Taxonomic and community composition of epigeal arthropods in monoculture and mixed tree species plantations in a deciduous forest of Ghana

Tropical forests provide several ecosystem services and functions and support approximately two-thirds of the world’s biodiversity but are seriously threatened by deforestation. Approaches to counteract this menace have revolved around afforestation with several or a single tree species. We thus investigated how plantation forests with either a single or several tree species influenced arthropod taxonomic and community composition using pitfall traps to sample selected groups of epigeal arthropods (Araneae, Coleoptera, Orthoptera and Hymenoptera) and with environmental variables assessed simultaneously. Our results revealed 54 taxonomic groups with significantly higher taxonomic richness, activity density, and diversity in the mixed stands than in the monoculture stands. The significant differences in community composition were mainly driven by families including Lycosidae, Formicidae, Staphylinidae, Scotylidae, Hydrophilidae, Gryllidae and Scarabaeidae and were explained by distinct habitat characteristics (canopy openness, litter depth, deadwood volume, and tree height). While the diverse tree communities and heterogeneous vegetation structure offered food and habitat resources for diverse arthropod groups, the allelopathic nature coupled with homogenous stand characteristics of the Tectona grandis stands in the monoculture suppressed the growth of understorey vegetation that could otherwise serve as food and habitat resources for arthropods, which might have led to limited activities and diversity of arthropods in the monoculture plantation stands. The findings thus highlight the need to promote mixed tree plantations in degraded tropical areas, especially when restoring biodiversity is the prime management focus.


