Ephedra shrubs facilitate local arthropod communities in the Andean Puna: Implications for conservation and habitat restoration

A�������. In high-altitude deserts, the vegetation is spatially structured as a mosaic of patches with vegetation and interpatches of bare soil. Shrub patches create microhabitats that facilitate the persistence of other organisms by ameliorating climate extremes and increasing soil nutrients and moisture. Although this facilitative effect has been studied mainly on shrub understorey plants, the positive influence of desert shrubs may extend to the local arthropod community. In this study, we examined the hypothesis that plant facilitation by the desert shrub species Ephedra multiflora and E. breana positively influences the epigeal arthropod communities of the Andean Puna. We found that arthropod abundance, richness and diversity were all higher on Ephedra shrub microsites relative to paired open microsites. The harsh environmental conditions prevailing in the Puna ecosystems and the protection and resources provided by plants could jointly explain the positive influence of Ephedra shrubs over the arthropod community. The growth of tourism and the boom for lithium, as well as the increasing drought expected for the region under climate change scenarios, strengthens the importance of conserving shrub cover in these ecosystems to sustain biodiversity and ecosystem functioning. Our results are also relevant for restoration projects that consider the whole biological community, where using Ephedra as foundation shrubs to support desert endemics could be a strategy for land restoration.

Arthropods are the dominant component of biodiversity in drylands (Whitford 2000), and frequently used as bioindicators of environmental changes and ecosystem health (Kremen et al. 1993;Andersen and Majer 2004;Carvalho et al. 2020;Chowdhury et al. 2023).They sustain a variety of interactions and trophic connections, and play important roles as decomposers, herbivores, granivores, pollinators and predators, controlling the nutrient cycling and the energy flow through the different levels in the food chain (Ayal 2007;Noriega et al. 2018).Besides particular physiological adaptations, their persistence in deserts is conditioned to the existence of habitat structures that provide food, shelter or nesting sites.
Desert shrubs fulfil these habitat requirements, as they create milder microhabitats for arthropod communities by moderating solar radiation, changing soil moisture and temperature, and regulating extreme climatic factors (Brooker et al. 2008;Wright et al. 2021).They also provide key resources such as food (e.g., pollen, nectar, leaves, wood, prey/host), shelter, oviposition and mating places, affecting the longevity, reproduction and dispersion of arthropods (Gardarin et al. 2018).Arthropods, in turn, maintain a variety of interactions and trophic connections with shrubs, including those mutualistic such as pollination, seed dispersal, and herbivore predation (Scherber et al. 2010).The positive effect of shrub microhabitats on arthropod communities has been assessed in deserts around the world (e.g., Mazía et al. 2006;Li et al. 2013;Ruttan et al. 2016), and was found consistent with an 'arthropod island' effect when the abundance and diversity of arthropods is enhanced in shrub vegetation patches compared to bare soils (Sanchez and Parmenter 2002;Meloni and Martínez 2021;Braun et al. 2021).Therefore, shrubs have important implications for maintaining the biodiversity of desert arthropods, enhancing in turn the stability and complexity of arid ecosystems.
The Puna ('high and cold land' in Quechua) is part of the Andean plateau, the second largest in the world after the Tibet.It is located in the central Andes, and has an average altitude of 3500 m a. s. l. and peaks up to 6000 m a. s. l.A distinctive feature of the Puna is its arid climate, with low temperatures throughout the year, a large daily temperature range, and low rainfall.To this must be added a low air density (causing 'apunamiento'), high radiation and strong winds (Grau et al. 2019).It is also a biodiversity hotspot with high levels of endemism (Myers et al. 2000), and was declared by the UNESCO as one of the Global 200 priority conservation areas (Olson and Dinerstein 2002).Because life for humans in Puna is uneasy, for a long time their lands have been inhabited by few people, mainly descendents from indigenous populations, with a survival economy based on agriculture and camelids.During the last decades, the increasing tourism (so called 'adventure tourism') and, more importantly, the boom for lithium, have disrupted these remote landscapes (Izquierdo et al. 2015(Izquierdo et al. , 2018)).To this must be added that future climate change scenarios identify high-elevation ecosystems among the most vulnerable (Beniston et al. 1997), while climate reconstructions show decreasing trends of precipitation for the area (Morales et al. 2015).This context of vulnerability highlights the urgency to understand the key factors that sustain the biodiversity in the Puna, in order to establish conservation recommendations.
In this study, we evaluate the facilitative effect of two species of Ephedra shrubs (E.multiflora and E. breana) on the epigeal arthropods of the desert Puna.Ephedra (Gymnospermae: Gnetales: Ephedraceae) are plants of ancient lineage distributed in semiarid and arid habitats worldwide (Ickert-Bond and Renner 2016), known for some species that contain ephedrine alkaloids (not found as far in the South American species, Caveney et al. 2001).As gymnosperms, Ephedra are perennial,

