Journal of Hydrology 442–443 (2012) 2–14

Contents lists available at SciVerse ScienceDirect

Journal of Hydrology

journal homepage: www.elsevier .com/locate / jhydrol

Canopy transpiration of pure and mixed forest stands with variable abundanceof European beech

Tobias Gebauer ⇑, Viviana Horna, Christoph LeuschnerPlant Ecology, Albrecht-von-Haller-Institute of Plant Sciences, Georg-August-University of Göttingen, Untere Karspüle 2, 37073 Göttingen, Germany

a r t i c l e i n f o

Article history:Received 17 November 2010Received in revised form 30 December 2011Accepted 8 March 2012Available online 28 March 2012This manuscript was handled by LaurentCharlet, Editor-in-Chief, with the assistanceof Jiin-Shuh Jean, Associate Editor

Keywords:Tree diversityFagusFraxinusSap fluxHydraulic architectureSeasonality

0022-1694/$ - see front matter � 2012 Elsevier B.V. Ahttp://dx.doi.org/10.1016/j.jhydrol.2012.03.009

⇑ Corresponding author. Present address: AppliedHamburg, Ohnhorststr. 18, 22609 Hamburg, Germany+49 40 42816 396.

E-mail addresses: tgebaue@gwdg.de, tobi(T. Gebauer), vhorna@gwdg.de (V. Horna), cleusch@

s u m m a r y

The importance of tree species identity and diversity for biogeochemical cycles in forests is not wellunderstood. In the past, forestry has widely converted mixed forests to pure stands while contemporaryforest policy often prefers mixed stands again. However, the hydrological consequences of these changesremain unclear. We tested the hypotheses (i) that significant differences in water use per ground areaexist among the tree species of temperate mixed forests and that these differences are more relevantfor the amount of stand-level canopy transpiration (Ec) than putative complementarity effects of treewater use, and (ii) that the seasonal patterns of Ec in mixed stands are significantly influenced by theidentity of the present tree species. We measured xylem sap flux during 2005 (average precipitation)and 2006 (relatively dry) synchronously in three nearby old-growth forest stands on similar soil differingin the abundance of European beech (pure beech stand, 3-species stand with 70% beech, 5-species standwith <10% beech). In summer 2005 with average rainfall, Ec was 50% higher in the beech-poor 5-speciesstand than in the two stands with moderate to high beech presence (158 vs. 97 and 101 mm yr�1); in thedry summer 2006, all stands converged toward similar Ec totals (128–139 mm yr�1). Species differencesin Ec were large on a sapwood area basis, reflecting a considerable variation in hydraulic architecture andleaf conductance regulation among the co-existing species. Moreover, transpiration per crown projectionarea (ECA) also differed up to 5-fold among the different species in the mixed stands, probably reflectingcontrasting sapwood/crown area ratios. We conclude that Ec is not principally higher in mixed foreststhan in pure beech stands. However, tree species-specific traits have an important influence on the heightof Ec and affect its seasonal variation. Species with a relatively high ECA (notably Tilia) may exhaust soilwater reserves early in summer, thereby increasing drought stress in dry years and possibly reducing eco-system stability in mixed forests.

� 2012 Elsevier B.V. All rights reserved.

1. Introduction

In Central Europe, North America and elsewhere, large areas ofnatural forest have been replaced by monocultures of productivetree species, resulting not only in a reduction of tree species diver-sity, but also in a largely modified forest structure. Because mono-cultures are typically more threatened by herbivore damage andattack by pathogenic fungi (Jactel et al., 2005) and also harbor lessdiverse plant and animal communities than more species-richstands, the forest policy of many industrialized countries hasemphasized the establishment of mixed-species stands in the pastdecades (e.g. Otto, 1992). The large-scale establishment ofmonocultures replacing natural forests in the past and the recent

ll rights reserved.

Plant Ecology, University of. Tel.: +49 40 42816 577; fax:

as.gebauer@uni-hamburg.degwdg.de (C. Leuschner).

conversion of monospecific to mixed stands may have had pro-found consequences for biogeochemical cycles and the diversityof various organism groups (Ellenberg and Leuschner, 2010).

Since more than 100 yr, foresters have been comparing the yieldof pure and mixed stands, in most cases combinations of two spe-cies (e.g. Pretzsch, 2005; Kelty, 2006; Pretzsch et al., 2010). An-other research focus exists on the susceptibility of pure andmixed stands to windthrow (review in Dhote (2005)) and nutri-tional interactions in mixed stands (e.g. Rothe and Binckley,2001). In contrast, much less is known about the consequencesof tree species mixing for the hydrological cycle in forests. System-atic comparisons between pure and two-species mixed stands withrespect to tree water use have only been conducted in very fewstudies. We are aware of only two comparative hydrological inves-tigations in temperate monospecific forests and the respectivemixed stands (Anders et al., 2006; Schume et al., 2004). Baldocchi(2005) tackled the question by relating the normalized transpira-tion rates of six temperate forest stands to tree diversity. The firsttwo investigations reported a higher water use of the mixed stands

T. Gebauer et al. / Journal of Hydrology 442–443 (2012) 2–14 3

as compared to the pure stands, while the latter one found a de-crease of evapotranspiration with increasing tree species diversity.However, the analysis of Baldocchi (2005) included stands growingunder contrasting edaphic conditions which make a direct compar-ison difficult. Thus, it remains unclear whether mixed stands differsystematically from the respective pure stands in water use andwhether they consume more or less water. More recently, thewater use of mixed and monospecific Eucalyptus/Acacia-planta-tions was compared in semiarid subtropical Australia by Forresteret al. (2010) showing a higher water use by the mixed plantation.The water use of multi-species stands composed of more than twospecies with contrasting hydraulic architecture (e.g. diffuse- vs.ring-porous) has not yet been compared to pure stands, neitherin temperate nor in other climate regions.

From a theoretical point of view, mixed stands could differ fromthe respective pure stands in water use due to at least three mech-anisms. First, species mixing could introduce a tree species withparticularly high (or low) water use per canopy projection area intothe stand, thereby increasing (or decreasing) canopy transpiration(Ec) in comparison to Ec of the focal species’ pure stand. This wouldrepresent a simple sampling effect. Evidence for species differencesin water use and water flow regulation among temperate broad-leaved trees is discussed below. Second, mixed stands could havea higher Ec than the corresponding pure stands if the different treespecies would use soil water in a complementary way, for exampleby spatial stratification of their root systems or by a differing phe-nology of their foliage. Third, a tree species mixture could increasestand-level Ec through indirect effects such as a stimulation ofnutrient cycling, thereby increasing foliar nutrient contents andconsequently leaf conductance and transpiration rate of the spe-cies in the mixture.

Numerous hydrological studies provided evidence for the exis-tence of significant tree species differences in canopy transpirationin temperate forests. Stand-level studies on canopy transpiration(Ec) using xylem sap flux measurement, or investigations on standevapotranspiration applying the eddy covariance technique,microclimatological gradient studies or soil moisture budgetingapproaches, revealed a considerable variation in the water use offorest stands composed by different temperate broad-leaved orconiferous species growing under similar edaphic and climaticconditions. A well-studied example is the transpiration rate of ma-ture European beech (Fagus sylvatica L.) stands which in most casesis higher than that of nearby planted Norway spruce (Picea abiesKarst) stands when growing on similar soil (Benecke, 1984;Bücking and Krebs, 1986). In a literature survey for Central Europe,Ellenberg and Leuschner (2010) documented different Ec values forpure Fagus, Quercus, Betula, Picea, Pinus and Larix stands. Otherauthors reported up to fourfold differences in canopy transpirationper ground area of co-existing tree species in temperate and borealforests when largely different tree functional groups (e.g. broad-leaved vs. needle-leaved or diffuse- vs. ring-porous trees) werecontrasted (Baldocchi, 2005; Ewers et al., 2002; Granier et al.,1996; Wullschleger et al., 2001).

Tree species differences in Ec are mostly the consequence of spe-cies-specific differences in (i) the area of hydroactive sapwood in thestem, (ii) xylem anatomy (ring- vs. diffuse-porous, micro- vs. macro-porous), (iii) maximum rooting depth, (iv) leaf area index, (v) thesensitivity of stomatal regulation to vpd and leaf water status, and(vi) stem density in the stand (e.g. Baldocchi, 2005; Bush et al.,2008; Ewers et al., 2002; Granier et al., 2000; Vincke et al., 2005;Wullschleger et al., 2001). In addition, if tree species differ in the leafemergence and senescence patterns during the vegetation period orin the sensitivity of their earlywood vessels to embolism, the sea-sonal course of Ec may differ significantly between the species.

In this study, we measured canopy transpiration with the xylemsap flux method after Granier (1985, 1987) in three nearby temper-

ate broad-leaved forest stands that differed in the abundance ofEuropean beech (Fagus sylvatica L.) and the presence of otherbroad-leaved tree species with contrasting traits related to wateruse such as Tilia, Fraxinus, Carpinus and Acer species (Köcheret al., 2009). The three stands represent a series from a pure beechstand to a diverse five-species stand with Fagus being increasinglyreplaced by the other tree species (from 100% to 10% contributionof Fagus to stand basal area). The stands were selected in a forestarea where a variety of forest patches with different tree speciescomposition co-exist under more or less homogenous edaphicand climatic conditions, thus allowing to isolate the effect of treespecies composition from other environmental factors that mayinfluence Ec. The stands are part of the Hainich Tree Diversity Ma-trix (Leuschner et al., 2009), a set of old-growth forest standsencompassing plots with low to high tree species numbers (1 toP5 species) in close neighborhood to each other. The remarkableheterogeneity in forest structure is the consequence of a mosaicof different former land ownerships and management practicesthat coexisted in the area for centuries (Leuschner et al., 2009).

