Hydrol.EarthSyst.Sci.,21,1421–1438,2017 www.hydrol-earth-syst-sci.net/21/1421/2017/ doi:10.5194/hess-21-1421-2017 ©Author(s)2017.CCAttribution3.0License. Comparisons of stemflow and its bio-/abiotic influential factors between two xerophytic shrub species ChuanYuan1,2,GuangyaoGao1,3,andBojieFu1,3 1StateKeyLaboratoryofUrbanandRegionalEcology,ResearchCenterforEco-EnvironmentalSciences, ChineseAcademyofSciences,Beijing100085,China 2UniversityofChineseAcademyofSciences,Beijing100049,China 3JointCenterforGlobalChangeStudies,Beijing100875,China Correspondenceto:GuangyaoGao([email protected]) Received:18August2016–Discussionstarted:31August2016 Revised:16February2017–Accepted:24February2017–Published:9March2017 Abstract. Stemflow transports nutrient-enriched precipita- pattern) had a great influence during lighter rains≤10mm tion to the rhizosphere and functions as an efficient terres- and heavier rains>15mm, respectively. The lower precip- trial flux in water-stressed ecosystems. However, its eco- itation threshold to start stemflow allowed C. korshinskii logical significance has generally been underestimated be- (0.9mm vs. 2.1mm for S. psammophila) to employ more causeitisrelativelylimitedinamount,andthebioticmech- rainstoharvestwaterviastemflow.Thebeneficialleaftraits anismsthataffectithavenotbeenthoroughlystudiedatthe (e.g.,leafshape,arrangement,area,amount)mightpartlyex- leaf scale. This study was conducted during the 2014 and plainthegreaterstemflowproductionofC.korshinskii.Com- 2015 rainy seasons at the northern Loess Plateau of China. parisonofSF betweenthefoliatedandmanuallydefoliated b Wemeasuredthebranchstemflowvolume(SF ),shrubstem- shrubsduringthe2015rainyseasonindicatedthatthenewly b flow equivalent water depth (SF ), stemflow percentage of exposed branch surface at the defoliated period and the re- d incident precipitation (SF%), stemflow productivity (SFP), sulting rainfall intercepting effects might be an important funnellingratio(FR),themeteorologicalcharacteristicsand mechanismaffectingstemflowinthedormantseason. the plant traits of branches and leaves of C. korshinskii and S.psammophila.Thisstudyevaluatedstemflowefficiencyfor thefirsttimewiththecombinedresultsofSFPandFR,and 1 Introduction sought to determine the inter- and intra-specific differences of stemflow yield and efficiency between the two species, Stemflow delivers precipitation to the plant root zone more aswellasthespecificbio-/abioticmechanismsthataffected efficiently via preferential root paths, worm paths and soil stemflow. The results indicated that C. korshinskii had a macropores, compared with throughfall, another important greaterstemflowyieldandefficiencyatallprecipitationlev- element of rainfall redistribution. The double-funnelling ef- elsthanthatofS.psammophila.Thelargestinter-specificdif- fectsofstemflowandpreferentialflowcreate“hotspots”and ference generally occurred at the 5–10mm branches during “hotmoments”byenhancingnutrientcyclingratesatthesur- rainsof≤2mm.Precipitationamountwasthemostinfluen- facesoilmatrix(Mcclainetal.,2003;JohnsonandLehmann, tialmeteorologicalcharacteristicthataffectedstemflowyield 2006; Sponseller, 2007), thus substantially contributing to and efficiency in these two endemic shrub species. Branch theformationandmaintenanceof“fertileislands”(Whitford anglewasthemostinfluentialplanttraitonFR.ForSF ,stem b et al., 1997), “resource islands” (Reynolds et al., 1999) or biomass and leaf biomass were the most influential plant “hydrologicislands”(Rangoetal.,2006).Thiseffectisim- traitsforC.korshinskiiandS.psammophila,respectively.For portantforthenormalfunctionofrainfeddrylandecosystems SFPofthesetwoshrubspecies,leaftraits(theindividualleaf (Wangetal.,2011). area) and branch traits (branch size and biomass allocation PublishedbyCopernicusPublicationsonbehalfoftheEuropeanGeosciencesUnion. 1422 C.Yuanetal.:Comparisonsofstemflowanditsbio-/abioticinfluentialfactors Shrubs are the representative plant functional type (PFT) depth, percentage of incident precipitation, FR and SFP for indrylandecosystems.Theyhavedevelopedeffectivephys- a comprehensive description of the inter- and intra-specific iological drought tolerance by reducing water loss, e.g., stemflowyieldandefficiencyatbranchandshrubscales. throughadjustingtheirphotosyntheticandtranspirationrate The precipitation amount has been generally recognized byregulatingstomatalconductanceandabscisicacid(ABA), as the single most influential rainfall characteristic affect- titlingtheirosmoticequilibriumbyregulatingtheconcentra- ing stemflow production (Clements, 1972; André et al., tionofsolublesugarsandinorganicions,andremovingfree 2008; Van Stan II et al., 2014). However, as to biotic radicals (Ma et al., 2004, 2008). The stemflow, a vital eco- mechanisms,althoughthecanopystructure(Mauchampand hydrological flux, is involved in replenishing soil water at Janeau,1993;CrockfordandRichardson,2000;Pypkeretal., shallow and deep layers (Pressland, 1973), particularly the 2011) and branch architecture (Herwitz, 1987; Murakami, root zone (Whitford et al., 1997; Dunkerley, 2000; Yang, 2009; Carlyle-Moses and Schooling, 2015) have been stud- 2010), even during light rains (Li et al., 2009). It might al- ied for years, the most important plant traits vary with lo- low the endemic shrubs to remain physically active during cation