Occurrence of bees and bumblebees in bark beetle slit traps from spruce and Fir Woodlands of Central Dinaric Alps

The paper analysed bees by-catch collected in 259 bark beetle slit traps, from eleven localities in Bosnia and Herzegovina. Sampling was carried out in spruce and fir forests in 2020 and 2021. As a by-catch from bark beetle slit traps 84 bee individuals from four families and 13 genera were collected. In the bark beetle slit traps sample, out of 29 bee taxa, 22 species were identified at the species level and eight specimens were left at the genus/subgenus level. The most dominant genera were Megachile with 34 specimens and Osmia represented by 20 specimens in the total sample. The research identified 14 bee species new to the fauna of Bosnia and Herzegovina. The bee species collected in the bark beetle slit-traps were dominated by nesters in cavities, above the ground-nesting bees.

In Europe, bees represent the main group of pollinators (Drossart and Gérard 2020) and in the European Union pollination has an estimated economic value of €15 billion per year (Nieto et al. 2014). The number of bee species in the wild has declined globally, the decrease occurred in the last 10-15 years, with roughly 25 per cent fewer species recorded (Potts et al. 2010, Zattara andAisen 2021). The loss of bee diversity has intensified efforts to develop methods of standardized sampling and the assessment of bee diversity. The three most commonly used trapping methods for collecting bees for biodiversity studies are bowl traps, vane traps, and Malaise traps. Methods of passive bee sampling with traps and statistical models for the purpose of monitoring are still being developed and attempts are being made to understand the scope of bee diversity that is included through sampling with coloured pan traps (McCravy 2018). Bees also find their way into traps that use both visual and olfactory cues to attract pest insects. Researchers work to improve pest monitoring tools to increase target captures and reduce bee by-catch. The bee by-catch composition analysis can help assess biodiversity, determine population fluctuations and range expansions, support monitoring efforts, and identify patterns and processes of broader ecological interest (Spears et al. 2021). Forestry implements traps to control bark beetles, ambrosia beetles, woodboring insects, and wood wasps, thereby minimizing their populations. Additionally, traps in forestry are employed to identify and monitor the presence of pests and invasive insects. Introduced in the late 1970s, pheromone traps were implemented as a replacement for trap trees that had been utilized for over two centuries, serving as a protective measure against the spruce bark beetle (Zahradnik 2015).
Standardised bark beetle slit traps used for the mass trapping of bark beetles can be used with or without an attractant. Bark beetle slit traps are considered to have relatively few wider ecosystem effects on the woodland environment, but this is rarely tested in field conditions. In Bosnia and Herzegovina, monitoring of wild bees by standardized sampling has not been performed and data on the composition of solitary bee and bumblebee fauna from traps are unknown.
The importance of by-catch bees in traps for monitoring of pests has already been confirmed and data were used for biodiversity assessment (Buchholz et al. 2011, Spears andRamirez 2015). By-catch from lepidopteran traps and pitfall traps were used to determine the abundance and diversity of bees (Hatten et al. 2013, Hung et al. 2015, Parys et al. 2021.
Considering the importance of solitary bees and population decline trends, data on collected species from non-target catches are a significant source of data on the distribution and diversity of bees, especially if one takes into account the standardized collection method and longterm monitoring of bark beetles in forestry. This study aimed to determine the composition of honey bee, solitary bees and bumblebees that occur in the bark beetle slit traps in the spruce and fir forests of central Bosnia and Herzegovina.

MAtErIALs AND MEtHODs
Control and monitoring of bark beetles are carried out by the University of Sarajevo -Faculty of Forestry, Plant Protection Laboratory. Collected bees were separated from the sample within a bark beetle monitoring program. The bycatch was processed at the Faculty of Science, Department of Biology, where taxonomic analysis of the bees was carried out. Bees from the by-catch were washed, dried and placed on entomological pins. The identification was carried out using stereo zoom microscope with 90X magnification and taxonomic keys (Friese 1895a, 1985b, 1897, Brohmer et al. 1930, Warncke 1968, Mauss 1994, Amiet et al. 2001, 2004, Michez et al. 2019, Rasmont et al. 2021. The current systematic and species status follows Kuhlman et al. (2023).
The sampling was carried out from May to June 2020 and May to September 2021. The pheromonebaited traps were emptied weekly.

study sites
Sampling was performed at 11 monitoring sites in central Bosnia and Herzegovina ( Figure 1, Table 1).
The monitoring sites are located on Mt. Bjelašnica, Mt. Igman, Mt. Ozren, Mt. Trebević, and Mt. Zvijezda. The traps were used for annual bark beetle control and monitoring. All selected sampling sites were within Picea abies (L.) Karst. and Abies alba Mill. forests in the Dinaric, Pre-Alpine region (Table  1). Local habitat parameters were not estimated on the field due to a lack of field protocol. Values of climatic parameters for each sampling site were extracted from WorldClim raster for each locality (Fick et al. 2017). The parameters used for obtaining bioclimatic data were extracted from WorldClim raster with a spatial resolution of 1 km 2 : Annual Mean Temperature -bio1, Mean Temperature of Warmest Quarter -bio10, Mean Temperature of Coldest Quarter -bio11, Annual Precipitation -bio12 (Fick et al. 2017).