This study has the objectives (i) to compare a pure beech standwith two mixed stands with respect to annual totals of canopytranspiration and (ii) to quantify the contribution of Fagus sylvaticaand of the other four tree species to stand transpiration in relationto their abundance in a species-rich temperate broad-leaved forest.Sap flux measurements were conducted in two consecutive yearswith contrasting precipitation amounts (average and relativelydry summer) covering 44 trees of the five species. We hypothe-sized (1) that significant differences in tree water use per groundarea exist among the tree species in the mixed stands and thatthese differences are more relevant for the amount of stand-levelEc than putative complementarity effects of tree water use, and(2) that the seasonal patterns of Ec in the mixed stands are signif-icantly influenced by the identity of the present tree species.

2. Materials and methods

2.1. Study sites and tree layer diversity

The study sites are located in the north-eastern part of HainichNational Park, Thuringia, Central Germany, between 295 and355 m a.s.l. (51�040N, 10�300E). Hainich National Park is coveredby temperate mixed broad-leaved forest dominated by Europeanbeech (Fagus sylvatica L.). Linden (Tilia cordata Mill. and T. platy-phyllos Scop.), common ash (Fraxinus excelsior L.), European horn-beam (Carpinus betulus L.) and different maple species (Acerpseudoplatanus L., Acer platanoides L. and Acer campestre L.) areco-occurring in different densities; further deciduous tree specieslike elm (Ulmus glabra L.), oak (Quercus robur L. and Q. petraea(Matt.) Liebl.), cherry (Prunus avium L.) and service tree (Sorbustorminalis L.) are interspersed in lower numbers within the forest.In the study region, hybrids of Tilia cordata and T. platyphyllos arealso occurring. Because of variable degrees of hybridization be-tween these two species, we did not differentiate between themat the species level, but refer solely to the genus Tilia.

The climate is sub-continental (Klaus and Reisinger, 1995) witha mean annual precipitation of 590 mm and 7.5 �C as mean annualair temperature (1973–2004, Deutscher Wetterdienst, Offenbach,Germany). The study year 2005 received average rainfall amounts(601 mm), while 2006 was drier than the average (518 mm, Mete-omedia AG, Germany).

The soils in the study region developed from loess which isunderlain by Triassic limestone (Muschelkalk). The loess cover var-ies between 75 and 120 cm in thickness. The soil texture in theupper 30 cm of the mineral soil is characterized by high silt(�75%) and clay contents (16–25%) but a low sand content (<5%).

4 T. Gebauer et al. / Journal of Hydrology 442–443 (2012) 2–14

The dominant soil type is a Luvisol showing stagnant propertiesduring winter and spring, while the soils are drying out stronglyduring summer (Guckland et al., 2009). The terrain is level toslightly sloping (0–5�) and located far above the ground water level(Leuschner et al., 2009).

The study was conducted in three stands of the Hainich TreeDiversity Matrix (Leuschner et al., 2009), a set of forest plots withdifferent tree species composition and tree species diversity thatserve for comparative ecological research in species-poor and spe-cies-rich forests. The three study plots with a size of 50 m � 50 mrepresent a sequence of stands with decreasing abundance ofbeech and increasing abundance of linden, ash and other broad-leafed species. The stand DL1a is a pure beech stand with >95%of the canopy trees belonging to Fagus, stand DL2c is a 3-speciesstand dominated by beech, but with about 30% admixture of Frax-inus and Tilia, and stand DL3a is a species-rich (5 species) standwith dominant Tilia and sparse contribution of Fraxinus, Carpinus,Acer and Fagus (<10% of the canopy trees; Table 1, Fig. 1). The de-crease in beech dominance is associated with an increase in treespecies diversity (the Shannon diversity index H0 is 0.31 in DL1a,0.82 in DL2c and 1.16 in DL3a). H0 considers the number of treespecies present in a stand and their relative abundance in termsof stems (DBH >7 cm) reaching the upper canopy.

All stands had a closed canopy (gap fraction <0.1) and roughlycomparable stand basal areas (36–45 m2 ha�1) and mean tree ages(83–116 yr, Schmidt et al., 2009). Leaf area index (LAI, unit:m2 m�2) varied between 6.5 and 7.3 in 2005, and between 6.5and 7.6 in 2006 (Jacob et al., 2010). While the DL2c plot had a con-siderably higher total number of stems per hectare than the DL1aand DL3a plots due to abundant beech trees in the subcanopy layer,the number of tree individuals participating in the upper canopylayer was more similar among the stands (188–376 trees ha�1).

Tree selection for sap flux measurement followed the objectiveto reach at reliable estimates of stand-level transpiration at thethree plots and to compare the water use of the species present.We selected stems of the five species in numbers and dimensionsthat were representative for the relative abundance of the speciesand stem diameter (DBH) classes in the three stands, following rec-ommendations made by Granier et al. (1996). In the DL1a plot, 8beech trees were instrumentated, in DL2c 8 beech, 3 linden and5 ash trees (16 in total), and in DL3a 3 beech, 8 linden, 3 ash, 3 ma-ple and 3 hornbeam trees (20 in total). Thus, the total number ofmeasured trees in the stands increased with stand diversity from8 in DL1a to 20 in DL3a, adding up to 44 in the whole study. Alltrees were individuals that reached the middle or upper canopy(tree height 25–35 m).

2.2. Stand structure

The diameter at breast height, basal area at 1.3 m height (AB),and projected crown area (CAp) of all trees in the 2500 m2 plotswere recorded by DBH measurements with dendrometer tapesand by determining the crown radii in 8 directions (8-point crownprojections). LAI and tree height data were taken from Jacob et al.(2010). The LAI of the plots was calculated from the leaf biomasscollected in autumn in each 10 litter traps per stand. The leaveswere sorted by species and measured for size, dried and weighed.LAI was calculated by multiplying mean specific leaf area (SLA) withleaf mass for all species present. The hydroactive sapwood area (AS)at breast height was calculated from relationships between DBHand AS, that had been established earlier for the five investigatedtree species in the same stands by dyeing and subsequent woodcoring (Gebauer et al., 2008). Basal area index (BAI), sapwood areaindex (SAI) and crown area index (CAI) (units: m2 ha�1) were calcu-lated from stand- and species-specific AB, AS and CAp values dividedby ground area (AG). The phenologies of the tree species were

inspected regularly in both study years for determining the exactlength of the vegetation period in 2005 and 2006.

2.3. Sap flux measurements

We measured xylem sap flux density (Js, unit: g m�2 s�1) in thestem xylem using Granier-type heat dissipation sensors (Granier,1985, 1987) in trees >10 cm in diameter at 1.3 m height aboveground. Pairs of 20 mm-long and 2.0 mm-wide heating probeswere inserted in northern and southern trunk directions into thestem sapwood. For the ring-porous species ash, probes with10 mm heating spiral length were used because of the smaller sap-wood thickness compared to the diffuse-porous species. Theprobes were manufactured according to the original design proto-col given by A. Granier (1996, and pers. communication). The twopaired sensors were identical in construction. The upper probe washeated with constant current of 0.12 A and a heating power of0.2 W. The lower probe was unheated and served as a referenceto the upper probe. The distance between the two probes wasabout 15 cm whereby thermal interference especially at zero sapflux should be avoided. The temperature difference between thetwo probes was recorded with copper-constantan thermocouplesplaced at the centre of the heating spirals every 30 s with a datalogger (CR10X; Campbell Scientific Ltd., UK) equipped with a 16/32-channel multiplexer (AM16/32, AM416; Campbell ScientificLtd., UK). 30-min averages were calculated from the 30-s readingsand stored in the data logger. The temperature difference was usedto calculate sap flux density Js (in g m�2 s�1) according to theempirical calibration equation given by Granier (1985, 1987):

Js ¼ 119 � K1:231 ð1Þ

where K = (DTM–DT)/DT�DTM is the maximum temperature differ-ence when sap flux is assumed to be zero. In general, DTM was cal-culated for every day from the predawn temperature readings,given that the VPD data indicated zero flux or very low flux in thenight. When microclimatic data indicated significant nighttime fluxto occur, DTM was calculated by averaging the DTM values of thedays before and after that day (compare Lu et al., 2004).

In both summers (2005 and 2006), sap flux was measured at thesame trees. The quality of sensor readings was regularly controlledand, if necessary, sensors were replaced.

2.4. Canopy transpiration and canopy conductance

Up-scaling of sensor-level sap flux Js to whole-tree sap flow re-quires information on the sapwood cross-sectional area (AS) of themeasured tree which was estimated from relationships betweenDBH and AS established for 12–25 trees per species in these foreststands by Gebauer et al. (2008). Furthermore, the radial patterns ofxylem sap flux density within the hydroactive xylem were ob-tained by analyzing species-specific radial flux profiles in the xy-lem of 1–3 stems each of the five species with sensors placed infour different depths of the sapwood (Gebauer et al., 2008). Theflux data were expressed as relative flux density along the sap-wood profile (scaled in relative units) and applied to all stemswhere sap flux was only measured in the outermost xylem. The re-sults of the dyeing experiments were used for estimating the sap-wood depth of the studied trees. Based on this information, wecalculated mean tree xylem flux density Jst as follows:

Jst ¼Pn

n¼1Js � BSðxiÞ �WðxiÞAS

ð2Þ

where Jst is the mean sap flux density in the entire sapwood of a tree(g m�2 sapwood s�1), Js is the mean sap flux density at the outermostsensor position (0–2 or 0–1 cm of xylem depth), BS (xi) is the area of

Table 1Species composition and stand structural characteristics of the three study plots. Mean tree age was determined in wood cores (Schmidt et al., 2009). Stem density includes alltrees >7 cm in DBH. Shannon-Wiener index (H0) bases on stem density data. Leaf area index (LAI) for 2005 and 2006 was estimated from leaf litter samples (Jacob et al., 2010).H0 – Shannon-Wiener index; BAI– basal area index; SAI – sapwood area index; CAI – crown area index; LAI – leaf area index.

