Transpiration in a sub-tropical ridge-top cloud forest

Journal of Hydrology 462–463 (2012) 42–52

Contents lists available at SciVerse ScienceDirect

Journal of Hydrology journal homepage: www.elsevier.com/locate/jhydrol

Transpiration in a sub-tropical ridge-top cloud forest G. García-Santos ⇑ Department of Hydrology and Geo-Environmental Sciences, VU University, Amsterdam, The Netherlands

a r t i c l e

i n f o

Article history: Available online 4 November 2011 Keywords: Sap flow Fog Laurisilva Cloud forest Transpiration Climate

s u m m a r y Laurel forests in the Canary Islands (Spain) survive where humid conditions are guaranteed throughout the year. On peaks and ridges, laurel forest gives way to mixed evergreen tree-heath/beech forest of low stature (‘‘fayal–brezal’’) that has to cope with rapidly changing light, temperature and humidity conditions due to the occurrence of intermittent sunny and foggy periods during the mostly rainless summer. These conditions are poorly understood and there is a lack of information on the interrelations between tree physiological behavior and ambient climatic and soil water conditions in fayal–brezal. In this study sap velocities were measured for 2 years in two dominant tree species (Myrica faya and Erica arborea) in a ridge-top forest in the National Park of Garajonay on the island of La Gomera. The resulted average daily stand transpiration was 1.2 ± 0.12 mm (416 mm year1). However, the narrow-leaved E. arborea exhibited higher sap velocities than the broad-leaved M. faya. Also, sap velocity increased with stem diameter in E. arborea but not in M. faya. Nocturnal flow activity was observed throughout the year and reflected ambient conditions on some occasions, and stem water storage recovery on others. Strong stomatal control in response to increases in vapor pressure deficit was seen in both species. Fog reduced sap velocity from 10% up to 90% but no consistent pattern was found. Soil water uptake during the dry summer (246 mm) was much larger than atmospheric water inputs (41 mm, rain and fog). The low moisture levels in the top 0.3 m of the soil had limited influence on transpiration rates indicating that vegetation must have had access to moisture in deeper layers. Ó 2011 Elsevier B.V. All rights reserved.

1. Introduction Laurel-dominated forest, or ‘‘Laurisilva’’, is a sub-tropical evergreen vegetation type occurring predominantly on the northern slopes of the so-called Macaronesian Islands (Azores, Canaries, Cabo Verde and Madeira) between 800 and 1300 m.a.s.l., where stratocumulus clouds transported by the trade winds produce a rather uniform climate with moderate temperature variation and high relative humidity throughout the year (Dorta, 1996). The National Park of Garajonay in central La Gomera (Canary Islands, Spain) contains the largest single representative example of this ecosystem within the Canaries (Pérez de Paz et al., 1990). The summer is mostly rainless, but humid atmospheric conditions occur frequently because of the characteristic ‘‘sea of clouds’’ (Marzol, this volume). However, the humid conditions prevailing at summit and ridge-top locations are often reduced by a lowering of the position of the thermal inversion in summer (Dorta, 1996; García-Santos and Bruijnzeel, 2011). Thus, apart from the general absence of rainfall in summer, the vegetation also has to cope with sudden light variations and rapidly changing temperature and humidity conditions. ⇑ Present address: Remote Sensing Laboratories (RSL), Department of Geography, University of Zürich, Switzerland. Tel.: +41 789 503 978. E-mail address: [email protected] 0022-1694/$ - see front matter Ó 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jhydrol.2011.08.069

The laurel-dominated forests of the slopes and valley bottoms give way to a mixed tree-heath/beech forest on the ridges (socalled ‘‘fayal–brezal’’) which is composed mainly of wax myrtle or beech (Myrica faya Ait.) and tree heath (Erica arborea L.) (Golubic, 2001; Pérez de Paz et al., 1990). There is little information on annual transpiration totals by the fayal– brezal (García-Santos et al., 2004) but studies in laurel-dominated forest on nearby Tenerife suggested annual stand transpiration totals ranging between 300 and 640 mm (Morales et al., 2003). Obtaining reliable estimates of seasonal and annual transpiration totals for this type of forest, 23% of the total surface of the island (Consejería de Obras Públicas, 2003), is essential for the assessment of the National Park’s overall water budget and its importance for regional water supply (García-Santos et al., 2004). Moreover, in view of expected changes in the archipelago’s climate (Sperling et al., 2004) baseline studies of the interactions between meteorological variables, soil moisture content, and vegetation water use are needed. Results on rainfalland cloud water-interception by this type of forest are given by García-Santos and Bruijnzeel (2011). The present work aims to determine transpiration at the stand level, and to identify possible traits developed by the vegetation in the face of potential drought stress during the summer. To achieve these objectives, sap velocities in the two dominant tree species (M. faya and E. arborea) were studied separately, as well as their relationships with meteorological variables and soil moisture levels.

