The impact of forest characteristics, and bird and insect diversity on the occurrence of the Great Spotted Woodpecker Dendrocopos major and Grey-headed Woodpecker Picus canus in South Korea

ABSTRACT

Capsule

Although the Great Spotted Woodpecker Dendrocopos major and Grey-headed Woodpecker Picus canus have overlapping ranges in South Korea, they occur in areas with different forest structures and the occurrence of each species is associated with their main foods.

Aims

To investigate the biotic factors related to the occurrence of the Great Spotted Woodpecker and the Grey-headed Woodpecker, two species that frequently inhabit overlapping regions in South Korea.

Methods

Correlations between the occurrence of the two woodpecker species, forest characteristics, and bird and insect species richness were evaluated based on large-scale public data from the National Ecosystem Survey and a Forest Type Map.

Results

The occurrence of Great Spotted Woodpeckers was associated with forest stands with at least 75% coniferous trees, whereas the occurrence of Grey-headed Woodpeckers was negatively associated with forested areas with canopy cover exceeding 50%. The occurrence of both species was strongly and positively correlated with the species richness of forest birds. The occurrence of Great Spotted Woodpeckers was correlated with the species richness of Coleoptera, while that of the Grey-headed Woodpecker was correlated with the species richness of Hymenoptera and Diptera.

Conclusion

The co-occurrence of the two woodpecker species within shared habitats can be explained by low competition due to differences in preferred forest characteristics and in the food sources used by each species.

KEYWORDS:

• Dendrocopos major
• Picus canus
• occurrence
• forest characteristics
• forest birds
• insects

Identifying factors that determine species occurrence is a prerequisite for successful species preservation (Luck 2002, Oppel et al. 2004) and a key goal in ecological research, as it improves our understanding of fitness under various environmental and ecological conditions (Vickery et al. 2001, Atkinson et al. 2005, Buckingham & Peach 2005, Sexton et al. 2009). Models predicting the probability of species occurrence are commonly used owing to their ability to incorporate a number of variables (Carvalho & Gomes 2003). Such models are advantageous with respect to time and cost efficiency for the differentiation of positive and negative variables related to the occurrence of a species (Guisan & Zimmermann 2000), including abiotic (e.g. temperature, precipitation, climate, and altitude; Porej et al. 2004, Van Buskirk 2005) and biotic factors (e.g. vegetation, competition, predation, and mutualism; Holbrook & Schmitt 1988, Lima & Dill 1990, Haila et al. 1996, Araújo & Luoto 2007, Hof et al. 2012, Wisz et al. 2013). To date, species occurrence has generally been viewed as a function of abiotic factors (Rahbek & Graves 2001, Hawkins et al. 2003); however, many studies have demonstrated that biotic factors are also powerful determinants of species distribution (Matthews et al. 2011, Giannini et al. 2013, de Araújo et al. 2014, Freeman & Mason 2015, Andradas et al. 2019).

Strategies for coexistence are needed when multiple species inhabit a single macro habitat (Leibold 1995). In particular, competition can occur between species which have similar ecological niches, such as overlapping food or breeding site requirements (Pianka 1976). A mechanism for maximizing resource availability while avoiding competition ensures that highly correlated biotic factors do not overlap with each other (Bastolla et al. 2005). This is possible because there are differences between species in their ability to respond to biotic factors (Chesson 2000, Leibold & McPeek 2006). Therefore, information on biotic factors can clearly identify whether species inhabiting the same region compete (Arlettaz et al. 2000, Kusch & Schmitz 2013), and might be important for conservation strategies (Nie et al. 2019).