₉₄₀ NI Y��� �� ��
Ecología Austral 33:938-949 drought tolerant long-lived plants, and some species are used for ecological restoration and desertification control, mainly in the deserts of North America and Asia (Ackerman 1979;Derbel et al. 2010;Lortie et al. 2018;He et al. 2021).The Ephedra species of this study are common and dominant in the area, and found in plain steppes between 3300 and 3500 m a. s. l., where hydric stress and harsh climate impose particular adaptations for survival.
As a hypothesis, we establish that the epigeal arthropod community reflects the effects of plant facilitation.Therefore, we expect a higher abundance, species richness and diversity of arthropods in the Ephedra shrub microsites relative to paired non canopied (open) microsites.

Study area
We conducted this study in the Desert Puna of Catamarca Province, northwest Argentina (Carilla et al. 2019).This is a high-elevation cold desert (above 3000 m a. s. l.), where the dominant vegetation type is the shrubsteppe (Cabrera 1968).Climate is arid, with low temperatures, a large daily temperature range (more than 15 °C difference between day and night), and low rainfall concentrated in summer (<100 mm annually) (Morales et al. 2019).In addition to scarce rainfall, the combination of elevated evapotranspiration, strong winds and high solar radiation results in a negative water balance throughout the year (Izquierdo et al. 2018).
We selected two sampling sites based on the presence of each Ephedra species: 1) the E. multiflora site, near the town El Peñón (26°29' S -67°15' W; 3400 m a. s. l.), and 2) the E. breana site, located 50 km apart to the east in the Reserva de la Biosfera Laguna Blanca, a protected natural area near the town of Laguna Blanca (26°43' S -66°55' W; 3300 m a. s. l.) (Figure 1).Rainfall decreases in a Northeast-Southwest direction and, while both sites lie at similar elevations (3300-3400 m a. s. l.), they are separated by a mountain range with peaks as high as 6000 m a. s. l. (Cerro Laguna Blanca).
Preliminary site-level environmental characterization (climate, aridity index and soil properties) indicates that the E. multiflora site have more stressful environmental conditions than the E. breana site (Supplementary Material-Table S1).Indeed, all climatic parameters show a decrease in precipitation and temperature, as well as higher aridity for the E. multiflora site.Both sites have sandy soils with ~90% of sand, but in the E. multiflora site they have lower field capacity and organic matter.The vegetation of both sites is also quite different, since at the E. multiflora site the only other shrub found other than E. multiflora is Aloysia deserticola, while at the E. breana site, the plant community is more diverse and includes other shrub species such as Fabiana densa, Junellia seriphioides and Adesmia horrida.

Plant species
Ephedra shrubs are dioecious gymnosperms that produce seed cones, have small scale-like leaves, and photosynthetic stems.The two species of this study are easily differentiated because E. multiflora produce dry winged cone bracts and the seeds are wind-dispersed, while E. breana have succulent, brightly red coloured cone bracts (Hunziker 1995;Ickert-Bond and Renner 2016).During the pollination stage, the cones of female plants produce pollination drops rich in sugar that capture airborne pollen released from male plants (Gelbart and von Aderkas 2002).Both are conspicuous plants with a dense canopy (at maturity, they can reach more than 1 m in diameter and height).Besides their morphological differences, the two species do not coexist in the same plant communities.The actual distribution of E. multiflora, based on the collection sites found in the herbarium data, appears to be restricted to extreme arid habitats where very few other species grow.This suggests that E. multiflora occupies a narrow environmental niche, resulting in a relatively patchy distribution in the desert Puna.Ephedra breana, in turn, is widely distributed in the shrubland steppes of the Andean Puna from Ecuador to northwest Argentina (Hunziker 1995).We note that here we follow the Ephedra taxonomical classification of Hunziker (1995), where E. breana is considered a different species from E. chilensis.

