2D and 3D territories and territory overlap of Willow Warblers Phylloscopus trochilus in Arctic forests

doi:10.21203/rs.3.rs-2611464/v1

Download PDF

Research Article

2D and 3D territories and territory overlap of Willow Warblers Phylloscopus trochilus in Arctic forests

https://doi.org/10.21203/rs.3.rs-2611464/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 01 Jul, 2023

Read the published version in Polar Biology →

You are reading this latest preprint version

Space use by animals is usually considered in two horizontal dimensions, less often with a separate analysis of the vertical stratification. Only recently, studies on three-dimensional (3D) space use appeared with modelling of individual movements simultaneously in three dimensions. In ornithology, such works were done when studying home ranges of large birds tagged with radio transmitters and GPS locators or during visual observations of colour-banded passerine wintering birds and residents in tropics. To our knowledge, such methods have not been used for studying 3D territories advertised by singing in passerines during the breeding period in the Palearctic Region. We studied the 3D territoriality of Willow Warblers Phylloscopus trochilus during the breeding period in the Kola Peninsula, Arctic Region, NW Russia. Using the method of Cooper et al. (2014), we located points with 3D coordinates at which individually marked males sang and analysed 2D and 3D territories and their overlap in R. We found that 2D methods, compared to 3D methods, overestimated individual spatial overlap by 8.5% and total spatial overlap by 16.7%. Thus, Willow Warblers in Arctic forests can partition their territorial space in 3 dimensions. With the overlap indices adopted for use with 3D data, we also proved that they could avoid areas of overlap. Furthermore, it was demonstrated for the first time that the overestimation of the territory overlaps by 2D versus 3D methods grew higher as the number of neighbour territories increased. In dense populations, 3D territories are arranged like layers of roof tiles, avoiding substantial overlaps.

Phylloscopus trochilus

space use

3D territoriality

territory

overlap

Arctic

Space use patterns are among the key characteristics of each animal species (Ivanitsky 1998) and description of the models of space use by various species is essential for understanding their ecology and for adequate conservation planning (Cantrell et al. 2007; Anich et al. 2009; McKellar et al. 2015; Aliperti et al. 2021; Buderman et al. 2021; Hirt et al. 2021; Peters et al. 2022). A concept closely related to the concept of space use is territoriality. Yet, despite the century-long history of animal territoriality research (Howard 1920), there still remain some poorly studied aspects, including analysis of 3D territories. In fact, starting from the very first papers on territoriality in animals (Howard 1920; Noble 1939; Nice 1941), territory has been construed and analysed as an area
(defended or advertised) and, hence, as two-dimensional. One of the first methods that have been mostly commonly applied to collect and treat data in such studies is the method of minimum convex polygons (Mohr 1947) permitting territories to be outlined around their extreme points and their areas and perimeters to be analysed. In turn, among the first methods to be suggested for assessing the uneven use of the studied space were techniques that in addition to outlining this space permitted estimating the probability of finding an individual in each given point of the space (utilisation distribution) (Van Winkle 1975). This approach was later applied in the widely recognised kernel analysis, although initially it was only used to analyse two-dimensional data (Belant et al. 2012).

Meanwhile, some groups of animals (living in the water column or soil, inhabiting vegetation canopies, and capable of flying) actively utilise their habitats not only in 2D but also in 3D space. The first conclusions regarding the spatial (or, less often, spatial-temporal) division of the habitat space between individuals were widely represented in the ecological niche concept (Grinnell 1917; Hutchinson 1957; MacArthur 1968; Pianka 1974; Schoener 1974; Chesson 1991). In ornithology, similar ideas are communicated in studies on the vertical stratification of birds in different vegetation types (Pearson 1971; Cody 1974, 1985; Anderson et al. 1979; Terborgh 1980; Walther 2002; Acharya and Vijayan 2017; Cooper et al. 2021; Pires et al. 2022, etc.). The concepts of ecological niches and habitat stratification, however, are usually applied to different species or, at least, to groups of organisms rather than to individuals or reproductive pairs.

As regards the study of territories belonging to solo individuals or breeding pairs, some researchers, specifically in ornithology, have quite long ago pointed out the fact that territories of forest birds are 3D rather than 2D structures (Williamson 1971; Askins 1987). However, only a few previous studies took into account the vertical distribution of birds when actually analysing individual territories. Examples of such early studies were those by Zubtsovsky and Matantsev (1992), whose registration of advertised territories included the height at which the bird was in each recorded point. These studies, as well as their extensions (Zubtsovsky et al. 2006; Matantseva and Simonov 2008; Matantseva et al. 2017) have demonstrated that, as opposed to the perception of territories as planar objects, 3D mapping of territorial spaces can contribute significantly to our knowledge of the matter, since the patterns of individuals’ distribution in the habitat space can be a weighty factor shaping the bird settlement structure. Such research, however, was hindered by the lack of specialised techniques for collecting, processing, and representing data on 3D territories of birds.

It is only recently that papers on so-called 3D territoriality and 3D space use (not vertical stratification only) were published describing specific approaches to the analysis of space use by animals, for instance, in the study of aquatic animals (e.g. Ehrenberg and Steig 2003; Zhu and Weng 2007; Lee et al. 2017; Vivancos et al. 2017; Aspillaga et al. 2019; Matley et al. 2021); soil invertebrates (Bastardie et al. 2003), jumping or flying insects (Ahmed et al. 2021; Chen et al. 2021), and arboreal mammals (Chandler et al. 2020). In some studies, researchers offer and use more widely applicable methods (Smouse et al. 2010; Simpfendorfer et al. 2012; Tracey et al. 2014). Where monitoring of temporal changes in animal settlement structure was needed, it was suggested to complement the analysis of the 3D coordinate system of mapping the animal’s location in space with the analysis of a fourth dimension, i.e., time (e.g. Smouse et al. 2010; Ferrarini et al. 2018). In ornithology, such works were performed when studying home ranges of large birds tagged with radio transmitters and GPS locators (Tracey et al. 2014; Ferrarini et al. 2018) or in the course of visual observations of colour-banded passerine wintering birds and residents in the tropics (Cooper et al. 2014; Powell et al. 2021). As far as we know, such methods have not been used for studying the territorial spaces advertised by singing (3D territories) in passerines during the breeding period in the Palearctic region.

