New records and detailed distribution and abundance of selected arthropod species collected between 1999 and 2011 in Azorean native forests

Abstract Background In this contribution we present detailed distribution and abundance data for arthropod species identified during the BALA – Biodiversity of Arthropods from the Laurisilva of the Azores (1999-2004) and BALA2 projects (2010-2011) from 18 native forest fragments in seven of the nine Azorean islands (all excluding Graciosa and Corvo islands, which have no native forest left). New information Of the total 286 species identified, 81% were captured between 1999 and 2000, a period during which only 39% of all the samples were collected. On average, arthropod richness for each island increased by 10% during the time frame of these projects. The classes Arachnida, Chilopoda and Diplopoda represent the most remarkable cases of new island records, with more than 30% of the records being novelties. This study stresses the need to expand the approaches applied in these projects to other habitats in the Azores, and more importantly to other less surveyed taxonomic groups (e.g. Diptera and Hymenoptera). These steps are fundamental for getting a more accurate assessment of biodiversity in the archipelago.


Introduction
In 1999 a group of researchers from the University of the Azores and the University of Lisbon started a long-term (1999)(2000)(2001)(2002)(2003)(2004) standardized sampling program to inventory the arthropod biodiversity in native forest remnants of the Azores -the BALA I project -B iodiversity of Arthropods from the Laurisilva of the Azores (Borges et al. 2000, Borges et al. 2005a, Ribeiro et al. 2005, Gaspar et al. 2008. More recently, this project was extended by researchers from the Universities of the Azores, Athens and Oxford, by surveying part of the same native forest plots almost 10 years later - BALA II project (2010.
Eight years of standardized survey of the native forest in seven of the nine Azorean islands resulted in a major improvement on the knowledge of the Azorean arthropod fauna, in particular concerning Araneae, Opiliones, Pseudoscorpionida, Diplopoda, Chilopoda and Insecta (excluding Collembola, Diptera and Hymenoptera). As a consequence, several new endemic taxa were described for the archipelago (e.g. Blas and Borges 1999, Ribes and Borges 2001, Platia and Borges 2002, Quartau and Borges 2003, Borges et al. 2004, Borges and Wunderlich 2008, Crespo et al. 2013, Crespo et al. 2014 or are in the process of being described (Borges et al. 2016 in press). In fact, after examining the shape and characteristics of discovery curves, Lobo and  clearly show that it is very likely that many new species of arthropods remain to be discovered in the Azores particularly for less studied groups in this archipelago such as Diptera and Hymenoptera. Besides purely faunistic results, the BALA data was also used to evaluate abundance, spatial variance and occupancy of arthropods (Gaston et al. 2006, the effects of disturbance and biotic integrity of the native forests on arthropod assemblages (Cardoso et al. 2007, Florencio et al. 2013, Florencio et al. 2015, the extinction debt of Azorean forest specialist species (Triantis et al. 2010) and the performance of species richness estimators (Hortal et al. 2006). Moreover, such data allowed the ranking of conservation priorities for the fauna and flora of the Azores (e.g. Borges et al. 2005a, Martín et al. 2010) and allowed the estimation of extinction debt in Azores (Terzopoulou et al. 2015, Triantis et al. 2010.
During this period, two complete checklists of Azorean arthropod fauna were produced (Borges et al. 2005b, which included the distribution of each species per island. In this paper we compile and synthesize the faunistic results of both BALA projects, highlighting novel distribution records and presenting not only detailed distribution but also abundance data for each species, adding taxonomical and biogeographical information whenever possible. Finally, we provide a general and updated overview on the diversity of the Azorean arthropods.

Area of study: The Azores
The remote Azores archipelago extends for 615 km in the North Atlantic Ocean (37-40 °N, 25-31 °W), 1584 km to the east (southern Europe) and 2150 km to the west (northern America) of the nearest mainland. It comprises nine main islands and some small islets, all of volcanic origin, and is located at the triple junction of the Eurasian, African and American tectonic plates. The nine islands are divided into three groups: the western group (Corvo and Flores isls.), the central group (Faial, Pico, Graciosa, São Jorge and Terceira isls.), and the eastern group (São Miguel and Santa Maria isls) (Fig. 1). The climate is temperate and oceanic, strongly influenced by the ocean and island topography, which together produce high relative atmospheric humidity, above 95% on average on native forests.

