Introduction
Structure of the original European temperate forest biome was historically characterized by heterogeneity rather than uniform dense forest - with up to 50% open habitats (Pearce et al. 2023). This explains why richer insect communities can be linked to open habitats. Moreover, even higher species richness of plants have been found in heterogeneous forests that include small or large clearings (Dormann et al. 2020).The strict relationship between insects and plants has been largely investigated (Tscharntke and Greiler 1995; Samways et al. 2020). Plant changes have already been shown to affect insect abundance and communities in agricultural landscapes (e.g. Outhwaite et al. 2022), grasslands (Bonari et al. 2017), forests (e.g. Schuldt et al. 2019; Uhl et al. 2021) and urban areas (e.g. Mata et al. 2017). Most Lepidoptera, Orthoptera and Hemiptera species and some Coleoptera (e.g. xylophagous beetles) directly depend on plants due to their feeding behavior. Indeed, plants represent a trophic source and a shelter, influencing insect diversity and behavior (Rasmann et al. 2014). On the other hand, 78% of wild plants in temperate areas are pollinated by insects (Ollerton et al. 2011) and an estimated 75% of global crop species depend on pollination (Klein et al. 2007).
Among all insects, Lepidoptera are particularly suitable to be indicator species for ecological processes (Erhardt and Thomas 1991; Moise 2011). Thanks to the large taxonomical and ecological knowledge on this taxon, moths can be used as an indicator of environmental changes (Fleishman and Murphy 2009) and biotope quality (Maes and Van Dyck 2005). In addition, they react quite fast to environmental change, due to their short life cycle. Moreover, Lepidoptera can be indicators and umbrella species even for other insect taxa (Thomas 2005). Within the order, about 90% are moths, consisting of 129 families (Lees and Zilli 2019). Moths have already been used as indicator taxa for forest (e.g. Summerville et al. 2004) and grassland changes (Pöyry et al. 2005; Rákosy and Schmitt 2011). Moreover, they are also considered pivotal for ecosystem functioning as herbivores (Damestoy et al. 2020), nocturnal pollinators (Macgregor et al. 2015; Anderson et al. 2023) and food source for higher trophic levels, such as bats (Corcoran and Conner 2016).
Advertisement
To support higher moth abundance and species richness, low intensity management of forests and semi-natural grasslands, as well as habitat connectivity can be crucial (Pöyry et al. 2009; Fuentes-Montemayor et al. 2011; Kühne et al. 2022). However, in the last decades, semi-natural grasslands are experiencing encroachment by woody species (e.g. Wölfling et al. 2019), especially in mountain areas due to the increase of ecological succession following the abandonment of agricultural activities (Orlandi et al. 2016). This phenomenon is also affecting moth communities; indeed, typical specialist species are decreasing and in some cases are already locally extinct (Habel et al. 2019; Roth et al. 2021). On the other hand, natural forest disturbances, such as bark beetle outbreaks, create open areas within forests, increasing richness of moth communities by favoring some species linked to open plants and shrubs (Uhl et al. 2023). Ride widening, which opens up dense forest structures, is also valuable woodland conservation tools for macro-moths (Merckx et al. 2012). Some moth species, even forest-specialist species (Dulaurent et al. 2011), benefit from those open areas, for feeding and as possible corridors (Kozel et al. 2021).
Considering the importance of open habitats within forested areas, we investigated plant and macro-moth communities in recently shaped clearings and surroundings in a forested area of the Western Italian Alps. In this area, some management activities (e.g. ride widening and thinning) were carried out in 2020 in order to improve plant biodiversity and habitat quality for several open-habitat species, including saproxylic beetles, bats and butterflies, through the realization of an ecological corridor made of stepping-stones clearings (Piccini et al. 2022). These conservation actions provide the opportunity to study the community of moths in both the previously existing habitat (forest) and the newly formed ones (clearings). The aim of the study is the understanding of the relationship between plant (including trees, shrubs and grasses) and moth communities. We hypothesized that newly shaped clearings in forested areas overall increase plant species richness, moth abundance and species richness by introducing new habitats within the forests (e.g. xerophile or mesophile clearings). Then, we hypothesized that changes in plant communities towards open habitats might reflect changes in moth communities with the presence of indicator species for the newly identified habitats. An additional hypothesis was that an association between distinct plant and moth communities across different habitats (clearings and forests) could be detected through specialist species being more associated with one habitat over the other.