The vast Palearctic expanses offer habitats for a great number of birds, with conditions essentially different from those in tropics. Research on 3D territories of Palearctic birds and comparison of the results with data from the tropical belt can provide a deeper insight into the birds’ adaptations to different environments. Furthermore, many species that spend the winter in the tropics breed in the Palearctic region – a fact that offers unique opportunities for studying 3D territories in particular and, in general, space use by members of the same species under different environmental conditions and in different stages of the annual cycle. Besides, studies of 3D territories being scarce, there is a high change of discovering new data on this matter in most varied regions, including those with extreme conditions. In particular, studies in the Arctic can deepen our understanding of the mechanisms of adaptation to high latitudes in birds. Generally speaking, research on 3D space use can also produce novel data on other aspects of ecology and ethology. This approach can be useful, specifically, in the analysis of ecological niche differentiation, especially in vertically complex environments, intra- and interspecies competition, habitat selection and use, and wildlife conservation (Cunha and Vieira 2004; Cooper et al. 2014, 2021; Chandler et al. 2020).

Here, we investigate the space use and overlap in territorial Willow Warbler Phylloscopus trochilus males during the breeding period in the Kola Peninsula, the Arctic Region of NW Russia. The Willow Warbler Phylloscopus trochilus is a migratory songbird species which breeds in northern Europe and Asia, winters in sub-Saharan Africa, and occupies a broad range of light forest and scrub habitats (Cramp et al. 1994; Hume 2002). During the breeding period, males of this species are highly territorial, advertise their territories with singing and defend them from other conspecific males (May 1949; Lawn 1982, 1994; Kopij 2009; Simms 1985). Nevertheless, overlaps of their 2D territories were sometimes noted in this species (e.g. Enemar et al. 1979; Arvidsson and Klaesson 1986). We undertook to investigate how much the territories of Willow Warblers in their dense populations can really overlap and how the birds use the shared space in such cases.

Study system

We collected data on territorial Willow Warbler males during the breeding period in the Kola Peninsula, Murmansk Region, Northwest Russia. The study areas were located in three different points on the North-East of the Kola Peninsula (Figure 1) – on the banks of rivers Kolmak (67°10′N, 39°56′E), Acherjok (67°23'N, 39°26'E), and Drozdovka (68°17'N, 38°26'E), and were named Kolmak, Acherjok, and Drozdovka respectively. In the first two study areas, Kolmak and Acherjok, dense populations of Willow Warblers (up to 20–25 pairs/km²) occupy Arctic riparian forests (Figure 2), which form green belts up to 300–600 m wide along rivers with forested tongues along groundwater springs protruding into the dwarf-shrub tundra predominantly represented by downy willow (Salix lapponum) with dwarf birch (Betula nana) (Figure 3), where the population density of these birds is significantly lower (2–4 pairs/km²). We established a 2000 m x 400 m study plot in Kolmak area and 900 m x 800 m study plot in Acherjok area. Both these plots were dominated by downy birch (Betula pubescens) with canopy height averaging 8–10 m and occasional tall spruce trees (Picea sp.) rising up to 20–25 m above the main canopy and with a scarce understory composed of common juniper (Juniperus communis) and deciduous trees and shrubs such as alder (Alnus kolaensis), aspen (Populus tremula), and rowan (Sorbus gorodkovii). The ground was predominantly covered by a thick layer of moss (Dicranum, Pleurozium schreberi, Barbilophozia hatcheri, and Ptilidium ciliare), where Willow Warblers preferred to nest, and with occasional rock outcrops (Figure 1).

In the most northern study area Drozdovka (Figure 1), the population density of Willow Warblers reached 16 pairs/km². The width of the riparian forests here was only 50–300 m. Some forests were represented by isolated fragments of 900–1000 m². Unlike the first two study areas, the forest habitats here were surrounded by the predominantly moss-lichen, and not shrub, tundra, which was not settled by Willow Warblers at all (Figure 4). In this region, we established a 1km x 1km study plot, which covered several fragments of forest and a forest line along the bank of the Drozdovka River. By species composition, it was close to the study plots in the Kolmak and Acherjok areas, except that there were no conifers in the tree layer, and the average height of the top forest canopy reached 10 m.

Data collection

We collected data from 06 to 25 June 2019 in Kolmak study area, 02 to 22 June 2021 in Acherjok study area, and 13 to 29 June 2022 in Drozdovka study area (Figure 1). We defined a territory conceptually in the traditional and most common sense as a defended space (Howard 1920; Noble 1939; Maher and Lott 1995) and operationally – as the space advertised by singing, which is conventional for bird studies involving passerines (Maher and Lott 1995). It was not possible to record data blind because our study involved focal animals in the field. Following Cooper et al. (2014) indicating that utilization distribution methods are robust to spatial autocorrelation (Swihart and Slade 1997), we have decided to collect data in a manner that was biologically rather than statistically independent (see Lair 1987).

We marked all males in the study plot with plastic coloured rings and then identified and located them visually. We estimated territories and their overlap for focal males and as many of their neighbours as possible. Observers visited focal territories as often as possible, but not less often than once every three days, depending on the periods of active singing and weather conditions (usually in the early morning and when fair weather returned after rain and sometimes snow). Using mobile devices with the Android OS and Orux Maps program (©Jose Vazquez) with highly detailed satellite maps and GPS, we located the positions of every singing male as a point with geographical coordinates complemented by the registration of height. While we used the GPS integrated in mobile devices, which cannot deliver needed precision, the GPS only served to find our bearings by determining the approximate location in the map. It was after that, relying on fixed landmarks, that we accurately positioned each individual, thus getting exact “map to location” referencing for each case and avoiding GPS-only positioning errors. With this approach, the location capture error did not exceed 1 m.

