Environmental Management DOI 10.1007/s00267-015-0457-5
Importance of Farmland in Urbanized Areas as a Landscape Component for Barn Swallows (Hirundo rustica) Nesting on Concrete Buildings Takeshi Osawa1
Received: 20 December 2013 / Accepted: 19 March 2015 Ó Springer Science+Business Media New York 2015
Abstract Urbanization is one of the key factors in the population declines of many species. Conversely, some species may favor urbanized areas. The barn swallow Hirundo rustica is well known to breed in urban areas of Japan, and uses both urban and farmland areas as habitat during the breeding season. Specifically, this species often nests on concrete buildings and feeds in surrounding farmland. Therefore, it was hypothesized that H. rustica is not strongly influenced by heavy urbanization and benefits from farmland areas, even if they are not near its nests. In this study, I evaluated the landscape components around H. rustica nests situated on concrete buildings, focusing on both urbanized and farmland areas. In particular, I explored the occurrence of H. rustica nests at train stations in the Kinki region of Japan. Assisted by 124 citizen scientists, I analyzed the landscape components around the train stations at multiple spatial scales. Results showed that the occurrence of H. rustica nests was negatively influenced by both urbanized land area and road density, whereas nest occurrence was positively influenced by farmland area and river density. These results suggest that H. rustica does not prefer urbanized areas overall, but can rather utilize urbanized areas primarily as nesting spots. Therefore, H. rustica cannot breed in heavily urbanized areas without feeding sites such as farmland or riparian areas.
Electronic supplementary material The online version of this article (doi:10.1007/s00267-015-0457-5) contains supplementary material, which is available to authorized users. & Takeshi Osawa [email protected] 1
National Institute for Agro-Environmental Sciences, 3-1-3, Kannondai, Tsukuba, Ibaraki 305-8604, Japan
Keywords Citizen scientist Green space Hirundo rustica Landscape components Landscape ecology Station Urban ecosystems
Introduction Over half of the global human population now resides in urban areas, and urbanization is related to declining biodiversity (Goddard et al. 2010). In many cases, urbanization causes changes in habitat conditions and promotes the introduction of alien species, which in turn results in the decline of native species (Gibbs et al. 2005; Goddard et al. 2010; McKinney 2002). However, multiple studies have found that urban and suburban areas can contain relatively high levels of biodiversity (Araujo 2003; Balmford et al. 2001; Cornelis and Hermy 2004; Kuhn et al. 2004). For example, Kuhn et al. (2004) showed that both native and nonnative plant species richness was higher in urban areas than in nonurban areas in Germany. Araujo (2003) also found a positive correlation between human population density and plant, mammal, reptile, and amphibian species richness, including endemic species, throughout Europe. Additionally, Aronson et al. (2014) showed that urban areas are home to several endemic native species globally. Currently, known patterns of association between human activities and biodiversity include positive, negative, and unimodal relationships (Lepczyk et al. 2008). For example, urbanization may not only cause the loss of natural habitat, which would negatively influence biodiversity (Lehtinen et al. 1999; Soule et al. 1988), but may also enhance the mosaic of environments available at the landscape scale, which would positively influence biodiversity (Antrop 2004; Lehtinen et al. 1999; Luck et al. 2004; Soule et al. 1988). Indeed, there is evidence that such
123
Environmental Management
environmental heterogeneity is an important factor in promoting species richness (Miyashita et al. 2012; Van Rensburg et al. 2002). Furthermore, unimodal relationships between urbanization and a number of plant and animal taxa have been observed (reviewed in McKinney 2002), in accordance with the intermediate disturbance hypothesis (Connell 1978). Promoting and preserving biodiversity within urban areas is one method of decelerating the rapid rate of biodiversity loss (Alvey 2006). Clarifying the main factors involved in maintaining biodiversity in urbanized areas may aid biodiversity conservation; therefore, it is important to understand how urban areas can help maintain biodiversity. In this study, I focused on assessing the relationship between urban areas and environmental heterogeneity at the landscape scale. Some bird species are convenient for studying a variety of environmental problems and can also be used as indicators of biodiversity (Khera et al. 2010; Newton 1995; Schulze et al. 2004). In addition, many bird species require habitat heterogeneity at the landscape scale (Brady et al. 1979; Geary et al. 2013). The barn swallow H. rustica L. 1758 is a well-known avian insectivore in agricultural landscapes (Ambrosini and Saino 2010; Ambrosini et al. 2002a, 2002b; Geary et al. 2013; Turner 2006) and is distributed worldwide (Turner 2006; GBIF portal search: http://www.gbif.org/species/5230791, accessed 4 June 2014). In Japan, H. rustica uses both urban and farmland areas as habitat during its breeding season (Takagawa et al. 2012), and it is common to find nesting sites on artificial structures even in urbanized areas (Ambrosini and Saino 2010; Suzuki 1998). Therefore, H. rustica likely uses a mix of both urban and rural habitats. H. rustica’s range and distribution were relatively stable in Japan from 1980 to 2004 (Takagawa et al. 2012) despite land-use changes, in particular decreasing farmland (farmland area in 1980: 546,100 ha; 2004: 471,400 ha; Ministry of Agriculture, Forestry and Fisheries, Japan http://www.e-stat.go.jp/SG1/ estat/List.do?lid=000001058399 accessed 4 June 2014). I aimed to identify the landscape attributes that most influence H. rustica nest occurrence, with a particular focus on both farmland and urbanized areas. Highly mobile species may not be markedly affected by regional habitat loss and degradation of surrounding land because they can avoid unsuitable spots due to their mobility (Tsuji et al. 2011). Therefore, I hypothesized that H. rustica is not strongly influenced by heavy urbanization, which might otherwise constitute unsuitable habitat, because it can use nonurbanized areas such as farmland, even if distant from its nests. I also hypothesized that H. rustica nests benefit from farmland areas, which represent the species’ main feeding sites, with a preference for nearby sites because of usability. Based on these results, I discuss the importance of landscape composition in maintaining biodiversity in urbanized areas.
123
Methods Study Area and Sites I investigated the presence/absence of H. rustica nests in train stations throughout the Kinki region of Japan (Fig. 1). Train stations in Japan have similar small-scale environmental conditions across several regions consisting of open, concrete buildings. The Kinki region is characterized by strong environmental gradients, from heavily urbanized areas to farmland- and forest-dominated areas. The flatland paddy fields in this region have been lost and fragmented by rapid urbanization since the 1980s (Saizen et al. 2006). Citizen Science Approach I used a citizen science approach to record the presence/ absence of H. rustica nests (Dickinson et al. 2010; Silvertown 2009). Many ecologically based citizen science projects collect important baseline data and contribute to environmental management (Conrad and Hilchey 2011; Dickinson et al. 2012; cf. Freitag and Pfeffer 2013). Citizen science approaches can produce large longitudinal datasets, although their potential for error and bias is poorly understood (Dickinson et al. 2010, 2012). I predicted there would be few identification errors in H. rustica citizen scientist research in Japan because this species is one of the most popular Japanese birds, and observations have been reported previously (e.g., Bird Research Japan; http://www. bird-research.jp/1_katsudo/kisetu/index_kisetsu_taisho.html; http://www.tsubame-map.jp/index.html: accessed 4 June 2014). Site Surveys I observed the presence/absence of H. rustica nests during the reproductive season (20 April to 22 July) in 2012 at 1536 train station sites using a citizen science approach. In total, 124 citizen scientists were recruited using Twitter and nature-friendly mailing lists managed by the Osaka Prefectural Museum of Natural History. No face-to-face meetings on the project were conducted with participants. Citizen scientists searched for nests at station buildings and platforms; I excluded subway stations from this study. I received results from citizen scientists through e-mail and requested that they report (1) the date of the search, (2) station name, (3) the presence/absence of H. rustica nests, not considering nest number, (4) the status of nests (such as new, living, or vacant), and (5) their name in the report e-mail. Ambiguous reports were excluded. The study sites and the presence/absence records were then input as point data into ArcGIS 10.1 (ESRI, Redlands, California, USA).