and shrub species and have not yet been determined. drought spells (Navar and Bryan, 1990; Navar, 2011). The Theeffectsoftheleaveshavebeenstudiedmorerecentlyata stemflow is an important potential source for available wa- smallerscale,e.g.,leaforientation(CrockfordandRichard- teratrainfeddrylandecosystem(Lietal.,2013).Therefore, son, 2000), shape (Xu et al., 2005), arrangement pattern producingstemflowwithagreateramountinamoreefficient (Owens et al., 2006), pubescence (Garcia-Estringana et al., mannermightbeaneffectivestrategytoutilizeprecipitation 2010),area(Sellinetal.,2012),epidermismicrorelief(Roth- by reducing the evaporation loss (Devitt and Smith, 2002; Nebelsicketal.,2012),amount(Lietal.,2016),orbiomass Li et al., 2009), acquire water (Murakami, 2009) and with- (Yuan et al., 2016). Comparisons of stemflow yield during stand drought (Martinez-Meza and Whitford, 1996). How- summer(thegrowingorfoliatedseason)andwinter(thedor- ever,becausestemflowoccursinasmallamount,somestud- mantordefoliatedseason)generallyindicatenegativeeffects ies neglected the dynamics of stemflow yield by setting a of leaves because more stemflow occurs during the leafless fixedpercentageofincidentprecipitationintherangeof1– period (Dolman, 1987; Masukata et al., 1990; Neal et al., 8% (Dykes, 1997; Germer et al., 2006; Hagy et al., 2006) 1993; Muz˙yło et al., 2012). However, both negligible and or even ignored stemflow while computing the water bal- positiveeffectshavealsobeenconfirmedbyMartinez-Meza anceofterrestrialecosystems(LlorensandDomingo,2007; and Whitford (1996), Deguchi et al. (2006) and Liang et Zhang et al., 2016). That underestimated its disproportion- al.(2009).Thevalidityoftheseconflictingfindingshasbeen ately high influence on xerophytic shrub species (Anders- called into question as a result of the seasonal variation of son, 1991; Levia and Frost, 2003; Li, 2011). Therefore, it meteorological conditions and plant traits, e.g., wind speed isimportanttoquantifytheinter-andintra-specificstemflow (Andréetal.,2008),rainfallintensity(Dunkerley,2014a,b), yield,toassessthestemflowproductionefficiencyandtoelu- airtemperatureandconsequentprecipitationtype(snow-to- cidatetheunderlyingbio-/abioticmechanisms. rainvs.snow)(LeviaandUnderwood,2004).Moreover,they Stemflow yield includes the stemflow volume and depth, ignore the effects of the exposed stems at leafless period, and it describes the total flux delivered down to the base of whichsubstitutetheleavestointerceptraindropsandmight abranchoratrunk.However,stemflowdataareunavailable playasignificantroleinstemflowproduction.Furthermore, for comparison of inter-specific differences caused by vari- although the rainfall simulator made an identical and gradi- ations in the branch architecture, the canopy structure, the entchangeofrainfallcharacteristicspossible,thelaboratory shrub species and the ecozone. Herwitz (1986) introduced experiment ignored the dynamics of rainfall characteristics thefunnellingratio(FR),whichexpressedasthequotientof andmeteorologicalfeatures(e.g.,windspeed,vaporpressure the volume of stemflow yield and the product of the base deficit, air temperature and humidity) during rainfall events areaandtheprecipitationamount.Itindicatestheefficiency at field conditions. Therefore, a controlled field experiment with which individual branches or shrubs capture raindrops with the foliated and manually defoliated plants under the and deliver the water to the root zone (Siegert and Levia, same stand conditions is needed to resolve these uncertain- 2014).TheFRallowsforacomparisonoftheinter-andintra- ties. specific stemflow yield under different precipitation condi- Inthisstudy,thebranchstemflowvolume(SF ),theshrub b tions. However, the FR does not provide a good connection stemflow depth (SF ), the stemflow percentage of the inci- d betweenhydrologicalprocesses(e.g.,rainfallredistribution) dentprecipitationamount(SF%),theSFPandtheFRwere and the plant growth processes (e.g., biomass accumulation measuredintwoxerophyticshrubspecies(C.korshinskiiand and allocation). Recently, Yuan et al. (2016) introduced the S. psammophila) during the 2014 and 2015 rainy seasons. parameter of stemflow productivity (SFP), expressed as the Furthermore,acontrolledfieldexperimentwithfoliatedand volume of stemflow yield per unit of branch biomass. The manually defoliated shrubs was also conducted for the two SFPdescribestheefficiencybycomparingthestemflowyield shrub species during the 2015 rainy season. The detailed ofunitbiomassincrementatdifferentsizedbranches.Hence, objectives were to (1) quantify the inter- and intra-specific it is necessary to combine the results of stemflow volume, stemflowyield(SF ,SF andSF%)andefficiency(SFPand b d Hydrol.EarthSyst.Sci.,21,1421–1438,2017 www.hydrol-earth-syst-sci.net/21/1421/2017/ C.Yuanetal.:Comparisonsofstemflowanditsbio-/abioticinfluentialfactors 1423 Figure1.LocationoftheexperimentalstandsandfacilitiesforstemflowmeasurementsofC.korshinskiiandS.psammophilaattheLiudao- goucatchmentintheLoessPlateauofChina. FR) at different precipitation levels; (2) identify the most nual temperature is 9.0◦C (Zhao et al., 2010). The coldest influential meteorological characteristics affecting stemflow and warmest months are January and July, with an average yieldand(3)investigatethebioticinfluentialmechanismof