## Abstract

Background

Tropical montane cloud forests (TMCFs) are characterized by a unique set of biological and hydroclimatic features, including frequent and/or persistent fog, cool temperatures, and high biodiversity and endemism. These forests are one of the most vulnerable ecosystems to climate change given their small geographic range, high endemism and dependence on a rare microclimatic envelope. The frequency of atmospheric water deficits for some TMCFs is likely to increase in the future, but the consequences for the integrity and distribution of these ecosystems are uncertain. In order to investigate plant and ecosystem responses to climate change, we need to know how TMCF species function in response to current climate, which factors shape function and ecology most and how these will change into the future.

Scope

This review focuses on recent advances in ecophysiological research of TMCF plants to establish a link between TMCF hydrometeorological conditions and vegetation distribution, functioning and survival. The hydraulic characteristics of TMCF trees are discussed, together with the prevalence and ecological consequences of foliar uptake of fog water (FWU) in TMCFs, a key process that allows efficient acquisition of water during cloud immersion periods, minimizing water deficits and favouring survival of species prone to drought-induced hydraulic failure.

Conclusions

Fog occurrence is the single most important microclimatic feature affecting the distribution and function of TMCF plants. Plants in TMCFs are very vulnerable to drought (possessing a small hydraulic safety margin), and the presence of fog and FWU minimizes the occurrence of tree water deficits and thus favours the survival of TMCF trees where such deficits may occur. Characterizing the interplay between microclimatic dynamics and plant water relations is key to foster more realistic projections about climate change effects on TMCF functioning and distribution.

## INTRODUCTION

The climatic conditions associated with elevation in tropical landscapes favour the occurrence of a unique and endangered ecosystem known as tropical montane cloud forests (TMCFs). Despite the occurrence of TMCFs in a wide range of climatic envelopes (Jarvis and Mulligan, 2010), the main common climatic attribute for every TMCF is frequent and persistent cloud immersion (i.e. fog; Scholl et al., 2010; Bruijnzeel et al., 2011). Fog frequency and intensity is an important factor determining several structural features of TMCFs (Grubb and Whitmore, 1966; Bruijnzeel and Veneklaas, 1998; Bruijnzeel and Hamilton, 2000; Bruijnzeel, 2001). As a general rule, there is an increase in epiphyte cover and decrease in tree height, canopy stratification and leaf area index in high-altitude TMCFs (also called upper montane cloud forests). TMCFs located at lower altitudes (lower montane cloud forests) are closer structurally to lowland tropical forests (Bruijnzeel and Hamilton, 2000; Bruijnzeel, 2001; Bruijnzeel et al., 2011).

The climatic and structural characteristics of TMCFs are widely assumed to be responsible for some of the ecosystem services provided by TMCFs. The environments of TMCFs are thought to increase streamflow volume, not only because of the additional inputs of cloud water interception (CWI), but also because of the low average atmospheric demand and thus low evapotranspiration, caused by the frequent cloud immersion (Bruijnzeel et al., 2011). Water quality may also be improved by the role of the TMCF cover in reducing soil erosion and landslides compared with other land uses (Sidle et al., 2006). These ecosystem services might be extremely valuable in some regions in which significant populations occur downslope and downstream of cloud forests, and are sometimes considered as a basis for TMCF conservation programmes through ‘payment for ecosystem services’ schemes (Bruijnzeel et al., 2011).

Cloud forests are also extremely valuable from a biological conservation point of view; the uniqueness of TMCF environments is also reflected in their high biodiversity and endemism levels (Bruijnzeel et al., 2010a, b). These ecosystems have a unique floristic composition, significantly distinct from that of lowland tropical forest (Grubb and Whitmore, 1966; Bertoncello et al., 2011). Neotropical TMCFs present an abundance of temperate-climate taxa, such as Podocarpus, Alnus, Drimys, Weinmannia and Magnoliaceae (Webster, 1995; Bertoncello et al., 2011). Based on the disjunct distribution of these taxa in tropical landscapes and palinological records, some authors suggest that the modern floristic composition and distribution of Neotropical TMCFs could be explained by Pleistocene climatic fluctuations, causing expansions and retractions in vegetation (Webster, 1995; Meireles, 2003; Bertoncello et al., 2011) and the reconnection between North and South America during the Pliocene, which allowed the migration of Andean and cordilleran taxa between north and south (Webster, 1995). The relatively low endemism at species level, despite high generic endemism, suggests recent and rapid speciation in TMCFs (Webster, 1995).

Various assessments of the distribution of TMCFs exist. The most comprehensive assessment of the distribution of cloud forests throughout the tropics is that compiled under the auspices of UNEP–WCMC by Aldrich et al. (1997). This is a database comprising >560 point observations distributed throughout the tropics and representing areas that have been defined as cloud forests in the literature or by local experts. These point observations have been used to help develop spatial assessments of cloud forest distribution on the basis of nationally or regionally defined elevational bands and remotely sensed forest cover assessments (Bubb et al., 2004; Scatena et al., 2010). The derived total cover of TMCFs was estimated to be in the order of 215 000 km2 (1·4 % of the total area of all tropical forests). However, TMCFs are defined by the frequency and persistence of cloud cover, not by elevation, and Jarvis and Mulligan (2011) stress the very wide range of climatic and landscape situations (temperature, rainfall, altitude, distance to sea and mountain size) represented by the >560 observed UNEP–WCMC cloud forest sites. Because this climatic variability is not just controlled by elevation, elevation-based approaches to estimate cloud forest distribution will be able to indicate the major cloud forest areas but they are not likely to identify all cloud-affected forests and may thus represent an underestimate of the true cloud forest distribution and extent.

The cloud frequency-based pan-tropical assessment of Mulligan (2010) models the distributions of cloud forest hydroclimatically to define the distribution of hydroclimatic cloud-affected forests (CAFs) rather than elevationally or ecologically defined TMCFs (Fig. 1). The most affected CAFs will have ecological adaptations that are characteristic of TMCFs (ecologically or elevationally defined), but lesser CAFs may still be hydrologically and ecologically distinct from forests that are not cloud affected but might not be considered as cloud forest structurally or ecologically. CAFs represent some 14·2 % of all tropical forests and cover an area of 2·21 Mkm2 between 23·5°N and 35°S (Mulligan, 2010).

Fig. 1.

Global distribution of cloud-affected forests (CAFs) defined hydroclimatically (Mulligan, 2010) in South-east Asia and Oceania (A), Paleotropics (B) and Neotropics (C). Areas with >40 % tree cover are shown; the darkest shades are 100 % tree cover.

Fig. 1.

Global distribution of cloud-affected forests (CAFs) defined hydroclimatically (Mulligan, 2010) in South-east Asia and Oceania (A), Paleotropics (B) and Neotropics (C). Areas with >40 % tree cover are shown; the darkest shades are 100 % tree cover.

The archipelagic distribution of TMCFs (Luna-Vega et al., 2001) and the relationship between altitude and TMCF structure and composition (Grubb and Whitmore, 1966; Bruijnzeel and Hamilton, 2000; Bruijnzeel, 2001; Bertoncello et al., 2011; Bruijnzeel et al., 2011) raises the question of which ecophysiological traits TMCF plants possess that allow and restrict their distributions to these specific hydroclimatic conditions. Addressing this question will help provide a mechanistic basis to investigate how these ecosystems will respond to the climatic changes projected to affect tropical montane regions.

Temperature projections of general circulation models (GCMs) agree reasonably well that tropical mountains will see warming over the next decades. Some models project an increase in the height of cloud formation (‘cloud uplift’) and higher evapotranspiration in tropical montane regions as a consequence of increasing earth surface temperatures (Still et al., 1999). These changes may affect TMCF structure and functioning in a number of ways, from drought-induced mortality of some tree species (Lowry et al., 1973; Werner, 1988) to an upward shift in lowland fauna and flora and invasion of pre-montane and lowland tropical species (Pounds et al., 1999). There is much less agreement between GCMs concerning the projected distribution of rainfall in tropical mountains (Mulligan et al., 2011), and different GCMs disagree in both the magnitude and direction of change of rainfall at the regional scale (Bruijnzeel et al., 2011). Given the spatial complexity of climate in general and rainfall in particular in tropical mountains, the local scale impacts of these rainfall changes are impossible to project (Oliveira et al., 2014). Given their limited geographic extent, island isolation by elevation and surrounding land use change, and strong dependence on a unique set of climate characteristics, it is clear that changes in rainfall and temperature will lead to significant stress on these systems.

In this review, we intend to link TMCF unique hydrometeorological conditions with TMCF vegetation distribution, functioning and survival in current and future climates. We will do that by coupling published and new data about TMCF plant water relations, including recent advances regarding foliar water uptake (FWU; Eller et al., 2013; Goldsmith et al., 2013) and the hydraulic safety margin (Choat et al., 2012), with published and new data about current and projected TMCF microclimate.

