Introduction

Dry and semiarid ecosystems in the Mediterranean region are characterized by summers with long water deficit periods and high values of global radiation, temperature and vapour pressure deficit (Di Castri 1973). These environmental conditions affect the physiological activity of plants (Larcher 1995; Gazal et al. 2006) producing changes in plant water consumption through stomatal regulation (Salleo et al. 2000; Mediavilla and Escudero 2004) which, in turn, influence the sap flow rate (Badalotti et al. 2000; David et al. 2007; Gartner et al. 2009).

On drought stress conditions, the species have developed adaptations for escaping or resisting to drought. Drought resistance can be by means of two strategies: drought avoiding or drought tolerating. There is two distinct kinds of drought avoidance mechanisms: (1) water saver, that avoid drought by water conservation, and (2) the water spender, that avoid drought by absorbing water sufficiently rapidly to keep up with their extremely rapid water loss. These concepts and definitions exposed by Levitt (1980) have been applied in several studies with the aim of assess the drought resistance strategy of species. In this sense, previous papers have assessed some species of the present study. Pinus halepensis, Erica multiflora and Pistacia lentiscus have been assessed as drought avoidance species (Vilagrosa 2002; Ferrio et al. 2003; Lorens et al. 2003b; Baquedano and Castillo 2007); while Quercus coccifera has been assessed as a drought-tolerant species (Martínez-Ferri et al. 2000; Baquedano and Castillo 2006).

In semiarid areas of the Mediterranean Spain, some changes in the reforestation strategy have been observed in the last decades. In several cases, has been prioritized the introduction of native shrubs and grasses, unlike of extensive conifer reforestation (e.g. Aleppo pine) carried out during the second part of the twentieth century. In this context, knowledge about plants strategies to cope with drought, as well as the drought avoidance mechanisms of the species introduced acquires great importance; because would allow to link the selection of species for forest restoration with the soil water availability on the plantation site and the water demand of the species (Vilagrosa et al. 2003a; Vallejo and Alloza 2004).

Taking all the above into consideration, in this paper, we analyzed the daily sap flow rates in one tree species (Pinus halepensis), three shrub species (Quercus coccifera, Pistacia lentiscus and Erica multiflora), and an alpha grass steppe (Stipa tenacissima) in a semiarid ecosystem, under two slope aspects (north- and south-facing). The hypotheses of this work were: (1) Sap flow rate is affected by microclimatic conditions and soil water availability, therefore hopefully observe: (a) a common seasonal pattern in the sap flow rate by species, (b) higher sap flow rate in south-facing than north-facing slope, and (c) higher sap flow rate during the wetter year; and (2) Sap flow response of plants to cope with water deficit stress, expressed by means of predawn water potential, can be linked with drought avoidance mechanisms (water saving or water spending). To test these hypotheses, the aims of this paper were (1) evaluate the effect of microclimatic conditions and soil water content on the sap flow rate by species in the time (seasonal pattern and year) and slope aspect; and (2) establish a link between the daily sap flow of the different species as response to variations in predawn leaf water potential and drought avoidance mechanisms of plants (water saver or water spender).

