Outcomes of a Bat-Box Mitigation Scheme in a Mining Development in South Wales, United Kingdom Open Access

All species of bats are legally protected in the United Kingdom against harm and disturbance by the Habitats and Species Regulations 2017 (Legislation.Gov.UK nd a), which covers disturbance of, damage to, and loss of roosts and resting places. Hence, there is a legal requirement for the competent authorities to ensure that applications for developments are fully compliant and have adequate mitigation measures in place and that the developer has the necessary licenses from the regulating body prior to the granting of planning consent (Legislation.Gov.UK nd b; Natural Resources Wales 2023). Given the widespread occurrence of the more common bat species, it is not unusual that these regulations are needed.

The study site was the 52-ha mine extension to the large operational Selar surface mine located on the southern side of the Neath Valley, about 1 km to the south of Glynneath (decimal grid reference 51.738550 N, −3.612563 E). It has a temperate oceanic climate of relatively warm winters and cool summers with typical low and high winter temperatures of 2–9 °C and summer temperatures of 11–19 °C.

The extent of the extension was defined in the west by the deeply incised and wooded Nant Rhyd-y-Gau Valley, to the east by the deeply incised and wooded Nant Gwerlych Valley, to the south by the former Parish Road, and to the north by a natural ridge line.

Most of the extension was ground disturbed by much earlier underground and surface mining. Much of this land was subsequently planted as commercial coniferous forestry. In the northeast corner, the former Pont Walby Colliery spoil heaps were largely covered in scrub, whereas the north-central and central-east parts were undisturbed land comprising rush-dominated fields with remnant hedgerows and broadleaved copses. To the north, on the valley slopes outside of the extension, the landscape was a patchwork of small woodland and hedges with trees and pasture between the wooded incised valleys (Fig. 1) connecting to the wooded River Neath valley and areas of urbanization. Following the completion of the extension and the larger mine site in 2018, these areas were restored to a patchwork of woodland, hedgerows, agriculture pasture, marshy grassland, and wetland linking the retained and unaffected habitats outside the mine site.

Bat activity surveys prior to the mine development had revealed the presence of Common Pipistrelle (Pipistrellus pipistrellus) and Soprano Pipistrelle (Pipistrellus pygmaeus), Myotis species likely to be Daubenton’s Bat (Myotis daubentonii), and Noctule Bat (Nyctalus noctula) within the temperate deciduous oak-birch woodland of the proposed extension (Celtic Energy 2009). The Pipistrelle bat species are crevice dwelling, whereas the others generally favor voids (Bat Tree Habitat Key 2018).

Within the extension there were five mature oak trees, which had crevice features suitable for bats (Bat Tree Habitat Key 2018) and, according to the then survey and assessment guidelines (Bat Conservation Trust2007), and had moderate to high potential bat roost features.

Here we report on the effectiveness of the bat boxes installed as the mitigation for the loss of the potential bat-resting places along with the possible implications for use of these boxes in future similar schemes.

The preparation for the extension to the existing mine started in 2011 following the issuing of planning consent. To comply with the law, the five potential bat roost trees were checked by a licensed bat ecologist for occupancy by bats prior to their “soft felling” in January 2012. The planning consent stipulated that bat boxes were also to be installed on suitable trees and monitored annually while the mining operation was in progress.

Ten commercially available hardwood boxes of a single-cavity design (Fig. 2) with access from the underside were installed in January 2011 at a height of 3 to 4 m on the southern aspect of 10 selected trees in the woodland and hedgerows near the Nant-yr-Eithin and Nant Rhyd-y-Gau valleys (51.740557 N, −3.616232 E), just northwest and outside the mine extension. The height of the boxes corresponded to the typical height of occupied tree roosts (Bat Tree Habitat Key 2018). Ten more boxes were installed in the northeast on 10 trees at 3 to 4 m high in August 2011 in woodland of the Nant Hir Valley (51.742324 N, −3.609259 E), and a further 10 were installed on 10 trees in the woodland of the Nant Gwerlych Valley (51.738829 N, −3.600804 E).