## HYDROCLIMATIC CONDITIONS AND HYDRAULIC FUNCTIONING OF TMCF TREES

### Temporal and spatial patterns of fog occurrence in TMCFs

Mulligan (2010) calculates the lifting condensation level (LCL) for four periods of the day for each month on the basis of pan-tropical climatological data and finds very high frequencies at which LCL is at ground level (i.e. fog is possible) in the Andes and Central America, but also in Africa and to a lesser extent parts of South-east Asia. However, elevation was not a good surrogate for satellite-observed cloud frequency across the tropics. Although the minimum observed cloud frequency does increase linearly with altitude (areas close to sea level having cloud frequencies of around 30 % in the tropics), sites at a particular altitude can show a range of cloud frequencies, depending on other factors. Nevertheless, at altitudes >1400 m a.s.l., cloud frequencies are generally >65 % (Mulligan, 2010).

Jarvis and Mulligan (2011) found the climate of the UNEP–WCMC TMCF sites to be highly variable, with an average rainfall of 2000mm year–1 and an average temperature of 17·7 °C. They also found TMCFs to be wetter (rainfall being 184 mm year–1 higher on average), cooler (by 4·2 °C on average) and less seasonally variable than the average for all montane forests (defined as all tropical forests at >500 m elevation). These global generalizations hide significant variability within and between sites.

Fog tends to occur much more frequently in the afternoon and night (Mulligan, 2010) and may persist through the dry season when rainfall is low or zero. This may be important hydrologically and ecophysiologically in seasonally dry environments (Bruijnzeel et al., 2011). Observations of cloud frequency (2001–2006) based on the MODIS cloud climatology developed by Mulligan (2010) for CAF areas in Colombia (forest cover >40 %) show an area-average frequency of 0·66. Rainfall extracted from WorldClim for the same CAF areas and months show 179 mm month–1 (Table 1). Diurnality of fog frequency for Colombian TMCFs defined using the elevational limits of Bubb et al. (2004), CAFs defined by Mulligan (2010) and all land in Colombia is shown in Table 2. Clearly CAFs do not have significantly greater cloud frequency than all land in Colombia except in the evening – whereas the elevationally defined TMCFs have very low observed cloud frequency at this time. High cloud frequency during the day will lead to a lower incident solar radiation and photosynthetically active radiation (PAR) loads, with an increased diffuse fraction of light radiation (Letts and Mulligan, 2005; Mercado et al., 2009), whereas high night-time cloud frequency will tend to reduce outgoing long-wave radiation and thus daily temperature range. In contrast to the pan-tropical mean, for Colombia the mean annual rainfall for CAFs and TMCFs is lower than for all land, though the monthly minimum for CAFs (97 mm) is higher than the minimum for all land (75 mm). Mean annual temperature for CAFs in Colombia is 18 °C (lower than all land at 24 °C) but not as low as for TMCFs at 13·25 °C. Jarvis and Mulligan (2011) show that rainfall seasonality is highly variable between TMCF sites, with most showing low seasonality but some having a strong seasonality of rainfall.

Table 1.

Climatic characteristics of CAFs, TMCFs and all land in Colombia

Variable Area Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mean
Cloud frequency (fraction) TMCFs* 0·36 0·51 0·63 0·76 0·79 0·74 0·72 0·66 0·77 0·74 0·62 0·57 0·66
CAFs 0·39 0·53 0·63 0·78 0·79 0·74 0·74 0·66 0·76 0·76 0·63 0·58 0·67
All land 0·36 0·53 0·65 0·78 0·79 0·75 0·74 0·69 0·76 0·75 0·63 0·58 0·67
Rainfall (mm h–1TMCFs* 75 86 110 180 180 140 130 120 130 190 160 110 134·25
CAFs 97 110 140 230 240 200 170 170 190 250 210 140 178·92
All land 91 110 140 250 300 280 260 250 240 280 220 140 213·42
Temperature (°C) TMCF* 13 13 14 14 14 13 13 13 13 13 13 13 13·25
CAFs 18 19 19 19 19 18 18 18 18 18 18 18 18·33
All land 24 24 25 24 24 24 23 24 24 24 24 24 24·00
Variable Area Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mean
Cloud frequency (fraction) TMCFs* 0·36 0·51 0·63 0·76 0·79 0·74 0·72 0·66 0·77 0·74 0·62 0·57 0·66
CAFs 0·39 0·53 0·63 0·78 0·79 0·74 0·74 0·66 0·76 0·76 0·63 0·58 0·67
All land 0·36 0·53 0·65 0·78 0·79 0·75 0·74 0·69 0·76 0·75 0·63 0·58 0·67
Rainfall (mm h–1TMCFs* 75 86 110 180 180 140 130 120 130 190 160 110 134·25
CAFs 97 110 140 230 240 200 170 170 190 250 210 140 178·92
All land 91 110 140 250 300 280 260 250 240 280 220 140 213·42
Temperature (°C) TMCF* 13 13 14 14 14 13 13 13 13 13 13 13 13·25
CAFs 18 19 19 19 19 18 18 18 18 18 18 18 18·33
All land 24 24 25 24 24 24 23 24 24 24 24 24 24·00

*Bubb et al. (2004); 2000–3500 m a.s.l.

Table 2.

Diurnality of satellite observed cloud frequency (2000–2006) for CAFs, TMCF and all land in Colombia

Local time TMCFs* CAFs All land
0600–1200 0·62 0·66 0·67
1200–1800 0·78 0·75 0·75
1800–2400 0·56 0·8 0·76
2400–0600 0·62 0·6 0·62
Local time TMCFs* CAFs All land
0600–1200 0·62 0·66 0·67
1200–1800 0·78 0·75 0·75
1800–2400 0·56 0·8 0·76
2400–0600 0·62 0·6 0·62

*Bubb et al. (2004); 2000–3500 m a.s.l.

Fog impacts the solar radiation, temperature and precipitation mean and seasonal behaviour of TMCFs, but perhaps the key component of TMCF climate of relevance to climate change studies is the altitudinally and topographically controlled spatial variability of climate, which means that cloud forests occur over highly restricted ranges with sharp climatic gradients. Table 3 shows change in temperature, precipitation and cloud frequency for all land, TMCFs and CAFs (>40 % tree cover) in Colombia and indicates that, though gradients of cloud frequency are only slightly steeper in TMCFs and CAFs compared with all land, gradients of rainfall and temperature are much steeper and it is these gradients that create sensitivity to climate change in cloud forest ecosystems since these gradients create barriers to dispersal and migration as cloud forest climates change.

Table 3.

Spatial variability in the climate characteristics of TMCFs, CAFs and all land in Colombia expressed as the mean gradient for each variable in each zone

Variable Area Mean gradient (units per km)
Cloud frequency (fraction) TMCFs* 0·012
CAFs 0·012
All land 0·011
Rainfall (mm h–1TMCFs* 90
CAFs 110
All land 46
Temperature (°C) TMCFs* 0·9
CAFs 0·8
All land 0·3
Variable Area Mean gradient (units per km)
Cloud frequency (fraction) TMCFs* 0·012
CAFs 0·012
All land 0·011
Rainfall (mm h–1TMCFs* 90
CAFs 110
All land 46
Temperature (°C) TMCFs* 0·9
CAFs 0·8
All land 0·3

*Bubb et al. (2004); 2000–3500 m a.s.l.

### Water use patterns of TMCF trees

The linkages between the highly variable hydrometeorological conditions in TMCFs and vegetation water use remain poorly explored. Though there are a paucity of studies quantifying tree transpiration in TMCFs compared with other systems, Bruijnzeel et al. (2011) are able to show a general negative relationship between TMCF vegetation water use and altitude (Bruijnzeel et al., 2011). Forests located at higher altitudes are more affected by fog (upper montane cloud forests and elfin cloud forests) and transpire less (380·4 ± 31·8 mm year–1) than lower montane cloud forests (646 ± 38·8 mm year–1) and lowland evergreen rain forests (1004 ± 81·6 mm year–1). This negative relationship between vegetation water use and altitude could be attributed mostly to increased cloud cover (and thus reduced evaporative demand) at higher altitudes (Zotz et al., 1998) as well as reduced leaf area index (Bruijnzeel et al., 2011). Cavelier (1996) proposed that hydraulic inefficiency could constrain TMCF tree transpiration, but several studies showed that peak transpiration rates of TMCF trees are comparable with those of lowland forests (Zotz et al., 1998; Feild and Holbrook, 2000; Santiago et al., 2010). Santiago et al. (2010) even showed that xylem area per unit of leaf area increased with altitude in the Hawaiian tree species Metrosideros polymorpha.

The TMCFs located at higher altitudes are usually exposed to more persistent fog events (Grubb and Whitmore, 1966; Jarvis and Mulligan, 2011), a microclimatic condition that affects tree water relations through transpiration suppression and by the addition of a water subsidy to the ecosystem (Fig. 2). Transpiration suppression caused by fog has been described in several fog-affected ecosystems (Burgess and Dawson, 2004; Reinhardt and Smith, 2008; Limm et al., 2009), including TMCFs (Gotsch et al., 2014; C. B. Eller et al., unpubl. data). The mechanism behind this suppression is probably the decrease in atmospheric vapour pressure deficit (VPD) and PAR associated with fog events (Reinhardt and Smith, 2008), which decreases the driving gradient for water loss by the vegetation. The formation of a water film on leaves also limits gas exchange and contributes to transpiration suppression (Smith and McClean, 1989; Brewer and Smith, 1997; Letts and Mulligan, 2005). Moreover, high-altitude TMCFs are subjected to lower mean air temperatures when compared with lowland forests (Bruijnzeel et al., 2011; Jarvis and Mulligan, 2011), which leads to lower VPD and, consequently, lower plant transpiration rates.