Introduction
Globally, forests and woodlands serve as reservoirs for terrestrial and aquatic biodiversity (Thompson et al. 2009;Gibson et al. 2011) and are an important part of Earth's biogeochemical systems (Gonzalez et al. 2005), supplying essential ecosystem services for human well-being (Brockerhoff et al. 2017). Among the high proportion of the world's Abstract Tropical forests provide several ecosystem services and functions and support approximately two-thirds of the world's biodiversity but are seriously threatened by deforestation. Approaches to counteract this menace have revolved around afforestation with several or a single tree species. We thus investigated how plantation forests with either a single or several tree species influenced arthropod taxonomic and community composition using pitfall traps to sample selected groups of epigeal arthropods (Araneae, Coleoptera, Orthoptera and Hymenoptera) and 1 3 biodiversity supported by forest ecosystems (Coote et al. 2013;Giam 2017), tropical forest ecosystems alone support approximately two-thirds, serving as a hotspot and home for myriad plants and animal species (Raven 1988;Merritt et al. 2019) and maintaining several other ecosystem processes (Orians 2000). For example, in Ghana, forests remain home to over 3600 flora species, including approximately 728 species of birds (of which 15 occur in internationally important numbers and 7 threatened), butterflies (23 species classified as endemic or near-endemic) and several other species of arthropods (Ministry of Environment, Science, Technology, and Innovation 2016; Kondra 2019).
Globally forests provide an essential livelihood for over 1.745 billion people, mostly in developing countries (Langat et al. 2016), with the level of people dependence varying according to the geographical area defined by socioeconomic conditions (Widianingsih et al. 2016). In Ghana, over 2.5 million people depend on forests for sustenance and revenue (Boafo 2013). People and local communities depend on forests for the supply of nontimber forest products (NTFPs), which include mushrooms, snails, plant-derived medicine, fuelwood, and game meat (Edusah 2011), and household income generation through the sale of forest products (Mantey and Teye 2021). In addition, they supply numerous ecosystem goods and services, such as food, fodder, and fuelwood, and regulate climate as well as pest activities that are beneficial from a food production point of view (Felipe-Lucia et al. 2018;Damptey et al. 2021).
Notwithstanding the benefits of forests worldwide, including serving as critical reservoirs for many groups of biotas, forests are being deforested at an alarming rate (FAO and UNEP 2020;Ritchie and Roser 2021). The global annual net rate of deforestation was estimated at 4.7 million hectares in 2010 and 10 million hectares between 2015 and 2020 (FAO and UNEP 2020). In Ghana, over 794,214 hectares of forest were lost annually between 2013 and 2015 (Forestry Commission 2017) due to multifaceted factors such as logging, farming, infrastructural development, mining and wildfire (Acheampong et al. 2019;Damptey et al. 2020;Weisse and Goldman 2020). Hence, deforestation contributes to climate change and biodiversity loss and negatively impacts the provision of several essential ecosystem services ( Austin et al. 2017;Eguiguren et al. 2019;Prevedello et al. 2019), making it a major concern for most developing countries (Allen and Barnes 1985;Hosonuma et al. 2012), including Ghana.
Aiming to combat deforestation worldwide, several policies and initiatives have been proposed, including forest protection and restoration, payments for ecosystem service programs that compensate people for conserving forests, community engagement and social inclusion, as well as strengthening land tenure systems (Busch and Ferretti-Gallon 2014;Tuttleman et al. 2019). In Ghana, through the Forestry Commission and in line with Ghana's Medium Term Development Plan, the government has proposed and implemented several plantation initiatives targeting restoring forest cover losses (Brown et al. 2016). Among the many comprehensive reform programs in the forestry sector initiated is the Ghana Government initiative in plantation development (Baatuuwie et al. 2011). Through the New York declaration on forests, the Government of Ghana pledged to restore 2 million hectares of degraded forests by 2030 to continuously provide goods and services to communities (Foli et al. 2018).
Forest plantations provide not only goods (e.g., timber, food, water, medicinal resources) and services essential for the environment and mitigating climate change (e.g., carbon storage, flood control, clean air; Bauhus et al. 2010;Bampoh and Damnyag 2020), but also serve as habitats for many species, including arthropods, thereby enhancing the biodiversity situation in such ecosystems (Brockerhoff et al. 2017). Depending on the aims of a particular plantation program and local context, afforestation programs have focused on using either mixed or single tree species (Martin et al. 2021) which should support different levels of ecosystem services and biodiversity. For instance, monoculture plantations focus more on timber production, while mixed-species plantations are usually more beneficial for biodiversity (Liu et al. 2018).
In most regions of Ghana, especially deciduous forest zones experiencing rampant annual wildfire events, Tectona grandis L.f. (Teak) has been the most preferred and successful plantation species used in afforestation programs (Hawthorne and Abu-Juam 1995;Nunifu and Murchison 1999). In addition to being fire-resistant, the tree has a short rotation period and high-quality timber, making it the preferred timber species for local use and export (Djagbletey and Adu-Bredu 2007;Kumi et al. 2021;Restrepo et al. 2021). However, mixed species are widely used in most afforestation programs from a biodiversity perspective. They are known to be resistant to insect attacks and diseases and could also increase the species composition of other life forms (e.g., birds, insects; Li et al. 2012;Liu et al. 2018).
Arthropods are also essential biodiversity components, as they provide several ecosystem services and functions (Birkhofer et al. 2016;Noriega et al. 2018;Dangles and Casas 2019;Damptey et al. 2021). For example, they offer pollination, regulatory (e.g., pest regulation), and supporting services (e.g., soil formation) for biodiversity and local communities (Rader et al. 2016;Schowalter 2017;Birkhofer et al. 2018;Dangles and Casas 2019;Elizalde and Superina 2019). Hence, it is imperative to mention that both monoculture and mixed plantation approaches come with a cost or gains for local biodiversity, including arthropods. Here, in an ecological study in a deciduous forest in Ghana, we investigated the diversity and community composition of arthropods associated with tree plantations. Specifically, we assessed (1) whether the arthropod communities differed between the mixed and monoculture plantation stands, (2) whether arthropod taxonomic and community composition in the different plantations is driven by plant attributes (e.g., diversity, habitat structural heterogeneity) and other environmental characteristics and (3) identified the arthropods community associated with each plantation type. We hypothesised that areas with diverse plant communities and complex vegetation structures would provide sufficient food resources and create appropriate habitat requirements to support diverse groups of arthropods.