monthlytemperatureof−9.7and23.7◦C,respectively.The planttraitsespeciallyatthefinerleafscale.Giventhatonly two soil types Aeolian sandy soil and Ust-Sandiic Entisol the aboveground ecohydrological process was involved, we dominatethiscatchment(Jiaetal.,2011).Soilparticlescon- focused on stemflow in this study, its interaction with soil sist of 11.2–14.3% clay, 30.1–44.5% silt and 45.4–50.9% moisturewouldbediscussedinnextstudy.Theachievement sandintermsofthesoilclassificationsystemofUnitedStates of these research objectives would advance our understand- DepartmentofAgriculture(ZhuandShao,2008).Theorigi- ingoftheinfluentialmechanismofstemflowproduction,its nalplantsarescarcelypresent,exceptforveryfewsurviving ecological importance for dryland shrubs, and the signifi- shrub species, e.g., Ulmus macrocarpa, Xanthoceras sorb- canceofleavesfromanecohydrologicalperspective. ifolia, Rosa xanthina or Spiraea salicifolia. The currently predominant shrub species were planted decades ago, e.g., S. psammophila, C. Korshinskii or Amorpha fruticosa, and thepredominantgrassspeciesincludeMedicagosativa,Stipa 2 Materialsandmethods bungeana,Artemisiacapillaris,Artemisiasacrorum,etc.(Ai 2.1 Studyarea etal.,2015). Two representative experimental stands of C. Korshinskii This study was conducted at the Liudaogou catchment andS.psammophilawereestablishedinthesouthwestofthe (110◦21(cid:48)–110◦23(cid:48)E, 38◦46(cid:48)–38◦51(cid:48)N) in Shenmu County Liudaogoucatchmentinthisstudy(Fig.1).Astheendemic in the Shaanxi Province of China. It is 6.9km2 and 1094– shrubspeciesinaridandsemiaridnorthernChina,theywere 1273ma.s.l.(abovesealevel).Thisareahasasemiaridcon- generally planted for wind proofing and dune stabilizing. tinentalclimatewithwell-definedrainyanddryseasons.The BothC.korshinskiiandS.psammophilaaremulti-stemmed meanannualprecipitation(MAP)between1971and2013is shrubsthathaveaninvertedconecanopyandnotrunk,with 414mm,withapproximately77%oftheannualprecipitation thebranchesrunningobliquelyfromthebase.C.korshinskii amount occurring fromJuly to September (Jiaet al., 2013). usuallygrowsto2mandhaspinnatecompoundleaveswith Thepotentialevaporationis1337mmyr−1,andthemeanan- 12–16 foliates in an opposite or sub-opposite arrangement www.hydrol-earth-syst-sci.net/21/1421/2017/ Hydrol.EarthSyst.Sci.,21,1421–1438,2017 1424 C.Yuanetal.:Comparisonsofstemflowanditsbio-/abioticinfluentialfactors (Wang et al., 2013). The leaf of C. korshinskii is concave ple shrubs that had a similar canopy structure. Four mature and lanceolate shaped, with an acute leaf apex and an ob- shrubs were selected for C. korshinskii (designated as C1, tusebase.Bothsidesoftheleavesaredenselysericeouswith C2, C3 and C4) and S. psammophila (designated as S1, appressed hairs (Liu et al., 2010). In comparison, S. psam- S2, S3 and S4) for the stemflow measurements. They had mophila usually growsto 3–4m, andhas an odd numberof isolated canopies, similar intra-specific canopy heights and stripshapedleaveswith2–4mminwidthand40–80mmin areas, e.g., 2.1±0.2m and 5.1±0.3m2 for C1–C4, and length. The young leaves are pubescent and gradually be- 3.5±0.2m and 21.4±5.2m2 for S1–S4. We measured the comesubglabrous(ChaoandGong,1999).Thesetwoshrub morphologicalcharacteristicsofallthe180branchesofC1– species were planted approximately 20 years ago, and the C4andallthe261branchesofS1–S4,includingthebranch twostandssharedasimilarslopeof13–18◦,asizeof3294– basal diameter (BD; mm), branch length (BL; cm) and 4056m2andanelevationof1179–1207ma.s.l.However,the branch inclination angle (BA; ◦). The leaf area index (LAI) C. korshinskii experimental stand has a 224◦ aspect with a and the foliage orientation (MTA, the mean tilt angle of loess ground surface, whereas the S. psammophila experi- leaves)weremeasuredusingtheLiCor®(LiCorBiosciences mentalstandhasa113◦aspectwithasandgroundsurface. Inc., Lincoln, NE, USA) 2200C plant canopy analyzer ap- proximatelytwiceamonth. 2.2 Fieldexperiments Atotalof53branchesofC.korshinskii(17,21,7,8forthe BDcategoriesof5–10,10–15,15–18and>18mm,respec- Field experiments were conducted during the rainy sea- tively) and 98 branches of S. psammophila (20, 30, 20 and sons of 2014 (1 July to 3 October) and 2015 (1 June 28branchesattheBDcategoriesof5–10,10–15,15–18and to 30 September) to measure the meteorological charac- >18mm,respectively)wereselectedforstemflowmeasure- teristics, plant traits and stemflow. To avoid the effects of ments following these criteria: (1) no intercrossing stems, gully micro-geomorphology on meteorological recording, (2) no turning point in height from branch tip to the base we installed an Onset® (Onset Computer Corp., Bourne, (Dongetal.,1987)and(3)representativenessinamountand MA, USA) RG3-M tipping bucket rain gauge (0.2mm per branch size. Stemflow was collected using aluminum foil tip) at each experimental stand. Three 20cm diameter rain collars, which were more accurate than the spiral tubes be- gauges were placed around to adjust the inherent under- cause the tubes outlet were more liable to be blocked by estimating of automatic precipitation recording (Groisman vegetationlitter(Wright,1977;Durocher,1990).Thecollar and Legates, 1994). Then, the rainfall characteristics, e.g., wasfittedaroundtheentirebranchcircumferenceandclose rainfall duration (RD; h), rainfall interval (RI; h), the av- to the branch