Fig. 2.

Scenarios illustrating the direction and magnitude of water fluxes in tropical montane cloud forests (TMCFs) under contrasting micrometeorological conditions. In scenario (A), clear days and nights, TMCF trees lose water to the atmosphere by transpiration (E). In scenarios (B) and (C), leaf-wetting events suppress transpiration of TMCF trees and provide additional water supply to the vegetation by cloud water interception (CWI), which is the water intercepted by the plant aerial tissue that then drips to the soil, and by foliar water uptake (FWU) that is the water directly intercepted and absorbed by plant leaves which may be redistributed downwards through the plant xylem to the soil (see Eller et al., 2013). The magnitude of FWU, CWI and water drip to the soil will depend on: (1) the duration and magnitude of fog events; (2) canopy water storage capacity; and (3) atmosphere–soil water potential gradient (WPG). In scenario (B), fog events of high magnitude and long duration saturate canopy water storage capacity and increase CWI, causing an increase in soil water potential and a decrease in FWU. In scenario (C), hydrological inputs of low magnitude and/or duration wet the canopy but not the soil, increasing the WPG and the magnitude of FWU. However, during the wet season or in very humid TMCFs, when the soil has high water potential, the FWU should be minor regardless of fog event intensity, because of the small WPG. Soil water potential values are monthly means of the wettest month (–0·19 MPa) and driest month (–0·77 MPa) in a Brazilian cloud forest stand.

Fig. 2.

Scenarios illustrating the direction and magnitude of water fluxes in tropical montane cloud forests (TMCFs) under contrasting micrometeorological conditions. In scenario (A), clear days and nights, TMCF trees lose water to the atmosphere by transpiration (E). In scenarios (B) and (C), leaf-wetting events suppress transpiration of TMCF trees and provide additional water supply to the vegetation by cloud water interception (CWI), which is the water intercepted by the plant aerial tissue that then drips to the soil, and by foliar water uptake (FWU) that is the water directly intercepted and absorbed by plant leaves which may be redistributed downwards through the plant xylem to the soil (see Eller et al., 2013). The magnitude of FWU, CWI and water drip to the soil will depend on: (1) the duration and magnitude of fog events; (2) canopy water storage capacity; and (3) atmosphere–soil water potential gradient (WPG). In scenario (B), fog events of high magnitude and long duration saturate canopy water storage capacity and increase CWI, causing an increase in soil water potential and a decrease in FWU. In scenario (C), hydrological inputs of low magnitude and/or duration wet the canopy but not the soil, increasing the WPG and the magnitude of FWU. However, during the wet season or in very humid TMCFs, when the soil has high water potential, the FWU should be minor regardless of fog event intensity, because of the small WPG. Soil water potential values are monthly means of the wettest month (–0·19 MPa) and driest month (–0·77 MPa) in a Brazilian cloud forest stand.

Night-time transpiration is another common and important component of tree and ecosystem water balance in TMCFs (Dawson et al., 2007). The few studies of night-time transpiration in TMCFs show moderate to very high water losses at night (Feild and Holbrook, 2000; Rosado et al., 2012; Gotsch et al., 2014). The functional meaning of night-time transpiration is not completely clear, but it is often suggested that it can contribute to nutrient acquisition (Scholz et al., 2007; Snyder et al., 2008). Nocturnal sap flow in TMCF trees during drier nights could compensate for the lack of nutrient acquisition during periods in which transpiration is suppressed by leaf-wetting events.

Soil water conditions might pose additional constraints on the water use of TMCF vegetation. Extreme conditions, such as water logging, constrain plant transpiration in some TMCFs because of poorly developed root systems or lower leaf area of trees inhabiting anoxic soils (Jane and Green, 1985; Santiago et al., 2000). Soil water deficits, documented in seasonally dry TMCF areas (Jarvis and Mulligan, 2011), can cause a decrease in tree crown conductance and constrain plant transpiration (Kumagai et al., 2004; 2005; Chu et al., 2014).

Stomatal behaviour of TMCF trees might be quite conservative, closing in response to relatively low VPD (Jane and Green, 1985), leading to the inhibitory effects of high VPD (>1–1·2 kPa) on tree transpiration even under non-limiting soil water conditions (Motzer, 2005). This type of stomatal behaviour is usually associated with plants vulnerable to hydraulic failure (McDowell et al., 2008). Despite the paucity of tree hydraulic data for TMCFs, Santiago et al. (2000) demonstrated that M. polymorpha trees from TMCFs are more susceptible to xylem cavitation than lowland forest trees. Drimys brasiliensis, one of the most abundant and ubiquitous tree species in Brazilian TMCFs (Bertoncello et al., 2011), also has a hydraulic system that is a very vulnerable to drought, losing 50 % of its hydraulic conductivity at –1·56 MPa (Fig. 3), a very high value when compared with the average –2·6 MPa for tropical forests (Choat et al., 2012). In addition, this species has a very narrow xylem hydraulic safety margin, indicating that this species operates close to the steepest point of its xylem vulnerability curve and is therefore very prone to catastrophic embolism (Fig. 3). These results support the view that TMCF trees are particularly vulnerable to droughts and might depend on alternative water sources, such as cloud water, to avoid hydraulic failure.

Fig. 3.

Embolism vulnerability curve showing loss of hydraulic conductivity (PLC, %) as a function of xylem water potential (Ψx, MPa) for branches of Drimys brasiliensis (Winteraceae), a dominant species in Brazilian tropical montane cloud forests. Ψ50 (–1·55 MPa) and Ψ88 are the xylem water potential inducing 50 and 88 % embolism, respectively. Ψmin (–1·54 MPa) is the minimum xylem water potential measured in the field during 24 months. Ψf is the increase in Ψmin due to fog occurrence. The difference between Ψ50 and Ψmin (vertical red bar) represents the ‘safety margin’ that the plant operates in the driest conditions, which is 0·01 MPa. The blue arrow represents the increase in leaf water potential and hydraulic safety margin after fog exposure and foliar water uptake (FWU) (from Eller et al., 2013). The curve was fitted using an exponential sigmoidal equation: PLC = 100/{1 – exp[a(ΨxΨ50)]}, where a is the slope of the curve. The R2 value for the fit (–0·72) was obtained with linear regression of the transformed data (Pammenter and Van der Willigen, 1998).

Fig. 3.

Embolism vulnerability curve showing loss of hydraulic conductivity (PLC, %) as a function of xylem water potential (Ψx, MPa) for branches of Drimys brasiliensis (Winteraceae), a dominant species in Brazilian tropical montane cloud forests. Ψ50 (–1·55 MPa) and Ψ88 are the xylem water potential inducing 50 and 88 % embolism, respectively. Ψmin (–1·54 MPa) is the minimum xylem water potential measured in the field during 24 months. Ψf is the increase in Ψmin due to fog occurrence. The difference between Ψ50 and Ψmin (vertical red bar) represents the ‘safety margin’ that the plant operates in the driest conditions, which is 0·01 MPa. The blue arrow represents the increase in leaf water potential and hydraulic safety margin after fog exposure and foliar water uptake (FWU) (from Eller et al., 2013). The curve was fitted using an exponential sigmoidal equation: PLC = 100/{1 – exp[a(ΨxΨ50)]}, where a is the slope of the curve. The R2 value for the fit (–0·72) was obtained with linear regression of the transformed data (Pammenter and Van der Willigen, 1998).

## CLOUD WATER INPUTS IN TMCFs

### Cloud water interception

Cloud water interception (CWI) and its subsequent precipitation as fog drip may represent a major hydrological input to TMCFs. There is a general trend of higher altitude TMCFs presenting higher CWI values (Giambelluca and Gerold, 2011); however, the relative importance of this hydrological input varies considerably between sites because of the importance of vegetation structure and epiphytism, fog frequency, fog water content, topographic exposure, wind direction and wind speed (Bruijnzeel et al., 2011). Holwerda (2010) found CWI values as low as 0·15 mm d–1 (1·7 % of the rainfall at the site) in a Mexican lower montane cloud forest, while Takahashi et al. (2010) found values as high as 3·3 mm d–1 (37 % of the rainfall at the site) in a Hawaiian lower montane cloud forest.

The hydrological relevance of CWI might vary seasonally and peak during dry seasons when rainfall inputs are lowest. Brown (1996) used throughfall data (water captured below the canopy during fog or rainfall) to investigate seasonal variation in CWI in a TMCF in Guatemala. He found that throughfall in an upper montane cloud forest, despite being relatively high during the entire year, can exceed rainfall by 147 mm during the dry season. Holder (2004) estimates that the contribution of fog precipitation to the hydrological budget in Guatemalan TMCF is 1 mm d–1 during the dry season and 0·5 mm d–1 during the rainy season. The impact of the seasonality of this water input on vegetation water use has yet to be demonstrated directly in TMCFs. Plants from redwood forests, a non-montane fog-affected ecosystem, use significantly more fog water during the dry season, when fog incidence is higher (Dawson, 1998). It is likely that fog inputs have the greatest impacts hydrologically in low rainfall, seasonally dry but frequently foggy and highly exposed forests (Bruijnzeel et al., 2011).

