Alejandro A.Schaaf·Román A.Ruggera·Ever Tallei·Constanza G.Vivanco·Luis Rivera·Natalia Politi

Abstract In tropical and subtropical forest ecosystems,cavities formed by decay processes are a key but scarce resource for birds that nest and roost in them,which makes them a highly sensitive group to logging.The piedmont forest of northwestern Argentina is a complex ecosystem with 113 tree and 120 bird species.It has high logging pressure on the few, well-conserved forest remnants,complicating the delineation of sustainable management guidelines for each tree or bird species in a short time.Our objective was to reduce the complexity of subtropical forests by grouping tree species according to the characteristics used by secondary cavity-nesting birds(i.e.non excavators).In the piedmont forest,50 plots of 0.25 ha were sampled to record cavity trees and cavity characteristics.These were then used in a cluster analysis to form tree groups.Additionally,cavities were searched to identify the bird species using the decay-formed cavities.A total of 187 cavity trees,comprising 23 tree species,were recorded,and these formed four tree groups or clusters.We recorded 86 cavities that were used by secondary cavitynesting bird species.The four tree groups were unequally used by secondary cavity nesters.The tree group that included valuable timber species(Myroxylon peruiferum,Anadenanthera colubrina and Calycophyllum multiflorum)and had the greatest cavity availability represented 71% of total cavity use.Another tree group with valuable timber species (Cedrela balansae and Amburana cearensis),measured ＞73 cm DBH and ＞21 m tall, had cavity entrances ＞0.10 cm2,and contributed 14% of all cavity use by birds.A third group had no highly economically valuable tree species,and included the snag category(i.e.standing dead trees)as well as a 15% of cavity use.The fourth tree group had a DBH ＜0.40 cm,only one highly economically valuable tree species(Cordia trichotoma),and supported no cavity use.The clustering of subtropical trees can reduce the complexity of these forests,hence easing their management by focusing on those groups with tree species showing similar characteristics and providing suitable nesting sites for secondary cavity-nesting birds.

Keywords Biodiversity·Cavity-nesting birds·Conservation·Logging

Introduction

Sustainable forest management should consider both timber harvesting and biodiversity conservation(Lindenmayer et al.2006;M?nkk?nen et al.2014).Biodiversity information gathered in temperate forests enhances the understanding of how to maintain the functioning of these forest ecosystems and how to formulate sustainable forest management guidelines(Lindenmayer et al.2000;McComb 2015;Martinez Pastur et al.2017).Such advances are possible because temperate forests have low species diversity and low richness diversity(Myers et al.2013;Hansen et al.2013).In contrast,there is relatively little biodiversity information on functioning forest ecosystems at tropical and subtropical latitudes,making it difficult to design appropriate sustainable management guidelines for these forests(Drummond and Loveland 2010;Hansen et al.2013).At a global scale,at least 400 million hectares of tropical and subtropical forests worldwide(53% of the total forest area)are or will be under timber management(Pacala and Socolow 2004).Logging has steadily increased in neotropical forests during the last 50 years,particularly in the piedmont forest of northwestern Argentina,where local and international demand for timber has greatly expanded the areas under exploitation (Politi et al. 2012). It is imperative to develop management guidelines that can enhance conservation of the high biodiversity hosted in these forests.

Logging can alter forest structure,ecological processes,and affect animal populations if it is not conducted under sustainable management guidelines (Lindenmayer and Franklin 2002).Tree cavities are a key and scarce resource in forest ecosystems for several vertebrate species(Newton 1998;Politi et al.2009).Approximately 18% of all the bird species in the world use tree cavities as nesting sites(van der Hoek et al.2017).The birds that nest in cavities excavated by other bird species such as woodpeckers or by decomposition processes are called secondary-cavity-nesting birds(Aitken and Martin 2007).Not all of the cavities available in any forest are used by secondary cavity-nesting birds. Instead, birds select cavity types for nesting depending on the tree species,its diameter at breast height(DBH), height, the cavity origin (e.g., excavated or decayed),and/or the internal cavity temperature(Martin et al.2004;Cockle et al.2011b;Edworthy and Martin 2013;Ruggera et al.2016).Woodpeckers play an important role in excavating cavities that are later used by secondary cavity-nesting birds in temperate forests.However,in tropical and subtropical forests,most secondary cavity nesters use decomposition-generated cavities(Gibbons and Lindenmayer 2002;Cockle et al.2011a;Ruggera et al.2016).Thus,secondary cavity-nesting birds in subtropical and tropical forests can be highly sensitive to logging due to the removal of trees with decay-formed cavities(Cornelius et al.2008;van der Hoek et al.2017).