Environmental Management
135° E
35° N
0
160 (km)
Fig. 1 Location of train stations in Kinki region, Japan
Landscape Characteristics Around Each Study Site I established a digital map of urbanized areas, farmland, national roads, and rivers in the study area (Appendix Table 1). Map data resources were derived from a national government source, ‘‘National Land Numerical Information,’’ according to Japanese GIS standards (National and Regional Planning Bureau, Ministry of Land, Infrastructure, Transport and Tourism, Japan: http://nlftp.mlit.go.jp/ ksj-e/index.html: accessed 4 June 2014). All datasets were originally in lat/long format with JDG 2000 data. The coordinate system was converted to UTM zone 53 to use metric units. Urbanized areas and farmland consisted of GIS polygon data, whereas roads and rivers comprised line data. I integrated and used both road and river data because these are also likely to influence the distribution of H. rustica nests (Table 1). There was minor overlap between urbanized and farmland areas on the digital maps at some scales because these datasets coarsely categorize predominant land-use types into polygons based on photographic interpretation (National and Regional Planning Bureau, Ministry of Land, Infrastructure, Transport and
Tourism, Japan http://nlftp.mlit.go.jp/ksj-e/index.html: accessed 4 June 2014); these overlapping areas were not cut or divided. Next, the composition of the surrounding landscape areas was determined (the sum of urbanized and farmland areas and the sum of road and river lengths within each buffer) at each station at five spatial scales consisting of circular buffers with radii of 500, 1000, 1500, 2000, and 2500 m. These radii were based on a previous study showing that H. rustica moves several hundred meters from its nest (Turner 2006); similarly, in Japan, H. rustica moves from several hundred meters to over 1 km from the nests during the breeding period (Sugawara 1993). Additionally, I determined the sums of both urbanized and farmland areas within circular buffers of radius 5000 m for each station. These 5000-m-radius data were used in defining the characteristics of each station (see next section). Data Analyses Before analysis, I arranged the data according to land-use components within a 5000-m buffer around each station
123
Environmental Management Table 1 Summary of the 2 datasets defined within a 5000-m-radius buffer for land characteristics and pooled data
Data
Farmland area (km2)
Urbanized area (km2)
Rural area
31.03 (15.797)
63.04 (23.58)
768
168
Urban area
16.78 (18.66)
78.51 (0.13)
768
151
Pooled
16.36 (18.51)
66.06 (19.84)
1536
300
n
Presence no.
Units were used not meter (m) but kilo meter (km)
because the study areas contained several landscape types, such as extremely rural and urbanized areas. It was possible that no relationships would be detected between H. rustica nest locations and landscape characteristics when the data from all stations were analyzed together. Thus, I selected two landscape types for analysis: rural and urban areas. I prepared two datasets consisting of (1) rural areas with farmland area in the upper 50 % and (2) areas in the upper 50 % of urbanization, within their 5000-m buffers. Some data points occurred in both datasets. Thus, I used the two excerpted datasets, each with half of the total data, as well as the whole, pooled dataset in the analysis. Subsequently, effective extent was used as a scale dependency index for H. rustica nest locations to determine the higher explanatory power scale, i.e., the effective extent scale of each land-use type in terms of H. rustica nest presence. To do so, I used generalized linear models (GLMs) with binomial distributions (logit-link). This analysis was conducted both on the excerpted datasets and for the pooled data. In all models, nest presence (presence, 1; absence, 0) was the response variable. The models had four landscape explanatory variables: urbanized area, farmland area, road length, and river length. Measurements of the variables were obtained at five scales within a circular area of radius 500, 1000, 1500, 2000, or 2500 m from each station. I selected one scale from the total of five and incorporated it on candidate models for each variable. The scale associated with each explanatory variable was identified based on Akaike’s information criterion (AIC) derived from the GLM results. I constructed 625 models (5 9 5 9 5 9 5) in total, each with the four explanatory variables for each of the five scales (500, 1000, 1500, 2000, and 2500 m), for both the excerpted and pooled datasets. I also incorporated latitude, longitude, and their respective squared values as explanatory variables into each model to regulate spatial autocorrelation among study sites. The five models with the lowest AIC for each dataset were retained from the 625 candidate models, and the majority scale of each explanatory variable was considered the effective extent scale. Finally, Moran’s I residual errors were calculated for these five models to evaluate the regulation of spatial autocorrelation (Moran 1950). All statistical analyses were performed using the statistical software package R 2.14.2 (R Development Core Team 2012).
123
Results Sites and Classes In total, of the 1536 train stations surveyed in the Kinki region, citizen scientists found 300 stations with H. rustica nests (Fig. 2). The Pearson product–moment correlation coefficient of land cover between the sum of urbanized and farmland areas within the 5000-m buffer of each station was -0.091. Effective Scale and Effects of Landscape Composition Farmland area and river length both positively influenced nest occurrence in the models for all three datasets, excluding some rural areas (Table 2). By contrast, urbanized area and road length negatively influenced nest occurrence in all datasets (Table 2). The rural area model showed that farmland area both positively and negatively influenced nest occurrence, depending on the spatial scale (Table 2). The dominant scale for river length was 1000 m (all cases) in rural areas, whereas the urban area model showed that the dominant scale for farmland was 2500 m (4 of 5 cases) and that of river length was 1000 m (4 of 5). In the pooled model, farmland within 500 m (2 of 5), rivers within 2000 m (all cases), urbanized areas within 1000 m (4 of 5), and road length within 1000 m (all cases) were all influential (Table 2). The estimated values and McFadden’s pseudo R2 values of the lowest AIC models in each dataset were shown in Appendix Table 2. Moran’s I statistics (Moran 1950) for the residual errors of all models were approximately 0 (rural area models: 0.039, 0.038, 0.039, 0.039, and 0.039; urban area models: 0.074, 0.075, 0.071, 0.071, and 0.071; pooled data: 0.0059, -0.0054, -0.0060, -0.0058, and -0.0055 from the top model to the fifth model, respectively).
Discussion The occurrence of H. rustica nests was negatively influenced by urbanization and essentially positively influenced by farmland in both the excerpted and pooled datasets.