base and sealed by neutral silicone caulking erage rainfall intensity (I, mmh−1), the maximum rainfall (Fig. 1). Nearly all sample branches were selected on the intensity in 5min (I , mmh−1), 10min (I , mmh−1) and skirts of the crown, which were more convenient for instal- 5 10 30min (I , mmh−1) could be calculated accordingly. In lation and limited the amount of shading by other branches 30 this study, the individual rainfall events were greater than lyingabove.Associatedwiththelimitedexternaldiameterof 0.2mm and separated by a period of at least 4h without foilcollars,theyminimizedtheaccessingoftheprecipitation rain (Giacomin and Trucchi, 1992). Furthermore, a mete- and throughfall (both free and released). A 0.5cm diameter orological station was also installed at each experimental PVC hose led the stemflow to lidded containers. The col- standtorecordothermeteorologicalcharacteristics(Fig.1), larsandhoseswerecheckedperiodicallyagainstanyleakage e.g., wind speed (WS; ms−1) and wind direction (WD; ◦) and blockage. The stemflow was measured within 2h after (Model03002,R.M.YoungCompany,TraverseCity,Michi- the rainfall ended during the daytime; if the rainfall ended gan,USA),theairtemperature(T;◦C)andhumidity(H;%) at night, we took the measurement early the next morning. (ModelHMP155, Vaisala,Helsinki,Finland),and solarra- Aftercompletingmeasurements,wereturnedstemflowback diation(SR;kWm−2)(ModelCNR4,Kipp&ZonenB.V., tothebranchbasetomitigatetheunnecessarydroughtstress Delft, the Netherlands). Moreover, raindrops attributes, in- forthesamplebranches.Bydoingso,wetriedbesttomiti- cluding raindrop diameter (D, mm), raindrop terminal ve- gatetheinfluencesoftheprecipitationandthroughfall,which locity (V, ms−1), and raindrop inclination angle from the might lead to an overestimation of stemflow yield and ef- vertical (A, ◦), werealso computed toinvestigate the possi- ficiency. Nevertheless, these errors might not be eradicated bleeffectsofraindropstriking,theobliqueandwind-driven atfieldconditionsafterall.Thecarefulexperimentpractices rainonstemflowyieldandefficiency. wereespeciallyneededinthisstudy,andmorethoughtfulex- C. korshinskii and S. psammophila, as modular organ- perimentdesignswererequiredinfuturestudies. isms and multi-stemmed shrub species, have branches that The controlled field experiment with foliated and manu- seek their own survival goals and compete with each other ally defoliated shrubs was conducted during the 2015 rainy for light and water (Firn, 2004; Allaby, 2010). They were season for C. korshinskii (five rainfall events from 18 to ideal experiment objects to conduct a stemflow study at the 30 September) and for S. psammophila (10 rainfall events branch scale. Therefore, we focused on branch stemflow from 2 August to 30 September) (Fig. 2). Considering the and ignored the canopy variance by experimenting on sam- workload to remove all the leaves of 85 branches on C. ko- Hydrol.EarthSyst.Sci.,21,1421–1438,2017 www.hydrol-earth-syst-sci.net/21/1421/2017/ C.Yuanetal.:Comparisonsofstemflowanditsbio-/abioticinfluentialfactors 1425 Figure2.Thecontrolledfieldexperimentforstemflowyieldbetweenthefoliatedandmanuallydefoliatedshrubs. rshinskii (designated as C5) and 94 branches on S. psam- of 66 branches for C6–C8 and 61 branches for S6–S8 were mophila (designated as S5) nearly twice a month, only measuredonceduringmid-Augustforthebiomassofleaves oneshrubindividualwasselectedwithsimilarintra-specific andstems(BMLandBMS,g),theleafareaofthebranches canopyheightandarea(2.1mand5.8m2 forC5,3.3mand (LAB, cm2), and the leaf numbers of the branches (LNB), 19.9m2forS5)asotherfoliatedexperimentalshrubs.Atotal when the shrubs showed maximum vegetative growth. The of 10 branches of C5 (3, 3 and 4 branches at the BD cat- BML and BMS were weighted after oven-drying for 48h. egories 5–10, 10–15 and >15mm), and 17 branches of S5 The detailed measurements have been reported in Yuan et (4, 5 and 7 branches at the BD categories 5–10, 10–15 and al.(2016).Thevalidityoftheallometricmodelswasverified >15mm) were selected for stemflow measurements. Ac- bymeasuringanother13branchesofC6–C8and14branches cording to the in situ measurement of branch morphology ofS6–S8. and the laboratory measurement of biomass, these sample branches had similar BD, BL, BA and biomass of leaves 2.3 Calculations (BML) with those in the foliated shrubs (C1–C4 and S1– Theraindropattributes(D,V andA)werecalculatedonthe S4) (see the values at Sect. 3.2). Given a limited amount of basisofthebest-fitequationsdevelopedfromrainfallinten- samplebranchesandrainfallevents,theexperimentalresults sity and wind speed (Laws and Parsons, 1943; Gunn and were just used for a comparison with those of the foliated Kinzer, 1949; Herwitz and Slye, 1995; Van Stan II et al., shrubs,butnotforastatisticalanalysiswithmeteorological 2011;Carlyle-MosesandSchooling,2015). characteristicsandplanttraits.Ifnotspecificallystated,itis importanttonoticethatthestemflowyieldandefficiencyin D=2.23·(0.03937·I)0.102 (1) thisstudyreferredtothoseofthefoliatedshrubs. Another three shrubs of each species were destructively V =(3.378·ln(D))+4.213 (2) measured for biomass and leaf traits. They had similar tanA=WS/V, (3) canopy heights and areas as those of the shrubs for which thestemflowwasmeasured,andweredesignatedasC6–C8 where D is the average raindrop