sign in sign up
<<
Home , Acronicta strigosa, Coppicing, Eriogaster catax, Eupithecia dodoneata

Journal for Nature Conservation 43 (2018) 75–84

Contents lists available at ScienceDirect

Journal for Nature Conservation

journal homepage: www.elsevier.com/locate/jnc

Successful maintenance of by government-funded management T of coppiced ⁎ Matthias Doleka, , Ádám Kőrösia,b, Anja Freese-Hagera a Büro Geyer und Dolek, Alpenblick 12, D-82237 Wörthsee, b MTA-ELTE-MTM Ecology Research Group, Pázmány Péter s. 1/c, 1117 Budapest,

ARTICLE INFO ABSTRACT

Keywords: Coppice and coppice with standards are important habitats for species of light forests, because regular cutting of Conservation re-growth applied in this type of -use produces forest structures that fulfil habitat requirements of these Management species. Coppicing had been a traditional forest-use system in Central Europe for centuries, but it has drama- Coppice with standards tically declined in the last 100 years. As a conservation strategy, a contract-based conservation program for Lepidoptera forests (CBCP Forest) was introduced in 2005 in Bavaria, Germany, that supports among other activities cop- Euphydryas maturna piced forests. We studied the effects of CBCP Forest on the occurrence and abundance of two Lepidoptera species catax scheme of European conservation concern (Euphydryas maturna, Eriogaster catax) as umbrella species in ten forest pat- Contract-based conservation program for forest ches. With the introduction of the CBCP Forest no further loss of coppiced forests occurred in our study area, in (CBCP forest) contrast, a new coppiced forest was initiated. The occurrence and number of larval webs of our studied species Natura 2000 were highest after 5–10 (E. catax) and 10–15 (E. maturna) years of cutting. This means that coppicing can maintain a dynamically changing forest where both study species can find patches of optimal age and thus coppicing ensures their long-term persistence. We also found larval webs in high forests adjoining coppiced forests, but in much lower number and larval webs occurred always around forest gaps (e.g. roads) or edges. Onset of specific management including coppicing in these high forests highly increased the occurrence of E. maturna and E. catax. The survival of the populations of E. maturna and E. catax in our study area is closely connected to the persistence of coppicing as supported by CBCP Forest. As they are considered as umbrella species for numerous inhabitants of light forests, we assume a positive effect on the whole community. CBCP Forest is a valuable conservation tool for coppices and should be actively advertised to promote greater uptake by forestry practitioners and stakeholders.

1. Introduction are left to grow older to provide for building purposes. “Coppice” is the same system without leaving old trees. In reality, in- For a long time, conservation efforts in Central Europe concentrated termediate systems exist, e.g. some trees are not cut during one rotation on open habitats such as grasslands, fens, bogs, -rows, while cycle, but regularly felled during the next cycle for fire-wood or other forests were rather neglected. Financial support for open habitats purposes. started in the German Federal state of Bavaria, for example, in 1982 by In contrast with Central Europe, coppice have been a offering farmers payments if they met prescribed conservation-friendly conservation issue for some years in the United Kingdom (Buckley practices, since 1992 such agri-environmental schemes are part of the 1992; Crowther & Evans 1984; Waring & Haggett 1991; Warren & Key European policy (Güthler et al. 2012). The equivalent funding of forests 1991). Butterfly conservation was of particular interest in that system was established in Bavaria in 2005. In general, conservation in forests (e.g. Shreeve 1981; Warren & Thomas 1992), but other invertebrates has become an increasingly accepted and discussed issue over the last have been considered as well (e.g. Hill et al. 1990; Greatorex-Davies & 10–15 years. Important conservation targets and funding tools have Marrs 1992). General recommendations and suggestions for conserva- been developed, including funding of coppice and coppice with stan- tion to emerge from these studies are similar to the recommendations dards (overview see e.g. Häusler et al. 2008; Güthler et al. 2005). and suggestions of the Bavarian strategy as presented by Liegl & Dolek, “Coppice with standards” is defined as a forest that is regularly cut at an 2008 and Liegl et al. (2008). interval of about 20–35 years, mostly to produce fire-wood, while some In recent years, many aspects of coppicing for conservation have

