Shrub facilitation drives tree establishment in a semiarid fog-dependent ecosystem

Received Date : 14-Jul-2015 Revised Date : 25-Dec-2016 Accepted Article Accepted Date : 03-Jan-2017 Article type : Research Article Co-ordinating Editor : Ralf Ohlemuller Running head: Specific plant interactions drive fog-dependent forest formation Shrub facilitation drives tree establishment in a semiarid fog-dependent ecosystem Petr Macek1,2*, Christian Schöb3, Mariela Núñez-Ávila4,5, Iván R. Hernández Gentina6, Francisco I. Pugnaire2, Juan J. Armesto4,7 1 Faculty of Science, University of South Bohemia, Branišovská 1760, CZ-37005 České Budějovice, Czech Republic 2 LINCGlobal, Estación Experimental de Zonas Áridas, Consejo Superior de Investigaciones Científicas, Ctra. Sacramento s/n, 04120 La Cañada, Almería, Spain 3 Department of Evolutionary Biology and Environmental Studies, University of Zurich, Winterthurerstrasse 190, CH-8057 Zürich, Switzerland 4 LINCGlobal, Departamento de Ecología, Facultad Ciencias Biológicas, Pontificia Universidad Católica de Chile, Casilla 114-D, Santiago, Chile 5 Departamento Manejo de Bosques y Medio Ambiente, Facultad de Ciencias Forestales, Universidad de Concepción, Victoria 631, Concepción, Chile This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/avsc.12301 This article is protected by copyright. All rights reserved.  6 Corporación Nacional Forestal, Vicuña Mackenna 310, Ovalle, Coquimbo. 7 Millennium Institute of Ecology and Biodiversity, Alameda 340, Santiago, Chile Accepted Article * Correspondence: P.M., Faculty of Science, University of South Bohemia, Branišovská 31, CZ-37005 České Budějovice, Czech Republic; Tel.: +42 387776303; E-mail: maca@prf.jcu.cz Abstract Questions: The exceptional occurrence of tall rainforest patches on foggy coastal mountaintops, surrounded by extensive xerophytic shrublands, suggests an important role of plant–plant interactions in the origin and persistence of these patches in semiarid Chile. We asked whether facilitation by shrubs can explain the growth and survival of rainforest tree species, and whether shrub effects depend on the identity of the shrub species itself, the drought tolerance of the tree species, and the position of shrubs in regards to wind direction. Locations: Open area-shrubland-forest matrix at Fray Jorge Forest National Park, Chile. Methods: We recorded survival after 12 years of a ~3600 tree saplings plantation (originally ~30 cm tall individuals) of Aextoxicon punctatum, Myrceugenia correifolia and Drimys winteri placed outside forests underneath the shrub Baccharis vernalis and in open (shrub- free) areas. We assessed the effects of neighbouring shrubs and soil humidity on survival and growth along a gradient related to the direction of fog movement. Results: Baccharis vernalis had a clear facilitative effect on tree establishment and survival since, after ~12 years, saplings only survived underneath the shrub canopy. Long-term survival strongly depended on tree species identity, drought tolerance, and position along the soil moisture gradient, with higher survival of A. punctatum (>35%) and M. correifolia This article is protected by copyright. All rights reserved.  (>14%) at sites on wind- and fog-exposed shrubland areas. Sites occupied by the shrub Aristeguetia salvia were unsuitable for trees, presumably due to the drier conditions than Accepted Article under B. vernalis. Conclusions: Interactions between shrubs and fog-dependent tree species in dry areas revealed a strong, long-lasting facilitation effect on planted trees survival and growth. Shrubs acted as benefactors providing sites suitable for tree growth. Sapling mortality in the shrubland interior was caused by lower soil moisture, the consequence of lower fog loads in the air and thus insufficient facilitation. While B. vernalis was a key ecosystem engineer (nurse) and intercepted fog water that dripped to trees planted underneath, drier sites with A. salvia were unsuitable for trees. Consequently, nurse effects related to water input are strongly site- and species-specific, with facilitation by shrubs providing a plausible explanation for the initiation of forest patches in this semiarid landscape. Key words: Aextoxicon punctatum; Baccharis vernalis; Drought stress gradient; Ecosystem engineers; Facilitation; Fragmented rainforest; Landscape patchiness; Plant-plant interactions; Succession. Introduction Positive plant-plant interactions, i.e. facilitation, became integrated into ecological theory over the last few decades and their prominent influence on shaping plant communities has been widely accepted (Bruno et al. 2003; He et al. 2013). Facilitation should gain importance with increasing environmental stress, i.e. the stress gradient hypothesis (Bertness & Callaway 1994; Brooker & Callaghan 1998), up to a point where its effect was likely to wane under extremely harsh conditions, such as extreme aridity (Michalet et al. 2006, 2014). Positive interactions could play a main role in the persistence and possible origin of isolated forest This article is protected by copyright. All rights reserved.  