Study site
We carried out these investigations in the Bosomkese forest reserve (BFR), located in the Tano North Municipal district of the Ahafo region. The BFR lies in Ghana's semideciduous southeast forest zone ( Fig. 1; longitude 2°14.782′ W, latitude 7°6.338′ N) and is characterised by two pronounced seasons: the hot-dry harmattan (November to March) and the rainy (April to October) seasons, with rainfall ranging between 900 and 1500 mm (Addai and Baidoo 2013). BFR is a protected forest managed by the Bechem Forest District to ensure sustainable timber production and increase the timber resource base through enrichment planting and other plantation development (Ghana Forestry Commission 2013). Based on the above aims, several plantation initiatives involving mixed tree species (both exotic and indigenous) have been initiated in the area. In addition to the regular mixed tree planting style, single tree species are used in specific degraded compartments where wildfire incidence is high and annual. One tree species that can withstand annual wildfire events in the semideciduous forest zone is T. grandis (teak). The mixed plantation includes tree species such as Cedrela odorata L., Terminalia ivorensis A. Chev, Senna siamia (Lam.) H.S. Irwin & Barneby, and Ceiba pentandra (L.) Gaertn. (Table 1). The management directive of such plantation initiatives takes the form of agroforestry, where forest trees (both exotic and indigenous) are interplanted with arable crops to meet both environmental and resource (e.g., food, fuelwood, fodder) necessities of local communities.

Sampling design
We demarcated and sampled 12 plots in the teak plantation (monoculture stand) and 12 in the mixed plantation stands (24 sampling plots). The monoculture stand is located in compartment 64, and the mixed plantation is in compartment 35 of BFR. Plots were 20 m × 20 m, placed randomly, and at least 200 m from each other.

Arthropod sampling
To sample the activity density of ground-dwelling (epigeal) arthropods in each plot, five pitfall traps constructed from transparent disposable cups (11.5 cm diameter, 12 cm depth) were installed flush with the soil surface at least 5 m apart. Traps were filled with a 50:50 mixture of propylene glycol and water and sheltered by small roofs (biodegradable disposable plates with a size of 15 m × 15 cm) to avoid rain dilution of the trap liquid and litter fall (Underwood and Quinn 2010;Schmidt et al. 2006). Traps remained unused for a week to prevent any digging-in effect (Greenslade 1973), then traps were emptied weekly for 12 weeks (February to May 2020) and stored in 70% ethanol until they were sorted into taxonomic groups (order, suborder or family). Approximately 89% of the catch samples after 12 weeks belonged to the order Coleoptera, Araneae, Hymenoptera and Orthoptera, which became surrogate taxa for this research. Because of our limited taxonomic skills to identify samples to lower taxonomic resolution (genus or species), we relied on family-level identification based on Picker et al. (2002) and Dippenaar-Schoeman and Jocqué (1997) of these four groups for further analysis.

Vegetation attribute sampling
We counted, measured or estimated seven vegetation attributes (Table 1) to determine their relationship with the sampled arthropod groups. In each plot, we identified all trees to species following Hawthorne and Gyakari (2006) and counted all trees ≥ 10 cm diameter at breast height (dbh). We measured tree heights with a Nikon Forestry Pro II Laser Rangefinder, measured their diameter using a Vernier calliper and estimated their basal area from the measured dbh.
Canopy openness was estimated based on digital hemispherical photographs taken with a smartphone (iPhone 8 Plus) fitted with a fish-eye lens (MACTREM Phone Camera Lens Kit with a 205° angle-of-view) mounted on a tripod  (Tichý 2002). In addition, litter depth was measured using the horizontal bar method (Marimon-Junior and Hay 2008), and deadwood volume was estimated based on Böhl and Brändli (2007).

Data analyses
The total abundance of arthropods for each plantation type was first estimated, after which the plot-level abundances were pooled and log-transformed [log (x + 1)] to generate the activity density for each taxonomic group per plantation type. From the transformed dataset, we estimated diversity indices (based on Hill numbers [N1]; Hill 1973) by applying the "Diverse" function in Primer vs 7 (Clarke and Gorley 2015). Statistically significant differences between plantation types for arthropod activity densities, taxonomic richness and diversity indices were measured with permutational multivariate analysis of variance (PERMANOVA) based on Bray-Curtis similarity (Clarke et al. 2014;Bray and Curtis 1957) and an unrestricted permutation of the raw data (N = 9999; Anderson et al. 2008). Plantation type (2 levels) was used as a fixed factor, and plot (12 levels) nested in the plantation type was used as a random factor. The extent of the variation between plantation types for arthropod taxonomic groups and order was estimated based on the effect size estimation with Hedges' g (Cohen 1988). We employed a resampling distribution from 5000 resamples and bias-corrected for 95% bootstrap confidence intervals (Ho et al. 2019). The effect size could be small (d = 0.2), medium (d = 0.5) or large (d = 0.8; Cohen 1988).
Variations in arthropod composition between plantation types were visualised using non-metric multidimensional scaling ordination (NMDS) with Bray-Curtis similarities and goodness of fit evaluated using the two-dimensional stress value (Clarke et al. 2014). Characteristic taxa were further superimposed as vectors (correlation = 0.2) with axis scores on the NMDS cluster (Clarke et al. 2014). Individual taxonomic contributions to the dissimilarity between plantation types were evaluated using similarity percentage analysis (SIMPER) with a 70% cut-off value for the total taxon contribution (Somerfield and Clarke 2013).
The explanatory power of vegetation attributes on the differences in arthropod communities was evaluated with a distance-based linear model (DistLM) using the BEST selection procedure and R 2 criterion at a permutation of 9999 (Clarke and Gorley 2006). We performed all statistical analyses and visualised our results with PRIMER version 7 (Clarke and Gorley 2015) and an online statistical tool (estimationstats.com; Ho et al. 2019).

