Forest structure determines nest box use by Central European boreal owls

Nest boxes represent a popular tool to support secondary cavity-nesting species. Surprisingly, the benefits and limitations of nest boxes for target species in different environments are poorly understood. We performed a 3-years experimental study in two different Central European forests to evaluate nest box use and breeding performance of boreal owl (Aegolius funereus) — a species well known for its readiness to occupy nest boxes. Based on territorial vocalisation, two boreal owl populations 200 km apart were similarly abundant in their environments. However, only the boreal owl population in young restored Norway (Picea abies) and blue (Picea pungens) spruce-dominated forests on mountain plateaus readily occupied nest boxes with the occupancy reaching 8–15%. Nest boxes lost their supporting function for the boreal owl in mature Scots pine (Pinus sylvestris)-dominated forests in the lowland, where the nest box occupancy reached 0–1%. As a result, the population of boreal owls that used nest boxes in the young restored forests produced 10 times more fledglings than the population inhabiting mature Scots pine forests. We explain the differences by the contrasting availability of natural tree cavities between the two study areas being much higher in mature Scots pine forests. For the first time, this study documents differences in nest box use despite similar food availability and population size of the target species. The study provides the findings-related recommendations for deploying nest boxes for boreal owls and points out a general lack of practical guides.

. Nest box use by the boreal owl in Europe, North America, and Asia; the state, locality, elevation, period, the number of boxes, nests, box-years, and the nest box occupancy (expressed as the proportion of occupied nest boxes of checked boxes). The way how authors described their results varied hugely. Therefore, we present the information on the number of boxes, nests, and nest box occupancy from long-term studies either as a sum counted for the entire study period (marked as 'Total') or a yearly mean (marked as 'Mean') with minimum ('Min') and maximum ('Max') values.   Table 2. Basic information about the study areas. Geographical position, the size of the area, elevation, climate conditions (mean daily temperature, mean daily rainfall, and the yearly mean number of days with continual snow cover), habitat composition (based on CORINE Land Cover 2018), and estimates of the number of black woodpecker's territories (per 1 km 2 ) and tree cavities (per 1 ha) excavated by black woodpeckers are shown. Forest structure comparison. Forest habitats dominated both study areas, covering 87% and 47% of the Ore Mts. and Trebon Basin study area, respectively. The area of any of the other habitat types (pasture and grasslands, agriculture, water, and urban areas) did not exceed 15% (for details, see Table 2; data based on CORINE Land Cover 2018). We used vegetation maps of forest stands (the Czech Forestry Institute, 2015-2017) to compare species and age forest structures between the study areas. First, we grouped forest stands into seven categories according to the dominant tree species (i.e., Norway spruce, blue spruce, Scots pine, European beech, other coniferous stands, other deciduous stands, and clear-cuts). We counted the proportion of each species category within each study area. We found that the structure of forest stands differed significantly between the areas (Chisquared test: Chi = 722.59, P < 0.001, df = 6). Norway (26.0%) and blue (15.4%) spruces were dominant tree species in the Ore Mountains, while Scots pine (54.5%) and Norway spruce (31.3%) were dominant tree species in the Trebon Basin (Fig. 2a). Second, we grouped forest stands into three categories according to age (0-40 years, 40-80 years, > 80 years). We counted the proportion of each age category within each study area. We found the age of forest stands significantly differed between the areas (Chi-squared test: Chi = 32.67, P < 0.001, df = 2). The less-than-40 years habitats dominated in the Ore Mts. On the contrary, all age categories of forest stands in the Trebon Basin occurred evenly (Fig. 2b).

Black woodpecker populations.
We recorded a substantially more abundant population of black woodpeckers (assessed based on territorial calls 46 ) and number of cavities excavated by this species in the Trebon Basin than the Ore Mts. (Table 2).

