Introduction
A large number of long-distance migratory songbirds exhibit decreasing population trends in Eurasia (Sanderson et al. Reference Sanderson, Donald, Pain, Burfield and van Bommel2006, Laaksonen and Lehikoinen Reference Laaksonen and Lehikoinen2013, Vickery et al. Reference Vickery, Ewing, Smith, Pain, Bairlain, Skorpilova and Gregory2014). Amongst the 170 long-distance migratory songbird species using the East Asian flyway, for example, one third show declining trends and 12% are classified as threatened or near-threatened (Yong et al. Reference Yong, Liu, Low, Española, Choi and Kawakami2015). Whilst the drivers of population declines are comparatively well studied in the Afro-Palearctic migratory system (Vickery et al. Reference Vickery, Ewing, Smith, Pain, Bairlain, Skorpilova and Gregory2014), such knowledge is scarce for the East Asian flyway (Yong et al. Reference Yong, Liu, Low, Española, Choi and Kawakami2015). Asia has seen a strong intensification in agriculture during recent decades, and witnessed increasing urbanisation and industrialisation (Alauddin and Quiggin Reference Alauddin and Quiggin2008). Anthropogenic and climate-related stressors in the form of logging, draining, and forest fires have also increased (Goldammer and Furyaev Reference Goldammer, Furyaev, Goldammer and Furyaev1996, Achard et al. Reference Achard, Molliconi, Stibig, Aksenov, Laestadius, Li, Popatov and Yaroshenko2006) as has trapping of songbirds for consumption and religious merit release (Gilbert et al. Reference Gilbert, Sokha, Joyner, Thomson and Poole2012, Townsend Reference Townsend2015, Yong et al. Reference Yong, Liu, Low, Española, Choi and Kawakami2015). Stronger environmental pressures on breeding and wintering habitats and increased levels of persecution have therefore emerged as serious threats to long-distance migratory songbird populations in East Asia.
Habitat loss and persecution are considered as threats to rare and range-restricted species, but the scale of these factors may now have reached levels whereby even superabundant and wide-ranging species are becoming threatened. For example, data for Yellow-breasted Bunting Emberiza aureola, compiled across its distribution in Eurasia, showed that the global population of this once very abundant species, virtually collapsed over a time period of only 25 years (Kamp et al. Reference Kamp, Oppel, Ananin, Durnev, Gashev, Hölzel, Mischenko, Pessa, Smirenski, Strelnikov, Timonen, Wolanska and Chan2015). The Rustic Bunting Emberiza rustica shares many ecological traits with the Yellow-breasted Bunting. For example, both have very large and almost overlapping breeding ranges, being superabundant long-distance migrants, and use similar migration routes (Byers et al. Reference Byers, Olsson and Curson2013). There are some signals of a strong population decline also in the Rustic Bunting (BirdLife International 2015a), but the details and causes of the decline remain unclear.
The distribution of the Rustic Bunting spans roughly 170 degrees of longitude, across Eurasia from Fennoscandia in the west to the Kamchatka peninsula in the east (Cramp and Perrins Reference Cramp and Perrins1994). The breeding range, which is estimated at 218 Mha (BirdLife International 2015a), is intimately linked to the boreal forest, wherein the typical breeding habitat is wet coniferous forest with birch Betula spp. and willow Salix spp. growth along slow-flowing water (Öhrn Reference Öhrn and Olsson1963, Pulliainen and Saari Reference Pulliainen and Saari1989, Kretchmar Reference Kretchmar2000). Current evidence suggests that the winter range of the Rustic Bunting is confined to East Asia comprising Japan, the Korean Peninsula, and eastern and central mainland China north of the tropic of Cancer (Byers et al. Reference Byers, Olsson and Curson2013). For the bulk of the population, autumn migration follows forested regions eastwards, turning south in Asia east of Mongolia (Cramp and Perrins Reference Cramp and Perrins1994, Byers et al. Reference Byers, Olsson and Curson2013). While just a few birds migrate through or winter in Kazakhstan (Berezovikov and Levinskiy Reference Berezovikov and Levinskiy2008), the Rustic Bunting is one of the most common species during migration in the Russian Far East (Averin et al. Reference Averin, Kalinin, Malikova, Osipov, Rybzova and Strelzov2012, Heim et al. Reference Heim, Smirenski, Siegmund and Eidam2012, Heim and Smirenski Reference Heim and Smirenski2013). Birds from Kamchatka (sometimes regarded as a separate subspecies latifascia) are believed to migrate south-west across the Sea of Okhotsk or south to Japan across the Sea of Japan (Valchuk et al. Reference Valchuk, Yuasa and Morosova2005). The spring migration in the Rustic Bunting appears to mirror the autumn migration, but the details of its migratory patterns including stopover sites are inadequately known.