Plot Treespecies

Meanage (yr)

Stem density (alltrees) (ha�1)

Stem density (uppercanopy trees) (ha�1)

H0 Mean treeheight (m)

BAI(m2 ha�1)

SAI(m2 ha�1)

CAI(m2 ha�1)

LAI (m m�2) Soil profiledepth (cm)

2005 2006

DL1a Stand 109(±12.0)

428 232 0.31 33.3 (±2.2) 46.13 33.35 12446.8 6.5 7.3 120

Fagussylvatica

400 224 44.01 31.78 11753.2

DL2c Stand 83(±17.8)

776 376 0.82 29.2 (±2.1) 45.00 29.02 14874.0 6.6 6.5 120

Fagussylvatica

572 264 30.12 22.45 10832.4

Tilia sp. 84 40 5.68 4.48 2050.4Fraxinusexcelsior

100 64 7.98 1.62 1473.2

DL3a Stand 116(±16.8)

392 188 1.16 27.4 (±1.9) 35.73 23.64 12557.6 7.3 7.6 75

Fagussylvatica

12 12 3.76 2.67 875.6

Tilia sp. 264 108 19.59 13.58 6851.6Fraxinusexcelsior

28 20 3.35 1.31 1163.6

Carpinusbetulus

36 20 3.58 2.42 1770.4

Acer sp. 32 12 2.26 1.65 724.8

T. Gebauer et al. / Journal of Hydrology 442–443 (2012) 2–14 5

concentric rings of 1 cm width between the cambium (xi) and theheartwood boundary (xi+n), where sap flux reaches zero (unit: m2),n is the number of rings with index i, W (xi) is a species-specific pro-portionality factor (unitless) which expresses flux density at a givensapwood depth (xi) in relation to flux density at the outermost sen-sor position, and AS is the cross-sectional sapwood area of the tree. Wwas obtained from 4-parametric Weibull functions fitted to the ra-dial flux profile data of Gebauer et al. (2008).

Various methods are available for up-scaling from mean treewater flux to stand-level flow (Cermak et al., 2004). We used sap-wood area index (SAI) for extrapolation, which relates cumulativesapwood area to ground area (unit: m2 ha�1), and did this separatelyfor the various DBH classes and different species. Before calculation,Jst was integrated over a day by multiplying the half-hour meanvalues by 1800 and adding them to obtain values in the unitg m�2 sapwood d�1. The daily stand-level transpiration rate Ec (in -mm d�1) of a given species was then calculated for m DBH classes as:

Ecj ¼ASj

AG

� �� 1m

Xm

m¼1

Jst ð3Þ

with Ecj being the daily canopy transpiration of the DBH class j of aspecies, 1

m

Pmm¼1Jst the daily mean tree sap flux (in g m�2 sap-

wood d�1) of the DBH class j of a species in the stand, ASj the cumu-lative sapwood area of all stems of a species in the DBH class j, andAG the ground area of the plot. The different DBH classes weresummed to give the species-specific canopy transpiration of thestand. The total canopy transpiration of the stand (Ec) was then cal-culated as the sum of the Ecj values of all species being present inthe plot (unit: mm d�1). For days where data was missing, Ec valueswere estimated using the regression of single tree Jst on a daily basiswith corresponding vapor pressure deficit (VPD) values and furtherfollowing the procedure for up-scaling presented above. All Ec datawere related to the length of the vegetation period which extendedin the Hainich forest from about April 20 to October 31 (unit: mm orL m�2 per vegetation period).

In order to compare the species with respect to their water use,we related the canopy transpiration of a species to the respectivespecies-specific CAI, BAI or SAI values in the stand.

The canopy conductance (Gc) was calculated from Ec and micro-meteorological data according to Köstner (2001) using theexpression

Gc ¼Ec

VPDqwGvTK ð4Þ

where Ec is the measured canopy transpiration in mm d�1, VPD thevapor pressure deficit of the air (kPa), qw the density of water(998 kg m�3), Gv the gas constant for water vapor(0.462 m3 kPa kg�1 K�1) and TK the air temperature (Kelvin).

2.5. Microclimatological and hydrological measurements

Microclimatological (air temperature, relative air humidity (RH)and atmospheric vapor pressure deficit (VPD)) and precipitationdata were obtained from the Weberstedt/Hainich meteorologicalstation (station No. 9460, Meteomedia AG, Germany) located2 km northeast from our study plots. Incident shortwave radiation(R) data were taken from satellite measurements regionalized tothe study region (Meteosat). All variables were recorded at hourlyintervals.

Volumetric soil water content (h) was recorded half-hourly inthe three stands at depths of 10, 20, and 30 cm by Krämer andHölscher (2010) using EnviroSCAN FDR sensors (Sentek Pty Ltd.,Stepney, Australia). The measurements started in June/July 2005and were continued throughout the whole year 2006. These datawere used to calculate the relative extractable water (REW) inthe soil profiles of the three stands using Eq. (4) (Bréda et al.,1995, 2006; Granier 1987; Granier et al., 1999):

REW ¼ ðh� hminÞðhFC � hminÞ

ð4Þ

where h is the actual soil water content, hmin the minimum soilwater content observed in the years 2005 and 2006, and hFC the soilwater content at field capacity. Field capacity (FC) was estimatedfrom laboratory desorption curves characterizing the water content– water potential relationship with �100 hPa being defined as FC(U. Talkner, unpublished data). REW was uniformly calculated for

Fig. 1. Maps of stem positions (dots) and projected crown areas of the trees in the study plots DL1a, DL2c and DL3a (GIS maps created by K.M. Daenner). Plot size was50 m � 50 m. Shaded crowns are the trees instrumented with sap flux sensors.

6 T. Gebauer et al. / Journal of Hydrology 442–443 (2012) 2–14

the 0–30 cm soil profile in the three stands; this soil layer contained>80% of the profile total of fine root biomass (Meinen et al., 2009).

2.6. Data analysis

To test for significant differences in Ec between the pure andmixed stands, we applied a model II simple linear regression anal-ysis combined with the major axis method using the lmodel2 func-tion of the statistical software R, version 2.8.1 (R Development CoreTeam, Vienna, Austria; Legendre and Legendre, 1998; Legendre,2008; Warton et al., 2006) in order to obtain a quantitative mea-sure of similarity or dissimilarity between the sets of Ec daily totalsof the three plots. A model II regression was selected because bothsets of variables (Ec values of two plots to be compared) are ran-dom. In the lmodel2 function, a permutation test is included todetermine the significance of differences in the slopes of the majoraxis method. For the years 2005 and 2006, we analyzed the slopesof linear regression fits of daily Ec totals of one plot on the daily Ec

totals of another plot (i.e., DL1a vs. DL2c; DL1a vs. DL3a; DL2c vs.DL3a) and contrasted the slopes with the 1:1 line (45�) whichstands for complete congruity of the Ec time courses of the twoplots compared. The analysis gave parametric 95% confidenceintervals (C.I.) for the regression slopes in all six possible plot com-binations and the p-values and coefficient of determination (R2) ofthe respective regressions.

In a second analysis, we compared the patterns of temporalchange in daily canopy transpiration in the first half of the vegeta-tion period from early May (leaf flushing) to July 2006 (peak valuesof Ec) in more detail. Following Legendre (2008), we calculated inthis model II regression analysis the inverse of the mean slope fac-tor of the 3-month-period (unit: the inverse of mm d�1/d = d2 mm�1) and compared their upper and lower 95% confidenceintervals and R2 values among the three plots in order to detect sig-nificant differences in the seasonal evolution of Ec in the threestands (second hypothesis). To obtain a quantitative measure ofthe size of species effects on Ec (first hypothesis), we partitioned

Fig. 2. Seasonal course of incident global radiation (daily totals) and atmospheric water vapor pressure deficit (VPD, daily means) at the Weberstedt meteorological station(upper panel) and daily precipitation totals together with the seasonal course of relative extractable water (REW) in the 0–30 cm profile in the plots DL1a, DL2c and DL3a inthe period April to November 2005 (lower panel). The dotted line indicates 0.4 � REW which is thought to represent a critical minimum threshold of soil water availability(REWc) in temperate forests (Granier et al., 1999, 2007; Bernier et al., 2002). No soil water content data were available before mid of June 2005.

T. Gebauer et al. / Journal of Hydrology 442–443 (2012) 2–14 7

the daily totals of canopy transpiration to the species level andanalyzed the seasonal change in the contribution of the variousspecies.