Nest box occupancy.
To evaluate nest box occupancy, we used an existing nest box system in the Ore Mts.
(for details, see the chapter Study areas) in which 246 nest boxes were evenly distributed in forest habitats over an area of 150 km 2 with a density (mean ± SD) 1.65 ± 0.05 nest boxes/km 2 . In 2012-2014, we erected a similar nest-box distribution scheme in the Trebon Basin. Forest habitats of this area were separated by other habitat types (i.e., pastures, agriculture, water, and urban areas) to a greater extent than in the Ore Mts., covering a total of 400 km 2 . In the Trebon Basin, we evenly distributed 245 nest boxes with a density (mean ± SD) of 0.60 ± 0.01 nest boxes/km 2 . We provided the same type of nest boxes in both study areas. The nest boxes were made of 2-cm thick wooden planks. The bottom area dimensions counted for 20-25 × 20-25 cm, the height of the walls was 40 cm, and the roof exceeded the front wall by 5 cm. The entrance had a diameter of 8 cm. All nest boxes were painted with dark brown colour and filled with a 3-5 cm layer of sawdust, resulting in an effective distance of 20-22 cm from the www.nature.com/scientificreports/ bottom (i.e., top of the sawdust) to the lowest part of the entrance. We positioned nest boxes 3-5 m above the ground and kept them in good conditions, i.e., they were repaired, cleaned, and filled with new sawdust after each breeding season. The surrounding of all nest box entrances was kept free of vegetation. The mean age of nest boxes (i.e., the number of years since the nest box installation or reinstallation counted for all nest boxes in all study years) was (mean ± SD) 4.9 ± 4.4 years in the Ore Mts. (n = 722 boxes) and 3.2 ± 1.1 years in the Trebon Basin (n = 724 boxes). We identified nest box use by boreal owls and other animals based on present animals and their active nests or reliable traces of animal activities (e.g., an abandoned nest or nest material) inside the boxes. Each year, we checked 230-246 boxes in the Ore Mts. and 237-245 boxes in the Trebon Basin to identify the number of boreal owl nests. We included nest boxes with the presence or absence (0/1) of boreal owl nests in the GLMM analysis (see below). We further conducted synchronized spring (i.e., in April 2015 and May 2016-2017) and autumn (in September 2016 and October 2017) nest box inspections in both study areas to identify nest box use by boreal owls and other animal taxa. We synchronously checked 230-246 nest boxes (i.e., 230, 246, and 246 boxes in both spring and autumn) and 56-245 nest boxes (i.e., 245, 242, and 237 boxes in spring and 0, 56, and 200 boxes in autumn) during spring and autumn nest box inspections in 2015-2017 in the Ore Mts. and Trebon Basin, respectively. We included nest boxes with the presence or absence (0/1) of other taxa in the CCA analysis (see below). We also calculated the nest box occupancy rate as the number of boxes with present animal taxa or their traces per all checked nest boxes within a specific period and year. Alternatively, we presented the nest box occupancy as the proportion of occupied nest boxes of all erected nest boxes.

Boreal owl nests. We accomplished additional inspections in the boxes where boreal owl nests occurred
to determine the number of eggs, fledglings, nests depredated by martens (Martes sp.), and nests abandoned by boreal owl females with no fledgling produced (according to 19 ). We counted boreal owl breeding productivity based on these data; however, we did not include additional inspections in statistical analyses.