During the breeding season, the diet of Rustic Bunting mostly consists of invertebrates, whereas it shifts to seeds, grasses, and other vegetable food in other seasons (Cramp and Perrins Reference Cramp and Perrins1994). A variety of semi-open habitats are used during stopovers and in the winter range, including woodlands, willow thickets, clearings, rice stubble, reed beds and riverbanks (Cramp and Perrins Reference Cramp and Perrins1994, Fujioka et al. Reference Fujioka, Lee, Kurechi and Yoshida2010, Yoo et al. Reference Yoo, Lee, Kim and Park2011, Reference Schäfer, Meffert and HeimSchäfer et al. in prep). At its winter feeding sites, the Rustic Bunting prefers wet grassy places with plenty of seeds and presence of dense shrubs or trees (Iijima Reference Iijima1973, Maeda Reference Maeda2001). Maeda (Reference Maeda1973) observed spatially segregated habitat use in the Rustic Bunting during winter, with flocks of several hundred birds feeding in rice paddies during daytime and roosting in loosely dispersed groups in shrubby or grassy patches in woodland areas up to 1 km from the feeding sites.
The Rustic Bunting is classified as a species of ‘Least Concern’ (LC) on the IUCN Red List, i.e. it is not considered globally threatened (BirdlLife International 2012). However, it has recently been upgraded to ‘Vulnerable’ (VU) on the European Red List (BirdLife International 2015b). Population trends have been estimated for Fennoscandia (Dale and Hansen Reference Dale and Hansen2013) and Europe (BirdLife International 2015a) but not for the whole distribution range. Trend assessments and expert opinions suggest that the European breeding population of Rustic Bunting has declined by 30–49% over the last 10 years in Europe, but only around one-fifth of the species’ global breeding range is within Europe (Symes Reference Symes2015). For large parts of the breeding range, population trends remain unknown. While earlier population assessments have focused on restricted parts of the distribution range and included a limited amount of time series data, we here present the first range-wide assessment of population trends in the Rustic Bunting based on several short and long term time series.
We compiled Rustic Bunting data from different countries to determine the range-wide rates of population change and hence assess the species’s global status. We also present an overview of possible drivers underlying the decline and propose key research activities to help overcome the current lack of knowledge about the bunting’s ecology which hampers effective conservation policies.
Material and methods
Data collection
We collated time series data on the Rustic Bunting from countries in the breeding range (Norway, Sweden and Finland) and from stop-over sites during migration (Japan and north-eastern China). Standardised breeding survey data were preferentially used, but since such data are scarce we also included ringing data. In case of ringing data we only included standardised time series data, i.e. data controlled for capture effort. We used nationwide breeding survey data from Sweden and Finland and breeding territory count data for the very small Norwegian population (Green and Lindström Reference Green and Lindström2015a, Hansen Reference Hansen2015, Aleksi Lehikoinen pers. comm.). In Sweden, breeding survey counts have been carried out on fixed routes systematically distributed across Sweden since 1998 (Green and Lindström Reference Green and Lindström2015a). However, the number of surveyed fixed routes within the Rustic Bunting’s breeding range in northern Sweden was low during the first years. Therefore, we used Swedish breeding bird data only from the last 10 years (2005–2014). To increase the length of the overall time series we complemented the Swedish breeding survey data with Rustic Bunting data from ringing stations. The Japanese and Chinese data consisted of the number of Rustic Buntings ringed during autumn migration at ringing stations. In total, we compiled seven time series of data of up to 30 years in length, see Table 1 and Appendix S1 in the online supplementary materials for details.
Table 1. TRIM-estimated annual decline rates for national data sets 1985–2014. P-levels denote significance of deviation of slope values from 1 (= no change).