Results
In total, 3593 individuals belonging to 54 taxonomic groups (Table S1: 19 families of Araneae, 24 families of Coleoptera, 3 families of Hymenoptera, and 5 families of Orthoptera) were collected from the two forest plantation (monoculture and mixed) stands. In the mixed stands, there were 15 Araneae, 24 Coleoptera, 3 Hymenoptera and 5 Orthoptera families, whereas the monoculture stands had 12 Araneae, 16 Coleoptera, 3 Hymenoptera and 3 Orthoptera families.
Forest plantation types differed significantly in taxonomic richness ( Fig. 2a; F 1 Arthropod community composition differed significantly between plantation types for the taxonomic groups (F 1, 23 = 6.18; P = 0.001), Araneae (F 1, 23 = 8.24; P = 0.001), Fig. 2 Box plots of A taxonomic richness, B activity density and C Hill numbers (N1) of arthropods in mixed and monoculture forest plantations. The line represents the median value; the 25th and 75th percentiles represent the box limits. Error bars show 10th and 90th percentiles Coleoptera (F 1, 23 = 5.23; P = 0.001), Hymenoptera (F 1, 23 = 6.98; P = 0.002) and Orthoptera (F 1, 23 = 2.84; P = 0.041). For the taxonomic groups, the monoculture stands were characterised by a higher abundance of Apionidae and Idiopidae. In contrast, the mixed stands were mostly dominated by Formicidae, Gryllidae, Carabidae, Scarabaeidae and Lycosidae (Fig. 4 A). An average dissimilarity of 41% was observed in taxonomic composition between the monoculture and the mixed tree species plantations. This dissimilarity was mainly driven by higher abundances of Lycosidae, Formicidae, Staphylinidae, Scotylidae, Hydrophilidae and Gryllidae in the mixed tree species plantation compared to the monoculture stands ( Table 2). The family groups of Araneae (Barychelidae, Clubionidae, Cyrtaucheniidae, Gnaphosidae, Miturgidae, Oonopidae, and Pholcidae), Coleoptera (Aderidae, Chrysomelidae, Curculionidae, Dytiscidae, Endomychide, Mycetophagidae, Pselaphidae, and Ptiliidae) and Orthoptera (Tridactylidae and Telligomidae) were completely absent from the monoculture community (Table S1). The composition of the Araneae community differed by approximately 52% between the two plantation types, with the families Idiopidae and Araneidae showing a higher affinity to the monoculture stands and the remaining families showing a preference for the mixed tree species stands (Fig. 4B). The Coleoptera community composition differed by approximately 48% (Fig. 4C), Hymenoptera differed by 26% (Fig. 4D), and Orthoptera differed by 26% (Fig. 4E) between the mixed and monoculture plantations. The measured habitat attributes revealed the mixed plantation to be more diverse in terms of tree species than the monoculture stands. Tree richness for the mixed stands ranged between 3 and 15 trees per plot. Apart from tree abundance and basal area, the other measured attributes differed significantly between the two plantation types, with higher deadwood volume and more canopy openness in the mixed plantation stands (Table 1). The distance linear model revealed that approximately 66% of the explained variations between the mixed and monoculture stands in terms of arthropod community composition were influenced by habitat attributes. Of the seven attributes, five showed significant relationships (Table 3). Tree richness, tree height and litter depth showed the highest explanatory power for the observed relationship between habitat attributes and arthropod community composition. The dataset that supports the findings of this study is openly available at https:// figsh are. com/s/ 65638 b6d41 8794f 9f733.