Most woodpeckers live in forests and frequently occur in areas with dynamic forest elements, such as old trees, dead wood, structural diversity, natural edges, openings, etc. (Mikusiński & Angelstam 1997). The cavities on trees created by woodpeckers provide nesting and roosting sites for many secondary cavity users (birds, mammals, amphibians, reptiles, and invertebrates), thus supporting the role of woodpeckers as keystone species in the forest ecosystem (Gorman 2014). Acquiring reliable population information is facilitated by easily-detected unique woodpecker behaviours, such as drumming, pecking, and excavation (Drever et al. 2008). Among 11 native species of woodpecker (NIBR 2022) in South Korea, the Great Spotted Woodpecker Dendrocopos major and the Grey-headed Woodpecker Picus canus are the most common medium-sized residents, distributed throughout the country (Park 2014). The National Ecosystem Survey on birds for 2006–2018 reported approximately 40% habitat overlap for the two species, with cohabitation in 666 of 1668 total grids. When multiple species of woodpeckers coexist in a region, competition is often reduced by differentiation of food sources or tree characteristics (e.g. height and diameter at breast height, DBH) (Török 1990, Kruszyk 2003). However, only one study has evaluated the variables related to the occurrence of the Great Spotted Woodpecker in South Korea (Lee 2013), and that identified the two key variables as forest naturalness and the normalized difference vegetation index (NDVI). The habitat use of the Grey-headed Woodpecker in South Korea has not been evaluated to date, which limits our understanding of the co-occurrence of these two species. Moreover, the model developed by Lee (2013) did not include biotic factors other than forest characteristics.

Therefore, this study aimed to determine the biotic factors influencing the occurrence of Great Spotted and Grey-headed Woodpeckers in South Korea using large-scale public data, including forest characteristics, incidence of other forest bird species in various mutual relationships with woodpeckers, and insects, which form much of their main food source. This analysis provides insight into the coexistence of the two species and the role of reduced competition in their distribution.

Methods

Climate and forest characteristics of South Korea

South Korea is located in the mid-latitude temperate climate zone and has four distinct seasons. In summer, the weather is hot due to the high temperature and humidity of the North Pacific High, while in winter, it is cold and dry due to the influence of cold and dry continental high pressure. In spring and autumn, there are many clear and dry days due to the influence of migratory anticyclones (KMA 2023).

The forests of South Korea are characterized by steep slopes in the east and gentle slopes in the west, as the Taebaek Mountains lean towards the east. According to Lee (2015), the country has a high proportion of forests, exceeding 60% of the national territory. Coniferous forests account for 40.5% of the total forests, while broad-leaved forests account for 27.0%. The Korean Red Pine Pinus densiflora dominates the vegetation community and accounts for 27.9% of land cover, followed by the Mongolian Oak Quercus mongolica with 21.0% (Lee 2015).

Identification of occurrence sites of target species

The data were obtained from the National Ecosystem Survey (NES) conducted by the Ministry of Environment for the period 2006–2018. In accordance with the Natural Environment Conservation Act, the NES divided the country of South Korea into 5–10 year cycles, and the landform, vegetation, terrestrial flora and fauna (flora, mammals, birds, herptiles, and insects), aquatic fauna (fish and benthic invertebrates) are investigated annually (NIE 2019). They are investigated at a 1:25,000 scale using topographic maps (map size 11.3 km × 13.9 km) with each map divided into a grid of nine rectangles of 3.8 km × 4.6 km (Kwon et al. 2020). In the case of birds, since they inhabit almost all habitat types (Mardiastuti 2019), species and individuals are investigated seasonally in all habitat types in the grid using methods of total counts, line transects and point counts (NIE 2019).

Grids were selected that contained at least one species of forest bird. Grids including islands or coastal areas were excluded as possible migration routes. Owing to the lack of an official list of forest birds in South Korea, the habitats described by Lee et al. (2020) were used for reference, and forests and woods were taken as relevant areas. The period of investigation varied slightly each year, although the survey was generally conducted between April and October. Only grids investigated in all three periods: spring (April–May), summer (June–August), and autumn (September–October), were selected for analysis. Among the selected grids, those coinciding with insect surveys as part of the NES in the same period were identified because the richness of the insect community is generally correlated with that of the avian community (Beşkardeş et al. 2018) and woodpeckers feed on insects throughout the year (Arslangundogdu 2010). For reference, NES insect surveys are performed twice a season, except for the winter season, to survey species and individuals using line transects, sweep netting and malaise traps diurnally, and light traps and pitfall traps nocturnally (NIE 2019).

