DISCUSSION
In this study, after evaluating 42 species and 20 functional traits in one paramo site on the eastern range of the Northern Andes in Colombia, we found that many of the species studied shared similar attributes or functional traits, and they were clustered in three groups or plant functional types (PFTs). Traits that were powerful functional descriptors of PFTs in the paramo were traits related to the leaf nutrient status (C, N and P), leaf toughness, thickness and leaf dry matter content (LDMC), seed dispersal syndromes and physiological traits related to water use efficiency and tolerance to water deficit like Ψnoon, WUE and transpiration rate. Cuticular conductance did not differ among species, among growth forms or among PFTs, and given how time consuming this trait is to measure it could be omitted for future functional classification in the paramo.
We did find some evidence that growth forms correspond to PFTs and could be used to define groups with similar ecological and physiological traits, as was found for the Venezuelan paramos (Rada, Azócar and García-Núñez, 2019) and the tundra (Chapin et al. , 1996). However, for shrubs the story is complex; some species shared traits with the forbs while others form their own group. The classification of upright and prostrate shrub appears to be impractical since both forms were clustered together, and according to our data it would make more sense to classify them into tough and tender leaved shrubs. Where shrubs with tough leaves present a series of traits that suggest higher tolerance to water deficit.
One limitation of our study is that we did not include grasses or cushion plants in our analysis and these groups could potentially increase the number of PFTs. However, the increase of functional types is likely to be small if grasses belong to just one PFT as in the Venezuelan paramo where all nine species of grasses studied shared similar water and temperature resistant strategies (Rada, Azócar and García-Núñez, 2019). Similarly, in Chingaza, a different Colombian paramo, Cárdenas-Arévalo and Vargas-Ríos (2008) identified only three PFTs after analyzing 70 species, including rosettes, shrubs, forbs, grasses and cushion plants. Although all these findings from different paramos might suggest that paramos have reduced functional responses, we should keep in mind that mountains have complex environmental gradients and various microclimates (Scherrer and Körner, 2011; Sklenář et al. , 2016; Körner and Hiltbrunner, 2018) and that many species show significant variation in functional traits across elevation gradients (Baruch and Smith, 1979; Baruch, 1984; Meinzer, Goldstein and Rundel, 1985; Rada et al. , 1998; Wang et al. , 2016), which suggests a wide acclimation capacity for many species which was not explored here. Our study presents a picture of the functional diversity at a single elevation, but is possible that many of these species can acclimate and modify their traits; understanding species plasticity is key to fully evaluate if the few PFTs identified in paramos imply a limited capacity to respond to changes in the environment.
The paramo hosts the highest plant diversity of alpine ecosystems worldwide (Sklenář, Hedberg and Cleef, 2014), which has hindered our ability to understand the ecology and physiology of most of their species and the ecosystem services they provide. Paramos are entirely different from alpine temperate ecosystems so we cannot assume a priori that the same functional responses that we know from alpine areas apply to the paramo and here we provide some insights into the functional ecology of the paramo. To understand the adaptations and traits of paramo plants we have to take into account that in this tropical alpine ecosystem, temperature variation is not seasonal but diurnal; in a single day differences between absolute minimum and absolute maximum temperatures may exceed 20 to 30K (Sklenář et al. , 2016; Leon-Garcia and Lasso, 2019; Rada, Azócar and García-Núñez, 2019), including freezing temperatures. These species need to withstand great daily temperature changes and freezing temperatures. From studies on the plant response to freezing tolerance in the Venezuelan paramo (Llambí and Rada, 2019), we know that the strategy to deal with freezing temperatures varies among growth forms (Rada, Azócar and García-Núñez, 2019). Rosettes avoid freezing with insulating structure and supercooling, whereas forbs, cushion plants and grasses are freeze tolerant (Azocar, Rada and Goldstein, 1988; Squeo et al. , 1991). However, as we show in our study, shrubs are hard to place in a single strategy; some species avoid freezing (Squeo et al. , 1991), while others tolerate freezing (Azócar, 2006). Less is known about species’ tolerance to high temperature, which is the new challenge for plants in a warming world, but from a screening on heat tolerance in 21 paramo species we know that all species can tolerate high temperatures; and irreversible damages to the photosynthetic apparatus occurs when the leaves reaches 45oC, in some species, or in other species this happens at temperatures as high as 53.9oC, and rosettes seems to be especially tolerant (Leon-Garcia and Lasso, 2019). Again, growth forms appear to be associated with the strategies to deal with high temperatures, as well as with freezing temperature.