Owls' vocalisation.
At the beginning of April and May in 2015-2017 (a total of six recording events), we distributed audio recordings across both study areas to monitor the presence and absence of all owl species based on territorial vocal activity. In the Ore Mts., we placed 36 audio recorders during each recording event on sampling points within a regular grid with a span of (mean ± SD) 2.0 ± 0.3 km (min-max: 1.6-2.6 km, 25-75% range: 1.8-2.1 km, see also 47 ). In Trebon Basin, we placed 32 audio recorders within a span of 2.5 ± 0.6 km (min-max: 1.7-3.4 km, 25-75% range: 1.9-2.7 km). We installed audio recorders (Olympus DM650, Olympus Corporation, Japan) with automatic time-recording settings during suitable weather conditions (i.e., without strong wind and precipitations) on tree trunks at the height of 1.5 m above the ground. We left audio recorders exposed for one night during each recording event (April, May 2015-2017). We set the recording time of each audio recorder (at least) from 8 PM to 6 AM, allowing us to evaluate owl vocalisation during continuous 10-h recording at each sampling point (recorder) and recording event. We collected a total of 216 and 192 recordings (sampling points) in the Ore Mts. and Trebon Basin, respectively, out of which we excluded 13 and 14 recordings due to technical failures. As a result, we used a total of 2030 h (i.e., 203 sampling points) in the Ore Mts. and 1780 h (i.e., 178 sampling points) in the Trebon Basin to evaluate territorial owl calls.
We transformed particular audio recordings into spectrograms and analysed them using AMSrv software 48 , setting the spectrum at 1 min with FFT length, the window size of 4096 × 1366. We used no filters to remove background noise. We recognized the vocal activities of owls based on territorial calls of particular species 46 . We included sampling points with vocal presence or absence (0/1) of each owl species into subsequent analyses. Simultaneously, we calculated the vocal occupancy rate as the number of sampling points with the vocal presence of owl species per all sampling points, sorted according to the sampling period and year. Based on vocal activity, we also estimated the density for each owl species per 10 km 2 , considering the area of 3.14 km 2 (radius 1 km) per one sampling point.

Small mammal abundances. The boreal owl primarily feeds on small mammals, mainly Microtus
and Myodes voles, Apodemus mice, and Sorex shrews 18,49 . The abundance of voles and mice affects breeding characteristics, including territorial vocal activity 47 , breeding performance 19 , home range size 50 , and parental investment 51 . Therefore, we assessed the abundance of small mammals using snap trapping carried out every year at the beginning of June by setting up snap traps at six blocks (30 × 90 m, 4 × 10 traps, 10-m span) within representative habitats of each study area. The traps were left for three days in the field and checked every morning. We determined all captured small mammals (n = 102 individuals) to species and grouped them into four categories: Apodemus mice, Microtus voles, Myodes voles, and Sorex shrews. We also calculated the abundance index as the number of captured individuals of each category per 100 trap nights for each trapping site (for details, see 18 ).

Statistical analyses.
We used generalized linear mixed models (GLMM) in R 4.0.2 software 52 to assess the effect of the study area and year on the abundance of small mammals. Using lmer function (package Lme4), we created four models with Apodemus mouse, Myodes, and Microtus vole, and Sorex shrew abundances as dependent variables. We calculated the number of trapped individuals per 100 trap nights for each trapping site, study area, and year (trapping index) for each dependent variable. The study area, year, and the interaction of the study area and year were used as independent variables and the trapping site as a random effect. We also calculated partial relationships between the study areas and years using a post-hoc test (function lsmeans in lsmeans package). To compare the nest box occupancy by boreal owls between the study areas, we used GLMM analysis (lmer function) with a binomial distribution of the dependent variable. We examined the effect of the study area on the occurrence of boreal owl nests (0/1). We included the density of nest boxes (i.e., the number of nest boxes within a buffer of 1000 m of each nest box) as a covariate and ID box and year as random effects.
We performed a multivariate canonical correspondence analysis (CCA) in Canoco 5 software 53 to examine the effect of primary predictors, i.e., the study area, year (2015-2017), and period (April, May) on the vocal occurrence (0/1) of each owl species. We used the sampling point of each audio recording as a categorical covariate. Further, we computed another CCA analysis to examine the effect of primary predictors, i.e., the study area, year (2015-2017), and period (spring, autumn) on nest box occupancy (0/1) by boreal owls and other animal taxa (i.e., pine marten Martes martes, passerines, common goldeneye Bucephala clangula, insects, bats). We used the number of boxes in a buffer of 1000 m around each box (continuous variable) and ID box (categorical variable) as covariates.  Table S1). The abundance of Myodes voles, Microtus voles, and Sorex shrews did not differ between study areas and years (Supplementary Information, Table S2). However, the abundance of Apodemus mice was significantly affected by the interaction of the study area and year (GLMM; Chi = 13.45, df = 32, P = 0.004, Supplementary Information, Table S2). A post-hoc test showed that Apodemus mice were significantly more abundant in 2017 in the Ore Mts. compared to the Trebon Basin (post-hoc test; t = 3.58, P = 0.011, Supplementary Information, Table S3. In the Ore Mts., the abundance of Apodemus mice also differed significantly between 2016 and 2017 (post-hoc test; t = − 4.21, P = 0.004, Supplementary Information, Table S3).

