Quantifying water use by temperate deciduous forests in South Korea : roles of species diversity, canopy structure, and complex terrain

Quantifying water use by temperate deciduous forests in South Korea: roles of species diversity, canopy structure, and complex terrain Dissertation zur Erlangung des Doktorwürde (Dr. rer. nat.) der Fakultät für Biologie, Chemie und Geowissenschaften der Universität Bayreuth von Eunyoung Jung aus Deajeon, Süd-Korea Bayreuth, Juni 2013 ! i Abstract About seventy percent of South Korea is covered with forests, most of which are found in the mountain regions since mountains receive more rainfall and are difficult terrains not suitable for agriculture. Because mountains are important water sources for cities and human population downstream, performing water balance for forest catchments has become a research priority. The ongoing shift from coniferous to speciesrich deciduous forests due to a changing government policy and the anticipated changes in future climate, associated with increasing amount of rainfall and temperature will also impact forest water use, calling for an urgent need to understand how forests, in their current status, use water. The knowledge is vital for predicting water requirements for the future forest. The warm-deciduous temperate forests found in South Korea, however, have a high diversity of tree species, have multi-layered canopies and are mostly located on rugged mountainous terrains, which make it difficult to quantify forest water use, a basic requirement for catchment water budgeting. The main objectives of this study were to: (1) identify the roles of species diversity in tree and forest water use, (2) examine the impact of canopy structure on forest transpiration, and (3) evaluate the influence of terrain on forest water use. Site-specific studies were carried out in three different natural deciduous forests, namely, Gyebang (GB), Gwangneung (GN) and Haean (HA) forest sites, representing the general structure of S. Korean forests. GB site is known for its high species diversity, GN site is an old forest growth at climax, with clearly defined understory and overstory canopy layers while the HA site was located with in a catchment, with strong elevation changes within short horizontal distances, rising from 400 to 1,000 m a.s.l., and in different aspects. Four locations with varying elevations and aspects were chosen in the HA site. Tree water use (TWU) and canopy transpiration (EC) were estimated from sap flux density measured with thermal dissipation probes. Understory transpiration (EU) was measured using stem heat balance while ecosystem evapotranspiration (Eeco) was determined using eddy covariance technique. Air temperatures (Ta), precipitation, solar radiation, vapor pressure deficit (VPD), wind speed were measured from weather stations and soil water content was measured from frequency domain reflectometry (FDR) sensors at the respective study sites. Vegetation surveys, including diameter at breast height (DBH), tree density, species composition, sapwood area (AS), and leaf area index were performed in all the sites. Canopy conductance (GC) and stomatal sensitivity to VPD were assessed based on transpiration and microclimate measured at each site. A functional allometric relationship was established between AS and DBH, and also between TWU and DBH for all the study sites; first for single species and then combining all the species either in a single site or in all the sites. Irrespective of tree species, AS and maximum TWU were significantly correlated with DBH in a power function for AS (R2 = 0.77, P <0.0001) and both in power (R2 = 0.63, P <0.0001) and sigmoid functions (R2 = 0.66, P <0.0001) for TWU, for the co-occurring species as well as across the sites, ! ii suggesting that DBH can be a good predictor of stand AS and maximum TWU, based on the established allometric functions. Early bud break and development of the understory compared to the overstory canopy resulted in an earlier onset of forest transpiration, with EU contributing 22% and 14% between April and May to the total forest transpiration. This high contribution was favored by high radiation and VPD in the understory, since the overstory was still undeveloped and open. Despite diminishing VPD and light conditions in the understory between June and August, the understory continued to transpire a substantial amount of water, contributing 10% of the total transpiration. The seasonal patterns of both EO and EU were synchronized to canopy development, while VPD and radiation determined daily trends. EO and EU accounted for 80% of Eeco in spring but only 60% during the monsoon period due to lowered radiation input, VPD, and plant area index (PAI). Thus, Eeco is largely influenced by transpiration rate and its seasonal variation and also canopy structure. Early saturation of EC at relatively low VPD and also a rapid decrease in GC with increasing VPD were observed in the forest stand located at the highest elevation studied (950 m) in the HA site, compared to the GN and the other forest stands in HA. These differences in transpiration rates and stomatal response can be explained by greater stomatal sensitivity to VPD of 0.83 found at the 950 m site compared to 0.63–0.66 in the other study sites. However, the main controlling factor of the change in stomatal sensitivity at the 950 m stand is uncertain. Although maximum daily EC were correlated with AS of the forest stands at different sites (R2 = 0.78, P <0.01), annual EC declined with increasing elevation, i.e., 176 >175 >110 >90 mm year−1 at 340 >450 >650 >950 m, respectively. Decline in total EC was due to the decline in annual Ta, daytime VPD, and length of growing season at higher elevations. The GB site, which was located at 960 m elevation, however, did not display a same response pattern as those observed at the 950 m site. It is likely because these sites were under different environmental conditions, i.e., GB site is exposed to higher Ta and higher humidity, and is sheltered (lower wind speeds). These observations emphasize the complexity associated with estimation of transpiration in rugged terrains, since general principles do not always apply and the spatial patterns of forest transpiration are complex. Complexity arising from multiple tree species composition when estimating forest water use can be reduced by applying functional allometric relationship linking tree size and water use. Forest canopy structure and physical location should be taken into account since they influence the way forests use water resources by altering microclimate and plant physiology. Based on our findings, estimation of forest water use on rugged terrains require repeated measurements at relatively small spatial scales since the driving factors change rapidly over very narrow vertical distances. ! iii Zusammenfassung In Südkorea ist ca. 70% der Fläche von Wald bedeckt, welcher sich hauptsächlich über gebirgige und durch hohe Niederschläge gekennzeichnete, landwirtschaftlich nicht nutzbare Gebiete erstreckt. Diese Gebirgsregionen stellen eine wichtige Wasserressource für die städtische Bevölkerung dar, sodass die Wasserbilanzierung von bewaldeten Einzugsgebieten zu einem primären Forschungsgegenstand geworden ist. Die gegenwärtige Verlagerung von Nadelwäldern hin zu artenreichen Laubwäldern als Folge einer sich verändernden Strategie der südkoreanischen Regierung, zur Anpassung an die sich verändernden klimatischen Bedingungen, wie steigende Niederschläge und Temperaturen, führt zu einem veränderten Wasserverbrauch der Waldbestände. Daher sollte der Status des gegenwärtigen Wasserverbrauchs der Bestände umgehend untersucht werden, um Vorhersagen über den zukünftigen Verbrauch treffen zu können. Die im warm-gemäßigten Klima in Südkorea verbreiteten Wälder sind durch eine hohe Artendiversität und einen vielschichtigen Aufbau in ihrer Struktur gekennzeichnet. Durch ihre Lage im schroffen, zerklüfteten Gelände, gestaltet sich die quantitative Erfassung des Wasserverbrauchs dieser Wälder umso schwieriger. Seine Erfassung ist jedoch für eine Wasserbilanzierung in den Waldbeständen eine Voraussetzung. Das übergeordnete Ziel dieser Studie ist (i) zu identifizieren, welche Rolle die Artendiversität in Bezug auf den Wasserverbrauch sowohl einzelner Baumindividuen als auch des gesamten Waldbestandes spielt, (ii) den Einfluss der Kronendachstruktur auf die Transpiration des Waldbestandes zu untersuchen und (iii) den Einfluss des Geländes auf den Wasserverbrauch des Waldes zu evaluieren. Für die Untersuchungen dieser Studie wurden drei, für Südkorea repräsentative, natürliche Laubwälder in Gyebang (GB), Gwangneung (GN) und Haean (HA) ausgesucht. Das GB-Waldgebiet zeichnet sich besonders durch eine hohe Artenvielfalt aus. Das GN-Waldgebiet besteht aus einer alten KlimaxWaldgesellschaft und lässt sich strukturell in eine Unterholzschicht und eine Baumschicht gliedern. Haean (HA) zeichnet sich hingegen durch einen starken Höhengradient (400 m bis 1000 m ü. NN) über kurze Distanzen und durch eine Exposition in alle Himmelrichtungen aus, sodass in HA insgesamt vier Standorte mit unterschiedlichen Höhenlagen und Expositionen für die Untersuchungen ausgewählt wurden. Der Wasserverbrauch von Baumindividuen (TWU) und die Kronendachtranspiration (EC) wurden mittels der Saftflussmethode, welche die Saftflussdichte misst, untersucht. Die Unterholztranspiration (EU) wurde mit der „stem heat balance“ Methode (SHB) gemessen. Am Standort GN wurde zur Erfassung der Ökosystemevapotranspiration (Eeco) die Eddy-Kovarianz-Methode benutzt. Die Installation von Wetterstationen diente der Erfassung von Lufttemperatur (Ta), Niederschlag, Solarstrahlung, Wasserdampfsättigungsdefizit (VPD) und der Windgeschwindigkeit. Zusätzlich wurde FDR-Sonden installiert, um den Bodenwassergehalt zu messen. Zur Untersuchung der Vegetation wurde der Stammdurchmesser in Brusthöhe (DBH), die Bestandesdichte, die Splintholzfläche (AS), der ! iv Blattflächenindex und die Artenzusammensetzung an allen Standorten erfasst. Basierend auf der Transpiration und dem Mikroklima wurde die Kronendachleitfähigkeit (GC) und die stomatäre Empfindlichkeit bezüglich des VPD untersucht. Nicht nur die Berücksichtigung einzelner Arten, sondern auch die Einbeziehung aller Arten an einem Standort sowie aller Standorte ergab eine funktionale allometrische Beziehung, sowohl zwischen As und DBH, als auch zwischen TWU und DBH. Unabhängig von der jeweiligen Baumart zeigten die Analysen einen signifikanten Zusammenhang zwischen As und BHD, ausgedrückt in einer Potenzfunktion (P <0.0001). Auch der maximale TWU korrelierte signifikant mit dem DBH im Sinne von Potenzund sigmoidalen Funktionen (P <0.0001), sowohl für einzelne Arten als auch standortübergreifend. Aufgrund dieser Ergebnisse ist der DBH unter Berücksichtigung der eingeführten allometrischen Funktionen ein gute Schätzgröße für die Bestandessplintholzfläche As und dem maximalen Wasserverbrauch (TWU). Der im Vergleich zur Baumschicht frühe Blattaustrieb im Unterholz führte zu einem erhöhten Anteil der Unterholztranspiration von 22% im April und 17% im Mai an der Gesamttranspiration des Waldbestandes. Dieser hohe Anteil wurde durch die hohe Einstrahlung und das hohe Wasserdampfsättigungsdefizit (VPD) im Unterholz aufgrund der lichten Baumschicht begünstigt. Allerdings trug das Unterholz auch von Juni bis August mit einem beträchtlichen Anteil von 10% zur Gesamttranspiration bei, obwohl die solare Einstrahlung und das VPD durch die entwickelte Baumschicht geringer waren. Der saisonale Verlauf von EO und EU verlief synchron zur Entwicklung des Kronendaches, während hingegen der Tagesgang durch das Mikroklima, die solare Einstrahlung und das VPD gesteuert wurde. Im Frühling trugen EO und EU mit einem Anteil von 80% zur gesamten Eeco bei. Während der Monsunperiode verringerte sich dieser Anteil jedoch aufgrund einer geringeren Einstrahlung, einem geringerem VPD und einem minimierten Pflanzenflächenindex (PAI) auf 60%. Die Eeco wird daher stark durch den saisonalen Verlauf der Transpirationsrate als auch durch die Kronendachstruktur beeinflusst. Im Vergleich zu GN, GB, und den niedrigeren Standorten im HA-Einzugsgebiet konnte eine frühe Sättigung von Ec bei relativ geringem VPD und auch eine relativ starke Abnahme von GC bei steigendem VPD am höchsten Standort in 950 m ü.NN in HA beobachtet werden. Die unterschiedlichen Transpirationsraten und die stomatäre Reaktion können durch eine größere stomatäre Empfindlichkeit in Bezug auf ein VPD von 0.82 erklärt werden. Im Vergleich dazu wiesen die anderen Standorte nur ein VPD zwischen 0.63−0.66 auf. Dennoch sind die verursachenden Faktoren für eine Veränderung der stomatären Empfindlichkeit in größeren Höhenlagen mit gewisser Unsicherheit behaftet. Obwohl die maximale tägliche EC mit AS (R²=0.78, P <0.01) der Waldbestände an verschiedenen Standorten korrelierte, konnte eine Abnahme der jährlichen Ec, mit 176 > 175 > 110 > 90 mm Jahr −1, bezüglich der Höhenlage, 340 > 450 > 650 > 950 m ü.NN beobachtet werden. Eine Abnahme der gesamten EC läßt sich auf die Abnahme der Jahresdurchschnittstemperatur, des tageszeitlichen VPD und der Länge der Vegetationsperiode in größeren Höhenlagen zurückführen. Die Ergebnisse des Standortes GB, welcher sich in ähnlicher Höhenlage auf 960 ! v m befand, zeigten jedoch im Vergleich zum HA-Standort in 950 m ü.NN unterschiedliche Reaktionsmuster, welche auf die geschützte Lage mit höheren Temperaturen, höherer Luftfeuchtigkeit und geringeren Windgeschwindigkeiten zurückzuführen sind. Diese Beobachtungen unterstreichen die Schwierigkeiten bei der Schätzung der Bestandestranspiration im zerklüfteten, steilem Gelände, da aufgrund der komplexen räumlichen Muster der Waldbestandtranspiration nicht immer allgemeingültige Prinzipien abzuleiten sind. Durch die multiple Zusammensetzung der Baumarten in einem Bestand gestaltet sich die Schätzung des Wasserverbrauchs als schwierig, dennoch können diese Schwierigkeiten durch die Anwendung funktionaler allometrischer Beziehungen zwischen dem Baumumfang und dem Wasserverbrauch reduziert werden. Insbesondere sollte die Kronendachsstruktur und die geografische Lage ausreichend berücksichtigt werden, da diese Faktoren den Wasserverbrauch der Waldbestände durch die Veränderung des Mikroklimas und der Pflanzenphysiologie beeinflussen. In Anbetracht der Ergebnisse sollte die Schätzung des Wasserverbrauchs der Bestände anhand wiederholter Messungen auf relativ kleinräumiger Skala erfolgen, da sich die treibenden Faktoren schnell und auf relativ kurzer Distanz ändern können. ! vi Acknowledgements First of all, I would like to thank PD Dr. Dennis Otieno, Department of Plant Ecology, University of Bayreuth, for his guidance on how to be a scientist during my Ph. D. He is not only the best teacher in my whole life but also the best friend who I was able to discuss with about anything. I am very honored to be the first Ph. D. graduate under him in Germany. I also thank Prof. Dr. John Tenhunen, Department of Plant Ecology, University of Bayreuth, for giving me a lot of opportunities to learn and to see more in the world. I admire his enthusiasm on the research, positive attitude in life, and wise leadership. I would like to thank Dr. Hyojung Kwon, Oregon State University, for her encouragement and thoughtful comments on analyzing the data, writing the manuscripts, and also having a meaningful life. I learned a lot from her through uncountable meetings and discussions while she stayed in Bayreuth. I am grateful to Margarete Wartinger for her excellent support on preparation for the field work in Korea and Africa. I believe that I could not accomplish all the field work without her help in solving unexpected technical problems. I also thank Fiederike Rothe, Bärbel Heindl-Tenhunen, and Sandra Thomas for their tremendous cares for me to have a pleasant stay in Bayreuth. I thank all my colleagues of the TERRECO and KiLi projects, who gave me inspiration for the possible future research and enjoyable life in Bayreuth, Haean, and Nkweseko. Special thanks to Steve Linder for taking care of me in every possible ways, to Bora Lee for just being there for me, to Bumsuk Seo for stimulating me to study our subject in more depth, Saem Lee for listening to me attentively all the time, Marianne for her help on German summary, Emily for English corrections, Thomas for German corrections, and Sina for encouraging me and teaching me promising spirit of her. I would like to thank all my friends who I spent unforgettable time together in Bayreuth: Family Gärditz, Family Jeoung, Family Mader, Family Otto, Family Park Hoseon, Frank, Heera, Yangmin, and Yoolim. Finally, I would like to express my special thanks to my husband Jae-Woo and my parents for their endless support, trust, patience, and love. ! vii Table of contents Abstract……………………………………………………………………………………………………..i Zusammenfassung………………………………………………………………………………………...iii Acknowledgements………………………………………………………………………………………..vi Table of contents………………………………………………………………………………………….vii List of figures………………………………………………………………………………………………x List of tables……………………………………………………………………………………………..xiii List of abbreviations and symbols………………………………………………………………………..xv 1 Detailed summary.