Standardised inventories of spiders (Arachnida, Araneae) on touristic trails of the native forests of the Azores (Portugal)

Abstract Background The sharp increase in tourist visitation of the Azores Archipelago from 2015 onwards raised concerns about the impacts of recreational tourism on native habitats. In response, a project was financed by the Azorean Government to investigate the drivers of biodiversity erosion associated with recreational tourism. Here, we present the data on spider biodiversity found on trails located within the native Azorean forests as they are home to several endemic species of great conservation value. We applied an optimised and standardised sampling protocol (COBRA) in twenty-three plots located in five trails on Terceira and São Miguel Islands and assessed diversity and abundance of spider species at different distances from the trail head and the trail itself. New information Of the 45 species (12435 specimens) collected, 13 were endemic to the Azores (9690 specimens), 10 native non-endemic (2047 specimens) and 22 introduced (698 specimens). This database will be the baseline of a long-term monitoring project for the assessment of touristic impacts on native forest trails. This methodology can also be used on other habitats and biogeograhical regions.


Introduction
In the Azores, as in many other temperate, semi-tropical and tropical islands, historical patterns of habitat loss have typically resulted in lowland clearance, meaning that the last remnants of the pre-human pristine forest that covered the major parts of oceanic islands are in the mountain areas (Gaspar et al. 2011). The communities of these mountain forests are of critical importance for the protection of current island biodiversity since they are home to many Azorean endemic species , Borges et al. 2018, Malumbres-Olarte et al. 2019 and provide a variety of ecosystem services (e.g. water storage, erosion control, pollination, pest-control, food supply, recreation and tourism), contributing to the local economy and welfare ).
The recent increase in recreational tourist activities in native habitats of the Azores (SREA 2018) raises concerns about the use of trails being a threat to the already imperilled native forest biodiversity. Hiking trails in particular have been found to be promoting the spread of invasive plants ( Barros and Pickering 2014), which may cause adverse cascading effects on arthropods.
The spider communities of the Azores are exceptionally well known due to ongoing inventorying and monitoring projects carried out since 1999 (Borges et al. 2016, Malumbres-Olarte et al. 2019. The protocol used in NETBIOME ISLANDBIODIV and in this project is part of a long term monitoring proposal for oceanic islands (Borges et al. 2018).

General description
Purpose: We aimed to characterise the richness and abundance of spiders in areas surrounding trails in native Azorean forest and to assess if the distance to the head of hiking trails or to the trail itself explains shifts in spider community composition, compared with areas undisturbed by tourists.

Study area description:
We focused on the Azorean forests of Terceira and São Miguel Islands, as they have pedestrian trails going through native forests with a relevant level of visitation (Fig. 1). Terceira Island (area: 400.6 km²; elevation: 1021.14 m) and São Miguel Island (area: 744.6 km²; elevation: 1103 m) are two of the nine islands from the Azores Archipelago. The climate in the Azores is temperate oceanic, with regular and abundant rainfall, with high levels of relative humidity and persistent winds, mainly during winter and autumn seasons. Terceira Island is known for the presence of some very important pristine areas at high elevation (Gaspar et al. 2011).
Funding: This research was supported by the Rui Carvalho Ph.D. DRCT scholarship from the Azores Government (M3.1.a/F/135/2015). Data was obtained mostly during the Ph.D. DRCT scholarship, but some samples ("Control 250", see below) are from a previous project (ERA-Net NetBiome research framework, financed through Portuguese FCT-NETBIOME ISLANDBIODIV grant 0003/2011).

Sampling methods
Study extent: We selected six 50 x 50 m sampling sites in native forest patches along the studied trails, at increasing distances from the trail head: 0 m, 50 m and 250 m. Another sampling site, termed Max, was set independently from distance -it was located where the forest adjacent to the trail was most pristine. Finally, two control sites were selected at 50 m and 250 m from the closest trail point (Table 1). This setup is repeated at each trail. In Terceira, the Control 250 data were retrieved from NETBIOME-ISLANDBIODIV samples from 2012 (see Malumbres-Olarte et al. 2019 Sampling description: The inventory COBRA (Conservation Oriented Biodiversity Rapid Assessment) protocol (Cardoso 2009) was used at the most pristine area in the studied fragment, firstly to assess whether completeness is sufficient to use the less time-intensive protocols; secondly, in order to be used as alpha and beta diversity baselines (Borges et al. 2018). It is composed of four hours of aerial search, four hours of tree beating, four hours of vegetation sweeping and pitfall sampling using 48 traps. The traps containing propylene glycol were active for 15 days and, during sample collection, they were arranged in groups of four to make 12 sample units. For the remaining sampling areas of each trail, the much less intensive monitoring COBRA protocol was used. It is composed of four hours of aerial search and two hours of beating trees using a drop cloth (see Borges et al. 2018 for details). The COBRA protocol has been proposed as part of standard inventorying and monitoring programmes targeting spiders and beetles and has been used on island and continental ecosystems, from subarctic regions to the tropics (Cardoso 2009   Description: The following data table includes all the records for which a taxonomic identification of the species was possible. The dataset submitted to GBIF (Global Biodiversity Information Facility) is structured as a sample event dataset, with two tables: event (as core) and occurrences. The data in this sampling event resource have been published as a Darwin Core Archive (DwCA), which is a standardised format for sharing biodiversity data as a set of one or more data tables. The core data file contains 194 records (eventID) and the occurrences file 1290 records (occurrenceID). This IPT (integrated publishing toolkit) archives the data and thus serves as the data repository. The data and resource metadata are available for download from Carvalho et al. (2021).

Results
We collected a total of 12435 specimens belonging to 45 species of spiders. A total of 13 species are endemic to the Azores Archipelago (9690 specimens), 10 are native nonendemic (2047 specimens) and 22 are introduced (698 specimens) (   This database will be used in future studies where the variation of the spider communites amongst the various sites will be tested against variables that are known to be relevant for understanding the impact of touristic activities, such as the distance to the trail head and the distance from the sampling area to the nearest trail point. We will use GLMMs, where the trail identity will be used as random effect and the edge effect will be added as an independent variable in order to avoid spurious results. This will respond to the questions of whether there is a detectable effect of recreational activities on the spiders community structure and what is contributing to this ecological shift.
Contrary to Canary Islands and Madeira, the Azorean Archipelago has not yet experienced continuous high levels of visitation. This sampling was made at the early times of a noticeably higher touristic pressure in the Azores and will allow for future monitoring events to have a comparable baseline and better isolate the touristic factors from others, thus improving the management outlook on tourism's ecological effects on spider communities.