Materials and methods
Study area
This research was carried out in Clarea valley (45°08′07.7” N 6°59′37.8” E), between 765 and 983 m a.s.l. in Giaglione (TO, NW Italy). The study area is characterized by abandoned chestnut orchards (habitat 9260; listed in Annex I of Habitats Directive 92/43/EEC), where the chestnut (Castanea sativa Miller, 1768) is still the dominating species. Other broad-leaved tree species are abundant, such as oaks (Quercus spp.), wild cherry (Prunus avium (L.) L.) and the alien locust tree (Robinia pseudoacacia L.). In April 2020, around 10 ha of forest were managed to increase biodiversity. Therefore, forest management (as thinning) over approximately 20 hectares were made within the forest in 2020, accompanied by the realization of ten clearings. Overall, the ecological corridor extended to approximately 500 m in length. Within each clearing, all trees were cut and all woody material was left in the place as deadwood for saproxylic beetles. The study was carried out in 16 patches: eight small clearings (min 108 m2; max 236 m2) belonging to the ecological corridor and eight corresponding forest patches without intervention (Fig. 1). Average horizontal distance between clearings and forested patches was 78 m, while the slope distance was around 150 m (measured in the field using hand-held GPS). Elevation ranged between 650 and 950 m a.s.l. Topography and other features of clearings and forest patches are described in Table 1.
Fig. 1
Map of the study area. Plant surveys and light-trap moth sampling were conducted within clearing and within forest, respectively in yellow and in red in the map. Green patches represent the clearings belonging to the ecological corridor where plants and moths were not collected for this study. The top-left box shows the location of the study area inside of Europe. Figure made using ESRI satellite in QGIS
×
![]()
Advertisement
Table 1
Features of clearings and forest patches belonging to the study area
ID | Habitat | Perimeter (m) | Area (m2) | Elevation (m a.s.l.) | Tree Cover (%) | Shrub cover (%) | Herbaceous cover (%) | Rock and Bare ground cover (%) |
|---|---|---|---|---|---|---|---|---|
F02 | Forest | 157 | 1963 | 765 | 75 | 0 | 40 | 60 |
F03 | Forest | 157 | 1963 | 782 | 85 | 1 | 40 | 59 |
F04 | Forest | 157 | 1963 | 794 | 90 | 0 | 35 | 65 |
F05 | Forest | 157 | 1963 | 817 | 95 | 0 | 30 | 70 |
F07 | Forest | 157 | 1963 | 850 | 95 | 0 | 15 | 85 |
F08 | Forest | 157 | 1963 | 874 | 70 | 15 | 20 | 65 |
F09 | Forest | 157 | 1963 | 909 | 80 | 0 | 20 | 80 |
F10 | Forest | 157 | 1963 | 951 | 90 | 0 | 15 | 85 |
C02 | Clearing | 70 | 251 | 766 | 0 | 1 | 75 | 24 |
C03 | Clearing | 88 | 269 | 769 | 0 | 0 | 80 | 20 |
C04 | Clearing | 85 | 324 | 770 | 0 | 1 | 71 | 28 |
C05 | Clearing | 57 | 108 | 796 | 0 | 0 | 45 | 55 |
C07 | Clearing | 72 | 241 | 852 | 0 | 0 | 58 | 42 |
C08 | Clearing | 78 | 215 | 880 | 0 | 1 | 77 | 22 |
C09 | Clearing | 81 | 236 | 915 | 0 | 5 | 65 | 30 |
C10 | Clearing | 78 | 218 | 944 | 0 | 5 | 83 | 12 |
Data collection
Plant communities
Plant species composition of the eight clearings and eight forest sites was assessed with the phytosociological method (Braun-Blanquet 1932). According to this methodology, every plant species found within the survey area was identified at the species level, and its abundance was recorded (expressed in percentage cover). In the case of the clearings, the plant plot encompassed the entire surface area (recorded using a hand-held GPS device) where the forest had been cut, while for forested sites, the plant plot was defined as a circular area with a 25-meter radius (accordingly to the estimated attractiveness of the light trap), with its center coinciding with the position of the moth light trap. To calculate in the field the area we used the measuring tape. The total herbaceous, shrub and tree covers were visually estimated in percentage. Plant species nomenclature followed Aeschimann et al. (2004). In order to describe the relationships between plants and moths, the first step was to group plant species into homogeneous groups based on their auto-ecological characteristics, so as to characterize the ecological conditions of the habitats starting from the botanical composition. Identification of the main ecological groups of plant species (i.e. social behavior types or SBT, sensu Troiani et al. 2016; and Tardella et al. 2018) was carried out through a two-step process. First of all, each plant species was classified based on its phytosociological optimum at the class level according to Aeschimann et al. (2004). Subsequently, species from various phytosociological classes, characterized by similarities in physiognomy, ecology, and floristics, were grouped together into SBT in accordance with Theurillat et al. (1995). In total, the following eight SBT were detected: woodland (species with phytosociological optimum belonging to Carpino-Fagetea sylvaticae, Erico-Pinetea, Pyrolo-Pinetea, Quercetea pubescentis, Quercetea robori-sessiliflorae, Robinietea, Vaccinio-Piceetea excelsae classes), dry grassland (Festuco-Brometea and Lygeo-Stipetea classes), ruderal (Stellarietea mediae, Agropyretea intermedii-repentis, and Artemisietea vulgaris classes), shrubland (Crataego-Prunetea class), fringe (Epilobietea angustifolii, Mulgedio-Aconitetea, and Trifolio-Geranietea classes), scree (Asplenietea trichomanis and Thlaspietea rotundifolii classes), meso-eutrophic meadows (Molinio-Arrhenatheretea class), and therophitic species (Koelerio-Corynephoretea and Thero-Brachypodietea classes). For data analysis, only the SBTs with an average abundance greater than 5% were retained, i.e. woodland (62.5 ± 10.30, mean percentage cover ± standard error), dry grassland (16.0 ± 3.16), ruderal (6.6 ± 1.59), shrubland (5.8 ± 1.20), and fringe species SBTs (5.2 ± 1.09).