The single-cavity bat boxes were replaced or supplemented between 2013 and 2015 with boxes with a “Kent” two-slit hardwood design now recommended by the Bat Conservation Trust (2023) (Fig. 3). These boxes were installed in clusters of two or three (depending on the girth of the trees) to face southwest, south, or southeast where possible. In September 2013, 10 boxes were installed on the trees previously selected in woodland along the Nant Gwerlych. In June 2014, eight Kent boxes were installed in clusters of two on the previously selected trees in the Nant Rhyd-y-Gau woodland and hedgerows. Also in June 2014, two Kent boxes were installed in a cluster on the previously selected trees in the Nant Hir Valley woodland. A further 10 were installed in May 2015 as clusters of two or three on the previously selected trees in the Nant Hir woodland.

The direct monitoring by a licensed bat ecologist of the boxes for their occupation by bats and any other evidence of their use by bats was possible from the ground using binoculars and a bright light. This monitoring occurred during daylight in September 2013, June 2014, May 2015, June 2016, April 2017, May 2018, and April 2019. The ecologist was able to discern the genus of bat present from these ground-level observations but not the species. When bats were absent, the following signs were recorded as evidence for box use: the presence of droppings, urine beneath roosting positions, discoloring of entrance holes, scratch marks, grease or rub marks, and evidence of wear on potential access points. To understand the occupation of the boxes over a full year, monthly visits were carried out between November 2020 and October 2021.

In addition to the monthly inspections, audio bat activity was continuously recorded with four bat detectors (Songmeter Mini, Wildlife Acoustics, Maynard, MA) between 14 August 2020 and 8 November 2021. Kaleidoscope Pro ver. 5.6.3 (Wildlife Acoustics 2023) analytical software was used to determine the bat species calling.

Two detectors were installed in each of the Nant Hir Valley and Nant Gwerlych Valley woodlands within the group of trees where bat boxes were sited to identify the species using the woodland habitat where the bat boxes were located and to gauge the levels of bat activity at different times of year.

Table 1 lists the numbers of bat boxes remaining in place and intact 2, 5, and 7 years after installation. The attachment bracket on the original single-cavity boxes was prone to failure, with boxes falling to the ground, but other boxes disappeared, suggesting their removal by unknown parties. The Kent boxes seemed to be more robust and to attract less attention and generally remained in situ longer.

Bird nests were recorded in all but 1 of the remaining 16 single-cavity boxes in 2013, in 2 of the 8 boxes in 2016, and in all of the remaining 4 boxes in 2018. Individual pipistrelle bat species were recorded after 7 years in only two of the single-cavity boxes, one at Nant Rhyd-y-Gau and the other at Nant Gwerlych.

No evidence of nesting birds were found in the Kent boxes on any monitoring occasion. Pipistrelle bats were the only genus recorded, except for a single Noctule on one occasion. The number of boxes with pipistrelles was generally lower in the initial two seasons of installation than later in the survey (Table 2).

Repeated years of occupancy occurred in three boxes at Nant Rhyd-y-Gau, five boxes at Nant Hir, and one box at Nant Gwerlych. However, the pattern of Pipistrelle occupancy differed between box locations, with occupancy at Nant Hir and Nant Rhyd-y-Gau locations occurring sooner after installation than at Nant Gwerlych (Table 3). The Pipistrelle bats appeared to favor particular boxes; they were found in one particular box at Nant Gwerlych and in one at Nant Rhyd-y-Gau in three consecutive years and in five of the boxes at Nant Hir.

Consistent with the annual monitoring program, low numbers of bats were found in the monthly surveys at the Nant Hir and Nant Gwerlych locations between November 2020 and October 2021. At the former location, the trees had three boxes, whereas at the latter the smaller trees had two boxes. The occupation of the boxes by bats differed between the two woodland locations, between the winter (Table 4) and summer seasons (Table 5), between individual trees, and between boxes on individual trees.

At Nant Gwerlych, boxes on two of the trees did not host bats during the winter, but boxes on the other three trees were occupied during at least 1 month. During the summer, the same pattern was recorded, but occupied boxes were on different trees. A similar pattern was recorded for the winter at the Nant Hir location; no bats were recorded in the boxes on two of the trees, but bats were recorded in the summer in at least one box for all trees.