Landscape Pattern Analysis
Land cover maps of the study area were generated from Corrine Land Cover maps with 100X100 meters resolution. The classification of the study region was made according to CLC classification. From the CLC map, 11 forest patches were selected. The centre of each CLC patch coincides with the centroid of traps in the investigated localities, calculated using a centroid point layer in QGIS. The forest patch had a diameter of 1600 meters and an area of 2.0 km 2 . Landscape pattern analysis was conducted with QGIS raster to calculate forest cover and landscape heterogeneity using the Shannon diversity index.
The Shannon diversity index considers the number of different types of environments and their proportion in each landscape (Table 2.). If two landscapes are covered by exactly the same types of habitats, that with the highest Shannon-Weaver value will be the one with the highest category evenness (McGarigal et al. 2012).

Data Processing
For each sampling site, we estimated two bee community level variables: richness (number of species), and  abundance (total bee amount), from which we calculate the Shannon diversity index. The frequency of bee species in the investigated locality was analysed as an indicator of the diversity and composition of bee communities. Based on the data, the similarity between the samples was compared using Beta Diversity and Pairwise comparison Whittaker. The diversity indexes by sampling locality and the number of bees collected in the traps were used to calculate: Taxa_S, Individuals, Dominance_D, Simpson_1-D, Shannon_H, Margalef, Fisher_alpha, Berger-Parker, chao1. The correlation between altitude, the number of individuals and the number of species was tested using linear correlation.
Ordinary Least Squares Regression was used for the correlation between the number of bees and the number of traps per locality.