Materials and methods

Site description

The research was carried out at the “El Ventós” experimental station, Alicante, Spain. Altitude 600 m asl and slopes between 23° and 26°. Loam and loamy limestone support calcareous Cambisol and Lythic Leptosol (FAO-UNESCO), with average soil depth of 20 cm, organic matter content of 4.7%, clay content between 48 and 58%, total porosity of 50.2%, field capacity of 25% and average infiltration rates measured with a double-ring infiltrometer were between 192 and 438 mm h−1 (Bellot et al. 2001; Chirino et al. 2006). The climate is semi-arid Mediterranean, with a very high inter-annual variability. Mean annual temperature is 18.2°C and mean annual rainfall is 275 mm (1976–2006 period). The current vegetation landscape in the study area is composed of different land cover types (patch-mosaic) which can be situated along a successional gradient of increasing structural complexity of vegetation: degraded open land or bare soil, sometimes occupied by microphytic crusts (Maestre et al. 2002); dry grassland formations of Brachypodium retusum Pers. Beauv., with dwarf scrubs (Anthyllis cytisoides L., Helianthemum syriacum (Jacq.) Dum.-Cours. and Thymus vulgaris L.); and more mature landscape patches composed of scattered thorn and sclerophyllous shrublands with Quercus coccifera L., Pistacia lentiscus L., Erica multiflora L., Rhamnus lyciodes L. and Rosmarinus officinalis L. (Bonet et al. 2001). Other anthropogenic components of the present landscape mosaic are afforestations with Pinus halepensis Mill. on dry grasslands and thorn shrublands. On south-facing slopes, the predominant vegetation is a rather interspersed mosaic of open Stipa tenacissima steppes and dwarf shrubland with gradual transitions between them (Bautista et al. 2007). A description of the land cover types has been presented by Chirino et al. (2006).

Experimental design for sap flow measurements and data processing

Sap flow rate measurements were carried out in sets of five consecutive days during the months of January (winter), May (spring), August (summer) and October (autumn) in the years 1998 and 1999. These measurements were taken in four species on a north-facing slope (Pinus halepensis Mill., Quercus coccifera L., Pistacia lentiscus L. and Erica multiflora L.) and four species on a south-facing slope (Pinus halepensis Mill., Quercus coccifera L., Pistacia lentiscus L. and Stipa tenacissima L.). On both slopes, Aleppo pines (Pinus halepensis) were planted 40 years ago, while the other species belonged to the natural succession. For each slope aspect (north- and south-facing) three individuals were selected randomly by species for the sap flow measurements. The height of the individuals is shown in Table 1. Leaf Area Index (LAI, Table 1) was measured in individuals similar to those where the sap flow was monitored. LAI was measured by means of a destructive method based on the leaf area of the shoot biomass in a rectangular prism with square base of 50 × 50 cm and a height according to the individuals sampled.

Table 1 Height of the individuals selected by species for sap flow measurements
Full size table

To measure the sap flow we used the stem heat balance method (SHB system, van Bavel 1984). In each of the plants selected we installed three sap flow gauges (Dynagauge, Dynamax Inc., Houston, USA, models SGA2, SGA3 or SGA5) on upper third crown on branches oriented in three directions (N, SE, SW). In the case of P. halepensis, it was necessary to use a 5 m tall scaffolding to reach the upper third of the plant crown. Moreover, additional insulation and an exterior wrapping of tinfoil were also used to provide higher isolation from meteorological conditions. Measurements of sap flow velocity (Q; g H2O h−1) were programmed with a scan interval of 1 min and signals were averaged every 15 min. After removing the gauges, the leaf area of each branch was measured destructively to normalize sap flow velocity per unit of leaf area (Q 1; g H2O cm−2 leaf area h−1). Leaves were scanned with a professional scanner (Epson Expression 1680 Pro, Seiko Epson Corporation, Nagano, Japan). The images obtained were analyzed with the specific image processor WinRHIZO software (Regent Inst., Canada) to obtain leaf area (cm2). FLOW32 software requests a time interval of reduced power, and during this time interval, the flow rate is set to zero (van Bavel, 1984). In our case, we defined the least time possible (power-down time: 00:00 AM and power-up time: 3:00 AM). Thus, the daily sap flow rate by individual (Q d; g H2O cm−2 leaf area day−1) was computed by means of the Q 1 accumulated from 3:00 to 24:00. Mean daily sap flow rate (Q md; g H2O cm−2 leaf area day−1) was computed by averaging the Q d value obtained from three individuals selected in each sampled periods. Subsequently, we estimated the mean daily sap flow rate for each species by month (January, May, August and October), year (1998 and 1999) and slope aspect (north- and south-facing).