Having sighted each singing individual, two observers consensually located it in the detailed map as described above and determined the height at which it was singing. We estimated the height visually (with down to 1 m precision) using marks on trees in each territory to calibrate the estimates. Then, with reference to the first point, the observers recorded all the points to which the male visibly moved (covering 1 m or more in horizontal or vertical directions). In doing so, at least one of the observers kept an eye on the focal bird. In the case both observers lost sight of the individual, they resumed observations after spotting the bird. Thus, the frequency of location registrations for a singing bird depended on the frequency of its movements in the process of singing, but as a rule it was at least 10 locations within 10 minutes, which was the minimum duration of observations of each individual in a day. As a rule, we got 20–30 locations for each monitored bird within a day. In total, we managed to collect 106–250 locations for each male (N = 51) in the study plots, which corresponds to the minimum sample size required for analysing 3D territories of birds – 80–110 locations (Cooper et al. 2014).

Minimum convex polygons (MCP)

This paper includes a discussion of the results obtained by the minimum convex polygon (MCP, the smallest polygon which encloses all points where an individual was registered) method, since it was the groundwork for a majority of early studies on territoriality in animals and is still used by some researchers. An additional task for us was to assess the performance of this method in studying territoriality in small passerine birds, namely in Willow Warblers. For plotting MCPs and calculating their areas we used the packages ‘adehabitatHR’ (Calenge 2006), ‘sp’ (Pebesma and Bivand 2005; Bivand et al. 2013), ‘scales’ (Wickham and Seidel 2020).

Kernel density estimation (KDE)

Following Cooper et al. (2014), we used the package ‘ks’’ (Duong 2007) to create 2D and 3D utilisation distributions (for sample code, see Cooper et al. 2014) and defined each male’s territory as the 95th isopleth.

Area and volume of territories

We quantified the area (2D) for each Willow Warbler territory as the MCP and 95% KDE. The volume (3D) was calculated only as 95% KDE. As stated above, we identified territory as the space advertised by singing.

Spatial overlap

We estimated spatial overlap for pairs of neighbouring males as the area or volume of overlap divided by each individual’s territory area or volume, creating a proportion (‘individual spatial overlap’ by Cooper et al. 2014). In case of multiple overlapping of several neighbouring territories, we also assessed the total share of area or volume overlapped by the neighbours (‘total spatial overlap’ by Cooper et al. 2014). The area of overlap between MCPs of neighbouring territories was calculated using the ‘rgeos’ package (Bivand et al. 2021). Combining the polygon geometry functions ‘gArea’ and ‘gIntersection’ included in the package, we plotted the overlap polygon and then measured the area of the polygon. We also assessed differences between 2D and 3D estimates (as the differences in average values) to evaluate possible mistakes in calculating space partitioning by territorial males during the breeding period on the basis of 2D territories only.

To check whether Willow Warblers could avoid areas of overlap, we applied the overlap indices adopted for the use with 3D data by Cooper et al. (2014): Volume of Intersection Index (VI) and Utilisation Distribution Overlap Index (UDOI). These indices show how frequently areas of overlap are used by each male. The VI varies between 0 (no overlap) and 1 (identical space use, Seidel 1992). The UDOI also can vary between 0 (no overlap) and 1 (overlap is complete and both utilisation distributions are uniform) and can sometimes be > 1 (overlap is high and the utilisation distributions are not uniform) (Fieberg and Kochanny 2005). We calculated the UDOI adapted for the use with 3D data by Cooper et al. (2014) from Fieberg and Kochanny (2005). Following their recommendations, we divided the utilisation distribution by the actual sum of all the density estimates within the 95th isopleth for each male (for more details, see Cooper et al. 2014).

Data analysis

Statistical analyses were performed using RStudio 2021.09.1+372 Ghost Orchid (© 2009-2021 RStudio, PBC). Since the data were not normally distributed in some cases, we indicated the median (x͂) as a measure of central tendency and the interquartile range (IQR) as a measure of dispersion. Following Cooper et al. (2014), we used paired-sample Wilcoxon signed-rank tests to compare estimates of the overlap. We estimated the effect sizes (r) using ‘rstatix’ package (Kassambara 2021) and the method by Field (2005), according to which the effect size r is calculated as Z statistic divided by square root of the sample size (N). Besides, also following the recommendation of Cooper et al. (2014), we used the Kendall’s coefficient of concordance (Kendall’s W, Kendall and Smith 1939) and ‘DescTools’ package (Signorell et al. 2021) to find out whether application of individual spatial overlap, VI and UDOI ranked pairs of birds similarly. As Cooper et al. (2014) pointed out, VI and UDOI are non-directional (give one value for every pair of males) whereas the individual spatial overlap is directional. Therefore, one should average the two individual spatial overlap values for each pair of neighbouring males to get a single value (Fieberg and Kochanny 2005) in order to compare the individual spatial overlap with these indices.

Area and volume of territories in the study plot

According to our results calculated with 95% KDE method, the areas of 2D territories of Willow Warblers in the studied populations widely varied from 1587.8 to 32427.9 m² (x͂ = 6122.4 m², IQR = 8147.9, N = 51), and the volumes of 3D territories varied from 2257.2 to 129171.6 m³(x͂ = 11604.2 m³, IQR = 19638.2, N = 51). Figure 5 shows the bagplot (Rousseeuw et al. 1999; Wolf 2019) of the relationship between the territory areas and volumes.

Some male Willow Warblers in the studied population were observed to perform rare visits far beyond the sites in which they sang most of the time. During such distant visits, the males could sing, which explains the singing-marked locations found far out of the main data cloud composed for each male. Hence, when territories were outlined around the extreme points of singing events (as MCPs), their area could grow disproportionately. Indeed, the territory area calculated as MCP area enclosing singular distant visits by singing males could be extra large – 2925.4 to 141065.0 m² (x͂ = 16597.3 m², IQR = 42146.4, N = 51), while for most of the time the males stayed in the core parts of their MCP territories (an example is given in Figure 6). It is also noteworthy that a male’s singing points in most cases formed several distinct blobs defining its territory (Figure 6). A vast majority of the outlined territories were characterised by a clustered structure.