patches in the semiarid regions of Central Chile, where unique ecosystems strongly dependent on oceanic fog occupy mountaintops in a dry setting (Gutiérrez et al. 2008; Accepted Article Johnstone & Dawson 2010; Stanton et al. 2013). Indeed, the whole ecosystem and its dynamics would be very different without horizontal precipitation, as fog water intercepted by tree and shrub canopies represents more than one third of the total water supply to these systems (Dawson 1998; Uclés et al. 2014). In fog-dependent forests, plant-plant interactions acquire greater importance, as some species can use soil water derived from fog-drip and supply surplus water to other plants (Rigg et al. 2002; Ewing et al. 2009; Anthelme et al. 2014), and through their engineering effects drive successional processes. Understory plants in the California redwood forest, for instance, receive up to two thirds of their annual water budget in the form of fog-drip from dominant canopy species (Dawson 1998). On the other hand, negative plant-plant interactions, i.e. competition, often play important roles in shaping interaction outcomes in such water-limited environments (Macek et al. 2016). Eventually, competition for water with other species may even reverse the facilitation effect, resulting in a specific spatial pattern of vegetation (Tielbörger & Kadmon 2000). Relict forest patches in semiarid north-central Chile (30ºS) are excellent examples of fog-dependent systems in a low rainfall environment. These rainforests on coastal mountaintops represent important biodiversity hotspots, harboring a rich assemblage of rainforest species (del-Val et al. 2006; Núñez-Ávila et al. 2006, 2013). Patches are restricted to coastal range summits, where advection brings dampened air from the Pacific Ocean developing a persistent fog belt above 400 m elevation (Barbosa et al. 2010). Forest patches form a mosaic in a landscape dominated by xerophytic shrubs and open, herbaceous- dominated areas (Fig. 1A). This article is protected by copyright. All rights reserved.  This system may be under threat from ongoing climate change due to the potential reduction of fog influx (Gutiérrez et al. 2008). Conservation concerns have been raised in Accepted Article recent decades regarding the conservation of these relict, fog-dependent forest patches. For instance, a recent afforestation effort led by the Chilean Forest Service (CONAF) was aimed at connecting forest patches trying to expand their area (Hernández & Vita 2004). The success of such efforts is uncertain due to the reduced interception of fog by shrubs and trees outside forest patches. The planted tree species, which occur within forest patches, have variable drought tolerance (Salgado-Negret et al. 2013), allowing to test the hypothesis that stress-sensitive species are more likely facilitated by shrubs than stress-tolerant species (cf. Liancourt et al. 2005). In addition, the plantation was designed to use dominant shrub species within a shrubland matrix outside the forest patches as potential “nurses” (sensu Turner et al. 1966), specifically, to improve conditions for fog condensation. In this way, facilitative interactions may assist ecosystem restoration in stressful environments (Gómez-Aparicio et al. 2004; Padilla & Pugnaire 2009). In Fray Jorge Forest National Park (Fray Jorge hereafter), facilitation could provide a mechanism to explain the formation of new forest patches, thereby increasing the connectivity among existing patches. In addition, secondary succession would eventually contribute to forest expansion. Here we analyzed the success of a 12-year experimental plantation established along an environmental gradient with respect to fog direction, to test the facilitation effects of shrub cover on tree survival. We asked whether facilitation could potentially provide the starting point for new forest patches, eventually leading to the expansion and interconnection of forest patches. We estimated the survival and growth of planted saplings underneath and outside shrub canopies, and assessed the role of shrub position in relation to other dominant woody species in the dry areas outside forest patches. We tested the following hypotheses: 1) Nurse shrubs will provide a more suitable environment (e.g. wetter soils), enhancing target tree This article is protected by copyright. All rights reserved.  species survival and growth relative to open sites without shrubs; 2) Survival and growth of target tree species will vary according to their differential tolerance to drought (Salgado- Accepted Article Negret et al. 2013), so that the species with the lowest drought tolerance will be influenced the most by the nurse shrub; and 3) Net interaction outcome will also depend on the presence of shrub species other than the nurse, meaning that nurses will further differ in their facilitation effects depending on the surrounding species. Methods Study area Fray Jorge Forest National Park is located in the semiarid zone of the Chilean coast (30°38’S, 71°40’W). Annual rainfall averages 127 mm (1983–2013), with 95% of the rain falling in the austral winter (i.e., June-August; Montecinos et al. 2016; López-Cortés & López 2004). Fog- driven water influx (Stanton et al. 2013), which depends on fog-interception by both leaves and stems of trees and shrubs (Fig. 1A), is nearly constant or enhanced during the drier spring and summer months (October-March; Garreaud et al. 2008). Fray Jorge contains an ancient and naturally fragmented mosaic of rainforest patches surviving a long process of aridization (del-Val et al. 2006; Gutiérrez et al. 2008; Núñez-Ávila et al. 2013). Remnants of relict preglacial rainforest (Villagrán et al. 2004) are scattered on coastal summits (Fig. 1A) and dominated mainly by three broadleaved evergreen trees, Aextoxicon punctatum Ruiz et Pav. (Aextoxicaceae), Myrceugenia correifolia Hook et Arn. (Myrtaceae), and Drimys winteri J.R. Forst. et G. Forst. (Winteraceae). Species studied Drimys winteri is the most hygrophilous species, and M. correifolia is the most xerophytic species, while A. punctatum displays an intermediate drought tolerance (Muñoz & Pisano 1947; Salgado-Negret et al. 2013). Drimys winteri occurs along an extensive This article is protected by copyright. All rights reserved.  