3. Results

3.1. Climatic conditions

The two study summers 2005 and 2006 differed with respect toVPD, incident radiation and total amount of precipitation receivedin the vegetation period (363.5 and 314.5 mm of rainfall, respec-tively, in the period April 20 to October 31) (Figs. 2 and 3, upperpanels). 2005 was an average year with continuous precipitationduring the whole vegetation period. VPD did never exceed1.3 kPa. Rainfall was lower and rainfall distribution much moreirregular in summer 2006 with pronounced rainless periods occur-ring in June, July and September. The dry spells were related toperiods of elevated vapor pressure deficits with VPD maximareaching 2.1 kPa in July 2006. In contrast, August 2006 receivedmore rainfall than the long-term average. Volumetric soil watercontent decreased more or less continuously during the summerof 2005 on the three plots and reached seasonal minima of 13–20 vol.% in September and October. The more regular precipitationdistribution during this year resulted in a less extreme depletion ofthe extractable water resources. In 2006, the moisture reductionwas more rapid in the first half of the summer with minima of11–20 vol.% already appearing in early August. High rainfall in Au-gust 2006 resulted in the temporal refilling of the soil water re-serves that were depleted again in an extended rainless period inSeptember. Throughout the summers of both years, the DL1a plotwith the monospecific stand showed a higher soil water contentthan the DL2c and DL3a plots (Figs. 2 and 3). The moisture decrease

in rainless June 2006 was more pronounced under the species-richest stand DL3a than in the other two stands (Fig. 3).

3.2. Canopy transpiration

Up-scaling from tree to stand level gave totals of canopy tran-spiration (Ec) for the vegetation periods 2005 and 2006 between97 and 158 mm in the three stands (Table 2). With regard to ourfirst study objective, we found no clear evidence of a principal in-crease or decrease of Ec with a large change in tree species compo-sition between the pure and mixed stands. In the summer 2005without a pronounced rainless period, Ec was by 50% higher inthe beech-poor species-rich stand DL3a (158.4 mm) than in themonospecific and the DL2c mixed stand (97.3 and 100.6 mm). Asimilar difference between the species-richest stand DL3a andthe two stands with higher beech dominance existed for mean dai-ly Ec (0.89 mm d�1 vs. 0.52 and 0.54 mm d�1) and maximum dailytranspiration (2.45 mm d�1 vs. 1.14 and 1.25 mm d�1). In compar-ison to 2005 (190 days), the vegetation period (bud burst to leaffall) was 17 days longer in 2006. In this summer with several ex-tended dry spells, we calculated similar transpiration rates in thevegetation period for all three stands (128–139 mm). The 3-speciesstand DL2c with medium abundance of beech showed the highestannual canopy transpiration (139.3 mm) of the three stands, eventhough the observed maximum daily transpiration rates were low-er in this stand (1.52 mm d�1) than in the two others (1.91 and1.75 mm d�1).

Highest daily transpiration occurred in the pure stand DL1a andthe DL2c stand with medium beech abundance in July and August2005, while the peak occurred about 4 weeks earlier in June/July2005 in the beech-poor species-rich stand DL3a (Fig. 4). Similarlyin 2006, Ec peaked earlier in the DL3a stand (around end of June)

Fig. 3. Seasonal course of incident global radiation (daily totals) and atmospheric water vapor pressure deficit (VPD, daily means) at the Weberstedt meteorological station(upper panel) and daily precipitation totals together with the seasonal course of relative extractable water (REW) in the 0–30 cm profile in the plots DL1a, DL2c and DL3a inthe period April to November 2006 (lower panel). The dotted line indicates 0.4 � REW.

8 T. Gebauer et al. / Journal of Hydrology 442–443 (2012) 2–14

than in the DL1a and DL2c stands (mid of July, Fig. 5). Comparingthe initial slopes (May–July) of the seasonal increase in daily tran-spiration rates in 2006 among the three stands revealed that theDL1a and DL2c stands had similar time courses of canopy transpira-tion with a slow but steady increase of Ec over time (lower andupper limits of the 95% confidence interval of daily Ec: 0.81–0.99and 0.79–0.95 mm d�1, respectively, Table 3b). In the beech-poorDL3a stand, the increase in Ec was much steeper in early summer(confidence interval limits for Ec: 1.09–1.19 mm d�1). In 2006, high-er water losses early in summer were reflected by a more rapid dropof soil water content in late June and early July in stand DL3a ascompared to the other stands: the critical threshold of 0.4 REW(Granier et al., 1999, 2007; Bernier et al., 2002) was reached about10 days earlier (July 2 or 3) in DL3a than in the two stands withhigher beech dominance (July 11 and 13) (Fig. 3). Correspondingdata for 2005 do not exist in the case of stand DL3a.

Table 3 summarizes the results of regression analyses con-ducted to compare the slopes of linear regressions of daily Ec totalsof one stand on the Ec of another stand in order to compare thetranspiration rates of the three stands. From the confidence inter-vals of the slopes, it is evident that the daily transpiration rateswere significantly different between the three stands in both2005 and 2006 (Table 3a). Only in summer 2006, the lower C.I. ofthe slope was close to 45� (which is the 1:1 line of the relationship)indicating that the stands DL1a and DL2c had rather similar canopytranspiration rates during the growing season. In summer 2005, Ec

was higher in the beech-poor stand DL3a than in the DL1a andDL2c stands, while in 2006, the daily transpiration rates in theDL1a, DL2c and DL3a stands were similar to each other, when com-paring the whole vegetation period. During the first half of the veg-etation period in 2006, the species-richest stand DL3a experiencedan earlier increase and the maximal Ec rate was reached earlierthan in the stands DL1a and DL2c with higher beech dominance(May–July, Table 3b, Fig. 5).

The mean canopy conductances (Gc) of the DL1a, DL2c and DL3astands were 1.84 (±0.08), 1.85 (±0.11) and 2.7 (±0.08) mm s�1 withmaximum values of 6.9, 9.9 and 7.7 mm s�1 in the relatively moistsummer 2005. In the drier summer of 2006, Gc reached two timeshigher Gc means than in 2005 (3.3, 3.3 and 4.4 mm s�1 with max-ima of 13.4, 7.8 and 19.8 mm s�1 for the 1-, 3- and 5-speciesstands). While the DL1a and DL2c stands were remarkably similarin their Gc values during both summers, higher conductances werecalculated for the species-rich DL3a stand in June 2005 and in earlysummer 2006 (end of April to early July) (Fig. 6).

3.3. Relative contribution of the five tree species to canopytranspiration

The relative contribution of the five tree species to canopy tran-spiration in the mixed stands DL2c and DL3a deviated considerablyfrom the relative importance of these species in the stands whencrown area index (CAI), sapwood area index (SAI) and basal areaindex (BAI) are considered (Tables 1 and 2). Trees of the genus Tiliacontributed more to canopy transpiration than would be expectedfrom its relative crown, sapwood or basal area in stand DL2c (and,to a lesser extent, also in stand DL3a, Table 2). Similarly, Acer sp. instand DL3a had much higher Ec/CAI-, Ec/SAI- and Ec/BAI-quotientsthan the average of all stems in the stand. In contrast, Fraxinusexcelsior tended to transpire less than the stand average in themixed stands, as did Fagus sylvatica in stand DL2c in 2006. Whennormalized to the respective canopy projection area, the four dif-fuse-porous species differed more than fivefold in their water useduring the vegetation period (57–290 L m�2); ring-porous ashhad an about ten times smaller Ec rate per canopy projection areathan the average of the diffuse-porous species. The variation be-tween the species was even larger when Ec was normalized to ba-sal area (20.500–104.800 L m�2 yr�1 for the diffuse-porous speciesand 4.800–16.700 L m�2 yr�1 for F. excelsior). When related to

Tabl

e2

Cano

pytr

ansp

irat

ion

(Ec)

ofal

lind

ivid

uals

ofa

spec

ies

and

ofth

ew

hole

stan

din

the

year

s20

05an

d20

06in

the

plot

sD

L1a,

DL2

can

dD

L3a.

The

tran

spir

atio

nra

tes

are

also

expr

esse

din

rela

tion

toba

sala

rea

inde

x(B

AI)

,sap

woo

dar

eain

dex

(SA

I)an

dcr

own

area

inde

x(C

AI)

(uni

tsof

indi

ces:

m2

ha�

1).

The

last

colu

mn

give

sca

nopy

tran

spir

atio

nas

apr

opor

tion

ofpr

ecip

itat

ion

duri

ngth

eve

geta

tion

peri

od(P

v)fo

rbo

thye

ars.

The

indi

ces

BAI,

SAIa

ndCA

Iwer

eca

lcul

ated

from

stan

d-an

dsp

ecie

s-le

vel

AB,A

San

dCA

pva

lues

each

divi

ded

bygr

ound

area

(AG

).

2005

2006

Plot

Tree

spec

ies

E c (mm

)Pr

opor

tion

ofst

and

E c(%

)E c

/BA

I(�

103

Lm�

2)

E c/S

AI

(�10

3L

m�

2)

E c/C

AI

(Lm�

2)

E c/P

v

(mm

mm�

1)

E c (mm

)Pr

opor

tion

ofst

and

E c(%

)E c

/BA

I(�

103

Lm�

2)

E c/S

AI

(�10

3L

m�

2)

E c/C

AI

(Lm�

2)

E c/P

v

(mm

mm�

1)

DL1

aSt

and

100.

610

0.0

21.8

30.2

80.8

0.28

134.

310

0.0

29.1

40.3

107.

90.

41Fa

gus

sylv

atic

a99

.999

.322

.731

.485

.013

3.4

99.3

30.3

42.0

113.

5

DL2

cSt

and

97.3

100.

021

.633

.565

.40.

2813

9.3

100.

031

.048

.093

.70.

42Fa

gus

sylv

atic

a61

.763

.520

.527

.557

.066

.647

.822

.129

.761

.5

Tilia

sp.

22.4

23.1

39.4

50.0

109.

359

.542

.710

4.8

132.

829

0.2

Frax

inus

exce

lsio

r13

.113

.416

.680

.988

.913

.29.