Moth communities
Moths were collected using light traps with UV-LED light (emission peak 398 nm light angle per LED 120°; EPIS- TAR Corporation, Hsinchu, Taiwan). Light traps were designed following Parenzan and De Marzo (1981) and modified to accommodate the UV LEDs, see Infusino et al. (2017) for further information. Each month between May and October 2020 (6 temporal replicates), data were collected in patches (8 clearings and 8 forests) activating four traps on four consecutive nights. To investigate moth communities in both paired open and forested patches, two forest and two-paired clearing patches were sampled per night (Piccini et al. 2023). Samplings were carried out between 9 p.m. and 6 a.m. on favorable nights for moth activity (i.e. low wind intensity, no full-moon interference; following Greco et al. 2018). We avoided periods with unstable weather conditions to ensure constant meteorological conditions over consecutive nights. Sampled Lepidoptera individuals died within traps, using ethyl acetate as the killing agent (Infusino et al. 2017). We identified macro-moths (hereafter “moths”), belonging to the families Saturniidae, Sphingidae, Lasiocampidae, Drepanidae, Geometridae, Notodontidae, Noctuidae, Erebidae, and Nolidae and also moths belonging to Hepialidae, Cossidae and Limacodidae. Species identification was performed in the laboratory using the reference collection at CREA Research Centre for Forestry and Wood, Rende (Cosenza, Italy), for some cryptic taxonomic groups the dissection of genitalia was necessary to identify the species. Nomenclature follows Karsholt and Nieukerken (2013).
Openness canopy cover
Close to the light trap, the openness of each plot was measured by taking hemispherical photographs oriented to the zenith at 0.3 m height above the ground with the support of a camera tripod (CANON EOS 350D camera equipped with a LENSBABY Circular Fisheye— 5.8 mm f/3.5). Hemispherical photographs were processed with Gap Light Analyzer (Frazer et al. 1999) to obtain the values of openness.
Topographic variables
Within each clearing and forest plot, the topographical variables were measured (elevation, aspect and slope), using a hand-held GPS device, a clinometer, and a compass. Slope was measured in degrees.
Data analysis
All statistical analyses were carried out using the statistical software R v.3.2.1 (R Core Team 2017).
To test if the realization of the clearings within the forest area increased plant, moth abundance and species richness, we conducted a Generalized Linear Mixed Model (GLMM) analysis. To assess differences in terms of plant species richness (i.e. total species richness, species richness of each retained SBT) and abundance (i.e. percentage cover of each retained SBT) and species richness and abundance (i.e. total number of individuals) of seasonal patch-pooled insect communities between clearings and forest, a set of GLMMs was carried out using the glmmTMB function of glmmTMB package (Brooks et al. 2017). Count variables (i.e. number of species and number of individuals) were modelled by specifying the habitat type as a fixed factor. To account for the pairing of sampling units, the ID of the clearing-forest pairing was set as a random factor, whereas to adjust the count data in relation to the size of the surveyed area an offset term of area extent was specified. Initially, the counts were modeled using a Poisson distribution, and in cases of overdispersion, tested with the check overdispersion function from the DHarma package (Hartig 2019), a negative binomial distribution was chosen. The abundance of plant species (i.e. percentage cover) were modeled using a Beta distribution being percentage variables in a 0–100 interval. Because this distribution doesn’t allow for values of 0 and 100, and given that 0 occurred frequently, cover values were adjusted using the transformation method outlined by Smithson and Verkuilen (2006). Only the Woodland SBT was modeled with a Gaussian distribution as its cover exceeded the 100% due to the sum of species covers belonging to different plant layers. Similarly to the count variables, the habitat type was set as a fixed factor, and the ID of the clearing-forest pairing was set as a random factor to account for the pairing of sampling units.