Meteorological conditions, soil moisture and predawn water potential

Meteorological data were monitored during 1998 and 1999. Rainfall, air temperature (Tª), air humidity (H) and global radiation (R g) were measured continuously with an automatic weather station Campbell Scientific CR10 (Campbell Scientific Ltd. UK). The automatic weather station on north-facing slope was installed since 1995, while on the south-facing slope was installed in March, 1998. Tª, H and R g were taken at 10 min intervals and averaged every hour. Vapour pressure deficit (VPD) was calculated from Tª and H data. Soil moisture was monitored using the Time Domain Reflectometry System (Reflectometer Tektronic 1502C, Metallic TDR cable Tester, Tektronix, Beaverton, OR, USA) from 0 to 30 cm (maximum soil depth) by means of 12 probes inserted in a patch-mosaic of three vegetation cover types: pine with dry grass and sclerophyllous shrub on the north-facing slopes and alpha grass steppes of S. tenacissima on the south-facing slopes (Chirino 2003). Relative Extractable Water (REW) in the root zone was calculated from the soil moisture monitored in these vegetation types by means of the equation REW = (θ – θ mín)/(θ FC – θ mín) described by Granier (1987) and Bréda et al. (1995), where θ is the actual soil water content and θ mín is the minimum soil water content observed during the soil moisture monitoring, and θ FC = θ at field capacity. All the variables are expressed in L m−2. REW = 0.4 was identified as the threshold for soil water deficit (Granier 1987; Bréda et al. 1995). Leaf water potential before sunrise (predawn leaf water potential: Ψpd) was measured by using the Scholander pressure chamber (Soil moisture Equipment Corp., Santa Barbara, CA, USA) in three branches of three individuals selected by species. This variable was measured in all species, except in Stipa tenacissima due to technical difficulties.

Statistical analysis

Statistical analysis was carried out with the STATISTICA 6.0 package (Statsoft Inc., Tulsa, USA). Means monthly Q md were compared using a General lineal Model (GLM) Repeated Measures by means of multi-way within-subject designs. We used the years (1998 and 1999) and monthly Q md value as within-subjects factors, and slope aspect and species as between-subjects factors. Due to interactions observed in the results of GLM Repeated Measures (multi-way within-subject designs), subsequently was carried out a GLM Repeated Measures with one within-subject factor (monthly Q md) and only a between-subject factor (slope aspect, year or species). In the case of species analysis, a post hoc Tukey HSD was used. In order to know the differences or patterns commons between monthly Q md of species, a two-factor Univariate ANOVA (year and month) was carried out, using a post hoc Tukey HSD. Meteorological data on the annual periods (full data on north-facing slope) were compared by means of analysis of variance (one way ANOVA). Data on Tª (°C), H (%), R g (W m−2), VPD (kPa) and rainfall by month and slope aspect were compared using nonparametric statistical methods (Mann–Whitney test and “t” test) due to differences in the number of observations. The north-facing slope had full data for the years 1998 and 1999; in contrast, the south-facing slope had data only from March, 1998 to December, 1999, when the automatic weather station was installed. The relationship between Q md (g H2O cm−2 day−1) and Ψpd (MPa) in each species was calculated by means of a exponential regression. In the case of the relationship between Q 1 (g H2O cm−2 h−1) and the hourly value of R g (W m−2) and VPD (kPa) throughout the day we used exponential, quadratic and linear regressions.

Results

Meteorological variables and soil moisture

Monthly global radiation (R g) was higher on the south-facing slope than on the north-facing slope (Fig. 1; p < 0.001). No significant differences between slope aspects were found with respect to rainfall and VPD (Fig. 1). Nevertheless, both slope aspects showed significant differences (p < 0.001) between measuring months for the variables Tª, R g, and VPD. On an annual time scale, Tª (p = 0.028) and VPD (p = 0.011) were higher in 1998 than in 1999. Annual rainfall was 1.2 times higher in 1998 (288.3 mm) than in 1999 (241.1 mm), resulting in higher soil moisture values and REW during the first year (Fig. 2). Soil moisture average on north-facing slope ranged between 18.1% in pine with dry grass and 15.2% in sclerophyllous shrub; while, on south-facing slope represented by alpha grass steppe of S. tenacissima the soil moisture average was 12.7% (Fig. 2, above). REW values on both slope aspects were below the threshold established for Mediterranean vegetation cover (REW = 0.4), which in our case represents a soil water content of 40% below field capacity (Fig. 2, below). These results point to the long duration of soil water deficit conditions in the study area. Only from September, 1997 to March, 1998 were the REW values higher than the threshold, as a result of the extraordinary rains in the autumn of 1997.