Environmental Management
: Rural area with nest : Rural area without nest : Urban area with nest : Urban area without nest 0
160 (km)
Fig. 2 Locations of both the rural and urban areas of train stations. Circles indicate rural areas; and triangles indicate urban areas. Gray colors indicate the station has nest
These results suggest that H. rustica does not prefer urbanized areas, but that it can use them for nesting despite preferring farmland. Additionally, the effective scales for farmland were not necessarily close to the nests. Therefore, the hypothesis that H. rustica is not influenced by urbanized areas but is positively influenced by proximity to farmland was partially supported. Effects of Urbanized Areas and Road Density Hirundo rustica often nests in urbanized areas (Ambrosini and Saino 2010; Suzuki 1998). Thus, I hypothesized that urbanization would not noticeably affect H. rustica nesting preferences. However, this hypothesis was not supported, as both urbanized areas and the presence of national roads negatively influenced H. rustica nesting in all cases. This result indicates that H. rustica avoids heavily urbanized areas (i.e., areas of concentrated urbanization), at least in the selection of nesting sites. Although my hypothesis was not supported, these results are not unexpected because of food availability. Heavily urbanized areas generally have fewer insects, which provide food for H. rustica, compared to natural areas, and at least one previous population
decline in a bird species has been attributed to a lack of food (Benton et al. 2002; Evans et al. 2007). H. rustica might avoid areas of concentrated urbanization for the simple reason that these areas lack sufficient food resources. On the other hand, H. rustica benefited from concrete buildings, which are the main component of urbanized areas from the perspective of nesting spots. According to my results, H. rustica used concrete buildings solely as nesting spots, and previous studies indicate that declines in bird numbers may also be attributable to a scarcity of nesting sites (Benton et al. 2002; Evans et al. 2007). Meanwhile, the breeding conditions of H. rustica are determined by both large and small habitat scales (Ambrosini and Saino 2010; Ambrosini et al. 2002a; Turner 2006). Thus, H. rustica may be positively affected by urbanization at a small scale for nesting alone, but negatively affected by urbanization at a landscape scale. In terms of road density, previous studies indicate that high-traffic roads, such as main roads or highways, have reduced bird densities due to factors such as habitat loss and fragmentation (Brotons et al. 2005; Salek et al. 2010). H. rustica does not appear to be an exception to this
123
Environmental Management Table 2 Generalized linear model (GLM) and model selection tests based on Akaike’s Information Criterion (AIC) for the effects of scales on farmland, urbanized areas, river length, and road length in the presence/absence of nests Data
Farmland (m)
Rural area
Urban area
All
Urbanized area (m) ?
1500
River length (m) -
Road length (m)
AIC
1000
1
1000
-
1500
-
1500
-
1000
?
1000
-
500
?
1500
-
1000
?
1000
-
799.272 799.305
2000
-
1500
-
1000
?
1000
-
799.3301
2500
?
1500
-
1000
?
1000
-
799.3305
1000
?
500
-
1000
?
1000
-
799.793
2500 2500
1 ?
500 500
-
1000 1000
1 ?
500 500
-
764.390
2000
?
500
-
1000
?
500
-
764.607
2500
?
1500
-
1000
?
500
-
764.6296
2500
?
2000
-
1000
?
500
-
764.686
2500
?
500
-
500
?
500
-
764.928
500
1
1000
-
2000
1
1000
-
500
?
1000
-
2000
?
1000
-
1505.896
2500
?
1000
-
2000
?
1000
-
1505.992
500
?
500
-
2000
?
1000
-
1506.010
1000
?
1000
-
2000
?
1000
-
1506.044
2000
?
1000
-
2000
?
1000
-
1506.084
Positive (?) or negative (-) model contributions with the scales of the top five AIC models are shown. Positive (?) or negative (-) means that a variable positively/negatively contributed to the presence of a nest, respectively. Bold results are the main scales among the five models
postulate. Road density itself had only negative effects on H. rustica because roads cannot serve as H. rustica nesting spots and may not provide insects for food. The effective scales for urbanized area and road density were relatively large in the rural area model, relatively small in the urban area model, and intermediate in the pooled data model. These results also suggest that H. rustica tends to avoid heavily concentrated urbanized areas and high-density road regions in particular. In rural areas, both urbanized regions and road density were relatively sparse. As a result, H. rustica in such areas tended to avoid concentrated urban areas at relatively large scales. By contrast, in urban areas, both urbanized area and road density were relatively dense, and thus H. rustica tended to avoid such concentrated urbanized areas at relatively small scales. The effective scale for the pooled data was 1000 m, which was between the rural and urban effective scales. This trend also supported the results from both the rural and urban models, which used portions of the pooled data. Thus, H. rustica clearly avoids heavily urbanized areas. Effects of Farmland In contrast to the results above, farmland had a positive effect on H. rustica nesting in all cases except in rural area models. This result was also not unexpected due to the importance of food availability: farmland generally has more insects than urbanized areas (Denys and Schmidt
123
1998; McIntyre 2000). Previous studies have shown that areas of livestock farming positively affect barn swallows at the landscape level mainly due to improved feeding conditions (Gruebler et al. 2010). Therefore, farmland areas may provide an increase in insect prey availability (Gruebler et al. 2010). During the breeding season, birds require additional insects to feed juveniles, and farmland provides a good hunting site for H. rustica. In the rural area model, there are generally large farmland areas around stations; thus, the model could not detect clear trends. I discuss this result in the context of scale dependency below. The effective scales for farmland showed intriguing trends. The most effective scale for the urban area models was 2500 m (the largest scale), whereas the most effective scale for the pooled data model was 500 m (the smallest scale). Rural areas showed no effect at any scale. One potential explanation for this finding is competition avoidance. Robillard et al. (2013) suggested that nesting sites near farms represent a greater competition risk for tree swallows. Therefore, both rural areas, which have high food resources, and urbanized areas, which have relatively low food resources, may present a high competition risk. Food-rich areas tend to breed too many H. rustica, which may lead to territorial competition. To avoid such competition, H. rustica may avoid specific scales in farmland areas. Conversely, food-poor areas may cause food competition through other mechanisms, such as high searching
Environmental Management
costs. To obtain sufficient food resources, H. rustica may search over large scales. Future research should test this issue with detailed observations of behavior and/or interactions with potential competitors. The effective scale for farmland in all data models was 500 m, the smallest scale. This suggests that H. rustica generally prefers areas near farmland, in line with my prediction. However, the top five models pertain to relatively large scales (2000 and 2500 m). These results may be explained by food competition in both rural and urban areas. Effects of River Density River lengths had a positive influence in all cases at relatively small scales. This result suggests that H. rustica fundamentally prefers rivers and/or riparian areas. The effective scale in the urban area model was 500 m, the smallest scale, but was 1000 m in both the rural and pooled models. Although urbanized, concrete-walled rivers have less biodiversity (Urban et al. 2006), they generally have large numbers of insects (Naiman et al. 1993), and thus H. rustica might use rivers as feeding sites. Consequently, river length could reflect the number of feeding sites for H. rustica in both rural and urban areas. Especially in urban regions, H. rustica might select riparian areas as the main feeding sites because of the relative scarcity of farmland. In urban areas, H. rustica might use farmland areas as secondary feeding sites. Future investigations should also examine the feeding-site preferences and feeding behavior of H. rustica.
Conclusion My results suggest that H. rustica uses urbanized areas solely as a nesting microenvironment and does not prefer urbanized areas overall. Additionally, H. rustica may require both farmland and riparian areas as feeding sites. Therefore, H. rustica cannot breed in heavily urbanized areas without feeding sites such as farmland and riparian areas, and consequently, green space (such as farmland) within urbanized areas is important for promoting H. rustica breeding. This type of heterogeneous habitat is often called a mosaic environment, but the role of each component within the mosaic, such as urbanized area and farmland, differs ecologically by scale. These findings will aid land managers in improving heavily urbanized areas as habitats for bird species. Acknowledgments I would like to thank Dr. T. Wada and all 124 citizen scientists for joining this study. Especially, Dr. Wada strongly contributed for the management of the citizen scientist activities. I also would like to thank the member of urbanized bird research circle for their valuable support. I also would like to thank Dr. N. Katayama for
commenting on an earlier version of this manuscript. Two reviewers and the editor gave me several useful comments and suggestions. This study was partially supported by a grant-in-aid for young scientists (No. 24710038) from the Japan Society for the Promotion of Science.