⁎ Corresponding author. E-mail address: [email protected] (M. Dolek). https://doi.org/10.1016/j.jnc.2018.02.001 Received 7 March 2017; Received in revised form 1 February 2018; Accepted 1 February 2018 1617-1381/ © 2018 The Authors. Published by Elsevier GmbH. This is an open access article under the CC BY license (http://creativecommons.org/licenses/BY/4.0/). M. Dolek et al. Journal for Nature Conservation 43 (2018) 75–84 become an issue in European countries and beyond. Müllerová et al. around 1600 decrees were released to avoid over-use of forests, re- (2014, 2015) described the long-term history of coppicing, its aban- sulting in the concept of “sustainable forestry” (“sustainability” first donment and its implications on plant species diversity in Southern mentioned by Carlowitz, 1713), which is still applied. Coppiced forests Moravia (). The importance of light and sun penetration were also treated in detail (e.g. Hamm 1895). In 1900 about 12% of the in the forest as well as open-grown trees as key habitats for saproxylic forests in Bavaria were still coppices or coppices with standards (cal- has been highlighted by several authors (e.g. Cizek & Vodka culation based on data in Borchert 2008). The coppiced forest area 2013; Schmidl & Bußler 2008; Sebek et al. 2016; Seibold et al. 2016). decreased steadily over the last 100 years, resulting in about 0.15–0.2% Fartmann et al (2013) compared butterflies and especially of the present forest area (calculation based on data in Albrecht & butterflies of different stages of the coppicing cycle in the Alsace (NE Müller 2008; BWI 2017, Liegl & Dolek, 2008). Because of sustainable France). The importance of historical decisions in the development and forestry with permanent cover (“Dauerwald”) and the loss of his- maintenance of coppicing was described for the Osnabrück Region (NW torical forestry practices the forest interior is darker and less exposed to Germany) by Mölder (2016). Vacik et al. (2009) gave an overview on the sun than in previous centuries; species that favour these conditions the importance of coppicing for and its distribution in tend to be listed in Red Data books. , , Macedonia, and Croatia. Coppicing as an ancient The number of studies on coppicing and its influence on in- forestry practice and its application in the context of biodiversity con- vertebrates is limited, but these studies usually conclude that coppiced servation is not only a European issue. Kobayashi et al. (2010) studied forests form important habitats for them (Benes et al. 2006; Dolek et al. butterflies in Japanese forests that are coppiced to produce logs on 2008a; Fartmann et al., 2013; Spitzer et al., 2008; Vandekerkhove et al. which the Japanese forest mushroom Lentinus edodes grows. 2016, 2009). Butterflies, and especially certain butterfl ies of light for- Experimental re-creations of coppices have also been established. In ests, seem to be typical for properly performed conservation coppices. Southern Moravia (Czech Republic) coppice and coppice with standards Settele et al. (2009) mention species of European conservation concern, is applied on an experimental scale to an initially 98-year-old forest some of these species occur in our study forests. For the butterfly Eu- comprising 54% Quercus petraea, 18% Picea abies, 15% , phydryas maturna (Linnaeus, 1758) (Scarce Fritillary, Lepidoptera: and 10% Larix decidua (Kadavý et al. 2011). In Krumlov forest (Czech Nymphalidae) and the Eriogaster catax (Linnaeus, 1758) (Eastern Republic) experimental was applied in 1999 to a former Eggar, Lepidoptera: ) the number of known sites is ex- coppice with standards. Ten years after the treatment, light-demanding tremely limited in Bavaria and Germany after they declined drastically oligotrophic plant-species significantly increased, indicating positive due to habitat loss (Dolek et al. 2017a). These species were chosen as consequences of restoration, while heavy thinning also brought about focal species during the establishment of CBCP Forest (Dolek et al. the expansion of native ruderal species (Vild et al. 2013). Together with 2008b) and in the present study of its application. the regained interest in coppices, certain improvements are also sug- In this study, we analysed the relationship between forests, financial gested, such as the application of a tree-oriented (originally support according to CBCP Forest and the populations of our focal designed for high forests) to coppices (Manetti et al. 2016). In 2005, the species as indicators of biodiversity. Do CBCP Forest and coppicing help Bavarian Contract-Based Conservation Program of Forests (CBCP to preserve our focal species? As coppicing with standards produces a Forest; Vertragsnaturschutz Programm Wald, abbreviated to VNP Wald) dynamic habitat, we considered the financially supported forests in was initiated. CBCP Forest includes several measures that are funded, their full extent as well as the yearly coppiced areas. In this context, we e.g. for old forest stands, dead wood, and bog forests. The program (see analysed distribution and dynamic occurrence of our focal species. We StMUV 2017) is regularly revised and adapted to new, improved finally draw conclusions on the effects of the present management and knowledge and experiences. Right from the beginning, coppiced forests give recommendations for improvement. were supported, most are coppices with standards. This part of the program did build on extensive work, which defined how different 2. Sites and methods kinds of management can create desired forest structures and habitats for threatened species and species guilds (Liegl & Dolek, 2008). 2.1. Sites The present funding is based on a guideline published in 2014 (StMELF & StMUV, 2014), its intention and basics are described in The study was conducted in the Steigerwald region in northern StMUV (2015). The funding of coppiced forests according to this Bavaria (Fig. 1), which is an important stronghold for many endangered guideline consists of two parts: (i) For the whole forest managed as butterfly and moth species of light forests, but also for endangered coppice funding is received for not transferring it to . This species of other groups such as saproxylic beetles and ants (Liegl funding amounts to 80 € /ha*year if it is a coppice with standards and & Dolek, 2008, Dolek et al. 2008a, 2009). We concentrated on those rotation time is ≤ 30 years, to 55 €/ha*year if rotation time is > 30 forests that hold populations of Euphydryas maturna and/or Eriogaster years and to 50 €/ha*year if it is a coppice without standards and ro- catax, which are considered as important umbrella species for this kind tation time is ≤ 25 years. There is a commitment of 5 years; and, (ii) of habitat. At present, there are 10 parcels of forest that are used as For cutting the wood there is additionally a funding of 750 €/ha and for coppice or coppice with standards in our study area (Table 1). Most of separate management (usually at about half of the rotation period or a them are traditionally used as commons with complex rights of forest bit later) 600 €/ha. The coppiced area is measured every year to cal- use. Moreover, these regulations are more or less unique to each com- culate the exact amount of financial support. For a 120 ha forest man- munity (Bärnthol 2003). They are usually surrounded or bordered by aged as coppice with standards with a rotation period of 30 years and forests under present standard forestry (high forest). For one continuous no separate management this results, for example, to yearly payments forest of about 1500 ha, which includes several parcels of coppiced of (approximately) 120/30 = 4 * 750,- € plus 120 * 80,- €. On a vo- forests, we analysed the larval web distribution of E. maturna and E. luntary base, private and selected public forest-owners and other pri- catax between coppiced forests (481 ha) and high forests (1047 ha). vate law and public bodies and their associations may apply for con- tracts to receive this support. 2.2. Study species Historically, coppicing was for centuries one of the most widespread kinds of forest-use in Central Europe. Written decrees exist since the Euphydryas maturna is a Palaearctic species and was widely dis- end of the Middle Ages, which usually restrict forest-use in a way that tributed in Germany (Dolek et al. 2017a), but has declined dramatically provided for coppice and coppice with standards (Bärnthol 2003; during the past few decades in Germany and Europe; it is listed in all Poschlod 2017). Especially grazing of and the timing and the relevant Red Data books (Europe: van Swaay et al. 2010, Germany: right of the forest-use were regulated at that time. In general, starting Reinhardt & Bolz 2011, Bavaria: Voith et al. 2016) and in Annex II and

76 M. Dolek et al. Journal for Nature Conservation 43 (2018) 75–84 + Yes No Yes E. maturna with CBCP Yes No Yes No 2014 2014 Yes – – 2010, 2012 2010, 2012, 2014 Yes 2016 2010, 2012, 2014 Yes 2010, 2012 – – – – – 2005-2016 2005-2016 Study years E. maturna with CBCP 2014-2016 2005-2010, 2012-2016 (partim) 2005 2005 (partim) 2005 2013 2005 2005, 2014 E. catax with CBCP Yes Yes Yes Yes Yes Yes Yes Yes Yes 2005, 2006, 2009, 2005, 2006, 2009, Study years E. catax with CBCP 2005, 2006, 2009, 2005, 2014, 2016 Yes 2012, 2014 2012, 2014, 2016 (partim) 2014 2012, 2014, 2016 (partim) 2005, 2006, 2009, 2005, 2006, 2009, 2014 2005, 2006, 2009, 2012, 2014, 2016 2012, 2014 2005, 2006, 2009, 2012, 2014 2012-2014 29 – < 30 < 30 < 30 < 30 29-30 25-26 Rotational time with CBCP (years) na 28 25 28

Fig. 1. Location of study area in Bavaria and Germany.