Statistical analysis
For the Swedish and Finnish breeding bird survey data, we used TRIM-estimated (Pannekoek and van Strien Reference Pannekoek and van Strien2013) yearly population indices and linear trends provided to us by the national programme organisers. To make the ringing data comparable with these breeding survey data, we also used TRIM (ver. 3.54) to calculate indices and trends. In the case of the Norwegian breeding survey data we applied TRIM to the numbers of occupied territories 2008–2014 (Hansen Reference Hansen2015; Table 1, Figure 1b). The widespread TRIM (TRends and Indices for Monitoring data) software uses log-linear models with a Poisson error distribution and allows for missing counts. Yearly population indices were scaled (indexed) to 1 for the starting year. We employed models that controlled for over-dispersion and serial correlation, and used trend estimates based on the imputed population indices (to account for missing values) as recommended by Pannekoek and van Strien (Reference Pannekoek and van Strien2013). Rangewide trends were estimated by averaging the national time-series trends. Trends were estimated for 30 years (1985–2014, long-term) and 10 years (2005–2014, short-term), respectively. Missing values amounted to 2% (n = 2) and 7% (n = 5) of the data points in the long- and short-term datasets, respectively. The small population in Norway, where the Rustic Bunting is dwindling and is estimated currently at less than 10 pairs (Hansen Reference Hansen2015), may have a disproportionate impact on the range-wide trend estimates. Therefore, we analysed the time-series data also without the Norwegian data.
Figure 1. TRIM estimated yearly indices of Rustic Bunting numbers in long-term (a) and short-term (b) data series.
Results
The TRIM-estimated yearly indices of the national data series revealed consistent population declines across the range (Table 1, Figure 1). For the long-term data (1985–2014), the average range-wide annual decline rate was estimated at 5.5% (± 1.0% 95% CI), which corresponds to an 82% (75–87%) reduction in population size over 30 years. The short-term (2005–2014) annual decline rate was 12.4% (± 8.6% 95% CI), corresponding to a 73% (32–91%) reduction over 10 years (Figure 1b). When we re-ran the analyses after excluding the Norwegian data, it yielded a 10-year annual decline rate of 8.9% (± 6.3% 95% CI), i.e. a 60% (23–81%) overall decline.
Discussion
Our analysis provides compelling evidence for a strong range-wide decline in Rustic Bunting populations in recent times. To put the magnitude of the population decline into perspective, we can compare it with two better-known examples of dramatic range-wide population declines within the same genus: the Yellow-breasted Bunting and the Ortolan Bunting E. hortulana. The decline in the Yellow-breasted Bunting was estimated at 84–95% over 34 years (Kamp et al. Reference Kamp, Oppel, Ananin, Durnev, Gashev, Hölzel, Mischenko, Pessa, Smirenski, Strelnikov, Timonen, Wolanska and Chan2015). An extrapolation of our 82% (75–87%) long-term (30 year) trend in the Rustic Bunting yields an 85% range-wide decline over 34 years. For the Ortolan Bunting, Vickery et al. (Reference Vickery, Ewing, Smith, Pain, Bairlain, Skorpilova and Gregory2014) reported an 84% decline over 30 years (1980–2009), which was the strongest population decline of all long-distance migratory species in the Afro-Palearctic migratory system they studied. In the case of the Yellow-breasted Bunting, the population collapse was accompanied by a strong eastward range contraction. Although such a dramatic global range contraction has not yet been observed in the Rustic Bunting, the range contraction towards north-east within Fennoscandia has been dramatic. On the western edge of the global distribution huge areas of habitat on the former Rustic Bunting distribution are now more or less empty of the species (Valkama et al. Reference Valkama, Vepsäläinen and Lehikoinen2011, Dale and Hansen Reference Dale and Hansen2013, Green and Lindström Reference Green and Lindström2015b).
The decline rate estimates have strong relevance for the ongoing discussion about the conservation status of the Rustic Bunting on the IUCN global Red List where it is currently listed in the ‘Least Concern’ (LC) category. Our range-wide estimate of a 61% (73% with the Norwegian data) population decline for 2005–2014 (a time span roughly corresponding to three generations in the Rustic Bunting; Symes Reference Symes2015) indicates a decline well over the 30% threshold required for classification as ‘Vulnerable’ (VU). Moreover, the observed population reduction exceeds the 50% threshold for the category ‘Endangered’ (EN) on the global Red List.