Relationship between environment and moths
To ascertain the multivariate correlation between moths and plant species a mantel test (function mantel.rtest with ‘vegan’ package; Dixon 2003) with 999 freely exchangeable permutations was carried out on log-transformed (log1p function) dissimilarity matrices of moth and plant compositions. Bray–Curtis distance was used to compute the moth and plant dissimilarity matrices. Mantel statistics were based on Pearson’s product-moment correlation.
Relationships between habitats and moths
Having established the relationship between plants and moths, we first wanted to describe the habitats based on the plants by grouping the 16 plots through a cluster analysis.
To classify plots into plant communities with similar botanical composition a hierarchical cluster analysis was performed. The percentage cover visually estimated in the field was used as data input, whereas the Pearson correlation was used to compute the similarity matrix (Dist function of ‘amap’ package), with UPGMA (Unweighted Pair Group Method using Arithmetic Averages) as grouping method (hclust function of ‘base’ package).
Then, we proceeded to test more specifically whether the habitats identified based on plants also determined equally differentiated groups in terms of moth communities. For this purpose, non-metric multidimensional scaling (NMDS) and Permutational Multivariate Analysis of Variance (PERMANOVA) were performed. The first analysis had a purely descriptive and graphical function, to visually highlight that the habitats identified by the cluster also correspond to moths. The second analysis was to statistically test whether the plant and moth communities were significantly different between habitats.
In the NMDS, Bray–Curtis distance was used to compute similarity matrices of log-transformed (log1p function) abundance of plant and moth species, respectively. To visually validate the habitat types defined through plant cluster analysis, the surveys were grouped using convex hulls on the NMDS plot of plant (axis 1 and axis 2). Additionally, in order to ascertain any potential correspondence between the habitats defined based on plant and moth communities, the plots were also grouped with convex hulls representative of the habitats identified from plant in the NMDS plot (axis 1 and axis 2) with insect data. The NMDS was carried out with the metaMDS function of the ‘vegan’ package (Oksanen et al. 2020).
To statistically validate the differences in plant and moth community composition observed among habitats identified using the cluster analysis, we performed a PERMANOVA using a Bray-Curtis distance matrix with the adonis function of the ‘vegan’ package and by setting 999 free-exchangeable permutations. We tested log-transformed (log1p function) plant and moth (calculated as seasonal patch-pooled moth communities) community species compositions for both plants and moths between ecological groups derived by plant clusters (explanatory categorical variable).
For descriptive purposes, we also wanted to identify the indicator moth species of the habitats through an Indicator Species Analysis. An indicator species analysis (ISA) was used to identify moth species associated with plant clusters. A species that occurs only in a single plant cluster is highly ranked as a potential indicator. The analysis was performed through the indicator value (IndVal) function (Dufrêne and Legendre 1997) by using the multipatt function of the ‘indicspecies’ package of R (De Cáceres et al. 2016). The IndVal is higher for species occurring in only one group (specificity) and with a high-abundance rate in all the samples belonging to that group (fidelity). The statistical significance of the association of a species with a group was obtained by 999 freely exchangeable permutations.
Once the correspondence between plants and moths was established, we also wanted to understand which factors are responsible for the differences between moth communities. Therefore, the purpose was not descriptive but to test cause-effect relationships. Hence, an RDA and a variance partitioning were conducted, which actually belong to the same analysis (since variance partitioning is a development of the RDA). With the RDA, we aimed to test which factors were significantly responsible for influencing the moth community. Therefore, factors related to the botanical composition of the clearings, their structure, and topographical characteristics were considered. The variance partitioning was mainly used to understand which of these factors were the most important.
To assess the relations between moth communities and plant composition, habitat structure, and topography a redundancy analysis (RDA) was performed using the rda function of the ‘vegan’ package. Two main matrices were arranged: (1) a matrix of Hellinger-transformed moth community and (2) a matrix with three explanatory variable groups: (i) plant SBTs, (ii) habitat structure (i.e. openness) and (iii) topographic variables (i.e. elevation and slope). The significance of the analysis was assessed with the Monte Carlo test (999 freely exchangeable permutations). The RDA was carried out with the rda function of the ‘vegan’ package (Oksanen et al. 2020).
A variationpartitioning (Legendre et al. 2012) was carried out to assess how much variability was explained by plant composition (i.e. Woodland, Dry grassland, Ruderal, Scrub, and Fringe SBT), habitat structure (openness), and topography (i.e. slope and elevation), on moth communities. A Monte Carlo permutation test (n = 999 freely exchangeable permutations) was used to test the significant differences between our trends and the random ones. To avoid issues with multicollinearity, it was ensured that the variables had a VIF (Variance Inflation Factor) ≤ 10 (Zuur et al. 2010; Montgomery and Peck 1992). The marginal and conditional effects of the different components were computed and tested for significance. The significance level was determined at 0.05. The analysis was carried out with the varpart function of the ‘vegan’ package (Oksanen et al. 2020).