Individual spatial overlap of neighbouring territories

In total, for neighbouring 2D territories, we registered 101 cases of the overlap by the MCP method and 41 cases of the overlap by the 95% KDE method (Figures 6, 7). For 3D territories, we found 40 cases of the overlap according to the 95% KDE method. The values of individual spatial overlap of territories calculated as MCPs varied immensely, from a fraction of a percent to 100% (x͂ = 15.8%, IQR = 42.1%, N = 202). For 2D 95% KDE territories this parameter varied from 0.3 to 85.4% (x͂ = 8.8%, IQR = 13.6%, N = 82) and for 3D 95% KDE territories it varied from 0.01 to 42.7% (x͂ = 3.5%, IQR = 5.9%, N = 80) (Figure 7).

In 3 cases out of 41 (i.e., 7% of cases), when the projections of the displayed spaces on the ground surface overlapped (2D 95% KDE territories), these spaces did not really overlap in 3D (an example of this is shown in Figure 6). The difference was even greater when comparing 3D overlap and 2D overlap counts for the territories identified as MCPs – there was no overlap of 3D territories in 62.4% of cases (63 out of 101) where 2D MCP territories overlapped (free points in Figure 7). Meanwhile, where territories overlapped both in 2D projections (95% KDE) and in 3D space (95% KDE), the percentage of 3D overlap was significantly lower than the percentage of 2D overlap (2D KDE overlap: x͂ = 8.48%, IQR = 13.8%, 3D KDE overlap: x͂ = 2.65%, IQR = 5.52%, Wilcoxon signed-rank exact test: V = 3618, N = 86, Z = - 8.84, r = 0.674, p < 0.001; Figure 7).

Total spatial overlap of neighbouring territories

According to the 95% KDE, the total spatial overlap of territories was higher in 2D projections (x͂ = 12.2%, IQR = 40.1%) than in 3D space (x͂ = 4.4%, IQR = 15.5%) (paired-sample Wilcoxon signed-rank exact tests, V = 898, N = 42, Z = -6.91, effect size r = 0.76, p < 0.001; Figure 8). Maximum values of total overlap for 3D data were 62.2% per territory, whereas for 2D data in some cases it could be as high as 100.0% per territory (Table 1, Figure 8).

Overlap indices

To study potential avoidance of overlap areas by Willow Warblers, we compared the values of individual overlap with VI and UDOI (Cooper et al. 2014). VI and UDOI proved to be lower than individual spatial overlap (Table 1) both for 2D and 3D estimates. We can conclude from this that Willow Warblers avoided areas of overlap. Each of the 3 measures, averaged individual overlap, VI and UDOI, ranked pairs of birds quite differently in terms of spatial overlap (2D: Kendall’s W = 1, df = 2, P = <0.001, 3D: Kendall’s W = 1, df = 2, P = <0.001).

Overestimation of territorial overlap with 2D methods versus 3D methods

Overall, according to the results of data analysis by 95% KDE methods, 2D methods, as opposed to 3D methods, overestimated individual (pairwise) spatial overlap by 8.5% and total (all neighbours combined) spatial overlap by 16.7%, calculated as the difference in average values. Such overestimation by 2D methods compared to 3D methods was found also for other parameters in the analysis (Table 1). We also observed that the overestimation of the territory overlaps by 2D methods versus 3D methods grew higher as the number of neighbour territories increased (Table 2).

Table 1 Median ± interquartile range (x͂±IQR) of Individual Overlap (IO), Total Overlap (TO), Averaged Proportional Overlap (APO), Volume of Intersection Index (VI), and Utilization Distribution Overlap Index (UDOI) and results of paired-sample Wilcoxon signed-rank exact tests comparing 2D and 3D 95% KDE measures of spatial overlap

Value	2D x͂ ± IQR	3D x͂ ± IQR	2D limits	3D limits	n	V	Z	r	p
IO	0.085 ± 0.138	0.027 ± 0.055	0.003 – 0.854	<0.001 – 0.427	84	3618	-8.84	0.67	<0.001
TO	0.122 ± 0.401	0.044 ± 0.155	0.013 – 1	0.0001 – 0.622	45	898	-6.92	0.76	<0.001
APO	0.091 ± 0.144	0.036 ± 0.061	0.005 – 0.683	<0.001 – 0.342	40	933	-6.71	0.72	<0.001
VI	0.027 ± 0.061	0.011 ± 0.021	0.001 – 0.477	<0.001 – 0.183	41	936	-6.81	0.73	<0.001
UDOI	0.002 ± 0.018	0.0004 ± 0.001	<0.001 – 0.466	<0.001 – 0.089	41	931	-6.64	0.72	<0.001

Table 2 Spearman rank correlation between the number of neighbours in 2D and 3D and overestimation of total overlap in 2D vs 3D

Parameter	S	Rho	p-value
Number of neighbours in 2D	2876.8	0.767	<0.001
Number of neighbours in 3D	3276.6	0.735	<0.001

Specific features of Willow Warbler territories in the study area

Having studied the case of Willow Warblers, we produced a first description of 3D territories of passerine birds during the breeding period in the Palearctic region. It turned out that in our study area, i.e. the Arctic region, where forest vegetation patches are small and fragmentary, Willow Warblers can form dense populations with closely neighbouring and even overlapping territories. The overlaps were particularly numerous where territories were outlined as MCPs but were observed also for 2D 95% KDE territories. It might seem at a first glance that territorial males in the studied populations actively share space with neighbour males. Whether this was really so and to what extent, was studied by analysing the sizes of 2D and 3D territories and their overlaps.

We found that in the studied dense populations of Willow Warblers during the breeding period in the Arctic forests, the area of 2D territories as well as the volume of 3D territories varied significantly (Figure 5). The volume of a territory increased proportionately to the increase in its area, although the volume increased at a greater scale (Figure 5). Significant variations in the territory size are, in fact, characteristic of this species in general (May 1949; Arvidsson and Klaesson 1986; Ryabitsev 1993). Another noteworthy fact is that often territories were not uniform but consisted of clusters – aggregations of locations (Figure 6). Something of the kind has been reported also for tropical regions (Cooper et al. 2014; Powell et al. 2021) – some of the territories fell into distinct clusters. A comparison of the spatial distribution of such clusters against elements of the vegetation structure in the occupied habitat in our study area has demonstrated that these clusters represent the birds’ favourite singing locations, such as certain crown parts or solitary shrubs. Hypothetically, such clustered structure of 3D territories occurs in different bird species inhabiting non-uniform environments. Further studies will help gather more data on this matter as well.