516

.781

.589

.6

DL3

aSt

and

158.

410

0.0

44.3

67.0

126.

10.

4712

8.0

100.

035

.854

.110

1.9

0.41

Fagu

ssy

lvat

ica

15.6

9.8

41.5

58.4

178.

215

.712

.341

.858

.817

9.3

Tilia

sp.

105.

366

.553

.877

.515

3.7

78.1

61.0

39.9

57.5

114.

0Fr

axin

usex

cels

ior

1.9

1.2

5.7

14.5

16.3

1.6

1.3

4.8

12.2

13.8

Carp

inus

betu

lus

18.4

11.6

51.4

76.0

103.

917

.313

.548

.371

.597

.7

Ace

rsp

.17

.110

.875

.710

3.6

235.

915

.312

.067

.792

.721

1.1

T. Gebauer et al. / Journal of Hydrology 442–443 (2012) 2–14 9

hydroactive sapwood area, Ec varied more than fourfold among thediffuse-porous species, but the ranking among the species changedaccording to their sapwood area/basal area ratios.

The Ec/CAI-, Ec/SAI- and Ec/BAI-quotients of the species differedmarkedly between the two hydrologically different years. Largeinterannual differences in the relative contribution to stand Ec oc-curred in Tilia, which had a much higher transpiration per crownarea in stand DL2c in the summer 2006 as compared to 2005,and in F. sylvatica and F. excelsior, which tended to transpire rela-tively less in 2006 than in 2005. However, when related to theircontribution to stand crown area, sapwood or basal area, the inter-annual differences were small in these two species.

4. Discussion

4.1. Canopy transpiration of pure and mixed stands and the possiblerole of species diversity

While our data support hypothesis 1, which postulates signifi-cant differences in Ec per ground area among the species, the evi-dence for a principally higher canopy transpiration of the mixedstands was contradictory. We found a ca. 50% higher Ec in thebeech-poor species-rich DL3a stand in summer 2005 as comparedto the pure DL1a stand and the DL2c stand with moderate beechabundance, but calculated similar transpiration rates for all standsin the relatively dry summer 2006. Thus, the interannual variationin Ec was large and stand differences in canopy transpiration seemto be visible only in certain years. Data from several further sum-mers would be needed to confirm a putative positive effect of cer-tain tree species (such as Tilia sp.), or of tree species diversity, onstand-level Ec in the Hainich forest.

If complementary resource use in mixed forests leads to a high-er productivity of species-rich than monospecific stands, forestwater consumption could, in theory, also be higher in mixed thanin pure stands. Baldocchi (2005) conducted a meta-analysis of for-est evapotranspiration data from the FLUXNET program to test thishypothesis (see also Baldocchi et al. (2001)). However, this analysisrevealed a decrease, and not an increase, in evapotranspirationwith increasing number of tree species per stand, which he ex-plained with an assumed greater proportion of ring-porous speciesin species-richer stands. Another explanation given by the authorrefers to the nitrogen economy of mixed stands, which could leadto a reduced stomatal conductance (Schulze et al., 1994) and leafarea index in the species-rich forests.

Species-rich stands could also transpire more than species-poorforests if tree species with high mean leaf conductances or largeleaf areas were present, which could be species with deep-reachingroots. In the Hainich forest, we can rule out an effect of differentialrooting patterns of the species on tree water use, because we foundno indication of a vertical segregation of the fine root systems ofthe coexisting species, but a highly coherent depth distributionof the roots of the five species (Meinen et al., 2009). The absenceof significant root system stratification in this mixed forest is prob-ably caused by the rather high soil bulk density and a rather shal-low soil depth which effectively hinder the exploration of deepersoil layers. This makes complementary use of soil water by differ-ent species unlikely. Species differences in water absorption andtransport, which could lead to elevated canopy transpiration rates,are discussed below.

We found some indication for a higher canopy conductance ofthe species-rich DL3a stand early in the season in 2006 (and partlyalso in 2005), which most likely was caused by species differencesin leaf conductance and not by differences in canopy roughness be-tween the three stands. In fact, the spatial heterogeneity of thecanopy of the three stands was similar as shown by similar

10 T. Gebauer et al. / Journal of Hydrology 442–443 (2012) 2–14

coefficients of variation of tree height in the plots (for dominantand co-dominat trees: 6.5%, 7.2% and 6.7% in the DL1a, DL2c andDL3a stands, respectively).

4.2. Species effects on canopy transpiration

The different tree species in the mixed stands contributed notequally to canopy transpiration. Acer sp. and Tilia sp., in general,had a relatively high transpiration per projected crown area, F.excelsior relatively low specific Ec rates. In fact, not only diffuse-and ring-porous species, but also different diffuse-porous treeswere found to differ largely in their water use on a sapwood area,basal area or canopy projection area basis, supporting our firsthypothesis. Hölscher et al. (2005) also found large species differ-ences in water use during their measurements in the Hainich forestin 2001 when the sap flux densities were related to the hydroactivesapwood area.

However, the species contribution to canopy transpiration wasnot constant but apparently varied between the stands and years.For example, the canopy transpiration per sapwood area of F. sylv-atica was higher in the beech-poor species-rich DL3a stand than inthe DL2c stand (58 vs. 27–30 � 103 L m�2 yr�1) and reached muchhigher values for Tilia in the DL2c stand in 2006 than in 2005 (133vs. 50 � 103 L m�2 yr�1). Since available energy and soil moistureresources are more or less similar in the three stands, the differ-ences in sap flux density per sapwood area between the threestands may partly reflect differences in the hydraulic conductanceof the soil-to-leaf conducting pathway or in canopy conductanceamong different tree individuals of the same species, as were re-ported by Köcher et al. (2009) in the Hainich mixed forest. Anotherinterpretation of differences in canopy transpiration of F. sylvaticabetween the DL2c and DL3a stands could be that the few beechtrees in the DL3a stand possess a particularly large crown area withhigh transpirative water loss, probably caused by a high input ofincident radiation.

DL3a

0.0

0.5

1.0

1.5

2.0

DL2c

E c (m

m d

-1)

0.0

0.5

1.0

1.5

2.0

DL1a

0.0

0.5

1.0

1.5

2.0

Apr May Jun Jul

Fig. 4. Annual course of canopy transpiration (Ec) in the three plots DL1a, DL2c and DL3stand Ec.

In our study, the Acer species (mostly A. pseudoplatanus and A.platanoides) and the Tilia species (T. cordata and T. platyphyllos)exhibited higher canopy transpiration rates than the stand average,whereas Fraxinus excelsior showed substantially smaller canopytranspiration per ground area, at least in the beech-poor standDL3a. Plant traits which have been found to cause tree species dif-ferences in water use at the stand level include tree age and treeheight (Köstner, 2001; Köstner et al., 1998, 2002; Mencuccini andGrace, 1996; Roberts, 2000; Ryan et al., 2000; Schäfer et al.,2000; Vertessy et al., 1994, 1995, 1997), the anatomy of the con-ducting tissue (e.g. Oren and Pataki, 2001), leaf area (Ellenbergand Leuschner, 2010), and leaf conductance (e.g. Köcher et al.2009). Wullschleger et al. (1998, 2001) proposed that the sapwoodarea per unit ground area of a species should be a key trait deter-mining Ec. However, in the Hainich forest, the monospecific beechstand (DL1a) had a larger cumulative sapwood area (33.4 m2 ha�1)than the 3-species (29.0 m2 ha�1) and the 5-species (23.6 m2 ha�1)mixed stands, while the summed canopy projection areas of thetrees in these stands were more similar (1.26–1.49 ha ha�1). Thus,the relatively high transpiration rate of the beech-poor stand DL3ain 2005 is not explained by a particularly large sapwood area, thuscontradicting the assumption of Wullschleger et al. (1998, 2001).

Oren and Pataki (2001) concluded that forest stands composedmainly of ring-porous species have a smaller stand sapwood area, alower mean canopy conductance and, thus, a smaller canopy tran-spiration than stands composed primarily of diffuse-porous specieswith a larger sapwood area. The significant contribution of ring-porous F. excelsior in the DL2c and DL3a stands with 100 and28 stems ha�1, respectively, may partly explain the lower sapwoodareas of these stands as compared to the monospecific Fagus standDL1a. If the conclusion of Oren and Pataki (2001) is more generallyvalid, species-rich stands should exhibit higher sapwood areas andalso transpiration rates only when diffuse-porous trees are abun-dant in the mixed stands and ring-porous trees are dominatingthe species-poor stands. This would be a rare case in temperate

2005

Fagus sylvaticaTilia sp.Carpinus betulusAcer sp.Fraxinus excelsior

Fagus sylvaticaTilia sp.Fraxinus excelsior

Fagus sylvatica

Aug Sep Oct Nov

a during 2005. Different hatching indicates the contribution of different species to

DL3a

0.0

0.5

1.0

1.5

2.0 Fagus sylvaticaTilia sp.Carpinus betulusAcer sp.Fraxinus excelsior

DL2c

E c (m

m d

-1)

0.0

0.5

1.0

1.5

2.0 Fagus sylvaticaTilia sp.Fraxinus excelsior

DL1a

0.0

0.5

1.0

1.5

2.0Fagus sylvatica

Apr May Jun Jul Aug Sep Oct Nov

2006

Fig. 5. Annual course of canopy transpiration (Ec) in the three plots DL1a, DL2c and DL3a during 2006. Different hatching indicates the contribution of different species tostand Ec.