Owls' vocalisation.
Based on territorial vocal activity, we identified the same five owl species -boreal owl, tawny owl Strix aluco, long-eared owl Asio otus, Eurasian pygmy owl, and Eurasian eagle owl Bubo bubowithin each study area. The vocal occurrence rate of owl species (all owl species included) varied from 0 to 0.59 and 0 to 0.58 per sampling point in the Ore Mts. and Trebon Basin, respectively (Supplementary Information, Table S4). Using CCA, we found a significant effect of the study area and year on the vocal occurrence of owl species, but the sampling period (April, May) had no effect ( Table 3). The first and second ordination axes explained together 94.7% of the variability. Boreal and tawny owls were the most frequently recorded owl species in both study areas and all years ( Supplementary Information, Table S4). Comparing the study areas showed that the boreal owl vocal occupancy rate was similar and simultaneously the highest in 2015 in both study areas (Fig. 3, Supplementary Information, Table S4). However, it was slightly higher in 2016 and 2017 in the Ore Mts. compared to the Trebon Basin (Fig. 3, Supplementary Information, Table S4). Contrary, the vocal occupancy rate of the tawny owls was similar in 2016 in both study areas, but it was higher in 2015 and 2017 in the Trebon Basin compared to the Ore Mts. ( Supplementary Information, Table S4). Pygmy owls occurred more frequently in the Trebon Basin than in the Ore Mts.; however, long-eared owls showed an opposite trend (Fig. 3, Supplementary  Information, Table S4). Eagle owl was a rare species in both study areas; it was absent in 2015 and 2017 in the Ore Mts. and 2016 in the Trebon Basin (Fig. 3, Supplementary Information, Table S4).
The density of vocalising owl populations reached yearly 0-1.87 and 0-1.85 individuals per 10 km 2 in the Ore Mts. and Trebon Basin, respectively (for details, see Supplementary Information, Table S4). Boreal and tawny owls were the most abundant species in the Ore Mts., followed by long-eared owls, pygmy owls, and eagle owls. Tawny and boreal owl were the most abundant species in the Trebon Basin, followed by pygmy owl, long-eared owl, and eagle owl ( Supplementary Information, Table S4). Table 3. The results of CCA analyses. First analysis: the effect of the study area (Ore Mts., Trebon Basin), year (2015-2017), and sampling period (April, May) on the vocal occurrence (0/1) of each owl species. Second analysis: the effect of the study area, year, and sampling period (spring, autumn) on nest box occupancy (0/1) by animal taxa (boreal owl, goldeneye, pine marten, passerines, bats, and insects). We identified six main differences. First, boreal owls nested in boxes (pooled spring and autumn inspections) more often in the Ore Mts. than the Trebon Basin (GLMM; Chi = 43.27, df = 1441, P < 0.001, Table 4). Second, during the spring inspections, the nest box occupancy reached 7-10% in the Ore Mts., but only 0-1% in the Trebon Basin ( Supplementary Information, Table S5). Third, the total number of boreal owl nests (both active and abandoned) found during spring and autumn nest box inspections was more than 10 times higher in the Ore Mts. (76 nests) than in the Trebon Basin (5 nests, Table 5). Thus, the total nest box occupancy reached 8-15% in the Ore Mts. and only 0-1% in the Trebon Basin. Fourth, the boreal owl females breeding in nest boxes produced yearly 128.3 ± 49.4 eggs and 63.3 ± 52.1 fledglings (mean ± SD, n = 76 nests) in the Ore Mts., but only 6.0 ± 5.6 eggs and 5.3 ± 4.7 fledglings (n = 5 nests) in the Trebon Basin (Table 5). Fifth, pine martens predated six boreal owl nests in the Ore Mts. and one nest in the Trebon Basin (Table 5). Finally, the boreal owl females abandoned 24 nests in the Ore Mts., and no in the Trebon Basin (Table 5).