IV of the Habitats Directive of the European Union. In Central Europe, +* + No Yes E. maturna before CBCP No No No* No Yes Yes the females lay eggs in batches on leaves of excelsior, or very Yes No* rarely Ligustrum vulgare and Viburnum opulus (Dolek et al. 2013). Young larvae feed communally in silk-woven nests, stop feeding in late summer and diapause in the leaf litter until early spring. These silk-woven nests are the best developmental stage to survey,

fi : no suitable habitat. 2004 monitor, or study E. maturna. The best period to nd these larval webs 2004 + – – is roughly around mid or end of July / beginning of August. Until this Study years E. maturna before CBCP – – 2002 2000 - 2004 2002 - 2004 2001 2002 2003 2002 period, larvae have grown long enough and have built a conspicuous 2002 web. At earlier dates, the web is small and may easily be overlooked, later in the year larvae leave the web and it gradually degrades. 2016, introduction of CBCP in 2005 in all forests except F1 (2012)). Rotational time: rotation of cuts of re-growth: the number of years before further coppicing on the

Eriogaster catax is a Palaearctic species and in Central Europe is – extinct in large areas (Dolek et al. 2017b); it is listed in the Red Data + + No E. catax before CBCP Yes Yes Yes Yes Yes Yes No Yes No books of Germany and Bavaria as critically endangered (Rennwald et al.

2011; Wolf 1992) and in Annex II and IV of the Habitats Directive of the – 2004 European Union. Females lay egg batches on branches of the larval food – plant, which in Central Europe is predominantly spinosa (Dolek et al. 2008b), while spp. may dominate in other regions. Egg 2002 2003 batches hibernate and hatch around April. The larvae usually form a – silk-woven web on which they bask in the sun or hide in the shade, they – Study years E. catax before CBCP 2004 2001 - 2002, 2003 2004 (partim) 2001 (partim) 2002 2004 2002 2002 2002 – never rest or feed in the web (Ruf et al., 2003), as E. maturna does. Nevertheless, the larval web is normally conspicuous and used for monitoring and surveys. The best time is roughly around beginning of May, when larvae had enough time to build the web, but did not yet leave it for their solitary stage, which starts at about the 3rd instar. After leaving the web, solitary larvae may also be found, as they be-

come rather large, before they pupate in the leaf litter. CBCP (years) Both our focal species are specialists of light-penetrated forests and similar habitats that consist of a mixture of grassy patches, shrubs, and trees that receive enough sunshine and air humidity. As these are the conditions that shall be created by the funding programme, E. maturna F1 80 timber Site Size (ha) Rotational time before F2 58 na F3 310F4 ca. 35 38 na F5 159 33 F6 62F7 65F8 na 41 32 38 and E. catax were selected as indicator species. F9 182F10 23 32 na Table 1 Overview of characteristics of coppiced forests analysed in this study (2000 same piece of land again. Abbreviations: na: data not available, *: old records before 2000;

77 M. Dolek et al. Journal for Nature Conservation 43 (2018) 75–84

At the start of the funding program in 2005, the distribution of E. age category as an explanatory variable. We used a binomial error maturna and E. catax populations and their size was known from our distribution and sub-patch ID was the random factor (Zuur et al., 2009). previous work, which started in 2000 (E. maturna) and 2001 (E. catax). Data analysis was conducted using the mgcv (Wood et al. 2016) and During the last 10 years we repeatedly visited these forests again and lme4 (Bates et al. 2015) packages of the R 3.4.2 statistical software (R searched and counted larval webs to determine if populations still exist Core Team 2017). (Table 1). Prior to field work we used orthophotos and information on the cycle of coppicing to detect appropriate habitats and concentrate 3. Results our field work on promising areas. Based on the field observations, we categorized each sampling site as dry, mesic or humid. 3.1. The system of coppice with standards

2.3. Analysis Since the funding started, no further coppice with standards was given up in our study area. This contrasts with an immense rate of land- We also collected data from forest users and owners as well as from use change in the decades before. Moreover, within our study area one funding authorities on the size and exact location of each years cut in new coppice with standards of 80 ha was initiated in a forest patch, the forest to allow predictions on the rotational time until a certain where coppicing had been given up several decades ago (Table 1). forest patch is coppiced again. This enabled us to analyse larval web As a further sign of regained interest, rotational time in the coppiced distribution in relation to age of coppiced forest patches. forest decreased from 32 to 38 years in the beginning of the 2000s to We divided all the 10 forest patches (both the coppiced areas and 25-30 years at present (Table 1). Long and increasing rotational cycles the non-coppiced high forests) into smaller sub-patches. Sub-patches of had previously been identified as a major threat to the persistence of the coppiced areas were cut in different years. Sub-patches in non- the coppiced forests, as they were considered as an early warning sign coppiced high forests represented potential habitats where the number of decreasing interest in this labour-intensive kind of forestry. of larval webs of the study species was regularly sampled. Each ob- servation in our dataset contains the number of larval webs of the study 3.2. Coppiced forests species in each year and sub-patch. In some cases, only one of the study species was sampled in a given year and location, therefore the sample Larval webs of E. catax were found in all 9 originally coppiced size is different for the two species. We measured the area of each sub- patches out of the 10 forest patches. In these 9 forest patches, webs patch using a GIS and calculated the time length in years since the last were present (at least in one year) in 97 out of the 174 sub-patches that cut of the sub-patch for each sample year. We aimed to assess the sta- were checked (n = 1147 observations). (Note that the sampling period tistical relationship between the presence and number of larval webs of and thus the number of observations were slightly different among E. catax and E. maturna as response variables and the area and the time forest patches.) In the binomial model, the smoother of years since last length since last cut as explanatory variables. cut was highly significant and sub-patch area was not included. The We analyzed the data of coppiced areas and high forests separately. smoothing spline indicates that the occurrence probability of E. catax For the coppiced areas, we applied generalized additive mixed models larval webs has a peak between 5–10 years after coppicing (Fig. 2, (GAMM), because we assumed a non-linear relationship between time Table 2). The best model for the log(x + 1) transformed number of since last cut and the response variables (Wood, 2017). We used a bi- larval webs included sub-patch area as well. The smoother of years since nomial error distribution in the models for presence data and a quasi- last cut was again highly significant, the smoothing spline shows a peak Poisson error structure when the log(x + 1)-transformed nest number around 5–7 years after coppicing (Fig. 2, Table 2). was the response variable. Forest patch ID was used as a random factor. Larval webs of E. maturna were found in 6 out of the 8 forest patches Most of the sub-patches were sampled in several subsequent years, thus that were checked for this species, but in one patch there was only one we incorporated an autoregressive moving average (ARMA) correlation observation so we included only four patches in the analysis. In these structure in the GAMM models to account for temporal autocorrelation four patches, webs were present (at least in one year) in 50 out of the (Zuur, Ieno, Walker, Savaliev, & Smith, 2009). We checked the model 100 sub-patches that were checked (n = 1271 observations). In the residuals for autocorrelation. We aimed to keep our models as simple as binomial model, the smoother of years since last cut was highly sig- possible in terms of the number of explanatory variables, so we fitted nificant and sub-patch area was not included. The smoothing spline only two models for each response variable: one included a smoothing indicates that the occurrence probability of E. maturna larval webs has a function of the time length in years since the last cut, and the other peak between 10–12 years after coppicing (Fig. 2, Table 2). Due to the additionally included the area of the sub-patch (without smoothing). high number of zeros in the E. maturna dataset, in the analysis of the The better model was selected based on the AIC value (Akaike 1973). number of larval webs we included only those 50 sub-patches where Our dataset was strongly zero-inflated, i.e. it contained many zero webs were observed at least in one year (n = 641). The best model for observations, that made model fitting and parameter estimations quite the log(x + 1) transformed number of larval webs did not include sub- difficult. We did not find any specific method to handle zero-inflation patch area and the smoother was also highly significant. Smoothing within GAMM models (but see Skaug et al. 2016 for GLMMs), therefore spline shows a peak between 12–15 years after coppicing (Fig. 2, we omitted all dry sub-patches, because none of our study species oc- Table 2). curred there, and those forest patches where a given species did not E. catax closely follows forest dynamics created by coppicing if occur (see Results for more details) to reduce the number of zero ob- abiotic conditions allow the development of suitable habitats (Fig. 3). servations. In each year, larval webs occur on those patches that have reached the In case of the high forests, time length since the last cut (i.e. forest proper age (1–3 years after coppicing) and are disappearing on those age) was usually not identifiable or applicable. During our study, patches that are growing older (beyond 20 years). Quite often, first however, some coppicing activities were launched in a few sub-patches webs already occur one year after the cut, while density peaks in the as a conservation measure and the date of the last cut was known there. following years. In older patches the number of webs decrease year We categorized the sub-patches according to the time since last cut as after year with some webs persisting or re-occurring along forest roads follows: very young forests (0–5 years since last cut); young forests (6- or in larger gaps. In sum, E. catax populations are spatially rather dy- 11 years since last cut); and, old high forests (> 30 years). Then we namic and inhabit different parts of the coppiced forest in different performed Fisher's exact tests on the presence of both species in the years. three categories. Finally, we applied a generalized linear mixed effects E. maturna follows forest dynamics as well, but not as closely as E. model with presence of nests of each species as a response and forest catax (Fig. 4). Most larval webs occurred consistently over many years