Fig. 1
figure 1

Comparison between north- and south-facing slopes with respect to global radiation (R g), vapour pressure deficit (VPD) and rainfall during 1998 and 1999. Data on R g and VPD are mean monthly ± standard error (north-facing: black circle and solid line; south-facing: white circle and solid line). Rainfall data are monthly accumulated values (north-facing: black bar; south-facing: grey bar)

Full size image
Fig. 2
figure 2

Daily rainfall and soil moisture (above), and Relative Extractable Water (REW, below) in two hydrological years 1997/1998 and 1998/1999. The data showed in the x-axis correspond to daily measurements. We show label of months to improve the understanding. On north-facing slope: pine with dry grass (solid line) and sclerophyllous shrub (dashed line). On south-facing slope: alpha grass steppes of S. tenacissima (long dash). REW = 0.4 is threshold for Mediterranean species (dotted line). Months of sap flow measurements: vertical arrow

Full size image

Relationship between sap flow and meteorological variables

During the course of the day, the sap flow velocity per unit of leaf area (Q 1) showed a significant relationship with R g and VPD. As an example, we present the results obtained for Quercus coccifera on two different dates in 1998, with contrasting soil moisture and meteorological conditions. The first case (Fig. 3a) shows the results of the May 28th sap flow measured on the south-facing slope, with high daily soil water content (19.8%) and high daily VPD (1.22 kPa). The second case (Fig. 3b) shows the August 19th sap flow measured in north-facing conditions, with a lower daily soil water content (10.3%) and lower daily VPD (0.67 kPa). In both cases, R g presents a significant relationship with Q 1, by means of an exponential equation \( \left[ {Q_{ 1} = {\text{a}}{*} \left( { 1 - e^{{ - {\text{b}}{*} R_{g} }} } \right)} \right] \) where extreme values of R g tend to produce Q 1 approaching the asymptote. VPD shows a quadratic polynomial equation \( \left[ {Q_{ 1} = a + b_{{0{*} \, }} {\text{VPD}} + b_{{ 1{*} \, }} {\text{VPD}}^{ 2} } \right] \) on May 28th and a linear equation \( \left[ {Q_{ 1} = a + b_{{0{*} \, }} {\text{VPD}}} \right] \) on August 19th. These differences in the regression equations are explained by the fact that the first case presented higher values of soil water content and VPD than the second case, which favoured higher Q 1 values. However, when VPD > 1.4 kPa, Q 1 decreased. Similar results were observed in the other species (data no shown).

Fig. 3
figure 3

Left side shows the hourly variation in sap flow velocity (Q 1 : g H2O cm−2 h−1), global radiation (R g : W m−2) and vapour pressure deficit (VPD: kPa) in Q. coccifera (solid line Q 1, short dash R g, dash-dot VPD. a South-facing data: May 28, 1998. b North-facing data: August 19, 1998. Right side shows the regression equations according to the variables on the left side of the figure (significance differences **p < 0.01 and ***p < 0.001)