Nest box occupancy by other animals.
We identified a total of 487 animal activities during synchronized 2,194 spring (April-May) and autumn (September-October) nest box inspections in the two study areas ( Supplementary Information, Table S5). Using CCA, we found a significant effect of the study area, year, and sampling period on the occurrence of animal species in nest boxes ( Table 3). The first and second ordination axes of the model explained 98.7% of the variability together. Boreal owls, passerine birds, and pine martens were more frequent users in the Ore Mts., while bats, insects, and common goldeneyes occurred more often in the Trebon Basin (Fig. 4). During the spring (April-May) nest box inspections, passerine birds (Paridae > 90%) and boreal owls were the most frequent dwellers of nest boxes in the Ore Mts. (Fig. 4). Their nest box occupancy reached 5-30% and 7-10%, respectively (for details, see Supplementary Information, Table S5). Pine martens rarely occupied nest boxes in the Ore Mts. during spring (nest box occupancy was ≤ 1%; Supplementary Information, Table S5. The   Table S5). During the autumn nest box inspections, no animals inhabited nest boxes in the Ore Mts. Only additional abandoned nests of passerine birds (nest box occupancy reached 6-8%) and one boreal owl nest (< 1%) were Table 5. Boreal owls nesting in 2015-2017 in the Ore Mts. and Trebon Basin: the vocal occupancy rate, the density of vocalising individuals, the proportion of spring (i.e., only nests during spring inspections) and total (i.e., all nests) nest box occupancy, the number of boxes, all nests, active nests (i.e., during spring nest box inspections), additional nests (i.e., during autumn nest box inspections), successful nests, nests predated by martens, nests abandoned by boreal owl females (including the proportion of all nests), and the total number of eggs and fledglings.  Table S5). Insects (Vespidae) appeared in only one nest box in the Ore Mts. ( Supplementary Information, Table S5). On the contrary, insects (Vespidae and Apidae > 90%) were most often dwellers of the nest boxes during autumn inspections in the Trebon Basin (Fig. 4); their occupancy reached 44-52% ( Supplementary Information, Table S5). Additional abandoned passerine nests also occurred during the autumn inspections in the Trebon Basin, reaching the occupancy of 0-19% (Supplementary Information, Table S5). Finally, bats occupied nest boxes in this study area (Fig. 4), with less than 2% ( Supplementary Information, Table S5).

Discussion
Population size and nest box use. Based on territorial vocal activity, we found that the two boreal owl populations in the Ore Mts. and Trebon Basin were of similar size in 2015, and that the one in the Ore Mountains was more abundant in 2016 and 2017 than in the Třeboň Basin. However, the Ore Mts. ' boreal owls used nest boxes much more frequently than their counterparts in the Trebon Basin in all study years; the yearly nest box occupancy reached 8-15% and 0-1%, respectively. As a result, the boreal owl population using nest boxes in the Ore Mts. produced 10 times more fledglings than the same population in the Trebon Basin.
In general, the readiness of boreal owls to use nest boxes is inconsistent over the Holarctic region, reaching the nest box occupancy of 0-66% (for details, see Table 1). Some studies have shown that boreal owls preferred using nest boxes to natural cavities; however, other studies have documented that they did not use nest boxes at all or only rarely. The authors have suggested various reasons for the variability in nest box occupancy rate, including the availability of natural cavities, different breeding success in nest boxes and tree cavities, food availability, and nest box age. Our experimental study has documented that the occupancy rate differed significantly between the two study habitats, even though food abundances and the boreal owl population sizes were comparable. We suggest that the main reason for this contrast lies in the different availability of natural cavities that resulted from different forest structures between the two study areas (the first hypothesis). The explanations are following: 1. Availability of natural cavities Natural nest-site opportunities for boreal owls are represented almost solely by the cavities excavated by black woodpeckers 3,16 . Their availability increases with increasing age of the forests, and simultaneously, it is higher in deciduous and pine forests than spruce forests 7,54,55 . In our study areas, the species and age forest structure differed substantially. It influenced the abundance of black woodpecker populations and the availability of tree cavities excavated by this species being much higher in mature Scots pine-dominated forests of the Trebon Basin than young restored forests of the Ore Mts. During our long-term study in the Ore Mts. (since 1999; Zárybnická, unpublished data), we recorded even a higher nest box occupancy rate by boreal owls, up to 26% (2001)(2002)(2003)(2004)(2005)(2006), than in the current study (2015)(2016)(2017). Over the years, this decrease in a nest box occupancy rate probably resulted from gradual vegetation succession accompanied by the increasing availability of tree cavities excavated by black woodpeckers (Zárybnická and Kilb, unpublished data). However, the abundance of the black woodpecker population and the availability of natural cavities are still much lower in the Ore Mts. than in the Trebon Basin nowadays. Therefore, we suggest that the lack of natural tree cavities was the main reason for a substantially higher readiness of boreal owls to occupy nest boxes in the Ore Mts. Following our third hypothesis, we also found a lower nest box occupancy rate by other secondary cavity-nesting birds, i.e., passerine birds, in the Trebon Basin compared to the Ore Mts. (0-19% vs. 5-30%). Additionally, pygmy owls -which almost solely use natural tree cavities for their breeding and avoid nest boxes in Central European conditions 4 -were substantially higher in their population size in the Trebon Basin than the Ore Mts. Riegert and Kloubec (unpublished data) have confirmed that pygmy owls frequently nest in natural cavities while avoiding nest boxes in the Trebon Basin.