### Foliar water uptake

Recent studies have suggested that direct foliar water uptake is an ecophysiologically important input in TMCFs (Eller et al., 2013; Goldsmith et al., 2013). Unlike CWI, FWU is a water flux within the plant, driven by water potential gradients between sources and sinks along the soil–plant–atmosphere continuum (SPAC) and the hydraulic conductivity between SPAC compartments (Fig. 2). Simonin et al. (2009) suggested that FWU can be described using a simple equation based on Darcy's law:

$$\hbox{FWU} = k_{{\rm Atm} - {\rm L}} \Delta {\it \psi}_{{\rm Atm} - {\rm L}}$$

where kAtm−L is the efficiency of leaf water uptake, which is basically the leaf surface total conductivity to water entry, and ΔψAtm−L is the water potential (ψH2O) gradient between the inside and the outside of the leaf. During fog events, the atmospheric boundary layer surrounding leaves is saturated with moisture and the ψH2O outside the leaf should be close to zero. If leaf ψH2O is negative, FWU should be higher than 0 during most leaf-wetting events provided that the leaf surface is hydrophilic enough to allow water film formation.

With constant kAtm−L, we should expect higher FWU rates in leaves experiencing water deficits, which should be more common during periods of low soil water availability (Fig. 2B, C). Supporting this prediction, Breshears et al. (2008) shows that the effect of FWU on leaf water potential is greater when the plant is subjected to water stress. Also, sap flow reversals of higher magnitude have been observed during the dry season in D. brasiliensis at a Brazilian TMCF (C. B. Eller et al., unpubl. data). However, Burgess and Dawson (2004) observed that well-watered leaves of Sequoia sempervirens absorbed more fog water than water-stressed leaves, implying that kAtm−L is more dynamic in some species than others. Therefore, FWU could be more controlled by ΔψAtm−L in some species, while in others the kAtm−L should play a larger role.

The kAtm−L should be largely determined by leaf cuticle permeability to water and the occurrence of structures that facilitate water uptake. Despite their role in restricting molecular diffusion and thus water loss, the cuticles of leaves are known to be permeable to various molecules (Schönherr and Riederer, 1989; Schreiber and Riederer, 1996; Niederl et al., 1998). Water might diffuse through a lipophilic pathway in the cuticle, with lipophilic cutin and wax domains forming its transport path (Schreiber, 2005), or aqueous pores, which are formed by the hydration of dipoles and ionic functional groups (Schönherr, 2006). It is important to note that cuticle permeability to water might vary by several orders of magnitude between species (Kerstiens, 1996), and might be quite sensitive to changes in environmental conditions, increasing under high temperature (Schreiber, 2001) and high atmospheric humidity (Schreiber et al., 2001; Eller et al., 2013).

The occurrence of structures on leaf epidermis that facilitate water uptake can increase kAtm−L even further. Trichomes (Schreiber et al., 2001; Schönherr, 2006), hydathodes (Martin and von Willert, 2000), guard cells (Schlegel et al., 2005) and stomatal plugs (Westhoff et al., 2009) are examples of epidermal structures that might be preferential paths to FWU in some species because of differential properties of the cuticle on these structures. For example, Schönherr (2006) shows that aqueous pores are more likely to occur at the base of trichomes (Schönherr, 2006). Spatial heterogeneity on the wax content of the cuticle might also strongly affect water permeability (Schönherr and Lendzian, 1981; Becker et al., 1986).

There is also substantial empirical evidence that water might enter into leaves through stomatal apertures (Eichert et al., 2008; Burkhardt et al., 2012). Until recently, direct water entry through stomata was considered physically impossible because of high water surface tension and the morphology of stomata (Schönherr and Bukovac, 1972), but the recent hypothesis of ‘hydraulic activation of stomata’ by Burkhardt (2010) provides a possible explanation for this process. Burkhardt (2010) suggested that the deposition of hygroscopic particles around the guard cells and sub-stomatal cavity might break water surface tension and allow the formation of thin water films along the stomata, establishing a hydraulic connection between the outside surfaces of the leaf and the apoplast.

Considering the multiple water entry pathways on the leaf, it is not surprising that the occurrence of FWU in plants is a very widespread phenomenon, confirmed in >70 species (>85 % of all the studied species; Goldsmith et al. 2013). To our knowledge, all the studies investigating FWU in TMCFs found that this mechanism was present at least to some extent in the studied trees (Lima, 2010; Cassana and Dillenburg, 2012; Eller et al., 2013; Goldsmith et al., 2013). The prevalence of this mechanism in TMCFs enhances vegetation survival during seasonal droughts (Eller et al., 2013), and might affect biotic interactions, foliar traits associated with fog interception efficiency (Martorell and Ezcurra, 2007) and perhaps even hydraulic niche differentiation (Silvertown et al., 1999) and community assembly patterns. We should also note that this process adds a potentially important biotic component to TMCF water fluxes that has been ignored in hydrometeorological models until now.

### Ecological consequences of cloud immersion

Cloud immersion generally has a positive effect on leaf, plant and forest water balance (Bruijnzeel et al., 2011; Eller et al., 2013; Goldsmith et al., 2013). Even if a certain tree species is not capable of significant FWU, the suppressive effect on plant transpiration (Limm et al., 2009; Gotsch et al., 2014) and additional soil water input by fog drip can provide an important water subsidy for plants (Dawson, 1998; Liu et al., 2004). However, there are important ecological differences in the water subsidy provided by FWU and fog drip. First, part of the water of some leaf-wetting events might not even reach the soil because of the canopy storage and subsequent evaporation. Thus, species capable of FWU could benefit from the water input even of a weak leaf-wetting event. Also, the water absorbed by FWU might be redistributed inside the plant and even reach the plant rhizosphere (Eller et al., 2013). The increase in root moisture associated with this transport should cause ecological consequences to plants similar to those when water is redistributed between roots in different soil layers (hydraulic redistribution; Burgess et al., 1998; Oliveira et al., 2005a). These consequences include a decrease in branch and root embolism, an increase on root life span (Domec et al., 2004, 2006; Bauerle et al., 2008), benefits to mycorrhizal development (Querejeta et al., 2007) and even increased nutrient availability in the soil close to the roots (Dawson, 1997; Pang et al., 2013). The effects of this water transport on biotic interactions and below-ground resource competition could also be significant (Dawson, 1993; Zou et al., 2005; Prieto et al., 2011). However, there are also a number of possibly negative ecological consequences associated with FWU. If FWU occurs directly through the cuticle, as seems to be the case in some species (Schönherr, 1976, 2006; Kerstiens, 2006; Eller et al., 2013), this additional water permeability could work both ways, leading to higher cuticular conductance, which can be detrimental to plant drought resistance, mostly because of the reduced capacity to control leaf water loss during droughts (Burkhardt and Riederer, 2003). Another possible cost associated with FWU comes from the potential negative relationship between FWU and leaf water repellency (LWR) (Fig. 4; Grammatikopoulos and Manetas, 1994; Rosado et al., 2010; Rosado and Holder, 2013). FWU is thought to be favoured in plants with lower LWR (i.e. plants that stay wet for longer). Comparative studies show that LWR in cloud forests is lower than in lowland forests (Holder, 2007a), which indirectly reinforces the proposed LWR–FWU relationship (Fig. 4), now that we have evidence that FWU occurs in TMCFs (Lima, 2010; Eller et al., 2013; Goldsmith et al., 2013). Low LWR might have several negative consequences for the leaf, such as facilitation of pathogen infection (Reynolds et al., 1989; Evans et al., 1992), foliar nutrient leaching (Cape, 1996), epiphyll growth (Holder, 2007b), decrease in leaf self-cleaning properties (Barthlott and Neinhuis, 1997) and decrease in leaf gas exchange (Smith and McClean, 1989; Brewer and Smith, 1997; Letts and Mulligan, 2005).

Fig. 4.

Hypothetical relationship between foliar water uptake (FWU) and leaf water repellency (LWR) (A); higher contact angles mean a more hydrophobic leaf surface. We propose a negative non-linear relationship between FWU and LWR. More hydrophobic cuticles could reduce FWU due both to a more impermeable biochemical structure and to reduced formation of water films on its surface. The region where the curve approaches its asymptote (close to 90°) represents a hypothetical point where the leaf would dry too quickly to allow significant FWU. The arrows in (A) represent ecosystem-level consequences of this relationship: a more hydrophilic canopy should increase canopy storage, while a more hydrophobic canopy should increase dripping during leaf-wetting events. At the plant level, plants with more hydrophilic leaves (B) should have higher FWU rates than plants with more hydrophobic leaves (C).

Fig. 4.

Hypothetical relationship between foliar water uptake (FWU) and leaf water repellency (LWR) (A); higher contact angles mean a more hydrophobic leaf surface. We propose a negative non-linear relationship between FWU and LWR. More hydrophobic cuticles could reduce FWU due both to a more impermeable biochemical structure and to reduced formation of water films on its surface. The region where the curve approaches its asymptote (close to 90°) represents a hypothetical point where the leaf would dry too quickly to allow significant FWU. The arrows in (A) represent ecosystem-level consequences of this relationship: a more hydrophilic canopy should increase canopy storage, while a more hydrophobic canopy should increase dripping during leaf-wetting events. At the plant level, plants with more hydrophilic leaves (B) should have higher FWU rates than plants with more hydrophobic leaves (C).