geographic range from Fray Jorge (30º S) to Cape Horn (56º S) in Patagonia, reaching much further south than A. punctatum or M. correifolia. Communities in areas outside forest Accepted Article patches are dominated by evergreen and deciduous shrubs (<1.5 m tall and with homogeneous density). Baccharis vernalis F.H. Hellwig is a common shrub species outside forest patches, a species which tolerates dry conditions outside forests and may grow isolated or along with other shrub species in open areas. In the experiment by CONAF, the evergreen B. vernalis was considered a potential “nurse” for forest trees in open areas thanks to its ability to intercept fog (Hernández & Vita 2004). Experimental design An experimental plantation was established at Fray Jorge by CONAF between 1999 and 2001, resulting in over 3600 planted saplings (~30 cm tall). Saplings of the three main tree species found in forest patches, A. punctatum (n = 1800), D. winteri (n = 720), and M. correifolia (n = 1080), were planted in numbers that followed a 5:2:3 ratio, corresponding with their overall abundances in existing forest patches. These plants were grown in a local nursery within Fray Jorge National Park, derived from seeds collected from forest patches. Saplings were planted in equal quantities both underneath B. vernalis shrubs (which forms irregularly shaped shrublands) and in areas between shrubs (hereafter called open areas); shrubs follow a random distribution within open areas and do not seem to fit any pre-existing environmental differences. Saplings underneath B. vernalis were planted near the stem. Although the main direct effect on saplings came from B. vernalis, there were other shrub species surrounding the planted saplings. All saplings were individually protected from herbivory (rodents and rabbits) with wire fences. Saplings were planted at least 3 m from each other or farther (see Hernández & Vita 2004 for details). Establishment was initially supported by watering; fog was intercepted by large mist nets set at the study sites, and the This article is protected by copyright. All rights reserved.  water was redistributed to the planted trees for two years, which resulted in high initial survival rates (97.3%, 74.3%, and 83.3% for A. punctatum, D. winteri and M. correifolia, Accepted Article respectively; Hernández & Vita 2004). We re-assessed the outcome of the experiment in 2012 (i.e., 12 years after plantation), and linked sapling survival rate to potential nurse shrubs and the position of planted trees with regard to the surrounding vegetation. The survival rates were calculated as the percent of live saplings out of the total number of found planting locations of the respective species (again following a 5:2:3 ratio for A. punctatum, D. winteri, and M. correifolia, respectively). We also compared the height of the surviving saplings (log-transformed prior to analyses to meet the normality and homoscedasticity criteria), assuming a fairly similar height at the time of plantation, and recorded all shrub and tree species present within 2 m from every target plant. To check for soil water status, we sampled surface soil (~5 cm) under B. vernalis individuals along a shrubland edge-interior gradient as well as in open areas (n = 32). Soil samples were weighed in the field and their water content estimated after drying for 48 h at 105°C. In addition, wind direction during the last two years of the experimental period were monitored at 15 min intervals. Data analyses A generalized linear model (GLM) with a binomial distribution of error terms was used for the survival data analysis, with site as the factor; i.e., saplings in open sites vs. saplings under B. vernalis. Saplings under B. vernalis were further separated into three groups of approximately equal sample sizes according to the distance from the wind-exposed edge (facing fog influx) of the shrubland; i.e., <2 m, 2-15 m, and >15 m towards the shrubland interior. Kruskal-Wallis ANOVA was used to compare sapling height at different distances. This article is protected by copyright. All rights reserved.  