Results
Plants
In total, 242 plant species were identified, 195 in the clearings of which 98 exclusively found in the clearings (41%) and 132 in the forest of which 35 exclusively to forest (15%). The overall species richness, as well as the number of species typical of dry grasslands and the ruderal species, were higher in the clearings (Table 2). The number of plant species typical of fringes, woodlands, and shrublands, on the other hand, was statistically the same between clearings and the forest (Table 2). Regarding the abundance of plant species, the species typical of fringes and the ruderal species were more abundant in the clearings, while species typical of woodlands were more abundant in the forest (Table 2). Furthermore, among the ruderal species observed after the creation of the clearings, new species, including some very rare ones in the NW-Italian Alps, have appeared that would not thrive within the original forest., i.e. Crucianella angustifolia L. (Bartolucci et al. 2021). Species typical of dry grasslands and shrublands, on the other hand, had comparable percentage cover between the two habitats (Table 2).
Table 2
Features of clearings and forest patches belonging to the study area. Composition of plants and species richness and abundance of moths is represented by mean density over the patch size (i.e. count divided by plot area, which is the way by which the count data are modeled in a poisson/negative binomial GLMM with the plot area as offset term), while abundance of plants is represented by mean cover
GLMM distribution family | Clearings (mean ± se) | Forest (mean ± se) | z | P-value | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
Species richness | ||||||||||
Plants | ||||||||||
Total species richness | poisson | 0.325 | ± | 0.0180 | 0.022 | ± | 0.0007 | -39.4 | p < 0.001 | *** |
Dry grassland species richness | poisson | 0.065 | ± | 0.0130 | 0.001 | ± | 0.0007 | -16.5 | p < 0.001 | *** |
Fringe species richness | poisson | 0.038 | ± | 0.0059 | 0.003 | ± | 0.0005 | -13.0 | p < 0.001 | *** |
Ruderal species richness | poisson | 0.057 | ± | 0.0089 | 0.004 | ± | 0.0006 | -16.9 | p < 0.001 | *** |
Shrubland species richness | nbinom2 | 0.019 | ± | 0.0021 | 0.002 | ± | 0.0002 | -9.4 | p < 0.001 | *** |
Woodland species richness | poisson | 0.074 | ± | 0.0104 | 0.008 | ± | 0.0004 | -17.3 | p < 0.001 | *** |
Moths | ||||||||||
Total species richness | poisson | 0.677 | ± | 0.1280 | 0.1 | ± | 0.003 | -39.4 | p < 0.001 | *** |
Abundance | ||||||||||
Plants | ||||||||||
Dry grassland % cover | beta | 22.00 | ± | 5.270 | 10.00 | ± | 2.190 | -2.0 | 0.044 | * |
Fringe % cover | beta | 7.70 | ± | 1.630 | 2.70 | ± | 0.810 | -4.4 | p < 0.001 | *** |
Ruderal % cover | beta | 9.00 | ± | 2.690 | 4.30 | ± | 1.430 | -3.0 | 0.003 | ** |
Shrubland % cover | beta | 6.90 | ± | 1.500 | 4.70 | ± | 1.890 | -1.1 | 0.261 | ns |
Woodland % cover | gaussian | 25.50 | ± | 6.470 | 99.50 | ± | 4.700 | 11.8 | p < 0.001 | *** |
Moths | ||||||||||
N° individuals | poisson | 3.520 | ± | 0.851 | 0.293 | ± | 0.020 | -122.2 | p < 0.001 | *** |
Hierarchical cluster analysis of plant surveys revealed three clusters belonging to the main plant communities of forest, xero-thermophilic clearings and mesophilic clearings (Fig. 2). The plots classified under the ‘forest’ group represent wooded habitats predominantly dominated by Chestnut (C. sativa), as until a few decades ago, they constituted forests primarily cultivated for chestnut production. Other reported tree species included Downy Oak (Quercus pubescens Willd.) and Wild Cherry (P. avium). The herbaceous layer of the undergrowth is dominated by species typical of wooded environments, such as Brachypodium sylvaticum (Huds.) P.Beauv. and Festuca heterophylla Lam. Generally, these communities are predominantly mesophilic, occurring on moderately developed and well-drained soils. Xero-thermic and mesophilic clearings represent the herbaceous undergrowth of recently cut forests, belonging respectively to the forest types of Scots Pine (Pinus sylvestris L.) stands mixed with meso-xerophyle hardwoods, and secondary hardwood forests under mesophilic conditions, dominated by Common ash (Fraxinus excelsior L.) and Sycamore maple (Acer pseudoplatanus L.). The most abundant herbaceous species in xero-thermophilic clearings were Achnatherum calamagrostis (L.) P.Beauv., Bromus erectus Huds, Teucrium chamaedrys L., and Carex humilis Leyss., while in mesophilic clearings, they include Poa nemoralis L., B. sylvaticum, and Brachypodium rupestre (Host) Roem. & Schult. Plots belonging to different clusters in PCA figure S1b are separated.