………………………………………………………………………………1 1.1 General introduction and literature review.…………………………………………………………..1 1.1.1 Temperate deciduous forests: distribution and structure.…………………………………………….…1 1.1.2 Forests in South Korea…………………………………………………………………………………..2 1.1.2.1 History of forests in South Korea………………………………………………………..2 1.1.2.2 Current status of forests in South Korea………………………………………………....2 1.1.2.3 The future of forests in South Korea…………………………………………………….3 1.1.3 Regulation of water use by forest ecosystems…………………………………………………………..4 1.1.4 Estimation of water use by forests in South Korea………………………………………………….…..5 1.1.5 Statement of research challenges………………………………………………………………………..6 1.1.6 Objectives of the research……………………………………………………………………………….7 1.2 General materials and methods……………………………………………………………………...10 1.2.1 Description of study sites……………………………………………………………………...10 1.2.2 Methods………………………………………………………………………………………..15 1.2.2.1 Sap flow measurements………………………………………………………………...15 1.2.2.2 Canopy conductance……………………………………………………………………18 1.2.2.3 Biometric measurements……………………………………………………………….19 1.2.2.4 Micrometeorological measurements……………………………………………………20 1.2.2.5 Eddy covariance flux measurements…………………………………………….……..20 1.2.2.6 Water use efficiency……………………………………………………………………20 1.3 General results and discussions……………………………………………………………………...21 1.3.1 Tree and forest water use in diverse species composition ………………………………………….….21 1.3.2 Impact of the forest structure on forest water use……………………………………………………...25 1.3.3 Forest water use in the complex terrain………………………………………………………………..26 1.4 General conclusions…………………………………………………………………………………30 1.5 List of manuscripts and specification of contributions……………………………………………...31 1.6 References…………………………………………………………………………………………...32 2 Up-scaling to stand transpiration of an Asian temperate mixed-deciduous forest from single tree sap flow measurements……………………………………………………………41 Abstract…………………………………………………………………………………………………...41 2.1 Introduction………………………………………………………………………………………….42 2.2 Materials and methods………………………………………………………………………………43 2.2.1 Study site…………………………………………………………………………………………….…43 2.2.2 Vegetation.…………………………………… ……………………………………………………….43 2.2.3 Micrometeorology……………………………………………………………………………………...44 2.2.4 Tree allometrics………………………………………………………………………………………...45 ! xiv Table 4.4 Annually averaged microclimates (solar radiation (RS), air temperature (Ta), mean daytime vapor pressure deficit (VPD), annual rainfall and wind speed) and soil water content (θ) during growing season of the study sites in Haean catchment, South Korea in 2010. ± are standard deviation (SD). Table 4.5 Maximum leaf area index (LAI), length of the growing season, total canopy transpiration (EC, mm) during the growing season and mean ring width for Quercus mongolica at each study site. ± are standard deviation (SD). Table 4.6 Canopy conductance (GC, mm s−1) on clear days in June, its respective vapor pressure deficit (VPD, kPa) for daytime (from 8:00h to 18:00h) and GCref (GC at VPD = 1 kPa) at each site. ± are standard deviation (SD). ! xv List of abbreviations and Symbols Abbreviation/Symbol Definition Unit AIC Akaike’s information criterion - AS sapwood area [m2] BA basal area [m2] cp specific heat at constant pressure [J kg−1 K−1] DBH diameter at breast height [cm] EC canopy transpiration [mm h−1, mm d−1] Eeco ecosystem evapotranspiration [mm h−1, mm d−1] Emax maximum stand transpiration [mm h−1, mm d−1] EO overstory transpiration [mm h−1, mm d−1] EU understory transpiration [mm h−1, mm d−1] Fd sap flux density [g m−2 s−1] GA aerodynamic conductance [m s−1] gb boundary layer conductance [m s−1] GC canopy conductance [mm s−1] GCref canopy conductance at VPD = 1 kPa [mm s−1] gt turbulent conductance [m s−1] GV gas conductance of water vapor [m3 kPa kg−1 K−1] KoFlux Korean Flux group - LAI leaf area index [-] LAImax maximum leaf area index [-] LAIU understory leaf area index [-] MLT modified lookup table - PAI plant area index [-] PAR photosynthetic active radiation [mol m−2 s−1] PPFD photosynthetic photon flux density [mol m−2 s−1] ! xvi RMSE Root-mean-square-error [-] RN net radiation [W m−2] RS solar radiation [W m−2] SHB stem heat balance - SLA specific leaf area [cm2 g−1] SWAT Soil and Water Assessment Tool - Ta air temperature [°C] TDP thermal dissipation probes - Tk air temperature in kelvin [K] U wind speed above the vegetation layer [m s−1] VPD vapor pressure deficit [kPa] WUE water use efficiency - z canopy height [m] z0 roughness length [m] α attenuation coefficient for wind speed inside the canopy [-] Δ change of saturation water vapor pressure with temperature [Pa K−1] γ psychometric constant [Pa K−1] θ soil water content [%] κ von Karman constant [-] λ latent heat of vaporization of water [J kg−1] ρ density of dry air [kg m−3] ρw density of water [kg m−3] ! Chapter 1 – Detailed summary ! 1 Chapter 1 Detailed summary 1.1 General introduction and literature review 1.1.1 Temperate deciduous forests: distribution and structure Temperate deciduous forests are widely distributed across the globe, covering an area of approximately 7.8 million km2 worldwide (Allaby 2006). They occur in Eastern North America, western and central Europe, eastern Asia, Near East, and in parts of South America such as Patagonia and Chile. These forests occur in relatively warm-temperate moist climates, with an average temperature of the coldest month ranging between −18 and −3°C and the warmest month ranging between 18 and 30°C (Röhrig 1991a; Allaby 2006). The mean annual precipitation ranges from 750 to 1,500 mm, and is allocated evenly throughout the season for most of the regions except for eastern Asia, which experiences severe rain storms during summer (Röhrig 1991a; Allaby 2006). Compared to the forests in Europe and North America, the temperate deciduous forests of eastern Asia are two or three times higher in plant species diversity (Latham and Ricklefs 1993; Qian and Ricklefs 1999). Historically, this substantial difference in plant diversity appears to result from greater physiographic heterogeneity in Asia, which allowed for allopatric speciation in response to climate and sea level fluctuations after temperate forest zones became disjunct in the late Tertiary (Qian and Ricklefs 2000). Moreover, repeated glaciation during the Pleistocene was more extreme in Europe and North America than in eastern Asia (Qian and Ricklefs 2000). In eastern North America, more uniform climate and simpler geography have not fostered evolutionary conditions among the same families as in eastern Asia (Qian and Ricklefs 2000). In the case of Europe, even lower diversity is found in forest vegetation due to the restrictions on refugia for migrating species that were imposed by the Alps barrier and the location of the Mediterranean Sea during Pleistocene glaciations (Ellenberg 1978), resulting in large numbers of extinctions. Generally, the temperate deciduous forests of eastern Asia can be classified into two main groups, namely cooland warm-temperate deciduous forests (Kira 1991). The cool-temperate deciduous forests mainly inhabit western Japan, while the warm-temperate deciduous forests occur in other parts of Japan, China and Korean Peninsula (Nakashizuka and Iida 1995). Characteristically, the cool-temperate deciduous forests are dominated by Fagus crenata, which accounts for more than 80% of the total forest basal area (Nakashizuka 1987). Warm-temperate deciduous forests, on the other hand, are species-rich, for example, Chapter 1 – Detailed summary ! 2 the dominant genus Quercus occur as 66 different species (Röhrig 1991b). This high species diversity likely resulted from the warm and humid summer conditions associated with high radiation and moisture inputs from the monsoon (Röhrig 1991a). Dominant genera include Quercus, Carpinus, Ulmus, and Tilia (Velichko and Spasskaya 2002). Thus, the warm-temperate deciduous forests of Korea and other parts of Asia represent an important ecosystem type with multi-layered physiognomy (Kim 2002), where both the overstory and understory are well developed and display diverse species compositions. As a result, large differences are expected in terms of water use between the cool-and warm-temperate deciduous forests. 1.1.2 Forests in South Korea 1.1.2.1 History of forests in South Korea South Korea is a mountainous country, with 70% of its land covered by mountains with elevations of up to 2,000 m a.s.l. separated by deep and narrow valleys. Two thirds of the land area is, therefore, difficult terrain that is not suitable for agriculture and is, instead, under forest cover. Most streams originate from the high elevation forests, and supply most of the water requirements for the population downslope. During the colonial period of 1910 to 1945 and the Korean War of 1950 to 1953, forests were excessively devastated, which led to frequent floods and landslides. In order to rehabilitate forests, the government initiated a nation-wide, large-scale reforestation program from 1973, with a goal of restoring 1 million ha of forests within a short period by planting fast growing tree species, mostly conifers (Pinus koraiensis, Abies holophylla and Larix leptolepis) and restricting the burning of forests to create land for cultivation. Through this action, about 730 thousand ha of degraded land was restored through an extensive plantation of about 10 billion trees on an area of over 350 thousand ha (Lee et al. 1997; Korea Forest Service 2009). 1.1.2.2 Current status of forests in South Korea In the recent past, forest coverage in South Korea has been declining at an alarming rate of 40 thousand ha per year (Korea Forest Service 2006), due to conversion into agricultural land, urbanization and expansion of the manufacturing industry (Youn et al. 2009). The proportional coverage of coniferous and deciduous forests has changed as well. Coniferous forests have declined from 55% to 42% (600 thousand ha), while the area covered with deciduous forests has increased from 17% to 40% (540 thousand ha), between 1972 and 2008 (Korea Forest Service 2009). Although total forest area has declined, the growing stock of trees has increased 11 times during the last 40 years (i.e., growing stock was 70 million m3 in 1972 and increased to 800 million m3 in 2010), as a result of natural re-growth and improved forest management practices (Korea Forest Service 2011). A substantial proportion of the forests is at early and mid succession stages. Chapter 1 – Detailed summary ! 3 According to the Korea Forest Service report of 2009, forests under 10 years and over 51 years of age cover 7% and 2% of the total forest area, respectively. Most (close to 70%) of the forest area is occupied by trees aged between 20 and 40 years. Oaks occupy 75% of the area under natural deciduous forests, with Q. mongolica being one of the dominant species growing from 100 m to 1,800 m a.s.l., but mostly abundant at around 700 m a.s.l. (Chung and Lee 1965). 1.1.2.3 The future of forests in South Korea The climate of South Korea has experienced a gradual warming during the 20th century (Oh et al. 2004; National Institute of Environmental Research 2011). In the last century alone, the average air temperature increased by about 1.5°C, which is twice as high compared to the global warming projections. This has been attributed to rapid industrialization (Oh et al. 2004; Kwon 2005). This temperature increase is contributing to the current shifts in plant species distribution ranges. For example, bamboos (Phyllostachys) have shifted from 35°N to 36°N in the continental region and 38°N in the eastern coastal region during the last 200 years (Gong 2001; Oh et al. 2004). Also, the tree line of Korean fir (Abies koreana), which grows only at high elevations, has been continuously moving upwards and the species is now threatened with high mortality rate of 20–50% in its natural range (Lim et al. 2008). Based on the high-resolution climate simulations for 2021 to 2050, under the B2 scenario (IPCC 2000), warming in the range of 1–4 °C is expected in the northern part of S. Korea during the cold season (Im et al. 2008). Precipitation will also be regionally variable, with increased summer rains in the north, but a decline in the southern regions (Im et al. 2008). Based on these climate projections, deciduous and mixed forests could increase by 60% and 10%, respectively, while coniferous forest cover could decrease by 10% by the year 2080 (Shin et al. 2012). The projected changes in climate and forest structure stimulate interests in assessing the ongoing changes and how they will influence forest water budget in the shortand long-term since most of the country’s water requirement is met by water from forested mountains. Already, hydrological simulations with the SWAT (Soil and Water Assessment Tool) model, using climate scenarios of ca. 4°C and 20–35% precipitation increase, respectively, show that an evapotranspiration (Eeco) increase of 15–20% is expected between 2000 and 2080 (Park et al. 2011). Most forests are located in mountains and they are perceived as water reservoirs for agricultural lands and populations downstream. Historically, rulers of the ancient Korean kingdoms gave a high priority to forest protection since they considered the water regulation function of forests as a fundamental service for agriculture (Youn et al. 2009), a concept that is widespread among South Koreans to date. Forest soils perform as reservoirs for water from precipitation and losing water by runoff and Eeco. Any vegetation changes strongly influence Eeco, which necessarily affects runoff, because Eeco is one of the large components of the forest hydrologic budget. For example, conifer forests produce less runoff than deciduous forests due to their higher rates and longer season for Eeco (Swank Chapter 1 – Detailed summary ! 4 and Douglass 1974). Understanding how forest catchments store rain water and also the regulation of water release from forests, either as river discharge or through Eeco is, therefore, critical for the management of natural water resources (Chapin et al. 2011). 