Among grids satisfying all of the above criteria, the Great Spotted Woodpecker and the Grey-headed Woodpecker were present in 201 and 182 eligible grids respectively, with both species present in 43 eligible grids. In these grids, the occurrence data of the two species was not split according to season because woodpeckers are sedentary residents (Mikusiński et al. 2001, Virkkala 2006, Wisz et al. 2013, Park 2014).

Variable selection

The species richness was estimated for forest birds, Lepidoptera (LEPI), Coleoptera (COLE), Hymenoptera (HYME) and Diptera (DIPT). These insect taxa were included because they form much of the woodpeckers’ diet in South Korea (Fennell 1965). For grids investigated over multiple years, the cumulative species richness was estimated. Arc Map v.10.8 software (Environmental Systems Research Institute Inc., Redlands, CA, USA) was used to superimpose the actual vegetation map (polygons) from the NES for the period 2014–2018 and the 5th Forest Type Map (FTM) at a 1:25,000 scale for the period 2006–2010. The FTM is a forest theme map that shows the distribution of forests throughout South Korea, and provides information on various attributes, such as forest stands, diameter class, age class and canopy cover, in connection with the National Forest Inventory (Park et al. 2019). The number of vegetation populations composed of trees (TREE) were determined from the actual vegetation map (NIER 2012). The forest area (AREA), forest stands with 75% or more coniferous trees (CONI), forest stands with 75% or more deciduous trees (DECI), forest stands with over 50% medium- to large-diameter trees (≥18 cm DBH) (MELA), and forest stands with canopy cover greater than 50% (DEBC) were estimated from the 5th Forest Type Map for conversion to the grid unit area (%) (KOFPY 2017).

Scope determination and analysis

To identify factors that influence the occurrence of the Great Spotted Woodpecker and the Grey-headed Woodpecker, a binary logistic model was constructed. Binary logistic models predict the probability of occurrence of a target species based on independent factors (Sarhangzadeh et al. 2013). To ensure the accuracy of the model, a two-fold greater number of randomly selected non-occurrence grids were added to obtain a total of 730 and 623 grids for the Great Spotted Woodpecker and the Grey-headed Woodpecker, respectively (Figure 1). Correlation analysis was performed to identify factors displaying high collinearity (Spearman’s r² > 0.7; Zar 1999), all of which were included in the covariance analysis owing to a lack of correlation among variables. The explanatory power of variables related to the occurrence of target species was evaluated using Nagelkerke $R^2$values, and the goodness of fit of the model was regarded as acceptable if the Hosmer–Lemeshow test indicated P > 0.05 (Hosmer & Lemeshow 2000). Statistical analysis was performed using the moonBook package in R v.4.1.2 (R Foundation, Vienna, Austria) and IBM SPSS Statistics v.20.0 software (IBM Corp., Armonk, NY, USA).

Figure 1. Distribution of the Great Spotted Woodpecker Dendrocopos major (A) and the Grey-headed Woodpecker Picus canus (B) in South Korea based on the National Ecosystem Survey conducted from 2006 to 2018. The grid squares and recording units were approximately 3.8 km × 4.6 km. Black indicates presence and grey indicates absence.

Results

The Hosmer-Lemeshow test for the binary logistic model revealed acceptable model fits for both the Great Spotted Woodpecker (${\chi }^2$ = 2.494, df = 8, P = 0.962) and the Grey-headed Woodpecker (${\chi }^2$ = 15.157, df = 8, P = 0.056). Nagelkerke $R^2$ values (an indicator of explanatory power) for the investigated variables were 0.396 for the Great Spotted Woodpecker and 0.352 for the Grey-headed Woodpecker. Among forest characteristics, AREA, MELA and TREE were not significantly related to the occurrence of either species. The occurrence of the Great Spotted Woodpecker was positively correlated with CONI (β = 0.014, n = 730, P < 0.05), whereas the occurrence of the Grey-headed Woodpecker was negatively correlated with DEBC (β = −0.079, n = 623, P < 0.05). The species richness of forest birds was positively correlated with the occurrence of both the Great Spotted Woodpecker (β = 0.152, n = 730, P < 0.001) and the Grey-headed Woodpecker (β = 0.124, n = 623, P < 0.001). The species richness of LEPI showed no significant relationship with the occurrence of either woodpecker species. However, the occurrence of the Great Spotted Woodpecker was positively correlated with the species richness of COLE (β = 0.027, n = 730, P < 0.001) and that of the Grey-headed Woodpecker was positively correlated with the species richness of HYME (β = 0.045, n = 623, P < 0.001) and DIPT (β = 0.046, n = 623, P < 0.001) (Figure 2).