78 M. Dolek et al. Journal for Nature Conservation 43 (2018) 75–84

Fig. 2. Estimated smoothing splines and point-wise 95% confidence bands from the GAMM models for the presence and log(number) of larval nests of both study species. All these splines were highly significant indicating a non-linear relationship between years since last cut (x-axes) and presence and abundance of larval nests. Note that the curves show their peaks at different forest age for the two studied Lepidoptera species. in the same area, as this area remained suitable and newly coppiced structure of the sites, where larval webs were found in high forests patches were not suitable as they covered dry hill-sides. Only from 2012 (Fig. 6). All these larval webs were situated in conditions that differ onwards newly created habitats on more humid patches were used. This from the normal forest structure in high forests, e.g. edges along forest occurred just in time, when habitat quality in the former main area roads (E. maturna) or outer forest edges and food plant patches outside diminished. In sum, E. maturna populations are also spatially dynamic the forest (E. catax). Additionally, both species used forest gaps that are and inhabit different parts of the coppiced forest in different years, created by a cut of trees in an intensity that is unusually high for timber although it seems more difficult to achieve suitable conditions. forests. Indeed, most of them were also initiated as conservation mea- sures to support the focal species also in a high forest. 3.3. High forests 4. Discussion We sampled 45 sub-patches in 7 forest patches in the high forests that were not supported by CBCP Forest. However, in 19 sub-patches, We used the presence and distribution of two Lepidoptera species some coppicing, usually under the funding scheme, was started during (E. maturna and E. catax) to indicate the qualitative conservation result our study. We found 28 webs of E. catax (n = 206 observations) and of supported coppicing. For both species, we had conducted previous 187 webs of E. maturna (n = 528 observations) in these 45 sub-patches. studies on their structural and ecological habitat needs and we hy- fi The Fisher's exact test was marginally non-signi cant for E. catax oc- pothesized that they would benefit from ongoing coppicing. This hy- fi currence (p = .078) and highly signi cant for E. maturna pothesis was largely supported by our results. In our study area, po- (p < < 0.0001)(Table 3). The binomial GLMM showed that the pulations of both species are rather restricted to coppiced forests, only a fi probability of occurrence of E. catax was marginally signi cantly higher few larval webs were found in high forests, all of them in places where in the young sub-patches than in the high forests. The probability of vegetation structure was atypical for high forests. However, coppicing fi occurrence of E. maturna was signi cantly higher both in the young and activities commenced in high forests resulted in highly increased oc- in the mid-aged sub-patches than in the high forests. We note that in currence of both species. Although the results were only marginally or both models the random term (sub-patch ID) explained a very high nearly significant for E. catax, it was probably due to the low number of proportion of the total variance (Table 2). sub-patches with coppicing (Table 3). In coppiced forests, populations of both species followed the dynamic creation of new habitats by yearly 3.4. Comparing coppice with standards with high forests coppices. However, the time profile of the two species' response to coppicing was different, since occurrence probability and larval web Distribution of larval webs of E. maturna and E. catax is extremely abundance of E. catax reached a peak at 5–10 years, while that of E. unequal between coppiced forests and timber forests with 1544 E. maturna was at 10–15 years after coppicing. To our knowledge, this is maturna webs and 1690 E. catax webs in coppiced forests and 187 re- the first study that demonstrates the effects of coppicing on population spectively 28 webs in high forests although the area of timber forest is persistence of two umbrella species in detail. more than double the size (Fig. 5). Although the high forest was not In sum, our results show very positive effects of coppicing and its each year as thoroughly checked for larval webs as the coppiced forest, funding on E. maturna and E. catax. As these two species are considered as we concentrated our search on obviously suitable habitats, this figure as umbrella species, we expect similar effects for many other en- is still extremely convincing. As a further support, we analysed the dangered species. Liegl & Dolek, 2008 and Liegl et al. (2008) described