We also found that determination of the territory area by the MCP method is not the best way to describe space use when the bird mostly stays within the core area of the territory and only from time to time performs remote visits out of their core singing-advertised site and sings at the outlying points. The events of singing at points situated far away from the core of the territory, even though very rare, can result in a disproportionate enlargement of the MCP outlined territory, considering that its periphery remains virtually out of use (as in the large MCP territory example in Figure 6). Methods based on the KDE analysis seem much more effective in portraying space use in such cases. That said, the choice of the method to apply is dictated by the goals of the study. It is obvious that the MCP method is not applicable in estimating space use unevenness, but it can be useful in targeted registration and assessment of all movements, including rare ones. Also admittedly, in a more homogenous environment and with more uniform behaviour of different individuals in a population, 2D methods (incl. MCP) can produce a more realistic output and can still be used. This should, however, be done with utmost caution and taking into account the behaviour of individuals under specific habitat conditions.

Ratio of 2D and 3D overlaps

Expectedly, the number of overlaps between neighbouring territories was the highest when the MCP method was applied to outline territories. As stated above, we consider this method unreliable in situations similar to ours and so believe the overlaps detected by this method were excessively overestimated. Indeed, in the output of the kernel method, which accounts for the probability of finding an individual in different points in the space (even at 2D level), the number and share of overlaps turned out to be much lower. Thus, application of the MCP method alone could lead to false conclusions about multiple and quite substantial overlaps between neighbouring territories in the given populations. The overlaps (both individual and total) detected by the KDE method in 3D environment were even smaller (Figures 6, 7, 8), indicating that territories in 3D space overlapped less often and to a smaller extent than it appeared after the analysis of 2D data.

Our data show that 2D methods, compared to 3D methods, overestimated the individual (pairwise) spatial overlap by 8.7% and total (all neighbours combined) spatial overlap by 16.7%. Thus, Willow Warblers can partition their territorial space in three spatial dimensions. Furthermore, overestimation of the overlap by 2D methods versus 3D methods grew higher as the number of neighbour territories increased (Table 2). Thus, in dense populations with many neighbours the error in calculating the proportion of shared space can be even more critical. Our results are generally comparable with the results obtained by Cooper et al. (2014) for wintering American Redstarts Setophaga ruticilla in tropics, which indicate that, depending on a habitat, 2D methods overestimated individual spatial overlap by 3–4% and total spatial overlap by 9–12%. Our figures are somewhat higher than the ones reported by Cooper et al. (2014), which can be explained by differences in the structure of habitats or by the different biological characteristics of the two species. To be able to draw some conclusions on these matters, further research producing data from comparisons of 2D and 3D methods for different species in different habitats is needed.

Sharing space and avoiding areas of overlap

So, what is the actual proportion of the space shared by different males? According to the 3D kernel method, which determines the probability of finding an individual in each point of the 3D space, the median value of the individual overlap in the studied population was only 2.7% (but with significant variation among males from fractions of a percent to 42.7%), and that of total overlap – 4.4% (62.2% at maximum) (Table 1). Furthermore, having applied overlap indices (Cooper et al. 2014), we found that neighbouring males could avoid areas of overlap (Table 1). In a manner of speaking, 3D territories in dense populations are arranged like layers of roof tiles, avoiding substantial overlaps. These results are consistent with the data from Cooper et al. (2014) and earlier conclusions that the distribution of territorial spaces by height can occur in warblers when their 2D territories seem significantly overlapped (Matantseva and Simonov 2008; Matantseva et al. 2017).

Thus, even a strictly territorial species such as the Willow Warbler can, in certain circumstances (viz. high population density where the area of preferred habitats is limited), practice shared use of some areas, and the birds avoid them. Indeed, full demarcation of territories is not the only way of dividing space between individuals even in the species considered strictly territorial. In certain situations, typical territoriality can be substituted with alternative ways of lowering competition for resources, e.g. a hierarchy of relationships and/or sharing of the space or parts thereof, as observed in many species, including birds (Ryzhanovsky 1977; Dewsbury 1978; Titus and Hans 1990; Ryabitsev 1993; Newton 1998; Both and Visser 2003; Bas et al. 2005; Cooper et al. 2014; Matantseva et al. 2017, etc.). In reality, such division (or shared use) of space takes place in the habitat volume rather than just in the ground surface projection of the territory. These questions are obviously also worth studying in 3D space.

An important observation is that Willow Warblers demonstrated high variation in the space use by different individuals, as evidenced by high variation in the size of territories and share of their overlap. This high individual behavioural variation in space use can hypothetically result in the formation of different spatial structures of the species populations in different conditions. To verify this hypothesis and find out how the behaviour of birds in this species can change in different conditions further studies are to be carried out in different regions, in populations with different densities, and in areas with different habitat structures. Such studies will also be especially useful when analysing 3D territories.

The analysis of 2D and 3D territories of Willow Warblers revealed that birds of this species actively utilise the habitat space in all three dimensions. In Arctic forests, where preferred habitats are in deficit, they may form a dense mosaic of neighbouring territories arranged like layers of roof tiles, avoiding substantial overlaps. Such data could not have been obtained by analysing 2D space alone. The territory overlaps, both individual and total, would be overestimated by 2D methods versus 3D methods. Overall, our studies corroborate the conclusion of previous research (Simpfendorfer et al. 2012; Cooper et al. 2014) that 3D analysis more accurately describes the spatial segregation of individuals, taking into account the vertical component, and prevents overestimation of the overlap between neighbouring territories, which may happen when the more conventional 2D analysis is applied. We have also demonstrated for the first time that the territory overlap estimation error increases with growing local population density. In other words, the higher the density of the population, the more cautious one should be about applying 2D methods and the more relevant 3D methods would be.