Statistical analyses were performed in R 3.1.1 (http://www.R-project.org). For each month, the centroid of the wind direction was calculated as the center of gravity of the oriented wind Accepted Article speed vectors resulting in a predominant monthly wind direction; these centroids (n = 24) were then analyzed using a Rayleigh test to check for their uniform distribution around a circle (Batschelet, 1981). Sapling survival of each tree species underneath shrubs was estimated for every possible species assemblage (i.e. combination of presence/absence data of B. vernalis and other shrub and tree species within 2 m from the transplant) and the number of samples was used as a weighting factor in a multivariate analysis. Redundancy Analysis (RDA, with standardized variables) was used to relate survival to the specific plant assemblage; only sapling locations underneath B. vernalis (n = 950) were used for this analysis. To reduce Type II error, we used shrub species data as response variables predicted by sapling survival probabilities (Lepš & Šmilauer 2003). RDA can be considered as an extension of multivariate linear regression for a multivariate response variable (Lepš & Šmilauer 2003), with the parametric test replaced by the Monte Carlo permutation test. Explanatory variables were selected by forward selection (499 permutations; p = 0.002). As the probability of M. correifolia survival was correlated to that of A. punctatum, it was added as a passive variable only, hence without any effect on the analysis. Similarly, the environmental variable “Distance from the wind-exposed shrubland edge” was considered as a passive variable. The RDA and visualization of its results were carried out with Canoco 4.5 and CanoDraw 4 (ter Braak & Šmilauer 2002). Results After 12 years we were able to find 2328 locations (i.e., planting sites with or without surviving saplings) out of the original ~3600 planted tree saplings, of which 950 were under the nurse shrub B. vernalis and 1378 in open areas without shrub cover. There was strong This article is protected by copyright. All rights reserved.  evidence of facilitation effects of shrubs on sapling survival (F(1,1163) = 155.40, p < 0.001 for A. punctatum; F(1,701) = 41.33, p < 0.001 for M. correifolia), as no saplings survived after 12 Accepted Article years in open areas, i.e. without B. vernalis cover. The lowest mortality rates (Fig. 2) after ~12 years were recorded for A. punctatum (18%) and M. correifolia (9%) saplings underneath B. vernalis. The saplings of D. winteri did not survive anywhere. The wind direction centroids were not uniformly distributed (Rayleigh test; z = 22.4, p < 0.001), and had a mean angle of 355°; i.e., fog-loaded winds blew predominantly from the south all year round (Fig. 1B). Hence, shrubs closer to the southern edge of the shrubland facing the wind have a higher possibility to intercept fog. Furthermore, survival probabilities declined significantly with distance from the wind-exposed edge to the shrubland interior for A. punctatum (F(2,449) = 36.7, p < 0.001), and were marginally significant for M. correifolia (F(2,270) = 2.98, p = 0.053; Fig. 3A); hence, sapling survival reached over 35% for A. punctatum and 14% for M. correifolia at the wind-exposed shrubland edge. SWC followed a similar trend, decreasing with distance from the wind-exposed edge to the shrubland interior (F3,28 = 35.9, p < 0.001; Fig. 3B). Distance from the shrubland wind-exposed edge to the interior also affected the height of A. punctatum, which decreased with increasing distance to the edge (H(2,81) = 6.58, p = 0.037; Fig. 4). In contrast, M. correifolia showed no such relationship (H(2,24) = 1.30, p = 0.522). Mean A. punctatum heights were 38 cm, 48 cm and 17 cm, and mean M. correifolia heights were 20 cm, 17 cm and 23 cm at <2 m, 2-15 m, and >15 m from the wind-exposed edge, respectively. Aextoxicon punctatum, but not M. correifolia, survival was significantly related to the shrub species surrounding each planted tree (eigenvalue = 12.8, F = 8.33, p = 0.002). Species scores on the first canonical axis were negative (-0.47) for Aristeguietia salvia (Colla) R. M. This article is protected by copyright. All rights reserved.  King. et H. Rob., and positive (0.43) for A. punctatum and (0.43) Griselinia scandens (Ruiz et Pav.) Taubert; other shrub species showed less correlation to the main canonical axis (Fig. Accepted Article 5). Discussion The importance of nurse shrubs for tree survival Fog-dependent ecosystems depend on fog interception and water drip to the soil (Johnston & Dawson 2010). Efforts to extend fog-dependent forest cover by interconnecting isolated forest patches in Fray Jorge included a large experimental plantation in the semiarid shrubland matrix. However, we conclude, based on sapling survival after 12 years, that this effort had very limited success in most sites. The large-scale transplant experiment provided, nonetheless, a great opportunity to test hypotheses about the long-term effects of plant-plant interactions in restoration experiments (Gómez-Aparicio 2009). The expectation of the Chilean Forest Service (CONAF) was to find positive effects of shrubs on tree establishment as a potential, inexpensive method for expanding tree cover (Hernández & Vita 2004). Here, we documented the significant positive effect (although varying; see below) of B. vernalis on sapling survival (van Zonneveld et al. 2012). No trees survived in open areas after irrigation was stopped after the first two years, suggesting that shrubs provide an additional source of soil moisture to plants growing underneath them, as reported in other fog-dependent systems (Dawson 1998; Dunne & Parker 1999; Rigg et al. 2002). The high tree mortality in open areas stresses the ecological significance of fog water interception and delivery to the soil by plants, an engineering effect which enhanced the overall survival of one of the target tree species by up to 18% in some places in the landscape after a decade without irrigation. Intercepted fog and dripping were noticed by the higher soil moisture under the shrub, and was evidenced in this and other fog-dependent systems using stable isotopes (Aravena et al. This article is protected by copyright. All rights reserved.  