Piedmont forests of northwestern Argentina host 113 tree and 120 bird species(Brown et al.2001,2009;Malizia et al.2005).This abundance of biodiversity makes it difficult to design management guidelines for individual tree and bird species.To address the forest complexity resulting from such high species diversity and to simplify the conservation strategies needed by forest managers,we aimed to(1)group tree species according to their tree and cavity characteristics, (2) relate the use of secondary cavitynesting birds to the tree groups formed,and(3)to use this information to develop forest management guidelines to ensure the conservation of secondary cavity-nesting birds in the piedmont forest of Argentina.

Materials and methods

Study area

This study took place in three sites of the piedmont forest of northwestern Argentina(Fig.1)that have not been logged for at least 45 years.Around 90% of the original piedmont forest range has been converted to urban and agricultural lands(Brown and Malizia 2004).Most of the remaining piedmont forest is logged,and only a few hundred hectares have been set aside for at least 45 years and can be considered reference sites(Politi et al.2009).The piedmont forest is part of the seasonal dry forests of South America (Prado 2000) at elevations between 400 and 900 m a.s.l.The climate is highly seasonal with an annual rainfall of between 800 and 1000 mm,concentrated during the Austral summer(Arias and Bianchi 1996).The dominant tree species are Calycophyllum multiflorum,Phyllostylon rhamnoides,A.colubrina,Myroxylon peruiferum,and Astronium urundeuva(Brown et al.2001;Brown and Malizia 2004).

Fieldwork and data analyses

Fieldwork took place from July 2014 to February 2018.In each site,an area of 100 ha was delimited for sampling.Inside each of these areas,from 10 to 20 plots of 0.25 ha were randomly demarcated (Fig.1). In 50 plots, we recorded all trees with cavities and classified the agent of cavity origin as excavation or decay.We measured cavity tree DBH and height,cavity height,cavity entrance area(using the formula of an ellipse),and cavity internal depth.Trees with cavities were identified by their species,while standing dead trees of different species were pooled into the single category of‘‘snags’’.Cavities with an entranceof ＞5 cm diameter and ＜16.8 m height were inspected with a mini-camera system attached to an extendable pole(Richardson et al.1999).In each plot,the number of cavity trees ＞10 cm DBH was recorded to calculate the cavity tree density,which was categorized into:＜1 tree ha-1,1-3 trees ha-1,or ＞3 trees ha-1.The cavity tree density was used as a measure of availability of cavities for secondary cavity nesting birds.

Fig.1 Study area used for field sampling.In each site,we established 10(black points)to 20 plots(blue points),reaching a total of 50 plots of 0.25 ha

In every 100-ha sampling site,during four consecutive breeding seasons from August to February(2014 to 2018),we intensively searched for cavities used by secondary cavity-nesting birds along 4-6 transects per site,each 1-1.5 km long and 20-50 m wide. The cavities were inspected to determine whether they were used for breeding or roosting.A cavity was deemed‘‘used’’when it contained adult birds,feathers,eggs or chicks.When the occupying bird species was unidentifiable,the cavity use was recorded as‘‘unknown species’’.The same variables of cavity tree species and cavity characteristics in plots were measured for all used cavities.Secondary cavity-nesting bird species were categorized according to their body weight:small(＜100 g),medium(101-500 g),and large(＞500 g).Since this study was focused on secondary cavity nesters, primary-cavity-nesting birds (mainly woodpeckers)were considered as cavity providers for the secondary nesters.

Tree and cavity characteristics were standardized,and cluster analysis was applied to form tree groups using Ward’s method and the Mahalanobis distance,which allows the use of correlated variables(see Spearman test[R2]in Supplementary material,Table S1).A multivariate analysis of variance(MANOVA and Hotelling post hoc test)was performed to test for statistical differences between clusters(i.e.,tree groups).We used canonical principal coordinate analysis(CAP)using a Mahalanobis distance to determine which physiognomic tree variables most influenced the tree group differentiation in terms of bird use(Casanoves et al.2011a).

For each tree group,we calculated the number of cavity tree species and the number of cavity trees used by secondary cavity-nesting birds.The differences between groups of trees used by secondary cavity nesters and tree availability were assessed by use of the Wilcoxon test.All analyses were performed using INFOSTAT(Di Rienzo et al.2008),except PAST 3.2 was used for the CAP,and the clusters were plotted with FDiversity(Casanoves et al.2011b).