Sampling protocol
Eighteen native forest fragments distributed across seven of the nine islands were sampled (Table 1; see also Gaspar et al. 2008). Graciosa and Corvo islands were excluded as they no longer present native forest. Human settlement in the Azores lead to considerable native forest destruction which has left the entire archipelago with little over 2% of the original forest cover. During the summer (June to September) 150 m long and 5 m wide transects were set up in 100 sites from 1999 to 2004 (BALA I: 18 native forest fragments) Location of the Azores and of native forest fragments in the archipelago. Codes for forest fragments as in Table 1. and some were sampled twice in that period totalling 123 samples; about 29 of those sites were resampled from 2010 to 2011 using the same protocol (BALA II project; 15 native forest fragments). Along each transect, arthropods from the soil (mainly epigean) and herbaceous vegetation were surveyed with pitfall traps, while arthropods from woody plants were sampled using a beating tray. Pitfall traps consisted of plastic cups with 4.2 cm diameter and 7.8 cm height. Thirty pitfall traps were set up per transect. Half of the traps were filled with a non-attractive ethylene glycol preservative solution (antifreeze solution), and the remaining with a general attractive solution, a modified version of Turquin (Turquin 1973) prepared mainly with dark beer and preservative agents. A few drops of dishwashing liquid were added to both solutions to reduce surface tension. Traps were sunk in the soil (cup rim at surface level) every 5 m along the transects, those filled with Turquin alternating with traps containing antifreeze solution. Traps were protected from rain using a plastic plate, placed about 5 cm above surface level and fixed to the ground by two pieces of wire. Accidental collection of small vertebrates and damage by rodents was prevented using a piece of plastic mesh placed on top of the trap and fixed to the ground by pieces of wire. The traps remained active in the field for two weeks.  Table 1.
Main characteristics of the Azorean islands (bold) and native forest fragments sampled from 1999 to 2011, including area (hectares), highest point (altitude in metres), distance to the nearest island/ fragment (isolation in kilometres) and the oldest geological age of emerged substrate (million years BP) (adapted from Gaspar et al. 2008). Canopy sampling was conducted during the trapping period, when the vegetation was dry. A 5 m wide square was established every 15 m (total of 10 squares per transect). Two woody plant specimens of the most abundant species (up to three species when available) were sampled in each square. For each selected plant, a branch was chosen at random and a beating tray placed beneath. The tray consisted of a 1 m wide and 60 cm deep cloth inverted pyramid, with a plastic bag at the vertex. Five beatings were made using a stick for each plant individual sampled.
The arthropod taxa considered in this study were selected based on the availability of expert taxonomists and ability to readily separate them by morphological criteria. All Araneae, Opiliones, Pseudoscorpionida, Diplopoda, Chilopoda and Insecta (excluding Collembola, Diptera and Hymenoptera) were assigned to morphospecies through comparison with a reference collection. Various taxonomists (PAVB, ARMS, LC, PC, HE, FI, VM, MTP, JR, AB, ABS, RzS, VV, JW, JAQ, and see also Acknowledgments) checked the assignment to morphospecies, performed species identifications and supplied additional ecological information. The taxonomic nomenclature follows the most recent list of Azorean arthropods ).
All specimens are deposited in the Entomological Collection Dalberto Teixeira Pombo at the University of the Azores (Portugal), under the curation of Paulo A. V. Borges (pborges @uac.pt).
In this contribution we list the 286 species for which we obtained an identification. The new records for each island are marked with *. For this list two families of Coleoptera were not considered since they will be presented elsewhere, Staphylinidae (Borges et al. in prep.) and Zopheridae (Borges et al. 2016). For detailed maps on the distribution of these species in Azores consult the Azores Bioportal.
All specimens were assigned a SITE CODE composed of several letters and numbers that read as follows (see Suppl. material 1 for complete data For the geographical location of transects within reserves (UTM coordinates) see Suppl. material 3.
Accumulation curves were obtained using the software "Species Diversity and Richness" V.4.

Analysis
Azorean Arthropod biodiversity -towards a more complete knowledge The ultimate goal of biodiversity assessments is documenting all species inhabiting a region. However, this has often proven impossible to achieve given the unfeasibility of collecting every single species that exists in a study area. This study focuses on the terrestrial arthropod diversity of the Azores and encompasses most orders of the phylum Arthropoda. A pool of a total of 1215 species and subspecies was surveyed, representing 53% of the whole arthropod fauna known from the Azores ). By deliberately not surveying Crustacea, Acari, Collembola, Diptera and Hymenoptera, we excluded 47% of the archipelago's species pool. Yet, this study added 10 endemic and at least 16 other species, mostly exotics, to the known Azorean arthropod fauna. More will be added soon after the on-going revision of Staphylinidae (in prep.) and Zopheridae (Borges et al. 2016, in press). Overall, at least 26 species that occur in native forests were added to the Azorean arthropod fauna list. The new 346 taxonomic records provided by this study (see Suppl. material 4 for the complete list of new records per island) represent on average an increase in species number of about 10% for each studied island (Table 2). However, the increment for São Jorge island was about 22%, while for São Miguel this represented only 3% (Table 2). 164 species were found in new islands, with an average of two islands per species. For 82 of those species only one new island was added to their known distribution contrasting with 27 species for which four or more islands were added (Fig. 2).
Notably, nine out of the 27 species with more than three island added to their previous distribution belong to Arachnida. In fact, arachnids but also millipedes and centipedes experienced a large proportion of new records (more than 30%) (see Table 3).  Table 3.
Total species and subspecies records for the Azores, new species and subspecies records during this study and increment for the most speciose classes and orders. Values for all islands are added, so richness may be up to 7 times higher than the archipelago's richness (as 7 islands were surveyed). (*)The Coleoptera families Staphylinidae and Zopheridae were not considered (see text).
The number of species identified for each of the 18 native forest fragments surveyed is shown in Fig. 3. The fragment with the highest species diversity is Serra de Santa Bárbara in Terceira island (S = 124), which is also the larger native forest area in the Azores.
Remarkably, one of the smallest fragments, Pico Alto in Santa Maria island, is the second most diverse (S = 121).  Number of species per native forest fragments. Island codes as in Table 1 BALA2 samples only added 4% of species to the previous BALA survey (Fig. 4). Interestingly, 59 samples collected in the first two years of survey (1999 and 2000) provided about 81% of the total species recorded in this study.