G. García-Santos / Journal of Hydrology 462–463 (2012) 42–52

2. Materials and methods 2.1. Study area and experimental plot The ridge-top cloud forest studied is located at Laguna Grande in the headwater area of the Jelima catchment at 1270 m.a.s.l. (28°070 4000 N, 17°150 2600 W) in the northern central sector of the Garajonay National Park, La Gomera. The climate is humid Mediterranean with mild temperatures throughout the year (average 13.0 °C, monthly range 8.9–19.7 °C). Annual rainfall in the area is 660 mm but varies strongly between years (Arévalo et al., 2002). Most of rain falls in the autumn and winter months. Whilst fog occurs throughout the year (1000 mm annual fog), its highest incidence is observed in spring (García-Santos and Bruijnzeel, 2011). The volcanic soils on the ridges are Leptosols (FAO, 1998) developed from trachy-basaltic lava flows and dykes. The soils are shallow (0.3–0.5 m) and typically consist of an A-horizon of sandy (clay) loam texture above a stony C-horizon containing numerous angular basalt fragments in a gray clayey matix (Rodríguez et al., 2002). Roots were concentrated in the top 0.2 m but, extended into the rocky sub-soil. The forest consisted mostly of evergreen sclerophyllous tree-heath (E. arborea L.), beech (M. faya Ait.), and to a lesser extent holly (Ilex canariensis Poivet.) and laurel (Laurus azorica Seub. Franco). The trees were partly covered with epiphytes. Beech, holly and laurel are all evergreen broad-leaved trees, whereas Erica has tiny needle-like leaves (García-Santos, 2007, p. 26). An experimental plot of 300 m2 was delineated on the upper ridge above a steep slope (30–40°) of north-easterly orientation. Tree density was 1266 trees ha1 with a mean canopy height of 9 m (for trees with a diameter at breast height DBH > 0.07 m). Mean leaf area index (LAI) was 4.2 ± 1 m2 m2 (Golubic, 2001). Average DBH was 0.20 ± 0.17 m and stand basal area was 68 m2 ha1. M. faya made up the greatest proportion (57%) of stand basal area, followed by E. arborea (33%) and L. azorica (10%). 2.2. Climatic measurements Actual climatic conditions were monitored by an automatic weather station placed on top of a 9 m high scaffolding tower and extension mast (3 m) above the canopy. Relative humidity (RH,%) and air temperature (T, °C) were recorded at 2 m above the canopy (Vaisala HMP45C T-RH probe), global radiation (Rg, W m2) was measured by a pyranometer (Campbell SP1110), and wind speed (u, m s1) by a switch anemometer (Vector Instruments A100R). Occurrence (hours with fog) and intensity of fog (mm h1) were derived from recorded amounts of water collected every 15 min by a quarter-sized variant of the standard fog collector (0.25 m2) of Schemenauer and Cereceda (1994). Wind direction corrections were made to obtain a more real estimate of the effective collection surface (García-Santos and Bruijnzeel, 2011). To convert fog water volumes (L) to depth equivalents (mm), they were divided by the estimated effective cross-sectional area of the screen. Rainfall was measured with a self empting recording gage at 0.25 mm resolution (Rain-o-matic type, Pronamic). Rainfall amounts were corrected for topographic and wind effects based on Yang et al. (1998) and Sharon (1980) as described by García-Santos and Bruijnzeel (2011). All variables were sampled at 3 min intervals and recorded every 15 min by a Combilog data-logger and then transmitted by GSM modem (see García-Santos (2007) for details). 2.3. Soil water content Volumetric soil water content hv (m3 m3) was measured continuously at 0–0.15 m and 0.15–0.30 m depth with two time

43

domain reflectometry (TDR) probes (TRIME-IT/EZ) inserted horizontally. The stoniness of the sub-soil prevented the use of sensors below 0.3 m. Given the stony and volcanic character of the soil, the sensors were calibrated gravimetrically in the laboratory using undisturbed soil columns taken from the field site. Measured water contents during the calibration were consistently higher than those predicted by the widely used equation of Topp et al. (1980), as has been reported for volcanic soils elsewhere (e.g. Miyamoto et al., 2003; see García-Santos (2007) for details). Relative extractable soil water (he, m3 m3) was calculated as the ratio of actually extractable water to maximum extractable water for plants (Black, 1979):

he ¼

hv  hr hfc  hr

ð1Þ

where hfc is the soil water content at field capacity, and hr the residual soil water content. hfc was determined as the gravimetric water content after 48 h of drainage from an initially saturated undisturbed sample (n = 4) whereas hr was determined after equilibrating wetted soil aggregates (n = 15) to a pressure of 1500 kPa. 2.4. Sap velocity and scaling up to stand level Sap velocity was measured continuously between February 2003 and January 2005 in seven trees selected from the two most abundant tree species present (M. faya and E. arborea) using the heat dissipation technique of Granier (1987) (sapflow sensor SFS2 TypM, UP GmbH). Three representative individuals of M. faya (in terms of DBH) were used plus four individuals of E. arborea (Fig. 1). The thermocouples were inserted horizontally and 0.10 m apart in the outer 20 mm of the stem of each sample tree. Data were recorded at 3 min intervals and stored every 15 min in a data-logger (Combilog) and then transmitted by GSM modem. The temperature difference (DT, °C) between the heated upper needle and non-heated lower reference needle during the day, combined with the temperature difference at night (usually taken as the peak nigh-time value of DT), allows the estimation of sap

Fig. 1. Distribution and diameter at breast height (DBH, m) of the trees (n = 38) within the experimental plot at Laguna Grande ridge, La Gomera. Erica arborea L. sample tree with one pair of sapflow sensors; Myrica faya Ait. sample tree with one pair of sapflow sensors; M. faya Ait. sample tree with two pairs of sapflow sensors. (xx) Location of TDR soil water probes.

44

G. García-Santos / Journal of Hydrology 462–463 (2012) 42–52

flux density (Granier, 1987) – also called sap velocity (v, ml cm2 min1; Edwards et al., 1996) – according to:

  DT night 1 DT

v ¼ 0:714

ð2Þ

Several systematic and random errors are known to affect the estimation of sap velocity. An example of the former is poor thermal contact between the thermocouples and the stem whereas potential random errors include the occurrence of thermal gradients between the stem and the environment, natural gradients along the trunk, and possibly excessive stem heating during the night. Poor thermal contact between the thermocouples and the stem and thermal gradients between the stem and the environment were avoided by careful installation of the probes and insulating them against rainfall and temperature variations. Furthermore, in view of the generally humid and shady conditions prevailing below the canopy, and the generally mild temperatures at the study site, the effects of natural gradients or changes in stem heat storage during the day (Grime and Sinclair, 1999) were considered negligible. Although Grime and Sinclair (1999) suggested that nocturnal overheating may be avoided by reducing the heat input by 50% between 1 h after sunset and 1 h before sunrise, this was not deemed necessary because the mild temperatures during the night were not likely to produce significant changes in stem heat storage. Sensors remained installed during the experimental time as no wounding effects tree reactions were observed. Tree sapflow, also called sapflow rate or sap flux (Qt, ml min1; Edwards et al., 1996), can be computed by multiplying sap velocity v, as measured in the outer 0.02 m, times the cross-sectional area (SA, cm2) of the active xylem of the tree (with radius r, in cm) assuming that sap velocity is constant across the xylem:

Q t ¼ v  SA

ð3Þ

However, v across the xylem is often non-uniform (Cermák and Nadezhdina, 1998; Jiménez et al., 2000; Lundblad et al., 2001); therefore Qt as obtained with Eq. (3), with sensors placed only in the outer parts of the sapwood is likely to be overestimated. Jiménez et al. (2000) studied radial profiles of sap velocities in M. faya, E. arborea and L. azorica trees in Tenerife. In the curved trunks of stunted E. arborea trees – very similar to the ones present in the study area – the outer part of the stem was shown to be conductive, whereas the inner (heartwood) part was non-conductive (Fig. 2, left-hand side). In the case of M. faya, the entire cross-sectional stem area in some trees was conducting (tree type 1C according to Jiménez et al., 2000), whereas only part was conductive in others (tree type 2B; Fig. 2, right-hand side). To obtain supplementary information on the changes in sap velocity with depth

within the xylem of M. faya at the study site, an additional pair of thermocouples was installed in the second 20 mm (20–40 mm) of xylem in one of the bigger M. faya trees (DBH = 0.30 m; see Fig. 1 for location). The results were compared with the patterns obtained by Jiménez et al. (2000) and showed an average decrease of 55% in sap velocity from the outer 20 mm to the next 20 mm of xylem (2 years of hourly data). The pattern was similar to that observed by Jiménez et al. (2000) for an open-crowned type 1C M. faya tree of DBH 0.27 m. Therefore, to estimate sap velocity vs. xylem depth in the three large open-crowned M. faya trees under study, two separate regressions were derived, describing relative sap velocity (% of maximum) as a function of position within the xylem, i.e. for the inner 67% of the xylem (according to Jiménez et al., 2000) and for the remaining portion (67–100%; this study) (see García-Santos (2007) for details). Thus, sapflow per tree Qt was obtained by integrating sap velocity at 0.02 m radial depth intervals:

Qt ¼

X

v  ci  SAi

ð4Þ

i

where v (ml min1) is the measured sap velocity in the outer 0.02 m of the xylem, and ci the correction factor as a function of the radius (ri). The relationship between ci and ri (both expressed as a percentage of the maximum value) for a small-crowned E. arborea tree with curved trunk was calculated from the sap flow pattern obtained by Jiménez et al. (2000) (cf. Fig. 2, left) as:

ci ¼ 0:001r3i þ 0:20r 2i  10:37ri þ 149:24

ð5Þ

and for a large open-crowned M. faya tree (type 2B, Fig. 2 right) as:

ci ¼ 6:1  105 r 4i  1:5  102 r 3i þ 1:5r 2i  60r i þ 913

ð6Þ

In the absence of direct sap velocity measurements in L. azorica in the study plot, sapflow for this species was approximated using the relationship between L. azorica and E. arborea trees derived from Jiménez et al. (2000), in which the total sapflow in an individual of L. azorica (DBH = 0.25 m) was 0.92 times the sapflow in a similarly sized E. arborea tree (DBH = 0.19 m). Scaling up of tree sapflow in individual trees belonging to each of the three main species to the stand as a whole (300 m2) was carried out in stages. First, empirical relationships were derived per species between measured Qt (ml min1) and DBH. Next, values of Qt were obtained for all 38 individuals in the plot by inserting their DBH values into the respective equations. Daily stand transpiration (Et, mm day1) was then obtained by adding up the individual daily values and dividing by plot area (A, m2):

Fig. 2. Changes in relative sapflow per unit cross-sectional stem area, or sap velocity (% of maximum flow as measured in the outer 0.02 m) vs. depth of cross-sectional stem area, based on Jiménez et al. (2000). Left: Erica arborea L. tree (0.19 m DBH). Right: Myrica faya Ait. trees type 2B (0.37 m DBH) and type 1C (0.27 m DBH) in a laurel-dominated forest in Tenerife.

45

G. García-Santos / Journal of Hydrology 462–463 (2012) 42–52

P Et ¼

i¼1 38Q t

A

 60  24  103

ð7Þ

Two potential errors in the estimation of Qt relate to the determination of sap velocity and sapwood cross-sectional area (Giambelluca et al., 2003). Error in the latter is supposed to exert less influence on total sap flow than errors in sap velocity across the xylem because v decreases with xylem depth. Assuming that v in the outer 20 mm was not subject to the systematic and random errors described above (which were either avoided or considered negligible), the error in v may be approximated by scaling v over the entire conducting area. A priori estimates of Qt in E. arborea will have a higher uncertainty because of the stronger radial gradient in sap velocity associated with this species (Fig. 2, left). An error of ± 5% in v throughout the conducting xylem was assumed to estimate the associated sensitivity in Qt per species. The resulting standard deviations for Qt were ±6.2% for M. faya and ±19.6% for E. arborea (García-Santos, 2007). 3. Results 3.1. Diurnal patterns of sap velocity in tree-heath and beech trees Hourly sap velocity patterns (v, in mm h1 or L m2 SA h1) during two representative bright days (30 and 31 March 2003) in the absence of soil water limitation (he > 0.8) are shown for M. faya and E. arborea in Fig. 3. During these 2 days, the mean temperature was 15.4 ± 3.2 °C (range: 10.8–21.3 °C), the mean VPD was 1.26 ± 0.3 kPa (range: 0.78–1.98 kPa), and wind speeds were low (<2 m s1). For trees with similar DBH (0.22–0.24 m), the higher daily sap velocity in E. arborea compared to that in M. faya (three times higher) is striking. Similar pattern is observed at annual base (p < 0.05, T-test two-tailed test) (Fig. 3). Very small differences in v were observed between M. faya trees with DBH of 0.24 and 0.31 m, both at the 2 days example and at annual base. However, in the case of E. arborea, v increased with DBH (Fig. 3). In both species the start of sapflow activity was very variable in time. On some days, sapflow started at sunrise, whereas on other days activity began between 1 and 3 h later. Maximum values were observed around mid-day in both species, after which sapflow started to decrease until 3–4 h after dusk. Nocturnal sapflow activity was observed throughout the year in both species (a 4 days example shown in Fig. 4). On some occasions, sapflow activity continued slowly after dusk until it stopped a few hours before sunrise. On other occasions, which tended to coincide with high wind speeds (García-Santos, 2007), nocturnal patterns of v followed the evaporative demand of the atmosphere. In such cases, maximum values of v at night were about 85% lower than the peak during the preceding day-time (example in Fig. 4). In E. arborea, night-time sap velocity increased with tree DBH as was observed for day-time conditions (cf. Figs. 3 and 4). Average sap velocity during such nights represented 6% and 4.8% of day-time v for M. faya and E. arborea, respectively.

Fig. 3. Diurnal course of mean sapflow per unit of conducting sapwood area per tree (i.e. sap velocity in mm h1 or L m2 SA h1) and species Myrica faya and Erica arborea as a function of DBH (m) on 30 and 31 March 2003 (up) and annual average of sap velocity (down) in the Laguna Grande ridge-top cloud forest.  Significantly different (p < 0.05) after a T-test (two-tailed test).