Because of the negative impact that cloud immersion might have on leaf gas exchange of plants with low LWR, we hypothesize that the relationship LWR–FWU (Fig. 4) should influence how leaf-wetting events affect plant carbon balance. In a scenario where plant carbon assimilation is not being limited by water, cloud immersion should decrease plant instantaneous gas exchange rates due to the formation of a water film on leaves (Fig. 4; Smith and McClean, 1989; Brewer and Smith, 1997; Letts and Mulligan, 2005) and the decrease in PAR (Reinhardt and Smith, 2008). In this scenario, water gains by FWU should be minor when the water potential gradient between the atmosphere and the soil is small (Fig. 2B); thus, plants with high LWR should be favoured because they can achieve their maximum assimilation rates after the fog events more rapidly than plants with low LWR/high FWU (Fig. 4A). However, in a scenario where carbon assimilation is limited by water, the additional water subsidy provided by FWU (in comparison with plants that only depend on fog drip to use cloud water) should allow plants with higher FWU capacity to achieve higher assimilation rates after the fog events (Fig. 4B). The LWR–FWU relationship will further depend on the predominant time of occurrence of leaf-wetting events (night-time or daytime) and on the relative importance of light energy limitation compared with CO2 supply limitation during the leaf-wetting events.

Because of the many ecological benefits of FWU in water-limited conditions, we hypothesize that FWU could have been selected in seasonally dry TMCFs. The presence of fog could favour the selection of a unique strategy for dealing with soil drought in these environments. The additional water supply could allow plants capable of FWU to maintain gas exchange even during drier seasons without presenting other drought resistance-related traits such as deep roots or cavitation-resistant xylem. However, the costs associated with FWU might make it a sub-optimal strategy in very humid TMCFs. Considering that all empirical evidence of FWU in TMCFs thus far comes from seasonally dry TMCFs (Eller et al., 2013; Goldsmith et al., 2013), studies investigating the prevalence of FWU in very humid TMCFs could help us clarify if FWU occurrence can be attributed to environmental selection or whether it is just a consequence of TMCF leaves not being completely impermeable.

## IMPACT OF CLIMATE CHANGE ON TMCFs

Given the importance of their unique hydrometeorological conditions to TMCF vegetation–water relations, climate change will probably affect TMCF functioning and structure, However, the exact response of TMCF ecosystems to climate change will depend on the nature of changes in the seasonal and diurnal distribution of climatic variables and their intensity–frequency distribution, none of which can be projected well by GCMs (Oliveira et al., 2014). Still et al. (1999) showed that increases in land surface temperature might decrease the frequency of cloud immersion events in tropical mountains because of ‘cloud uplift’. Given the importance of cloud water inputs to the TMCF water budget, a decrease in the frequency of ground-level cloud (fog) (assuming that there were no significant changes in other climatic parameters such as rainfall and temperature) will probably increase TMCF evapotranspiration, vegetation drought stress and, ultimately, plant mortality.

However, cloud uplift will probably be accompanied by changes in rainfall and temperature. We can use GCMs to examine the projections for changes in these variables in a typical cloud forest situation. Using the SRES A2a climate scenarios downscaled for 17 GCMs by Ramirez-Villegas and Jarvis (2010) and cut for the tropical montane areas of Colombia, we can calculate an ensemble mean temperature and rainfall for the 2050s. We compare the ensemble mean with the mean + 1 standard deviation (mean +1 s.d.) and the mean – 1 s.d. of the ensemble (Table 4). For all land areas in Colombia, baseline mean annual temperature (MAT) is 24 °C, rising to 26 °C for the SRES A2a ensemble mean, 27 °C for mean +1 s.d. and 25 °C for mean – 1 s.d. However, precipitation for the baseline is 2500 mm year–1, rising to 7800 mm year–1 for the ensemble mean, 8900 mm year–1 for mean +1 s.d. and 6800 mm year–1 for mean – 1 s.d. The patterns are similar when analyses are confined to the areas of the country defined as CAFs by Mulligan (2010) and the areas defined as TMCFs according to the elevation limits used by Bubb et al. (2004). Based on GCM results, one could postulate that temperature increases may potentially reduce TMCF distribution, while the increases in rainfall are likely to work in the opposite direction.

Table 4.

Climate change uncertainty for TMCFs, CAFS and all land in Colombia based on a 17 GCM ensemble for a SRES A2a scenario

TMCFs*

CAFs

All land

MAT (°C) Precipitation (mm year–1MAT (°C) Precipitation (mm year–1MAT (°C) Precipitation (mm year–1
Baseline (1950–2000) 13 1600 18 2200 24 2500
Mean of 17 GCMs A2a 2050s 16 5100 21 6700 26 7800
Mean of 17 GCMs +1 s.d. A2a 2050s 17 6000 22 7500 27 8900
Mean of 17 GCMs – 1 s.d. A2a 2050s 15 4300 20 5800 25 6800
TMCFs*

CAFs

All land

MAT (°C) Precipitation (mm year–1MAT (°C) Precipitation (mm year–1MAT (°C) Precipitation (mm year–1
Baseline (1950–2000) 13 1600 18 2200 24 2500
Mean of 17 GCMs A2a 2050s 16 5100 21 6700 26 7800
Mean of 17 GCMs +1 s.d. A2a 2050s 17 6000 22 7500 27 8900
Mean of 17 GCMs – 1 s.d. A2a 2050s 15 4300 20 5800 25 6800

*Bubb et al. (2004); 2000–3500 m a.s.l.

MAT, mean annual temperature.

Combining the Still et al. (1999) ‘cloud uplift’ predictions with the GCM results presented here, we propose two broad directions of response of TMCFs to climate changes: in one scenario, the increased rainfall would not be enough to offset the drying effects of the ‘cloud uplift’ and higher temperatures. This would probably lead to drought-induced mortality of the more vulnerable species. Drought-induced tree mortality has already been documented in TMCFs during extreme droughts (Lowry et al., 1973; Werner, 1988). As mentioned previously, TMCF tree species operate close to their limit of hydraulic failure (Fig. 3). This means that the dry-season changes in soil water availability and atmospheric demand expected in this drier scenario might seriously damage the species hydraulic system and increase the chance of large-scale vegetation mortality. Plants with high FWU capacity could be particularly vulnerable to the decrease in leaf-wetting events, not only because of the key role of FWU in the maintenance of ecophysiological performance during drought (Simonin et al., 2009; Eller et al., 2013), but also because of the role that FWU might play in hydraulic failure avoidance. The increase in leaf water potential associated with FWU (average of 0·4 MPa in D. brasiliensis; Eller et al., 2013) might decrease xylem tension (Brodersen and McElrone, 2013) and increase the plant hydraulic safety margin (Fig. 3). Foliar water uptake can also be an important mechanism responsible for successful embolism repair in leaves and stems of TMCF plants (Limm et al., 2009; Simonin et al., 2009; Eller et al., 2013). Cuticular absorption could reduce the tension on the xylem enough to allow for refilling (Burgess and Dawson, 2007; Limm et al., 2009, Oliveira et al., 2005b).

In another possible scenario, the increased rainfall completely offsets the reduced cloud water contribution to TMCF water budget and increased atmospheric demand caused by higher temperatures. This would expose TMCF vegetation to a warmer, less foggy but rainier climate, similar to the climatic envelope of lowland tropical forests. This kind of change could favour the invasion of TMCFs by lower elevation species (Loope and Giambelluca, 1998; Pauchard et al., 2009). Lowland species are likely to be better competitors in this climatic scenario, due to their higher leaf area index (Bruijnzeel et al., 2011) and higher optimum temperature for photosynthesis (Allen and Ort, 2001), which leads to faster growing rates and potentially higher seed output. The invasion of TMCFs by lowland animals observed by Pounds et al. (1999) and associated with climate change could also increase dispersion rates of lowland tree species upwards into the mountains.

In both of the hypothetical scenarios, TMCF functioning and structure would be altered. In the drier scenario, drought should induce widespread mortality of less drought-resistant species, while in the warmer scenario, TMCF species could be competitively displaced by lower elevation species. More knowledge about TMCF vegetation and ecosystem functioning is necessary in order to understand more precisely to what extent each particular scenario could affect TMCFs. It is possible that different TMCFs would be more or less vulnerable to a particular scenario depending on their current climate characteristics. For example, current seasonally dry TMCFs could be less vulnerable to a drier scenario, because one could assume that species from these TMCFs are already more drought resistant.

## CONCLUDING REMARKS AND PERSPECTIVES

In this review, we propose that TMCF distribution depends strongly on the relationship between particular plant ecophysiological traits, such as FWU (increasing the hydraulic safety margin), and unique hydrometeorological conditions of TMCFs. Changes in these conditions, especially related to cloud immersion events, could drastically change the costs and benefits associated with FWU and, consequently, TMCF structure and functioning. More information about the mechanisms behind drought-induced mortality in TMCF plants is needed to clarify how drought events might affect population dynamics and community structure of TMCFs under drier climates. Despite knowing that leaf-wetting events and FWU might be important to some TMCF species during droughts (Eller et al., 2013), we do not know what proportion of TMCF species depend on FWU for survival during drought. We also do not know if a small hydraulic safety margin (Fig. 3) is a widespread trait in different TMCF species.