diameter (mm), V is the (2.0–2.1mand5.8–6.8m2)andS6–S8(3.0–3.4mand15.4– terminal raindrop velocity (ms−1), A is the raindrop incli- 19.2m2). Therefore, the development of allometric models nation angle from the vertical (◦), I is the average inten- couldbedevelopedforestimatingthecorrespondingbiomass sity(mmh−1)andWSistheaveragewindspeed(ms−1). and leaf traits of C1–C5 and S1–S5 (Levia and Herwitz, 2005;Silesetal.,2010a,b;Stephensonetal.,2014).Atotal www.hydrol-earth-syst-sci.net/21/1421/2017/ Hydrol.EarthSyst.Sci.,21,1421–1438,2017 1426 C.Yuanetal.:Comparisonsofstemflowanditsbio-/abioticinfluentialfactors Biomassandleaftraitswereestimatedbyallometricmod- 2.4 Dataanalysis els as an exponential function of BD (Siles et al., 2010a, b; Jonardetal.,2006): APearsoncorrelationanalysiswasperformedtotestthere- lationshipbetweenSF andeachofthemeteorologicalchar- b PT =a·BDb, (4) acteristics (P, RD, RI, I, I , I , I , WS, T, H, SR, D, e 5 10 30 V andA)andplanttraits(BD,BL,BA,LAB,LNB,ILAB, where a and b are constants, and PTe refers to the esti- BML, BMS and PBMS). Significantly correlated variables mated plant traits BML, BMS, LAB and LNB. The other were further tested with a partial correlation analysis for plant traits could be calculated accordingly, including indi- their separate effects on SF . Then, the qualified variables b vidual leaf area of branch (ILAB=100·LAB/LNB, mm2), were fed into a stepwise regression with forward selection and the percentage of stem biomass to that of branch to identify the most influential bio-/abiotic factors (Carlyle- (PBMS==BMS/(BML+BMS)·100%, %). Furthermore, Moses and Schooling, 2015; Yuan et al., 2016). Similar to the total stem-surface area of individual branch (SA) a principal component analysis and ridge regression, step- was computed representing by that of the main stem, wiseregressionwascommonlyusedbecauseitgotalimited which was idealized as the cone (SA=π·BD·BL/20, effect of multicollinearity (Návar and Bryan, 1990; Honda cm2). Therefore, specific surface area represented with et al., 2015; Carlyle-Moses and Schooling, 2015). More- LAB (SSAL=LAB/(BML+BMS), cm2g−1) and in SA over,weexcludedvariablesthathadavarianceinflationfac- (SSAS=SA/(BML+BMS), cm2g−1) could be calculated. tor (VIF) greater than 10 to minimize the effects of multi- It was important to notice that this method underestimated collinearity (O’Brien, 2007), and kept the regression model the real stem-surface area by ignoring the collateral stems having the least AIC (Akaike information criteria) values andassumingmainstemasthestandardcorn,andtherefore andlargestR2.Theseparatecontributionofindividualvari- theSAandSSASwouldnotfeedintothestatisticalanalysis, ablestostemflowyieldandefficiencywascomputedbythe butratherbeappliedtoreflectageneralcorrelationwithSFb methodofvariancepartitioning.Thesameanalysismethods inthisstudy. werealsoappliedtoidentifythemostinfluentialbio-/abiotic In this study, stemflow yield was defined as the stemflow factors affecting SFP and FR. The level of significance was volume production of branch (hereafter “stemflow produc- set at 95% confidence interval (p=0.05). The SPSS 20.0 tion”,SFb,mL),theequivalentwaterdepthonbasisofshrub (IBMCorporation,Armonk,NY,USA),Origin8.5(Origin- canopyarea(hereafter“stemflowdepth”,SFd,mm)andthe LabCorporation,Northampton,MA,USA)andExcel2013 stemflow percentage of the incident precipitation amount (MicrosoftCorporation,Redmond,WA,USA)wereusedfor (hereafter“stemflowpercentage”,SF%,%): dataanalysis. n X SF =10· SF /CA (5) d bi 3 Results i=1 SF%=(SFd/P)·100%, (6) 3.1 Meteorologicalcharacteristics whereSFbi isthestemflowvolumeofbranchi (mL),CAis Stemflow was measured at 36 rainfall events in this study, thecanopyarea(cm2),nisthenumberofbranchesandP is 18 events (209.8mm) in 2014 and 18 events (205.3mm) theincidentprecipitationamount(mm). in 2015, which accounted for 32.7 and 46.2% of total rain- SFP (mLg−1) was expressed as the SFb (mL) of unit fallevents,and73.1and74.9%oftotalprecipitationamount branch biomass (g) and represented the stemflow efficiency during the experimental period of 2014 and 2015, respec- ofdifferent-sizedbranchesinassociationwithabiomassal- tively(Fig.3).Therewere4,7,10,5,4and6rainfallevents locationpattern: atprecipitationcategoriesof≤2,2–5,5–10,10–15,15–20, and>20mm,respectively.Theaveragerainfallintensityof SFP=SFb/(BML+BMS). (7) incidentrainfalleventswas6.3±1.5mmh−1,andtheaver- agevalueofI ,I andI were20.3±3.9,15.0±2.9and 5 10 30 The FR was computed as the quotient of SFb and the prod- 9.2±1.6mmh−1, respectively. RD and RI were averaged uctofP andBBA(branchbasalarea;cm2)(Herwitz,1986). 5.5±1.1 and 63.1±8.2h. The average T, H, SR, WS and The value of (P ·BBA) equals to the precipitation amount WD were 16.5±0.5◦C, 85.9±2.2%, 48.5±11.2kwm−2, that would have been caught by the rain gauge occupying 2.2±0.2ms−1 and 167.1±13.9, respectively. As to the thesamebasalareainaclearing.AFRwithavaluegreater raindropattributes,D,V andAwereaveraged1.8±0.4mm, than1indicatedapositiveeffectofthecanopyonthestem- 6.1±0.1ms−1and19.6±1.2◦,respectively. flowyield(Carlyle-MosesandPrice,2006): FR=10·SF /(P ·BBA). (8) b Hydrol.EarthSyst.Sci.,21,1421–1438,2017 www.hydrol-earth-syst-sci.net/21/1421/2017/ C.Yuanetal.:Comparisonsofstemflowanditsbio-/abioticinfluentialfactors 1427 Figure3.Meteorologicalcharacteristicsofrainfalleventsforstemflowmeasurementsduringthe2014and2015rainyseasons. Figure 4. Verification of the allometric models for estimating the biomass and leaf traits of C. korshinskii. BML and BMS refer to the biomassoftheleavesandstems,respectively,andLABandLNBrefertotheleafareaandthenumberofbranches,respectively. 