References Alvey AA (2006) Promoting and preserving biodiversity in the urban forest. Urban For Urban Green 5:195–201 Ambrosini R, Saino N (2010) Environmental effects at two nested spatial scales on habitat choice and breeding performance of barn swallow. Evol Ecol 24:491–508 Ambrosini R, Bolzern AM, Canova L, Arieni S, Moller AP, Saino N (2002a) The distribution and colony size of barn swallows in relation to agricultural land use. J Appl Ecol 39:524–534 Ambrosini R, Bolzern AM, Canova L, Saino N (2002b) Latency in response of barn swallow Hirundo rustica populations to changes in breeding habitat conditions. Ecol Lett 5:640–647 Antrop M (2004) Landscape change and the urbanization process in Europe. Landsc Urban Plan 67:9–26 Araujo MB (2003) The coincidence of people and biodiversity in Europe. Glob Ecol Biogeogr 12:5–12 Aronson MF, La Sorte FA, Nilon CH et al (2014) A global analysis of the impacts of urbanization on bird and plant diversity reveals key anthropogenic drivers. Proc R Soc B 281:20133330 Balmford A, Moore JL, Brooks T et al (2001) Conservation conflicts across Africa. Science 291:2616–2619 Benton TG, Bryant DM, Cole L, Crick HQP (2002) Linking agricultural practice to insect and bird populations: a historical study over three decades. J Appl Ecol 39:673–687 Brady R, Tobias T, Eagles PF et al (1979) A typology for the urban ecosystem and its relationship to larger biogeographical landscape units. Urban Ecol 4:11–28 Brotons L, Herrando S, Martin J-L (2005) Bird assemblages in forest fragments within Mediterranean mosaics created by wild fires. Landsc Ecol 19:663–675 Connell JH (1978) Diversity in tropical rain forests and coral reefs. Science 199:1302–1310 Conrad CC, Hilchey KG (2011) A review of citizen science and community-based environmental monitoring: issues and opportunities. Environ Monit Assess 176:273–291 Cornelis J, Hermy M (2004) Biodiversity relationships in urban and suburban parks in Flanders. Landsc Urban Plan 69:385–401 Denys C, Schmidt H (1998) Insect communities on experimental mugwort (Artemisia vulgaris L.) plots along an urban gradient. Oecologia 113:269–277 Dickinson JL, Zuckerberg B, Bonter DN (2010) Citizen science as an ecological research tool: challenges and benefits. Annu Rev Ecol Evol Syst 41:149–172 Dickinson JL, Shirk J, Bonter D et al (2012) The current state of citizen science as a tool for ecological research and public engagement. Front Ecol Environ 10:291–297 Evans KL, Wilson JD, Bradbury RB (2007) Effects of crop type and aerial invertebrate abundance on foraging barn swallows Hirundo rustica. Agric Ecosyst Environ 122:267–273 Freitag A, Pfeffer MJ (2013) Process, not product: investigating recommendations for improving citizen science ‘‘success’’. PLoS One 8:e64079 Geary M, Fielding AH, Marsden SJ (2013) Designing mosaic landscapes for Black Grouse Tetrao tetrix using multi-scaled models. Ibis 155:792–803 Gibbs JP, Whiteleather KK, Schueler FW (2005) Changes in frog and toad populations over 30 years in New York State. Ecol Appl 15:1148–1157
123
Environmental Management Goddard MA, Dougill AJ, Benton TG (2010) Scaling up from gardens: biodiversity conservation in urban environments. Trends Ecol Evol 25:90–98 Gruebler MU, Korner-Nievergelt F, Von Hirschheydt J (2010) The reproductive benefits of livestock farming in barn swallows Hirundo rustica: quality of nest site or foraging habitat? J Appl Ecol 47:1340–1347 Khera N, Das A, Srivasatava S, Jain S (2010) Habitat-wise distribution of the house sparrow (Passer domesticus) in Delhi, India. Urban Ecosyst 13:147–154 Kuhn I, Brandl R, Klotz S (2004) The flora of German cities is naturally species rich. Evol Ecol Res 6:749–764 Lehtinen RM, Galatowitsch SM, Tester JR (1999) Consequences of habitat loss and fragmentation for wetland amphibian assemblages. Wetlands 19:1–12 Lepczyk CA, Flather CH, Radeloff VC, Pidgeon AM, Hammer RB, Liu J (2008) Human impacts on regional avian diversity and abundance. Conserv Biol 22:405–416 Luck GW, Ricketts TH, Daily GC, Imhoff M (2004) Alleviating spatial conflict between people and biodiversity. Proc Natl Acad Sci USA 101:182–186 Mcintyre NE (2000) Ecology of urban arthropods: a review and a call to action. Ann Entomol Soc Am 93:825–835 Mckinney ML (2002) Urbanization, biodiversity, and conservation: the impacts of urbanization on native species are poorly studied, but educating a highly urbanized human population about these impacts can greatly improve species conservation in all ecosystems. Bioscience 52:883–890 Miyashita T, Chishiki Y, Takagi SR (2012) Landscape heterogeneity at multiple spatial scales enhances spider species richness in an agricultural landscape. Popul Ecol 54:573–581 Moran PaP (1950) Notes on continuous stochastic phenomena. Biometrika 37:17–23 Naiman RJ, Decamps H, Pollock M (1993) The role of riparian corridors in maintaining regional biodiversity. Ecol Appl 3:209–212 Newton I (1995) The contribution of some recent research on birds to ecological understanding. J Anim Ecol 64:675–695
123
Robillard A, Garant D, Belisle M (2013) The swallow and the sparrow: how agricultural intensification affects abundance, nest site selection and competitive interactions. Landsc Ecol 28:201–215 Saizen I, Mizuno K, Kobayashi S (2006) Effects of land-use master plans in the metropolitan fringe of Japan. Landsc Urban Plan 78:411–421 Salek M, Svobodova J, Zasadil P (2010) Edge effect of low-traffic forest roads on bird communities in secondary production forests in central Europe. Landsc Ecol 25:1113–1124 Schulze CH, Waltert M, Kessler PJA et al (2004) Biodiversity indicator groups of tropical land-use systems: comparing plants, birds, and insects. Ecol Appl 14:1321–1333 Silvertown J (2009) A new dawn for citizen science. Trends Ecol Evol 24:467–471 Soule ME, Bolger DT, Alberts AC, Wrights J, Sorice M, Hill S (1988) Reconstructed dynamics of rapid extinctions of haparral-requiring birds in urban habitat islands. Conserv Biol 2:75–92 Sugawara K (1993) Hirundo rustica and Sturnus cineraceus. Photo Book, Seibundo Shinkosha, Tokyo Suzuki H (1998) Nests situations on Hirundo rustica in Tahata river, Hiroshima. Strix 16:99–108 Takagawa S, Al E (2012) JAVIAN Database: a species-level database of life history, ecology and morphology of bird species in Japan. Bird Res 7:R9–R12 Tsuji M, Ushimaru A, Osawa T, Mitsuhashi H (2011) Paddyassociated frog declines via urbanization: a test of the dispersaldependent-decline hypothesis. Landsc Urban Plan 103:318–325 Turner A (2006) The barn swallow. A&C Black, London Urban MC, Skelly DK, Burchsted D, Price W, Lowry S (2006) Stream communities across a rural urban landscape gradient. Divers Distrib 12:337–350 Van Rensburg BJ, Chown SL, Gaston KJ (2002) Species richness, environmental correlates, and spatial scale: a test using South African birds. Am Nat 159:566–577
Importance of Farmland in Urbanized Areas as a Landscape Component for Barn Swallows (Hirundo rustica) Nesting on Concrete Buildings Takeshi Osawa1
Received: 20 December 2013 / Accepted: 19 March 2015 Ó Springer Science+Business Media New York 2015
Abstract Urbanization is one of the key factors in the population declines of many species. Conversely, some species may favor urbanized areas. The barn swallow Hirundo rustica is well known to breed in urban areas of Japan, and uses both urban and farmland areas as habitat during the breeding season. Specifically, this species often nests on concrete buildings and feeds in surrounding farmland. Therefore, it was hypothesized that H. rustica is not strongly influenced by heavy urbanization and benefits from farmland areas, even if they are not near its nests. In this study, I evaluated the landscape components around H. rustica nests situated on concrete buildings, focusing on both urbanized and farmland areas. In particular, I explored the occurrence of H. rustica nests at train stations in the Kinki region of Japan. Assisted by 124 citizen scientists, I analyzed the landscape components around the train stations at multiple spatial scales. Results showed that the occurrence of H. rustica nests was negatively influenced by both urbanized land area and road density, whereas nest occurrence was positively influenced by farmland area and river density. These results suggest that H. rustica does not prefer urbanized areas overall, but can rather utilize urbanized areas primarily as nesting spots. Therefore, H. rustica cannot breed in heavily urbanized areas without feeding sites such as farmland or riparian areas.