Potential effects of increased precipitation – which will vary highly between cloud forests in different topographic and continental settings – could either compensate for the reduction of leaf-wetting events or combine with warming to create a microclimatic envelope that could facilitate the invasion of TCMFs by lowland species. Competitive interactions between lowland forest species under different environmental conditions are also needed to illustrate TMCF vulnerability to lowland species invasion and the consequences for TMCF community structure and ecosystem-level processes.

We believe that the inclusion of non-standard climate variables (fog frequency and terrain exposure) and species functional attributes is essential for an accurate niche-based modelling of species distribution and also for more accurate predictions of ecophysiological models, especially under climate change. The FWU phenomenon in TMCFs, for example, adds an important component that needs to be taken into consideration in TMCF ecophysiological models, since it could increase the predicted contribution of fog to the ecophysiology of these ecosystems and also affect canopy water storage and re-evaporation to the atmosphere. The water subsidy provided by fog could also allow species capable of FWU to occur in places where they could not otherwise occur, if they depended only on soil water.

## ACKNOWLEDGEMENTS

This review was based, in part, on a plenary lecture presented at the ComBio2013, Perth, Australia, sponsored by the Annals of Botany. The authors would like to express their thanks to the Graduate Program in Ecology from the University of Campinas (UNICAMP), and to Hans Lambers and Tim Colmer (University of Western Autralia) for the invitation to present a lecture at ComBio2013. This work was supported by the São Paulo Research Foundation (FAPESP) (grant no. 10/17204-0), FAPESP/Microsoft Research (grant no. 11/52072-0) awarded to R.S.O., and the Higher Education Co-ordination Agency (CAPES) (scholarship to C.B.E. and P.L.B.). The cloud forest mapping was supported by the UK Department for International Development Forestry Research Programme (ZF0216).