The most abundant species
A total of 163744 individuals were identified as belonging to the 286 species (see Suppl. material 5 for the complete list of abundance per species). The ten most abundant species (Fig. 5) accommodate 56% of the total number of individuals and include mostly indigenous species (endemic or native non-endemic). The single introduced species is the millipede Ommatoiulus moreletii (Fig. 6). With exception of the millipede Ommatoiulus moreletii, the centipede Lithobius pilicornis pilicornis and the opilion Leiobunum blackwalli (Fig. 7) that are mostly soil epigean species, the other seven species live preferentially in the canopies of Azorean endemic trees. The moth Argyresthia atlanticella (Fig. 8) is particularly common in Juniperus brevifolia and Erica azorica; the spider Savigniorrhipis acoreensis (Fig. 9) is particularly abundant in Juniperus brevifolia, but can also be found in other plants    New records and detailed distribution and abundance of selected arthropod ...  identified seven impediments in invertebrate conservation. Three of them are particularly relevant for our study: most species are undescribed (the Linnean shortfall), the distribution of described species is mostly unknown (the Wallacean shortfall), and the abundance of species and its variation in space and time are unknown (the Prestonian shortfall). We argue that with the BALA project we were able to contribute to overcome some of these impediments in the Azores. In fact, we show that as a result of the standardized sampling performed in Azorean native forests we were able to: i) decrease the Linnean shortfall, by increasing the number of described Azorean endemics (e.g. Blas and Borges 1999, Ribes and Borges 2001, Platia and Borges 2002, Quartau and Borges 2003, Borges et al. 2004, Borges and Wunderlich 2008, Crespo et al. 2013, Crespo et al. 2014, Borges et al. 2016; ii) decrease the Wallacean shortfall, by increasing the known distribution of many endemic and exotic species in the archipelago (e.g. Borges et al. 2005a, Cardoso et al. 2009, Meijer et al. 2011; and iii) decrease the Prestonian shortfall, by using standardized sampling, which allowed the comparison of species abundances in space and time as many of the same sites were sampled in two different time periods.

Discussion
The increase in the number of islands from where each species is known and the distribution increase for many species within each island shows the importance of regional standardized surveys, which provided a major improvement in the knowledge of the distribution of arthropod species in the native forests of the Azores.
The fact that most diversity was captured during the first two years of the project reflects the importance of sampling a wide geographic range covering all the islands and the BALA2 project) had a lesser impact in increasing our knowledge about biodiversity (Fig. 4).
The future agenda for surveying and monitoring Azorean arthropod biodiversity includes: a) expanding the standardized survey of Azorean arthropods to other habitat types, mostly man-modified, an already on-going task for some of the islands (see e.g. Cardoso et al. 2009, Meijer et al. 2011, Florencio et al. 2013); b) selecting study areas along a comprehensive environmental gradient where an optimal sampling strategy will be applied in order to sample the entire arthropod communities (All Taxa Biodiversity Inventory -ATBI). ATBIs are intensive sampling efforts to identify and record all living species that exist within a given area and simultaneously create a common and standardized biodiversity database (Lawton and Gaston 2001); c) finishing the identification of many morphospecies. Good progress has been made with Staphylinidade (Borges et al. in prep.), but other taxa need further effort to reach proper identification; d) increase sampling and update the current list of Azorean Hymenoptera and Diptera, which is clearly incomplete . The shortage of taxonomists who can adequately identify species (i.e. the so-called Taxonomic Impediment) has prevented advances in the knowledge for many diverse groups in the Azores, including these two. This study advances the knowledge on the unique arthropod biodiversity of the Azores, but at the same time highlights the need for further surveys. We strongly believe that the BALA project will stimulate further research and conservation actions towards the preservation of Azorean biodiversity. Furthermore, we hope that all the taxa yet to be identified will entice taxonomist to join us in the endeavour of cataloguing all terrestrial arthropods of the most remote of the Macaronesian archipelagos, the Azores. The ongoing longterm research projects in Azores and the recent creation of the E-Repository ISLANDLAB will create new opportunities for biodiversity studies in Azores.

Supplementary materials
Suppl. material 1: Appendix 1 -Detailed data on the distribution and abundance of the studied species