Fig. 2
Dendrogram resulting from cluster analysis of 16 patches based on phytosociological data (correlation similarity matrix, UPGMA agglomeration method), with the identification of plant communities and their dominant species. Species abundance is expressed in percentage cover
×
![]()
Moths
In total, 10,478 specimens (5,874 in the clearings, 2369 in mesophilic while 3505 in xero-thermophilic clearings, and 4,604 in the forest) were collected and determined at the species level. Overall, 396 species were identified, 330 in the clearings of which 94 (24%) were exclusively found in clearings (241 and 277 in mesophilic and xero-thermophilic clearings, respectively) and 301 in the forest of which 65 (16%) were exclusively found in forest. The overall species richness was significantly higher in the clearings than in forests (Table 2). Regarding the moth abundance, clearings host significantly higher individuals than forests (Table 2).
Species indicator analysis
Species with an IndVal value greater than 0.70 were selected as the predictor set, being almost exclusively present in one habitat cluster (Table 3).
Table 3
Species identified by the IndVal analysis with an indicator species per habitat with values greater than 0.7. Species marked with * are those present exclusively in that habitat
Habitat | Families | Species | IndVal | p-value |
|---|---|---|---|---|
Forest | Noctuidae | Dypterygia scabriuscula | 0.879 | 0.005 ** |
Erebidae | Paracolax tristalis | 0.774 | 0.001 *** | |
Noctuidae | Cosmia trapezina | 0.754 | 0.012 * | |
Noctuidae | Caradrina aspersa | 0.703 | 0.037 * | |
Mesophile clearings | Noctuidae | Hadena luteocincta | 0.816 | 0.025 * |
Noctuidae | Amphipyra tragopoginis | 0.798 | 0.039 * | |
Noctuidae | Dichagyris signifera | 0.798 | 0.021 * | |
Geometridae | Idaea rusticata | 0.769 | 0.037 * | |
Xero-thermophile clearings | Geometridae | Gymnoscelis rufifasciata | 0.880 | 0.010 ** |
Noctuidae | Euxoa tritici* | 0.866 | 0.017 * | |
Geometridae | Selidosema taeniolaria | 0.836 | 0.027 * | |
Geometridae | Idaea degeneraria | 0.821 | 0.005 ** | |
Noctuidae | Hoplodrina ambigua | 0.798 | 0.030 * | |
Noctuidae | Mythimna sicula | 0.797 | 0.012 * | |
Noctuidae | Antitype chi | 0.791 | 0.021 * | |
Noctuidae | Mythimna albipuncta | 0.775 | 0.043 * | |
Noctuidae | Agrotis trux | 0.760 | 0.008 ** | |
Noctuidae | Mesogona acetosellae | 0.760 | 0.050 * | |
Geometridae | Catarhoe cuculata | 0.756 | 0.024 * | |
Geometridae | Eupithecia lariciata | 0.740 | 0.044 * | |
Geometridae | Peribatodes rhomboidaria | 0.728 | 0.019 * |
Relationship between plants and moths
Mantel test results showed that the matrix of plant structure was significantly positive correlated with the moth matrix (Mantel statistic r: 0.513; p = 0.001).
The three groups of plant communities identified based on cluster analysis were well separated, as statistically confirmed by PERMANOVA and visually in the NMDS for both plant data (F = 7.31, R2 = 0.53, p = 0.001***; Fig. 3A) and moth data ((F = 1.85, R2 = 0.22, p = 0.003**; Fig. 3B), suggesting that moth communities respond to plant habitats.
Fig. 3
Nonmetric multidimensional scaling (NMDS) patches (two dimensions) of samples of (A) plant (stress 0.04; R2 = 0.99) and (B) moth species (stress 0.07; R2 = 0.97)
×
![]()
The RDA was statistically significant (F = 1.489, p = 0 0.001***, R2adj 0.21) and the 32.39% and 17.15% of the total variance were explained by the first (F = 3.857, p = 0 0.001***) and second axis (F = 2.042, p = 0 0.089.), respectively (Fig. 4A). The patches separated along the first axis were mainly linked to habitat features (clearings vs. forest), while the separation along the second axis is mainly linked to the clearing communities (mesophile vs. xero-thermophile clearings; Fig. 3A). Particularly, the openness (F = 2.996, p = 0.001***) and elevation (F = 2.448, p = 0.004**) are significant variables in differentiating patches.