## LITERATURE CITED

Aldrich
M
Billington
C
Edwards
M
Laidlaw
R
A global directory of tropical montane cloud forests
,
1997
Cambridge, UK
UNEP–WCMC
Allen
DJ
Ort
DR
Impacts of chilling temperatures on photosynthesis in warm-climate plants
Trends in Plant Science
,
2001
, vol.
6
(pg.
36
-
42
)
Barthlott
W
Neinhuis
C
Purity of sacred lotus, or escape from contamination in biological surfaces
Planta
,
1997
, vol.
202
(pg.
1
-
8
)
Bauerle
TL
Richards
JH
Smart
DR
Eissenstat
DM
Importance of internal hydraulic redistribution for prolonging the lifespan of roots in dry soil
Plant, Cell and Environment
,
2008
, vol.
31
(pg.
177
-
186
)
Becker
M
Kerstiens
G
Schönherr
J
Water permeability of plant cuticles: permeance, diffusion and partition coefficients
Trees
,
1986
, vol.
1
(pg.
54
-
60
)
Bertoncello
R
Yamamoto
K
Meireles
LD
Shepherd
GJ
A phytogeographic analysis of cloud forests and other forest subtypes amidst the Atlantic forests in south and southeast Brazil
Biodiversity and Conservation
,
2011
, vol.
20
(pg.
3413
-
3433
)
Breshears
DD
McDowell
NG
Goddard
KL
, et al.  .
Foliar absorption of intercepted rainfall improves woody plant water status most during drought
Ecology
,
2008
, vol.
89
(pg.
41
-
47
)
Brewer
CA
Smith
WK
Patterns of leaf surface wetness for montane and subalpine plants
Plant, Cell and Environment
,
1997
, vol.
20
(pg.
1
-
11
)
Brodersen
CR
McElrone
AJ
Maintenance of xylem network transport capacity: a review of embolism repair in vascular plants
Frontiers in Plant Science
,
2013
, vol.
4
(pg.
1
-
11
)
Brown
MB
de la Roca
I
Vallejo
A
, et al.  .
A valuation analysis of the role of cloud forests in watershed protection. Sierra de las Minas Biosphere Reserve, Guatemala and Cusuco N.P., Honduras
1996
RARE Center for Tropical Conservation
Bruijnzeel
LA
Hydrology of tropical montane cloud forest: a reassessment
Land Use and Water Resources Research
,
2001
, vol.
1
(pg.
1
-
18
)
Bruijnzeel
LA
Hamilton
LS
Decision time for cloud forests. IHP Humid Tropics Programme Series no. 13
,
2000
Paris
IHP-UNESCO, Amsterdam: IUCN-NL, and Gland: WWF International
Bruijnzeel
LA
Veneklas
E
Climatic conditions and tropical montane forest productivity: the fog has not lifted yet
Ecology
,
1998
, vol.
79
(pg.
3
-
9
)
Bruijnzeel
LA
Scatena
FN
Hamilton
LS
Tropical montane cloud forests. Science for conservation and management
2010a
Cambridge
Cambridge University Press
Bruijnzeel
LA
Kappelle
M
Mulligan
M
Scatena
FN
Bruijnzeel
LA
Scatena
FN
Hamilton
LS
Tropical montane cloud forests: state of knowledge and sustainability perspectives in a changing world
Tropical montane cloud forests. Science for conservation and management
,
2010b
Cambridge
Cambridge University Press
(pg.
691
-
740
)
Bruijnzeel
LA
Mulligan
M
Scatena
FN
Hydrometeorology of tropical montane cloud forests: emerging patterns
Hydrological Processes
,
2011
, vol.
25
(pg.
465
-
498
)
Bubb
P
May
I
Miles
L
Sayer
J
Cloud forest agenda
,
2004
Cambridge
UNEP–WCMC
Burgess
SSO
Dawson
TE
The contribution of fog to the water relations of Sequoia sempervirens (D. Don): foliar uptake and prevention of dehydration
Plant, Cell and Environment
,
2004
, vol.
27
(pg.
1023
-
1034
)
Burgess
SSO
Dawson
TE
Predicting the limits to tree height using statistical regressions of leaf traits
New Phytologist
,
2007
, vol.
174
(pg.
626
-
636
)
Burgess
SSO
MA
Turner
NC
Ong
CK
The redistribution of soil water by tree root systems
Oecologia
,
1998
, vol.
115
(pg.
306
-
311
)
Burkhardt
M
Hygroscopic particles on leaves: nutrients or desiccants?
Ecological Monographs
,
2010
, vol.
80
(pg.
369
-
399
)
Burkhardt
M
Riederer
M
Ecophysiological relevance of cuticular transpiration of deciduous and evergreen plants in relation to stomatal closure and leaf water potential
Journal of Experimental Botany
,
2003
, vol.
54
(pg.
1941
-
1949
)
Burkhardt
J
Basi
S
Pariyar
S
Hunsche
M
Stomatal penetration by aqueous solutions – an update involving leaf surface particles
New Phytologist
,
2012
, vol.
196
(pg.
774
-
787
)
Cape
JN
Kerstiens
G
Surface wetness and pollutant deposition
Plant cuticles: an integrated approach
,
1996
Oxford
BIOS Scientific Publishers
(pg.
283
-
300
)
Cavelier
J
Tissue water relations in elfin cloud forest tree species of Serrania de Macuira, Guajira, Colombia
Trees
,
1996
, vol.
4
(pg.
155
-
163
)
Cassana
FF
Dillenburg
LR
The periodic wetting of leaves enhances water relations and growth of the long-lived conifer Araucaria angustifolia
Plant Biology
,
2012
, vol.
15
(pg.
75
-
83
)
Choat
B
Jansen
S
Brodribb
TJ
, et al.  .
Global convergence in the vulnerability of forests to drought
Nature
,
2012
, vol.
491
(pg.
752
-
755
)
Chu
H-S
Chang
S-C
Klemm
O
, et al.  .
Does canopy wetness matter? Evapotranspiration from a subtropical montane cloud forest in Taiwan
Hydrological Processes
,
2014
, vol.
28
(pg.
1190
-
1214
)
Dawson
TE
Hydraulic lift and the water use by plants: implications for water balance, perfomance and plant–plant interactions
Oecologia
,
1993
, vol.
95
(pg.
565
-
574
)
Dawson
TE
Flore
HE
Lynch
JP
Eissenstat
DM
Water loss from tree roots influences soil water and nutrient status and plant performance
Radical biology: advances and perspectives on the function of plant roots
,
1997
Rockville, MD
American Society of Plant Physiologists
(pg.
235
-
250
)
Dawson
TE
Fog in the California redwood forest: ecosystem inputs and use by plants
Oecologia
,
1998
, vol.
117
(pg.
476
-
485
)
Dawson
TE
Burgess
SS
Tu
KP
, et al.  .
Nighttime transpiration in woody plants from contrasting ecosystems
Tree Physiology
,
2007
, vol.
27
(pg.
561
-
575
)
Domec
JC
Warren
JM
Meinzer
FC
Native root xylem embolism and stomatal closure in stands of Douglas-fir and ponderosa pine: mitigation by hydraulic redistribution
Oecologia
,
2004
, vol.
141
(pg.
7
-
16
)
Domec
JC
Scholz
FG
Bucci
SJ
Meinzer
FC
Goldstein
G
Villalobos-Vega
R
Diurnal and seasonal changes in root xylem embolism in Neotropical savanna woody species: impact on stomatal control of plant water status
Plant, Cell and Environment
,
2006
, vol.
29
(pg.
26
-
35
)
Eichert
T
Kurtz
A
Steinerb
U
Goldbach
H
Size exclusion limits and lateral heterogeneity of the stomatal foliar uptake pathway for aqueous solutes and water suspended nanoparticles
Physiologia Plantarum
,
2008
, vol.
134
(pg.
151
-
160
)
Eller
CB
Lima
AL
Oliveira
RS
Foliar uptake of fog water and transport belowground alleviates drought effects in the cloud forest tree species, Drimys brasiliensis (Winteraceae)
New Phytologist
,
2013
, vol.
199
(pg.
151
-
162
)
Evans
KJ
Nyquist
WE
Latin
RX
A model based on temperature and leaf wetness duration for establishment of Alternaria leaf blight of muskmelon
Phytopathology
,
1992
, vol.
82
(pg.
890
-
895
)
Feild
TS
Holbrook
NM
Xylem sap flow and stem hydraulics of the vesselless angiosperm Drimys granadensis (Winteraceae) in a Costa Rican elfin forest
Plant, Cell and Environment
,
2000
, vol.
23
(pg.
1067
-
1077
)
Giambelluca
TW
Gerold
G
Levia
DF
Carlyle-Moses
D
Tanaka
T
Hydrology and biogeochemistry of tropical montane cloud forests
Hydrology and biogeochemistry of forest ecosystems
,
2011
New York
Springer Verlag
(pg.
221
-
259
)
Goldsmith
GR
Matzke
NJ
Dawson
TE
The incidence and implications of clouds for cloud forest plant water relations
Ecology Letters
,
2013
, vol.
16
(pg.
307
-
314
)
Gotsch
SG
Asbjornsen
H
Holwerda
F
Goldsmith
GR
Weintraub
AE
Dawson
TE
Foggy days and dry nights determine crown-level water balance in a seasonal tropical montane cloud forest
Plant, Cell and Environment
,
2014
, vol.
37
(pg.
261
-
272
)
Grammatikopoulos
G
Manetas
Y
Direct absorption of water by hairy leaves of Phlomis fruticosa and its contribution to drought avoidance
Canadian Journal of Botany
,
1994
, vol.
72
(pg.
1805
-
1811
)
Grubb
PJ
Whitmore
TC
A comparison of montane and lowland rain forest in Ecuador. II. The climate and its effects on the distribution and physiognomy of the forests
Journal of Ecology
,
1966
, vol.
54
(pg.
303
-
333
)
Holder
CD
Rainfall interception and fog precipitation in a tropical montane cloud forest of Guatemala
Forest Ecology and Management
,
2004
, vol.
190
(pg.
373
-
384
)
Holder
CD
Leaf water repellency of species in Guatemala and Colorado (USA) and its significance to forest hydrology studies
Journal of Hydrology
,
2007a
, vol.
336
(pg.
147
-
154
)
Holder
CD
Leaf water repellency as an adaptation to tropical montane cloud forest environments
Biotropica
,
2007b
, vol.
39
(pg.
767
-
770
)
Holwerda
F
Bruijnzeel
LA
Oord
AL
, et al.  .
Bruijnzeel
LA
Scatena
FN
Hamilton
LS
Fog interception in a Puerto Rican elﬁn cloud forest: a wet-canopy water budget approach
Tropical montane cloud forests. Science for conservation and management
,
2010
Cambridge
Cambridge University Press
(pg.
282
-
292
)
Jane
GT
Green
TGA
Patterns of stomatal conductance in six evergreen tree species from a New Zealand cloud forest
Botanical Gazette
,
1985
, vol.
146
(pg.
413
-
420
)
Jarvis
A
Mulligan
M
The climate of cloud forests
Hydrological Processes
,
2011
, vol.
25
(pg.
327
-
343
)
Kerstiens
G
Cuticular water permeability and its physiological significance
Journal of Experimental Botany
,
1996
, vol.
47
(pg.
1813
-
1832
)
Kerstiens
G
Water transport in plant cuticles: an update
Journal of Experimental Botany
,
2006
, vol.
57
(pg.
2493
-
2499
)
Kumagai
T
Saitoh
TM
Sato
Y
, et al.  .
Transpiration, canopy conductance and the decoupling coefﬁcient of a lowland mixed dipterocarp forest in Sarawak, Borneo: dry spell effects
Journal of Hydrology
,
2004
, vol.
287
(pg.
237
-
251
)
Kumagai
T
Saitoh
TM
Sato
Y
, et al.  .
Annual water balance and seasonality of evapotranspiration in a Bornean tropical rainforest
Agricultural and Forest Meteorology
,
2005
, vol.
128
(pg.
81
-
92
)
Letts
MG
Mulligan
M
The impact of light quality and leaf wetness on photosynthesis in north-west Andean tropical montane cloud forest
Journal of Tropical Ecology
,
2005
, vol.
21
(pg.
549
-
557
)
Lima
AL
The ecological role of fog and foliar water uptake in three woody species from Southeastern Brazilian Cloud Forest
2010
Brazil
Masters thesis, University of Campinas (UNICAMP)
Limm
E
Simonin
K
Bothman
A
Dawson
T
Foliar water uptake: a common water acquisition strategy for plants of the redwood forest
Oecologia
,
2009
, vol.
161
(pg.
449
-
459
)
Liu
W
Meng
F
Zhang
Y
Liu
Y
Li
H
Water input from fog drip in the tropical seasonal rain forest of Xishuangbanna, South-West China
Journal of Tropical Ecology
,
2004
, vol.
20
(pg.
517
-
524
)
Loope
LL
Giambelluca
TW
Vulnerability of island tropical montane cloud forests to climate change, with special reference to east Maui, Hawaii
Climatic Change
,
1998
, vol.
39
(pg.
503
-
517
)
Lowry
JB
Lee
DW
Stone
BC
Effects of drought on Mount Kinabalu
Malayan Nature Journal
,
1973
, vol.
26
(pg.
178
-
179
)
Luna-Veja
I
Morrone
JJ
Ayala
OAA
Organista
DE
Biogeographical affinities among Neotropical cloud forests
Plant Systematics and Evolution
,
2001
, vol.
228
(pg.
229
-
239
)
Martin
CE
von Willert
DJ
Leaf epidermal hydathodes and the ecophysiological consequences of foliar water uptake in species of Crassula from the Namib Desert in southern Africa
Plant Biology
,
2000
, vol.
2
(pg.
229
-
242
)
Martorell
C
Ezcurra
E
The narrow-leaf syndrome: a functional and evolutionary approach to the form of fog-harvesting rosette plants
Oecologia
,
2007
, vol.
151
(pg.
561
-
73
)
McDowell
N
Pockman
WT
Allen
CD
, et al.  .
Mechanisms of plant survival and mortality during drought: why do some plants survive while others succumb to drought?
New Phytologist
,
2008
, vol.
178
(pg.
719
-
739
)
Meireles
LD
Florística das fisionomias vegetacionais e estrutura da floresta alto-montana de Monte Verde, Serra da Mantiqueira, MG
2003
Brazil
Masters thesis, University of Campinas (UNICAMP)
LM
Bellouin
N
Sitch
S
, et al.  .
Impact of changes in diffuse radiation on the global land carbon sink
Nature
,
2009
, vol.
458
(pg.
1014
-
1017
)
Motzer
T
Micrometeorological aspects of a tropical mountain forest
Agricultural and Forest Meteorology
,
2005
, vol.
135
(pg.
230
-
240
)
Mulligan
M
Bruijnzeel
LA
Scatena
FN
Hamilton
LS
Modeling the tropics-wide extent and distribution of cloud forest and cloud forest loss, with implications for conservation priority
Tropical montane cloud forests. Science for conservation and management
,
2010
Cambridge
Cambridge University Press
(pg.
14
-
39
)
Mulligan
M
Fisher
M
Sharma
B
, et al.  .
The nature and impact of climate change in the Challenge Program on Water and Food (CPWF) basins
Water International
,
2011
, vol.
36
(pg.
96
-
124
)
Niederl
S
Kirsch
T
Riederer
M
Schreiber
L
Co-permeability of 3H-labelled water and 14C-labelled organic acids across isolated plant cuticles: investigating cuticular paths of diffusion and predicting cuticular transpiration
Plant Physiology
,
1998
, vol.
116
(pg.
117
-
123
)
Oliveira
RS
Dawson
TE
Burgess
SSO
DC
Hydraulic redistribution in three Amazonian trees
Oecologia
,
2005a
, vol.
145
(pg.
354
-
363
)
Oliveira
RS
Dawson
TE
Burgess
SSO
Evidence for direct water absorption by the shoot of the desiccation-tolerant plant Vellozia flavicans in the savannas of central Brazil
Journal of Tropical Ecology
,
2005b
, vol.
21
(pg.
585
-
588
)
Oliveira
RS
Christoffersen
BO
Barros
FV
, et al.  .
Changing precipitation regimes and the water and carbon economies of trees
2014