M. faya (0. 34 ± 0.22 L h1) were statistically different (p < 0.05, Ttest two-tailed test). 3.3. Sapflow at the stand level

3.2. Sapflow in tree-heath and beech trees Diurnal patterns of sapflow per tree (Qt, in L h1) and species were derived using Eq. (4). Diurnal course of total sap flow (30 and 31st of March 2003) and annual mean of sap flow per tree and species (2 year data) are shown in Fig. 5. Sapflow rates per tree in both M. faya and E. arborea are clearly related to tree DBH (p < 0.05, T-test two-tailed test). Thought Qt values for E. arborea trees are higher than for M. faya trees of similar diameter (0.22– 024 m) in the example shown, the differences at annual base are not significant (p > 0.05, T-test two-tailed test). Average values for the seven trees measured for E. arborea (0.15 ± 0.10 L h1) and

Mean daily sapflow for each DBH class (intervals of 0.02 m) and tree species was calculated by combining the regressions between Qt and DBH (Fig. 6a) for small trees, Erica arborea: DBH 6 0.22 m

Q t ¼ 0:0557exp24:45  DBH

r2 ¼ 0:99

ð8Þ

Myrica faya: DBH 6 0.31 m

Q t ¼ 202:49DBH2  13:62DBH þ 1:34 r 2 ¼ 0:99 and for larger trees only (Jiménez et al., 1996), Erica arbórea (n = 2): DBH > 0.22 m

ð9Þ

46

G. García-Santos / Journal of Hydrology 462–463 (2012) 42–52

Fig. 4. Above: Radiation (- Rg, W m2) and vapor pressure deficit (-s- VPD, kPa). Below: Diurnal course of sapflow per unit of conducting sapwood area (mm h1 or L m2 SA h1) for different tree species. Dotted lines: Myrica faya sample trees, solid lines: Erica arborea. Measurements made between 26 and 29 April 2003 at the Laguna Grande ridge-top cloud forest site.

Q t ¼  3186:82  DBH3 þ 3315:53  DBH2  710:21  DBH þ 42:81 r 2 ¼ 0:99

ð10Þ

Myrica faya (n = 3): DBH > 0.31 m

Q t ¼  0:00003  DBH5 þ 0:0061  DBH4  0:58  DBH3 þ 27:46  DBH2  623:53  DBH þ 5484:83 r2 ¼ 0:99 ð11Þ with the DBH distribution data shown in Fig. 6b. Mean total Qt values per DBH class and species are shown in Fig. 6c. Daily stand transpiration totals were obtained by summing Qt values per DBH class and species. Total amounts of water transpired per species were strongly controlled by the uptake of the few biggest individuals (Fig. 6c). Overall uncertainty in average daily stand transpiration over the 2 years of observation was estimated at ±22%, estimated as the sum of the errors estimated per tree and species within the plot. Error of Qt for sampled and non-sampled M. faya trees was ±6.2% and ±12.5%, respectively whereas error for sampled and non-sampled E. arborea trees was ±19.6% and ±39%, respectively. Error for (non-sampled) L. azorica trees was ±39%.

Fig. 5. Diurnal course of total sapflow per tree (L h1) on 30 and 31 March 2003 (up) and annual average of sapflow (down) in the Laguna Grande ridge-top cloud forest, La Gomera. Left: Myrica faya, right: Erica arborea.

3.4. Contribution of the three main tree species to overall stand transpiration

3.5. Sap velocity vs. meteorological and soil water conditions

Daily stand transpiration (experimental site, in Fig. 1) (Et, mm) was obtained by dividing total daily sapflow at the stand level (L) by plot surface area (m2). Average (measured) daily Et for the first year was 1.28 ± 0.22 mm (n = 297 days; range 0.1–3.9 mm day1) vs. 1.11 ± 0.73 mm day1 for the second year (n = 262 days; range 0.04–3.9 mm day1). The relative contributions of the three dominant tree species to total Et in the first year were 48% for M. faya, 40% for E. arborea, and 12% for L. azorica. To arrive at annual transpiration totals the gaps in the sap flow record needed to be filled. The Artificial Neural Network approach, used in García-Santos (2007), was the preferred method to fill these gaps. After filling the gaps in the sap flow data annual stand transpiration totals were remarkably close at 412 and 420 mm.

Global radiation reached maximum values between May and September, although many short-term oscillations were observed due to cloud occurrence; Rg declined again in fall and winter. Fluctuations in day-time temperature were relatively modest at this elevation, with a mean value of 14.9 ± 5.0 °C. Rainfall was mainly concentrated in autumn and winter, although numerous small showers occurred during the spring; the period between May and September was typically dry, but days with fog (and therefore conditions with low to zero VPD) at the study site happened throughout the year (Fig. 7a). Sap velocity in both M. faya and E. arborea was maintained throughout the year (Fig. 7c and d). Pronounced short-term fluctuations in v were observed during conditions of atmospheric satura-

G. García-Santos / Journal of Hydrology 462–463 (2012) 42–52

47

Fig. 6. (a) Mean daily sapflow per tree (L day1) vs. DBH class as calculated using empirical regressions for each species in the Laguna Grande ridge-top forest, La Gomera. (b) Frequency distribution of DBH classes within the plot per species. (c) Scaled-up total mean daily sapflow per DBH class (every 0.02 m) and tree species (Left: Erica arborea, Middel: Myrica faya, and Right: Laurus azorica) within the study plot (300 m2).

tion by fog. As shown in Fig. 7e, sapflow in June reached very low values when fog was registered (3–5 June 2003). Reductions in sapflow activity (hourly based) during times of fog ranged between 10% and 90% compared with preceding hours without fog or rain (July–August 2003). Monthly average sap velocity per species clearly decreased with amount of rainfall collected per month (r2 > 0.90) and in less extent with fog (r2 > 0.4) from July until November 2003 (Fig. 7f). However, this pattern was not maintained during the next year. On the other hand, maximum sapflow activity in both species was also observed in summer (up to 40–50 L m2 h1), even though extractable soil moisture in the top 0.30 m reached minimum values during that time of year (Fig. 7b). High sap velocity values in E. arborea occurred also during the flowering period in spring under conditions of ample soil moisture (Fig. 7b and d). Hourly sap velocities during the rainy season (low values of VPD and no soil water limitation) were related to Rg (Fig. 8). Under low radiation conditions, v was very low but it increased with light intensity. The radiation level needed to reach maximum v was 150 W m2 for M. faya (corresponding with a photon flux density of 284 lmol m2 s-1) and 100 W m2 (189.5 lmol m2 s1) for E. arborea (Fig. 8). Sapflow activity was related asymptotically to VPD for both species under light-saturated conditions (incoming radiation) (Fig. 9). However, although v in both species reached a plateau for VPD values above 0.7 kPa (E. arborea) and 0.9 kPa (M. faya) during wet conditions (March–April 2003), and above 1 kPa during times of soil water limitation (August 2003), sap velocities in E. arborea were markedly lower during the dry period. No such difference between wet and dry periods was observed for M. faya (Fig. 9). In the absence of limitations imposed by soil moisture (he > 0.8), Rg (>150 W m2 for M. faya and > 100 W m2 for E. arborea), or VPD (>0.9 kPa for M. faya and >0.7 kPa for E. arborea), v was linearly related to temperature (Fig. 10). However, under conditions of soil water limitation (he < 0.8), sap velocity showed a clear plateau at