Apart from the basic theoretical significance of the above-described studies on 3D space use (see Introduction), there is a practical dimension to such research. In particular, overestimation of potential territory overlap, especially in dense populations, can lead to fallacious conclusions of high behavioural lability of territorial animals. Clear understanding of the limits of behavioural lability in animals occupying different types of habitats, in turn, is necessary when planning actions to build and maintain certain bird numbers in natural and disturbed habitats. Data on space partitioning among conspecifics occupying habitats of limited size can also prove useful when assessing the feasibility of fragmented habitat occupation by members of a certain species. The question of whether different animals can settle in limited-size habitats is of special relevance in view of biodiversity conservation issues and the need to maintain normal ecosystem functioning under growing human pressure, shrinking and fragmentation of natural habitats. Another area where 3D models of space use by animals can be effective is the development of habitat conservation measures for some species, including rare and endangered ones. To conclude, the study of 3D territoriality in birds is turning into a hot topic in ornithology, since territoriality is one of the key ways of distributing individuals in space (1), real territories are three-dimensional and not flat (2), new methods of analysing data on the use of 3D space by animals are emerging (3), and analysis of 3D space use hold immense promise for ecological and ethological studies.

Acknowledgements We thank Konstantin Tirronen and Danila Panchenko for organizing the expeditions to the Kola Peninsula during which our fieldwork was carried out. We are also grateful to Olga Kislova for her help in translating the manuscript into English. The investigations comply with the current laws of the Russian Federation. The paper was prepared while implementing activities under Russian Science Foundation grant # 23-24-00092, https://rscf.ru/project/23-24-00092/.

Significance statement

3D mapping of territorial spaces can contribute significantly to our knowledge of the bird territoriality and space use, since the patterns of individuals’ distribution over in the space of the habitat can be a weighty factor shaping their settlement structure. With the publication of Cooper et al. (2014) a proper method for such research was offered and applied to 3D territories of passerines in tropics. Our work is the first research using this method to study advertised territories of songbirds during the breeding period in the Palearctic Region. By studying Willow Warblers during the breeding period, we proved that these birds actively utilise the habitat space in all three dimensions and avoid areas of neighbouring territories’ overlap. We also found that 2D methods, compared to 3D methods, can overestimate the territory size and overlap significantly and should be used carefully.

Ethical note All aspects of the study were planned and executed on the principles that observers must not injure birds in conformity with the “Guiding principles in the care and use of animals”. Capturing, banding and observations were conducted according to the Federal Law of the Russian Federation № 52 dated 24.04.1995 (ed. 11.06.2021) “regarding the animal kingdom” (24.04.1995).

Competing Interests There is no conflict of interest regarding the publication of this paper.

Author contributions MM and SS conceived, designed and conducted field research. SS analysed data. MM wrote the manuscript. Both authors read and approved the manuscript.

Data availability

The used unpublished script and datasets generated and analysed during the current study are available in the Figshare repository, https://doi.org/10.6084/m9.figshare.19361678.v3