1.1.3 Regulation of water use by forest ecosystems In most forest ecosystems, canopy transpiration is determined by the prevailing microclimate, soil moisture status and the plant characteristics (Körner 1994; Schulze et al. 2005). A gradient in vapor pressure between the intercellular spaces and the surrounding air outside the leaf surface determines the rate of water transfer from the leaf into the atmosphere and consequently, transpiration rate: as long as the stomata remain open. Under favorable soil moisture conditions, light intensity controls stomatal opening and stomatal conductance increases with higher light intensities (Schulze et al. 2005). On a daily basis, during ample soil moisture availability, therefore, canopy transpiration increases exponentially with increasing vapor pressure deficit (VPD), as long as the prevailing photosynthetic photon flux density (PPFD) is high enough to allow for full stomatal opening (Granier and Bréda 1996; Oren and Pataki 2001; Ewers et al. 2002). At higher VPD increases in transpiration are, however, regulated by the stomata such that the species-maximum capacity for hydraulic conductivity is not surpassed and cavitation does not occur. For example, saturation of daily stand transpiration for European beech forest stands occurred at mean daily VPD of 2.5–3.0 kPa, while for a spruce stand, transpiration saturated at VPD of 2.0–2.5 kPa (Köstner 2001). Similarly, more pronounced stomatal closure in response to increasing VPD (>2.5 kPa) in European beech (F. sylvatica) than in sessile oak (Q. petraea) has been reported, revealing larger sensitivity of beech to VPD changes (Aranda et al. 2000). Stomatal closure occurs to prevent the development of dangerously low water potentials, which can cause cavitation, and to protect the conducting vessels (Jones and Sutherland 1991). During a fully developed canopy stage in deciduous forests, soil water availability regulates the potential transpiration rates, while PPFD and VPD control diurnal patterns of transpiration water loss (Körner 1994). When soil moisture is limited, leaf water potential declines. Under such conditions, stomatal closure will occur in order to control transpiration water loss (Cochard et al. 1996). In European and North American temperate forests, several studies found that the critical value of relative extractable water from the soil was about 0.4, which is calculated as the ratio of extractable water (available soil water – minimum soil water) and maximum extractable water (soil water content at field capacity – minimum soil water), at which soil water content begins to limit maximum transpiration (Black 1979; Granier 1987; Granier et al. 1999; Wilson and Baldocchi 2000). While environmental drivers control forest stand transpiration in a relatively short period of time, forest structure, i.e., number and size of the trees, age and species composition, regulates stand transpiration over Chapter 1 – Detailed summary ! 5 longer time scales. Transpiration from monocultural forest stands is likely to be dependent on the total sapwood area (AS) of the stand or leaf area index (LAI) (Ewers et al. 2002; Wullschleger et al. 2001). For example, Zimmermann et al. (2000) found that canopy transpiration was correlated with stand AS which was determined by stand density and tree AS in pine forest monoculture stands, with diverse ages ranging from 28 to 383 years of age. Köstner (2001) showed an increasing pattern of the maximum stand transpiration with LAI for five European beech forests, regardless of the stand age. In mixed forest stands, however, differences in species composition may modify rates of transpiration, since species differ in water-resource acquisition, xylem anatomy, and phenology. For example, tree species with deep rooting systems like Q. alba show higher rates of transpiration as the soil dries, compared to shallow rooted ones such as Acer rubrum (Oren and Pataki 2001; Bovard et al. 2005). Oren et al. (1999) showed that the stomatal conductance of ring-porous species was less sensitive to variations in light and VPD than diffuse-porous species. This then resulted in lower mean canopy conductance and lower maximum canopy transpiration for the forest stands composed of a high proportion of ring-porous species than diffuse-porous species (Oren and Pataki 2001). The timing of leaf flushing and senescence also vary among deciduous species (Vitasse et al. 2009), which can have an impact on annual canopy transpiration, since it determines the period of active leaf transpiration. Complexity in terrain, such as along mountain slopes and valleys or exposition increases the complexity in patterns of tree transpiration and the quantities of water used by forest stands. At higher elevations transpiration rates are likely to decrease because of lower air temperatures (Ta), higher humidity (higher rainfall frequencies and amounts) and lower VPD (Kubota et al. 2005; Körner et al. 2007; Kumagai et al. 2008; McDowell et al. 2008; Matyssek et al. 2009). Soil characteristics also change with elevation, with relatively shallow soils found at higher elevations compared to down slope due to increased runoff and erosion at higher elevations and deposition at lower elevations (Hirobe et al. 1998; Tateno et al. 2004; Tromp-van Meerveld and McDonnell 2006). South-facing aspects receive higher solar radiation, which results in warmer and drier conditions on south than north-facing aspects (Van de Water et al. 2002). Spatial variability in microclimate and soil properties at different elevations in complex terrains are, therefore, likely to generate heterogeneous transpiration rates defined by the complex interactions of stand structure, edaphic and the prevailing microclimatic conditions above the forest stands. 1.1.4 Estimation of water use by forests in South Korea Using different approaches, a number of studies have attempted to ascertain water budgets for isolated forests in S. Korea. Kim and Woo (1988) and Lee et al. (1989) reported that direct interception water loss by the forest canopy in the planted coniferous forests was 15–20% higher compared to natural deciduous Chapter 1 – Detailed summary ! 6 forests by measuring throughfall and stemflow under the canopy. In comparison, Eeco was higher in coniferous forests than deciduous forests growing together at similar elevations, based on the calculations using the Thornthwaite method (Kim 1987). Since 2001, the Korean Flux group (KoFlux) has been assessing Eeco in different forest types using the eddy covariance technique in an attempt to obtain an overall water budget for the S. Korean forests (Kim et al. 2006; Kang et al. 2009; Kang et al. 2012). Kang et al. (2009) showed a characteristic seasonality, with mid-season depressions in Eeco that are associated with the reduced amount of available energy during the monsoon season. The application of the eddy covariance technique in the estimation of forest water use in S. Korea is, however, challenging and questions are raised regarding data accuracy, since most of the forests are located in mountainous landscapes that are highly heterogeneous and complex. One of the main assumptions of the eddy technique is a flat and homogeneous fetch for footprint measurement sites (Baldocchi et al. 1988), which is not met in most of these forest stands. Forest transpiration can be measured by sap flow techniques at relatively high temporal scales (Wullschleger et al. 1998). Since they apply at single tree level, the techniques are not limited by terrain complexity and tree species diversity (Wilson et al. 2001; Kumagai et al. 2008). Forest stand transpiration can be determined by summing up the values of transpiration by every single tree in the stand multiplied by its respective AS. A reasonable scaling process from tree to stand level requires accurate estimates of water use from a limited number of representative trees within the stand (Kumagai et al. 2008). The sap flow measurements are, therefore, suitable for estimation of water use by forest stands, the analysis of species effects on forest water use and for partitioning Eeco into transpiration and evaporation (Wilson et al. 2001; Ford et al. 2007). In S. Korea, sap flow measurements have been used to estimate tree transpiration of major species such as Quercus mongolica and Larix leptolepis (Han and Kim 1993; Han and Kim 1996), but not to estimate forest stand transpiration so far. Compared to Europe and North America where significant research in the temperate forests has been conducted, knowledge on the structure and function of the temperate forests in Asia is still lagging behind, both at regional and local scales. Thus, more studies are needed in order to fill the gaps in knowledge, which will allow for a more informed and sustainable forest management. 1.1.5 Statement of research challenges Natural regeneration of forests in South Korea has favored the expansion of the species-rich deciduous forests over conifers, which were massively planted in the previous restoration programs. This shift in forest composition is likely to change the hydrology of most S. Korean forests in a way that is not yet well understood. Climate projections for S. Korea (Im et al. 2008) also show future changes in rainfall patterns and amounts and increases in Ta, which are likely to significantly impact water use by forests. These forests Chapter 1 – Detailed summary ! 13 Table 1.2 Stand structure of the six natural deciduous forest sites in Gyebangsan (GB), Gwangneung (GN), and Haean (HA). LAI, BA, AS and DBH indicate leaf area index, basal area, sapwood area and diameter at breast height (DBH). Trees with diameter at breast height (DBH) ≥5 cm and trees with 2 cm ≤ DBH <5 cm are selected for the overstory (O/S) and the understory (U/S), respectively. Stand density and DBH-based data are interpolated from the inventory conducted in 2008 for GB and GN, and in 2010 for HA. GB GN HA 450N 650N 650S 950N Stand age [yrs] ca. 50 ca. 200 ca. 30 ca. 30 ca. 30 ca. 20 LAI Not measured 4.3 4.9 5.3 6.3 5.7 Tree density [trees ha-1] O/S 1,025 63 1,252 1,640 2,523 4,350 U/S 1,730 1,035 5,165 960 3,969 15,050 BA [m2 ha-1] O/S 24.0 38.5 21.8 17.5 20.5 22.4 U/S Not measured 0.9 2.3 0.8 2.5 1.0 Stand AS [m2 ha-1] O/S 5.7 16.5 12.5 13 15.5 10.5 Average DBH [cm] O/S 13.3 25.7 10.8 10.8 9.4 6.2 Max DBH [cm] O/S 40.7 64.0 28.4 33.5 23.5 9.9 Max tree height [m] 15 20 10 12 10 5 Species composition (BA cover [%]) O/S Tilia amurensis (31) Ulmus davidiana (14) Quercus mongolica (12) Acer mono (8) Acer pseudosieboldianum (5) Maackia amurensis (4) Cornus controversa (4) Quercus serrata (71) Carpinus laxiflora (22) Carpinus cordata (5) Quercus mongolica (24) Alnus sibirica (18) Quercus aliena (15) Quercus serrata (15) Ulmus laciniata (10) Quercus dentate (8) Tilia mandshurica (6) Qurcus dentate (65) Betula davurica (19) Quercus mongolica (14) Quercus mongolica (50) Tilia mandshurica (25) Quercus dentate (14) Fraxinus rhynchophylla (4) Quercus serrata (4) Quercus mongolica (72) Fraxinus rhynchophylla (13) Euonymus hamiltonianus (9) U/S Not measured Euonymus oxyphyllus Celtis jessoensis Sorbus alnifolia Styrax obassia Euonymus alatus Rhododendron yedoense Rhododendron schlippenbachii Qurcus dentate Stephanandra incisa Quercus mongolica Acer pseudosieboldianum Euonymus alatus Acer pseudosieboldianum Staphylea bumalda Chapter 1 – Detailed summary ! 14 Figure 1.2 Climatic charts following Walter and Lieth (1967) for Gyebangsan (GB), Gwangneung (GN), and Haean (HA) forest sites. Chapter 1 – Detailed summary ! 15 1.2.2 Methods 1.2.2.1 Sap flow measurements Sap flow techniques were used to measure transpiration of individual trees. For trees with stem diameter larger than 5 cm, thermal dissipation probes (TDP; Granier 1987) were applied (Figure 1.3a). The TDP is generally accepted as a reliable method for estimating tree transpiration and has been widely used in previous studies to estimate forest water use in various ecosystems (Barbour et al. 2005; Granier and Bréda 1996; Matyssek et al. 2009; Oren and Pataki 2001; Wullschleger et al. 2001; Zeppel et al. 2006). For the understory trees with stem/branch diameters ranging from 9 to 13 mm, stem heat balance (SHB) technique (Sakuratani 1981; 1984 and improved by Weibel and de Vos 1994) was employed (Figure 1.3b). The SHB method has been successfully employed to estimate transpiration of whole saplings (Lei et al. 2010; Weibel and de Vos 1994) and branches of large trees (Otieno et al. 2007). The TDP was applied in all the study sites, while the SHB was used only in the GN site. Sample trees were selected according to species and tree size distribution in the study plots. Characteristics of the sample trees for sap flow measurements such as tree species, DBH, tree height, and sapwood area, are summarized in Table 2.1 (chapter 2), Table 3.1 (chapter 3), and Table 4.2 (chapter 4). Sap flow measurements were carried out in June and July 2008 and repeated over the same period in 2009 at the GB site (chapter 2), April to September in 2008 at the GN site (chapter 3), and from May to October in 2010 at the HA sites (chapter 4). Chapter 1 – Detailed summary ! 16 Figure 1.3 Sap flow methods: (a) A schematic representation of thermal dissipation probe (TDP, Granier 1987) installed on to a tree. The upper probe was heated with a constant current power supply while the lower one (reference) was not heated. (b) A schematic representation of stem heat balance (SHB, Sakuratani 1981; 1984), showing arrangement of the thermocouples around the tree stem. A, B and C represent the respective temperature differences recorded at the logger. Qflow, Qv, Qr and Pin represent convective heat loss by the sap flow, vertical heat conduction, radial heat conduction and heating power, respectively. (c) Installation of TDP in the trees with DBH >5 cm and (d) installation of SHB in the understory trees with stem/branch diameters 9−13 mm. Chapter 1 – Detailed summary ! 17 A modified Jarvis-Stewart model as defined by Whitley et al. (2008) was used to estimate canopy transpiration (EC) for the period when data gaps occurred due to power failure. Whitley et al. (2008; 2009) expressed E in the same way as GC described by Jarvis (1976) and Stewart (1988). EC=ECmax ⋅f1(RS)⋅f2(VPD)⋅f3( θ )⋅f4(LAI) (1) The functions of solar radiation (RS), vapor pressure deficit (VPD), soil water content (θ), and leaf area index (LAI) are a set of scaling terms reducing a maximum EC (ECmax) in response to changes in each one of the variables. Daily EC was determined by the functions using the optimal estimates of parameters. A function describing a radiation response was, f1(RS)=RS 1000 ! " #$ % &⋅1000 +k1 RS+k1 ! " #$ % & (2) where, k1 is an empirical coefficient describing the curvature of the relationship. A function response of EC to VPD was, f2(VPD)=k2⋅VPD⋅exp(−k3⋅VPD) (3) where, k2 and k3 are the parameters describing the rate of changes in VPD. A function describing soil moisture response was expressed as three-phase relationships, f3( θ )= ! " # $ # 0, θ < θ w θ − θ w θ c− θ w , θ w< θ < θ c 1, θ > θ c (4) where, θw is the wilting point and θc is the field capacity. A function of LAI response to EC was, f4(LAI)=LAI LAImax (5) where, LAImax is maximum LAI. Model parameterization was performed using measured data on daily basis via nonlinear least squares analysis using R (R development Core Team, 2009). To avoid errors of division by zero and conditions of wet canopy, daytime data between 8 and 18h was used for the calculations, and the data of the rainy days was excluded. Root-mean-square-error (RMSE) and an agreement index developed by Willmott (1981) were used to evaluate the agreement between the predicted EC and the observed EC (see Table 4.3 in chapter 4). The best estimations of parameters (k1, k2, and k3) are shown in Table 4.3 as well. Total gap-filled periods of each site were 27, 54, 68, and 8 days for 450N, 650N, 650S, and 950N, respectively. Predicted EC was only used to calculate annual EC of each site, but it was not used for calculations of canopy conductance as well as for analyses of the relationship between EC and controlling factors. Chapter 1 – Detailed summary ! 