Besides temperature, water availability is also considered one of the most important environmental drivers in the paramo (Rada, Azócar and García-Núñez, 2019). The sudden daily changes in temperatures typical of the paramos also influence water availability. If temperatures descend below zero, water freezes and becomes unavailable, and when temperatures increase on clear days, evaporative demand increases (Ramirez, Rada and Llambí, 2014). Therefore, even in relatively wet paramos, water deficit can occasionally occur, especially in the dry season. The strategies of paramo plants to deal with water deficit also seem to be somewhat related to the growth form (Rada, Azócar and García-Núñez, 2019). In the Venezuelan drier paramos, many species of grasses and shrubs can reach quite low water potentials at noon which suggests they are tolerant of water deficit, while rosettes maintain higher (less negative) water potentials during the day, even in the dry season, which indicates they rely on avoidance mechanisms (Rada, Azócar and García-Núñez, 2019). In fact, many of these stem rosettes are known to store water in their stem pith allowing them to sustain higher water potentials thorough the day (Goldstein, Meinzer and Monasterio, 1984; Meinzer, Goldstein and Rundel, 1985).
In our study site, a relatively wet paramo, we found that all growth forms analyzed seem to behave more like avoiders with relatively high values (less negative) of noon water potential. However, one of the functional groups identified, the PFT-3, consisting of shrubs with tough leaves, had the lowest water potential at noon, slightly higher values of WUE and higher LDMC (Figure 3; Table A1 ). A higher LDMC is also associated with having leaves with higher modulus of elasticity (Niinemets, 2001) and more rigid lignified cell walls (Niinemets and Kull, 1998), which could allow them to support lower leaf water potentials and maintain leaf turgor even in periods of lower water availability. High LDMC has also been found to be associated with higher cavitation resistance against drought (Markesteijn et al. , 2011) and is a typical trait associated with higher survival in drier conditions (Martínez-Garza, Bongers and Poorter, 2013) suggesting that this group may be more tolerant of water deficit than the other two groups. In a future drier world paramo species in PFT-3 may have a higher probability of surviving than species in the other PFTs, especially those in the groups of rosettes. On the other hand, based on the Dim. 2 that separates species according to their seed dispersal strategy, shrubs in the PFT-3 are mainly dispersed by endo-zoochory and could migrate longer distances perhaps being more successful in following their climate niche as environmental conditions change, while forbs and rosettes in PFT-1 and 2 whose seeds are dispersed mainly by wind, gravity or water might be too slow to migrate upward.
High mountain regions are warming faster than the lowlands and could be more sensitive to climate change than other ecosystems at the same latitudes (Diaz and Bradley, 1997; Liu and Chen, 2000; Rangwala, Sinsky and Miller, 2013; Ning and Bradley, 2014). Therefore, predicting the responses to climate change of mountain ecosystems has become a major priority. For the tropical high Andes, projections suggest a 3.0 ± 1.5°C rise in the next century (Buytaert, Cuesta and Tobon, 2011), and a decreasing precipitation trend in Northern Colombia and Venezuela, but a small increase in precipitation for Ecuador and Southern Colombia in the next hundred years (Urrutia and Vuille, 2009; Buytaert, Cuesta and Tobon, 2011). While most paramo plants can deal with somewhat higher temperatures (Rada et al. , 1992; Körner, 2003; Rada, Briceño and Azocar, 2008; Leon-Garcia and Lasso, 2019), their response to water deficit could differ greatly among groups and future changes in water availability could change the plant community. The combination of lower precipitation, as projected for the Northern Andes, and higher temperatures, which may increase air evaporative demands and reduce soil water content, could greatly affect plant water relations and gas exchange. For the drier Venezuelan paramo, where many grasses and shrubs are more tolerant of water deficit, the shrub-herbaceous plant community is likely to show some degree of resilience (Llambí and Rada, 2019; Rada, Azócar and García-Núñez, 2019). However, the iconic giant rosettes, common in many of the paramos of the Northern Andes, could suffer in a drier world given their low capacity to tolerate water deficit (Azócar, Rada and García-Núñez, 2000). These giant rosettes are nurse plants with strong facilitation effects and are considered key species in the maintenance of paramo plant diversity (Mora, Llambí and Ramírez, 2019). Additionally, their rosette-like arrangement of heavily pubescent leaves, sometimes forming a basket, facilitates water interception from rain and fog which is then slowly released to the soil, helping in one of the key ecosystem services of the paramo, regulation and provision of water. A paramo without the PFT-2 made by rosettes would definitively lose some of its functionality.