Figure 2. Results of the binary logistic model based on large-scale public data for South Korea with the occurrence of the Great Spotted Woodpecker Dendrocopos major and the Grey-headed Woodpecker Picus canus as dependent variables and forest characteristics and species richness of forest birds and insects as independent factors. AREA: the forest area, CONI: forest stands with ≥ 75% coniferous trees, DECI: forest stands with ≥ 75% deciduous trees, MELA: forest stands with ≥ 50% medium- to large-diameter trees (≥18 cm diameter at breast height), DEBC: forest stands with canopy cover greater than 50%, TREE: the number of vegetation populations composed of trees, FOBI: the species richness of forest birds, LEPI: the species richness of Lepidoptera, COLE: the species richness of Coleoptera, HYME: the species richness of Hymenoptera, and DIPT: the species richness of Diptera. Black indicates a positive correlation and grey indicates a negative correlation. *P < 0.05, ***P < 0.001.

Discussion

Forest characteristics are an important determinant of the occurrence of woodpecker species (Stański et al. 2020) and preferences for forest attributes vary among species (MacArthur & MacArthur 1961). The occurrence of Great Spotted Woodpeckers in South Korea was significantly correlated with forested areas containing at least 75% coniferous trees, whereas that of the Grey-headed Woodpecker was not correlated with the abundance of coniferous or deciduous trees. The close association of the Great Spotted Woodpecker with coniferous trees has been reported in previous studies. Hogstad (1978) claimed that coniferous trees comprised 96% of the feeding sites of the Great Spotted Woodpecker, and Stański et al. (2020) reported that coniferous trees were used as feeding sites by Great Spotted Woodpeckers of both sexes, likely because these trees provide habitats for numerous insects found in the woodpecker’s diet (Hilszczański 2008). Great Spotted Woodpeckers frequently visit and forage on the surface of coniferous trees near the nest during the breeding season but display unique seasonal variation in food sources (Rolstad et al. 1995). During the summer and breeding season, they mainly feeds on insects, but can feed on seeds from pine cones during the winter by removing the cone from the branch, wedging it in an anvil and extracting the inner seeds (Osiejuk 1994, 1998, Kędra & Mazgajski 2001, Dylewski & Myczko 2019). Great Spotted Woodpeckers can feed on as many as 50 cones a day (Winkler & Christie 2002) and considering that coniferous trees produce as many as 650 cones annually, the species may be one of the major consumers of cones (Myczko & Benkman 2011). One study reported that coniferous trees may enable range expansion and colonization of new areas by the Great Spotted Woodpecker (La Mantia et al. 2002). In contrast, our findings suggest that the occurrence of the Grey-headed Woodpecker may not be related to forest type. Previous studies have reported that the Grey-headed Woodpecker inhabits mixed forests (Koskimies 1989), has a preference for deciduous over coniferous forests (Rassati 2014), and its abundance is unaffected by the proportion of coniferous trees in a forest (Pakkala et al. 2020).

Canopy cover is an important determinant of the composition and abundance of the forest bird community (Beşkardeş et al. 2018, Menon & Shahabuddin 2021). A high canopy cover offers the necessary microclimate for breeding (Şekercioḡlu et al. 2002) and lowers the probability of a nest being discovered by predators (Liebezeit & George 2002). Notably, canopy cover can also provide an indicator of the abundance of insects, which act as a food source for many forest birds (Kumar et al. 2011). However, in this study, canopy cover was not significantly correlated with the occurrence of Great Spotted Woodpeckers, which concurred with the results of previous studies (Beşkardeş et al. 2018; Beşkardeş 2020). In contrast, the occurrence of Grey-headed Woodpeckers was negatively correlated with canopy cover. Despite the previously described benefits, high canopy cover may reduce biodiversity on the forest floor (Lohr et al. 2002) and the Grey-headed Woodpecker has a characteristic preference for an open environment (Koskimies 1989, Gorman 2004, Alder & Marsden, 2010).