Full size image

Daily sap flow rate of each species by slope aspect, year and month

Average daily sap flow rates (Q md) of the different species by months and slope aspects during 2 years, 1998 and 1999 (Fig. 4) showed values between 0.001 and 0.202 g H2O cm−2 leaf area day−1. General lineal Model (GLM) Repeated Measures by means of multi-way within-subject designs indicated significant differences (p < 0.05) in Q md values between the factors: years, slope aspect and species (Table 2). In this analysis were observed interactions between all factors. Subsequent analysis using GLM Repeated Measures with only a within-subject and between-subject factor determined the effect of each factor (Table 3). Q. coccifera and P. lentiscus showed higher Q md values on the south-facing slope than on the north-facing slope during period of measurements (Table 3, Time factor p < 0.05), while P. halepensis showed an opposite result. A higher R g on south-facing slope could produce stress in P. halepensis, induce a stomatal closure and reduce the sap flow. In the cases of P. halepensis and P. lentiscus were observed interactions between the factors time and slope aspect (p < 0.05). The test of between-subjects effects indicated that Q. coccifera and P. lentiscus presented two times higher Q md in south- than north-facing (Table 3, slope aspect factor p < 0.01). P. halepensis not showed significance differences.

Fig. 4
figure 4

Monthly sap flow rate by species in each slope aspect (north- and south-facing) and during two years (1998 and 1999). Mean ± standard error, north-facing: black circle and solid line; south-facing: white circle and solid line. See in the last figure we show two species, E. multiflora from north-facing (black circle and solid line) and S. tenacissima from south-facing (white circle and solid line)

Full size image
Table 2 Results of General lineal Model (GLM) Repeated measures by means of multi-way within-subject designs
Full size table
Table 3 Results of General lineal Model (GLM) repeated measures by means of one-way within-subject designs (Greenhouse-Geisser test was used to compare the within-subject effects
Full size table

The inter-annual variability analysis on the north-facing slope indicated that Q md values were higher in 1998 than in 1999 for P. halepensis, P. lentiscus and E. multiflora (Table 3, Time factor p < 0.05). Only, were observed interactions between time and years in P. halepensis and P. lentiscus. On south-facing slope S. tenacissima showed higher Q md in 1998, while an opposite result was observed in P. lentiscus (Table 3, Time factor p < 0.05) showing Q md higher in 1999 than in 1998. P. halepensis (on south-facing) and Q. coccifera (in both slope aspect) no showed significance differences in the time (Table 3, Time factor p > 0.05). Several interactions were observed (Table 3, time × year factor, p < 0.05). The year factor (test of between-subjects effects) ratified the results of previous analysis, and indicated significance differences in the annual Q md in Q. coccifera.

Comparative analysis of Q md values between species within each slope aspect indicated significant differences (Table 3, Time factor, p < 0.001). On the north-facing slope Q. coccifera and P. lentiscus showed higher Q md than E. multiflora; while P. halepensis presented an intermediate result, no showing significance differences with the others species. On the south-facing slope Q. coccifera showed higher Q md than S. tenacissima and P. lentiscus. In this slope aspect, P. halepensis showed the lowest values (Table 3, Time factor, p < 0.001). Both time x species, like the species factor showed interactions and significance differences respectively (Table 3). The analysis of sap flow by species in each slope aspect no showed a common pattern between monthly Q md (Fig. 5). The month of maximum and minimum Q md during the 2 years of measurements was different in each species (Fig. 5). However, on south-facing slope P. halepensis, P. lentiscus and S. tenacissima tended to show the highest Q md values in May and August, and lowest values in January (Fig. 5). On north-facing slope, P. halepensis and P. lentiscus showed the highest Q md in May and lowest values Q md in August, while E. multiflora show the highest and lowest Q md in August and January respectively.

Fig. 5
figure 5

Average monthly sap flow rate by species in each slope aspect (north- and south-facing). Results of two-factor Univariate ANOVA (post hoc Tukey HSD). Sap flow values followed by the same letter are not significantly different at p < 0.05

Full size image

Daily sap flow rate by species under water deficit conditions: drought avoidance mechanisms of species