18 1.2.2.2 Canopy conductance Biological regulation of transpiration occurs at the stomata and is measured by stomatal conductance. Canopy conductance (GC) was estimated from estimated stand transpiration derived from sap flow measurements and microclimate (Ta and VPD) according to the Penman−Monteith equation (Granier et al. 1996; Köstner 1992; Oren et al. 1998), GC=( ρ w⋅Gv⋅Tk)⋅EC VPD (6) where, ρw is density of water, Gv is gas conductance of water vapor, Tk is air temperature in kelvin, and EC is canopy transpiration. This simplification of the Penman−Monteith equation is based on the assumption that tree canopies are well coupled to the atmosphere, when leaves are exposed to sufficiently high wind speeds, which results in larger aerodynamic conductance (GA) than GC. Thus, VPD can be used as an approximation of the total driving force for transpiration. The assumption of strong coupling was tested in all the study sites by comparing GA and GC. The results were in agreement with the assumption, i.e., GC is close to 1% of GA, for GB and HA sites, but not for GN, particularly the understory layer. Therefore, this simplified equation was used for the studies in GB (chapter 2) and HA (chapter 4), whereas a separate equation arranged from the Penman−Monteith equation was used for GN site (chapter 3) (Monteith 1965; Herbst et al. 2008). GC= λ ⋅E⋅ γ ⋅GA Δ ⋅ RN+ ρ ⋅cp⋅VPD⋅GA− λ ⋅E⋅(Δ+ γ ) (7) where, λ is the latent heat of vaporization of water, γ is the psychometric constant, Δ is the change of satuaration water vapor pressure with temperature, RN is net radiation, ρ is the density of dry air, and cp is the specific heat of air at constant pressure. Using this equation, canopy conductance for the overstory and the understory were calculated separately. GA of the overstory and the understory were also estimated using the boundary layer conductance (gb) and the turbulent conductance (gt) according to Köstner et al. (1992), Mangnani et al. (1998), and Tateishi et al. (2010) as, GA −1=gb −1+gt −1 (8) gb=b⋅U⋅[1−exp(− α / 2)] dm⋅ α (9) gt= κ 2⋅U LAI ⋅{ln[(z−d) / z0]}2 (10) where, b is the proportionality coefficient, U is the wind speed above the vegetation layers, α is the attenuation coefficient for wind speed inside the canopy, dm is the characteristic dimension calculated as the square root of a leaf area, κ is the von Karman constant, LAI is the leaf area index of the canopy, z is the canopy height, d is the zero-plane displacement, and z0 is the roughness length. To calculate GC based on these two models, data from the measurements between 11h and 17h (daytime) were used. Chapter 1 – Detailed summary ! 19 To assess stomatal sensitivity to VPD, a modified Lohammar’s function was applied as, GC(VPD)=GCref −m⋅ln(VPD) (11) where, GCref is the canopy conductance at VPD = 1 kPa and –m is the sensitivity of GC response to VPD (Oren et al. 1999). This analysis only took into account VPD larger than 0.6 kPa to keep the errors in GC estimates under 10% (Ewers and Oren 2000). GCref was also used to compare the capacity of water use response to VPD among different tree species or forest stands (Oren et al. 1999). 1.2.2.3 Biometric measurements To examine seasonal changes of the forest cover and quantify the maximum leaf area of the forests, we measured plant area index (PAI) and maximum leaf area index (LAI) at GN and HA sites. PAI was measured every month during the vegetative period using a plant canopy analyzer (LAI-2000, LI-COR Inc., Lincoln, USA) under diffuse light conditions at fixed 6−12 sampling points. The maximum LAI for the forest sites in HA was estimated from leaf litter collected with 1 m high, 0.5 m × 0.5 m litter traps randomly placed at five points above the forest floor. The litter was transferred to the laboratory, sorted according to species, dried for 48 hours at 75°C and weighed. Specific leaf area (SLA) was determined from the ratio of single leaf area and dry mass for each species. Total leaf area of each species from each site was computed from total leaf dry weight over the season multiplied by SLA and divided by the area of the litter trap. The maximum understory LAI (LAIU) for GN site was measured by collecting leaf samples of all the understory trees in three plots with 2 m × 2 m size in early July when LAIU was at the peak. The total area of the leaf samples was measured using a leaf area meter (LI-3100, LI-COR Inc.). We performed vegetation surveys of all tree stems larger than 5 cm in DBH in a 40 m × 50 m plot at the GB site, in a 30 m × 30 m plot at the GN, and in 25 m × 25 m plots at the HA sites. Based on the survey, mean DBH, basal area (BA), tree density, and species composition were determined for each plot (see Table 1.2). To calculate sapwood area (AS) of the sample trees, we measured bark and sapwood depths from the extracted cores taken from the stems at sap flow sensor height at the end of the measurements using an increment corer. Sapwood was identified by dying the core samples using bromocresol green (Sigma Chemicals, Germany) (Burrows 1980). AS was calculated from the measured DBH and depths of bark and sapwood. Based on the calculated AS and measured DBH, an allometric function was established (Vertessy et al. 1995, Meinzer et al. 2005) as, AS= α DBH β (12) where, α is a constant and β is the allometric scaling exponent. Allometric relationships were made for each one of the studied species as well as for all the species together. Since the differences in the estimations of stand AS (total AS of all trees in the plot) from separate regressions and from a combined species regression Chapter 1 – Detailed summary ! 20 were less than 10% at all sites, we chose the general regression for further analyses related to AS. For example, the values of stand AS determined by the general regression were used for the estimations of canopy transpiration of each site. 1.2.2.4 Micrometeorological measurements Air temperature (Ta), precipitation, net radiation (RN) or solar radiation (RS), photosynthetic active radiation (PAR), relative humidity or water vapor density, and wind speed (U) were measured with a 20 m tower in GB, with a 40 m tower in GN, and at 2 m above the ground in the open space next to the forest sites in HA. Microclimates below the canopy of Ta, PAR, and humidity were measured at 5 m in GB, at 4 m (but at 2 m for PAR) in GN, and at 2 m in HA sites. VPD was derived from air temperature and relative humidity (Murray 1967). In the GN site, U in the understory at a height of 4 m was additionally measured. Soil water content (θ) and temperature were measured at 30 cm depth in all the sites. Data were read every 30 seconds, averaged, and stored every 30 minutes using data loggers. Soil water retention was determined from measured θ and soil characteristics (i.e., soil texture and bulk density). 1.2.2.5 Eddy covariance flux measurements At the GN site, evapotranspiration (Eeco) was measured with an eddy covariance system installed on a 40 m high tower in GN site for the same period as sap flow measurements (chapter 3). Data quality was controlled using the standardized KoFlux protocol including planar fit rotation, Webb-Pearman-Leuning correction, spike detection, and gap filling (Hong et al. 2009). Data gaps were filled using a modified lookup table following the method proposed by FLUXNET (Reichstein et al. 2005) and modified by Kang et al. (2012). 1.2.2.6 Water use efficiency From each site, five sunand five shade-leaves each from five Q. mongolica canopy trees were collected on 24 June 2010, during mid season. The samples were oven-dried at 75°C for 48 hours and then ball-milled before subjected to 13C/12C isotopic ratio analysis at BayCEER – Laboratory of Isotope Biogeochemistry, Germany. Analyses were conducted with an elemental analyzer NA 1108 (CE Instruments, Milan, Italy) coupled to an isotope ratio mass spectrometer delta S (Finnigan MAT, Bremen, Germany) via an open split interface ConFlo III (Finnigan MAT, Brement, Germany) as described by Bidartondo et al. (2004). Standard CO2 gas was calibrated with respect to international standard (CO2 in Pee Dee Belemnite) by use of the reference substance NBS 16 to 20 for carbon isotopic ratio provided by the international Atomic Energy Chapter 1 – Detailed summary ! 21 Agency IAEA, Vienna, Austria. The 13C/12C isotopic ratios, denoted as delta values were calculated according to the equation δ 13C=Rsample Rstd −1 " # $% & '×1000 (13) where δ13C is the isotope ratio of carbon in delta units relative to the PDB standard. Rsample and Rstd are the 13C/12C of the samples and the PDB standard, respectively. δ13C was used as an index of seasonally integrated water use efficiency (WUE) (Tieszman and Archer 1990). 1.3 General results and discussions 1.3.1 Tree and forest water use in diverse species composition An allometric equation was used to find a general relationship between DBH and AS for different species co-occurring in each study site. A significant relationship, AS = αDBHβ, linking AS to DBH, was established (see Figure 2.5 in chapter 2, Figure 3.1 in chapter 3, and Figure 4.2 in chapter 4). These regression models were applied to calculate AS for all the trees in each study plot. To examine the applicability of this general relationship across the sites, we combined the data of 13 deciduous tree species growing in different study sites. We found that all the species, regardless of their location, significantly (R2 = 0.77, P <0.0001) fitted into a single power curve (Figure 1.4a). This indicates that the general regression model established in this study can be applicable to similar forests in S. Korea, outside our study sites located in other places. To reduce complexity in estimation of transpiration in mixed deciduous forests, a similar analysis relating TWU and DBH was performed. Maximum daily TWU was significantly related to DBH in a power function for the species co-occurring at each study site (P <0.001 for all sites). To check if this relationship is applicable across the studied sites, the 12 tree species measured in the three different study sites were combined in the analysis. A significant (R2 = 0.77, P <0.0001) single power function was established for the combined data (Figure 1.4b), although there was a tendency of higher TWU for the trees growing in the HA site, which can be explained by greater values of maximum Fd, compared to the others growing in GB and GN sites (Figure 1.5). This result was consistent with the previous studies conducted in temperate forest in North America (Wullschleger et al. 2001), tropical rain forest (Meinzer et al. 2005), and open woodland in Australia (Zeppel and Eamus 2008). Our findings support the hypothesis that different tree species growing together in common locations have converging function in water use determined by tree sizes (Kallarakal et al. 2013). As long as soil water is not limiting, transient changes in tree transpiration are due to the prevailing microclimate, while the potential maximum transpiration is determined by the xylem transport capacity (Oren and Pataki 2001). This implies that the same power scaling parameters, determining AS and Chapter 1 – Detailed summary ! 22 TWU of single trees by the DBH sizes, can be applied to different tree species growing in deciduous forests in S. Korea as long as trees are exposed to similar environments. Thus, our findings can be applied to estimate maximum daily TWU and stand transpiration using a simple empirical allometry model and DBH of the target sites, irrespective of the species. The power regressions and the statistics of these analyses are shown in Table 1.3 and 1.4. To identify the most suitable function explaining the relationship between TWU and DBH, the same data set was fitted to the three-parameter sigmoid function (Meinzer et al. 2005) (Figure 1.4b). We found similar goodness of fit between a sigmoid and power function (see AIC values in Table 1.4). Meinzer et al. (2005) found superior fitting from a sigmoid function than power function for angiosperm trees, which was supported by the evidences that transpiration and photosynthesis are limited as trees grow above a threshold size (McDowell et al. 2002, Niinemets 2002). Based on our study, however, both power and sigmoid functions can be used for the estimation of TWU from DBH. Unlike AS and TWU, there was no significant relationship between maximum Fd and DBH (Figure 1.5). This is consistent with the study conducted by Phillips et al. (1999). Meinzer et al. (2001), however, obtained a negative correlation, while Oren et al. (1998) found positive correlation between maximum Fd and DBH suggesting that DBH can be a good predictor of Fd among species from diverse locations. Several studies have also considered wood density as an alternative predictor not only for Fd but also for hydraulic conductivity and stomatal conductance (Bucci et al. 2004; O’Grady et al. 2009; Kallarackal et al. 2013). This is an alternative approach that we recommend in future studies, since wood density combines both the tree age and growth conditions. Chapter 1 – Detailed summary ! 29 Figure 1.7 Cumulative canopy transpiration (EC) from Gwangneung (GN) and Haean (HA) sites. Figure 1.8 Relationships between elevation and (a) annual mean Ta (b) annual daytime mean VPD, (c) length of growing season, and (d) total canopy transpiration (EC) for the Gwangneug (GN) and Haean (HA) sites. Chapter 1 – Detailed summary ! 30 1.4 General conclusions This study investigated the challenges imposed by multiple tree species, multiple canopy layering and complex terrain when attempting to determine water use by warm-temperate forest species in S. Korea. Our main findings and suggestions are summarized below:  DBH was correlated with AS (in a power function) and maximum TWU (both in power and sigmoid functions) not only for the tree species co-occurring in a single forest stand but also growing in differently aged and structured deciduous forests.  This functional relationship between DBH and TWU provides a relatively simple but accurate approach for the prediction of water use by trees and forest stands in mixed deciduous forests, thus reducing the complexity arising from multiple tree species.  Temporal differences in the overstory and understory developments, with earlier bud break and leaf expansion of the understory trees, altered the microclimate in the understory and hence the rate of EU before and after the overstory canopy maturity.  The seasonal pattern of both EO and EU are regulated by canopy development and microclimate, primarily VPD and PAR, which determine their daily trend.  EU significantly contributed to total forest transpiration during the whole growing season, with the highest contributions in April and May, which resulted in relatively high total transpiration early in the season.  Since Eeco is strongly influenced by EO and EU, both the overstory and understory should be considered when estimating Eeco for forests consisting of a well-developed understory layer.  High stomatal sensitivity to VPD of the forest located at higher elevation resulted in early saturation and midday depression of EC at relatively low daytime VPD and possibly high WUE, although the main driver of the shift in stomatal sensitivity is still unclear.  Total EC during the growing season decreased with increasing elevation, corresponding to the decrease in Ta, daytime VPD, and length of growing season at the higher elevation. These variables should be carefully taken into account for the estimation of forest water use in mountainous regions in complex terrains. Our study addressed the challenges involved in estimating water use by the warm-temperate deciduous forests in S. Korea, comprised of diverse tree species, multi-layered canopy, and rugged mountainous terrains. We observed that the complexity arising from multiple tree species can be solved using a functional relationship between DBH and TWU. We also found that for forests with a multi-layered canopy, both EO and EU should be taken into consideration when estimating Eeco. Finally, in complex terrains, elevation alone Chapter 1 – Detailed summary ! 