Differences in arthropod taxonomic and community composition between monoculture and mixed tree plantations
Mixed tree species plantations supported higher arthropod taxonomic and community characteristics than monoculture plantations and showed higher densities of the most characteristic taxa, with the differences in arthropod community being explained by several vegetation-related factors (e.g., the diversity of trees, vegetation structure; (Basset et al. 2012;Zhang et al. 2016;Wang et al. 2019;Knuff et al. 2020;Damptey et al. 2022), microclimate created by the plantation type, niche differences, resource availability and differences in management between the plantation types (Andersen 2019;Méndez-Rojas et al. 2021). Several studies have already affirmed the relationship between tree plantation types and arthropod fauna in the tropics (Barnes et al. 2014;Perry et al. 2016;Stephens et al. 2016) and temperate regions (Stamps and Linit 1997). Wang et al. (2019) and Ghazali et al. (2016) revealed higher arthropod species or order richness in mixed tree plantations than in monoculture stands. However, Oxbrough et al. (2012) also discussed no relationship between arthropod assemblages (e.g., spiders, beetles or moths) and forest type (mixed or monoculture stands). Although most taxonomic groups showed higher affinity to the mixed plantation stands, other groups from the Araneae (Araneidae, Idiopidae, Salticidae), Coleoptera (Apionidae, Cerambycidae, Elateridae, Scydmaenidae) and Orthoptera (Tetrigidae) also revealed higher preference for the monoculture plantation stands, reflecting the relationships between specific arthropod groups and individual tree species identity (Kriegel et al. 2021). From a biodiversity point of view, mixed tree species are a better option than monocultures in providing support for arthropod assemblages (Oxbrough et al. 2012). This support is seen in the diverse microhabitat requirements, ecological niches, and food resources mixed trees compared to monoculture plantations provide (Esquivel-Gómez et al. 2017;Ampoorter et al. 2020).

Effect of habitat characteristics on arthropod taxonomic and community composition
Our study revealed heterogeneity in the physical structure and diversity of plants in the mixed plantation, with a homogenous structure and single tree species characterising the monoculture stands. These findings are seen from the differences in vertical tree strata of the mixed tree species, the larger amount of deadwood volume, different tree sizes and the degree to which canopies are opened, permitting light transmission to the understorey and litter zones necessary for arthropod activities. Tree diversity in the mixed tree plantation ranged between three and fifteen species per plot and supported the activity of most taxonomic groups compared to the monoculture plantation. Multiple trees as hosts for phytophagous arthropods are poised to increase arthropod community characteristics by modifying environmental conditions and habitat space favourable for arthropod activities (Langellotto and Denno 2004;Schuldt et al. 2019. In addition, such communities offer a diversity of resources to support a wider range of feeding needs and improve the overall diet and fitness (Campos-Navarrete et al. 2015;O'Brien et al. 2017;Staab et al. 2021), which could limit competitive exclusion and enhance the co-existence of different arthropod groups (Levine and HilleRisLambers 2010). Thus, diverse plant communities offer wide niche differences that provide more niche opportunities through increased resource diversity (Levine and HilleRisLambers 2009;Staab et al. 2021) and a range of environmental requirements and various food sources that could support diverse arthropods with a wider feeding range and environmental resource needs (Levine and HilleRisLambers 2010). In addition, the leaf architecture of these diverse species could host several nesting sites or serve as a hiding area or a place for oviposition, subsequently enhancing the arthropod population and composition (Campos et al. 2006). However, the monoculture stands with just one tree species probably offered just a host plant supplying limited resources for some specialised groups of arthropods. Monoculture is known to simplify the complex nature of the plant system to a single species community, rendering it less suitable for most arthropod groups (Stamps and Linit 1997), while a community with diverse tree species provides numerous ecological niches to support a greater number of associated species (Liu et al. 2018). Our results agree with other similar studies. For instance, Wang et al. (2019) revealed higher levels of arthropod diversity in mixed plantations characterised by a more diverse vegetation structure and species composition than in monoculture plantations with homogenous vegetation structures. Similarly, Stamps and Linit (1997) revealed an increase in arthropod diversity in mixed plantations compared with monoculture plantations because of the former's greater niche diversity and complexity. Furthermore, Skarbek et al. (2020) discussed the superiority of mixed forests over monoculture stands in promoting litter ant diversity.
Structural complexity in an ecosystem is advantageous for arthropod communities because it provides a range of niches for species co-existence due to increased microhabitat availability and offers a refuge for vulnerable species that could have otherwise become prey for predators (Kovalenko et al. 2012). The mixed species in our study probably offered complex canopy structures, resulting in an increase in available nesting and hiding sites for arthropods (Basset et al. 2001;Campos-Navarrete et al. 2015). In addition, the canopy condition of mixed-species might have facilitated light penetration, thereby enhancing and modifying undergrowth vegetation attributes (e.g., litter and seedlings) essential for most arthropod activities (Oxbrough et al. 2012).
Other vegetation attributes driving arthropod taxonomic and community compositional differences between plantation types are vegetation height (Keten et al. 2015), the depth and availability of leaf litter (Skarbek et al. 2020) and deadwood volume (Seibold 2015). Deadwood volume was significantly higher in the mixed than in the monoculture plantation stands and probably could serve as a vital microhabitat, offering refuge from either birds or other predating arthropods and food resources essential for most arthropod groups (Parisi et al. 2018;Dufour-Pelletier et al. 2020;Haeler et al. 2021). Moreover, the system of management in the monoculture plantation might have limited food resources and modified microhabitats serving as a key refuge and breeding sites for most arthropods (Huuskonen et al. 2021). Usually, the monoculture plantation in the study region takes the form of agroforestry, whereby several arable crops are interplanted with a single tree species. This monoculture farming system is very intensive, involving the removal of microhabitats and excessive application of agrochemicals that kill host plants and limit food resources essential for most arthropods (Nagy et al. 2015).