1989; Dawson 1998). Comparing the differences in sapling survival between open areas and the understory of B. vernalis clearly showed that even well-established saplings (up to 50 cm Accepted Article tall) were unable to intercept enough water to sustain their own growth and thus survive in open areas, highlighting the role of fog-capturing structures (trees or shrubs) and emphasizing the importance of facilitative interactions. While other experimental studies have focused on shrub effects on herbs (e.g. Rolhauser & Pucheta 2015; Liczner et al. 2016) or short-term seedling survival (Sthultz et al. 2006; Leiva et al. 2015), our data provide strong experimental evidence of the long-term facilitative effect of shrubs on other woody species. Species-specificity of plant-plant interactions Facilitation mechanisms associated with the interception of fog by shrubs were highly species-specific. The shrubs were unable to support young Drimys winteri trees, a relatively drought-intolerant species restricted to the largest forest patches in Fray Jorge (Gutiérrez et al. 2008) and in fact D. winteri was the least drought-tolerant species used in the experiment (Salgado-Negret et al. 2013). Our prediction that this species would be facilitated the most was rejected, presumably because conditions generated by water dripping underneath B. vernalis did not fit D. winteri habitat requirements. On the other hand, the intermediately drought-tolerant species A. punctatum survived better than the more xerophytic species M. correifolia. Therefore, our results stress species-specificity of plant-plant interactions, and are in line with previous reports suggesting that the degree of deviation from the physiological optimum can predict species’ responses to plant interactions (Gross et al. 2010). As water scarcity limits plant growth, facilitation can only occur if competition for water is weaker than benefits from interacting species (Holmgren et al. 1997; Tielbörger & Kadmon 2000), or if the nurse shrub increases soil water availability for the facilitated species (Zou et al. 2005; Maestre et al. 2009; Prieto et al. 2010). Directional fog loads imply This article is protected by copyright. All rights reserved.  that B. vernalis shrubs closer to the wind-exposed edge of the shrubland intercepted a greater amount of fog than shrubs in the interior, as was also documented for forest patches (Stanton Accepted Article et al. 2013). Indeed, drier soils under B. vernalis in the shrubland interior suggest that this species decreased its positive effect away from the wind-exposed edges. For instance, survival of A. punctatum reached 35% at the wind-exposed edge while it was virtually null 15 m away from the windward edge. Therefore, we suggest a facilitation cessation towards the shrubland interior due to a large fraction of air humidity being removed by dripping at the wind-exposed edge, resulting in drier soils in the interior (Figs. 1A, 3B). We show here that ocean-facing (wind- and fog-exposed) edges hosted species requiring greater humidity, such as the vine G. scandens. Aextoxicon punctatum and M. correifolia showed opposing trends along the shrubland gradient, i.e. while A. punctatum showed significantly less growth towards the interior, growth of M. correifolia did not change. These opposing trends may indicate that the increased drought stress in the interior is better tolerated by the xerophytic M. correifolia (Muñoz & Pisano 1947; Salgado-Negret et al. 2013) than by A. punctatum. We therefore found consistent support for our hypothesis that facilitation could provide a starting point for new forest patches in open areas. Succession may then drive community assembly from the presence of more drought tolerant species (M. correifolia, A. punctatum) to more water- demanding species (G. scandens, D. winteri) during patch growth, which is supported by the differential species composition of differently sized forest patches (del Val et al. 2006). Succession towards forest patches Consequently, nurses provide an efficient and inexpensive way to increase the chances of success in planting fog-dependent tree species, which are presently restricted almost entirely to forest patches. This process would be limited to wind-exposed areas and wet This article is protected by copyright. All rights reserved.  