Age of nest boxes
The occupancy rate of nest boxes by boreal owls significantly decreases with their age 3,56 . The increased risk of nest predation by martens in older boxes 57,58 or food depletion in the vicinity of old nest boxes 59 have been suggested to explain this effect in boreal owls. However, we found a significantly higher nest box occupancy rate in the Ore Mts. than in the Trebon Basin, even though the mean age of nest boxes (all boxes in all study years) was about 1.7 years higher in the Ore Mts. than in the Trebon Basin. Therefore, the age of nest boxes was not a reason for differences in the nest box occupancy rate between the study areas.

Maintaining nest boxes
The quality of nest boxes and their maintenance can affect nest box use by boreal owls. For example, it is essential to install nest boxes high enough above the ground, regularly relocate and maintain nest boxes, keep nest box content dry and clean, and keep the surrounding of the nest box entrance free of branches (Zárybnická, unpublished data). Still, experimental studies are lacking. We used the same quality, maintenance, and installation of nest boxes in both study areas, preventing these factors from consideration as relevant reasons for differences in boreal owls' nest box occupancy between the study areas.  63 . In the Ore Mts., hornets were absent in nest boxes, and only wasps occupied one nest box during the autumn inspections. Thus, insects did not limit the boreal owl population in the Ore Mts. In the Trebon Basin, hornets only occupied 0-11% of our erected nest boxes during the spring inspections but up to 44-52% during autumn inspections. The period of nest box occupancy by insects did not overlap with that of the boreal owl in the Trebon Basin, avoiding interspecific competitions. Further, martens can raise their young in nest boxes and thus potentially compete for nesting sites with boreal owls. However, we discovered only three boxes with pine marten adults providing their young during spring inspections in the Ore Mts., and one such case was recorded in the Trebon Basin. These rare cases document that pine martens did not compete for nest boxes with boreal owls. Finally, passerine birds occupied up to 30% and 8% of erected boxes during spring inspections in the Ore Mts. and Trebon Basin, respectively. Since passerines are alternative prey for the boreal owl 3,18 , we are convinced that these birds would rather avoid the competition for nest boxes with the boreal owl. Overall, it seems clear that the competition with other animals was not a reason for the differences in nest box use between the study areas (the fourth hypothesis).
Breeding performance in nest boxes. Martens depredated only 8% of nests in the Ore Mts., and one nest failed for the same reason in the Trebon Basin. The rate of nest predation by martens in nest boxes and natural cavities varies across regions. For example, in Germany, martens depredated 59% of owl nests placed in boxes, but only 24% in natural tree cavities 64 . In the Spain Pyrenees, the trend was the opposite; martens destroyed 50% of nests in tree cavities, but only 15% in nest boxes 65 . During the long-term study in the Ore Mts., we found out that the predation rate of boreal owl nests by martens varied hugely among years, reaching 0-50%; and the availability of Apodemus mice was its main drive 66 . In addition to nest failure due to predation, we also recorded that females abandoned 20-52% of nests in the Ore Mts., probably due to insufficient providing of females and nestlings by males 67 . Our study did not compare owl breeding success in nest boxes and tree cavities. However, Hruška 60 performed long-term research (2006-2020, 123 nests) on a boreal owl population nesting in natural tree cavities in the Vysočina Hills in the central Czech Republic. Hruška 60 found that 33-63% of boreal owl nests were successful each year, martens depredated 6-50% of nests, and up to 25% failed due to water flooding, including one nest where nestlings drowned (see also 68 ). Water flooding does not happen in maintained nest boxes and has never been identified as a reason for nest failure in our study areas (Zárybnická, unpublished data). These findings indicate that breeding success of boreal owls nesting in wooden boxes and natural cavities is similar. Alternatively, maintaining and preventing nest boxes from undesirable environmental processes can increase owl breeding success in nest boxes compared to natural cavities.