Results

Tree characteristics used by secondary cavity nesting bird species

A total of 187 cavity trees of 23 tree species were recorded in all plots with 14.9±7.5 cavity trees ha-1.Of the complete set of cavities,93% were generated either by decomposition and/or wood decay,and the remaining 7% were excavated by woodpeckers.In total,58 cavities were used by 15 identified bird species and 28 cavities were used by unidentified bird species.Ninety percent of the nests and roosts were found in cavities formed by decomposition.

Nests and roots were mainly found in trees of the species C.multiflorum(40% ),standing dead trees(19% ),Amburana cearensis(9% ),A.colubrina(7% ),and A.urundeuva(7% ).Cavity-nest trees had a mean DBH of 0.5±0.3 m,cavity height of 9.7±3.6 m,tree height of 18.0±4.6 m,and cavity depth of 0.3±0.1 m.

Tree groups

Four tree groups were designated by cluster analysis.Group 3 had eight tree species and included C.multiflorum and P.rhamnoides,which had the greatest cavity tree density(availability).Group 1 had seven tree species and included snags.Groups 2 and 4 each included four tree species(Fig.2).Tree groups differed significantly in terms of physiognomic characteristics (F=26.76; p ＜0.01).The first two axes(CAP 1 and CAP 2)of the CAP explained 59.2% of the total variability in tree physiognomic features.Axis CAP1 accounted for 48.7% of the total variability.DBH and the cavity entrance area revealed separation among the four tree groups(eigenvector=0.61 and 0.44,respectively)(Fig.3).

When we compared the measured characteristics between the trees with available cavities and those with cavities used by birds,we found a significant difference in the height of both the trees and the cavities in groups 1 and 3.In group 1,the available trees were significantly lower than those used by birds (8.22 m±3.58 vs.14.18 m±5.26;p=0.0015),and the cavities were at a lower height (4.34 m±2.83 vs. 9.15 m±3.44;p=0.0001),while in group 3 the cavities used by birds were significantly higher(7.77±3.59 vs.9.60±3.57;p ＜0.0018).No significant differences were found in the rest of the measured characteristics,and group 4 did not show significant differences between trees with available cavities and those used by birds(Table 1).

Most of the nests and roosts(83% )used by large birds were found in tree group 4.Nests and roosts of mediumsized birds(80% ),small-sized birds(50% ),and unknown bird species(89% )were found in tree group 3.Tree group 1 had 14% of the nests and roosts of secondary-cavity nesters small,medium,large and unknown bird species.Tree group 2 was not used by secondary cavity-nesting birds(Table 2).

Discussion

Cavity tree groups were not equally used by secondary cavity-nesting birds.No nests were found in tree group 2,suggesting that this tree group is not key for avian secondary-cavity nesters.Tree groups 1 and 4 had a low number of nests(14% and 15% of the total,respectively).Tree group 3 was the most important for secondary cavitynesting birds(71% of total nests),particularly for small-and medium-sized birds.The role that tree group 3 could have for the complete guild of secondary-cavity nesters(regardless of bird body size)could be greater than estimated here given that 50% of the nests were of‘unknown’’bird species and could include bird species of any size.Tree group 3 also included tree species of high economic value,including M.peruiferum,A.colubrina,C.multiflorum,and Handroanthus impetiginosus(Minetti et al.2009).The combination of abundant cavity-nesting birds and many high-value tree species suggests that a balance is needed to mitigate the potential conflict between biodiversity conservation and logging(Imbeau et al.2000;Gibbons et al.2002).For example,cutting M.peruiferum-one of the most economically valuable tree species in the piedmont forest-could be compensated for by retaining more trees of C.multiflorum,which has lower economic value but is of importance for secondary cavity-nesting birds.A similar biodiversity offset has been proposed for other forests as a management strategy(Le Roux et al.2015).