temperatures above 17–20 °C for both species. Temperatures below 12 °C and above 30 °C seem to constrain sapflow activity (decreasing slope) in both species (Fig. 10). Similarly, without limitations imposed by Rg (>150 W m2 for M. faya and >100 W m2 for E. arborea), or VPD (>0.9 kPa for M. faya and >0.7 kPa for E. arborea), no clear relationship was found between sap velocity and extractable moisture he in the top 30 cm of soil (Fig. 11). In addition, it would seem as though a value for he of only 0.2 would be sufficient to produce maximum sapflow activity (Fig. 11). In view of the fact that this value of he is close to the residual moisture content for the upper layer (Fig. 7b), this demonstrates once again that the trees must have had access to moisture in deeper layers. 4. Discussion 4.1. Sap velocity and effects of meteorological and soil moisture conditions Sap velocity in Erica arborea clearly increased with increasing DBH, which may be explained by the fact that bigger trees of this species (characterized by their narrow crowns) tend to intercept a higher portion of incident radiation in forest canopies (Kelliher et al., 1992). By contrast, for Myrica faya, little variation in v was observed between differently sized trees (Fig. 3), possibly because of its broader crown. A slightly lower light intensity was needed to produce maximum sap velocity in E. arborea compared to M. faya (Fig. 8), which in combination with a more exposed position might contribute to increase sapflow activity in E. arborea. However, other physiological characteristics must be studied in order to explain the higher sapflow activity in Erica. Alternatively, the relationship between foliage mass (or LAI) and stem diameter may be inherently less strong for Myrica than for Erica (cf. Waring et al., 1982). Mean maximum v (55–65 L m2 h1) in E. arborea compared well with maximum rates obtained for this species in Tenerife by

48

G. García-Santos / Journal of Hydrology 462–463 (2012) 42–52

Fig. 7. (a) Water inputs to the Laguna Grande ridge top forest: rainfall (black bars, righthand axis), and fog intercepted by a quarter-sized standard fog collector above the canopy (gray bars, left-hand axis). (b) Temporal variability of relative extractable soil water content (from 0 to 1) in the top 0–0.15 and 0.5–0.30 m depth. (c) and (d) Annual course of mean sap velocity (mm h1 or L m2 SA h1) in Myrica faya and Erica arborea between February 2003 and January 2005. (e) Example of changes in hourly sap velocity as affected by fog occurrence between 3 and 9 June 2003. (f) Monthly sap flow velocity vs. water inputs from rainfall and fog from July until November 2003. E. arborea: v = 0.32 Fog + 21.53 (r2 = 0.44), v = 0.1  103 Rainfall 0.06 Rainfall + 17.51 (r2 = 0.92); M. faya: v = 0.32 Fog + 19.35 (r2 = 0.50), v = 0.1  103 Rainfall 0.07 Rainfall + 15.24 (r2 = 0.94).

Jiménez et al. (1999), but no comparative information appears to be available for M. faya (vmax = 25–30 L m2 h1). Sap velocities in E. arborea trees were three times higher than those in M. faya, which was unexpected as often narrow leaves are more water efficiency. Also in terms of sapflow per unit surface area of conducting sapwood, E. arborea trees transpired more water even though their sapwood area made up only half of the total cross-sectional area (cf. Fig. 2; Jiménez et al., 2000). The E. arborea trees in the study forest had a lower average DBH value (0.19 m, range: 0.04–0.59 m) compared to M. faya trees (average DBH 0.39 m, range: 0.16–0.69 m).

Both species showed delay in the start of sapflow activity observed in the early morning hours. This is probably caused by the presence of dew or fog droplets on the leaves, although in some cases it may reflect the refilling of previously depleted storage in the trunks (Waring et al., 1982; Goldstein et al., 1998; Motzer et al., this volume). Sap velocity was controlled largely by the prevailing evaporative conditions, with a maximum occurring around mid-day in both species. According to Zohlen et al. (1995) the stomata of some M. faya leaves remained closed in the absence of light. For this reason, the presently observed sapflow activity during the night

G. García-Santos / Journal of Hydrology 462–463 (2012) 42–52

49

Fig. 8. Relationship between sap velocity (mm h1 or L m2 SA h1) and global radiation (Rg, W m2) for Myrica faya (left) and Erica arborea (right) at the Laguna Grande ridgetop cloud forest site during a wet period (March 2003). Dashed line represents the light saturation curve.

Fig. 9. Relationship between sap velocity (mm h1 or L m2 SA h1) and vapor pressure deficit (VPD, kPa) for Myrica faya (left) and Erica arborea (right) at the Laguna Grande ridge-top cloud forest site during a wet period (March 2003) (d), and during a dry period in which soil moisture was likely to be limited (August 2003) (D).

Fig. 10. Relationship between hourly sap velocity (mm h1 or L m2 SA h1) and air temperature (°C) for Myrica faya (left) and Erica arborea (right) at the Laguna Grande ridge-top cloud forest site in the absence of limitations imposed by radiation, vapor pressure deficit or soil moisture (N), and during conditions of limitation by soil moisture (D).

Fig. 11. Relationship between mean daytime sap velocity (mm d1 or L m2 SA d1) and mean relative extractable soil water in the upper 0.30 m in Myrica faya (N) and Erica arborea (s) between February and October 2003 at the Laguna Grande ridgetop cloud forest site.