periods, like ENSO years (Holmgren & Scheffer 2001; Squeo et al. 2007). Not all dominant tree species in forest patches, however, are suitable for planting, or can initiate a forest patch. Accepted Article Drought-tolerant tree species like M. correifolia could be used at the start of plantation programs, or can initiate patch establishment in open areas, followed later by tree species such as A. punctatum, which have a more uncertain outcome. Once these species establish there is a higher probability of developing incipient forest patches with suitable conditions for spontaneous tree establishment of other, less drought-tolerant species. Eventually, incipient forest patches should help to increase the connectivity among older forest patches and start a successional trajectory, which may however take a long time if we are to judge by the growth rates of the established saplings. Therefore, long-term fluctuations in both fog influx and rainfall may become significant drivers of the process. However, the use of nurse plants may be restricted to certain species, and a good understanding of species is necessary before planning large-scale manipulations (Padilla & Pugnaire 2006). In Fray Jorge, sites covered by B. vernalis, but not A. salvia, would be suitable. Careful selection of planting locations (e.g., the presence of nurse shrubs) and tree species selection according to their drought tolerance, as well as the influence of surrounding vegetation, must be taken into account. Plantations can only be successful on a limited spatial scale given the present patterns of forest sizes and their distribution in the landscape (Gutiérrez et al. 2008). Acknowledgments Many thanks are due to Juan Monardez for field assistance. We thank CONAF for allowing research at Fray Jorge. We thank Ondřej Mudrák and three anonymous reviewers for helpful comments on the manuscript and Keith Edwards for linguistic revision. Martin Hanáček kindly provided the drawing in figure 1A. This work is a contribution of the LINCGlobal Project. P.M. was additionally supported by LM2015078 and a CSIC-JAEDoc This article is protected by copyright. All rights reserved.  Program co-financed by ESF. C.S. was supported by the Swiss National Science Foundation (PBBEP3_128361, PZ00P3_148261) and CSIC (PA1003183). Additional funds were Accepted Article provided by MICINN (grant CGL2014-59010-R). This publication is also part of the research program of the Long-term Socio-Ecological Research (LTSER) Network-Chile, supported by the Institute of Ecology and Biodiversity with funding from CONICYT grant PFB-23 and Millennium Scientific Initiative grant P05-002. References Anthelme, F., Gómez-Aparicio, L. & Montufar, R. 2014. Nurse-based restoration of degraded tropical forests with tussock grasses: experimental support from the Andean cloud forest. Journal of Applied Ecology 51: 1534–1543. Aravena, R., Suzuki, O. & Pollastri, A. 1989. Coastal fog and its relation to groundwater in the IV region of northern Chile. Chemical Geology 79: 83–91. Barbosa, O., Marquet, P.A., Bacigalupe, L.D., Christie, D.A., del-Val, E., Gutierrez, A.G., Jones, C.G., Weathers, K.C. & Armesto, J.J. 2010. Interactions among patch area, forest structure and water fluxes in a fog-inundated forest ecosystem in semi-arid Chile. Functional Ecology 24: 909–917. Batschelet, E. 1981. Circulat Statistics in Biology. Academin Press, London, UK. Bertness, M.D. & Callaway, R. 1994. Positive interactions in communities. Trends in Ecology & Evolution 9: 191–193. Brooker, R.W. & Callaghan, T.V. 1998. The balance between positive and negative plant interactions and its relationship to environmental gradients: a model. Oikos 81: 196– 207. Bruno, J.F., Stachowicz, J.J. & Bertness, M.D. 2003. Inclusion of facilitation into ecological theory. Trends in Ecology & Evolution 18: 119–125. This article is protected by copyright. All rights reserved.  Dawson, T.E. 1998. Fog in the California redwood forest: ecosystem inputs and use by plants. Oecologia 117: 476–485. Accepted Article del-Val, E., Armesto, J.J., Barbosa, O., Christie, D.A., Gutiérrez, A.G., Jones, C.G., Marquet, P.A. & Weathers, K.C. 2006. Rain forest islands in the Chilean semiarid region: Fog- dependency, ecosystem persistence and tree regeneration. Ecosystems 9: 598–608. Dunne, J.A. & Parker, V.T. 1999. Species-mediated soil moisture availability and patchy establishment of Pseudotsuga menziesii in chaparral. Oecologia 119: 36–45. Ewing, H.A., Weathers, K.C., Templer, P.H., Dawson, T.E., Firestone, M.K., Elliott, A.M. & Boukili, V.K.S. 2009. Fog water and ecosystem function: Heterogeneity in a California redwood forest. Ecosystems 12: 417–433. Garreaud, R., Barichivich, J., Christie, D.A. & Maldonado, A. 2008. Interannual variability of the coastal fog at Fray Jorge relict forests in semiarid Chile. Journal of Geophysical Research-Biogeosciences 113. Gómez-Aparicio, L., Zamora, R., Gómez, J.M., Hodar, J.A., Castro, J. & Baraza, E. 2004. Applying plant facilitation to forest restoration: A meta-analysis of the use of shrubs as nurse plants. Ecological Applications 14: 1128–1138. Gómez-Aparicio, L. 2009. The role of plant interactions in the restoration of degraded ecosystems: a meta-analysis across life-forms and ecosystems. Journal of Ecology 97: 1202–1214. Gross, N., Liancourt, P., Choler, P., Suding, K.N. & Lavorel, S. 2010. Strain and vegetation effects on local limiting resources explain the outcomes of biotic interactions. Perspectives in Plant Ecology Evolution and Systematics 12: 9–19. Gutiérrez, A.G., Barbosa, O., Christie, D.A., del-Val, E., Ewing, H.A., Jones, C.G., Marquet, P.A., Weathers, K.C. & Armesto, J.J. 2008. Regeneration patterns and persistence of This article is protected by copyright. All rights reserved.  