Similar results were obtained for bat acoustic activity recorded at both locations. High bat activity was recorded for Soprano Pipistrelle and Common Pipistrelle, and there was evidence for a wider range of species occasionally present at the locality (Table 6): two species of horseshoe bat, the Lesser Horseshoe Bat (Rhinolophus hipposideros) and the Greater Horseshoe Bat (Rhinolophus ferrumequinum), the Serotine Bat (Eptesicus serotinus), Brown Long-eared Bat (Plecotus auritus), Nathusius Pipistrelle (Pipistrellus nathusii), and Leisler’s Bat (Nyctalus leisleri).

Most bat activity was recorded between April and October, with markedly fewer calls during the November to March hibernation period (Fig. 4). The few calls in the winter period were probably pipistrelles (Bat Tree Habitat Key 2018).

It has long been an established conservation practice to install bat boxes as artificial roosts to support bat populations (Rueegger 2016). The practice as mitigation for the loss of roosts in trees and woodlands has been both advocated and criticized as creating suboptimal roosting conditions and being ineffective substitutes for some bat species (Crawford & O’Keefe 2023; Rueegger et al. 2019).

A range of types and design of boxes are commercially available, some of which are said to be selectively preferred by one or other woodland crevice-dwelling species (Dodds & Bilston 2013; Rueegger 2016). However, the effectiveness of different types of bat boxes remains inconclusive due to the lack of consistent evidence, and longer studies are necessary (Berthinussen et al. 2021).

Three outcomes of this study were notable: (i) the increase in the use of the bat boxes over time, (ii) the differences in effectiveness and resilience between the two box types, and (iii) the effectiveness of these boxes as mitigation for the loss of woodland resting places.

The findings of this study should be interpreted with caution because box monitoring was restricted to a single annual visit from 2011 to 2019 and to one visit per month from November 2020 to October 2021. Bats are known to frequently move from one roost to another. However, the staining and other indications of occupation in some boxes but not others support the findings that the bats were selectively using certain boxes. The findings reinforce the need for long-term and regular monitoring of bat boxes. The occupancy varied between years and months, between individual trees, and between boxes on individual trees; thus, the assumed absence of bats may have been a function of monitoring effort, which is most likely when the numbers of bats are low and only a few resting places are involved. This situation occurs at the Selar Mine, where the greater part of the bat population and principal winter and maternity roosts are likely to be in buildings within the nearby urban areas in the Neath Valley. Here, the woodlands adjacent to the mine provide, from time to time, commuting routes and foraging habitat for some of the bats.

The apparent increase in pipistrelle bats occupying the Kent boxes over the monitoring period at Selar is similar to that reported by Boyd and Stebbings (1989), who found that over a 12-year period the occupation of boxes by Brown Long-eared Bats doubled from 15% to 29%. These authors also recorded bats both roosting individually and in clusters, as was also found at Selar. It is not clear whether the increase in numbers is due to the bats becoming more familiar over time with the physical presence of the boxes and/or the ageing of the wooden boxes (the staining with urine and excreta being an attractant). Meddings et al. (2011) stated that increasing the number of boxes to more than eight per tree did not increase the occupancy beyond the 30% reported by Boyd and Stebbings. Boyd and Stebbings also concluded that the installation of three boxes per tree was the best practice. The boxes should be installed 3 years before the intended beneficial effect is required to enable the bats to become familiar with the availability of the boxes and for the boxes to age. The lack of familiarity and aging may explain why Rueegger et al. (2019) found no bats using their newly installed boxes as mitigation at a mine site.

As might have been expected, the three locations differed in the frequency of box occupation, which suggests that the siting of boxes in relation to bat activity is important to obtain a benefit from the mitigation. The simplistic provision of boxes at any location may not be adequate.

For the Selar Mine development, the single-cavity box was initially selected as the most likely to be favorable for the commonly occurring bat species (pipistrelle, myotis, and Noctule). However, the box design was ineffective owing to occupation by birds; hence, they were replaced with the Kent two-slit box, which excludes birds. Competition between bats and birds for the occupation of boxes has been widely reported (Dodds and Bilston 2013) and can be a major factor in determining the number of boxes required and the amount of benefit achieved. Some researchers have installed both bird boxes and bat boxes as a means of control. Meddings et al. (2011) found that bird occupancy could be reduced by up to 50% with this approach. However, the selection of boxes such as the Kent design is a more effective measure, as demonstrated here, by excluding occupation by birds. Hence, the change to the Kent design negated the need for further compensatory installations.