Our data samples are biased towards Fennoscandia. This is an unavoidable consequence of the fact that more data series are available from that part of the distribution range. From Russia, which holds the bulk of the breeding population, we found only scattered information on breeding densities (e.g. Rogacheva Reference Rogacheva1992 – Central Siberia, Kretchmar Reference Kretchmar2000 – noth-east Siberia), but no standardised data on long-term population development. However, numbers of Rustic Buntings captured during standardised ringing at Muraviovka Park in the Amur region of Russia show a decrease of > 90% in recent years (2012–2015) (Wieland Heim pers. obs.). The Gaofeng Bird Ringing Station located on the Chinese side of the Amur River also reports a strong recent decline (Table 1, Figure 1b). This suggests that the decline in numbers of Rustic Buntings passing through this part of East Asia during migration has been particularly strong. The negative trend expressed by the Fukushimagata Ringing Station data in Japan (Figure 1a) is also seen in the non-standardised national ringing sums of Rustic Bunting in that country for the same time period (Kiyoaki Ozaki pers. obs.). In South Korea no long-term standardised data series exist, but non-standardised short-term ringing data show that capture rates of Rustic Bunting (measured as the proportion of total ringing sums) are now about 10 times lower than they were in the mid-1960s (Chang-Yong Choi pers. obs.). Also in China there is a shortage of standardised ringing data (Bo Pettersson pers. comm.). We included the most comprehensive data series from China, from the Gaofeng Bird Ringing Station, Heilongjiang Province (Appendix S1) in our analysis, and it showed a strong decline in number of Rustic Buntings over the last 10 years (Table 1, Figure 1b). To conclude, both available standardised data and anecdotal information give unequivocal support for a strong range-wide population decline in Rustic Bunting over the last decades.
Possible drivers and knowledge gaps
Little is known about the magnitude of the impact of different threat factors and their interactions with regard to the population decline of long-distance migratory songbirds in the East Asian Flyway (Yong et al. Reference Yong, Liu, Low, Española, Choi and Kawakami2015). In Table 2 we have compiled threat factors related to changes in land use practices, natural processes and persecution that may affect population dynamics in the Rustic Bunting. To untangle the effects of different threats we need better data on population demographics. For example, using data modelling Kamp et al. (Reference Kamp, Oppel, Ananin, Durnev, Gashev, Hölzel, Mischenko, Pessa, Smirenski, Strelnikov, Timonen, Wolanska and Chan2015) concluded that an initial harvest rate of 2% followed by a 0.2% annual increase was sufficient to produce a population collapse in the Yellow-breasted Bunting over a time period of 34 years. Currently we do not know to what extent survival in the Rustic Bunting is affected by habitat change and persecution. Similarly, our knowledge about the effects of habitat change on Rustic Bunting productivity is limited. Poor nesting success coupled to habitat loss and degradation could be a population limiting factor, as has been suggested in case of Jankowski’s Bunting E. jankowskii (Jiang et al. Reference Jiang, Gao, Lei, Wan, Zhao and Wang2008).
Table 2. Changes in land use practices, natural processes and persecution, and their potential impacts on Rustic Bunting populations.
Assessment of the impacts of different threat factors is also hampered by our limited understanding of the connectivity between breeding and wintering ranges in the Rustic Bunting. Lightweight, light-level geolocators and stable isotope analysis provide new tools for studying such questions, but to date they have not been used to any larger degree on migratory songbirds in Asia (Yong et al. Reference Yong, Liu, Low, Española, Choi and Kawakami2015). Knowledge about the migration routes and the wintering areas of the Rustic Bunting will be vital for prioritizing sites for new monitoring schemes in the Far East and for planning international conservation efforts.
Supplementary Material
To view supplementary material for this article, please visit https://doi.org/10.1017/S0959270916000046
Acknowledgements
Jesper Hornskov, Aleksi Lehikoinen, Åke Lindström, Bo Pettersson, Philip Round, Sergei M. Smirenski, and Terry Townsend kindly responded to personal inquiries and provided population data. We appreciate the comments from two anonymous reviewers and the associate editor. Funding was received from Göran Gustafsson Foundation (to LE), the Finnish Cultural Foundation (to TJ) and the Kempe Foundation (to JMR).