Acharya B, Vijayan L (2017) Vertical stratification of birds in different vegetation types along an elevation gradient in the Eastern Himalaya, India. Ornithol Sci 16:131–140. https://doi.org/10.2326/osj.16.131
Ahmed AD, Benhamou S, Bonsall M, Petrovskii S (2021) Three-dimensional random walk models of individual animal movement and their application to trap counts modelling. J Theor Biol 524:110728. https://doi.org/10.1016/j.jtbi.2021.110728
Aliperti JR, Davis BE, Fangue NA, Todgham AE, Van Vuren DH (2021) Bridging animal personality with space use and resource use in a free-ranging population of an asocial ground squirrel. Anim Behav 180:291–306. https://doi.org/10.1016/j.anbehav.2021.07.019
Anderson SH, Shugart HH, Smith TM (1979) Vertical and temporal habitat utilization within a breeding bird community. In: Dickson JD, Conner RN, Fleet RR, Jackson JA, Kroll JC (Eds.) The role of insectivorous birds in forest ecosystems, New York: 203–216
Anich NM, Benson TJ, Bednarz JC (2009) Estimating territory and home-range sizes: do singing locations alone provide an accurate estimate of space use? Auk 126(3):626–634. https://doi.org/10.1525/auk.2009.08219
Arvidsson B, Klaesson P (1986) Territory size in a Willow Warbler Phylloscopus trochilus population in mountain birch forest in Swedish Lapland. Ornis Scand 17(1):24–30. https://doi.org/10.2307/3676749
Askins RA (1987) Bird territories: A key to understanding bird behavior. Am Birds 41: 35–40
Aspillaga E, Safi K, Hereu B, Bartumeus F (2019) Modelling the three-dimensional space use of aquatic animals combining topography and Eulerian telemetry data. Methods Ecol Evol 10:1551–1557. https://doi.org/10.1111/2041-210X.13232
Bas JM, Pons P, Gómez C (2005) Home range and territory of the Sardinian Warbler Sylvia melanocephala in Mediterranean shrubland: capsule singing territories were well separated. Bird Study 52(2):137–144. https://doi.org/10.1080/00063650509461383
Bastardie F, Capowiez Y, Cluzeau D (2003) Burrowing behaviour of radio-labelled earthworm revealed by analysis of 3D-trajectories in artificial soil cores. Pedobiologia 47:554–559
Belant JL, Millspaugh JJ, Martin JA, Gitzen RA (2012) Multi-dimensional space use: The final frontier. Front Ecol Environ 10:11–12
Bivand RS, Pebesma E, Gomez-Rubio V (2013) Applied spatial data analysis with R, Second edition. Springer, NY. https://asdar-book.org/
Bivand R, Rundel C (2021) rgeos: Interface to Geometry Engine – Open Source ('GEOS'). R package version 0.5-8. https://CRAN.R-project.org/package=rgeos
Both C, Visser M (2003) Density dependence, territoriality, and divisibility of resources: from optimality models to population processes. Am Nat 161:326–36. https://doi.org/10.1086/346098
Buderman F, Gingery T, Diefenbach D, Gigliotti L, Begley-Miller D, McDill M, Wallingford B, Rosenberry C, Drohan P (2021) Caution is warranted when using animal space-use and movement to infer behavioral states. Mov Ecol 9. https://doi.org/10.1186/s40462-021-00264-8
Calenge C (2006) The package adehabitat for the R software: tool for the analysis of space and habitat use by animals. Ecol Model 197:1035
Cantrell R, Cosner C, Deangelis D, Padrón V (2007) The ideal free distribution as an evolutionarily stable strategy. J Biol Dyn 1:249–71. https://doi.org/10.1080/17513750701450227
Chandler CJ, Van Helden BE, Close PG, Speldewinde PC (2020) 2D or not 2D? Three-dimensional home range analysis better represents space use by an arboreal mammal. Acta Oecol 105:103576. https://doi.org/10.1016/j.actao.2020.103576
Chen CH, Chiang AS, Tsai HY (2021) Three-dimensional tracking of multiple small insects by a single camera. J Insect Sci 21(6). https://doi.org/10.1093/jisesa/ieab079
Chesson P (1991) A need for niches? Trees 6(1):26–28
Cody ML (1974) Competition and the Structure of Bird Communities. Monogr Pop Biol 7, Princeton Univ Press, Princeton, New Jersey
Cody ML (1985) Habitat selection in birds. Academic press, San Diego
Cooper NW, Sherry TW, Marra P (2014) Modeling three-dimensional space use and overlap in birds. Auk 131(4):681–693. https://doi.org/10.1642/AUK-14-17.1
Cooper NW, Thomas MA, Marra PP (2021) Vertical sexual habitat segregation in a wintering migratory songbird. Ornithology 138:1–11. https://doi.org/10.1093/ornithology/ukaa080
Cramp S, Perrins CM, Brooks DJ, Dunn E, Gillmor R, Hall-Craggs J, Hillcoat B, Hollom PAD, Nicholson EM, Roselaar CS, Seale WTC, Sellar PJ, Simmons KEL, Snow DW, Vincent D, Voous KH, Wallace DIM, Wilson MG (1994) Handbook of the Birds of Europe and Middle East and North Africa. Vol. 8. Oxford Univ. Press
Cunha AA, Vieira MV (2004) Two bodies cannot occupy the same place at the same time, or the importance of space in the ecological niche. Bull Ecol Soc Am 116:25–26.
Dewsbury DA (1978) Comparative Animal Behavior. New York: McGraw Hill
Duong T (2007) ks: kernel density estimation and kernel discriminant analysis for multivariate data in R. J Stat Softw 21(7):1–16
Ehrenberg JE, Steig TW (2003) Improved techniques for studying the temporal and spatial behavior of fish in a fixed location. Int J Mar Sci 60:700–706
Enemar A, Klaesson P, Sjöstrand B (1979) Accuracy and efficiency of mapping territorial Willow Warblers Phylloscopus trochilus: a case study. Oikos 33(2):176–181. https://doi.org/10.2307/3543995
Ferrarini A, Giglio G, Pellegrino SC, Frassanito AG, Gustin M (2018) A new methodology for computing birds’ 3D home ranges. Avian Res 9:19. https://doi.org/10.1186/s40657-018-0109-6
Fieberg J, Kochanny CO (2005) Quantifying home-range overlap: The importance of the utilization distribution. J Wildl Manage 69:1346–1359
Field A (2005) Discovering Statistics Using SPSS, second edition. Sage, London, UK
Hirt MR, Barnes AD, Gentile A, Pollock LJ, Rosenbaum B, Thuiller W et al (2021) Environmental and anthropogenic constraints on animal space use drive extinction risk worldwide. Ecol Lett 24:2576–2585. https://doi.org/10.1111/ele.13872
Grinnell J (1917) The niche relationships of the California thrasher. The Auk 34:427–433
Howard HE (1920) Territory in bird life. London, John Murray, Albemarle Street, W
Hume R (2002) Complete birds of Britain and Europe. London: Dorling Kindersley Limited
Hutchinson GE (1957) Concluding remarks. Cold Spring Harbor Symp. Quant Biol 22:415–427
Ivanitsky VV (1998) Spatially-oriented behavior of birds. Ornithology 28:6–24 (in Russian)
Kassambara A (2021) rstatix: Pipe-Friendly Framework for Basic Statistical Tests. R package version 0.7.0. https://CRAN.R-project.org/package=rstatix
Kendall MG, Smith BB (1939) The problem of m rankings. Ann Math Stat 10:275–287. https://doi.org/10.1214/aoms/1177732186
Kopij G (2009) Territoriality in the Willow Warbler Phylloscopus trochilus in its winter quarter in Lesotho. Ornis Svec 19: 50–54. https://doi.org/10.34080/os.v19.22658
Lair H (1987) Estimating the location of the focal center in red squirrel home ranges. Ecology 68:1092–1101
Lawn MR (1982) Pairing systems and side tenacity of the Willow Warbler Phylloscopus trochilus in Southern England. Ornis Scand 13:193–199