In a future drier paramo the species of shrubs belonging to PFT-3 are the species with a highest chance of withstanding the drier conditions given their higher tolerance to water deficit, and high LDMC values. Yet, even in this group the leaf water potential values were never too negative, indicating that all species, even the more tolerant ones, are probably living on the edge of their safety margin. In previous work in this same paramo, we found that two shrubs species in the genusHypericum were highly vulnerable to xylem cavitation; they lost 50% of water conductivity at values as high as -1.50 MPa, very close to the lowest leaf water potential measured in the dry season (Ayarza, Garzón-López and Lasso, 2018), an indication of a very narrow safety margin. How changes on CO2 atmospheric concentrations, which has been steadily rising (IPCC, 2014) will interact with the expected changes in water availability in the paramo is unclear, but based on what is known for other ecosystems, an increase in CO2 is likely to mitigate drought stress if plants can reduce their stomatal conductance (g s), improve their water use efficiency and increase photosynthesis (Bazzaz, 1990; Körner and Diemer, 1994; Long et al. , 2004). More physiological studies on the vulnerability of water deficit of paramo plants and the interaction with CO2 are urgently needed given that water deficit is likely to be one of the threatening environmental driver in a changing paramo. Using this PFT classification we can also guide drought experiments to evaluate whether this classification and set of traits are indeed good indicators of species responses to water deficit and vulnerability to future changes.
Paramos are well known for their role regulating and provisioning water to major cities in South America (Carrillo-Rojas et al. , 2016). They are also important carbon reservoirs in tropical mountains (Hofstede, 1999). There is a growing consensus that the impact of species on ecosystem properties is partially mediated by the traits of their component species (Chapin et al. , 2000; Díaz and Cabido, 2001; Kazakou et al. , 2006) and here we can start discussing how these paramo species from different taxonomic groups might share similar roles in ecosystem functionality (Hubbell, 2005). For example, LDMC and C:N ratios appear to be crucial traits of living leaves that influence the quality of the litter produced, and therefore their ‘after‐life effects’ on ecosystem properties (Kazakou et al. , 2006). Species belonging to PFT-1 with tender leaves with high N and P content and low LDMC are probably fast-growing species whose dead leaves will decompose rapidly and help in the rapid turnover of carbon and nutrients. Species in PFT-3 with high LDMC and species in PFT-2 with thicker and tougher leaves and higher C:N ratios decompose slowly, decreasing the liberation of carbon and nutrients to the system, two processes that are important for the role of paramos as carbon reservoirs. In terms of their physiological strategies and their putative role in sequestering carbon, species of forbs and shrubs in PFT-1 show an acquisitive strategy, presenting the highest leaf nitrogen and phosphorus content and slightly higher photosynthetic rate (although not statistically significant), while the rosettes in PFT-2 behave more as conservative species (Reichet al. , 2003; Diaz et al. , 2004; Wright et al. , 2004). Photosynthesis is the basis for vegetation growth, but several parameters related to carboxylation efficiency (e.g. Vcmax) at this high elevation are still unknown and are necessary to accurately simulate gross primary production (GPP) and the role of the paramo on the carbon cycle under climate change using Earth System Model predictions. Here we only measured the light-saturated photosynthetic rate and the Ci/Ca ratio. However, they both consistently showed no difference among growth forms or PFTs, supporting the optimization of photosynthetic traits hypothesis (Wang et al. , 2016) which suggests that the incorporation of the paramo taxa into global vegetation models (Woodward and Cramer, 1996) could be relatively simplified despite the immense species diversity of the paramo.
Although there is a widespread consensus that growth forms are related to functional adaptations, and that plant traits should reveal the response of these plants to biotic and abiotic stresses in the ecosystem (Westoby et al. , 2002; Reich et al. , 2003; Kattge et al. , 2011), our data suggest some caution in classifying species into PFTs based only on growth forms, especially in the case of shrubs. Clearly shrubs are a complex group that varies from paramo to paramo and even within the same paramo, as our data and those from the Venezuelan paramo indicate. Shrubs presented the largest variance for many traits, suggesting a greater capacity to respond to a broader range of conditions. More growth forms should be analyzed; especially grasses, and seedlings and seed responses need to be evaluated because they could differ from their adult counterparts. This data set represents a baseline to start discussing how species from different taxonomic groups can perhaps share similar roles in the paramo functionality and how much they differ in their strategies to acquire resources and withstand environmental stress. In our study site in the mid paramo, dominated by grasses, shrubs, forbs, and giant rosettes, we did find that rosettes and forbs group into their own functional groups, but shrubs belong to different groups. As in previous screening of functional diversity in the paramos (Cárdenas-Arévalo and Vargas-Ríos, 2008; Rada, Azócar and García-Núñez, 2019), we also found high redundancy and apparent limited physiological response despite the extremely high plant diversity typical of the paramo.
CONCLUSIONS
Our results add to a series of studies, which suggest that the paramo, despite being very diverse, has few functional types, and high redundancy, where many species share physiological strategies and traits. However, a careful examination of shrub species and other growth forms such as grasses, sedges and cushion plants is indispensable to fully understand the diversity of functional strategies in the paramo. Future studies should expose species from the different PFTs to drier and warmer conditions to understand the differences among PFTs in terms of vulnerability to environmental changes. We also need to better relate PFTs and functional traits to properties and services in the paramo ecosystem like water supply and carbon storage.