(Fig. 4) was unexpected. Nocturnal sapflow activity was observed throughout the year at Laguna Grande, and has also been reported for several coniferous and broad-leaved species on tropical mountains in East Africa (Fetene and Beck, 2004) and for a lower montane cloud forest in Ecuador (Motzer et al., this volume). Average nocturnal water uptake was small (4.6% and 5.7%) compared to day-time values, but was similar to values published for other species, e.g. 5% for Eucalyptus grandis (Beyon, 1999), 6% in Malus sylvestris (Green et al., 1989), and 10% in paper birch (Daley and Phillips, 2006). Nights with sapflow activity were more frequent during the wet season (not shown), which may be tentatively explained by the absence of any limitation in soil moisture. Two different sap velocity patterns were distinguished in the absence of light, which were controlled by different factors in the two species. On some

50

G. García-Santos / Journal of Hydrology 462–463 (2012) 42–52

days, v decreased steadily after the mid-day peak until stopping entirely several hours after dusk. This pattern may reflect the gradual depletion of water stored in the stem tissue, which may take several hours (Fetene and Beck, 2004; Motzer et al., this volume). In other cases, nocturnal sapflow was observed to be increased from zero or low activity to a secondary maximum that was always significantly lower than the preceding peak during the day-time. This can be explained in terms of transpiration and water storage dynamics in stem tissue under conditions of ample soil moisture. In those cases where night-time sapflow activity was positively correlated with VPD and high wind speeds (not shown), sapflow may be indicative of nocturnal transpiration, as has also been observed in eucalyptus (Beyon, 1999), poplar (Hogg and Hurdle, 1997) and paper birch (Daley and Phillips, 2006). By contrast, nights in which the timing and magnitude of sapflow was not affected by environmental variables (not shown) may indicate stem water recharge, as also observed in red oak and red maple by Daley and Phillips (2006). Zero sapflow activity may be part of a nocturnal flow reversal. Although the direction of sapflow could not be distinguished by the presently used method, zero activity may have occurred after downward sapflow occurred for a few hours and before normal upward flow was resumed. The purpose of reverse sapflow may be seen as a strategy to facilitate root growth in deeper layers and to transfer water away from shallow-rooted competitors (Hultine et al., 2003) although it is unclear whether this applies to the Laguna Grande ridge-top forest. M. faya reached its maximum sapflow activity in summer (Fig. 7c) during its flowering period, even though extractable soil moisture in the top layers was at a minimum at this time of year (Fig. 7b). Therefore, the roots must have had access to water in deeper layers. Because canopy conductance decreased under highly evaporative conditions (García-Santos et al., 2009), transpiration was maintained at a more or less constant rate despite the higher evaporative demand during the summer months. The shape of the relationship between canopy conductance in M. faya and VPD, as well as the threshold value at which sap velocity reached a maximum (Fig. 9), were similar to those reported for another species (Persea indica) found in the Canary laurel forest (González-Rodríguez et al., 2002), as well as for Canary pine (Luís et al., 2005). However, these results contrast with the weak stomatal response to evaporative conditions found for M. faya in Tenerife by Zohlen et al. (1995) using porometry. In contrast to M. faya, E. arborea exhibited its highest sapflow activity during its spring-time flowering, with a decrease was observed in August when extractable soil moisture in the upper 0.30 m was at a minimum (Figs. 7b–d and 11), this may be explained by shalowed root sytem in this species. Canopy conductance for E. arborea also decreased with VPD in summer (García-Santos et al., 2009), as also observed for E. arborea in Mediterranean maquis (Tognetti et al., 2000). A direct correlation between VPD and sap velocity was found, with v in both species reaching a plateau for VPD values above 0.7 kPa when soil moisture was limiting, indicating stomatal limitation of transpiration (Fig. 9). Saturation conditions were reached at sligthly higher values (1 kPa) for both species during the driest period (August 2003). Whilst sap flow activity in E. arborea showed a slight reduction during the dry season (indicating soil water limitation), this was not observed in M. faya (Fig. 9). Reaching saturation at higher VPD values, and reducing sapflow activity during low soil water conditions is a drought adaptation mechanism that has also been observed during summer in E. arborea trees in Mediterranean maquis (Tognetti et al., 2000). A similar adaptation was recently reported for Pinus canariensis in the Canaries as well (Luís et al., 2005).

4.2. Stand transpiration The forest at Laguna Grande transpired throughout the year, and high values (around 2.5–3 mm day1) occurred on clear days with high VPD. Much higher values (up to 7 mm day1) were obtained by Jiménez et al. (1996) for a Laurus azorica forest in Tenerife. The average transpiration observed at the stand level (1.2 ± 0.12 mm day1 for the 2 years) resembles Et values obtained by micro-meteorlogical techniques for equally fog-ridden (but much wetter in terms of rainfall) upper montane cloud forests in Puerto Rico (1.33 ± 0.95 mm day1 at 800 m.a.s.l.; Holwerda, 2005) and Queensland (1.0 mm day1 at 1560 m.a.s.l.; McJannet et al., 2007), but it exceeds water uptake for a stunted ridge-top elfin cloud forest at 1.010 m.a.s.l. in Puerto Rico (0.81 ± 0.96 mm day1; Holwerda, 2005). Interestingly, Et expressed as a fraction of net radiation varied little between the Laguna Grande, Puerto Rican and Australian forests (0.27–0.31) despite major contrasts in incoming radiation and rainfall totals (cf. Bruijnzeel et al., this volume). Sap velocity was much reduced during times of fog (Fig. 7e), an observation been documented for redwood forest in California (Burgess and Dawson, 2004) and windward cloud forest in Hawai’i (Santiago et al., 2000). Interestingly, Burgess and Dawson (2004) demonstrated that fog water was actually absorbed by the redwood foliage, thereby reducing dehydration during times of low rainfall and occasionally even producing reversed sap flow (cf. Lai et al., 2007). Although it is not known whether the same process also occurs at Laguna Grande, the present results illustrate the importance of fog for the reduction of tree water use (cf. Reinhardt and Smith, 2008; Mildenberger et al., 2009), even if actual amounts of fog drip may seem limited (García-Santos and Bruijnzeel, 2011). Although the ridge-top forest examined in this study was characterized by relatively thin trunks (DBH between 0.03 and 0.7 m), the contribution of a few dominant trees (DBH > 0.3 m) to overall stand transpiration represented 90% (cf. Küppers et al., 2008; Motzer et al., this volume). As for the dominant species, transpiration by M. faya made up nearly half of the total stand transpiration (48%) despite contributing 57% of plot basal area. Conversely, E. arborea (33% relative basal area) contributed 40% to overall stand transpiration whereas the contribution by L. azorica (11.5%) matched its relative basal area (10%). Finally, the forest’s principal loss of water during the five mostly dry summer months was via transpiration. Total net water inputs to the forest floor between May and September 2003 were only 41 mm, which was insufficient to satisfy vegetation water demand during that period (c. 246 mm) (Fig. 7). Therefore, the remaining water must have been taken up from the soil beyond 0.3 m depth (Fig. 7b) (García-Santos, 2007).