79 M. Dolek et al. Journal for Nature Conservation 43 (2018) 75–84

several core resources and stages in succession that form important habitats within coppiced forests. Our focal species belong to the group “thermophilous species of half-open forest structures (early to inter- 2 = -0.63

θ mediate successional stages) with a close relation to humid sites (hy- grophilous)”, which includes more butterfly and moth species (butter- flies: Coenonympha hero (Linnaeus, 1761), Lopinga achine (Scopoli, ff 1 = -0.25, 2 = 0.27 2 = 0.38 2 = 0.38 1763); : strigosa ([Denis & Schi ermüller], 1775), Ste- θ θ θ θ gania cararia (Hübner, 1790), dodoneata (Guenée, 1857), Rheumaptera hastata (Linnaeus, 1758)). If these species are present,

1 = 0.98, most likely their habitat needs are also met on our study sites. More- 1 = 0.40, 1 = 0.45, 1 = 0.60, θ θ Φ θ over, core resources of abiotically different conditions or including dead wood and later successional stages are also defined. Based on the forest dynamics created by coppicing we expect these to be supported as well. Clearly, the support of and management by coppicing should be carried on for conservation purposes. Nevertheless, the populations of E. maturna decreased extremely in 2015 and 2016. This decrease is not connected to habitat quality as it is influenced by coppicing, but probably caused by other external influ- ences such as (i) ash () dieback and (ii) variable weather conditions with a pronounced drought in 2015. Ash dieback is caused by a fungal pathogen (the ascomycete Hymenoscyphus pseu- doalbidus) and effects all age classes of the trees (Pautasso et al. 2013). Many trees finally die and are lost as larval resource for E. maturna. Actually, for E. maturna the resource loss occurs even earlier as quite often those twigs that are preferred for egg-laying and young larval development (Dolek et al. 2008b) are usually infected first by the pa- thogen. This reduction of larval resources was followed in 2015 by a drought that led to wilting of trees in summer. We also found some egg batches dried out and we assume strong effects of drought on egg and young larval mortality as early instars are known to prefer humid conditions (Freese et al. 2006). Despite the creation of good habitat

zp quality and increasing area we might lose E. maturna due to these in- ects is given. In the binomial GLMM, high forest was the reference category. 4.79 < < 0.0001 74% 6.34 < < 0.0001 94% ff or fl ff

− − uences. There is no such e ect on E. catax which is less restricted to ff xed e humid conditions and has a di erent life cycle. fi Coppices and coppices with standards have been recognized as ect Proportion of variance explained by the random term Correlation parameters (ARMA) ff important strongholds of biodiversity, hosting specific guilds and spe- 3.31 (0.69) 4.61 (0.73) -value of cies groups of more or less open and light-penetrated forest structures. z 0.59 (1.23) 0.483.15 (0.90) 0.63 3.50 < 0.001 1.61 (0.80) 1.993.63 (0.73) 0.046 4.96 < < 0.0001 − − The resulting importance for conservation has now become generally accepted. Consequently, the amount of still existing coppiced forests was surveyed and estimated for Bavaria at about 4,000–5,000 ha (Albrecht & Müller 2008, Liegl & Dolek, 2008). A large proportion of these forests are situated in the Steigerwald natural region including our study area. This amount is only a tiny fraction of 1.5–2.0% of the -value, in GLMMs the t once existing area of coppiced forests in Bavaria, as in 1900 there were about 175,000ha used as coppice with standards and 86,000ha as coppice (Rossmann 1996). Until the start of CBCP Forest, coppiced edf F p estimate (SE) t 3.57 8.02 < 0.0001 < 1% 3.33 7.59 < 0.0001 6.2% Smooth term Fixed e forests had been transformed to high forest at an immense rate, the last large parcel even during the negotiations that resulted in the start of the funding program in 2005. Considering this background, the re-start of coppicing in an 80ha forest parcel is a great success: not only the loss of coppices was stopped, but also the trend was reversed and a new cop- pice was created. Without CBCP Forest this would not have happened. Beside the financial support by the program, public appreciation gen- cant terms are bold. In GAMMs the fi sub-patch areas(years since last cut) 0.06 (0.013) 4.51 < < 0.0001 s(years since last cut) mid-aged mid-aged young young erated by this funding is of unmeasurable importance. The re-introduction of coppicing in one forest is a good example that this change of forest use is possible if the product (fire wood) can be utilized. The re-growth of stools is especially good if deer density is kept occurrence s(years since last cut) 4.72 23.1 < < 0.0001 32.4% nest number s(years since last cut) 5.88 29.77 < < 0.0001 22.4% occurrence nest number occurrence High forest (intercept) occurrence High forest (intercept) low enough. Pyttel et al. (2013) also reported “that the re-sprouting ability of 80–100 year old oak trees originating from former coppice management is still high and little influenced by harvesting methods”. This contrasts with former observations that stools lose re-growth E. catax E. maturna E. catax E. maturna ability with age beyond normal coppicing cycles and hopefully may encourage further re-introductions of coppicing for conservation pur- poses. Coppiced Forest type Response variable Explanatory variable(s) High forest Table 2 Parameter estimates of the best models. Signi

80 M. Dolek et al. Journal for Nature Conservation 43 (2018) 75–84

Fig. 3. E. catax distribution in one coppiced forest in 2005, 2009, 2012 and 2014 in relation to age of the coppiced patches.

Fig. 4. E. maturna distribution in one coppiced forest in 2005, 2010, 2012 and 2013 in relation to age of the coppiced patches. Diagonally hatched: dry, unsuitable habitat; horizontally hatched: conservation measurement areas in high forest, vertically hatched: converted to coppice with standards in 2012.

81 M. Dolek et al. Journal for Nature Conservation 43 (2018) 75–84

Table 3 positive effects were an enlargement of the light forest phase, which is Number of sub-patches in different age classes of high forest with and without larval nests not clearly visible under our general monitoring. Nevertheless, it might of the study species. allow shorter rotation cycles and still produce fire-wood thick enough,

Forest age thus indirectly playing an important role. This part of the funding program should be reconsidered and investigated through further stu- Very young Young (6–11 Old (> 30 y) dies. (< 6 y) y) As the program CBCP Forest works well and our results show that