Experimental setup
At each site, we established 30 replicate plots, each including a sampling Ephedra shrub.The distance between the plots was >15 m to ensure sampling independence.For each sampling plot, there were two paired sampling microsites: the shrub microsite, which includes both the canopy and the H���-A����� ������ ���������� ��������� ����������� ₉₄₁ ground shrub understory, and the open microsite, randomly selected open spaces without vegetation placed at least 2 m away.Shrubs selected for sampling were of similar size and well-developed, and included both female (N=20) and male plants (N=10).We measured the dimensions of each sampling shrub (width at the widest point and height of the canopy) to account for a potential influence of the plant size on the arthropod assemblage.

Arthropod sampling
Samplings were carried out during late spring (November 2022), which corresponds to the pollination stage of both Ephedra species.At each site, we collected epigeal arthropods at each pair of microsites (shrub and open) using two sampling methods: pitfall traps and suction.While pitfall traps are expected to collect mainly ground-dwelling or nearground arthropods, the suction method is more effective for flying insects or those associated to the shrub canopy.Pitfall traps consisted of plastic cups (8 cm upper diameter x 5.5 cm tall) filled with a 50% propylene glycol and water mixture; traps were active for three days.Suction sampling was done using a leaf blower set to suction mode (Lüsqtoff LSA-26), with a gauze-bag inserted into the suction tube.Each suction sample was defined as the suction during one minute over an area of one square metre placed on the ground, and up to 1.5 m height.Pitfall traps and suction methods were applied at both microsites (shrub and open), totalizing 60 samples for each method in 30 pairs of microsites at each site.Suction sampling at the shrub microsites was performed on the shrub canopy; at the open microsites suction was applied on the same area and height above ground but without vegetation.Collected samples were placed in containers with 70% ethanol and transported to the lab, where the arthropods were counted and sorted by order, family, genus, and species/morphospecies using available keys (Grissell and Schauff 1990;Acosta and Maury 1998;Fernández and Sharkey 2006;Brown et al. 2009;Fletcher 2009;Lawrence et al. 2010;Cigliano et al. 2023).Specimens that were identified up to family were sorted into putative species based on morphology, or morphospecies.Larval stages and hemipteran nymphs, as well as Acari and Collembola were excluded from the analyses.Collected specimens were deposited in the CRILAR Entomological Collection (CRILAR-En-Ar).

Data analyses
At each site, the arthropod community sampled was characterized by: 1) abundance (total number of arthropods); 2) species richness (total number of taxonomical groups), and 3) diversity, expressed by the Shannon diversity index.To test whether Ephedra shrubs facilitate arthropod communities, we analysed differences between microsites (shrub and open) by fitting Generalized Linear Mixed Models (GLMM) in R (function glmer.nb,package lme4).We performed three separate models, with 1) arthropod abundance; 2) species richness, and 3) Shannon diversity per sampling unit as response variables.Two analyses were performed: a) using as sampling unit the pooled data from the two sampling methods (pitfall and suction), and b) using only the data from pitfall traps.As predictor factors we used microsite (shrub

₉₄₂ NI Y��� �� ��
Ecología Austral 33: [938][939][940][941][942][943][944][945][946][947][948][949] and open), plant species, plant dimensions (plant height and width), plant sex (male and female), sampling unit (pitfall and suction) and their interactions.We treated plant ID as random effect.We used a negative binomial distribution because the arthropod abundance was represented by discrete counts that were overdispersed (Lindén and Mäntyniemi 2011).To analyse differences between the two sampling methods (pitfall traps and suction), we compared the arthropod abundance, species richness and Shannon diversity by fitting GLMMs for each response variable.The sampling method, plant species and their interaction were used as predictors, and plant ID as a random effect.Because suction sampling in the open microsites captured only one individual arthropod at each site, this analysis was performed only for shrub microsites.The best predictors for each model were selected based on Akaike's information criterion corrected for small samples (AICc) using the dredge function in the MuMIn package.Statistical analyses were performed in R version 4.2.2 (R Core Team 2022).

