Methods Ecol Evol 13:577–584
Kéry M (2010) Introduction to WinBUGS for Ecologists: A Bayesian Approach to Regression, ANOVA, Mixed Models and Related Analyses. Academic, pp 237–246
Kwok HK, Corlett RT (2000) The bird communities of a natural secondary forest and a Lophostemon confertus plantation in Hong Kong, South China. Ecol Manag 130:227–234
MacKenzie DI, Nichols JD, Lachman GB, Droege S, Royle JA, Langtimm CA (2002) Estimating site occupancy rates when detection probabilities are less than one. Ecology 83:2248–2255
Martin-Albarracin VL, Amico GC, Simberloff D, Nuñez MA (2015) Impact of non-native birds on native ecosystems: a global analysis. PLoS ONE 10:e0143070
Matsuda H (2023) Sannjaku ni tsuite (About the Red-billed Blue Magpie). Komadori (Bulletin of Ehime Branch of Wild Bird Society of Japan) 273: 2 (in Japanese)
McKinnon JR, Phillipps K (2000) A Field Guide to the Birds of China. Oxford University Press, Oxford, UK, pp 264–265
R Core Team (2022) R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. https://www.R-project.org/
Robson C (2015) Birds of South-east Asia. Concise Edition. Christopher Helm, London, p 166
Rota CT, Ferreira MAR, Kays RW, Forrester TD, Kalies EL, McShea WJ et al (2016) A multispecies occupancy model for two or more interacting species. Methods Ecol Evol 7:1164–1173
Sato S, Hamada T, Tanioka H (2018) Naturalization of the Red-billed Blue Magpie in western Shikoku Island, Japan. Bird Res 14:S1–S5 (in Japanese with English abstract)
Shirley SM, Kark S (2009) The role of species traits and taxonomic patterns in alien bird impacts. Glob Ecol Biogeogr 18:450–459
Someya T, Takemura S, Miyamoto S, Kamada M (2010) Predictions of bamboo forest distribution and associated environmental factors using natural environmental information GIS and digital national land information in Japan. Landsc Ecol Manag 15:41–54 (in Japanese with English abstract)
Tanaka N, Matsui T (2007) -) PRDB: Phytosociological Relevé Database, Forestry and Forest Products Research Institute. URL: http://www.ffpri.affrc.go.jp/labs/prdb/index.html
Tanioka H, Fukuda M (2023) An observation of nest site of the Red-billed blue magpie Urocissa erythrorhyncha in Ozu City, Ehime Prefecture, Japan. Bull Shikoku Inst Nat Hist 16:67–71 (in Japanese with English abstract)
Tan X, Yang X, Chen C, Wang AY (2021) Nestedness and underlying processes of bird assemblages in Nanjing urban parks. Curr Zool 67:383–392
Twining JP, Montgomery WI, Tosh DG (2021) Declining invasive grey squirrel populations may persist in refugia as native predator recovery reverses squirrel species replacement. J Appl Ecol 58:248–260
Ueta M, Fukui A, Yamaura Y, Yamamoto Y (2011) Status of forest birds in Japan. Jpn J Ornithol 60:19–34 (in Japanese with English abstract)
Watari Y (2019) Roadmap and checklist of invasive species management: Learning from the mongoose eradication project on Amami-Oshima. Jpn J Ornithol 68:263–272 (in Japanese with English abstract)
Yonekura N, Kaizuka S, Nogami M, Chinzei K (2001) Nihon no chikei 1: sosetsu (Topography of Japan 1: Review). University of Tokyo, Tokyo. (in Japanese)
Zellweger F, Braunisch V, Baltensweiler A, Bollmann K (2013) Remotely sensed forest structural complexity predicts multi species occurrence at the landscape scale. Ecol Manag 307:303–312
Zuberogoitia I, Martínez JE, González-Oreja JA, de Buitrago CG, Belamendia G, Zabala J et al (2020) Maximizing detection probability for effective large‐scale nocturnal bird monitoring. Divers Distrib 26:1034–1050

Table 1

Parameter estimates of the final model without random visit effects, derived using the “occu” function in the unmarked package (a), and the final model with random visit effects, derived using the “stan_occu” function in the ubms package (b). Forest600, forest cover within 600 m of the survey point; Elev, average elevation within 100 m of the survey point; Call.A–F, playback of each call type; Date, survey date; Time (evening), survey time in the evening (reference category: morning); 95% CI.l and 95% CI.u, lower and upper 95% credible interval limits, respectively. (b) Rhat = 1 for all variables.
	(a) unmarked without random effects				(b) ubms with random effects
Occupancy
Variables	Coefficient	SE	z	Wald p	Coefficient	95%CI.l	95%CI.u
Intercept	-0.17	0.76	-0.23	0.82	0.15	-1.21	1.68
Forest600	-1.41	0.90	-1.57	0.12	-1.15	-3.02	0.41
Forest600²	-2.87	1.41	-2.03	0.04	-1.93	-4.10	0.13
Elev	-3.50	1.38	-2.54	0.01	-2.91	-5.70	-0.96
Detection
Variables	Coefficient	SE	z	Wald p	Coefficient	95%CI.l	95%CI.u
Intercept	-2.23	0.51	-4.38	0.00	-8.54	-13.10	-4.62
Call.A	0.30	0.48	0.62	0.53	0.59	-1.02	2.34
Call.B	0.22	0.43	0.51	0.61	0.44	-1.08	2.04
Call.C	0.47	0.52	0.91	0.36	1.82	0.09	3.84
Call.D	0.75	0.43	1.75	0.08	1.95	0.41	3.68
Call.E	-0.08	0.50	-0.16	0.87	0.02	-1.61	1.71
Call.F	0.41	0.45	0.92	0.36	1.78	0.12	3.68
Date	-0.54	0.18	-3.02	0.00	-1.14	-2.92	0.57
Date²	-0.43	0.17	-2.51	0.01	-1.74	-3.83	0.02
Time (evening)	-0.79	0.30	-2.65	0.01	-1.82	-4.88	0.70
					Random visit effect
					SD
					6.12

The authors declare no competing interests.

SI240419submitclean.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Non-native Red-billed Blue Magpie Urocissa erythrorhyncha expanded in lowlands with moderate forest cover, with no significant impact on native common bird occupancy, in Shikoku, southern Japan

Status:

Version 1

Abstract

Figures

INTRODUCTION

MATERIALS AND METHODS

RESULTS

DISCUSSION

Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Version 1