the fog-dependent Fray Jorge forest in semiarid Chile during the past two centuries. Global Change Biology 14: 161–176. Accepted Article He, Q., Bertness, M.D. & Altieri, A.H. 2013. Global shifts towards positive species interactions with increasing environmental stress. Ecology Letters 16, 695-706. Hernández, I.R. & Vita, A. 2004. Reforestación para la expansión de los bosquetes de Olivillo. In: Squeo, F.A., Gutiérrez, J.R. & Hernández, I.R. (eds.) Historia Natural del Parque Nacional Bosque Fray Jorge, pp. 307–319. Ediciones Universidad de La Serena, La Serena, Chile. Holmgren, M. & Scheffer, M. 2001. El Niño as a window of opportunity for the restoration of degraded arid ecosystems. Ecosystems 4, 151‒159. Holmgren, M., Scheffer, M. & Huston, M.A. 1997. The interplay of facilitation and competition in plant communities. Ecology 78: 1966–1975. Johnstone, J.A. & Dawson, T.E. 2010. Climatic context and ecological implications of summer fog decline in the coast redwood region. Proceedings of the National Academy of Sciences of the United States of America 107: 4533–4538. Leiva, M.J., Mancilla-Leytón, J.M. & Martín Vicente, A. 2015. Differences in the facilitative ability of two Mediterranean shrubs on holm-oak seedling recruitment in Mediterranean savanna-forest ecosystems. Ecological Engineering 82: 349–354 Lepš, J. & Šmilauer, P. 2003. Multivariate Analysis of Ecological Data Using CANOCO. Cambridge University Press, Cambridge. Liancourt, P., Callaway, R.M. & Michalet, R. 2005. Stress tolerance and competitive- response ability determine the outcome of biotic interactions. Ecology 86: 1611–1618. Liczner, A.R., Sotomayor, D.A., Filazzola, A. & Lortie, C.J. 2016. Germination response of desert annuals to shrub facilitation is species specific but not ecotypic. Journal of Plant Ecology, doi:10.1093/jpe/rtw030. This article is protected by copyright. All rights reserved.  López-Cortés, F. & López, D. 2004. Antecedentes bioclimáticos del Parque Nacional Bosque Fray Jorge. In: Squeo, F.A., Gutiérrez, J.R. & Hernández, I.R. (eds.) Historia Natural Accepted Article del Parque Nacional Bosque Fray Jorge, pp. 45–60. Ediciones Universidad de La Serena, La Serena, Chile. Macek, P., Prieto, I., Macková, J., Pistón, N. & Pugnaire, F.I. 2016. Functional plant types drive plant interactions in a Mediterranean mountain range. Frontiers in Plant Science 7: 662. Maestre, F.T., Callaway, R.M., Valladares, F. & Lortie, C.J. 2009. Refining the stress- gradient hypothesis for competition and facilitation in plant communities. Journal of Ecology 97: 199–205. Michalet, R., Brooker, R.W., Cavieres, L.A., Kikvidze, Z., Lortie, C.J., Pugnaire, F.I., Valiente-Banuet, A. & Callaway, R.M. 2006. Do biotic interactions shape both sides of the humped-back model of species richness in plant communities? Ecology Letters 9: 767–773. Michalet, R., Le Bagousse-Pinguet, Y., Maalouf, J.P. & Lortie, C.J. 2014. Two alternatives to the stress-gradient hypothesis at the edge of life: the collapse of facilitation and the switch from facilitation to competition. Journal of Vegetation Science 25: 609–613. Montecinos, S., Guttiérez, J.R., López-Cortés, F. & López, D. 2016. Climatic characteristics of the semi-arid Coquimbo Region in Chile. Journal of Arid Environments 126: 7‒11. Muñoz, C. & Pisano, E. 1947. Estudio de la vegetación y flora de los Parques Nacionales de Fray Jorge y Talinay. Agricultura Técnica 7: 71–190. Núñez-Ávila, M.C. & Armesto, J.J. 2006. Relict islands of the temperate rainforest tree Aextoxicon punctatum (Aextoxicaceae) in semi-arid Chile: genetic diversity and biogeographic history. Australian Journal of Botany 54: 733–743. This article is protected by copyright. All rights reserved.  Núñez-Ávila, M.C., Uriarte, M., Marquet, P.A. & Armesto, J.J. 2013. Decomposing recruitment limitation for an avian-dispersed rain forest tree in an anciently fragmented Accepted Article landscape. Journal of Ecology 101: 1439–1448. Padilla, F.M. & Pugnaire, F.I. 2006. The role of nurse plants in the restoration of degraded environments. Frontiers in Ecology and the Environment 4: 196–202. Padilla, F.M. & Pugnaire, F.I. 2009. Species identity and water availability determine establishment success under the canopy of Retama sphaerocarpa shrubs in a dry environment. Restoration Ecology 17: 900–907. Prieto, I., Kikvidze, Z. & Pugnaire, F.I. 2010. Hydraulic lift: soil processes and transpiration in the Mediterranean leguminous shrub Retama sphaerocarpa (L.) Boiss. Plant and Soil 329: 447–456. Rigg, L.S., Enright, N.J., Perry, G.L.W. & Miller, B.P. 2002. The role of cloud combing and shading by isolated trees in the succession from maquis to rain forest in New Caledonia. Biotropica 34: 199–210. Rolhauser, A.G. & Pucheta, E. 2015. Annual plant functional traits explain shrub facilitation in a desert community. Journal of Vegetation Science 27: 60‒68. Salgado-Negret, B., Pérez, F., Markesteijn, L., Castillo, M.J. & Armesto, J.J. 2013. Diverging drought-tolerance strategies explain tree species distribution along a fog-dependent moisture gradient in a temperate rain forest. Oecologia 173: 625–635. Squeo, F.A., Holmgren, M., Jiménez, M., Albán, L., Reyes, J. & Gutiérrez, J.R. 2007. Tree establishment along an ENSO experimental gradient in the Atacama desert. Journal of Vegetation Science 18: 195‒202. Stanton, D.E., Negret, B.S., Armesto, J.J. & Hedin, L.O. 2013. Forest patch symmetry depends on direction of limiting resource delivery. Ecosphere 4. This article is protected by copyright. All rights reserved.  