Eriogaster catax Presence 4 1 10 large areas are needed for the long-term survival of populations the Absence 17 11 163 program should be advertised actively to attract more forest owners / Euphydryas maturna Presence 24 8 26 users to adopt this management. When considering the size of the ha- Absence 48 12 410 bitat, always keep in mind that at a given time, only part of the habitat is in a suitable successional stage and large parts might not be suitable at all depending on abiotic conditions (see calculations by Freese et al. 5. Recommendations 2006). Most promising for the adoption of coppicing are areas where coppiced forests still exist or where certain light-demanding forest The present program of CBCP Forest works well and fulfils ex- species (such as the butterflies Coenonympha hero, Lopinga achine, Sa- pectations concerning maintenance of biodiversity as indicated by our tyrium ilicis etc.) are known to survive. focal species. Nevertheless, small adaptations could be introduced. Nowadays, fire-wood use and production are changing. Some forest Forest users are concerned about their fire-wood getting too thin if users do not use fire-wood any more, but chopped wood pieces they follow rotational times of 28 years or less, but they consider 30 (Hackschnitzelheizung). They use the whole tree including all branches, years as acceptable. Our results indicate that a rotational time of about thus no substrate is left for saproxylic beetles. As a contrasting positive 25 years would be optimal for our focal Lepidoptera species, as the role effect, no nutrients are recycled by decaying branches and nutrient- as reproduction habitat ends at this age for E. maturna and for E. catax poor conditions are preserved. As long as only some of the forest users even a few years earlier (Fig. 2). Nevertheless, coppicing is only per- adopt this technique (as it is the case), no actions are needed, but the formed if a suitable product is produced. About 29–30 years could be an development has to be observed. acceptable compromise between fire- and habitat Although the program is well accepted and produces the desired creation. In oak-dominated forests this should be applied for both results, some changes from outside the system disturb the positive de- coppice and coppice with standards. This means a considerable change velopment. The above-mentioned ash dieback is a big problem for E. for coppice, where the program now sets a maximum of 25 years, while maturna and there is no long-lasting solution so far. Despite re- it remains the same in coppice with standards (maximum 30 years for commendations against ash planting in forestry, because saplings will higher funding, lower support for longer cycles). In areas with fast die in few years due to ash dieback, these trees should be planted to be growing tree species dominating the coppice, such as Alnus incana along used as food plant until they wither. In the long run, resistant ashes that rivers in southern Bavaria, 25 years might be justified. also support E. maturna are needed, otherwise Central European po- As a minor part within the program, a secondary cut is supported at pulations of this butterfly will be lost if they do not shift host-plant. about the middle of the rotational period. Originally this aims to pro- Climatic fluctuations with extremely variable weather conditions duce better quality fire-wood by removing shrubs and thin stems from and a pronounced drought in 2015 form further threats which are stools so that remaining stems may grow thicker. This is applied in one difficult to handle within the system. Keeping water in catchments studied forest only and forest users are convinced that it helps a lot, but within the forest after rainfall may reduce effects to some extent and there are no comparative studies. Within CBCP Forest the assumed have been introduced in present planning. As young larvae still need a

Fig. 5. Comparison of number of larval webs ofE. maturna and E. catax in coppiced forest and high forest. Summary of all study years.