Predawn leaf water potential (Ψpd) values during the study period ranged between −0.2 and −3.2 MPa (Fig. 6). With the exception of P. halepensis, all the individuals sampled on the south-facing slope showed slightly higher Ψpd values than those sampled on the north-facing slope, but statistically significant differences were found only in Q. coccifera (p = 0.047). The relationship between Q md and Ψpd was verified by means of a exponential regression (\( Q_{\text{md}} = a{*} e^{{b{*} }} {\Uppsi }_{\text{pd}} \)). The Q md values in each species decreased according to the decrease in Ψpd (Fig. 6). The slope of the regression equations indicated the different response of the species studied under the different soil water availability conditions present during the monitoring period, expressed by means of predawn leaf water potential. In conditions unaffected by water deficit (Ψpd >−1.0 MPa), Q. coccifera showed the highest Q md values, while P. halepensis and P. lentiscus showed Q md values that were relatively high, but lower than Q. coccifera. Under water deficit conditions (Ψpd < −2.0 MPa), both P. halepensis and P. lentiscus presented decreased Q md values, more notably so in P. halepensis. On the other hand, E. multiflora showed low Q md values in all conditions as well as a somewhat reduced sap flow with decreases in Ψpd. It should also be observed that under the lowest Ψpd (<−2.8 MPa), Q. coccifera showed the highest Q md values, which were 36 times higher than the values for P. halepensis.

Fig. 6
figure 6

Relationship between mean daily sap flow rate (Q md) and mean predawn leaf water potential (Ψpd). Exponential regression equation by species. Data in each measurement period and slope aspect (significance differences *p < 0.05, **p < 0.01 and ***p < 0.001)

Full size image

Under different water deficit conditions during the study period, the daily sap flow rate of each species has been linked with a specific drought avoidance mechanism. In general, all the species studied showed a decrease in Q md values according to the decrease in Ψpd; nevertheless, the responses of the species to cope with water deficit conditions were different. A species-by-species comparison of the variations in daily sap flow rate with respect to decreases in predawn leaf water potential (Fig. 6), shows that the response of P. halepensis, P. lentiscus and E. multiflora can be linked with water-savers mechanisms. In contrast, Q. coccifera’s response can be linked with water-spenders mechanisms.

Discussion

Meteorological variables and sap flow

Results show higher R g on the south-facing slope than on the north-facing slope (Fig. 1) and a significant relationship between Q 1 and R g and VPD (Fig. 3), which can explain why most of the individuals studied on the south-facing slope presented a higher Q md values than those on the north-facing slope (Table 3). In this sense, previous works have shown, by means of simple equations or multiplicative models, that sap flow is affected by meteorological variables such as Tª, R g, PAR, rainfall and other computed variables like VPD (Badalotti et al. 2000; Infante et al. 2003; Moro et al. 2004; Wang et al. 2006; Conejero et al. 2007; Poyatos et al. 2007; Yue et al. 2008; Pfautsch et al. 2010), even at night (Fisher et al. 2007). In fact, some authors have demonstrated that south-oriented branches present higher sap flow rates than north-oriented branches because of higher exposure to radiation (Steinberg et al. 1990). Other previous works have reported a decrease in sap flow rate when VPD is higher than 1.2 kPa (Luis et al. 2005; Poyatos et al. 2008; Dzikiti et al. 2010), as shown in our results (Fig. 3) for sap flow on the south-facing slope.

On the other hand, some studies have also been presented evidences on the relationship between sap flow and weather seasons, based on the interaction between soil water availability, Tª, R g , VPD and rain distribution; which is observed more easily in regions where clear seasonal differences exist (Farrington et al. 1994). Several species (Q. ilex, Q. agrifolia, Q. durata, P. latifolia, A. unedo and P. australis) show the maximum sap flow rate in spring or early summer (Goulden 1996; Infante et al. 2003; Martínez-Vilalta et al. 2003; Moro et al. 2004) when the soil water content is optimum and the VPD does not show stress values. Minimum sap flow is linked with low soil water content (e.g. summer), or low VPD, although there is an optimum soil water content as can be observed in winter. In our study, on north-facing slope P. halepensis and P. lentiscus showed a seasonal pattern similar to previously described; while on south-facing slope these same species jointly to S. tenacissima showed a different seasonal pattern. In general, we do not observe a single seasonal pattern between all species (Fig. 5) as was reported by Martínez-Vilalta et al. (2003); in fact, Luis et al. (2005) also no found a clear seasonal pattern in P. canariensis. In our case, we think that although edaphic and meteorological factors impose similar environmental conditions during the measurement period, the species themselves determine the sap flow rate response according to their particular water use strategy, morphology, structural or functional mechanisms to cope with environmental conditions. The species diversity can produce responses diversity.