31 does not determine forest water use, but works in tandem with microclimate and plant growth characteristics, factors that need to be determined at relatively small spatial scales since they change rapidly. 1.5 List of manuscripts and specification of contributions This dissertation includes three manuscripts. The first manuscript is published in Plant Ecology and the second manuscript is published in Journal of Plant Research. The third manuscript is submitted to Plant and Soil and is currently ‘under review’. Specific contributions by the co-authors of each manuscript are listed below. Manuscript 1 (Chapter 2) Authors EY Jung, D Otieno, B Lee, JH Lim, SK Kang, MWT Schmidt, J Tenhunen Title Up-scaling to stand transpiration of an Asian temperate mixed-deciduous forest from single tree sap flow measurements Status Published in Plant Ecology (2011) 212:383-395 Contributions EY Jung: concepts, field work, discussion and presentation of results, manuscript preparation, corresponding author D Otieno: concepts, field work, discussion of results, manuscript editing B Lee: field work JH Lim: field work, logistics in Korea SK Kang: logistics in Korea MWT Schmidt: discussion of results J Tenhunen: concepts, manuscript editing Manuscript 2 (Chapter 3) Authors EY Jung, D Otieno, H Kwon, B Lee, JH Lim, J Kim, J Tenhunen Title Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: contributions of the overstory and understory to forest water use Status Published in the Journal of Plant Research (2013) DOU 10.1007/s10265-013-0563-5 Contributions EY Jung: concepts, field work, presentation and discussion of results, manuscript preparation, corresponding author D Otieno: concepts, field work, discussion of results, manuscript editing H Kwon: field work, discussion of results, manuscript editing B Lee: field work Chapter 1 – Detailed summary ! 32 JH Lim: field work, logistics in Korea J Kim: logistics in Korea J Tenhunen: concepts, manuscript editing Manuscript 3 (Chapter 4) Authors EY Jung, D Otieno, H Kwon, S Berger, M Hauer, J Tenhunen Title Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea Status Submitted to Plant and Soil; current status ‘in revision’ Contributions EY Jung: concepts, field work, presentation and discussion of results, manuscript preparation, corresponding author D Otieno: concepts, field work, discussion of results, manuscript editing H Kwon: discussion of results, manuscript editing S Berger and M Hauer: field work J Tenhunen: concepts, logistics in Korea 1.6 References Allaby M (2006) Biomes of the Earth: Temperate Forests. New York, NY: Chelsea House Aranda I, Gil L, Pardos JA (2000) Water relations and gas exchange in Fagus sylvatica L. and Quercus petraea (Mattuschka) Liebl. In a mixed stand at their southern limit of distribution in Europe. Trees 14:344–352 Baldocchi DD, Hicks BB, Meyers TP (1988) Measuring biosphere-atmosphere exchanges of biologically related gases with micrometeorological methods. Ecology 69:1331–1340 Baldocchi DD, Vogel C (1996) A comparative study of water vapor, energy and CO2 flux densities above and below a temperate broadleaf and a boreal pine forest. Tree Physiol 16:5–16 Baldocchi DD, Wilson KB, Gu L (2002) How the environment, canopy structure and canopy physical functioning influence carbon, water and energy fluxes of a temperate broad-leaved deciduous forest– an assessment with the biophysical model CANOAK. Tree Physiol 22:1065–1077 Barbier S, Gosselin F, Balandier P (2008) Influence of tree species on understory vegetation diversity and mechanisms involved – A critical review for temperate and boreal forests. For Ecol Manage 254:1–15 Barbour MM, Hunt JE, Walcroft AS, Rogers GND, McSeveny TM, Whitehead D (2005) Components of ecosystem evaporation in a temperate coniferous rainforest, with canopy transpiration scaled using sapwood density. New Phytol 165:549–558 Chapter 1 – Detailed summary ! 33 Berbigier P, Diawara A, Loustau D (1991) A microclimatic study of the effect of drought on evapotranspiration in a maritime pine stand and its understory. Ann Sci For 48:157–177 Bidartondo MI, Burghardt B, Gebauer G, Bruns TD, Read DJ (2004) Changing partners in the dark: isotopic and molecular evidence of ectomycorrhizal liaisons between forest orchids and trees. Proc R Soc Lond B 271:1799−1806 Black TA (1979) Evapotranspiration from Douglas fir stands exposed to soil water deficits. Water Resour Res 15:164–170 Bovard BD, Curtis PS, Vogel CS, Su H-B, Schimid HP (2005) Environmental controls on sap flow in a northern hardwood forest. Tree Physiol 25:31–38 Bucci SJ, Goldstein G, Meinzer FC, Scholz FG, Franco AC, Bustamante M (2004) Functional convergence in hydraulic architecture and water relations of tropical savanna trees: from leaf to whole plant. Tree Physiol 24:891−899 Burrows LE (1980) Differentiating sapwood, hearwood and pathological wood in live mountain beech. New Zealand Forest Service, Forest Research Institute, Protection Forestry Report 172 Campbell-Clause M (1998) Stomatal response of grapevines to wind. Aust J Exp Agric 38:77−82 Canham DC, Finzi AC, Pacala SW, Burbank DH (1994) Causes and consequences of resource heterogeneity in forests: interspecific variation in light transmission by canopy trees. Can J For Res 24:337−349 Chapin FS, Matson PA, Vitousek PM (2011) Principles of Terrestrial Ecosystem Ecology. Second edition. Springer, New York Chung TH, Lee WC (1965) A study of the Korean woody plant zone and favorable region for the growth and proper species. Thesis Collection of Sungkyunkwan University 10:329−435 Cochard H, Bréda N, Granier A (1996) Whole tree hydraulic conductance and water loss regulation in Quercus during drought: evidence for stomatal control of embolism? Ann Sci For 53:197−206 Cordell S, Goldstein G, Mueller-Dombois D, Webb D, Vitousek PM (1998) Physiological and morphological variation in Metrosideros polymorpha, a dominant Hawaiian tree species, along an altitudinal gradient: the role of phenotypic plasticity. Oecologia 113:188-196 Dittmar C, Elling W (2006) Phenological phases of common beech (Fagus sylvatica L.) and their dependence on region and altitude in Southern Germany. Eur J For Res 125:181−188 Ehleringer JR (1993) Carbon and water relations in desert plants: an isotopic perspective. In Ehleringer JR, Hall AE, Farquhar GD (Eds.), Stable isotopes and plant carbon-water relations, Academic Press, San Diego, California, USA Ellenberg H (1978) Zur Populationsökologie des Rehes (Capreolus capreolus L., Cervidae) in Mitteleuropa. Spixiana 2, München Ewers BE, Oren R (2000) Analyses of assumptions and errors in the calculation of stomatal conductance from sap flux measurements. Tree Physiol 20:579−589 Chapter 1 – Detailed summary ! 34 Ewers BE, Mackay DS, Gowers ST, Ahl DE, Burrows SN, Samanta SS (2002) Tree species effects on stand transpiration in northern Wisconsin. Water Resources Res 38:1103 Ford GR, Hubbard RM, Koeppel BD, Vose JM (2007) A comparison of sap flux-based evapotranspiration estimates with catchment-scale water balance. Agric For Meteorol 145:176−185 Gong WS (2001) Temporal and spatial changes in the distribution of Phyllostachys habitation (in Korean). Kor Geograph Soc 36:444−457 Granier A (1987) Evaluation of transpiration in a Douglas-fir stand by means of sap flow measurements. Tree Physiol 3:309−320 Granier A, Bréda N (1996) Modelling canopy conductance and stand transpiration of an oak forest from sap flow measurements. Ann Sci For 53:537−546 Granier A, Huc R, Barigah ST (1996) Transpiration of natural rain forest and its dependence on climatic factors. Agric For Meteorol 78:19−29 Granier A, Bréda N, Biron P, Villette S (1999) A lumped water balance model to evaluate duration and intensity of drough constraints in forest stands. Ecol Mod 116:269−283 Han SS, Kim SH (1993) Ecophysiological interpretations on the water relations parameters of trees (VII) − Measurement of water flow by the heat pulse method in a Larix leptolepis stand. J Kor For Soc 82: 152–165 Han SS, Kim SH (1996) Ecophysiological interpretations on the water relations parameters of trees (IX) − Measurement of the transpiration rate by the heat pulse method in a Quercus mongolica stand. J Kor For Soc 85: 288–299 Hatton TJ, Ruprecht J, George RJ (2003) Preclearing hydrology of the Western Australia wheatbelt: Target for the future. Plant Soil 257:341–356 Herbst M, Rosier PTW, Morecroft MD, Gowing DJ (2008) Comparative measurements of transpiration and canopy conductance in two mixed deciduous woodlands differing in structure and species composition. Tree Physiol 28:959−970 Hirobe H, Naoko T, Iwatsubo G (1998) Spatial variability of soil nitrogen transformation patterns along a forest slope in a Cryptomeria japonica D. Don plantation. Eur J Soil Biol 34:123−131 Hoch G, Körner C (2005) Growth, demography and carbon relations of Polylepis trees at the world’s highest tree line. Funct Ecol 19:941−951 Hong J, Kwon H, Lim JH, Byun YH, Lee J, Kim J (2009) Standardization of KoFlux eddy-covariance data processing. Kor J Agric For Meteorol 2:12−26 Hubick KT, Farquhar GD, Shorter R (1986) Correlation between water-use efficiency and carbon isotope discrimination in diverse peanut (Arachis) germplasm. Australian J Plant Physiol 13:803−816 Chapter 1 – Detailed summary ! 35 Im E-S, Ahn J-B, Kwon W-T, Giorgi F (2008) Multi-decadal scenario simulation over Korea using a onway double-nested regional climate model system. Part 2: future climate projection (2021−2050). Clim Dym 30:239−254 Ito A, Saigusa N, Murayama S, Yamamoto S (2005) Modeling of gross and net carbon dioxide exchange over a cool-temperate deciduous broad-leaved forest in Japan: Analysis of seasonal and interannual change. Agric For Meteorol 134:122−134 IPCC (2000) Special report on emissions scenarios. A Special report of working group III of the intergovernmental panel on climate change (Nakicenovic N and lead authors). Cambridge University Press, Cambridge Jarvis PG (1976) The interpretation of the variations in leaf water potential and stomatal conductance found in canopies in the field. Phil Trans Roy Soc Lond B 273:593−610 Jones HG, Sutherland RA (1991) Stomatal control of xylem embolism. Plant Cell Environ 14: 607−612 Kallarackal J, Otieno DO, Reineking B, Jung E, Schmidt MWT, Granier A, Tenhunen JD (2013) Functional convergence in water use of trees from different geographical regions: a meta-analysis. Trees 27: 787−799 Kang M, Park S, Kwon H, Choi HT, Choi YJ, Kim J (2009) Evapotranspiration from a deciduous forest in a complex terrain and a heterogeneous farmland under monsoon climate. Asia-Pacific J Atmos Sci 45:175−191 Kang M, Kwon H, Cheon JH, Kim J (2012) On estimating wet canopy evaporation from deciduous and coniferous forests from Asian monsoon climate. J Hydrometeorl 13:950−965 Kelliher FM, Leuning R, Schulze ED (1993) Evaporation and canopy characteristic of coniferous forests and grasslands. Oecologia 95:153−163 Kim JS (1987) Forest effects on the flood discharges and the estimation of the evapotranspiration in the small watershed. Korea Forest Research Institute Reports, Vol.35 Kim JH (2002) Community ecological view of the natural deciduous forest in Korea. In: Ecology of Korea, Bumwoo, Seoul Kim KH, Woo BM (1988) Study on rainfall interception loss from canopy in forest (I). Journal of Korean Forestry Society, Vol.77, No.3 Kim J, Lee D, Hong J, Kang S, Kim S-J, Moon S-K, Lim J-H, Son Y, Lee J, Kim S, Woo N, Kim K, Lee B, Lee B-L, Kim S (2006) HydroKorea and CarboKorea: cross-scale studies of ecohydrology and biogeochemistry in a heterogeneous and complex forest catchment of Korea. Ecol Res 21:881−889 Kira T (1991) Forest ecosystems of east and southeast Asia in a global perspective. Ecological Research 6:185−200 Korea Forest Service (2006) Statistical Yearbook of Forestry. Seoul, KFS Chapter 1 – Detailed summary ! 36 Korea Forest Service (2009) National Report on Sustainable Forest Management in Korea, 2009. Daejeon, KFS Korea Forest Service (2011) 2010 Basic Statistics of Forest. Daejeon, KFS Körner C, Farquhar GD, Roksandik Z (1988) A global survey of carbon isotope discrimination in plants from high altitude. Oecologia 74:623−632 Körner C, Farquhar GD, Wong SC (1991) Carbon isotope discrimination by plants follows latitudinal and altitudinal trends. Oecologia 88:30−40 Körner C (1994) Leaf diffusive conductances in the major vegetation types of the globe. In: Schulse E-D, Caldwell MM (eds.) Ecophysiology of Photosynthesis. Springer-Verlag, Heidelberg Körner C (2003) Alpine Plant Life (2nd ed.), Springer-Verlag, Berlin Heidelberg New York Körner C (2007) The use of ‘altitude’ in ecological research. Trends Ecol Evol 22:569–574 Köstner BMM, Schulze E-D, Kelliher FM, Hollinger DY, Byers JN, Hunt JE, McSeveny TM, Meserth R, Weir PL (1992) Transpiration and canopy conductance in a pristine broad-leaved forest of Nothofagus: an analysis of xylem sap flow and eddy correlation measurements. Oecologia 91:350−359 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 Kubota M, Tenhunen J, Zimmermann R, Schmid M, Adiku S, Kakubari Y (2005) Influences of environmental factors on the radial profile of sap flux density in Fagus crenata growing at different elevations in the Naeba Mountains, Japan. Tree Physiol 25:545−556 Kumagai T, Tateishi M, Shimizu T, Otsuki K (2008) Transpiration and canopy conductance at two slope positions in a Japanese cedar forest watershed. Agric For Meteorol 148:1444−1455 Kwon W-T (2005) Current status and perspectives of climate change sciences (in Korean with English abstract). J Kor Meteorol Soc 41:325−336 Latham RE, Ricklefs RE (1993) Global patterns of tree species richness in moist forests: energy-diversity theory does not account for variation in species richness. Oikos 67:325−333 Lauscher F (1976) Weltweite Typen der Höhennabhängigkeit des Niederschlags. Wetter Leben 28:80−90 Lee JH, Kim TH, Lee WK, Choi K, Lee CY, Joo JS (1989) A study on regulating discharges by forest. Korea Forest Research Institute Reports, Vol.38, Seoul Lee YH, Park TS, Youn YC, Ryu TK, Lim YJ, Choi JC, Byun WH, Kim EG (1997) History of Korean Forestry Policy. Korea Forest Service, Seoul Lee IK, Lim JH, Kim C, Kim YK (2006) Nutrient dynamics in decomposing leaf litter and litter production at the long-term ecological research site I Mt. Gyebangsan (in Korean with English abstract). J Ecol Field Biol 29:585−591 Chapter 1 – Detailed summary ! 