Electronic supplementary material The online version of this article (doi:10.1007/s00267-015-0457-5) contains supplementary material, which is available to authorized users. & Takeshi Osawa [email protected] 1
National Institute for Agro-Environmental Sciences, 3-1-3, Kannondai, Tsukuba, Ibaraki 305-8604, Japan
Keywords Citizen scientist Green space Hirundo rustica Landscape components Landscape ecology Station Urban ecosystems
Introduction Over half of the global human population now resides in urban areas, and urbanization is related to declining biodiversity (Goddard et al. 2010). In many cases, urbanization causes changes in habitat conditions and promotes the introduction of alien species, which in turn results in the decline of native species (Gibbs et al. 2005; Goddard et al. 2010; McKinney 2002). However, multiple studies have found that urban and suburban areas can contain relatively high levels of biodiversity (Araujo 2003; Balmford et al. 2001; Cornelis and Hermy 2004; Kuhn et al. 2004). For example, Kuhn et al. (2004) showed that both native and nonnative plant species richness was higher in urban areas than in nonurban areas in Germany. Araujo (2003) also found a positive correlation between human population density and plant, mammal, reptile, and amphibian species richness, including endemic species, throughout Europe. Additionally, Aronson et al. (2014) showed that urban areas are home to several endemic native species globally. Currently, known patterns of association between human activities and biodiversity include positive, negative, and unimodal relationships (Lepczyk et al. 2008). For example, urbanization may not only cause the loss of natural habitat, which would negatively influence biodiversity (Lehtinen et al. 1999; Soule et al. 1988), but may also enhance the mosaic of environments available at the landscape scale, which would positively influence biodiversity (Antrop 2004; Lehtinen et al. 1999; Luck et al. 2004; Soule et al. 1988). Indeed, there is evidence that such
123
Environmental Management
environmental heterogeneity is an important factor in promoting species richness (Miyashita et al. 2012; Van Rensburg et al. 2002). Furthermore, unimodal relationships between urbanization and a number of plant and animal taxa have been observed (reviewed in McKinney 2002), in accordance with the intermediate disturbance hypothesis (Connell 1978). Promoting and preserving biodiversity within urban areas is one method of decelerating the rapid rate of biodiversity loss (Alvey 2006). Clarifying the main factors involved in maintaining biodiversity in urbanized areas may aid biodiversity conservation; therefore, it is important to understand how urban areas can help maintain biodiversity. In this study, I focused on assessing the relationship between urban areas and environmental heterogeneity at the landscape scale. Some bird species are convenient for studying a variety of environmental problems and can also be used as indicators of biodiversity (Khera et al. 2010; Newton 1995; Schulze et al. 2004). In addition, many bird species require habitat heterogeneity at the landscape scale (Brady et al. 1979; Geary et al. 2013). The barn swallow H. rustica L. 1758 is a well-known avian insectivore in agricultural landscapes (Ambrosini and Saino 2010; Ambrosini et al. 2002a, 2002b; Geary et al. 2013; Turner 2006) and is distributed worldwide (Turner 2006; GBIF portal search: http://www.gbif.org/species/5230791, accessed 4 June 2014). In Japan, H. rustica uses both urban and farmland areas as habitat during its breeding season (Takagawa et al. 2012), and it is common to find nesting sites on artificial structures even in urbanized areas (Ambrosini and Saino 2010; Suzuki 1998). Therefore, H. rustica likely uses a mix of both urban and rural habitats. H. rustica’s range and distribution were relatively stable in Japan from 1980 to 2004 (Takagawa et al. 2012) despite land-use changes, in particular decreasing farmland (farmland area in 1980: 546,100 ha; 2004: 471,400 ha; Ministry of Agriculture, Forestry and Fisheries, Japan http://www.e-stat.go.jp/SG1/ estat/List.do?lid=000001058399 accessed 4 June 2014). I aimed to identify the landscape attributes that most influence H. rustica nest occurrence, with a particular focus on both farmland and urbanized areas. Highly mobile species may not be markedly affected by regional habitat loss and degradation of surrounding land because they can avoid unsuitable spots due to their mobility (Tsuji et al. 2011). Therefore, I hypothesized that H. rustica is not strongly influenced by heavy urbanization, which might otherwise constitute unsuitable habitat, because it can use nonurbanized areas such as farmland, even if distant from its nests. I also hypothesized that H. rustica nests benefit from farmland areas, which represent the species’ main feeding sites, with a preference for nearby sites because of usability. Based on these results, I discuss the importance of landscape composition in maintaining biodiversity in urbanized areas.
123
Methods Study Area and Sites I investigated the presence/absence of H. rustica nests in train stations throughout the Kinki region of Japan (Fig. 1). Train stations in Japan have similar small-scale environmental conditions across several regions consisting of open, concrete buildings. The Kinki region is characterized by strong environmental gradients, from heavily urbanized areas to farmland- and forest-dominated areas. The flatland paddy fields in this region have been lost and fragmented by rapid urbanization since the 1980s (Saizen et al. 2006). Citizen Science Approach I used a citizen science approach to record the presence/ absence of H. rustica nests (Dickinson et al. 2010; Silvertown 2009). Many ecologically based citizen science projects collect important baseline data and contribute to environmental management (Conrad and Hilchey 2011; Dickinson et al. 2012; cf. Freitag and Pfeffer 2013). Citizen science approaches can produce large longitudinal datasets, although their potential for error and bias is poorly understood (Dickinson et al. 2010, 2012). I predicted there would be few identification errors in H. rustica citizen scientist research in Japan because this species is one of the most popular Japanese birds, and observations have been reported previously (e.g., Bird Research Japan; http://www. bird-research.jp/1_katsudo/kisetu/index_kisetsu_taisho.html; http://www.tsubame-map.jp/index.html: accessed 4 June 2014). Site Surveys I observed the presence/absence of H. rustica nests during the reproductive season (20 April to 22 July) in 2012 at 1536 train station sites using a citizen science approach. In total, 124 citizen scientists were recruited using Twitter and nature-friendly mailing lists managed by the Osaka Prefectural Museum of Natural History. No face-to-face meetings on the project were conducted with participants. Citizen scientists searched for nests at station buildings and platforms; I excluded subway stations from this study. I received results from citizen scientists through e-mail and requested that they report (1) the date of the search, (2) station name, (3) the presence/absence of H. rustica nests, not considering nest number, (4) the status of nests (such as new, living, or vacant), and (5) their name in the report e-mail. Ambiguous reports were excluded. The study sites and the presence/absence records were then input as point data into ArcGIS 10.1 (ESRI, Redlands, California, USA).