Table 3Results of two statistical tests on differences between daily transpiration totals (Ec) in the three stands (DL1a, DL2c, DL3a). (a) Simple linear regression analysis, model II withmajor axis method, on the degree of similarity between Ec in the three plots (6 plot combinations in 2 yr). Given are the lower and upper confidence intervals (C.I.) of the slopes ofthe regression lines between the Ec of one plot on the Ec of a second plot for the years 2005 and 2006 (in degrees, 45 degrees is the 1:1 line) together with the p-value and R2 forthe regressions. (b) Detailed regression analysis for the early-summer period 2006 in the three stands presenting the lower and upper C.I. of the regression slope (given as theinverse in the unit: d2/mm) together with the p-value and R2 of the regressions. The lower and upper C.I. of the Ec values itself is also presented (in mm d�1).

(a)

Plot combinations Lower and upper C.I. of regression slope (�) n p-value R2

DL1a–DL2c (2005) 33.72–38.72 168 <0.0001 0.74DL1a–DL3a (2005) 13.14–21.40 154 <0.0001 0.55DL2c–DL3a (2005) 25.45–30.47 146 <0.0001 0.71DL1a–DL2c (2006) 45.08–48.38 206 <0.0001 0.85DL1a–DL3a (2006) 33.26–44.39 160 <0.0001 0.39DL2c–DL3a (2006) 24.64–38.35 160 <0.0001 0.26

(b)

Plots Lower and upper C.I. of regression slope n p-value R2 Lower and upper C.I. ofEc values (in mm d�1)

DL1a (May–July 2006) 0.0068–0.0132 72 <0.0001 0.36 0.87 0.93DL2c (May–July 2006) 0.0078–0.0133 73 <0.0001 0.45 0.84 0.89DL3a (May–July 2006) 0.0163–0.0253 56 <0.0001 0.62 1.05 1.12

T. Gebauer et al. / Journal of Hydrology 442–443 (2012) 2–14 11

forests. Indeed, we are not aware of data supporting the existenceof higher sapwood areas in mixed stands as compared to monospe-cific stands of the same species. Finally, different cavitation risks ofring- and diffuse-porous trees could also affect xylem flow patternsat high VPD, thus resulting in different Ec rates of the two hydrau-lically different types of trees (Bush et al., 2008).

Investigations by Köcher et al. (2009) in the Hainich forestshowed that A. pseudoplatanus trees had a much higher hydraulicconductivity in the soil-to-leaf flow path than the other specieswhich would explain the large relative contribution of maple tocanopy transpiration in stand DL3a. The low Ec rates per crownprojection area of F. excelsior are explained by the small hydroac-tive sapwood area of ash with only the youngest annual rings (typ-ically 3–10) being involved in water transport (Gebauer et al.,2008). On the other hand, F. excelsior reached higher maximum leaf

conductances (�270 mmol m�2 s�1) than most other tree speciesin this mixed stand (160–190 mmol m�2 s�1) and showed no sig-nificant reduction in sap flux density between moist and dry peri-ods. In contrast, such reductions were large in T. cordata, F. sylvaticaand A. pseudoplatanus, and moderate in C. betulus with lower max-imum leaf conductances (Köcher et al., 2009). This seems to con-flict with the assumption that tree species with higher maximumleaf conductances generally show a greater sensitivity to VPD(Oren et al., 1999), thereby reducing the risk of cavitation underhigh VPD. We assume that both traits, hydraulic architecture andsensitivity of leaf conductance regulation, are key functional attri-butes determining species differences in canopy transpiration inthe Hainich forest.

When discussing species effects, it should be kept in mind thatthe flux scaling from sensor to tree and from tree to stand may

2005C

anop

y co

nduc

tanc

e G

c (m

m s

-1)

0

2

4

6

8

10

DL1aDL2cDL3a

2006

0

2

4

6

8

10

May Jun Jul Aug Sep Oct

Fig. 6. Annual course of canopy conductance (Gc) in the three plots DL1a, DL2c andDL3a during 2005 (upper panel) and 2006 (lower panel). The data were smoothedwith a running-average procedure covering four days.

12 T. Gebauer et al. / Journal of Hydrology 442–443 (2012) 2–14

have introduced a certain uncertainty in our Ec estimates as it isindicated by rather high coefficients of variation (23–47% at thespecies and 29–33% at the stand level). Even though we operated88 sensors in total on 44 trees and stratified the measured treesaccording to stem diameter, the minimum of three instrumentedindividuals per species and stand may have been too low to reachat reliable species-level estimates for the rarer species.

4.3. Interannual variability in canopy transpiration

Interannual differences in moisture supply had a profound ef-fect on the seasonal course and total amount of canopy transpira-tion in the three stands. During the summer 2005, rainfall wasmore evenly distributed throughout the growing season, while inthe drier summer 2006, two rainless periods in July and Septembercaused a pronounced decrease in relative plant-extractable water(REW) in the soil. The pure and mixed stands responded differentlyto this variation in soil moisture. Physiological measurements byKöcher et al. (2009) in the sun canopies of the trees indicated thata soil matrix potential of about�0.11 MPa is representing a thresh-old in the drought response of most tree species (except for F.excelsior) in the Hainich forest, resulting in a pronounced reductionof leaf conductance and a marked decrease of sap flux. The con-trasting rainfall regimes in the two years affected the canopy tran-spiration of monospecific and mixed stands differently. In thestands DL1a and DL2c, Ec increased from the moderately wet sum-mer 2005 to the drier summer 2006 by 34% and 43%, respectively.We assume that this is a consequence of the more extended peri-ods with high radiation and VPD in 2006 than in 2005. A higherevaporative demand in 2006 is reflected in the DL1a and DL2cstands with abundant beech by larger transpiration rates of Fagusand Tilia in this summer as compared to 2005. Another factorwhich may have contributed to the increase in Ec from 2005 to2006 in the monospecific DL1a stand is the increase in LAI from2005 to 2006 (Table 1). In contrast, the beech-poor species-richDL3a stand with highest canopy transpiration of all stands in

2005 responded to the drought periods in July and September2006 with a reduction in transpiration, resulting in a by 11% lowerEc total in 2006 than in 2005. It appears that the beech-poor DL3astand was more sensitive in its transpiration regulation to the ex-tended drought periods than the monospecific DL1a stand and theDL2c stand with 70% beech.

A closer look on the seasonal course of soil water availabilitymay explain the deviating behavior of Ec in the DL3a stand: in bothsummers, canopy transpiration increased more rapidly in May andJune in this stand as compared to the DL1a and DL2c stands, reach-ing early Ec peaks already in late June. This was mainly a conse-quence of the higher water use of Tilia per ground area early insummer. In the DL1a and DL2c stands with a higher Fagus and low-er Tilia contribution, in contrast, Ec reached its peak later in sum-mer in July or early August, which is probably a consequence of amore conservative water loss regulation of beech. This species isknown for its sensitive leaf conductance regulation to variationin VPD (e.g. Backes and Leuschner, 2000; Oren et al., 1999; Rennen-berg et al., 2006; Köcher et al., 2009). Thus, it appears that thebeech-poor species-rich DL3a stand had extracted the bulk ofplant-available water already in June 2006, mainly because of thehigh consumption of the dominant Tilia trees in this stand. Conse-quently, soil moisture dropped in DL3a to lower values in July, Au-gust and September than in the DL1a and DL2c stands, restrictingcanopy transpiration in the second half of the summer more se-verely in this Tilia-rich stand. This species effect on water extrac-tion may have been enhanced by the somewhat smaller storagecapacity for extractable water (hFC–hmin) in the DL3a stand(170 mm) as compared to the DL1a and DL2c stands (200 and240 mm).

In the moister summer 2005, soil water content remained athigher levels in the DL3a stand, thus supporting a high transpira-tion rate of Tilia throughout the summer. The lower VPD may havefavoured Tilia by enabling this species with a relatively high stoma-tal conductance in moist atmospheres to maintain higher photo-synthesis and transpiration rates than in the other species.Species with larger maximum leaf conductances (such as beech)may profit less from wet summers (Oren et al., 1999). The abun-dant Fraxinus trees in the stand DL2c may have contributed tothe fact that Ec did not pass through a minimum in mid-summer2006, since ash was found to maintain high sap flux rates even be-yond the �0.11 MPa threshold of soil matrix potential (Köcheret al., 2009).

Compared to other studies on canopy transpiration in temper-ate broad-leaved forests, we found rather low Ec totals for the veg-etation periods of 2005 and 2006 (97–158 mm). For a set of CentralEuropean monospecific beech forests, Schipka et al. (2005) ob-tained a mean Ec value of 289 (±58) mm in the vegetation period.For stands dominated by beech, oak, maple or ash, Granier et al.(1996, 2000), Herbst et al. (2008), Peck and Mayer (1996), Vinckeet al. (2005) and Wullschleger and Hanson (2006) reported Ec val-ues in the vegetation period of 212–397 mm. On the other hand,Poyatos et al. (2007) measured low Ec totals of 118 and 164 mmfor oak-dominated stands in the exceptionally dry summer 2003and the subsequent year, rates that are similar to our values. Lowcanopy transpiration rates of the Hainich stands in 2005 and2006 may have been caused by the specific edaphic situation ofthis forest with clay-rich soils that restrict root water uptake insummer by very low hydraulic conductivities of the soil matrix.Moreover, the extraordinary drought of 2003 resulted in a pro-nounced after-effect in 2004 and also in 2005 with reduced stoma-tal conductances over extended periods in beech and other species(N. Legner, pers. commun.) and higher canopy defoliation rates inmany Central European broad-leaved forests as compared to aver-age summers (Bréda et al., 2006; Ciais et al., 2005; Granier et al.,2007; ICP Forests, 2007). The similar annual totals of Ec in all three

T. Gebauer et al. / Journal of Hydrology 442–443 (2012) 2–14 13

Hainich stands in the relative dry year 2006 seem to support theconclusion by Huxman et al. (2004) that different forest standsconverge in their water use efficiency in drier years.