5. Conclusions Stomatal control during the dry season was found to have a moderating influence on transpiration in both Myrica faya and Erica arborea at Laguna Grande in the Canary Islands. Estimated annual stand transpiration (416 mm, average Et 1.2 mm day1) was low compared to that obtained for typical laurel-dominated forest elsewhere in the Canaries, but very similar to equally fog-ridden upper montane forests in Puerto Rico and Queensland. M. Faya had a higher contribution to overall stand transpiration than the other two species investigated, which was mainly due to contributions by several bigger-sized individuals. E. Arborea trees were generally smaller but showed higher sap flow activity. Therefore, their relative contribution to overall stand transpiration (40%) exceeded their contribution to stand basal area (33%). The 57% basal area occupied by M. faya contributed 48%.

G. García-Santos / Journal of Hydrology 462–463 (2012) 42–52

Nocturnal sap flow activity was observed throughout the year in both species, but was small compared with day-time values (4.8–6.0%, depending on species). In this humid ridge-top forest, transpiration was reduced by available energy (i.e. fog occurrence) and stomatal control (mostly on dry summer days). Although the water balance during the dry season was strongly negative and inputs by fog during this time of year were marginal, the prime importance of fog relates to diminishing evaporative losses. Hourly transpiration rates during times of fog were reduced by 10–90% compared to clear-sky conditions. Acknowledgments The fieldwork phase of the research was sponsored by a predoctoral scholarship to the senior author from the Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria (INIA) of the Ministry of Science and Technology of Spain, and was partly carried out by the author in the Soil and Irrigation Department of the Instituto Canario de Investigaciones Agrarias (ICIA) in Tenerife (Canary Islands) within the framework of project RTA01-97. Reprocessing of the data and writing were carried out at VU University Amsterdam. The author gratefully acknowledge the support received from Prof. Marzol-Jaén (University of La Laguna, Tenerife) and the staff of the National Park of Garajonay. In addition, I thank Drs. M. Lubczynski and C. Tobón for comments on an earlier draft, S. Bruijnzeel for his editorial contribution and the three anonymous reviewers for their interesting comments. References FAO, 1998. World reference base for soil resources. World Soil Resour. Rep. 84, 21– 22 (FAO, Rome). Arévalo, J.R., Fernández, A., Montalvo J., 2002. Plan complementario meteorológico. TRAGSATEC-Parque Nacional de Garajonay, La Gomera, Internal report. Beyon, R.G., 1999. Night time water use in an irrigated Eucalyptus grandis plantation. Tree Physiol. 19, 853–859. Black, T.A., 1979. Evapotranspiration from Douglas fir stands exposed to soil water deficits. Water Resour. Res. 15, 164–170. Burgess, S.S.O., Dawson, T.E., 2004. The contribution of fog to the water relations of Sequoia sempervirens D. Don: Foliar uptake and prevention of dehydration. Plant Cell Environ. 27, 1023–1034. Cermák, J., Nadezhdina, N., 1998. Sapwood as the scaling parameter – defining according to xylem water content or radial pattern of sap flow? Ann. Sci. For. 55, 509–521. Daley, M.J., Phillips, N.G., 2006. Interspecific variation in nighttime transpiration and stomatal conductance in a mixed New England deciduous forest. Tree Physiol. 26, 411–419. Dorta, P., 1996. Las inversiones térmicas en Canarias. Invest. Geogr. 15, 109–124. Edwards, W.R.N., Becker, P., Cermák, J., 1996. A unified nomenclature for sap flow measurements. Tree Physiol. 17, 65–67. Fetene, M., Beck, E.H., 2004. Water relations of indigenous versus exotic tree species, growing at the same site in a tropical montane forest in southern Ethiopia. Tree 18, 428–435. García-Santos, G., 2007. An Ecohydrological and Soils Study in a Montane Cloud Forest in the National Park of Garajonay, La Gomera Canary Islands, Spain. PhD Thesis, VU University Amsterdam, Amsterdam, The Netherlands. . García-Santos, G., Bruijnzeel, L.A., 2011. Water inputs dynamics in a subtropical ridge top cloud forest, National Park of Garajonay (La Gomera, Canary Islands, Spain). J. Hydrol. Process. 09, 0567. doi:10.1002/hyp. 7760. García-Santos, G., Marzol, M.V., Aschan, G., 2004. Water dynamics in a montane cloud forest in the National Park of Garajonay. Hydrol. Earth Syst. Sci. 8, 1065– 1075. García-Santos, G., Bruijnzeel, L.A., Dolman, A.J., 2009. Modelling canopy conductance under wet and dry conditions in a subtropical cloud forest. Agric. For. Meteorol. 149 (10), 1565–1572. doi:10.1016/ j.agrformet.2009.03.008. Giambelluca, T.W., Ziegler, A.D., Nullet, M.A., Truong, D.M., Tran, L.T., 2003. Transpiration in a small tropical forest patch. Agric. For. Meteorol. 117, 1–22. Goldstein, G., Andrade, J.L., Meinzer, F.C., Holbrook, N.M., Cavelier, J., Jackson, P., Celis, A., 1998. Stem water storage and diurnal patterns of water use in tropical forest canopy trees. Plant Cell Environ. 21, 397–406.