3.2 Species-specificvariationofplanttraits and 1.04 for the leaf traits) and an R2 value of 0.93–0.95. According to Yuan et al. (2016), the regression of S. psam- mophila had a slope of 1.13 and an R2 of 0.92. Therefore, Allometric models were developed to estimate the biomass thoseallometricmodelswereappropriate. andleaftraitsofthebranchesofC.korshinskiiandS.psam- C.korshinskiihadasimilaraveragebranchsizeandangle, mophilameasuredforstemflow.Theestimationqualitywas but a shorter branch length than did S. psammophila, e.g., verifiedbylinearregression.AsshowninFig.4,theregres- 12.5±4.2mm vs. 13.7±4.4mm, 60±18◦ vs. 60±20◦, sion of LAB, LNB, BML and BMS of C. korshinskii had and 161.5±35.0cm vs. 267.3±49.7cm, respectively. Re- anapproximately1:1slope(0.99forthebiomassindicators www.hydrol-earth-syst-sci.net/21/1421/2017/ Hydrol.EarthSyst.Sci.,21,1421–1438,2017 1428 C.Yuanetal.:Comparisonsofstemflowanditsbio-/abioticinfluentialfactors Table1.Comparisonofleaftraits,branchmorphologyandbiomassindicatorsofC.korshinskiiandS.psammophila. Planttraits C.korshinskii(categorizedbyBD,mm) S.psammophila(categorizedbyBD,mm) 5–10 10–15 15–18 >18 Avg.(BD) 5–10 10–15 15–18 >18 Avg.(BD) Leaftraits LAB(cm2) 1202.7 2394.5 3791.2 5195.2 2509.1±1355.3 499.2 1317.7 2515.2 3533.6 1797.9±1118.0 LNB 4787 11326 20071 29802 12479±8409 392 1456 3478 5551 2404±1922 ILAB(mm2) 25.4 21.3 18.9 17.5 21.9±3.0 135.1 93.1 72.6 64.3 93.1±27.8 SSAL(cm2g−1) 22.8 17.3 14.3 12.6 18.2±0.5 18.4 13.6 10.8 8.6 12.7±0.4 SSAS(cm2g−1) 3.4 2.3 1.9 1.6 2.5±0.1 10.4 5.4 3.3 1.9 5.1±0.3 Branch BD(mm) 8.17 12.49 16.61 20.16 12.48±4.16 7.91 12.48 16.92 19.76 13.73±4.36 morphology BL(cm) 137.9 160.3 195.9 200.7 161.5±35.0 212.5 260.2 290.4 320.1 267.3±49.7 BA(◦) 63 56 63 64 60±18 64 63 51 60 60±20 SA(cm2) 176.8 314.1 508.6 630.7 326.1±20.6 268.0 514.1 827.7 1312.3 711.0±38.9 Biomass BML(g) 13.9 19.0 30.2 41.4 19.9±10.8 5.4 18.0 40.0 61.3 27.9±20.7 indicators BMS(g) 62.9 121.4 236.4 375.8 141.1±110.8 23.0 81.4 188.5 295.5 130.7±101.4 PBMS(%) 82.0 86.3 88.7 90.0 85.6±3.1 80.8 81.8 82.5 82.8 81.9±0.8 Note:LABandLNBareleafareaandnumberofbranch,respectively.ILABisindividualleafareaofbranch.SSALandSSASarethespecificsurfacearearepresentingwithLABandSA,respectively.BD, BLandBAareaveragebranchbasaldiameter,lengthandangle,respectively.SAisthesurfaceareaofstems.BMLandBMSarebiomassofleavesandstems,respectively.PBMSisthepercentageofstem biomasstothatofbranch.Theaveragevaluesmentionedaboveareexpressedasthemeans±SE. garding branch biomass accumulation, C. korshinskii had a 3.3 Stemflowyieldofthefoliatedanddefoliated smallerBML(anaverageof19.9±10.8g)andalargerBMS C.korshinskiiandS.psammophila (anaverage141.1±110.8g)thandidS.psammophila(anav- erageof27.9±20.7and130.7±101.4g,respectively).Both In this study, stemflow yield was expressed as SF on the the BML and BMS increased with increasing branch size b branch scale and SF and SF% on the shrub scale. For the for these two shrub species. When expressed as a propor- d foliated shrubs, SF was averaged 290.6 and 150.3mL for tion, C. korshinskii had a larger PBMS than did S. psam- b individual branches of C. korshinskii and S. psammophila, mophilainalltheBDcategories.ThePBMS-specificdiffer- respectively,perincidentrainfalleventsduringthe2014and enceincreasedwithanincreasingbranchsize,rangingfrom 2015 rainy seasons. The SF was positively correlated with 1.2% for the 5–10mm branches to 7.2% for the >18mm b thebranchsizeandprecipitationforthesetwoshrubspecies. branches. As the branch size increased, SF increased from the aver- An increase in LAB and LNB, and a decrease in b age of 119.0mL for the 5–10mm branches to 679.9mL for ILAB, SSAL and SSAS were observed with increasing the >18mm branches of C. korshinskii and from 43.0 to branch size for these two shrub species. But at each 281.8mL for the corresponding BD categories of S. psam- BD level, C. korshinskii had on average a larger LAB (2509.1±1355.3cm2), LNB (12479±8409) and SSAL mophila. However, with increasing precipitation, a larger (18.2±0.5cm2g−1), but a smaller ILAB (21.9±3.0mm2) intra-specific difference in SFb was observed, which in- and SSAS (2.5±0.1cm2g−1) than did S. psammophila creased from the average of 28.4mL during rains ≤2mm (1797.9±1118.0g, 2404±1922, 12.7±0.4cm2g−1, to 771.4mL during rains >20mm for C. korshinskii and 93.1±27.8mm2 and 5.1±0.3cm2g−1, respectively) from9.0to444.3mLforthecorrespondingprecipitationcat- egoriesofS.psammophila.TheSF variedsignificantlyfor (Table 1). The inter-specific differences in the leaf traits b different rainfall characteristics and plant traits. The aver- decreasedwithincreasingbranchsize.Thelargestdifference age SF of 2375.9mL occurred for the >18mm branches occurred for the 5–10 mm branches, e.g., LNB and LAB b of C. korshinskii during rains >20mm in the 2014 and were 12.2-fold and 2.4-fold larger for C. korshinskii, and 2015 rainy seasons. However, the SF of 6.8mL occurred ILABwas5.3-foldlargerforS.psammophila. for the 5–10mm branches during rains ≤2mm. Compara- In the controlled