82 M. Dolek et al. Journal for Nature Conservation 43 (2018) 75–84

Fig. 6. Exact kind of habitat ofE. catax (n = 62 webs) and E. maturna (n = 158 webs) used for reproduction outside coppiced forests. Summary of all study years. good amount of direct sunshine, a shift to more shaded areas or 304,33–41. northern expositions is not a solution. Crowther, R. E., & Evans, J. (1984). Coppice. Forestry commission leaflet. London: Her Majesty`s Stationary Office. In our study area, enhanced nutrient input is apparently not a Dolek, M., Bußler, H., Schmidl, J., Geyer, A., Bolz, R., & Liegl, A. (2008a). Vergleich der problem yet, but it is observed in many other parts of Bavaria, including Biodiversität verschiedener Eichenwälder anhand xylobionter Käfer, Nachtfalter und forests. New approaches are needed to counteract this large-scale pro- Ameisen. In Bayer. Landesamt für Umwelt (Ed.). Ökologische Bedeutung und Schutz von Mittelwäldern in Bayern. Tagungsband zur Fachtagung "Nutzung der Mittelwälder" am cess, otherwise even well-working conservation instruments such as 31.05/01.06.2006 in Bad Windsheim (pp. 7–37). Augsburg: LfU. CBCP Forest may not be as successful as they could be. Dolek, M., Freese-Hager, A., Geyer, A., & Liegl, A. (2008b). Die Habitatbindung von Maivogel und Heckenwollafter: Ein Vergleich von zwei Lichtwaldarten. In Bayer. Acknowledgements Landesamt für Umwelt (Ed.). Ökologische Bedeutung und Schutz von Mittelwäldern in Bayern. Tagungsband zur Fachtagung "Nutzung der Mittelwälder" am 31.05/01.06.2006 in Bad Windsheim (pp. 38– 56). Augsburg: LfU. The studies were conducted on behalf of the Bavarian Agency for Dolek, M., Freese-Hager, A., Bussler, H., Floren, A., Liegl, A., & Schmidl, J. (2009). Ants ff Environment (Bayerisches Landesamt für Umwelt, LfU) und the re- on : e ects of forest structure on species composition. Journal of Conservation, 13, 367–375. sponsible conservation authority of the regional administration (Höhere Dolek, M., Freese-Hager, A., Geyer, A., Balletto, E., & Bonelli, S. (2013). Multiple ovi- Naturschutzbehörde, Regierung von Mittelfranken). All conservation position and larval feeding strategies in Euphydryas maturna (Linné, 1758) and forestry authorities were very helpful and we are happy about the (Nymphalidae) at two disjoint European sites. Journal of Insect Conservation, 17, 357–366. good cooperation. Especially A. Liegl, Dr. G. Kluxen, and J. Voith Dolek, M., Anton, C., Beinlich, B., Bräu, M., Brunzel, S., Geyer, A., et al. (2017a). Maivogel contributed considerably to the success of the project in their role as Euphydryas maturna (Linnaeus, 1758). BfN: Internethandbuch zu den Arten der FFH- representatives of responsible conservation authorities. Numerous col- Richtlinie Anhang IV. Accessed 17.01.2017 http://www.ffh-anhang4.bfn.de/ffh- fi anhang4-eschenscheckenfalter.html. leagues helped during eld-work over the years. Dr. S. Heppner (LfU) Dolek, M., Anton, C., Beinlich, B., Bräu, M., Brunzel, S., Caspari, S., et al. (2017b). supported us during the compilation of this work; she, M. Gindele-Glasl Heckenwollafter Eriogaster catax. BfN: Internethandbuch zu den Arten der FFH-Richtlinie and R. Gerlach provided helpful comments to a former version of this Anhang IV. Accessed 17.01.2017 http://www.ffh-anhang4.bfn.de/ffh-anhang4- heckenwollafter.html. manuscript. Many thanks also to all forest owners and users for the Fartmann, T., Müller, C., & Poniatowski, D. (2013). Effects of coppicing on butterfly good cooperation, without their input this valuable forestry system communities of woodlands. Biological Conservation, 159, 396–404. would not exist. Freese, A., Benes, J., Bolz, R., Cizek, O., Dolek, M., Geyer, A., et al. (2006). Habitat use of the endangered butterfly Euphydryas maturna and forestry in Central Europe. Animal Conservation, 9, 388–397. References Greatorex-Davies, J. N., & Marrs, R. H. (1992). The quality of coppice as habitats for invertebrates. In G. P. Buckley (Ed.). Ecology and management of coppice woodlands. London: Chapman and Hall. Akaike, H. (1973). Information theory and an extension of the maximum likelihood Güthler, W., Market, R., Häusler, A., & Dolek, M. (2005). Vertragsnaturschutz im Wald - principle. In B. N. Petrov, & F. Csáki (Eds.). 2nd international symposium on information Bundesweite Bestandsaufnahme und Auswertung. BfN-Skripte, 146,1–180. theory (pp. 267–281). Budapest: Akadémiai Kiadó. Güthler, W., Heppner, S., Heusinger, G., & Joswig, W. (2012). Erfolgskontrollen zum Albrecht, L., & Müller, J. (2008). Ökologische Leistungen aktiver Mittelwälder. bayerischen Vertragsnaturschutzprogramm: Flora und Fauna auf VNP-Flächen mit Schatztruhen für seltene Tier- und Pflanzenarten, aber auch Anschauungsobjekt für Grünland, Acker und Teichen. Naturschutz und Landschaftsplanung, 44(7), 197–204. Waldbaukonzepte. LWF aktuell, 62,36–38. Häusler, A., Dolek, M., Güthler, W., & Market, R. (2008). Conservation efforts and stra- Bärnthol, R. (2003). Nieder- und Mittelwald in Franken. Waldwirtschaftsformen aus dem. tegies for forest canopies in Germany: A review of conservation programmes. In A. Mittelalter Bad: Windsheim Fränkisches Freilandmuseum. Floren, & J. Schmidl (Eds.). Canopy Arthropod Research in Central Europe: Basic and Bates, D., Maechler, M., Bolker, B., & Walker, S. (2015). Fitting linear mixed-effects applied studies from the high frontier (pp. 563–572). Nürnberg: Bioform. models using lme4. Journal of Statistical Software, 67,1–48. Hill, D., Roberts, P., & Stork, N. (1990). Densities and of invertebrates in stands Benes, J., Cizek, O., Dovala, J., & Konvicka, M. (2006). Intensive game keeping, coppicing of rotationally managed coppice woodlands. Biological Conservation, 51, 167–176. and butterflies: The story of Milovicky Wood, Czech Republic. FEM, 237, 353–365. Kadavý, J., Kneifl, M., & Knott, R. (2011). Establishment and selected characteristics of Buckley, G. P. (Ed.). (1992). Ecology and management of coppice woodlands. London: the Hády coppice and coppice-with-standards research plot (TARMAG I). Journal of Chapman and Hall. Forest Science, 57(10), 451–458. Borchert, H. (2008). Veränderung des Waldes in Bayern in den letzten 100 Jahren. LWF Kobayashi, T., Kitahara, M., Ohkubo, T., & Aizawa, M. (2010). Relationships between the Wissen, 58,42–49. age of northern Kantou plain (central Japan) coppice woods used for production of BWI (2017). Thünen-Institut, Dritte Bundeswaldinventur - Ergebnisdatenbank, Waldfläche Japanese forest mushroom logs and butterfly assemblage structure. Biodivers Conserv, [ha] nach Land und Waldspezifikation 2012. Accessed 04.10.2017. https://bwi.info. 19, 2147–2166. Carlowitz, & Carl von (1713). In J. F. Braun (Ed.). Sylvicultura oeconomica, oder Liegl, A., & Dolek, M. (2008). Conservation of coppice with standards for canopy ar- haußwirthliche Nachricht und naturmäßige Anweisung zur wilden Baum-Zucht Leipzig. thropods. In A. Floren, & J. Schmidl (Eds.). Canopy Arthropod Research in Central Cizek, L., & Vodka, S. (2013). The effects of edge-interior and understorey-canopy gra- Europe: Basic and applied studies from the high frontier (pp. 551–562). Nürnberg: dients on the distribution of saproxylic beetles in a temperate lowland forest. FEM,