Theoretical and Experimental Plant Physiology.
Pang
J
Wang
Y
Lambers
H
Tibbet
M
Siddique
KHM
Ryan
MH
Commensalism in an agroecosystem: hydraulic redistribution by deep-rooted legumes improves survival of a droughted shallow-rooted legume companion
Physiologia Plantarum
,
2013
, vol.
149
(pg.
79
-
90
)
Pammenter
NW
Van der Willigen
C
A mathematical and statistical analysis of the curves illustrating vulnerability of xylem to cavitation
Tree Physiology
,
1998
, vol.
18
(pg.
589
-
593
)
Pauchard
A
Kueffer
C
Dietz
H
, et al.  .
Ain't no mountain high enough: plant invasions reaching new elevations
Frontiers in Ecology and the Environment
,
2009
, vol.
7
(pg.
479
-
486
)
Pounds
JA
Fogden
MP
Campbell
JH
Biological response to climate change on a tropical mountain
Nature
,
1999
, vol.
398
(pg.
611
-
615
)
Prieto
I
FM
Armas
C
Pugnaire
FI
The role of hydraulic lift on seedling establishment under a nurse plant species in a semi-arid environment
Perspectives in Plant Ecology, Evolution and Systematics
,
2011
, vol.
13
(pg.
181
-
187
)
Querejeta
JI
Egerton-Warburton
LM
Allen
MF
Hydraulic lift may buffer rhizosphere hyphae against the negative effects of severe soil drying in a California oak savanna
Soil Biology and Biochemistry
,
2007
, vol.
39
(pg.
409
-
417
)
Ramirez-Villegas
J
Jarvis
A
Downscaling global circulation model outputs: the delta method decision and policy analysis
,
2010
Colombia
Decision and Policy Analysis, CIAT

Working Paper No. 1
Reinhardt
K
Smith
WK
Impacts of cloud immersion on microclimate, photosynthesis and water relations of Abies fraseri (Pursh.) Poiret in a temperate mountain cloud forest
Oecologia
,
2008
, vol.
158
(pg.
229
-
238
)
Reynolds
KM
LV
Richard
DL
Ellis
MA
Splash dispersal of Phytophthora cactorum from infected strawberry fruit by simulated canopy drip
Phytopathology
,
1989
, vol.
79
(pg.
425
-
432
)
BHP
Holder
CD
The significance of leaf water repellency in ecohydrological research: a review
Ecohydrology
,
2013
, vol.
6
(pg.
150
-
161
)
BHP
Oliveira
RS
Aidar
MPM
Is leaf water repellency related to vapor pressure deficit and crown exposure in tropical forests?
Acta Oecologica
,
2010
, vol.
36
(pg.
645
-
649
)
BHP
Oliveira
RS
Joly
CA
Aidar
MPM
Burgess
SSO
Diversity in nighttime transpiration behavior of woody species of the Atlantic Rain Forest, Brazil
Agricultural Forest Meteorology
,
2012
, vol.
158–159
(pg.
13
-
20
)
Santiago
LS
Jones
T
Goldstein
G
Bruijnzeel
LA
Scatena
FN
Hamilton
JG
Juvik
JO
Bubb
P
Physiological variation in Hawaiian Metrosideros polymorpha across a range of habitats: from dry forests to cloud forests
Mountains in the mist: science for conserving and managing tropical montane cloud forests
,
2010
Honolulu
University of Hawaii Press
(pg.
456
-
464
)
Santiago
LS
Goldstein
G
Meinzer
FC
Fownes
J
Mueller-Dombois
D
Transpiration and forest structure in relation to soil waterlogging in a Hawaiian montane cloud forest
Tree Physiology
,
2000
, vol.
20
(pg.
673
-
681
)
Scatena
FN
Bruijnzeel
LA
Bubb
P
Das
S
Bruijnzeel
LA
Scatena
FN
Hamilton
LS
Setting the stage
Tropical montane cloud forests. Science for conservation and management
,
2010
Cambridge
Cambridge University Press
(pg.
3
-
13
)
Schlegel
TK
Schönherr
J
Schreiber
L
Size selectivity of aqueous pores in stomatous cuticles of Vicia faba leaves
Planta
,
2005
, vol.
221
(pg.
648
-
655
)
Scholl
M
Eugster
W
Burkard
R
Understanding the role of fog in forest hydrology: stable isotopes as tools for determining input and partitioning of cloud water in montane forests
Hydrological Processes
,
2010
, vol.
25
(pg.
353
-
366
)
Scholz
FG
Bucci
SJ
Goldstein
G
Meinzer
FC
Franco
AC
Miralles-Wilhelm
F
Removal of nutrient limitations by long-term fertilization decreases nocturnal water loss in savanna trees
Tree Physiology
,
2007
, vol.
27
(pg.
551
-
559
)
Schönherr
J
Water permeability of isolated cuticular membranes: the effect of pH and cations on diffusion, hydrodynamic permeability and size of polar pores
Planta
,
1976
, vol.
128
(pg.
113
-
126
)
Schönherr
J
Characterization of aqueous pores in plant cuticles and permeation of ionic solutes
Journal of Experimental Botany
,
2006
, vol.
57
(pg.
2471
-
2491
)
Schönherr
J
Bukovac
MJ
Penetration of stomata by liquids: dependence on surface tension, wettability, and stomatal morphology
Plant Physiology
,
1972
, vol.
49
(pg.
813
-
819
)
Schönherr
J
Lendzian
K
A simple and inexpensive method of measuring water permeability of isolated plant cuticular membranes
Zeitschrift für Pflanzenphysiologie
,
1981
, vol.
102
(pg.
321
-
327
)
Schönherr
J
Riederer
M
Foliar penetration and accumulation of organic chemicals in plant cuticles
Reviews of Enviromental Contamination and Toxicology
,
1989
, vol.
108
(pg.
1
-
70
)
Schreiber
L
Effect of temperature on cuticular transpiration of isolated cuticular membranes and intact leaf disks
Journal of Experimental Botany
,
2001
, vol.
52
(pg.
1893
-
1900
)
Schreiber
L
Polar paths of diffusion across plant cuticles: new evidence for an old hypothesis
Annals of Botany
,
2005
, vol.
95
(pg.
1069
-
1073
)
Schreiber
L
Riederer
M
Ecophysiology of cuticular transpiration: comparative investigation of cuticular water permeability of plant species from different habitats
Oecologia
,
1996
, vol.
107
(pg.
426
-
432
)
Schreiber
L
Skrabs
M
Hartmann
KD
Diamantopoulos
P
Simanova
E
Santrucek
J
Effect of humidity on cuticular water permeability of isolated cuticular membranes and leaf disks
Planta
,
2001
, vol.
214
(pg.
274
-
282
)
Sidle
RC
Ziegler
Negishi
JN
Nik
AR
Siew
R
Turkelboom
F
Erosion processes in steep terrain: truths, myths, and uncertainties related to forest management in Southeast Asia
Forest Ecology and Management
,
2006
, vol.
224
(pg.
199
-
225
)
Silvertown
J
Dodd
ME
Gowing
D
Mountford
O
Hydrologically-defined niches reveal a basis for species-richness in plant communities
Nature
,
1999
, vol.
400
(pg.
61
-
63
)
Simonin
KA
Santiago
LS
Dawson
TE
Fog interception by Sequoia sempervirens (D.Don) crowns decouples physiology from soil water deficit
Plant, Cell and Environment
,
2009
, vol.
32
(pg.
882
-
892
)
Smith
WK
McClean
TM
Adaptive relationship between leaf water repellency, stomatal distribution, and gas exchange
American Journal of Botany
,
1989
, vol.
76
(pg.
465
-
469
)
Snyder
KA
James
JJ
Richards
JH
Donovan
LA
Does hydraulic lift or nighttime transpiration facilitate nitrogen acquisition?
Plant and Soil
,
2008
, vol.
306
(pg.
159
-
166
)
Still
CJ
Foster
PN
Schneider
SH
Simulating the effects of climate change on tropical montane cloud forests
Nature
,
1999
, vol.
398
(pg.
608
-
610
)
Takahashi
M
Giambelluca
TW
Mudd
RG
, et al.  .
Rainfall partitioning and cloud water interception in native forest and invaded forest in Hawai'i Volcanoes National Park
Hydrological Processes
,
2010
, vol.
25
(pg.
448
-
464
)
Westhoff
M
Zimmermann
D
Zimmermann
G
, et al.  .
Distribution and function of epistomatal mucilage plugs
Protoplasma
,
2009
, vol.
235
(pg.
101
-
105
)
Webster
GL
Churchill
SP
Balslev
H
Forero
E
Luteyn
JL
The panorama of Neotropical cloud forests
Biodiversity and conservation of Neotropical montane forests. Neotropical Montane Forest Biodiversity and Conservation Symposium 1
,
1995
(pg.
53
-
77
New York Botanical Garden
Werner
WL
Canopy dieback in the upper montane rain forests of Sri Lanka
GeoJournal
,
1988
, vol.
17
(pg.
245
-
248
)
Zotz
G
Tyree
MT
Patiño
S
Carlton
MR
Hydraulic architecture and water use of selected species from a lower montane cloud forest in Panama
Trees
,
1998
, vol.
12
(pg.
302
-
309
)
Zou
C
Barnes
P
Archer
S
McMurtry
C
Soil moisture redistribution as a mechanism of facilitation in savanna tree–shrub clusters
Oecologia
,
2005
, vol.
145
(pg.
32
-
40
)