Environmental Management
135° E
35° N
0
160 (km)
Fig. 1 Location of train stations in Kinki region, Japan
Landscape Characteristics Around Each Study Site I established a digital map of urbanized areas, farmland, national roads, and rivers in the study area (Appendix Table 1). Map data resources were derived from a national government source, ‘‘National Land Numerical Information,’’ according to Japanese GIS standards (National and Regional Planning Bureau, Ministry of Land, Infrastructure, Transport and Tourism, Japan: http://nlftp.mlit.go.jp/ ksj-e/index.html: accessed 4 June 2014). All datasets were originally in lat/long format with JDG 2000 data. The coordinate system was converted to UTM zone 53 to use metric units. Urbanized areas and farmland consisted of GIS polygon data, whereas roads and rivers comprised line data. I integrated and used both road and river data because these are also likely to influence the distribution of H. rustica nests (Table 1). There was minor overlap between urbanized and farmland areas on the digital maps at some scales because these datasets coarsely categorize predominant land-use types into polygons based on photographic interpretation (National and Regional Planning Bureau, Ministry of Land, Infrastructure, Transport and
Tourism, Japan http://nlftp.mlit.go.jp/ksj-e/index.html: accessed 4 June 2014); these overlapping areas were not cut or divided. Next, the composition of the surrounding landscape areas was determined (the sum of urbanized and farmland areas and the sum of road and river lengths within each buffer) at each station at five spatial scales consisting of circular buffers with radii of 500, 1000, 1500, 2000, and 2500 m. These radii were based on a previous study showing that H. rustica moves several hundred meters from its nest (Turner 2006); similarly, in Japan, H. rustica moves from several hundred meters to over 1 km from the nests during the breeding period (Sugawara 1993). Additionally, I determined the sums of both urbanized and farmland areas within circular buffers of radius 5000 m for each station. These 5000-m-radius data were used in defining the characteristics of each station (see next section). Data Analyses Before analysis, I arranged the data according to land-use components within a 5000-m buffer around each station
123
Environmental Management Table 1 Summary of the 2 datasets defined within a 5000-m-radius buffer for land characteristics and pooled data
Data
Farmland area (km2)
Urbanized area (km2)
Rural area
31.03 (15.797)
63.04 (23.58)
768
168
Urban area
16.78 (18.66)
78.51 (0.13)
768
151
Pooled
16.36 (18.51)
66.06 (19.84)
1536
300
n
Presence no.
Units were used not meter (m) but kilo meter (km)
because the study areas contained several landscape types, such as extremely rural and urbanized areas. It was possible that no relationships would be detected between H. rustica nest locations and landscape characteristics when the data from all stations were analyzed together. Thus, I selected two landscape types for analysis: rural and urban areas. I prepared two datasets consisting of (1) rural areas with farmland area in the upper 50 % and (2) areas in the upper 50 % of urbanization, within their 5000-m buffers. Some data points occurred in both datasets. Thus, I used the two excerpted datasets, each with half of the total data, as well as the whole, pooled dataset in the analysis. Subsequently, effective extent was used as a scale dependency index for H. rustica nest locations to determine the higher explanatory power scale, i.e., the effective extent scale of each land-use type in terms of H. rustica nest presence. To do so, I used generalized linear models (GLMs) with binomial distributions (logit-link). This analysis was conducted both on the excerpted datasets and for the pooled data. In all models, nest presence (presence, 1; absence, 0) was the response variable. The models had four landscape explanatory variables: urbanized area, farmland area, road length, and river length. Measurements of the variables were obtained at five scales within a circular area of radius 500, 1000, 1500, 2000, or 2500 m from each station. I selected one scale from the total of five and incorporated it on candidate models for each variable. The scale associated with each explanatory variable was identified based on Akaike’s information criterion (AIC) derived from the GLM results. I constructed 625 models (5 9 5 9 5 9 5) in total, each with the four explanatory variables for each of the five scales (500, 1000, 1500, 2000, and 2500 m), for both the excerpted and pooled datasets. I also incorporated latitude, longitude, and their respective squared values as explanatory variables into each model to regulate spatial autocorrelation among study sites. The five models with the lowest AIC for each dataset were retained from the 625 candidate models, and the majority scale of each explanatory variable was considered the effective extent scale. Finally, Moran’s I residual errors were calculated for these five models to evaluate the regulation of spatial autocorrelation (Moran 1950). All statistical analyses were performed using the statistical software package R 2.14.2 (R Development Core Team 2012).
123
Results Sites and Classes In total, of the 1536 train stations surveyed in the Kinki region, citizen scientists found 300 stations with H. rustica nests (Fig. 2). The Pearson product–moment correlation coefficient of land cover between the sum of urbanized and farmland areas within the 5000-m buffer of each station was -0.091. Effective Scale and Effects of Landscape Composition Farmland area and river length both positively influenced nest occurrence in the models for all three datasets, excluding some rural areas (Table 2). By contrast, urbanized area and road length negatively influenced nest occurrence in all datasets (Table 2). The rural area model showed that farmland area both positively and negatively influenced nest occurrence, depending on the spatial scale (Table 2). The dominant scale for river length was 1000 m (all cases) in rural areas, whereas the urban area model showed that the dominant scale for farmland was 2500 m (4 of 5 cases) and that of river length was 1000 m (4 of 5). In the pooled model, farmland within 500 m (2 of 5), rivers within 2000 m (all cases), urbanized areas within 1000 m (4 of 5), and road length within 1000 m (all cases) were all influential (Table 2). The estimated values and McFadden’s pseudo R2 values of the lowest AIC models in each dataset were shown in Appendix Table 2. Moran’s I statistics (Moran 1950) for the residual errors of all models were approximately 0 (rural area models: 0.039, 0.038, 0.039, 0.039, and 0.039; urban area models: 0.074, 0.075, 0.071, 0.071, and 0.071; pooled data: 0.0059, -0.0054, -0.0060, -0.0058, and -0.0055 from the top model to the fifth model, respectively).
Discussion The occurrence of H. rustica nests was negatively influenced by urbanization and essentially positively influenced by farmland in both the excerpted and pooled datasets.