rEsULts AND DIsCUssION
The relative abundance of spruce and fir forests according to CLC analysis within an area of 2 km 2 varies in the range of 0.07-0.88, median=0.31 (Table 2).
The mean values for 11 sampling sites were calculated using data obtained from WorldClim rasters mean±standard deviation (min, max): Average Annual Mean  The solitary bees and bumblebees collected in the bycatch sample from the bark beetle slit-traps were represented by 84 bees. In 2020, 19 bees were collected from 18 bark beetle slit traps and in 2021, 65 bees were collected from 39 slit traps.
The minimum number of bees in the bark beetle slit traps was collected in May 2021, the bees were most numerous in July 2020 -10 bees and in July 2021 -23 bees. The number of bees in traps decreases in August and September (Table 3). The maximum number of individuals from the genus Osmia was collected in the traps in July and June, while the genus Megachile was most dominant in July and August (Table 3.).
The sampled bees collected in bark beetle slit traps belong to 13 genera ( Table 3). The prevailing bee genera was Megachile represented by 34 individuals, Osmia represented by 20 bees and Apis -10 individuals ( Table 3). The genus Megachile, Osmia and Apis in the sample comprise 76% of the total sample. Apis mellifera was represented in 73% of the sampling localities. In the total sample, 49 bees were females and 35 were males. The ratio of males to females in the genus Megachile was in favor of males, while in the genus Osmia females were dominant (Table 4).
Genus Osmia was represented by eight species and two bees which were identified at the genus level. The genus Megachile was represented by seven species and one bee was identified at the genus level. The lowest number of identified individuals at the species level from the slit traps was in the genus Bombus. Due to the damage of the individuals in the bark beetle slit-traps, only one bumblebee was identified at the species level (Table 5.).
The relationship between altitude, the number of individuals and the number of bee species has a positive correlation r=0.61; p=0.037. Ordinary Least Squares Regression showed a statistically significant correlation between the number of individuals and the number of traps per locality: t=2.77, p=0.02, r=0.64, r 2 =0.41. The number of collected individuals and the number of traps per locality is different; the largest number of traps is in the Igman locality with 93 traps, and in the Trnovo, Crna Rijeka-Željeznica locality two traps were present (Table 7).
From 259 bark beetle slit traps, we collected 84 bees belonging to 26 species. The number of ground-nesting solitary bees and bumblebee species was seven. Overground cavity-nesting bees and dead wood nesters were represented by 18 species and Apis mellifera was a eusocial cavity-nesting bee. The number of dead wood nesting bees and over-ground cavity nesters (78) was greater than the number of groundnesting bees in the by-catch sample. We also collected parasitic bees: Coelioxys conica, Sphecodes majalis and Stelis punctulatissima. Megachilidae were the most abundant taxonomic group represented by 17 bees identified to the species level. The relationship between Megachilidae and bark beetle slit traps placed over the ground is based on the biology of a group that nests in cavities above ground, most often in pre-existing abandoned tunnels of saproxylic insects. The data regarding collected bees is significant from the aspect of understanding the diversity of local bee fauna. The checklist for bee fauna for Bosnia and Herzegovina lists 125 The relationship between altitude, the number of individuals and the number of bee species has a positive correlation r=0.61; p=0.037. Multivariate linear regression analysis correlated habitat heterogeneity with the number of bee species, the number of collected bees and Shannon -H diversity index for each investigated locality (Table 8). Linear regression analysis indicated a negative correlation between habitat heterogeneity and the number of bee species; habitat heterogeneity and the number of bees. The positive correlation was found between habitat heterogeneity and Shanon -H for bee diversity, and the correlations had no statistical significance.
The spruce and fir forests dying due to bark beetle infestation changes the forest structure and has a positive effect on the diversity of pollinators. Bark beetle infestation leads to the death of trees and loss of cover in the tree floor, a greater amount of light can stimulate the growth of herbaceous plants. Bees are positively associated with disturbed forest habitats and forest low tree cover with high floral richness, while abundant dead wood creates suitable conditions for bees (Moretti et al. 2009, Williams et al. 2010, Spears and Ramirez 2015. In addition to the aforementioned loss of leaf mass, measures to control bark beetles are reduced to bare cutting and cleaning of infested areas, which opens up forest habitats. The bee species richness in the forests increases with flower richness and clear-cut size (Taki et al. 2007, Watson et al. 2011, Schüepp et al. 2011. The landscapes with more forests and environmental heterogeneity can provide more resources for bees through resource complementation processes, maintaining their diversity in the landscape. The presence of forest patches close to open areas is of utmost importance for the conservation of bees and pollination services (Rubene et al. 2015). Data related to the wild bees by-catch composition in forest pest management traps, such as bark beetle slit traps in Bosnia and Herzegovina, have not been studied. The absence of long-term bee monitoring can most likely be compensated by the analysis of the composition of the bee fauna in the traps for e.g., bark beetle, as bycatch. By-catches, also, may cause a variety of adverse consequences on populations, food webs and conservation efforts (Revill et al. 2005).
Integrated Pest and Pollinator Management has been proposed as a new framework to further improve the compatibility of pest management practices with pollinator conservation strategies (Biddinger and Rajotte 2015). Considering that integrated pest management and the inclusion of beneficial insects leads to a higher number of insect groups requiring taxonomic identification, the need for collaboration with additional taxonomists for the identification of materials collected from traps becomes crucial. Due to the intricate nature of the task, the challenges highlight the importance of teamwork and synchronized efforts among multiple teams, as well as the regular exchange of materials with expert taxonomists (Spears et al. 2021).
Non-targeted catch has been used to facilitate taxonomic research and describe new species, identify novel invasive alien species, enhance stakeholder knowledge and conduct surveys of non-target insects. The significant presence of Megachile and Osmia solitary bee species in the sample indicates the need for additional investigation  into the effectiveness of bark beetle slit traps for monitoring these genera in the Dinaric Alps. Further research is needed to verify the species diversity of the genera Megachile and Osmia in pest monitoring traps compared to traps used in bee biodiversity research, such as pan traps. Establishing a correlation in the species diversity of the aforementioned genera enables the introduction of an integrative approach to the control of pests and bee monitoring using bark beetle slit traps. Collecting data from forest pest management traps is the first step towards better knowledge of the distribution of solitary bees and bumblebees. Taking into account the standard sizes of bark beetle slit traps and a larger number of traps, it is most likely possible to monitor bees and determine the correlation between the forest habitat heterogeneity and community and the diversity of bees.

CONCLUsIONs
Solitary bees (79.8%), bumblebees (8.4%) and honey bees (11.8%) were detected as by catches in bark beetle slit traps used for the control and monitoring of bark beetles. The sample of bees from traps is dominated by species nesting in existing tree cavities: Osmia 8/22 (36%) and Megachile 6/22 (27%). A positive correlation was established between the number of collected bees and the number of traps per locality. Slit traps for the control of bark beetles are important as a source of information about the bee fauna when there is no long-term research on bee fauna of Bosnia and Herzegovina.
The presence of 14 new species of bees was determined which were not previously recorded for the fauna of Bosnia and Herzegovina. The data on the connection between habitat structure and the number of bee species are contrived and do not correspond to earlier assumptions about the positive connection between the heterogeneity of forest habitats and the number of insects and bees. This study shows that bark beetle slit traps can be used for monitoring of Megachile and Osmia genera within woodland communities.

Author Contributions
OM, AV, MD and SI conceived and designed the research, DP collected sample for analysis, DK carried out the sample separation, AV performed taxonomic analysis, DP and AV processed the data and performed the statistical analysis, OM, AV, MD and SI secured the research funding, supervised the research and helped to draft the manuscript, AV, DK, OM and DP wrote the manuscript.

Funding
This research had no funding support.

Conflicts of Interest
The authors declare no conflict of interest.