The variation partitioning showed that the pure effect of the three components (marginal effect) was statistically significant. Therefore, each component, independently from the others, could explain the moth communities associated with habitats. Specifically, the component that explained the most variance, based on the Adjusted R2, was plants, followed by topographic variables and, finally, structural variables (Table 4). In contrast, the conditional variance was not significant for any of the three components, suggesting that, when considering the effect of two components, the third one under consideration does not add a significant amount of variance in explaining the insect community (Table 4 Fig. 4B).
Fig. 4
(A) RDA ordination bi-plot showing the relations between patches (large dots in different colors: red for forest patches, green for mesophile clearings and blue for xero-thermophile clearings) and moths (species are identified by small blue dots). Habitat structure (i.e. openness), plants (woodland, fringe, dry grassland and ruderal species) and topography (elevation and slope) variables are projected as passive variables (arrows). The length of the arrows is proportional to their importance and the direction of the arrows shows their correlation with the axes. (B) Partitioning difference of moth Hellinger distance matrix, as response variable, among three sets of explanatory variables. The square area corresponds to 100% of the variation. Values < 0 are not shown
×
![]()
Table 4
Partition table derived by the variation partitioning analysis. Marginal and conditional effects of the three groups of variables (e.g. Habitat structure, plants and topography) that explain moth communities
Variables | Df | R2 | Adj. R2 | Effect | p-value |
|---|---|---|---|---|---|
Total habitat structure | 1 | 0.159 | 0.098 | marginal | ** |
Total plants | 5 | 0.436 | 0.154 | marginal | ** |
Total topography | 2 | 0.256 | 0.141 | marginal | *** |
Habitat structure + Plants | 6 | 0.490 | 0.149 | ||
Habitat structure + Topography | 3 | 0.357 | 0.197 | ||
Plants + Topography | 7 | 0.582 | 0.215 | ||
Individual fractions - Habitat structure | 1 | -0.010 | conditional | ns | |
Individual fractions - Plants | 5 | 0.009 | conditional | ns | |
Individual fractions - Topography | 2 | 0.058 | conditional | ns | |
All | 8 | 0.629 | 0.207 |
Discussion
The plant analysis initially allowed for the characterization of these habitats: the wooded area exhibited homogeneity in terms of botanical composition, whereas among the clearings, two main plant types were identified —xero-thermophilic and mesophilic. Additionally, the creation of the clearings has triggered a process of colonization by herbaceous species that may have influenced the differentiation of the two plant communities within the clearings. The successful establishment of plants in a new area relies on the combination of a species’ dispersal capability and the specific ecological conditions of the local environment (indeed closer areas with similar ecological conditions have a similar plant composition), for which altitude and sun exposition make some differences (Ozinga et al. 2004). The ecological conditions serve as filters, favoring the establishment of plant species possessing traits well-adapted to the unique conditions of the site (Guisan and Rahbeck 2011). These differences in plant communities were reflected accordingly in moth communities. Several species were indicators for each identified habitat. Plants explained most of the variation on moth communities, followed by topography (i.e. elevation) and structural variables (i.e. openness) (in Fig. 4; Table 4. See plant marginal effect).
Clearing formation and its short-term effects on plants and moths
Formation of clearings in forested areas has proved to increase species richness of different biodiversity component even in short-term: moths (24%; this study), plants (41%; this study), butterflies, saproxylic beetles, reptiles (Sebek et al. 2015). Moreover, clearing formation supports also protected species, butterflies such as Euphydryas maturna, Parnassius mnemosyne and Zerynthia polyxena (Dolek et al. 2018; Habel et al. 2022). Moreover, ecological successional stages show different moth communities (Weiss et al. 2021), with species of open habitats dominating the community in the first four years after the coppicing (Broome et al. 2011). However, we also found some species associated with mature coppiced forests, such as Apoda limacodes that are strictly linked to woody plants for larvae feeding (Broom et al. 2011). Moth communities usually take more than one year to respond to forest management measures such as ride widening and coppicing (Sebek et al. 2015). In our study newly-shaped clearings are part of ecological corridors, connecting open habitats. This aspect might have influenced the high number of species and individual abundance, increasing the mobility of species between open areas along the corridor. Clearings in connection with open habitats support higher biodiversity than isolated ones (Sebek et al. 2015).