An unexpected consequence of the selection of the two box types was the relative difference in their resilience to disturbance. The single-cavity boxes were seemingly more prone to damage and removal by unknown parties, whereas the Kent boxes seemed to attract less attention and were generally left alone. Although the boxes were not located near public rights-of-way, the site was within 1 km of human habitation and was accessed by local residents, particularly youths, on weekends. Hence, the risk of interference near to urban areas is also a factor in the choice of the type of bat box. Researchers should not assume that once installed the boxes will remain in place throughout their projected lifespan of 10 or more years.

Given the nature of the landscape, it is not surprising that the Common and Soprano Pipistrelles were the most numerous bat species recorded within the woodlands at the Selar Mine site. Notably for the purpose of the mitigation implemented and compliance with the legislation, the Serotine Bat, Brown Long-eared Bat, Nathusius Pipistrelle, Leisler’s Bat, a Myotis species, and the Lesser and Greater Horseshoe Bats were also recorded, albeit at a very low frequency of occurrence. The two species of horseshoe bats, in terms of the mitigation deployed (see Bat Tree Habitat Key 2018), were reasonably not considered as needing mitigation because they would not have used the type of tree crevices lost or the Kent type of bat boxes provided (Bat Tree Habitat Key 2018). However, the tree crevices lost to the development could have provided resting places for Serotine, Brown Long-eared, Nathusius, and Leisler’s bats (Bat Tree Habitat Key 2018). For these species, the Kent box was less likely to provide resting places, as indicated by the presence of the single Noctule Bat; therefore, other measures would be required. The Kent box is not suitable for maternity roosts for any of the bat species recorded. For these roosts, other structures, such as purpose-built bat roost/houses, would be necessary (e.g., Downs & Wells 2021).

In the context of replacing lost habitat for crevice-dwelling bats, the general provision of the bat boxes might have been considered appropriate mitigation at the mine site, and their sole use complied with the law (Legislation.Gov.UK nd a), assuming that these boxes remained until an alternative became available. Given that the wooden bat boxes used have a finite lifespan of about 10 years and could be prone to disturbance, particularly near urban areas, they will require regular checking and, if necessary, replacement.

These issues raise the question of how long are these boxes needed? Should they be kept in place until natural crevices or voids develop in the surrounding trees or in the planted replacement trees as they mature on the restored land? Such a wait could last decades, well beyond the mining period and typically the short durations specified for bat box mitigation schemes. At the Selar Mine, the bat box mitigation period was specified by the consenting and legally competent authority to be only until the mining operation ceased. If this stipulation were literally followed, the mitigation period would be inadequate and would not comply with the intent of the legislation. Issues of adequacy of the boxes for support for the bat populations and the fitness of the boxes for bat health were not addressed.

This study adds to the existing literature on the use of bat boxes and has highlighted some shortcomings in their deployment as mitigation. These boxes should be considered only a short-term measure, limited to the provision of resting places. In addition to the box installation, the mitigation should consider the creation of future resting features (e.g., holes, crevices, and peeling bark) in other mature trees and transplanted tree monoliths (Herbert et al. 2022: Reason & Wray 2023).

For pipistrelle species, the Kent type of wooden box or a similar design could be used, preferably built with old or weathered wood, with a minimum of three boxes per host tree located on the sunny aspect. The boxes should be installed 2 to 3 years in advance of the loss of the original host trees and maintained until tree features are available to replace them. The boxes need to be made vandal proof and secured from dislodging. Where hibernation and maternity roosts are involved, the boxes are likely to be inadequate, and other mitigation, such as bat houses, need to be provided.

One of the most significant outcomes of the monitoring at the Selar Mine site was the need for a much greater and longer-term commitment when bat boxes are used as mitigation. The provisions for monitoring should include both the boxes and any future resting features created and naturally developing in mature trees. Monitoring should occur at least monthly until the boxes and the newly developed tree features are shown to host bats.

The study was funded by Celtic Energy Ltd. The help of Celtic Energy site staff with installing the box boxes is acknowledged. The premining bat surveys and supervision of the removal of the potential bat trees were undertaken by Dr. Darryn Nash under his NRW Bat Licence. David Price supervised the installation of the boxes and undertook their monitoring under his NRW Bat Licence. Dr. Carlos Abrahams undertook the bat activity recordings and analysis.