Environmental Management
: Rural area with nest : Rural area without nest : Urban area with nest : Urban area without nest 0
160 (km)
Fig. 2 Locations of both the rural and urban areas of train stations. Circles indicate rural areas; and triangles indicate urban areas. Gray colors indicate the station has nest
These results suggest that H. rustica does not prefer urbanized areas, but that it can use them for nesting despite preferring farmland. Additionally, the effective scales for farmland were not necessarily close to the nests. Therefore, the hypothesis that H. rustica is not influenced by urbanized areas but is positively influenced by proximity to farmland was partially supported. Effects of Urbanized Areas and Road Density Hirundo rustica often nests in urbanized areas (Ambrosini and Saino 2010; Suzuki 1998). Thus, I hypothesized that urbanization would not noticeably affect H. rustica nesting preferences. However, this hypothesis was not supported, as both urbanized areas and the presence of national roads negatively influenced H. rustica nesting in all cases. This result indicates that H. rustica avoids heavily urbanized areas (i.e., areas of concentrated urbanization), at least in the selection of nesting sites. Although my hypothesis was not supported, these results are not unexpected because of food availability. Heavily urbanized areas generally have fewer insects, which provide food for H. rustica, compared to natural areas, and at least one previous population
decline in a bird species has been attributed to a lack of food (Benton et al. 2002; Evans et al. 2007). H. rustica might avoid areas of concentrated urbanization for the simple reason that these areas lack sufficient food resources. On the other hand, H. rustica benefited from concrete buildings, which are the main component of urbanized areas from the perspective of nesting spots. According to my results, H. rustica used concrete buildings solely as nesting spots, and previous studies indicate that declines in bird numbers may also be attributable to a scarcity of nesting sites (Benton et al. 2002; Evans et al. 2007). Meanwhile, the breeding conditions of H. rustica are determined by both large and small habitat scales (Ambrosini and Saino 2010; Ambrosini et al. 2002a; Turner 2006). Thus, H. rustica may be positively affected by urbanization at a small scale for nesting alone, but negatively affected by urbanization at a landscape scale. In terms of road density, previous studies indicate that high-traffic roads, such as main roads or highways, have reduced bird densities due to factors such as habitat loss and fragmentation (Brotons et al. 2005; Salek et al. 2010). H. rustica does not appear to be an exception to this
123
Environmental Management Table 2 Generalized linear model (GLM) and model selection tests based on Akaike’s Information Criterion (AIC) for the effects of scales on farmland, urbanized areas, river length, and road length in the presence/absence of nests Data
Farmland (m)
Rural area
Urban area
All
Urbanized area (m) ?
1500
River length (m) -
Road length (m)
AIC
1000
1
1000
-
1500
-
1500
-
1000
?
1000
-
500
?
1500
-
1000
?
1000
-
799.272 799.305
2000
-
1500
-
1000
?
1000
-
799.3301
2500
?
1500
-
1000
?
1000
-
799.3305
1000
?
500
-
1000
?
1000
-
799.793
2500 2500
1 ?
500 500
-
1000 1000
1 ?
500 500
-
764.390
2000
?
500
-
1000
?
500
-
764.607
2500
?
1500
-
1000
?
500
-
764.6296
2500
?
2000
-
1000
?
500
-
764.686
2500
?
500
-
500
?
500
-
764.928
500
1
1000
-
2000
1
1000
-
500
?
1000
-
2000
?
1000
-
1505.896
2500
?
1000
-
2000
?
1000
-
1505.992
500
?
500
-
2000
?
1000
-
1506.010
1000
?
1000
-
2000
?
1000
-
1506.044
2000
?
1000
-
2000
?
1000
-
1506.084
Positive (?) or negative (-) model contributions with the scales of the top five AIC models are shown. Positive (?) or negative (-) means that a variable positively/negatively contributed to the presence of a nest, respectively. Bold results are the main scales among the five models
postulate. Road density itself had only negative effects on H. rustica because roads cannot serve as H. rustica nesting spots and may not provide insects for food. The effective scales for urbanized area and road density were relatively large in the rural area model, relatively small in the urban area model, and intermediate in the pooled data model. These results also suggest that H. rustica tends to avoid heavily concentrated urbanized areas and high-density road regions in particular. In rural areas, both urbanized regions and road density were relatively sparse. As a result, H. rustica in such areas tended to avoid concentrated urban areas at relatively large scales. By contrast, in urban areas, both urbanized area and road density were relatively dense, and thus H. rustica tended to avoid such concentrated urbanized areas at relatively small scales. The effective scale for the pooled data was 1000 m, which was between the rural and urban effective scales. This trend also supported the results from both the rural and urban models, which used portions of the pooled data. Thus, H. rustica clearly avoids heavily urbanized areas. Effects of Farmland In contrast to the results above, farmland had a positive effect on H. rustica nesting in all cases except in rural area models. This result was also not unexpected due to the importance of food availability: farmland generally has more insects than urbanized areas (Denys and Schmidt
123
1998; McIntyre 2000). Previous studies have shown that areas of livestock farming positively affect barn swallows at the landscape level mainly due to improved feeding conditions (Gruebler et al. 2010). Therefore, farmland areas may provide an increase in insect prey availability (Gruebler et al. 2010). During the breeding season, birds require additional insects to feed juveniles, and farmland provides a good hunting site for H. rustica. In the rural area model, there are generally large farmland areas around stations; thus, the model could not detect clear trends. I discuss this result in the context of scale dependency below. The effective scales for farmland showed intriguing trends. The most effective scale for the urban area models was 2500 m (the largest scale), whereas the most effective scale for the pooled data model was 500 m (the smallest scale). Rural areas showed no effect at any scale. One potential explanation for this finding is competition avoidance. Robillard et al. (2013) suggested that nesting sites near farms represent a greater competition risk for tree swallows. Therefore, both rural areas, which have high food resources, and urbanized areas, which have relatively low food resources, may present a high competition risk. Food-rich areas tend to breed too many H. rustica, which may lead to territorial competition. To avoid such competition, H. rustica may avoid specific scales in farmland areas. Conversely, food-poor areas may cause food competition through other mechanisms, such as high searching
Environmental Management
costs. To obtain sufficient food resources, H. rustica may search over large scales. Future research should test this issue with detailed observations of behavior and/or interactions with potential competitors. The effective scale for farmland in all data models was 500 m, the smallest scale. This suggests that H. rustica generally prefers areas near farmland, in line with my prediction. However, the top five models pertain to relatively large scales (2000 and 2500 m). These results may be explained by food competition in both rural and urban areas. Effects of River Density River lengths had a positive influence in all cases at relatively small scales. This result suggests that H. rustica fundamentally prefers rivers and/or riparian areas. The effective scale in the urban area model was 500 m, the smallest scale, but was 1000 m in both the rural and pooled models. Although urbanized, concrete-walled rivers have less biodiversity (Urban et al. 2006), they generally have large numbers of insects (Naiman et al. 1993), and thus H. rustica might use rivers as feeding sites. Consequently, river length could reflect the number of feeding sites for H. rustica in both rural and urban areas. Especially in urban regions, H. rustica might select riparian areas as the main feeding sites because of the relative scarcity of farmland. In urban areas, H. rustica might use farmland areas as secondary feeding sites. Future investigations should also examine the feeding-site preferences and feeding behavior of H. rustica.
Conclusion My results suggest that H. rustica uses urbanized areas solely as a nesting microenvironment and does not prefer urbanized areas overall. Additionally, H. rustica may require both farmland and riparian areas as feeding sites. Therefore, H. rustica cannot breed in heavily urbanized areas without feeding sites such as farmland and riparian areas, and consequently, green space (such as farmland) within urbanized areas is important for promoting H. rustica breeding. This type of heterogeneous habitat is often called a mosaic environment, but the role of each component within the mosaic, such as urbanized area and farmland, differs ecologically by scale. These findings will aid land managers in improving heavily urbanized areas as habitats for bird species. Acknowledgments I would like to thank Dr. T. Wada and all 124 citizen scientists for joining this study. Especially, Dr. Wada strongly contributed for the management of the citizen scientist activities. I also would like to thank the member of urbanized bird research circle for their valuable support. I also would like to thank Dr. N. Katayama for
commenting on an earlier version of this manuscript. Two reviewers and the editor gave me several useful comments and suggestions. This study was partially supported by a grant-in-aid for young scientists (No. 24710038) from the Japan Society for the Promotion of Science.