The significant marginal effects for plants, habitat structure, and topography individually in the variance partitioning suggest that each of these factors plays a role in shaping the moth community. However, the lack of significance in the conditional effects implies that, if it is known the influence of one variable (e.g. plants), the information about the other variables (structure and topography) does not significantly improve the ability to predict or explain the variation in the insect community. All three components had a significant marginal effect, indicating that each of these components can explain the patterns of moth communities. However, plants were found to be the component with a higher adjusted R2, suggesting a greater importance compared to topography and habitat structure. Indeed, it has already been proved that communities change along altitudinal gradients (Stefanescu et al. 2011; Keret et al. 2020) and in relation to microvariations of topography and vertical heterogeneity (Heidrich et al. 2020) but plants explained most of moth community changes.
Plant and moth species
Feeding and habitat specialists and stenotopic species are more affected by habitat degradation and loss (Coulthard et al. 2019; Wölfling et al. 2020). Therefore, the creation of new habitats for species having these characteristics is a desirable practice. In the newly formed clearings, we found as indicator species, habitat specialist taxa such as Selidosema taeniolaria within xero-thermophile clearings and Dichagyris signifera within mesophile ones, indicating a very short-term rearrangement of moth communities toward a specialized assemblage. This was probably possible also because of the presence of xero-thermophile and mesophile grasslands in the nearby areas of the Clarea Valley.
Despite the trends of European moths are unknown (e.g. lack of European Red List of moths), a large amount of common and widespread macro-moth species has recently declined in some European countries (Conrad et al. 2006; Groenendijk and Ellis 2011). This might be linked to habitat shrub encroachment, agricultural intensification, changing forest management and climate change (Fox 2013). Some common species have already experienced enlarging their geographic (latitude and longitude) and altitudinal distribution and/or increasing their generation numbers. For example, the pine processionary has enlarged its altitudinal ranges (Battisti et al. 2005) and Gymnoscelis rufifasciata has increased its number of generations per year in its northern range (Itämies et al. 2011). Even if the open habitat has been newly shaped by opening the forest, we found this species in the xero-thermophile clearings, meaning that even newly-shaped clearings can support highly specialist species.
Larval biology explains most of our results concerning indicator species, especially for forest habitat. The larva of Cosmia trapezina feeds on broadleaved trees and it was found to be very abundant in Apennine beech forests (Infusino and Scalercio 2018), those of Dypterigia scabriuscula on forest grasses with adults preferring closed chestnut woodlots (Infusino et al. 2018), and those of Paracolax tristalis on dead leaves, found yet as a common species in chestnut forests (Greco et al. 2018). Mesophile and xero-thermophile clearings are characterized by species preferring open habitats and trophically linked to herbaceous plants as expected. However, it is not easy to characterize the two groups of indicator species. We can hypothesize that microclimatic conditions of the two habitats are likely the main driver of moth communities as suggested by the presence of Amphipyra tragopoginis and Dichagyris signifera in the mesophile clearings, both found also in mountain forests of Southern Italy (Scalercio et al. 2022).
Conclusions
Even if the study was conducted in a relatively small area, we recorded a high species richness and abundance of moths, including generalist and specialist moths. Thus, small and connected open and forested areas in the Alps can support high diversity and should be considered as a possible priority action for conservation. Conversely to general negative ideas on forest reduction and fragmentation, small size patches may host rich communities (Riva and Fahrig 2023). Thus, the active maintenance of clearings (even those naturally formed by falling of dead trees; Uhl et al. 2023), through for example ride widening in closed forest, could serve as a valuable tool for preserving biodiversity in temperate woodlands even in a long-term scenario. Indeed, newly shaped clearings supported higher species richness and abundance of plant and moth species. However, the scale of investigation affects the results, at regional level we do not know the effects of local management actions (see Schall et al. 2020). Our findings might also be related to connectivity between open areas in wooded environments. Additional studies of these clearings are desirable to determine the impact of such interventions in the long-term on how communities change in relation to the ecological succession, or conversely, to what degree inhibiting succession through repeated cutting can modify these communities. Thus, preserving complex and diverse forests by actively managing them can be a conservation measure along with the rewilding approaches to mitigate biodiversity decrease.
Acknowledgements
All field and laboratory activities were financially supported by TELT—Tunnel Euralpin Lyon Turin SAS. We are thankful to Federica Paradiso, Francesca Cochis, Davide Giuliano, Patrick Artioli, Michela Audisio, Marco Bonifacino, Michele Zaccagno and Nicolò Chiappetta for their contribution to fieldwork and laboratory activities. SS activities were supported by NBFC to the Council for agricultural research and economics, Research Centre for Forestry and Wood, Rende, Italy, funded by the Italian Ministry of University and Research, PNRR, Missione 4 Componente 2, “Dalla ricerca all’impresa”, Investimento 1.4, Project CN00000033. The authors declare that they have no conflicts of interest.
Declarations
Competing interests
The authors declare no competing interests.
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://creativecommons.org/licenses/by/4.0/.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