Characteristic taxonomic groups for the different plantation types
Divergent trends were observed, with some groups of arthropods being completely absent or showing lower activity density in the monoculture plantation. For instance, most Araneae, Coleoptera and Orthoptera families were sparsely represented or completely absent in the monoculture stands, reflecting the limits of such environments to offer food resources and habitat needs. Therefore, we attribute the low arthropod diversity and activity density to the use of T. grandis in the monoculture plantation. Although several factors are considered when selecting tree species for plantation programs (e.g., species growth rate, resistance to diseases and pests, human perceptions and beliefs, industrial demand for the species coupled with other economic reasons), the resistance of T. grandis to wildfire is the primary reason for its preference in afforestation in deciduous forests in Ghana where wildfire is rampant yearly (Hall and Swaine 1981;Osei et al. 2018). Therefore, T. grandis is chosen for most plantation programs in Ghana, especially the semi-deciduous forest zone, merely because it is able to withstand annual wildfire events (Damptey et al. 2016;Stephens et al. 2016).
In addition, T. grandis is also allelopathic (Biswas and Das 2016), and very few to no plant species grow in its understorey, greatly decreasing resource diversity and abundance and thus niche opportunities (Staab et al. 2021) for different groups of arthropods to persist. Thus, only a specific specialised group of arthropods are likely to survive in the resource-limited environment of such a monoculture stand. Our results affirm the outcomes of other studies. For instance, Stephens et al. (2016) observed a lower diversity of ants (hymenoptera) in teak plantations in Ghana, similar to the lower diversity of ants we found in the teak monoculture stands (Appendix S1).

Conclusion
Our results affirm the positive relationship between arthropod community composition and diverse plant communities, coupled with the complexity of vegetation structural attributes. The diversity of plant communities and the complex habitat in the mixed stands offered more niche opportunities and resource diversity essential for species co-existence. The monoculture teak stands supported only a few specialist arthropod groups, highlighting the limits of such stands in offering resources for most arthropod groups. This limitation in resources was fuelled by the allopathic nature of the teak trees, which greatly restricted epigeal arthropod activity. Although T. grandis does well in fire-prone areas, we recommend that a mixture of native tree species be interplanted to enhance the faunal biodiversity of such degraded forest areas. The findings of this study show that planting tree mixtures, as opposed to monoculture plantations, provides a better environment for enhancing arthropod community composition and diversity.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.