5. Conclusions

Whether canopy transpiration was higher, or equal, in species-rich broad-leaved forests with low abundance of beech than in apure beech stand, depended on the amount of rainfall. Thus, our re-sults do not support the assumption of a higher water use of mixedstands as reported by e.g. Schume et al. (2004). One reason is thatthe edaphic conditions in the Hainich forest seem to prevent thevertical stratification of different root systems, making spatialcomplementarity in root water uptake in the mixed stands unli-kely. We found large species differences in water use on a basalarea, sapwood area and ground area basis among the coexistingtree species, supporting our first hypothesis. Substantial differ-ences in Ec existed not only between diffuse- and ring-porous trees,but also within the diffuse-porous group. Consequently, tree spe-cies identity and the related specific functional traits are muchmore important for canopy transpiration and its seasonal variabil-ity than is tree species diversity. We suggest that the sizes of sap-wood area and leaf area as morphological attributes and thehydraulic conductance in the root-to-leaf flow path together withleaf conductance as physiological traits are main factors causingdifferent seasonal transpiration patterns among the five treespecies.

Since soil water is a preemptable resource that cannot be storedin large amounts, stands with a higher species richness may sufferfrom a higher drought exposure in drier years, if the species mix-ture encompasses trees such as linden that tend to exhaust thewater reserves early in summer. In cases with significant soil mois-ture preemption in mixed stands, tree species diversity could wellenhance drought stress in dry years, thereby reducing the stabilityof mixed forests.

Acknowledgements

The authors are grateful to the management of Hainich NationalPark, Thuringia, for the research permit and the good cooperation.We thank Dr. Heinz Coners for technical support and Karl Maximil-ian Daenner for helpful discussions regarding statistical analysis.We also thank anonymous reviewers for helpful comments whichhelped to improve the manuscript. This study was conducted with-in the framework of the Graduiertenkolleg 1086 with funding fromthe German Research Foundation (DFG).

References

Anders, S., Müller, J., Augustin, S., Rust, S., 2006. Die Ressource Wasser imzweischichtigen Nadel-Laub-Mischwald. In: Fritz, P., (Ed.), ÖkologischerWaldumbau in Deutschland. oekom, München, pp. 152–183.

Backes, K., Leuschner, C., 2000. Leaf water relations of competitive Fagus sylvaticaand Quercus petraea trees during 4 years differing in soil drought. Can. J. ForestRes. 30, 335–346.

Baldocchi, D.D., Falge, E., Gu, L., Olson, R., Hollinger, D., Running, S., Anthoni, P.,Bernhofer, C., Davis, K., Evans, R., Fuentes, J., Goldstein, A., Katul, G., Law, B., Lee,X., Malhi, Y., Meyers, T., Munger, W., Oechel, W., Paw, U.K.T., Pilegaard, K.,Schmid, H.P., Valentini, R., Verma, S., Vesala, T., Wilson, K., Wofsy, S., 2001.FLUXNET: a new tool to study the temporal and spatial variability of ecosystem-scale carbon dioxide, water vapor, and energy flux densities. Bull. Am. Met. Soc.82, 2415–2434.

Baldocchi, D.D., 2005. The role of biodiversity on the evaporation of forests. In:Scherer-Lorenzen, M., Körner, C., Schulze, E.D. (Eds.), Forest Diversity andFunction: Temperate and Boreal Systems. Ecological Studies, vol. 176. Springer-Verlag, Berlin, pp. 131–145.

Benecke, P., 1984. Der Wasserhaushalt eines Buchen- und einesFichtenwaldökosystems im Hochsolling. Schriften aus der Forstlichen Fakultaetder Universitaet Goettingen und Mitteilungen der NiedersaechsischenForstlichenVersuchsanstalt 77, 1–158.

Bernier, P.Y., Bréda, N., Granier, A., Raulier, F., Mathieu, F., 2002. Validation of acanopy gas exchange model and derivation of a soil water modifier fortranspiration for sugar maple (Acer saccharum Marsh.) using sap flow densitymeasurements. Forest Ecol. Manage. 163, 185–196.

Bréda, N., Granier, A., Aussenac, G., 1995. Effects of thinning on soil and tree waterrelations, transpiration and growth in an oak forest (Quercus petraea (Matt.)Liebl.). Tree Physiol. 15, 295–306.

Bréda, N., Huc, R., Granier, A., Dreyer, E., 2006. Temperate forest trees and standsunder severe drought: a review of ecophysiological responses, adaptationprocesses and long-term consequences. Ann. Forest Sci. 63, 625–644.

Bush, S.E., Pataki, D.E., Hultine, K.R., West, A.G., Sperry, J.S., Ehleringer, J.R., 2008.Wood anatomy constrains stomatal responses to atmospheric vapor pressuredeficit in irrigated, urban trees. Oecologia 156, 13–20.

Bücking, W., Krebs, A., 1986. Interzeption und Bestandesniederschläge von Bucheund Fichte im Schönbuch. In: Einsele, G. (Ed.), Das landschaftsökologischeForschungsprojekt Naturpark Schönbuch. VCH, Weinheim, pp. 113–131.

Cermak, J., Kucera, J., Nadezhdina, N., 2004. Sapflow measurements with somethermodynamic methods, flow integration within trees and scaling up fromsample trees to entire forest stands. Trees 18, 529–546.

Ciais, P., Viovy, N., Reichstein, M., Granier, A., Ogée, J., Rambal, S., Ourcival, J.M.,Bernhofer, C., Grunwald, T., Pilegaard, K., Berbigier, P., Vesala, T., Aubinet, M.,Heinesch, B., Loustau, D., Seufert, G., Manca, G., Matteucci, G., Miglietta, F.,Soussana, J.F., Knohl, A., Sanz, M.J., Carrara, A., Friend, A., Chevallier, F., Schulze,E.D., Valentini, R., 2005. An unprecedented reduction in the primaryproductivity of Europe during 2003 caused by heat and drought. Nature 437,529–533.

Dhote, J.F., 2005. Implication of forest diversity in resistance to strong winds. In:Scherer-Lorenzen, M., Körner, C., Schulze, E.D. (Eds.), Forest Diversity andFunction: Temperate and Boreal Systems. Ecological Studies, vol. 176. Springer-Verlag, Berlin, pp. 291–307.

Ellenberg, H., Leuschner, C., 2010. Vegetation Mitteleuropas mit den Alpen, sixthed. Ulmer, Stuttgart, p. 1335.

Ewers, B.E., Mackay, D.S., Gower, S.T., Ahl, D.E., Burrows, S.N., 2002. Tree specieseffects on stand transpiration in northern Wisconsin. Water Resour. Res. 38, 1–11.

Forrester, D.I., Theivayanathan, S., Collopy, J.J., Marcar, N.E., 2010. Enhanced wateruse efficiency in a mixed Eucalyptus globulus and Acacia mearnsii plantation.Forest Ecol. Manage. 259, 1761–1770.

Gebauer, T., Horna, V., Leuschner, C., 2008. Variability in radial sap flux densitypatterns and sapwood area among seven co-occurring broad-leaved treespecies. Tree Physiol. 28, 1821–1830.

Granier, A., 1985. Une nouvelle methode pour la mesure du flux de seve brute dansle tronc des arbres. Ann. Forest Sci. 42, 193–200.

Granier, A., 1987. Evaluation of transpiration in a Douglas-fir stand by means of sapflow measurements. Tree Physiol. 3, 309–320.

Granier, A., Biron, P., Bréda, N., Pontailler, P.Y., Saugier, B., 1996. Transpiration oftrees and forest stands: short and long-term monitoring using sapflowmethods. Globol Change Biol. 2, 265–274.

Granier, A., Bréda, N., Biron, P., Villette, S., 1999. A lumped water balance model toevaluate duration and intensity of drought constraints in forest straints. Ecol.Modell. 116, 269–283.

Granier, A., Biron, P., Lemoine, L., 2000. Water balance, transpiration and canopyconductance in two beech stands. Agr. Forest Meteorol. 100, 291–308.

Granier, A., Reichstein, M., Bréda, N., Janssens, I.A., Falge, E., Ciais, P., Grünwald, T.,Aubinet, M., Berbigier, P., Bernhofer, C., Buchmann, N., Facini, O., Grassi, G.,Heinesch, B., Ilvesniemi, H., Keronen, P., Knohl, A., Köstner, B., Lagergren, F.,Lindroth, A., Longdoz, B., Loustau, D., Mateus, J., Montagnani, L., Nys, C., Moors,E., Papale, D., Pfeiffer, M., Pilegaard, K., Pita, G., Pumpanen, J., Rambal, S.,Rebmann, C., Rodrigues, A., Seufert, G., Tenhunen, J., Vesala, T., Wang, Q., 2007.Evidence for soil water control on carbon and water dynamics in Europeanforests during the extremely dry year: 2003. Agr. Forest Meteorol. 143, 123–145.

Guckland, A., Jacob, M., Flessa, H., Thomas, F.M., Leuschner, C., 2009. Acidity,nutrient stocks and organic-matter content in soils of a temperate deciduousforest with different abundance of European beech (Fagus sylvatica L.). J. PlantNutr. Soil Sci. 172, 500–511.

Herbst, M., Rosier, P.T.W., Morecroft, M.D., Gowing, D.J., 2008. Comparativemeasurements of transpiration and canopy conductance in two mixeddeciduous woodlands differing in structure and species composition. TreePhysiol. 28, 959–970.