field experiment, the defoliated sample tively,amaximumandminimumSF of2097.6and1.8mL branches of C. korshinskii and S. psammophila had sim- b occurredforS.psammophilaundersimilarbio-/abioticcon- ilar branch morphology and BML with those of the foli- ditions. ated branches. The average BD, BL, BA and BML were 10.5±4.4mm,168.5±39.5cm,65±15◦ and22.2±11.6g C. korshinskii produced a larger SFb than did S. psam- in C5, and 14.8±6.4mm, 258.6±39.0cm, 50±23◦ and mophilaforallBDandprecipitationcategories,andtheinter- 27.3±22.1ginS5,respectively. specificdifferencesinSFb alsovariedsubstantiallywiththe rainfall characteristics and the plant traits. A maximum dif- ference of 4.3-fold larger for the SF of C. korshinskii was b observed for the >18mm branches during rains ≤2mm at the 2014 and 2015 rainy seasons. As the precipitation in- Hydrol.EarthSyst.Sci.,21,1421–1438,2017 www.hydrol-earth-syst-sci.net/21/1421/2017/ C.Yuanetal.:Comparisonsofstemflowanditsbio-/abioticinfluentialfactors 1429 creased,theSF -specificdifferencedecreasedfrom3.2-fold from 1.64 to 0.80mLg−1 for the corresponding BD cate- b larger for C. korshinskii during rains ≤2mm to 1.7-fold goriesofS.psammophila.MaximumSFPvaluesof5.60and largerduringrains>20mm.ThelargestSF -specificdiffer- 4.59mLg−1 were recorded for C. korshinskii and S. psam- b enceoccurredforthe5–10mmbranchesinalmostallprecip- mophila,respectively.Additionally,C.korshinskiihadlarger itationcategories,butnocleartrendofchangewasobserved SFP than did S. psammophila for all precipitation and BD withincreasingbranchsize(Table2). categories. This inter-specific difference in SFP decreased SF and SF% averaged 1.0mm and 8.0% per incident with increasing precipitation from 2.7-fold larger for C. ko- d rainfalleventsduringthe2014and2015rainyseasonsforin- rshinskiiduringrains≤2mmto1.5-foldlargerduringrains dividualC.korshinskiishrubs,and0.8mmand5.5%forin- >20mm,anditincreasedwithincreasingbranchsize:from dividualS.psammophilashrubs,respectively.Theseparam- 1.3-fold larger for C. korshinskii for the 5–10mm branches eters increased with increasing precipitation, ranging from to2.0-foldlargerforthe>18mmbranches. 0.09mm and 5.8% during rains ≤2 to 2.6mm and 8.9% FR averaged 173.3 and 69.3 for the individual branches duringrains>20mmforC.korshinskii,andfromlessthan ofC.korshinskiiandS.psammophilaperrainfalleventsdur- 0.01mmand0.7%to2.2mmand7.9%forthecorrespond- ingthe2014and2015rainyseasons,respectively(Table5). ingprecipitationcategoriesofS.psammophila,respectively. As the precipitation increased, an increasing trend was ob- Additionally,theindividualC.korshinskiishrubshadalarger served, ranging from the average FR of 129.2 during rains stemflow yield than did S. psammophila in all precipitation ≤2mm to 190.3 during rains >20mm for C. korshinskii categories. The differences in SF and SF% maximized as and from the average FR of 36.7 to 96.1 during the corre- d an 8.5- and 8.3-fold larger for C. korshinskii during rains spondingprecipitationcategoriesforS.psammophila.FRin- ≤2mm and decreased with increasing precipitation to 1.2- creasedwithincreasingBAfromtheaverageof149.9forthe and1.1-foldlargerduringrains>20mm. ≤30◦ branches to 198.2 for the >80◦ branches of C. kor- While comparing the intra-specific difference of SF be- shinskii and from the average of 55.0 to 85.6 for the corre- b tween different leaf states, SF of the defoliated S. psam- sponding BA categories of S. psammophila. Maximum FR b mophilawas1.3-foldlargerthanthefoliatedS.psammophila values of 276.0 and 115.7 were recorded for C. korshinskii on average, ranging from 1.1-, 1.0- and 1.4-fold larger for and S. psammophila, respectively. Additionally, C. korshin- the 5–10, 10–15 and >15mm branches, respectively. A skii had a larger FR than S. psammophila for all precipita- largerdifferencewasnotedduringlighterrains(Table3).On tion and BA categories. The inter-specific difference in FR the contrary, SF of the defoliated C. korshinskii was aver- decreased with increasing precipitation from 3.5-fold larger b aged 2.5-fold smaller than the foliated C. korshinskii at all forC.korshinskiiduringrains≤2mmto2.0-foldlargerdur- rainfall events. Except for a 1.2-fold larger at the 5–10mm ing rains >20mm, and it decreased with an increase in the branches,the3.3-foldsmallerSF wasmeasuredatthe10– branchinclinationangle:from2.7-foldlargerforC.korshin- b 15mm and >15mm branches of the defoliated C. korshin- skii for the ≤30◦ branches to 2.3-fold larger for the >80◦ skiiasopposedtothefoliatedC.korshinskii(Table3).While branches. comparingtheSF -specificdifferenceatthesameleafstates, b asmallerSF ofthefoliatedS.psammophilawasnotedthan 3.5 Bio-/abioticinfluentialfactorsofstemflowyieldand b that of the foliated C. korshinskii. However, SF of the de- efficiency b foliatedS.psammophilawas2.0-foldlargerthanthatofthe defoliated C. korshinskii on average at nearly all BD cate- For both C. korshinskii and S. psammophila, BA was the goriesexceptforthe5–10mmbranches(Table3). only plant trait that had no significant correlation with SF b (r<0.13, p>0.05) as indicated by the Pearson correlation 3.4 StemflowefficiencyofC.korshinskiiand analysis. The separate effects of the remaining plant traits S.psammophila were verified by the partial correlation analysis, but BL, ILAB and PBMS failed this test. The rest of plant traits, With