R������
Overall, we recorded 2490 arthropods belonging to 130 species/morphospecies from 68 families distributed in 11 higher taxa (see Supplementary Material-Table S2 for full species list).The arthropod groups exhibiting the greatest abundance and species richness were dipterans (relative abundance 36.8% and 40 species) and hymenopterans (relative abundance 39.9% and 37 species), followed by arachnids (relative abundance 7.2% and 14 species), and coleopterans (relative abundance 6.4% and 12 species) (Table 1).Seven species of ants contributed to 80.46% of the total abundance within the hymenopterans.
The positive effect on the arthropod communities differed significantly between  2c).

Comparison between sampling methods
The absolute abundances and species richness of the arthropods collected by pitfall trap and suction sampling are shown in Supplementary Material-Table S4.In open microsites, the number of arthropods captured by suction was negligible (just one individual at each site).In shrub microsites, suction sampling contributed to approximately half of the collected arthropods (E.multiflora: 361 of 699 individuals, E. breana: 605 of 1120 individuals).We found no significant differences between the two sampling methods in terms of arthropod abundance (14.96±6.25 vs. 16.98±16.11,P=0.78, pitfall and suction, respectively, means±S.D.) and Shannon diversity index (1.85±0.34 vs. 1.31±0.59,P=0.24), though species richness was higher in pitfall traps (8±2.48 vs. 5.88±3.23,P<0.05).
Suction sampling yielded a number of species associated with the canopy that were captured solely by this collection method (19 species for E. breana and 17 species for E. multiflora).These included mainly dipterans (nine species at each site) and parasitoid wasps (seven species at each site).Certain species were particularly numerous in the suction samples, such as the parasitic wasp Ichneumonoidae sp.(189 individuals) in E. multiflora, and the dipteran Olcella sp. in E. breana (56 individuals).In contrast, all arachnids in the E. multiflora shrub microsites were collected with pitfall traps (scorpions, spiders, and camel spiders), while in E. breana this taxonomic group was dominated by jumping spiders (Salticidae) and was associated with the canopy.Ants were consistently more abundant in pitfall shrub and open), plant width and height (continuous variables), plant sex (male and female) and their interactions.Plant ID was modeled as random effect.Best model for each response variable was selected on AICc values.*P<0.05,**P<0.01,***P<0.001,n.s.: non-significant, P>0.05.All models were run with a negative binomial error distribution.Sampling unit pooled the data from the two sampling methods (pitfall and suction).Tabla 2. Resultados de la selección de modelos lineares generalizados mixtos analizando la respuesta de la abundancia, riqueza de especies e índice de diversidad de Shannon a la especie de Ephedra (E.breana y E. multiflora), micrositio (abierto y arbusto), ancho y alto de la planta (variables continuas), sexo de la planta (macho y hembra) y sus interacciones.La identidad de la planta fue incluida como efecto al azar.El mejor modelo para cada variable respuesta se seleccionó en base a los valores de AICc.*P<0.05,**P<0.01,***P<0.001,n.s.: no significativo P>0.05.Todos los modelos se corrieron con una distribución de error binomial negativa.La unidad de muestreo es la suma de los datos obtenidos con los dos métodos de muestreo (trampas de caída y succión).traps than in the suction samples, with the exception of Brachymyrmex bruchi, which was as numerous in the canopies of E. breana as on the ground.
Regarding differences in the composition of the arthropod communities between sites, 45 species (34.6% of 130 species, E. multiflora and E. breana sites combined) were found exclusively in the E. breana site and 27 (20.7%)exclusively in the E. multiflora site, while 58 (44.6%) were common to the two sites.