Fig.2 Cavity-tree groups in the piedmont forest of northwestern Argentina determined in the cluster analysis(Ward method and Mahalanobis distance)in 50 plots of 0.25 ha.Numbers and colors differentiate tree groups.Tree species used by secondary cavity-nesting birds are in bold.The box size indicates cavitytree density(availability):small ≤1 tree/ha,medium=1-3 trees/ha,large ≥3 trees/ha

Fig.3 The first two canonical principal coordinates(CAP)displaying the groups and the number of tree species per group(group 1=7 species;Group 2=4 species;group 3=8 species;group 4=4 species).The CAP was performed based on the cluster analysis,and the green lines represent the response variables

While tree group 4 had few nests of secondary cavity nesters,it was particularly important for large birds such as Coragyps atratus,Tyto alba,and Mycrastur sp.Tree group 4 also included two highly valuable tree species,viz.Cedrela balansae and A.cearensis.Selective logging in the piedmont forest has been mainly centered on extraction of C.balansae with the result that stocks have been greatly reduced and large trees are uncommon(Malizia et al.2006;Blundo and Malizia 2009).A scarcity of suitable cavities for large birds in logged sites is to be expected(i.e.,＞73 cm tree DBH and ＞21 m tree height). Amburana cearensis is an endangered tree species that occurs at low population density(Politi et al.2015)and is unlikely to provide as many suitable cavities for large birds as C.balansae does.Management in favor of C.balansae might be critical for large secondary cavity nesters.

Table 1 Mean values for cavity-tree variables in plots and number of cavity trees used by secondary-cavity-nesting birds within each tree group(see Fig.1)

Table 2 Number of nests of secondary cavity-nesting birds found in each cavity tree group(see Fig.1)in the Piedmont forest of northwestern Argentina

Tree group 2 included Cordia trichotoma,a highly valuable tree species(Malizia et al.2009)that was not used by secondary cavity-nesting birds.Logging of this highvalue timber species is unlikely to impact nesting by secondary cavity-nesting birds.Nevertheless,because of its endangered tree species status, other sustainable management guidelines are required(e.g.,regeneration and seed trees)to log this timber species(Lindenmayer et al.2012).

Finally,tree group 1 included no tree species of high economic value but included all snags.In logged sites,a decrease in the density of snags is to be expected due to logging and would reduce the number of cavities(Politi et al.2010).

Cavity and tree characteristics used for grouping tree species were highly consistent with the characteristics used by cavity-nesting birds. Therefore, our tree species grouping was based on biological support.Although not all of the tree species within a group were used by birds,they shared similar characteristics, making them potentially usable.Such trees could be important in logged forests for the replacement of trees removed by logging(Ruggera et al.2016).This potential replacement role of some tree species as alternative cavity trees for secondary cavitynesting birds should be assessed in logged forests.This study should be considered as a starting point to delineate management guidelines;however,some caveats should be noted.First,we only analyzed the value of tree species for secondary cavity-nesting birds,not for other guilds,components of biodiversity,or ecological processes.Second,cavity trees are a dynamic resource,and consideration should be given not only to trees that harbor cavities,but also to those trees that have a greater probability to develop cavities(Politi et al.2010).In the third place,only a subset of cavities(those ＜16.5 m high and with ＞5 cm diameter entrance)was examined,and some use by secondarycavity-nesting birds could have been overlooked as noted by Ruggera et al.(2016).Finally,it is necessary to address aspects in addition to the biodiversity conservation to achieve sustainable forest management(Lindenmayer et al.2000;Gadow et al.2012).

Conclusion

Reducing the complexity of the piedmont forest to less than 1/5(from 23 cavity tree species to four tree groups)might ease the development of management guidelines.This approach has not been broadly used in other forests(e.g.,Dennis and Westcott 2006; Bocanegra-González et al.2015)and is novel for subtropical forests under unsustainable logging.In spite of that,we deem it necessary to prevent population and species extinction(Hansen et al.2013).Retaining tree species of tree groups 3 and 4(such as C.multiflorum and C.balansae ＞40 cm tree DBH)can provide adequate sites for the use of secondary cavitynesting birds.Even if this kind of recommendation at the level of particular groups of tree species is crucial,strategies also need to include designating sectors or patches for no extraction/logging(Politi et al.2009).These recommendations might be easier to implement by foresters than the protection of long lists of bird or tree species(Lindenmayer and Likens 2010),and more tree species would be retained.This way,we would protect both the key tree species and those potentially usable by birds. Consequently,foresters would have two simple ways to mitigate logging impact on cavity-nesting birds.Currently,most of the harvested trees in the piedmont have a DBH in the intermediate range(between 30 and 60 cm)(Politi et al.2010),the size used by cavity-nesting birds.Thus,we consider that these two management guidelines are important for these forests, especially if we take into account their present degree of endangerment.

AcknowledgementsWe thank the assistants that contributed with the fieldwork.We thank Cecilia Garcia for assistance with writing and producing the manuscript.We also thank two anonymous reviewers for their valuable contributions to previous versions of this work.