83 M. Dolek et al. Journal for Nature Conservation 43 (2018) 75–84

Bioform. Skaug, H., Fournier, D., Bolker, B., Magnusson, A., & Nielsen, A. (2016). Generalized Liegl, A., Möller, K., Kluxen, G., & Dolek, M. (2008). Der Schutz der bayerischen Linear Mixed Models using’ AD Model Builder’_. R package version 0.8.3.3. Mittelwälder und das bayerische Vertragsnaturschutzprogramm Wald. In Bayer. Spitzer, L., Konvicka, M., Benes, J., Tropek, R., Tuf, I. H., & Tufova, J. (2008). Does Landesamt für Umwelt (Ed.). Ökologische Bedeutung und Schutz von Mittelwäldern in closure of traditionally managed open woodlands threaten epigeic invertebrates? Bayern. Tagungsband zur Fachtagung "Nutzung der Mittelwälder" am 31.05/01.06.2006 Effects of coppicing and high deer densities. Biological Conservation, 141, 827–837. in Bad Windsheim (pp. 57–68). Augsburg: LfU. StMUV Bayerisches Staatsministerium für Umwelt und Verbraucherschutz & StMELF Manetti, M. C., Becagli, C., Sansone, D., & Pelleri, F. (2016). Tree-oriented silviculture: a Bayerisches Staatsministerium für Ernährung, Landwirtschaft und Forsten (2014). new approach for coppice stands. iForest, 9, 791–800. Richtlinie über Zuwendungen nach dem Bayerischen Vertragsnaturschutzprogramm Wald Mölder, A. (2016). Small forest parcels, management diversity and valuable coppice (VNPWaldR 2015). http://www.stmelf.bayern.de/mam/cms01/wald/waldbesitzer/ habitats: an 18th century political compromise in the Osnabrück region (NW dateien/vnpwaldr_2015.pdf. Accessed 13.01.2017. Germany) and its long-lasting legacy. iForest, 9, 518–528. StMUV Bayerisches Staatsministerium für Umwelt und Verbraucherschutz (2015). Müllerová, J., Szabó, P., & Hédl, R. (2014). The rise and fall of traditional forest man- Bayerisches Vertragsnaturschutzprogramm. http://www.bestellen.bayern.de/ agement in southern Moravia: A history of the past 700 years. FEM, 331, 104–115. application/applstarter?APPL=STMUG&DIR=stmug&ACTIONxSETVAL(artdtl.htm, Müllerová, J., Hédl, R., & Szabó, P. (2015). Coppice abandonment and its implications for APGxNODENR:283761,AARTxNR:stmuv_natur_0006,USERxBODYURL:artdtl.htm, species diversity in forest vegetation. FEM, 343,88–100. KATALOG:StMUG,AKATxNAME:StMUG,ALLE:x)=X (Accessed 13.01.2017). Pautasso, M., Aas, G., Queloz, V., & Holdenrieder, O. (2013). European ash (Fraxinus StMUV Bayerisches Staatsministerium für Umwelt und Verbraucherschutz (2017). excelsior) dieback – A conservation biology challenge. Biological Conservation, 158, Bayerisches Vertragsnaturschutzprogramm Wald (VNP Wald). http://www.stmuv. 37–49. bayern.de/themen/naturschutz/foerderung/vnp_wald.htm (Accessed 13.01.2017). Poschlod, P. (2017). Geschichte der Kulturlandschaft (2nd edition). Ulmer: Stuttgart. Vacik, H., Zlatanov, T., Traikov, P., & Dekanic, S. (2009). Role of coppice forests in Pyttel, P. L., Fischer, U. F., Suchomel, C., Gärtner, S. M., & Bauhus, J. (2013). The effect of maintaining forest biodiversity. Silva Balcanica, 10(1), 35–45. harvesting on stump mortality and re-sprouting in aged oak coppice forests. FEM, van Swaay, C., Cuttelod, A., Collins, S., Maes, D., López Munguira, M., Šašić, M., et al. 289,18–27. (2010). European Red List of Butterflies. Luxembourg: Publications Office of the R Core Team (2017). R: A language and environment for statistical computing. Vienna, European Union. Austria: R Foundation for Statistical Computing. https://www.R-project.org/. Vandekerkhove, K., Thomaes, A., Crèvecoeur, L., De Keersmaeker, L., Leyman, A., & Reinhardt, R., & Bolz, R. (2011). Rote Liste und Gesamtartenliste der Tagfalter Köhler, F. (2016). Saproxylic beetles in non-intervention and coppice-with-standards (Rhopalocera) (Lepidoptera: Papilionoidea et Hesperioidea) Deutschlands. Bundesamt restoration management in Meerdaal forest (): an exploratory analysis. für Naturschutz, Naturschutz und Biologische Vielfalt, 70(3), 167–194. iForest, 9, 536–545. Rennwald, E., Sobczyk, T., & Hofmann, A. (2011). Rote Liste und Gesamtartenliste der Vild, O., Roleček, J., Hédl, R., Kopecky, M., & Utinek, D. (2013). Experimental restoration spinnerartigen Falter (Lepidoptera: Bombyces, Sphinges s.l.) Deutschlands. of coppice-with-standards: Response of understorey vegetation from the conservation Bundesamt für Naturschutz, Naturschutz und Biologische Vielfalt, 70(3), 243–283. perspective. FEM, 310, 234–241. Rossmann (1996). Lebensraumtyp Nieder- und Mittelwälder. Landschaftspflegekonzept Voith, J., Bräu, M., Dolek, M., Nunner, A., & Wolf, W. (2016). Rote Liste und Bayern. München & Laufen / S.: StMLU & ANL. Gesamtartenliste der Tagfalter (Lepidoptera: Rhopalocera) Bayerns. Bayerisches Ruf, C., Freese, A., & Fiedler, K. (2003). Larval sociality in three species of central-place Landesamt für Umwelt (LfU)19. Accessed 13.01.2017 https://www.lfu.bayern.de/ foraging lappet moths (Lepidoptera: Lasiocampidae): A comparative survey. natur/rote_liste_tiere/2016/doc/tagfalter_infoblatt.pdf. Zoologischer Anzeiger, 242, 209–222. Waring, P., & Haggett, G. (1991). Coppiced woodland habitats. In R. Fry, & D. Lonsdale Schmidl, J., & Bußler, H. (2008). Xylobiontic guild composition and diversity (Eds.). Habitat conservation for – a neglected green issue. Middlesex, England: driven by forest canopy structure and management. In A. Floren, & J. Schmidl (Eds.). The Amateur Entomologists` Society. Canopy Arthropod Research in Central Europe: Basic and applied studies from the high Warren, M. S., & Key, R. S. (1991). Woodlands: Past, present, and potential for insects. In frontier (pp. 299–323). Nürnberg: Bioform. N. M. Collins, & J. A. Thomas (Eds.). The conservation of insects and their habitats. Sebek, P., Vodka, S., Bogusch, P., Pech, P., Tropek, R., Weiss, M. K., et al. (2016). Open- London: Academic Press. grown trees as key habitats for in temperate woodlands: The diversity, Warren, M. S., & Thomas, J. A. (1992). Butterfly responses to coppicing. In G. P. Buckley composition, and conservation value of associated communities. FEM, 380, 172–181. (Ed.). Ecology and Management of coppice woodlands. London: Chapman and Hall. Seibold, S., Bässler, C., Brandl, R., Büche, B., Szallies, A., Thorn, S., et al. (2016). Wolf, W. (1992). Rote Liste gefährdeter Nachtfalter Bayerns, 111, Schriftenreihe Bayerisches Microclimate and habitat heterogeneity as the major drivers of beetle diversity in Landesamt für Umweltschutz214–236. dead wood. Journal of Applied Ecology, 53(3), 934–943. Wood, S. N. (2017). Generalized Additive Models: An Introduction with R (second edition). Settele, J., Dover, J., Dolek, M., & Konvicka, M. (2009). Butterflies of European ecosys- Chapman & Hall/CRC. tems: impact of land use and options for conservation management. In J. Settele, T. G. Wood, S. N., Pya, N., & Saefken, B. (2016). Smoothing parameter and model selection for Shreeve, M. Konvicka, & H. Van Dyck (Eds.). Ecology of Butterflies in Europe (pp. 353– general smooth models (with discussion). Journal of the American Statistical 370). Cambridge: Cambridge University Press. Association, 111, 1548–1575. Shreeve, T. G. (1981). Flight patterns of butterfly species in woodlands. Oecologia, 51, Zuur, A., Ieno, E. N., Walker, N., Savaliev, A. A., & Smith, G. M. (2009). Mixed effects 289–293. models and extensions in ecology with R. Heidelberg: Springer.

84