References Alvey AA (2006) Promoting and preserving biodiversity in the urban forest. Urban For Urban Green 5:195–201 Ambrosini R, Saino N (2010) Environmental effects at two nested spatial scales on habitat choice and breeding performance of barn swallow. Evol Ecol 24:491–508 Ambrosini R, Bolzern AM, Canova L, Arieni S, Moller AP, Saino N (2002a) The distribution and colony size of barn swallows in relation to agricultural land use. J Appl Ecol 39:524–534 Ambrosini R, Bolzern AM, Canova L, Saino N (2002b) Latency in response of barn swallow Hirundo rustica populations to changes in breeding habitat conditions. Ecol Lett 5:640–647 Antrop M (2004) Landscape change and the urbanization process in Europe. Landsc Urban Plan 67:9–26 Araujo MB (2003) The coincidence of people and biodiversity in Europe. Glob Ecol Biogeogr 12:5–12 Aronson MF, La Sorte FA, Nilon CH et al (2014) A global analysis of the impacts of urbanization on bird and plant diversity reveals key anthropogenic drivers. Proc R Soc B 281:20133330 Balmford A, Moore JL, Brooks T et al (2001) Conservation conflicts across Africa. Science 291:2616–2619 Benton TG, Bryant DM, Cole L, Crick HQP (2002) Linking agricultural practice to insect and bird populations: a historical study over three decades. J Appl Ecol 39:673–687 Brady R, Tobias T, Eagles PF et al (1979) A typology for the urban ecosystem and its relationship to larger biogeographical landscape units. Urban Ecol 4:11–28 Brotons L, Herrando S, Martin J-L (2005) Bird assemblages in forest fragments within Mediterranean mosaics created by wild fires. Landsc Ecol 19:663–675 Connell JH (1978) Diversity in tropical rain forests and coral reefs. Science 199:1302–1310 Conrad CC, Hilchey KG (2011) A review of citizen science and community-based environmental monitoring: issues and opportunities. Environ Monit Assess 176:273–291 Cornelis J, Hermy M (2004) Biodiversity relationships in urban and suburban parks in Flanders. Landsc Urban Plan 69:385–401 Denys C, Schmidt H (1998) Insect communities on experimental mugwort (Artemisia vulgaris L.) plots along an urban gradient. Oecologia 113:269–277 Dickinson JL, Zuckerberg B, Bonter DN (2010) Citizen science as an ecological research tool: challenges and benefits. Annu Rev Ecol Evol Syst 41:149–172 Dickinson JL, Shirk J, Bonter D et al (2012) The current state of citizen science as a tool for ecological research and public engagement. Front Ecol Environ 10:291–297 Evans KL, Wilson JD, Bradbury RB (2007) Effects of crop type and aerial invertebrate abundance on foraging barn swallows Hirundo rustica. Agric Ecosyst Environ 122:267–273 Freitag A, Pfeffer MJ (2013) Process, not product: investigating recommendations for improving citizen science ‘‘success’’. PLoS One 8:e64079 Geary M, Fielding AH, Marsden SJ (2013) Designing mosaic landscapes for Black Grouse Tetrao tetrix using multi-scaled models. Ibis 155:792–803 Gibbs JP, Whiteleather KK, Schueler FW (2005) Changes in frog and toad populations over 30 years in New York State. Ecol Appl 15:1148–1157
123
Environmental Management Goddard MA, Dougill AJ, Benton TG (2010) Scaling up from gardens: biodiversity conservation in urban environments. Trends Ecol Evol 25:90–98 Gruebler MU, Korner-Nievergelt F, Von Hirschheydt J (2010) The reproductive benefits of livestock farming in barn swallows Hirundo rustica: quality of nest site or foraging habitat? J Appl Ecol 47:1340–1347 Khera N, Das A, Srivasatava S, Jain S (2010) Habitat-wise distribution of the house sparrow (Passer domesticus) in Delhi, India. Urban Ecosyst 13:147–154 Kuhn I, Brandl R, Klotz S (2004) The flora of German cities is naturally species rich. Evol Ecol Res 6:749–764 Lehtinen RM, Galatowitsch SM, Tester JR (1999) Consequences of habitat loss and fragmentation for wetland amphibian assemblages. Wetlands 19:1–12 Lepczyk CA, Flather CH, Radeloff VC, Pidgeon AM, Hammer RB, Liu J (2008) Human impacts on regional avian diversity and abundance. Conserv Biol 22:405–416 Luck GW, Ricketts TH, Daily GC, Imhoff M (2004) Alleviating spatial conflict between people and biodiversity. Proc Natl Acad Sci USA 101:182–186 Mcintyre NE (2000) Ecology of urban arthropods: a review and a call to action. Ann Entomol Soc Am 93:825–835 Mckinney ML (2002) Urbanization, biodiversity, and conservation: the impacts of urbanization on native species are poorly studied, but educating a highly urbanized human population about these impacts can greatly improve species conservation in all ecosystems. Bioscience 52:883–890 Miyashita T, Chishiki Y, Takagi SR (2012) Landscape heterogeneity at multiple spatial scales enhances spider species richness in an agricultural landscape. Popul Ecol 54:573–581 Moran PaP (1950) Notes on continuous stochastic phenomena. Biometrika 37:17–23 Naiman RJ, Decamps H, Pollock M (1993) The role of riparian corridors in maintaining regional biodiversity. Ecol Appl 3:209–212 Newton I (1995) The contribution of some recent research on birds to ecological understanding. J Anim Ecol 64:675–695
123
Robillard A, Garant D, Belisle M (2013) The swallow and the sparrow: how agricultural intensification affects abundance, nest site selection and competitive interactions. Landsc Ecol 28:201–215 Saizen I, Mizuno K, Kobayashi S (2006) Effects of land-use master plans in the metropolitan fringe of Japan. Landsc Urban Plan 78:411–421 Salek M, Svobodova J, Zasadil P (2010) Edge effect of low-traffic forest roads on bird communities in secondary production forests in central Europe. Landsc Ecol 25:1113–1124 Schulze CH, Waltert M, Kessler PJA et al (2004) Biodiversity indicator groups of tropical land-use systems: comparing plants, birds, and insects. Ecol Appl 14:1321–1333 Silvertown J (2009) A new dawn for citizen science. Trends Ecol Evol 24:467–471 Soule ME, Bolger DT, Alberts AC, Wrights J, Sorice M, Hill S (1988) Reconstructed dynamics of rapid extinctions of haparral-requiring birds in urban habitat islands. Conserv Biol 2:75–92 Sugawara K (1993) Hirundo rustica and Sturnus cineraceus. Photo Book, Seibundo Shinkosha, Tokyo Suzuki H (1998) Nests situations on Hirundo rustica in Tahata river, Hiroshima. Strix 16:99–108 Takagawa S, Al E (2012) JAVIAN Database: a species-level database of life history, ecology and morphology of bird species in Japan. Bird Res 7:R9–R12 Tsuji M, Ushimaru A, Osawa T, Mitsuhashi H (2011) Paddyassociated frog declines via urbanization: a test of the dispersaldependent-decline hypothesis. Landsc Urban Plan 103:318–325 Turner A (2006) The barn swallow. A&C Black, London Urban MC, Skelly DK, Burchsted D, Price W, Lowry S (2006) Stream communities across a rural urban landscape gradient. Divers Distrib 12:337–350 Van Rensburg BJ, Chown SL, Gaston KJ (2002) Species richness, environmental correlates, and spatial scale: a test using South African birds. Am Nat 159:566–577