Recommendations for nest box deploying.
We have documented that nest boxes can effectively support boreal owl populations and be a suitable tool for collecting biological information and material. At the same time, nest boxes can lose their supporting function for the target species under specific environmental conditions. Based on our findings, we provide the following specific recommendations for nest box deploying in Central European conditions to support boreal owl populations. First, nest boxes should be distributed preferably in forest habitats that suffer from short-term or long-term lack of natural tree cavities. Examples of such habitats are young forest stands, intensively managed forest plantations, and forest stands exposed to other anthropogenic (e.g., atmospheric pollution), climatic (e.g., windthrow calamity, heavy snowfall), and biotic (e.g., an outbreak of insects or fungi) process followed by extensive forest losses. Such habitats suffer from the lack of black woodpeckers and other woodpeckers as well 3,4 . The size of the black woodpecker population can be a bioindicator of the availability of natural cavities for nocturnal boreal owls and a signal to provide additional nesting opportunities.
Second, nest boxes should be distributed preferably in coniferous forest habitats at high and middle elevations (above 600 m a. s. l.), providing optimal climate conditions for the study species 15,34,69 . Among owls in the Czech Republic, only the boreal owl prefers inhabiting the highest elevations 15,34 . Our findings support the elevational preference because the boreal owl was the most abundant owl species in all years on the Ore Mts. plateau. At the same time, the situation differed in the Trebon Basin lowlands, in which tawny owls dominated during two of three study years. Therefore, elevational optimum should be considered when practical support for the boreal owl is developed.
In general, during our study, we have identified a lack of specific recommendations for deploying nest boxes for the boreal owl. For example, it would be helpful to guide which habitats and environmental conditions to prefer for nest box deployment, which microhabitats to prefer for nest box installation, how nest boxes to maintain, when nest boxes to relocate, and which nest box dimensions and protection against predators to use. We conclude the comprehensive practical guides should be a topic for further review.

Conclusions
The abundances and community structures of small mammals were similar in the Ore Mts. and Trebon Basin during the whole study, apart from one year (2017) when Apodemus mice peaked in the Ore Mts. Boreal owl populations were comparable in size based on territorial vocalisation in both study areas, but their readiness to occupy nest boxes differed substantially. Nest boxes were an efficient tool to support the availability of nesting opportunities for the study species in young restored forests in the historically air-polluted Ore Mts. However, Scientific Reports | (2022) 12:4735 | https://doi.org/10.1038/s41598-022-08792-y www.nature.com/scientificreports/ nest boxes lost their supporting function for the boreal owl in mature Scots pine forests in the Trebon Basin. Therefore, deploying nest boxes should be considered carefully under local environmental conditions.

Data availability
All data produced from this study are available in the manuscript (Tables 1, 2