37 Lei H, Zhi-Shan Z, Xin-Rong L (2010) Sap flow of Artemisia ordosica and the influence of environmental factors in a revegetated desert area: Tengger Desert, China. Hydrol Process 24:1248−1253 Lim JH, Chun JH, Woo SY, Kim YK (2008) Increased declines of Korean fir forest caused by climate change in Mountain Halla, Korea. In: Oral Presentation At: International Conference “Adaptation of Forests and Forest Management to Changing Climate with Emphasis on Forest Health: A Review of Science, Policies, and Practices”, Umeá, Sweden, FAO/IUFRO, 25–28 August 2008 Magnani F, Leonardi S, Tognetti R, Grace J, Borghetti M (1998) Modeling the surface conductance of a broad-leaf canopy: effects of partial decoupling from the atmosphere. Plant Cell Environ 21:867−879 Matyssek R, Wieser G, Patzner K, Blaschke H, Häberle K-H (2009) Transpiration of forest trees and stands at different altitude: consistencies rather than contrasts? Eur J For Res 128−579596 McDowell NG, Phillips N, Lunch C, Bond BJ, Ryan MG (2002) An investigation of hydraulic limitation and compensation in large, old Douglas-fir trees. Tree Physiol 22:763−774 McDowell NG, White S, Pockman WT (2008) Transpiration and stomatal conductance across a steep climate gradient in the southern Rocky Mountains. Ecohydrol 1:193−204 Meinzer FC, Goldstein G, Andrade JL (2001) Regulation of water flux through tropical forest canopy trees: do universal rules apply? Tree Physiol 21:19−26 Meinzer FC, Bond BJ, Warren JM, Woodruff DR (2005) Does water transport scale universally with tree size? Funct Ecol 19:558−565 Miyajima Y, Takahashi K (2007) Changes with altitude of the stand structure of temperate forests on Mount Norikura, central Japan. J For Res 12:187−192 Monteith JL (1965) Evaporation and environment. Symp Soc Exp Biol XIX:205−234 Monteith JL, Unsworth MH (1990) Principles of environmental physics. Edward Arnold, London Muraoka H, Koizumi H (2005) Photosynthetic and structural characteristics of canopy and shrub trees in a cool-temperate deciduous broadleaved forest: Implication to the ecosystem carbon gain. Agric For Meteorol 134:39−59 Murray FW (1967) On the computation of saturation vapor pressure. J Appl Meteorol 6:203−204 Nakashizuka T (1987) Regeneration dynamics of beech forests in Japan. Vegetatio 69:169−175 Nakashizuka T, Iida S (1995) Composition, dynamics and disturbance regime of temperate deciduous forests in Monsoon Asia. Vegetatio 121:23−30 Nasahara KN, Muraoka H, Nagai S, Mikami H (2008) Vertical integration of leaf area index in Japanese deciduous broad-leaved forest. Agric For Meteorol 148:1136−1146 National Institute of Environental Research (2011) Korea Climate Change Valuation 2010, Incheon, NIER Niinemets U (2002) Stomatal conductance alone does not explain the decline in foliar photosynthetic rates with increasing tree age and size in Picea abies and Pinus sylvestris. Tree Physiol 22:515−535 Chapter 1 – Detailed summary ! 38 O’Grady AP, Cook PG, Eamus D, Duguid A, Wischusen JDH, Fass T, Worldege D (2009) Convergence of tree water use within an arid-zone woodland. Oecologia 160:643−655 Oh J-H, Kim T, Kim M-K, Lee S-H, Min S-K, Kwon W-T (2004) Regional climate simulation for Korea using dynamic downscaling and statistical adjustment. J Meteorol Soc Jpn 82:1629−1643 Oren R, Phillips N, Katul G, Ewers BE, Pataki DE (1998) Scaling xylem sap flux and soil water balance and calculating variance: a method for partitioning water flux in forests. Ann Sci For 55:191−216 Oren R, Phillips N, Ewers BE, Pataki DE, Megonigal JP (1999) Sap-flux-scaled transpiration responses to light, vapor pressure deficit, and leaf area reduction in a flooded Taxodium distichum forest. Tree Physiol 19:337−347 Oren R, Pataki DE (2001) Transpiration in response to variation in microclimate and soil moisture in southeastern deciduous forests. Oecologia 127:549−559 Otieno DO, Schmidt MWT, Kurz-Besson C, Lobo Do Vale R, Pereira JS, Tenhunen JD (2007) Regulation of transpirational water loss in Quercus suber trees in a Mediterranean-type ecosystem. Tree Physiol 27:1179−1187 Park J-Y, Park M-J, Ahn S-R, Park G-A, Yi J-E, Kim G-S, Sinivasan R, Kim SJ (2011) Assessment of future climate change impacts on water quantity and quality for a mountainous dam watershed using SWAT. Trans ASABE 54:1725−1737 Phillips N, Oren R, Zimmermann R, Wright SJ (1999) Temporal patterns of water flux in trees and lianas in a Panamanian moist forest. Tree 14:116−123 Qian H, Ricklefs RE (1999) A comparison of the taxonomic richness of vascular plants in China and the United States. Am Nat 154:160−181 Qian H, Ricklefs RE (2000) Large-scale processes and the Asian bias in species diversity of temperate plants. Nature 407:180−182 R Development Core Team (2009) R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing Reichstein M et al. (2005) On the separation of net ecosystem exchange into assimilation and ecosystem respiration: revies and improved algorithm. Glob Change Biol 11:1424−1439 Richardson AD, Bailey AS, Denny EG, Martin CW, O’Keefe J (2006) Phenology of a northern hardwood forest canopy. Glob Change Bio 12:1174−1188 Röhrig E (1991a) Climatic conditions. pp. 7−16. In: Röhrig E and Ulrich B (eds.) Temperate Deciduous Forests. Elsevier, Amsterdam Röhrig E (1991b) Floral composition and its evolutionary development. pp. 17−33. In: Röhrig E and Ulrich B (eds.) Temperate Deciduous Forests. Elsevier, Amsterdam Sakuratani T (1981) A heat balance method for measuring water flux in the stem of intact plants. J Agric Meteorol 37:9–17 Chapter 2 – Up-scaling to stand transpiration of an Asian temperate mixed-deciduous forest from single tree sap flow measurements ! 45 (UA-002-08, Onset, USA) at 1 m (below the canopy) and at 20 m height (above the canopy), photosynthetic active radiation (PAR) (LI-190, LI-COR, USA) at 20 m, air temperature and humidity (HMP35C, Cambell Scientific Inc., USA) at 5, 10, and 20 m height below, within and above the canopy, and soil water content and temperature (5TE, Decagon Devices, USA) at −5, −15 and −30 cm. These parameters were measured continuously during the experimental period. Data were averaged and logged every 30 min, either with loggers built into the sensors (light intensity) or a central data logger (DL2e, Delta-T Devices, UK). Vapor pressure deficit (VPD) was calculated from air temperature and relative humidity. Precipitation data were obtained from an automated weather station built by the Korea meteorological administration located 5.2 km away from our study site. 2.2.4 Tree allometrics The sample trees were selected according to species and tree size distribution in the plot (Figure 2.2). T. amurensis was the dominant species in the plot, occupying 31.5% of the total basal area. U. davidiana, Q. mongolica, A. mono, and C. controversa were co-dominant and occupied 13.6, 10.7, 8.1, and 4.4% of the total basal area of the plot, respectively. Thus, these dominant and co-dominant species covered almost 70% of the total basal area of the study plot. We chose five trees each from Q. mongolica, T. amurensis, and U. davidiana and three trees each from C. controversa and A. mono, respectively, for the sap flow measurements. DBH ranges of 13.2−38.2 cm were considered (Table 2.1). Figure 2.2 (a) Percentage of basal area of sample tree species in the study plot. The basal area (BA) for the five measured species occupied 70% of all trees in the plot. (b) Study site with specific location and relative size of trees. Total plot size was 2,000 m2. Chapter 2 – Up-scaling to stand transpiration of an Asian temperate mixed-deciduous forest from single tree sap flow measurements ! 46 Table 2.1 Studied sample trees and respective diameter of breast height (DBH), tree height, sapwood depth, canopy area, mean sap flux density (Fd), and maximum Fd at Mt. Gyebangsan, June and July in 2008 and 2009. DBH was sapwood depth was estimated by empirical regression models of DBH (see Figure 2.5). Projected canopy area was measured in late fall, 2008 and sap flux density (Fd) was measured during June 2008 and 2009 a Trees with a sapwood depth larger than 20 mm had two sensors in different depth Species Sample trees DBH (cm) Tree height (m) Sapwood depth (cm) Canopy area (m2) Mean Fd (g m−2 s−1) Max Fd (g m−2 s−1) Q. mongolica Q1a 27.6 14 2.9 16.0 21.9 ± 3.4 27.7 Q2a 28.4 14 3.1 37.8 15.9 ± 3.9 23.8 Q3 20.3 13 2.0 19.8 21.8 ± 3.9 30.8 Q4 13.3 12 1.4 9.1 4.3 ± 2.6 11.7 Q5a 38.2 15 4.9 46.5 24.4 ± 4.0 34.6 T. amurensis T1 29.2 17 1.5 18.8 38.9 ± 9.3 54.5 T2 18.9 14 0.8 9.0 7.7 ± 2.8 13.5 T3 26.8 15 1.3 14.7 38.7 ± 8.9 53.2 T4 13.2 12 0.6 16.8 21.4 ± 7.8 33.9 T5 17.8 13 0.8 18.5 18.6 ± 7.7 32.7 U. davidiana U1 23.1 15 1.6 11.2 22.9 ± 5.4 32.0 U2 23.4 14 1.7 37.2 14.3 ± 2.4 18.3 U3 28.6 16 2.2 22.7 23.3 ± 3.9 30.4 U4 26.1 15 1.9 37.1 21.3 ± 4.9 30.1 U5 18.7 15 1.3 18.5 28.9 ± 4.1 35.4 C. controversa C1 25.3 15 1.2 62.5 34.0 ± 7.6 48.8 C2 22.3 15 1.0 41.9 33.4 ± 7.0 49.3 C3 17.6 15 0.7 28.8 15.7 ± 5.7 27.9 A. mono A1 15.0 13 0.3 13.3 15.7 ± 5.7 27.0 A2 22.0 14 0.5 38.7 30.7 ± 12.4 50.2 A3 13.8 11 0.3 12.1 31.9 ± 13.9 57.8 Chapter 2 – Up-scaling to stand transpiration of an Asian temperate mixed-deciduous forest from single tree sap flow measurements ! 47 To estimate the sapwood area (AS) of the sample trees, an increment borer was used to extract cores of sapwood at the sensor installation height (about 1.3 m height) on same species, but different trees from those installed with the sap flow sensors. Sapwood depth was determined visually on those cores since sapwood and heartwood were clearly distinguishable. AS was determined from sapwood depth and tree DBH based on the equation (Vertessy et al. 1995; Meinzer et al. 2005): AS= α ⋅DBH β (1) where α is a constant and β is the allometric scaling exponent, and both species-specific coefficients. Coefficients of the regression models for each measured species, number of samples and R2 are provided in the legend of Figure 2.5. The ground-projected crown area (Acp, m2) of sample trees was measured in eight horizontal directions using a compass, crown mirror, and measuring tape. The octagonal area was calculated as the sum of eight triangles (Schmidt 2007). These results were used to compute canopy conductance (GC, mm s−1). 2.2.5 Tree sap flow Sap flux density (Fd) was measured in the tree stems of five trees per species using the thermal dissipation method (Granier 1987) during June and July 2008 and repeated during the same period in 2009. This period was chosen as it was considered the most active period in the context of plant water use, just before the onset of the Monsoon rains. All sensor installations were made on the north-facing side of the trees to avoid exposure to the sun and minimize direct short-wave radiation (Wilson et al. 2001; Wullschleger et al. 2001). In addition, the sensors were covered with a radiation shield (Styrofoam sheets with aluminium foil) to further minimize the direct thermal load. Power for heating the sensors was provided by lead-acid batteries that were recharged with solar panels via a charge controller. Each sensor consisted of a pair of 2 mm diameter probes vertically aligned ca. 15 cm apart. Each probe included a 0.2 mm diameter copperconstantan thermocouple. The two thermocouples were joined at the constantan leads, so that the voltage measured across the copper leads provided the temperature difference between the heated upper probe and the lower reference. Heating across the entire length of the 20 mm upper probe was achieved with a constant current of 120 mA supplied to a constantan heating wire, resulting in a heating power of 200 mW (Granier 1987). Sensors were placed in the outer 20 mm of the sapwood (annulus 1, 0−20 mm radial sapwood depth). In cases where the tree trunk was large with a sapwood radius greater than 20 mm (Table 2.1), a second sensor was implanted 20 to 40 mm into the sapwood. Sensors were spaced 10–15 cm circumferentially, away from the first sensor pair, on the same side of the stem to avoid azimuth differences. Temperature differences were measured every 5 min and a 30-min mean value was logged (DL2e with LAC-1 in single ended mode, Chapter 2 – Up-scaling to stand transpiration of an Asian temperate mixed-deciduous forest from single tree sap flow measurements ! 48 Delta-T Devices, England). Sap flux density (Fd, g m−2 s−1) for each sensor was calculated from ΔT in accordance with Granier (1987), assuming zero Fd (i.e., ΔTmax) at night and VPD near zero: Fd=119 ⋅K1.231 (2) where, K=(ΔTmax − ΔT) ΔT (3) Tree water use (TWU, kg h−1) was obtained by multiplying Fd by sapwood cross-sectional area (AS, m2). TWU =(Fdi ⋅ASi i=1 n ∑) (4) where, Fdi is sap flux density of the annulus i (g m−2 s−1) and ASi is sapwood area of the annulus i (m2). This took into account the second annulus ring, in case a second sensor was installed into the tree. For example, i = 1 was annulus ring 0−20 mm sapwood depth, i = 2 was annulus ring 20−40 mm sapwood depth. Canopy transpiration (EC, mm per day) was computed by summing the contributions from all the trees in the study plot: EC=TWUj j=1 n ∑×Aplot −1 (5) where, TWUj is tree water use of tree j (kg h−1) and Aplot is plot area (m2). TWU of the trees on which sensors were not installed was estimated from the relationship between Fd and the computed AS of each species (Eq. 1). 2.2.6 Estimation of canopy conductance Canopy conductance was calculated from the sap flow measurements or stand/canopy transpiration, in relation to climate variables: half hourly averaged air temperature, VPD, and canopy transpiration as described by Köstner et al. (1992): GC=( ρ w⋅Gv⋅Tk)⋅EC VPD (6) where, ρw is density of water (998 kg m−3) and Gv is gas constant of water vapor (0.462 m3 kPa kg−1 K−1), Tk is air temperature (K), and EC is canopy transpiration (mm s−1) (Schmidt 2007). To estimate GC based on this model, data from measurements between 10 and 15 h, when half hourly rates of EC were highest, were used. This model assumed that tree canopies were well coupled to the atmosphere, so that aerodynamic conductance (GA) was larger than GC (Köstner et al. 1992; Phillips and Oren 1998). Chapter 2 – Up-scaling to stand transpiration of an Asian temperate mixed-deciduous forest from single tree sap flow measurements ! 49 2.2.7 Statistical analyses Fd and environmental variables were recorded as half hourly averaged values. These variables, including TWU and GC estimated from Fd were converted into daily averages. Data are presented as mean ± standard deviation (SD). Fd, TWU, and GC were compared between years and also among tree species using one-way ANOVA. Where differences were found among species, a post-hoc Kruskal-Wallis test was carried out. Normality of samples was established by testing the residuals obtained from the ANOVA. Measured and estimated EC by the relationship between TWU and DBH were compared with t-test. Regression analysis was tested with Pearson correlation test. All statistical analyses including regression models were based on a 0.05 significance level and performed with R version 2.6.2 (R Development Core Team, 2008). 2.3 Results 2.3.1 Micrometeorological and soil moisture measurements Daily mean air temperatures over the measurement period of June and July were about 19.3°C in 2008 and 18.3°C in 2009. Averaged daily VPD were 0.34 kPa in 2008 and 0.28 kPa in 2009, while the summed daily PAR were 32.6 mol m−2 d−1 in 2008 and 25.9 mol m−2 d−1 in 2009 respectively (Figure 2.3a). The total amount of precipitation recorded during the measurement period was 89 mm in 2008, and 178.5 mm in 2009. Mean soil water content (θ) within the 30cm soil profile was 0.24 ± 0.04 m3 m−3 in 2008 and 0.21 ± 0.05 m3 m−3 in 2009 (Figure 2.3a). Before the onset of our experiments 2009 was comparatively drier than 2008, as demonstrated by lower θ at the beginning of measurements. A rainstorm event on June 3, 2009 amounting to 73 mm, however, significantly raised θ (from 0.11 to 0.29 m3 m−3), and θ thereafter was comparable to 2008. 2.3.2 Transpiration rate and canopy conductance Mean maximum Fd for the 21 trees measured was 247.5 ± 93.1 kg m−2 h−1 in 2008 and 271.5 ± 97.1 kg m−2 h−1 in 2009 (Table 2.2). There was no significant (F = 0.88, P = 0.35) difference in Fd between the two years. And also, mean maximum TWU was 21.0 ± 21.8 kg d−1 in 2008 and 32.9 ± 22.0 kg d−1 in 2009 (Table 2.2). A comparison of maximum TWU from different years showed same results (F = 1.00, P = 0.33). Mean daily Fd of Q. mongolica, T. amurensis, U. davidiana, C. controversa, and A. mono were 40.9 ± 17.8 kg m−2 h−1 (n = 5), 49.5 ± 26.1 kg m−2 h−1 (n = 5), 49.2 ± 8.4 kg m−2 h−1 (n = 5), 59.5 ± 24.5 kg m−2 h−1 (n = 3) and 55.4 ± 17.8 kg m−2 h−1 (n = 3). TWU and GC averaged over the measurement period are shown in Table 2.3. The mean daily TWU ranged from 1.2 kg d−1 for A. mono with DBH of 15.0 cm to 70.1 kg d−1 for Q. Chapter 2 – Up-scaling to stand transpiration of an Asian temperate mixed-deciduous forest from single tree sap flow measurements ! 