Hölscher, D., Koch, O., Korn, S., Leuschner, C., 2005. Sap flux of five co-occurring treespecies in a temperate broad-leaved forest during seasonal soil drought. TreesStruct. Funct. 19, 628–637.

Huxman, T.E., Smith, M.D., Fay, P.A., Knapps, A.K., Shaw, M.R., Loik, M.E., Smith, S.D.,Tissue, D.T., Zak, J.C., Weltzin, J.F., Pockman, W.T., Sala, O.E., Haddad, B.M., Harte,J., Koch, G.W., Schwinning, S., Small, E.E., Williams, D.G., 2004. Convergenceacross biomes to a common rain-use efficiency. Nature 429, 652–654.

ICP Forests, 2007. The condition of forests in Europe. 2007 Executive Report. FederalResearch Centre for Forestry and Forest Products, Hamburg.

Jacob, M., Leuschner, C., Thomas, F.M., 2010. Productivity of temperate broad-leavedforest stands differing in tree species diversity. Ann. Forest Sci. 67, 503.

Jactel, H., Brockerhoff, E., Duelli, P., 2005. A test of the biodiversity-stability theory:Meta-analysis of tree species diversity effects on onsect pest infestations, andre-examination of responsible factors. In: Scherer-Lorenzen, M., Körner, C.,Schulze, E.D. (Eds.), Forest Diversity and Function: Temperate and BorealSystems. Ecological Studies, 176. Springer-Verlag, London, pp. 235–262.

14 T. Gebauer et al. / Journal of Hydrology 442–443 (2012) 2–14

Kelty, M.J., 2006. The role of species mixtures in plantation forestry. Forest Ecol.Manage. 233, 195.

Klaus, S., Reisinger, E., 1995. Der Hainich – ein Weltnaturerbe. Landschaftspflegeund Naturschutz in Thüringen – Sonderheft. Thüringer Landesanstalt fürUmwelt, Jena.

Köcher, P., Gebauer, T., Horna, V., Leuschner, C., 2009. Leaf water status and stemxylem flux in relation to soil drought in five temperate broad-leaved treespecies with contrasting water use strategies. Ann. Forest Sci. 66, 101.

Köstner, B., 2001. Evaporation and transpiration from forests in Central Europe –relevance of patch-level studies for spatial scaling. Meteorol. Atmos. Phys. 76,69–82.

Köstner, B., Granier, A., Cermak, J., 1998. Sapflow measurements in forest stands:methods and uncertainties. Ann. Forest Sci. 55, 13–27.

Köstner, B., Falge, E., Tenhunen, J.D., 2002. Age-related effects on leaf area/sapwoodarea relationships, canopy transpiration and carbon gain of Norway sprucestands (Picea abies) in the Fichtelgebirge, Germany. Tree Physiol. 22, 567–574.

Krämer, I., Hölscher, D., 2010. Soil water dynamics along a tree diversity gradient ina deciduous forest in Central Germany. Ecohydrology 3, 262–271.

Legendre, P., 2008. Model II regression user’s guide, R edition. (A tutorial within Rsession, R package lmodel2), pp. 14.

Legendre, P., Legendre, L., 1998. Numerical Ecology. Elsevier, Amsterdam.Leuschner, C., Jungkunst, H.F., Fleck, S., 2009. Functional role of forest diversity: pros

and cons of synthetic stands and across-site comparisons in established forests.Basic Appl. Ecol. 10, 1–9.

Lu, P., Urban, L., Zhao, P., 2004. Granier’s thermal dissipation probe (TDP) methodfor measuring sap flow in trees: theorie and practice. Acta Bot. Sin. 46, 631–646.

Meinen, C., Ryan, N.T., Hertel, D., Leuschner, C., 2009. No evidence of spatial rootsystem segregation and elevated root biomass in species-rich temperate broad-leaved forests. Trees 23, 941–950.

Mencuccini, M., Grace, J., 1996. Hydraulic conductance, light interception andneedle nutrient concentration in Scots pine stands and their relation to netprimary production. Tree Physiol. 16, 459–468.

Oren, R., Sperry, J.S., Katul, G.G., Pataki, D.E., Ewers, B.E., Phillips, N., Schäfer, K.V.R.,1999. Survey and synthesis of intra- and interspecific variation in stomatalsensitivity to vapour pressure deficit. Plant Cell Environ. 22, 1515–1526.

Oren, R., Pataki, D.E., 2001. Transpiration in response to variation in microclimateand soil moisture in southeastern deciduous forests. Oecologia 127, 549–559.

Otto, H.J., 1992. Rahmenbedingungen und Möglichkeiten zur Verwirklichung derökologischen Weiterentwicklung in den niedersächsischen Landesforsten. Forstu. Holz 47, 75–78.

Peck, A., Mayer, H., 1996. Einfluß von Bestandesparametern auf die Verdunstungvon Wäldern. Forstwissenschaftliches Centralblatt 115, 1–9.

Poyatos, R., Cermak, J., Llorens, P., 2007. Variation in the radial patterns of sap fluxdensity in pubescent oak (Quercus pubescens) and its implications for tree andstand transpiration measurements. Tree Physiol. 27, 537–548.

Pretzsch, H., 2005. Diversity and productivity in forests: evidence from long-termexperimental plots. In: Scherer-Lorenzen, M., Körner, Ch., Schulze, E.-D. (Eds.),Forest Diversity and Function. Ecological Studies, 176. Springer, Berlin, pp. 41–64.

Pretzsch, H., Block, J., Dieler, J., Dong, P.H., Kohnle, U., Nagel, J., Spellmann, H., Zingg,A., 2010. Comparison between the productivity of pure and mixed stands of

Norway spruce and European beech along an ecological gradient. Ann. For. Sci.67, 712.

Rennenberg, H., Loreto, F., Polle, A., Brilli, F., Fares, S.R., Beniwal, R.S., Gessler, A.,2006. Physiological responses of forest trees to heat and drought. Plant Biol. 8,556–571.

R Development Core Team, 2009. R: A language and environment for statisticalcomputing. R Foundation for Statistical Computing, Vienna, Austria. <http://www.R-project.org>.

Roberts, J., 2000. The influence of physical and physiological characteristics ofvegetation on their hydrological response. Hydrol. Processes 162, 229–245.

Rothe, A., Binckley, D., 2001. Nutritional interactions in mixed species forests: asynthesis. Can. J. Forest Res. 31, 1855–1870.

Ryan, M.G., Bond, B.J., Law, B.E., Hubbard, R.M., Woodruff, D., Cienciala, E., Kucera, J.,2000. Transpiration and whole-tree conductance in ponderosa pine trees ofdifferent heights. Oecologia 124, 553–560.

Schäfer, K.V.R., Oren, R., Tenhunen, J.D., 2000. The effect of tree height on crownlevel stomatal conductance. Plant Cell Environ. 23, 365–375.

Schipka, F., Heimann, J., Leuschner, C., 2005. Regional variation in canopytranspiration of Central European beech forests. Oecologia 143, 260–270.

Schmidt, I., Leuschner, C., Mölder, A., Schmidt, W., 2009. Structure and compositionof the seed bank in monospecific and tree species-rich temperate broad-leavedforests. Forest Ecol. Manage. 257, 695–702.

Schulze, E.D., Kelliher, F.M., Körner, C., Lloyd, J., Leuning, R., 1994. Relationshipsamong maximum stomatal conductance, ecosystem surface conductance,carbon assimilation rate, and plant nitrogen nutrition: a global ecologyscaling exercise. Ann. Rev. Ecol. Evol. Syst. 25, 629–662.

Schume, H., Jost, G., Hager, H., 2004. Soil water depletion and recharge patterns inmixed and pure forest stands of European beech and Norway spruce. J. Hydrol.289, 258–274.

Vertessy, R., Benyon, R., Haydon, S., 1994. Melbourne’s forest catchments: effect ofage on water yield. Water 21, 17–20.

Vertessy, R.A., Benyon, R.G., O’Sullivan, S.K., Gribben, P.R., 1995. Relationshipbetween stem diameter, sapwood area, leaf area and transpiration in a youngmountain ash forest. Tree Physiol. 15, 559–568.

Vertessy, R.A., Hatton, T.J., Reece, P., O’Sullivan, S.K., Benyon, R.G., 1997. Estimatingstand water use of large mountain ash trees and validation of sap flowmeasurement technique. Tree Physiol. 17, 747–756.

Vincke, C., Granier, A., Bréda, N., Devillez, F., 2005. Evapotranspiration of a decliningQuercus robur (L.) stand from 1999 to 2001. II. Daily actual evapotranspirationand soil water reserve. Ann. Forest Sci. 62, 615–623.

Warton, D.I., Wright, I.J., Falster, D.S., Westoby, M., 2006. Bivariate line-fittingmethods for allometry. Biol. Rev. 81, 259–291.

Wullschleger, S.D., Hanson, P.J., Tschaplinski, T.J., 1998. Whole-plant water flux inunderstory red maple exposed to altered precipitation regimes. Tree Physiol. 18,71–79.

Wullschleger, S.D., Hanson, P.J., Todd, D.E., 2001. Transpiration from a multi-speciesdeciduous forest as estimated by xylem sap flow techniques. Forest Ecol.Manage. 143, 205–213.

Wullschleger, S.D., Hanson, P.J., 2006. Sensitivity of canopy transpiration to alteredprecipitation in an upland oak forest: evidence from a long-term fieldmanipulation study. Globol Change Biol. 12, 97–109.