the combined results of SFP and FR, stemflow effi- including BD, LAB, LNB, BML and BMS, were regressed ciencywasassessedforC.korshinskiiandS.psammophila. with SF using the forward selection method. Biomass was b SFP averaged 1.95 and 1.19mLg−1 for individual C. kor- finally identified as the most important biotic indicator that shinskii and S. psammophila branches, respectively, per in- affected stemflow, which behaved differently in C. korshin- cident rainfall events during the 2014 and 2015 rainy sea- skii for BMS and in S. psammophila for BML. The same sons (Table 4). As precipitation increased, SFP increased methods were applied to analyze the influence of meteoro- from 0.19mLg−1 during rains ≤2mm to 5.08mLg−1 dur- logical characteristics on SF of these two shrub species. b ing rains >20mm for C. korshinskii, and from 0.07 to TestedbythePearsoncorrelationandpartialcorrelationanal- 3.43mLg−1 for the corresponding precipitation categories yses, SF related significantly with P, I , RD and H for b 10 for S. psammophila. With an increase in branch size, SFP C. korshinskii, and with P, I , I , I for S. psammophila. 5 10 30 decreased from 2.19mLg−1 for the 5–10mm branches to The stepwise regression finally identified the precipitation 1.62mLg−1forthe>18mmbranchesofC.korshinskii,and amountasthemostinfluentialmeteorologicalcharacteristics www.hydrol-earth-syst-sci.net/21/1421/2017/ Hydrol.EarthSyst.Sci.,21,1421–1438,2017 1430 C.Yuanetal.:Comparisonsofstemflowanditsbio-/abioticinfluentialfactors Table2.Comparisonofstemflowyield(SFb,SFdandSF%)betweenthefoliatedC.korshinskiiandS.psammophila. Intra-andinter-specific Stemflow BD Precipitationcategories(mm) Avg. differences indicators categories ≤2 2–5 5–10 10–15 15–20 >20 (P) (mm) Intra-specificdifferencesin 5–10 10.7 29.8 73.5 109.9 227.6 306.1 119.0 C.korshinskii(CK) 10–15 26.0 64.0 166.1 236.0 478.6 689.7 262.4 SFb(mL) 15–18 44.3 103.3 279.9 416.6 826.0 1272.3 464.5 >18 69.5 145.4 424.4 631.4 1226.9 1811.7 679.9 Avg.(BD) 28.4 67.3 180.6 264.6 529.2 771.4 290.6 SFd(mm) n/a 0.1 0.2 0.6 0.9 1.9 2.6 1.0 SF%(%) n/a 5.8 6.6 8.8 7.5 10.1 8.9 8.0 Intra-specificdifferencesin 5–10 2.8 8.9 28.8 47.2 66.5 120.0 43.0 S.psammophila(SP) 10–15 7.6 23.2 76.6 134.6 188.3 353.5 121.8 SFb(mL) 15–18 12.0 35.9 121.6 223.4 319.4 592.6 201.5 >18 16.2 52.3 165.5 289.2 439.6 860.4 281.8 Avg.(BD) 9.0 28.0 91.6 162.2 234.8 444.3 150.3 SFd(mm) n/a <0.1 0.1 0.5 0.9 1.3 2.2 0.8 SF%(%) n/a 0.7 3.0 6.1 6.8 7.2 7.9 5.5 Inter-specificdifferences 5–10 3.8 3.3 2.6 2.3 3.4 2.6 2.8 (theratioofthestemflowyield 10–15 3.4 2.8 2.2 1.8 2.5 2.0 2.2 ofCKtothatofSP) SFb 15–18 3.7 2.9 2.3 1.9 2.6 2.2 2.3 >18 4.3 2.8 2.6 2.2 2.8 2.1 2.4 Avg.(BD) 3.2 2.4 2.0 1.6 2.3 1.7 1.9 SFd n/a 8.5 2.2 1.3 1.0 1.5 1.2 1.3 SF% n/a 8.3 2.2 1.4 1.1 1.4 1.1 1.4 Note:BDisthebranchbasaldiameter;Pistheprecipitationamount;CKandSParetheabbreviationsofC.korshinskiiandS.psammophila,respectively. n/ameansnotapplicable. forthetwoshrubspecies.AlthoughI wasanotherinfluen- important plant trait affecting SFP during rains ≤10mm of 10 tialfactorforC.korshinskii,itonlymadea15.6%contribu- these species. However, during heavier rains >15mm, BD tiontotheSF onaverage. andPBMSwerethemostsignificantbioticfactorsforC.ko- b SF andSF hadagoodlinearrelationshipswiththepre- rshinskii and S. psammophila, respectively. For these two b d cipitationamount(R2≥0.93)forbothshrubspecies(Fig.5). shrubs species, it was leaf trait (ILAB) and branch traits The >0.9 and >2.1mm rains were required to start SF (biomassallocationpatternandbranchsize)thatplayedbig- b for C. korshinskii and S. psammophila, respectively. This ger roles on SFP during lighter rains ≤10mm and heavier wasclosetothe0.8and2.0mmprecipitationthresholdcal- rains >15mm, respectively. Therefore, it seemed that the culated with SF . Moreover, the precipitation threshold in- rainfallinterceptionprocessofleavescontrolledSFPduring d creasedwithincreasingbranchsize.Theprecipitationthresh- the lighter rains, which functioned as the water resource to oldvalueswere0.7,0.7,1.4and0.8mmforthe5–10,10–15, producestemflow.Nevertheless,whilewatersupplywasad- 15–18and>18mmbranchesofC.korshinskii,and1.1,1.6, equateduringheavierrains,thestemflowdeliveringprocess 2.0and2.4mmforthebranchesofS.psammophila,respec- ofbranchesmightbethebottleneck. tively. SF%ofthetwoshrubspecieswereinverselyproportional totheprecipitationamount.Astheprecipitationincreased,it 4 Discussion graduallyapproachedasymptoticvaluesof9.1and7.7%for C. korshinskii and S. psammophila, respectively. As shown 4.1 Differencesofstemflowyieldandefficiency inFig.5,fastgrowthwasevidentduringrains≤10mm,but betweentwoshrubspecies SF%slightlyincreasedafterwardsforbothshrubspecies. Precipitation amount was the most important factor af- C. korshinskii produced stemflow in a larger quantity com- fecting SFP and FR for C. korshinskii and S. psammophila, pared with S. psammophila in all precipitation categories, but the most important biotic factor was different. BA was particularly at the 5–10mm young shoots during light rains the most influential plant trait that affected FR of these two ≤2mm(Table2).Althoughthegreateststemflowyieldwas shrubspeciesatallprecipitation levels.ILABwasthemost observed during rains >20mm for the two shrub species, Hydrol.EarthSyst.Sci.,21,1421–1438,2017 www.hydrol-earth-syst-sci.net/21/1421/2017/