Forest area was not significantly correlated with the occurrence of either woodpecker species, lending support to the observation that even small forests can provide bird habitats (Winkler 2005); both species are commonly found using small farms or parks with small areas of forest (O'Connor & Shrubb 1986, Lee et al. 2020). Similarly, the occurrence of the Great Spotted Woodpecker and the Grey-headed Woodpecker in the Niraj Valley in Romania is reportedly not influenced by the forest area (Domokos & Cristea 2014).

Although the size of trees can be an important determinant of woodpecker habitat stability in general (Salvati et al. 2001, Bai 2005, Domokos & Cristea 2014), the occurrence of the two species in this study showed no significant correlation with forested areas containing more than 50% medium- to large-diameter trees. An increase in tree size is associated with an increase in the abundance of insects, and so the availability of food for insectivorous birds (Torgersen & Bull 1995, Rolstad et al. 1995, Lõhmus et al. 2010, Sukovata & Jaworski 2010, Menon & Shahabuddin 2021), and also an increase in potential nesting space (Rolstad et al. 1995, La Mantia et al. 2002). The Great Spotted Woodpecker, however, has been shown to prefer environments with relatively small trees, of less than 15 cm diameter at breast height (Lee 2013), and its population tends to decrease in forests that are mainly comprised of trees with large diameters (Hong et al. 2013). Notably, Great Spotted Woodpeckers prefer trees with smaller diameters as foraging sites (Hogstad 1971). The Grey-headed Woodpecker has a larger body size than the Great Spotted Woodpecker, and has a preference for forest containing large-diameter aspen Populus or birch Betula trees in Northern Europe. Indeed, the preference of the Grey-headed Woodpecker for larger trees has previously been correlated with nesting sites in boreal forests in Finland (Pakkala et al. 2020). Nevertheless, in Norway, the Grey-headed Woodpecker has been shown to inhabit coastal areas which lack large trees, which implies that the species can inhabit some regions despite limited nest site availability (Gjerde et al. 2005). Additionally, Grey-headed Woodpeckers might have a foraging advantage within forest stands comprised of one-year-old saplings or trees of small diameters (Spitznagel 1990).

The occurrence of the two woodpecker species was not correlated with the number of vegetation populations composed of trees in the forested area. Both species have also displayed flexibility enabling them to inhabit managed forests with a single vegetation population (Angelstam & Mikusiński 1994), which may help explain the results of the current study.

The diversity of the forest bird community is highly sensitive to variations in the forest environment and shows a strong positive correlation with woodpecker habitats (Mikusiński et al. 2001, Heikkinen et al. 2007, Camprodon et al. 2008, Drever et al. 2008, Cockle et al. 2011, Kumar et al. 2011, Segura et al. 2014, Menon & Shahabuddin 2021). Likewise, the species richness of forest birds was strongly and positively correlated with the occurrence of both the Great Spotted Woodpecker and the Grey-headed Woodpecker in this study. This is possibly because excavation by woodpeckers creates foraging and nesting sites for other forest birds (Drever et al. 2008). Indeed, Angelstam & Mikusiński (1994) stated that specialized woodpeckers, such as the White-backed Woodpecker Dendrocopos leucostos and the Eurasian Three-toed Woodpecker Picoides tridactylus, were powerful predictors of the diversity of the forest bird community. The results of this study suggest that generalist species, such as Great Spotted and Grey-headed Woodpeckers, may have similar predictive value.