Sthultz, C.M., Gehring, C.A. & Whitham, T.G. 2006. Shifts from competition to facilitation between a foundation tree and a pioneer shrub across spatial and temporal scales in a Accepted Article semiarid woodland. New Phytologist 173: 135–145. ter Braak, C.J.F. & Šmilauer, P. 2002. CANOCO Reference Manual and CanoDraw for Windows User's Guide: Software for Canonical Community Ordination (version 4.5). Microcomputer Power, Ithaca, NY. Tielborger, K. & Kadmon, R. 2000. Temporal environmental variation tips the balance between facilitation and interference in desert plants. Ecology 81: 1544–1553. Turner, R.M., Alcorn, S.M., Olin, G. & Booth, J.A. 1966. The influence of shade, soil, and water on saguaro seedling establishment. Botanical Gazette 127: 95–102. Uclés, O., Villagarcía, L., Moro, M. J., Canton, Y. & Domingo, F. 2014. Role of dewfall in the water balance of a semiarid coastalsteppe ecosystem. Hydrological Processes 28: 2271–2280. van Zonneveld, M.J., Gutiérrez, J.R. & Holmgren, M. 2012. Shrub facilitation increases plant diversity along an arid scrubland-temperate rain forest boundary in South America. Journal of Vegetation Science 23: 541–551. Villagrán, C., Armesto, J.J., Hinojosa, L.F., Cuvertino, J., Pérez, C. & Medina, C. 2004. El enigmático origen del bosque relicto de Fray Jorge. In: Squeo, F.A., Gutiérrez, J.R. & Hernández, I.R. (eds.) Historia Natural del Parque Nacional Bosque Fray Jorge, pp. 3–43. Ediciones Universidad de La Serena, La Serena, Chile. Zou, C.B., Barnes, P.W., Archer, S. & McMurtry, C.R. 2005. Soil moisture redistribution as a mechanism of facilitation in Savanna tree-shrub clusters. Oecologia 145: 32–40. This article is protected by copyright. All rights reserved. ccepted Articl Figure 1 A B This article is protected by copyright. All rights reserved. ccepted Articl A) Study area with fragmented forest, shrublands and open areas (above) and a theoretical model of fog-interception by shrublands (below). Wind direction is indicated by an arrow, and expectations about decreasing air (fog) and soil humidity along the shrubland edge-interior gradient are shown. The theoretical model was then confirmed by our measurements (see results for details). B) Wind directions in Fray Jorge – on the basis of data from a meteorological station in between forest fragments separated by the shrubland-open area matrix studied (monthly values from April 2010 to March 2012). Centroids of wind direction were calculated as centers of gravity of the oriented wind speed vectors. Centroids were not uniformly distributed around a circle (Rayleigh test; z = 22.4, p < 0.001), and had a mean angle of 355° (dashed line), meaning the main direction of the wind was south. Concentric circle values represent wind speed (m s-1). This article is protected by copyright. All rights reserved.  Figure 2 Accepted Article Mortality (%) of planted Aextoxicon punctatum (Aexpun; grey), Drimys winteri (Driwin; black) and Myrceugenia correifolia (Myrcor; white) tree species in open areas and underneath the nurse shrub Baccharis vernalis. This article is protected by copyright. All rights reserved. ccepted Articl Figure 3 A B (A) Survival probabilities of Aextoxicon punctatum (grey) and Myrceugenia correifolia planted saplings (white) according to their planting location in open areas and underneath Baccharis vernalis shrubs further divided by distance (m) from the wind-exposed edge of the shrubland; mean ± Clopper-Pearson [binomial] 95% confidence intervals; non overlapping confidence intervals mean significant differences. (B) Soil water content (SWC) at different planting locations; different letters on the top of the figure correspond to significant differences in SWC (p < 0.05). This article is protected by copyright. All rights reserved.  Fig. 4 Accepted Article ab b a 80 70 Sapling height (cm) 60 50 40 30 20 10 0 <2 2-15 >15 Distance from shrubland edge (m) Sapling heights of Aextoxicon punctatum (grey) and Myrceugenia correifolia (white) along the shrubland edge to interior gradient. Mean ± SE (box) ± SD (whisker). Different letters correspond to significant differences in sapling height of A. punctatum (p < 0.05). This article is protected by copyright. All rights reserved.  Fig. 5 Accepted Article 0.8 Ribes Aristeguietia salvia punctatum Distance from wind- Colletia hystrix exposed shrubland edge Berberis actinacantha Axis 2 Aextoxicon survival Haplopappus foliosus Myrceugenia survival Senecio planiflorus Aextoxicon punctatum Griselinia scandens -0.8 Myrceugenia correifolia -1.0 Axis 1 1.0 RDA biplot showing the relationship between survival probabilities of planted saplings of Aextoxicon punctatum (bold line) and Myrceugenia correifolia, and the shrubland matrix composition (italics). Explanatory variables were selected by forward selection (499 permutations; p = 0.002). As the survival probability of Myrceugenia correifolia was correlated to that of A. punctatum, it was added to the analyses as a passive variable (dashed line) only, hence without an effect on the analysis. Similarly, the environmental variable “Distance from the wind-exposed shrubland edge” is shown as a passive variable. The first and second ordination axes are shown. This article is protected by copyright. All rights reserved.