D���������
We found that the abundance, species richness and diversity of epigeal arthropods were significantly higher in the Ephedra shrub microsites than in the open spaces between shrubs, thus supporting the hypothesis that the arthropod community reflects the effects of plant facilitation.
Many studies have addressed the effects of shrub microhabitats on arthropod communities and, as in our study, confirmed an 'arthropod island' effect generated by shrub cover on desert arthropods (e.g., Sanchez and Parmenter 2002;Mazía et al. 2006;Liu et al. 2013;Braun et al. 2021).The main underlying mechanism by which the shrub facilitative effect operates is the amelioration of the abiotic environmental conditions, this process resulting in a biased spatial distribution of arthropod individuals and species (Pugnaire et al. 2011).The microclimatic changes produced by vegetation benefit arthropods in both their mobility and foraging behavior, allowing them to thermoregulate better in relation to conditions outside of the vegetation (Molenda et al. 2012).Besides acting as a refuge, shrubs maintain soil moisture, incorporate organic matter into the soil, and provide an important source of food for pollinators, phytophagous, decomposers and predators (e.g., Whitford 2000;Rodríguez-Echeverría and Traveset 2015;Ruttan et al. 2021;Sagi and Hawlena 2021).The diversity of arthropods associated with the two species of Ephedra in our study shows that they support a wide range of arthropod groups that might provide a suite of ecosystem services locally, including those that contribute to the performance of plants.The harsh environmental conditions prevailing in the Puna ecosystems and the protection and resources provided by plants could jointly explain the influence exerted by shrubs over the arthropod community.
Shrub patches and bare soil interpatches can be seen as opposite sides of a continuous spatial system; therefore, it is expected to find arthropods in open spaces as a spillover effect of shrub facilitation (Michalet and Pugnaire 2016).Although the main benefits for arthropods would occur within the patch, the positive influence of shrub canopies could reach regions beyond the patch border, also affecting arthropods in bare soil (Meloni and Martínez 2021).Furthermore, while some arthropods are restricted to zones located under or very close to vegetation patches, other arthropods are relatively motile between microhabitats.Ants, for instance, were the most abundant group in open microsites.Desert ants possess mechanisms that allow them to cope with extreme temperatures, such as thermoregulatory nest architecture, time partitioning of activities to avoid extreme temperatures, and superorganism behaviors (Hölldobler and Wilson 1990;Yela et al. 2020).Therefore, in the open microsites there was an abundant but low diversity group of ants that live in bare soil but search for food in shrubs (such as invertebrate preys and pollination drops).
On the contrary, the dipterans were between five (for E. multiflora) and eleven times (for E. breana) more abundant in shrub microsites than in bare soils, suggesting that this group strongly depends on the habitat provided by Ephedra shrubs.Desert predators, represented mainly by spiders, scorpions, and camel spiders (Solifugae), as well as coleopterans and parasitoid wasps, were also more numerous in shrubs than in the open microsites.In this sense, the suction sampling methodology was proven to be useful in collecting arthropods strongly associated with the canopy that could ₉₄₆ NI Y��� �� �� Ecología Austral 33:938-949 not otherwise be collected by pitfall traps, thus providing a more complete picture of the entire arthropod community associated with Ephedra shrubs.The higher abundance and richness of arthropods in female plants compared to males could be explained by the pollination drops secreted by female cones (and absent in male cones), these representing a valuable food resource attractive to many arthropods (Aranda-Rickert et al. 2021) (Supplementary Material-Figure S1).
Although the two Ephedra species shared many species of the arthropod community, each had a particular species richness and assemblage exclusive to each site.For example, the most abundant parasitoid wasp in E. multiflora (Ichneumonidae sp.) was exclusively associated to the canopies of this species and absent at the E. breana site, suggesting that the local habitat or the identity of the plant species is of particular importance in determining the arthropod assemblage (Schaffers et al. 2008;Tobisch et al. 2023).Our results also show that the arthropod island effect was higher at the E. breana site, as shrub microsites differed from open microsites more strongly in terms of arthropod abundance (2.5 vs. 1.8 enhancement, E. breana vs. E. multiflora) and species richness (2.7 vs. 1.8) compared to the E. multiflora site.A possible explanation is that E. breana, which grows in relatively less stressful conditions and with higher plant cover and diversity, the increase in plant diversity should directly increase arthropod diversity (Borer et al. 2012).On the contrary, drought and stressful abiotic conditions, such as at the E. multiflora site, reduce plant and arthropod diversity, but also favor drought-adapted arthropods, thus facilitating a different arthropod assemblage (Prather et al. 2020).As the turnover of species at local sites within a given ecoregion significantly contributes to the full diversity in that ecoregion, our findings reinforce the importance of maintaining each of the different plant communities of the Desert Puna.
Future land management and restoration strategies for damaged natural habitats should consider the use of foundation shrubs such as Ephedra due to their positive effects on biodiversity, including their impacts on desert arthropods (Filazzola et al. 2019), as well as their tolerance to low water availability and preferences for well-drained soils.Differences in habitat requirements and plant characteristics between the two Ephedra species suggest that E. multiflora might be best suited for dune fixation and as foundation species in areas where environmental conditions are extremely harsh.On the other hand, E. breana is a valuable species for restoration because of its nutritional value as a forage resource for native fauna and livestock (Hunziker 1995); and has edible cones that are consumed by humans and native fauna (mainly birds) (Aranda-Rickert, unpublished).The lack of consumption of E. multiflora stems by native and exotic ungulates probably accounts for some kind of herbivore deterrent present in photosynthetic stems, which merits further study.
Many desert-adapted species of the Ephedra genus are currently used in desert restoration programs, such as E. nevadensis in the Mojave Desert, E. trifurca in the Sonoran Desert, and E. californica in the San Joaquin Desert of California (Ackerman 1979;Derbel et al. 2010;Lortie et al. 2018).In Argentina, E. ochreata, a species inhabiting the Patagonian steppes, has been also pointed out as a potential plant for productive (considering their value as food for livestock) restoration of degraded environments (Rodríguez-Araujo et al. 2019).
In general, all these Ephedra species have high and homogeneous germination without pretreatments, and relatively high rates of establishment, with individuals that can reach an age greater than 50 years (Goldberg and Turner 1986).In California, E. californica is resilient to some herbivores during establishment, being able to recover quickly after significant removals of the shrub canopy (Lortie et al. 2018).As other desert foundation shrubs, Ephedra shrubs avoid growing under other shrub canopies and full-sun conditions promote their establishment rates and growth (Ji et al. 2019).