50 mongolica with DBH of 38.2 cm. And mean GC amounted from 0.7 mm s−1 for Q. mongolica with DBH of 13.3 cm to 16.1 mm s−1 for T. amurensis with 29.2 cm. Mean maximum GC of the stand was 5.6 ± 4.8 mm s−1. The averaged EC was 0.64 ± 0.26 mm d−1 in 2008 and 0.70 ± 0.30 mm d−1 in 2009. The maximum EC occurred around day 177 in 2009 (June 26, 0.97 mm d−1, Figure 2.4), coinciding with the highest daily total PAR and VPD. There were no significant (F = 0.31, P = 0.73) differences in daily transpiration among Q. mongolica, T. amurensis, and U. davidiana. The percentage mean contribution of the three species was about 30% each, while C. controversa and A. mono each accounted for about 4% of the total transpiration. There was no significant influence of species on Fd (P = 0.82), TWU (P = 0.19) and GC (P = 0.23). Figure 2.3 (a) Daily mean vapor pressure deficit (VPD, kPa) and daily amounts of photosynthetic active radiation (PAR, mol m−2 d−1), (b) rainfall (mm d−1) and soil water content (θ, m3 m−3) recorded at the study site during June 2008 and 2009 when sap flow measurements were conducted. Chapter 2 – Up-scaling to stand transpiration of an Asian temperate mixed-deciduous forest from single tree sap flow measurements ! 51 Table 2.2 Maximum sap flux density (Fd) and maximum tree water use (TWU) averaged for 21 measured trees from June 2008 and June 2009. June 2008 June 2009 F-value P-value Max Fd (kg m−2 h−1) 247.48 271.46 0.8834 0.3529 SD 93.13 97.14 Max TWU (kg d−1) 21.02 32.91 0.9971 0.3264 SD 21.77 22.02 Table 2.3 Mean tree water use (TWU, kg d−1) and canopy conductance (GC, mm s−1) of individual tree species. Species Sample trees Mean TWU (kg d−1) Mean GC (mm s−1) Q. mongolica Q1 27.8 ± 11.6 13.1 ± 4.3 Q2 20.1 ± 9.1 3.4 ± 0.9 Q3 14.2 ± 6.4 4.3 ± 1.7 Q4 1.3 ± 0.7 0.7 ± 0.2 Q5 70.1 ± 23.4 9.7 ± 4.1 T. amurensis T1 20.4 ± 9.7 16.1 ± 4.3 T2 1.9 ± 0.9 2.0 ± 0.8 T3 17.2 ± 5.2 6.7 ± 2.2 T4 2.6 ± 1.3 2.1 ± 0.6 T5 3.7 ± 2.0 2.4 ± 0.6 U. davidiana U1 13.7 ± 5.7 13.4 ± 2.9 U2 9.2 ± 2.9 7.0 ± 1.2 U3 22.3 ± 7.7 9.2 ± 2.9 U4 17.3 ± 5.5 5.4 ± 2.3 U5 9.1 ± 3.7 2.9 ± 0.8 C. controversa C1 14.7 ± 5.4 4.2 ± 1.8 C2 10.8 ± 4.0 2.8 ± 0.9 C3 2.8 ± 1.3 1.5 ± 0.4 A. mono A1 1.2 ± 0.6 2.3 ± 0.6 A2 5.4 ± 3.1 2.6 ± 0.7 A3 2.0 ± 1.0 1.6 ± 0.5 Chapter 2 – Up-scaling to stand transpiration of an Asian temperate mixed-deciduous forest from single tree sap flow measurements ! 52 Figure 2.4 Estimated canopy transpiration (EC, mm d−1) and transpiration of measured species, Quercus mongolica (Q.m.), Tilia amurensis (T.a.), Ulmus davidiana (U.d.), Cornus controversa (C.c.), and Acer mono (A.m.). 2.3.3 Relationship between tree water use and tree size Parameters derived from the allometric equation (Eq. 1) relating AS and DBH of the five different species are shown in Figure 2.5. The converged regression model inverted from the five different species showed a strong relationship (n = 35, R2 = 0.81, P <0.001) between AS and DBH. This model was used to compute AS of non-measured species to arrive at AS for the whole study plot. Q. mongolica (0.15 m2 ha-1), T. amurensis (0.29 m2 ha-1), U. davidiana (0.16 m2 ha−1), C. controversa (0.03 m2 ha−1), and A. mono (0.06 m2 ha−1), accounted for 79.3% of total AS which was 0.87 m2 ha−1 (study plot = 0.2 ha). Observed Fd, TWU and GC were dependent on tree size, determined by DBH and AS. Fd and TWU had a stronger dependency on DBH than on AS. For example, a regression of Fd in individual trees against AS did not show any relationship (n = 21, R2 = 0.03, P >0.44), but a regression between Fd and DBH showed that DBH could explain 21% of the observed variability (n = 21, R2 = 0.21, P = 0.036) in Fd among the studied trees (Figure 2.6a). Moreover, the relationship between mean daily TWU and DBH was stronger (n = 21, R2 = 0.87, P <0.001) for all measured species (Figure 2.6b). SA was also significantly correlated with TWU, but less than DBH (n = 21, R2 = 0.83, P <0.001). GC had significant relationship with both DBH and AS, and also like other relations GC was correlated better with DBH (n = 17, R2 = 0.63, P <0.001) (Figure 2.6c) than AS (n = 17, R2 = 0.48, P = 0.002). EC estimated from measured TWU was compared with EC* estimated from modelled TWU based on the relationship of DBH and TWU (Figure 2.6b). In this study, the empirical model was TWU = Chapter 2 – Up-scaling to stand transpiration of an Asian temperate mixed-deciduous forest from single tree sap flow measurements ! 53 0.0002DBH3.4302 for mean daily TWU over the measurement period (P <0.001). There was a strong agreement (n = 22, R2 = 0.94, P <0.0001) between the measured and modeled EC*. Figure 2.5 Relationship between sapwood area (AS, cm2) determined from increment core extracted from more than five samples per species and their respective diameter at breast height (DBH) (n = 35, R2 = 0.81, P <0.001). The regression equation for Quercus mongolica was AS = 0.2067DBH2.1 (n = 9, R2 = 0.94), for Tilia amurensis was AS = 0.1846DBH2.0 (n = 10, R2 = 0.85), for Ulmus davidiana was AS = 0.1745DBH2.1 (n = 3, R2 = 0.98), for Cornus controversa was AS = 0.0757DBH2.2 (n = 3, R2 = 0.98), for Acer mono was AS = 0.0215DBH2.4 (n = 3, R2 = 0.98), and for all the studied species was AS = 0.053DBH2.4 (n = 35, R2 = 0.81) 2.3.4 Relationship between tree water use and climate factors Both VPD (R2 = 0.78 and P <0.0001) and PAR (R2 = 0.91, P <0.0001) (Figure 2.7) could explain most of the daily fluctuations in EC. EC increased with increasing VPD, attaining a maximum at VPD = 0.5 kPa, but later dropped at higher VPD (Figure 2.7a). On days when VPD was high, EC increased during morning hours with increasing PAR and reached a maximum at around midday, when PAR was >1,200 µmol m−2 s−1 (Figure 2.7b). GC of the stand corresponding to the light conditions more than 20 mol m−2 d−1 of PAR, was plotted against VPD (Figure 2.7c). The selected data showed a log-linear relationship between GC and VPD (R2 = 0.69, P <0.0001). This regression model agreed with the simplified model of Lohammar et al. (1980): GC=b−c⋅(lnVPD) (7) where, b is GC at a reference VPD = 1 kPa. Chapter 2 – Up-scaling to stand transpiration of an Asian temperate mixed-deciduous forest from single tree sap flow measurements ! 54 Figure 2.6 Relationship between DBH and (a) mean sap flux density (Fd, kg m−2 h−1) (n = 21, R2 = 0.21, P = 0.036), (b) mean tree water use (TWU, kg d−1) (n = 21, R2 = 0.87, P <0.001) and (c) canopy conductance (GC, mm s−1) significant (n = 17, R2 = 0.63, P <0.001), of all the measured trees. Chapter 2 – Up-scaling to stand transpiration of an Asian temperate mixed-deciduous forest from single tree sap flow measurements ! 61 Luis VC, Jiménez MS, Morales D, Kucera J, Wieser G (2005) Canopy transpiration of a Canary islands pine forest. Agric For Meteorol 135:117–123 Meinzer FC (2003) Fuctional convergence in plant responses to the environment. Oecologia 134:1–11 Meinzer FC, Goldstein G, Andrade JL (2001) Regulation of water flux through tropical forest canopy trees: Do universal rules apply? Tree Physiol 21:19–26 Meinzer FC, Bond BJ, Warren JM, Woodruff DR (2005) Does water transport scale universally with tree size? Funct Ecol 19:558–565 O’Brien JJ, Oberbauer SF, Clark DB (2004) Whole tree xylem sap flow responses to multiple environmental variables in a wet tropical forest. Plant Cell Environ 27:551–567 O’Grady AP, Cook PG, Eamus D, Duguid A, Wischusen JDH, Fass T, Worldege D (2009) Convergence of tree water use within an arid-zone woodland. Oecologia 160:643–655 Oren R, Phillips N, Katul G, Ewers B, Pataki DE (1998) Scaling xylem sap flux and soil water flux in forests. Ann Sci For 55:191–216 Oren R, Pataki DE (2001) Transpiration in response to variation in microclimate and soil moisture in southeastern deciduous forests. Oecologia 127:549–559 Phillips N, Oren R (1998) A comparison of daily representations of canopy conductance based on two conditional time-averaging methods and the dependence of daily conductance on environmental factors. Ann Sci Forest 55:217–235 Phillips N, Oren R, Zimmermann R, Wright SJ (1999) Temporal patterns of water flux in trees and lianas in a Panamanian moist forest. Trees 14:116–123 R Development Core Team (2009) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria, ISBN 3-900051-07-0. Accessed 17 Nov 2009 Schmidt MWT (2007) Canopy transpiration of beech forests in Northern Bavaria−Structure and function in pure and mixed stands with oak at colline and montane sites. Dotoral dissertation, University of Bayreuth, Germany, 239 pp. URL: http://opus.ub.uni-bayreuth.de/volltexte/2008/428 Sevanto S, Nikinmaa E, Riikonen A, Daley M, Pettijohn C, Mikkelsen TN, Phillips N, Holbrook NM (2008) Linking xylem diameter variations with sap flow measurements. Plant Soil 305:77−90 Vertessy RA, Benyon RG, O’sullivan SK, Gribben PR (1995) Relationships between stem diameter, sapwood area, leaf area and transpiration in a young mountain ash forest. Tree Physiol 15:559−567 Walter H, Lieth H (1967) Klimadiagramm-Weltatlas. VEB Gustav Fischer Verlag, Jena Wessman CA, Aber JD, Peterson DL, Melillo JM (1988) Remote sensing of canopy chemistry and nitrogen cycling in temperate forest ecosystems. Nature 335:154−156 Chapter 2 – Up-scaling to stand transpiration of an Asian temperate mixed-deciduous forest from single tree sap flow measurements ! 62 Wilson KB, Hanson PJ, Mulholland PJ, Baldocchi DD, Wullschleger SD (2001) A comparison of methods for determining forest evaporation and its components: sap-flow, soil water budget, eddy covariance and catchment water balance. Agric For Meteorol 106:153−168 Wullschleger SD, Hanson PJ, Todd DE (2001) Transpiration from a multi-species deciduous forest as estimated by xylem sap flow techniques. For Ecol Manage 143:205−213 Zeppel M, Eamus D (2008) Coordination of leaf area, sapwood area and canopy conductance leads to species convergence of tree water use in a remnant evergreen woodland. Aust J Bot 56:97−108 Chapter 3 – Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use ! 63 Chapter 3 Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use Journal of Plant Research (2013) DOI 10.1007/s10265-013-0563-5 Eun-Young Junga*, Dennis Otienoa, Hyojung Kwona,c, Bora Leea, Jong-Hwan Limb, Joon Kimc, John Tenhunena aDepartment of Plant Ecology, University of Bayreuth, D-95440 Bayreuth, Germany bDepartment of Forest Conservation, Korea Forest Research Institute, 130-712 Seoul, Republic of Korea cDepartment of Landscape Architecture and Rural Systems Engineering, Seoul National University, 151-742 Seoul, Republic of Korea Abstract Understanding the dynamics of transpiration of the overstory (EO) and understory (EU) in forest stands under the influence of the Asian monsoon is needed to improve estimation of forest water budget and identify key factors controlling forest water use under climate change. In this study, EO and EU of a temperate deciduous forest stand located in South Korea were measured during the growing season of 2008 using sap flow methods. The objectives of this study were to (1) quantify the total transpiration of the forest stand, (2) determine their relative contribution to ecosystem evapotranspiration, and (3) identify factors controlling the transpiration of each layer. EO and EU were 174 and 22 mm, respectively. Total transpiration accounted for 55% of the total Eeco, revealing the importance of unaccounted contributions to Eeco (i.e., soil evaporation and wet canopy evaporation). During the monsoon period, there was a strong reduction in the total transpiration, likely because of reductions in photosynthetic active radiation (PAR), vapor pressure deficit (VPD) and plant area index (PAI). The ratio of EU to EO declined during the same period, indicating an effect of monsoon on the partitioning of Eeco in its two components. The seasonal pattern of EO was synchronized with the overstory canopy development, while EU varied in function of the understory canopy development as long as the overstory canopy remained open. Following the overstory canopy closure, environmental conditions of the understory controlled the variations in EU. Chapter 3 – Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use ! 64 Key-words: Asian monsoon, Canopy conductance, Transpiration of overstory, Understory, Warmtemperate deciduous forest 3.1 Introduction Temperate deciduous forests in China and Korea composed of Quercus, Carpinus, Ulmus, and Tilia species have developed together with the establishment of warm-temperate summer conditions, a strong decrease in winter temperatures and an increase in monsoon-induced precipitations. These forests have been compared with other major deciduous forest types of the region and are described as warm-temperate deciduous forests (Nakashizuka and Iida 1995). In general, two to three times higher plant species diversity is found in the temperate forests of the Asian monsoon region than in North America (Latham and Ricklefs 1993; Qian and Ricklefs 1999). This difference in plant diversity appears to result from a greater physiographic heterogeneity in Asia, which allowed for an allopatric speciation in response to the sea level fluctuations after temperate forest zones became disjunct in the late Tertiary (Qian and Ricklefs 2000). Warm and humid summer conditions due to the monsoon also contribute to the great diversity of plant species in the eastern Asian forests (Röhrig and Ulrich 1991). Thus, the warm-temperate deciduous forests of Korea and other parts of Asia represent an important ecosystem type with multi-layered physiognomy (Kim 2002), where both the overstory and understory are well developed and display diverse species compositions. As a result, large differences are expected in the partitioning of water use between the overstory and understory, which may influence forest ecosystem functions. In particular, these differences can influence forest energy flows, nutrient cycling and niche partitioning within species (Jackson et al. 1995). The species composition of understory vegetation is in part determined by canopy tree species and structure which modify microclimate, light availability, soil water content, and soil nutrients inputs below the canopy (Canham et al. 1994; Augusto et al. 2003; Barbier et al. 2008). Simultaneously, understory vegetation competes with overstory trees for resources and may influence tree growth (Riegel et al. 1992). In temperate deciduous forests, light availability at the understory is high in early spring before canopy closure, but decreases in early summer with the leaf emergence of overstory trees. Low radiation and wind speed below the canopy create an environment of relatively low vapor pressure gradient, resulting in low transpiration rates of the understory (Landsberg and Gower 1997). However, even in relatively closed canopies, there is enough light reaching the understory to allow the physiological functioning of the vegetation (Lieffers et al. 1999). Following precipitation events, as the forest canopy dries, the rate of transpiration of the understory is expected to increase with the increasing vapor pressure gradient (Black and Kelliher 1989). However, it will still remain below the canopy transpiration. Thus, although the understory may be composed of plant species different from those of the canopy, the transpiration of the former may still be tightly coupled to canopy processes given that the canopy can strongly influence the understory Chapter 3 – Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use ! 