Despite strong correlations with the species richness of forest birds for both woodpecker species, the Great Spotted and Grey-headed Woodpeckers differed with respect to correlations with the species richness of forest insects belonging to various orders, which reflects their main food sources. In particular, the occurrence of the Great Spotted Woodpecker was positively correlated with the species richness of coleopteran insects while that of the Grey-headed Woodpecker was positively correlated with the species richness of hymenopteran and dipteran insects. While the Great Spotted Woodpecker shows substantial variation in its selection of food sources among regions and populations (Hannsson 1992, Michalek & Miettinen 2003), it is generally known to be an important predator of tree boring insects, such as bark and cerambycid beetles (Iwata et al. 1998, Jiao et al. 2008, Hu et al. 2009). Indeed, the presence of the Great Spotted Woodpecker was shown to reduce the survival rate of tree borers by 45–98% (Crockett & Hanoley 1978). In Sicily, plantation forestry of Eucalyptus trees has grown rapidly but while this tree provides a poor environment for most forest birds, the density of Great Spotted Woodpeckers tends to be elevated in such plantations (La Mantia et al. 2002). In fact, the importation of Eucalyptus trees was accompanied by the spread of the tree borer beetle Phoracantha semipunctata, which is used as a food source by the Great Spotted Woodpecker and contributed substantially to the expansion of the woodpecker population (La Mantia et al. 2002). Interestingly, coniferous trees, which were significantly correlated with the occurrence of the Great Spotted Woodpecker in this study, are also closely associated with tree boring beetles. Among 214 species of bark beetles inhabiting regions in Canada, approximately 73% (156/214 species) inhabit only coniferous trees (Bright 1976). Additionally, the main habitat of cerambycid beetles in Japan is coniferous trees (Iwata et al. 1998) and certain tree borer populations show less abundance, suggesting the tendency of woodpeckers to feed primarily on Coleoptera species (Iwata et al. 1997). Thus, our work adds to the growing evidence that supports a strong link between the occurrence of Great Spotted Woodpeckers, coniferous trees and coleopteran insects. Conversely, the Grey-headed Woodpecker has been shown to rely almost entirely on ants for food in boreal and hemiboreal forests (Angelstam & Mikusiński, 1994). Indeed, egg clusters and adult ants comprised approximately 97% of the diet of Grey-headed Woodpeckers during the breeding season in South Korea (Koo & Won 1986). The previously described preference of the Grey-headed Woodpecker for an open environment can be explained by the abundance and high diversity of ants in these areas (Dauber et al. 2006). In the central and southern regions of Scandinavia, the Grey-headed Woodpecker switched its main food source to Camponotus ants and dipteran insects when the winter cold persisted or snow depth exceeded 15 cm (Rolstad & Rolstad 1995), even though the abundance of Formica ants was high throughout the year. These observations aligned with the rapid decrease in the availability of Formica ants during this time, indicating that the Grey-headed Woodpeckers were using food sources that were hibernating under thin tree bark but which were relatively easy to obtain. In addition, analysis of the gastric contents of 16 dead Grey-headed Woodpeckers in South Korea revealed that approximately 94% of the insects were Hymenoptera species (with ants accounting for 87%) while the remaining 6% were Diptera species (Fennel 1965).

This study had some limitations. Data for dead trees, which are critical determinants of woodpecker distribution, were lacking (Bütler et al. 2004, Vergara & Schlatter 2004, Czeszczewik & Walankiewicz 2006, Smith 2007, Roberge et al. 2008, Garcia-del-Rey et al. 2009; Drever & Martin 2010, Cockle et al. 2011, Stański et al. 2020). The large-scale public data used in this study were collected by multiple investigators, including us, leading to some heterogeneity due to observer differences (Kwon et al. 2020). Additionally, the collected data were purely qualitative and noted only the presence/absence of target species (Nam et al. 2019), thus only species richness could be determined for forest bird and insect communities. Nevertheless, this study supports the results of previous studies dealing with the relationship between the Great Spotted Woodpecker and conifer trees and Coleoptera, and the Grey-headed Woodpecker and ants. Although Prendergast et al. (1993), Berg & Tjernberg (1996), and Pärt & Söderström (1999) performed their surveys at a different scale to the 1:25,000 scale used in this study, we suggest that a metacommunity approach at various scales can provide important insights in understanding the occurrence patterns of species (Leibold et al. 2004).

Although woodpeckers are sedentary, resident birds, their occurrence and determining factors vary during the breeding and non-breeding seasons (Cahill & Matthysen 2007, Kumar et al. 2011). Our results suggest that the shared habitats of the two species can be explained by low competition resulting from differences in preferred forest characteristics and food sources.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by a grant from the National Institute of Ecology (NIE) and funded by the Ministry of Environment (MOE) of South Korea (NIE-A-2023-01).