Final remarks
High-altitude deserts such as Puna are important ecological systems that support some of the highest levels of endemics relative to other ecosystems (Aagesen et al. 2012;Grau et al. 2019).Currently, climate change combined with growing tourism and an exponential growth of prospects and mining concessions for lithium exploitation are the main threats to biodiversity and hydrological function in Puna (Izquierdo et al. 2015).It can be expected that environmental pressures that lead to decreasing vegetation cover and/or connectivity of vegetation patches will result in additional indirect impacts on the arthropod community, amplifying the overall adverse H���-A����� ������ ���������� ��������� ����������� ₉₄₇ effect on ecosystem functioning.Our study shows the positive effect of Ephedra shrubs on the arthropod fauna of the desert Puna, and stresses the importance of local and landscape conditions to maintain diverse arthropod communities.Given the crucial role played by arthropods in multiple ecosystem processes, a holistic approach that considers not only the plant species but also their associated arthropod communities is paramount for conservation and restoration strategies.
A��������������.We are grateful to Virginia Cortés for assistance in the field, Carolina Rothen for soil analyses, and Segundo Núñez Campero, Gustavo E. Flores and Sergio A. Roig-Juñet for their help with the taxonomic identification of arthropods.Thanks are extended to the two anonymous reviewers that provided helpful comments on the earlier version of the manuscript.This research was funded by FONCYT (PICT-2019-1816 to AAR).NIY and AYDC were supported with postdoctoral and doctoral fellowships from CONICET, and JASM with a posdoctoral fellowship from FONCYT.JT and AAR are career researchers with CONICET.Field collection of entomological material in this study was carried out with the permission of the Secretaría de Medio Ambiente del Gobierno de la Provincia de Catamarca (Resol.No 905/15).

Figure 1 .
Figure 1.Ephedra multiflora (a) and E. breana (b) sampling sites in the Desert Puna of Catamarca, Argentina.Arrows show individual plants of each Ephedra species.Figura 1. Sitios de muestreo de Ephedra multiflora (a) y E. breana (b) en la Puna desértica de Catamarca, Argentina.Las flechas indican individuos de cada especie de Ephedra.

Table 2 .
Results of generalized linear mixed model selection testing for arthropod abundance, species richness and Shannon diversity index responses to Ephedra species (E.breana and E. multiflora), microsite (