On an annual time scale, the higher annual rainfall in 1998, preceded by a very wet winter in 97/98 (Fig. 1), contributed to higher soil water content in this year (Fig. 2), which produced that most species in both slope aspect presented a higher sap flow rate in 1998 than in 1999 (Fig. 4). Similar results have been reported by Farrington et al. (1994). The high values of Tª and VPD in 1998, but with lesser value respect to threshold values of stress for these variables, also could contribute to higher the sap flow rate in 1998. In contrast, on south-facing slope is possible that spite the high soil water content, the high Tª and VPD during 1998 were limiting factor of sap flow in some species.

Drought avoidance mechanisms of species

In semiarid ecosystems, water is a limiting resource and plants are often subjected to water deficit conditions (Di Castri 1973; Kramer and Boyer 1995), as evidenced by the predominance of periods where REW values are lower than the threshold for Mediterranean species (Fig. 2), reported by Granier (1987) and Bréda et al. (1995). Under these conditions, species tend to adjust their water consumption to avoid reaching water potential values that produce a loss of hydraulic conductivity, thus preventing irreversible damage (Vilagrosa et al. 2003a). In this same sense, our results indicate that the species studied show a decrease in Q md according to the decrease in Ψpd (Fig. 6). Similar results were reported by Goulden (1996) in Q. agrifolia, Nadezhdina (1999) in Malus domestica Borkh, Stöhr and Lösch (2004) in F. excelsior; and more recently in Mediterranean environment by David et al. (2007) in Q. ilex and Q. suber and Poyatos et al. (2008) in P. sylvestris and Q. pubescent.

Under soil water deficit conditions (low Ψpd values), plants can prevent or avoid the impacts of stress in their tissues by using stomatal closure to regulate water loss (Fotelli et al. 2000; Baquedano and Castillo 2006), which, in turn, affects the daily sap flow rate (David et al. 2007; Gartner et al. 2009). In these conditions, the sap flow rate will depend on the plant’s water use strategy for coping with water deficit through the drought avoidance mechanisms (water-saving or water-spending), which evidenced in osmotic potential, elasticity of cell walls, capacity of xylem conduits and stomatal control over water loss by transpiration (Levitt 1980; Monson and Smith 1982; Kalapos 1994). A plant follows the water-spenders mechanisms when it shows high stomatal conductance associated with high water losses while maintaining high carbon gain and thus growth (Heilmeier et al. 2002). In contrast, a plant with water-savers mechanisms will display an opposite behaviour, showing strong stomatal regulation to reduce water loss. Both the water-saver and the water-spender mechanisms have been described by Levitt (1980) and have been used by several authors (Dong and Zhang 2001; Vilagrosa 2002; Llorens et al. 2003b; Sakcali and Ozturk 2004).

In this context, under conditions unaffected by water deficit (Ψpd > −1.0 MPa), P. halepensis showed high Q md values. However, under water deficit conditions (Ψpd < −2.0 MPa) this species reduced its sap flow rate and registered the lowest Q md of all the species studied (see Fig. 6); we thus classified it as a water-saver species, in agreement with Baquedano and Castillo (2006). Several authors have described P. halepensis as very sensitive to water deficit, showing a strong stomatal control under drought conditions (Martínez-Ferri et al. 2000; Baquedano and Castillo 2007), as consequence, we observed that sap flow rate decreased more rapidly in P. halepensis than in the shrubs studied (Q. coccifera and P. lentiscus) when Ψpd decreased (see Fig. 6). This result tends to confirm the fact that, while conifers regulate their transpiration from Ψpd = −1.8 MPa, at Ψpd = –4.0 MPa shrubs are still able to carry out osmotic adjustments that allow them to maintain a continuous transpiration in conditions of low soil water availability (Royce and Barbour 2001). This adaptation mechanism of P. halepensis by means of water use extremely conservative (water-saver), contribute to this species shows a high drought resistant. Thus jointly with extensive reforestation programs and its reproductive strategy after fire, allows a wide distribution of P. halepensis in the Mediterranean basin.