51

Golubic, I., 2001. Vegetationskundliche Analyse von Lorbeerwald-und ähnlichen Beständen im Rahmen einer Untersuchung des Landschaftswasserhaushaltes im Garajonay Nationalpark La Gomera. MSc Thesis, University of Essen, Essen, Germany. González-Rodríguez, A.G., Morales, D., Jiménez, M.S., 2002. Leaf gas exchange characteristics of a Canarian laurel forest tree species [Persea indica L.K. Spreng.] under natural conditions. Plant Physiol. 159, 695–704. Granier, A., 1987. Evaluation of transpiration in a Douglas fir stand by means of sap flow measurements. Tree Physiol. 3, 309–320. Green, S.R., McNaughton, K.G., Clothier, B.E., 1989. Observations of nighttime water use in kiwifruit vines and apple trees. Agric. For. Meteorol. 48, 251–261. Grime, V.L., Sinclair, F.L., 1999. Sources of errors in stem heat balance sap flow measurements. Agric. For. Meteorol. 94, 103–121. Hogg, E.H., Hurdle, P.A., 1997. Sap flow in trembling aspen: implications for stomatal responses to vapor pressure deficit. Tree Physiol. 17, 501–509. Holwerda, F., 2005. Water and Energy Budgets of Rain Forests Along an Elevation Gradient Under Maritime Tropical Conditions. PhD Thesis, VU University Amsterdam, Amsterdam, The Netherlands. . Hultine, K.R., Cable, W.L., Burgess, S.S., Williams, D.G., 2003. Hydraulic redistribution by deep roots of a Chihuahuan Desert phreatophyte. Tree Physiol. 23, 353–360. Jiménez, M.S., Cermák, J., Kucera, J., Morales, D., 1996. Laurel forest in Tenerife, Canary Islands: the annual course of sap flow in Laurus trees and stand. J. Hydrol. 183, 307–321. Jiménez, M.S., Morales, D., Kucera, J., Cermák, J., 1999. The annual course of transpiration in a Laurel forest of Tenerife. Estimation with M. faya. Phyton (Austria) 39, 85–90. Jiménez, M.S., Nadezhdina, N., Cermák, J., Morales, D., 2000. Radial variation in sap flow in five laurel forest tree species in Tenerife, Canary Islands. Tree Physiol. 20, 1149–1156. Kelliher, F.M., Kostner, B., Hollinger, D., Byers, J.N., Hunt, J.E., et al., 1992. Evaporation, xylem sap flow, and tree transpiration in a New Zealand broadleaved forest. Agric. For. Meteorol. 62, 53–73. Küppers, M., Motzer, T., Schmitt, D., Ohlemacher, C., Zimmermann, R., et al., 2008. Stand structure, transpiration responses in trees and vines, and stand transpiration of different forest types within the mountain rainforest. In: Beck, E., Bendix, J., Kottke, I., Makeschin, F., Mosandl, R. (Eds.), Gradients in a Tropical Mountain Ecosystem of Ecuador. Ecological Studies, vol. 198. SpringerVerlag, New York, pp. 243–258. Lai, I.L., Schröder, W.H., Wu, J.T., Kuo-Huang, L.L., Mohl, C., Chou, C.H., 2007. Can fog contribute to the nutrition of Chamaecyparis obtusa var. formosana? Uptake of a fog solute tracer into foliage and transport to roots. Tree Physiol. 27, 1001– 1009. Luís, V.C., Jiménez, M.S., Morales, D., Kucera, J., Wieser, G., 2005. Canopy transpiration of a Canary Islands pine forest. Agric. For. Meteorol. 135, 117– 123. Lundblad, M., Lagergren, F., Lindroth, A., 2001. Evaluation of heat balance and heat dissipation methods for sapflow measurements in pine and spruce. Ann. Sci. For. 58, 625–638. McJannet, D., Fitch, P., Disher, M., Wallace, J., 2007. Measurements of transpiration in four tropical rainforest types of north Queensland, Australia. Hydrol. Process. 21, 3549–3564. Mildenberger, K., Beiderwieden, E., Hsia, Y.J., Klemm, O., 2009. CO2 and water vapor fluxes above a subtropical mountain cloud forest – the effect of light conditions and fog. Agric. For. Meteorol. 149, 1730–1736. Miyamoto, T., Annaka, T., Chikushi, J., 2003. Soil aggregate structure effects on dielectric permittivity of an Andisol measured by time domain reflectometry. Vadose Zone J. 2, 90–97. Morales, D., González-Rodríguez, A., Jiménez, M.S., 2003. Ecofisiología de la laurisilva canaria. Ecosistemas 2003/1, . Pérez de Paz, P.L., del Arco Aguilar, M., Acebes, J.R., Wildpret, W., 1990. La vegetación cormofítica vascular del P.N. de Garajonay. In: Parque Nacional de Garajonay. Patrimonio Mundial. Coleccion Técnica, ICONA, Madrid, pp. 137– 171. Reinhardt, K., Smith, W.K., 2008. Impacts of cloud immersion on microclimate, photosynthesis and water relations of Abies fraseri Pursh. Poiret in a temperate mountain cloud forest. Oecologia 158, 229–238. Rodríguez, A., Arbelo, C., Notario, J., Vargas, G., Mora, J. et al., 2002. Plan complementario edafológico del programa de seguimiento edafológico en el Parque Nacional de Garajonay. La Laguna, Parque Nacional de Garajonay, Organismo Autónomo Parques Nacionales, Ministerio de Medio Ambiente, Tenerife, Spain. Santiago, L.S., Goldstein, G., Meinzer, F.C., Fownes, J.H., Mueller-Dombois, D., 2000. Transpiration and forest structure in relation to soil waterlogging in a Hawaiian montane cloud forest. Tree Physiol. 20, 673–681. Schemenauer, R., Cereceda, P., 1994. A proposed standard fog collector for use in high elevation regions. J. Appl. Meteorol. 33, 1313–1322. Sharon, D., 1980. Distribution of hydrologically effective rainfall incident on sloping ground. J. Hydrol. 46, 165–188. Sperling, F.N., Washington, R., Whittaker, R.J., 2004. Future climate change of the subtropical north Atlantic: implications for the cloud forest of Tenerife. Clim. Change 65, 103–123.

52

G. García-Santos / Journal of Hydrology 462–463 (2012) 42–52

Tognetti, R., Minnocci, A., Penuelas, J., Raschi, A., Jones, M.B., 2000. Comparative field water relations of three Mediterranean shrub species co-occurring at a natural CO2 vent. J. Exp. Bot. 51, 1135–1146. Topp, G.C., Davis, J.L., Annan, A.P., 1980. Electromagnetic determination of soil water content: measurements in coaxial transmission lines. Water Resour. Res. 16, 574–582. Waring, R.G., Schroeder, P.E., Oren, R., 1982. Application of the pipe model theory to predict canopy leaf area. Can. J. For. Res. 3, 556–560.

Yang, D., Goodison, B.E., Ishida, S., Benson, C.S., 1998. Adjustment of daily precipitation data at 10 climate stations in Alaska: Application of World Meteorological Organization intercomparison results. Water Resour. Res. 34, 241–256. Zohlen, A., González-Rodríguez, A., Jiménez, M.S., Lösch, R., Morales, D., 1995. Transpiration and stomatal regulation in Laurel forest trees measured in the spring. Vieraea 24, 91–104.