Lawn MR (1994) Site fidelity and annual survival of territorial male Willow Warblers Phylloscopus trochilus at four adjacent sites in Surrey. Ringing Migr 15:1–7
Lee KA, Huveneers C, Duong T, Harcourt RG (2017) The ocean has depth: two- versus three-dimensional space use estimators in a demersal reef fish. Mar Ecol Prog Ser 572:223–241. https://doi.org/10.3354/meps12097
MacArthur RH (1968) The theory of niche. Population Biology and Evolution Syracuse, New York: Syracuse Univ. Press: 159–176
Maher CR, Lott DF (1995) Definitions of territoriality used in the study of variation in vertebrate spacing systems. Anim Behav 49(6):1581–1597
Matantseva MV, Simonov SA (2008) Ecological and ethological characteristics of Sylvia warblers’ colonies in patchy habitats on the Courland Spit, Baltic Sea. Russ J Ecol 39(5):353–358
Matantseva M, Simonov S, Lapshin N (2017) The variability in territorial behavior of Sylvia warblers depending on the structure of habitats and population density. Principy èkologii 6(3):101–117. https://doi.org/10.15393/j1.art.2017.7002 (in Russian, with English summary)
Matley JK, Johansen LK, Klinard NV, Eanes ST, Jobsis PD (2021) Habitat selection and 3D space use partitioning of resident juvenile hawksbill sea turtles in a small Caribbean Bay. Mar Biol 168:120. https://doi.org/10.1007/s00227-021-03912-0
May D (1949) Studies on a community of Willow Warblers. Ibis 91(1):24–54. https://doi.org/10.1111/j.1474-919X.1949.tb02234.x
McKellar AE, Langrock R, Walters JR, Kesler DC (2015) Using mixed hidden Markov models to examine behavioral states in a co-operatively breeding bird. Behav Ecol 26(1):148–157. https://doi.org/10.1093/beheco/aru171
Mohr CO (1947) Table of equivalent populations of North American small mammals. Am Midl Nat 37:223–249
Newton I (1998) Population limitation in birds. Academic Press, San Diego
Nice M (1941) The role of territory in bird life. Amer Midl Nat 26:441–487
Noble GK (1939) Dominance in the life of birds. Auk 56:263–273
Pearson DL (1971) Vertical stratification of birds in a tropical dry forest. Condor 73:46–55. https://doi.org/10.2307/1366123
Pebesma EJ, Bivand RS (2005) Classes and methods for spatial data in R. R News 5(2):9–13. https://CRAN.R-project.org/doc/Rnews/
Peters NM, Beale CM, Bracebridge C, Mgumba MP, Kendall CJ (2022) Combining models for animal tracking: Defining behavioural states to understand space use for conservation. J Biogeogr 49(11):2016–2027. https://doi.org/10.1111/jbi.14483
Pianka ER (1974) Niche overlap and diffuse competition. PNAS 71(5):2141–145
Pires LP, Paniago LPM, Santos YR, de Melo C (2022) Seasonality drives variation in the use of forest strata by adult males of a dimorphic frugivorous bird species. Austral Ecol 47:392–399. https://doi.org/10.1111/aec.13129
Powell LL, Ames EM, Wright JR, Matthiopoulos J, Marra PP (2021) Interspecific competition between resident and wintering birds: experimental evidence and consequences of coexistence. Ecology 102(2):e03208. https://doi.org/10.1002/ecy.3208
Rousseeuw P, Ruts I, Tukey J (1999) The Bagplot: A Bivariate Boxplot. American Statistician. 53:382–387. https://doi.org/10.1080/00031305.1999.10474494
Ryabitsev VK (1993) Territorial relations and communities’ dynamics of birds in Subarctic. Ekaterinburg: Nauka Ural. otd. (in Russian, with a table of contents in English)
Ryzhanovsky VN (1977) Ecology of the Meadow Pipit and Red-throated Pipit on the Yamal Peninsula. Comparative ecology of the bird background species of the Yamal Tundra. Sverdlovsk: 3–35 (in Russian)
Schoener TW (1974) Resource partitioning in ecological communities. Science 185:27–39
Seidel KS (1992) Statistical properties and applications of a new measure of joint space use for wildlife. M.S. thesis, University of Washington, Seattle, WA, USA
Signorell A et al (2021) DescTools: Tools for descriptive statistics. R package version 0.99.44. https://cran.r-project.org/package=DescTools
Simms E (1985) British Warblers. London: Collins
Simpfendorfer CA, Olsen EM, Heupel MR, Moland E (2012) Three-dimensional kernel utilization distributions improve estimates of space use in aquatic animals. Can J Fish Aquat Sci 69:565–572
Smouse PE, Focardi S, Moorcroft PR, Kie JG, Forester JD, Morales JM (2010) Stochastic modelling of animal movement. Philos Trans R Soc Lond B Biol Sci 365(1550):2201–2211. https://doi.org/10.1098/rstb.2010.0078
Swihart RK, Slade NA (1997) On testing for independence of animal movements. JABES 2:48–63
Terborgh L (1980) Vertical stratification of a Neo-tropical forest bird community. Acta XVII Cong Int Orn Berlin 1978:1005–1012
Titus RC, Hans CA (1990) Singing behavior of American Robins in linear and non-linear habitats. Wilson Bull. 102(2): 325–328
Tracey JA, Sheppard J, Zhu J, Wei F, Swaisgood RS, Fisher RN (2014) Movement-based estimation and visualization of space use in 3D for wildlife ecology and conservation. PLoS One 9:e101205. https://doi.org/10.1371/journal.pone.0101205
Van Winkle W (1975) Comparison of several probabilistic home-range models. J Wildl Manag 39:118–123
Vivancos A, Closs G, Tentelier C (2017) Are 2D space-use analyses adapted to animals living in 3D environments? A case study on a fish shoal. Behav Ecol 28(2):485–493. https://doi.org/10.1093/beheco/arw175
Walther BA (2002) Vertical stratification and use of vegetation and light habitats by Neotropical forest birds. J Ornithol 143:64–81. https://doi.org/10.1007/BF02465460
Wickham H, Seidel D (2020) scales: Scale Functions for Visualization. R package version 1.1.1. https://CRAN.R-project.org/package=scales
Williamson K (1971) The breeding birds of a century-old grove of coast and sierra redwoods in Wales. QJF 65:109–121
Wolf H (2019) aplpack: Another Plot Package (version 190512). https://cran.r-project.org/package=aplpack
Zhu L, Weng W (2007) Catadioptric stereo-vision system for the real-time monitoring of 3D behavior in aquatic animals. Physiol Behav 91:106–119
Zubtsovsky NE, Matantsev VA (1992) Results of the ornithological studies of the Department of Zoology for the last decade. Bulletin of Udmurt University 3:101–106 (in Russian)
Zubtsovsky NE, Matantsev VA, Matantseva MV (2006) Ethological lability of Sylvia warblers as a security mechanism of local populations’ stability. Transactions of the XII Intern. Ornithol. Conf. of Northern Eurasia. Stavropol: 587–600 (in Russian, with English summary)

No competing interests reported.

Download PDF

Journal Publication

published 01 Jul, 2023

Read the published version in Polar Biology →

Editorial decision: Major revision
26 Apr, 2023
Reviews received at journal
07 Mar, 2023
Reviewers agreed at journal
01 Mar, 2023
Reviewers invited by journal
27 Feb, 2023
Editor assigned by journal
27 Feb, 2023
Submission checks completed at journal
22 Feb, 2023
First submitted to journal
21 Feb, 2023

You are reading this latest preprint version

2D and 3D territories and territory overlap of Willow Warblers Phylloscopus trochilus in Arctic forests

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Material And Methods

Results

Discussion

Conclusions

Statements And Declarations

References

Additional Declarations

Status:

Journal Publication

Version 1