Quercus coccifera showed the highest daily sap flow rate under conditions affected or unaffected by soil water availability (Fig. 6). Unlike P. halepensis, this species is able to maintain a high daily sap flow rate regardless of the decline in predawn leaf water potential; showing a water-spending behaviour, which is manifested by a higher stomatal conductance under water deficit conditions (Ksontini et al. 1998; Baquedano and Castillo 2006). Other Quercus species (Q. pubescens) also tend to show a higher sap flow rate than conifer species (P. sylvestris) under drought conditions (Poyatos et al. 2008). Q. coccifera is well adapted to semiarid ecosystems, presenting a highly developed root system with shallow roots that enable it to use the water in the top soil as well as deep roots that allow it to extract water reserves from deeper layers (Rambal 1984; Baquedano and Castillo 2007). Nevertheless, some authors have assigned a water saver strategy to this species (Vilagrosa 2002; Sakcali and Ozturk 2004); this may be because these authors used seedlings or young individuals in the measurements.

Erica multiflora presents a shallow root system which favours harnessing the rainwater stored in the top soil (Filella and Peñuelas 2003); however, in a drought period this can be a disadvantage because it cannot explore and use the water sources in the deepest and wettest soil layers. In this sense, this species can show a marked reduction in the gas exchange rate in summer in the absence of changes in shoot water potential; it has thus been considered a drought avoidance species which uses water-savers mechanisms (Llorens et al. 2003b). This evaluation is in agreement with our results.

Pistacia lentiscus has been classified as a drought avoidance species, which under optimum soil water availability and by means of water-spenders mechanisms is capable of maintaining a high stomatal conductance value; however, when water deficit conditions increase, it reduces stomatal conductance and shows a drought avoidance strategy by means of water-savers mechanisms (Vilagrosa 2002). A similar behaviour has been observed in our case. This response of P. lentiscus could be due to its shallow root system, as indicated by Filella and Peñuelas (2003).

Stipa tenacissima shows high drought resistance even at very low leaf water potentials (Sánchez 1995; Haase et al. 1999; Domingo et al. 2003). It also shows an opportunistic response to pulse water (Pugnaire et al. 1996) and net water gain, probably because of water vapour adsorption (Ramírez et al. 2007). This theoretical background could explain the relatively high sap flow rates observed in this species in our study. Although we were not able to take joint measurements of sap flow and predawn leaf water potential in this species because of technical difficulties, its high daily sap flow rate, which was almost similar to the one found for Q. coccifera on the south-facing slope (Figs. 4, 5), has led us to classify S. tenacissima as a water-spender species.

In summary, we found a notable effect of meteorological variables (R g , Tª, VPD) and water deficit conditions on the sap flow rate. Several meteorological variables showed differences between slope aspect and year of monitoring. Consequently, the species on the south-facing slope showed higher sap flow rates than the species on the north-facing slope. Moreover, the most species also had higher sap flow rates during the wetter year (1998). However, a common seasonal pattern in the sap flow rate by species has been not found, and therefore do not meet one of the initially indicated hypothesis. On the other hand, we were able to link the drought avoidance mechanisms of the different species through the combined use of daily sap flow rate and predawn leaf water potential under different water deficit conditions. Based in this theoretical background, Pinus halepensis, Pistacia lentiscus and Erica multiflora were evaluated as water-saver species, while Quercus coccifera and Stipa tenacissima were evaluated as water-spender species.