65 microclimate. Trends and partitioning of transpiration at both levels, therefore, usually display strong seasonality according to the dynamics of canopy cover and leaf area index (Wullschleger et al. 2001). Evapotranspiration (Eeco) in such forest stands consists of four main components, namely, overstory transpiration (EO), understory transpiration (EU), soil evaporation (Eg), and evaporation from wet canopy surface (Ew) (Barbour et al. 2005). Quantifying the relative contribution of each component to Eeco is essential in order to determine their importance and to identify key factors controlling forest water use. In turn, this allows for accurate assessment of water use by forest ecosystems, provides empirical data for the parameterization and calibration or validation of forest hydrological models, which requires all these components, as well as information for the sustainable management of forests (Hatton et al. 2003; Zeppel et al. 2006). To the best of our knowledge, there have been no attempts to partition Eeco into EO and EU in any warm-temperate forest stands under the influence of the Asian monsoon climate, although Eeco has been widely measured in several studies including those of the Asia Flux groups (Kosugi et al. 2007; Shi et al. 2008; Tanaka et al. 2008; Kang et al. 2012). The Asian monsoon, which is characterized by strong southwest surface winds and heavy rains (Lau and Li 1984), leads to forest canopy disturbance that affects canopy structure and function, including canopy plant area index (PAI). The monsoon is also generally characterized by hot and humid conditions, a reduction in vapor pressure deficit (VPD) and low photosynthetic active radiation (PAR). All these changes are likely to influence forest water use differently than that of the dry-summer deciduous temperate forests located in Europe and North America. According to the climate simulations for South Korea (Im et al. 2008), the frequency, intensity and duration of the Asian monsoon are likely to increase in the next 30 years. This will alter the structure and function of the warm-temperate deciduous forests of Asia and how they are managed. There is, therefore, an urgent need to fill the gaps in the knowledge on the influence of the Asian monsoon on forest water use and its dynamics. In this study, we conducted transpiration measurements using sap flow methods at both the overstory and understory of a warm-temperate deciduous forest stand located in Pocheon-si, Gyeonggi-do, South Korea. Our objectives were to (1) quantify EO and EU over the growing season, (2) determine their relative contribution to Eeco, and (3) identify controlling factors on the total amount and rates of EO and EU. The specific questions addressed were: (1) How does overstory development influences understory microclimate? (2) How is the partitioning of Eeco into EO and EU influenced by overstory and understory development? and (3) How does the monsoon affect on the total amount and rates of EO and EU? We hypothesized that the contribution of both EO and EU to Eeco is strongly dependant on microclimatic factors, mainly PAR and VPD that control canopy transpiration and the status of the overstory and understory canopy development. Chapter 3 – Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use ! 66 3.2 Materials and methods 3.2.1 Study sites The study was conducted in a warm-temperate deciduous forest stand of the Gwangneung National Arboretum located in the central part of the Korean peninsula (37°45’25.37’’N, 127°9’11.62’’E) at an elevation of 340 m above sea level. The site is registered as a KoFlux Supersite (Kim et al. 2006) and longterm ecological monitoring station (Oh et al. 2000). Mean annual precipitation over the past 25 years was 1,436 mm and was mainly concentrated in the months of late June and late July due to the influence of the monsoon rainband. The Korean monsoon is characterized by persistent and intense rainfall and is known as Changma. In 2008, when the measurements were made, Changma started on June 17 and ended on July 26 (Figure 3.2; Korea Meteorological Administration 2011). For the last 30 years, mean annual temperature was 11.5°C and temperature ranged of −11.5 and 30.0°C (Kim et al. 2006). Soil depth ranges from 0.4 to 0.8 m and the soil texture is predominantly sandy loam. The bedrock primarily consists of granite gneiss and schist (MOST 1999). The site is located on a slope of 15°, facing southwest. The forest stand is at climax and is dominated by 80to 200-year-old Quercus serrata Thunb. ex Murray and Carpinus laxiflora (Siebold & Zucc.) Blume var. laxiflora of an average height of 18 m (Cho et al. 2007). The understory is composed of a high diversity of species of saplings and shrubs and has an average height of 2 m (Lim et al. 2003). 3.2.2 Micrometeorological measurements Air temperature (Ta), water vapor density, photosynthetically active radiation (PAR), net radiation (RN), wind speed (U) and rainfall were measured at 40-m height tower. Below the canopy, Ta, water vapor density, RN, and U were measured at a height of 4 m. Ta and U at both height were measured with a threedimensional sonic anemometer (model: CSAT3, Campbell Scientific Inc., Logan, Utah, USA). PAR was measured by photodiode sensors (model: BPW21, Osram Semiconductor GmbH, Regensburg, Germany) at a height of 2 m at six random locations in the plot of sap flow measurements. Light transmittance was calculated as the ratio of PAR at the 40-m height (PAR40) and PAR at the 2-m height (PAR2). Vapor pressure deficit at the 40-m height (VPD40) and at the 4-m height (VPD4) were derived from air temperature and water vapor density measured at the respective heights (Murray 1967). Soil water content integrated over 0 to 30 cm depth was measured using soil moisture probes (model: CS616, Campbell Scientific Inc.). The data from the eddy covariance system were sampled at 10 Hz and the meteorological data were monitored every 30 seconds, and half-hourly means of both data were calculated. The data was stored on three types of data loggers (model: CR-5000 and CR-3000, Campbell Scientific Inc. and DL2e, Delta-T Devices, Cambridge, UK). The measurements of meteorological variables at the overstory and understory Chapter 3 – Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use ! 67 were used to analyze the impacts of environmental conditions on EO and EU and to calculate canopy conductance (GC) of both layers. 3.2.3 Biometric measurements PAI, which includes understory and overstory trees, was measured every three weeks throughout the year using a plant canopy analyzer (model: LI-2000, LI-COR Inc., Lincoln, Nebraska, USA) under diffuse light conditions at 12 sampling points with 50 m x 50 m grid interval over the plot of sap flow measurements (Kwon et al. 2010). In order to measure maximum understory leaf area index (LAIU), leaf samples of all the saplings in three plots measuring 2 m × 2 m were collected in early July, and the area of the leaf samples was measured with a leaf area meter (model: LI-3100, LI-COR Inc.). LAIU reached its peak in May and did not decrease until early September. Therefore, we assumed that the leaf area measurement conducted in July reflected the maximum value of LAIU. In order to measure bark and sapwood depth of the sample trees of Q. serrata and C. laxiflora, an increment corer was used to extract cores from the stems at sap flow sensor installation height. Bark and sapwood were visually distinguishable for both species. Sapwood area (AS) was calculated from the measurements of diameter at breast height (DBH) and depths of bark and sapwood. A non-linear regression was established between AS and DBH for both species (Figure 3.1a). The regression was used to estimate total AS of all trees of the study plot (Ast) (Vertessy et al. 1995; Meinzer et al. 2005). Figure 3.1 Climate diagram Relationships (a) between stem diameter at breast height (DBH) and sapwood area (AS) in power function (AS = 1.664DBH1.483, n = 14, R2 = 0.90, P <0.001) and (b) between DBH and tree water use (TWU) in three-parameter sigmoidal function (TWU = 97.07/(1 + e−[(DBH−49.85)/14.22]), n= 11, R2 = 0.97, P <0.001) for canopy tree species of Quercus serrata (Q.s.) and Carpinus laxiflora (C.l.) from the years 2007 and 2008. Chapter 3 – Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use ! 68 In a 30 m x 30 m plot, the DBH of all stems larger than 5 cm in diameter (70 stems in total) was measured. Basal area was then calculated from the measured DBH. The total basal area of the plot was 38.5 m2 ha-1. The dominant tree species were Q. serrata and C. laxiflora, constituting for 71% and 22% of the total basal area, respectively. The average DBH of Q. serrata and C. laxiflora were 45.4 ± 13 and 33.8 ± 10 cm (hereafter standard deviation (SD) is indicated as ±), respectively. The understory was composed of saplings and shrubs of different species, such as Euonymus oxyphyllus, Celtis jessoensis, Styrax obassia, Cornus kousa and Sorbus alnifolia. The basal area of the understory saplings was about 7% of the total basal area. 3.2.4 Measurement of overstory and understory transpiration Overstory (EO) and understory (EU) transpiration were measured from April 1 to September 27 in 2008, and total transpiration (EO + EU) during this period was considered as total annual transpiration since leaf senescence started at the end of September and no evergreen species existed in the study area. EO was estimated from sap flux density (Fd, g m−2 s−1) measured using thermal dissipation probes (TDP), which were constructed at the technical laboratory of the Department of Plant Ecology, University of Bayreuth, based on the original design by Granier (1987). The TDP consisted of two probes (2 mm in diameter and 20 mm in length) aligned vertically 10 cm apart into the sapwood. Each probe contained a copper-constant thermocouple and was connected in parallel (Granier 1987). The temperature of the upper probe, constantly heated by a 0.2 W power supply, was influenced by the rate of convective heat transport away from the heat source, with the vertically flowing sap in the xylem. The lower probe, which was an unheated reference, reflected the ambient sap temperature. The probes were inserted in two depths between 0 and 20 mm of the outer ring and between 20 and 40 mm of the inner ring, at a height of 1.3 m above ground, on the north-facing side to minimize the effects of direct shortwave radiation (Wilson et al. 2001; Wullschleger et al. 2001). After installation, each probe was covered with Styrofoam sheaths and aluminum foil to avoid direct thermal load from radiation. Temperature differences (ΔT) between the upper and lower reference probes were measured every 30 seconds and half-hourly means were calculated. Data were recorded on data loggers (model DL2e, Delta-T Devices). Fd was calculated as a function of ΔT according to standard calibration of the TDPs (e.g., Granier, 1987), Fd=119 ⋅[(ΔTmax − ΔT)⋅ ΔT−1]1.231 (1) where ΔTmax is the maximum temperature difference between the two probes during a day when sap flow was null. Fd at different sapwood depths of the same individual was used to calculate the sapwood area weighted sap flow density (Fdt) with the following equation: Fdt =(Fdo ⋅Aso +Fdi ⋅Asi )⋅(Aso +Asi )−1 (2) Chapter 3 – Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use ! 69 where Fdo is the sap flux density of the outer ring, Aso is the area of the outer ring, Fdi is the sap flux density of the inner ring and Asi is the area of the inner ring. We used a weighted average Fdt to estimate total forest transpiration because Fdi was lower than Fdo in Q. serrata and C. laxiflora by 0.2 and 0.8, respectively. Fd in the sapwood deeper than 40 mm sapwood depth, i.e., outside the heating probe, was estimated by applying an empirical function described by Poyatos et al. (2007) for Quercus spp. and Gebauer et al. (2008) for Carpinus spp. EO (mm h−1) was calculated as: EO=Fdt ⋅Ast ⋅Aplot −1 (3) where Fdt is the mean Fdt of the sample trees, Ast is the total sapwood area of all trees in the plot, and Aplot is the plot area. In order to determine circumferential differences in sap flow, we installed multiple sap flow probes, i.e., three per species, within the outer 20 mm sapwood on the southand north-facing side of the stem from July to October. The two sets of the probes (on each tree) were placed at relatively similar heights, but carefully positioned apart so that heating from one sensor on the southern azimuth did not interfere with the other on the north-side. However, we observed no circumferential variation in Fd between the south and north sides of the overstory tree stems (P <0.0001), hence we did not perform any further corrections on EO. In order to measure EO, we selected three individuals of Q. serrata (dominant species) and three individuals of C. laxiflora (sub-dominant species) with DBH between 24 and 59 cm and heights between 12 and 18 m. For the measurement of EU, three individuals of E. oxyphyllus, one individual of C. jessoensis and one individual of S. alnifolia saplings with stem diameters ranging from 1.7 to 3.0 cm (at 1 m above the ground) and heights ranging from 1.8 to 2.5 m were selected (Table 3.1). The choice of a species was based on the health status and its dominance within the plot. The number of individuals used for the measurements was severely restricted by the National Arboratum management policy since the Gwangneung forest is one of the only natural forests at climax whithin Korea. These forests have been placed under strict protection. Any destruction of trees of this forest is, therefore, strongly prohibited. However, the scaling of sap flow measurements from plot-scale to stand-scale requires that a sample be representative of the spatial distribution of species and size classes within a stand (Köstner et al. 1998; Oren et al. 1998; Kumagai 2005). Therefore, we used the results of previous measurements for comparison. In the previous year (2007), sap flow measurements were comducted on five individuals of Q. serrata trees (dominant species), with DBH distribution within the classes of 10−20, 20−30, 30−40, 40−50 and 50−60 cm which are representative of the DBH classes in this forest stand (unpublished data). We examined the relationship between DBH and rates of sap flow after combining data from the years 2007 and 2008. DBH explained 97% of the variations in water use among trees in a sigmoidal function relationship described by Meinzer et al. (2005) (Figure 3.1b), which agrees with the findings of previous studies in Chapter 3 – Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use ! 70 similar forest types (Vertessy et al. 1995; Meinzer 2003; Meinzer et al. 2005; Jung et al. 2011). Thus, despite the limited sample size used in subsequent analyses, we are convinced that the data are well representative of the forest stand and that the type II error is within acceptable limits. Table 3.1 Characteristics of the sample trees for sap flow measurements. DBH, AS, and AL indicate diameter at breast height, sap wood area of the overstory trees, and leaf area of the understory trees, respectively. Sample trees Tree species Tree height [m] DBH [cm] AS/AL [m2] Q1 Quercus serrata 15 35.4 0.03 Q2 17 58.8 0.07 Q3 17 38.8 0.03 C1 Carpinus laxiflora 18 43.0 0.07 C2 16 33.0 0.04 C3 12 24.1 0.02 SHB1 Euonymus oxyphyllus 1.8 1.9 0.83 SHB2 2.0 2.1 0.98 SHB3 1.9 1.7 0.90 SHB4 Celtis jessoensis 2.5 2.0 1.08 SHB5 Sorbus alnifolia 2.5 3.0 1.39 EU was estimated from sap flow (F, g h−1), which was measured by the stem heat balance (SHB) technique (Sakuratani 1981). This technique has been successfully applied to accurately estimate sap flow of herbaceous stems (Baker and van Bavel 1987), saplings (Lei et al. 2010), and branches of large trees (Otieno et al. 2007) with diameters ranging from 2 to 125 mm. The SHB sensors used for the measurements were manufactured at the electronic workshop, University of Bayreuth, according to the original design by Sakuratani (1981), improved by Weibel and de Vos (1994) for stems or branches with diameter ranging Chapter 3 – Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use ! 77 Table 3.2 Monthly mean overstory transpiration (EO), understory transpiration (EU), evapotranspiration (Eeco) and monthly accumulated rainfall during the growing season of 2008. Numbers in parenthesis indicate standard deviation. Figure 3.4 Relationship between monthly cumulative rainfall and the ratio of monthly averaged stand transpiration (= EO + EU) to monthly averaged ecosystem evapotranspiration (Eeco). The relationship was significant (R2 = 0.71, P <0.01, y = −0.131ln(x) + 1.3247). Month EO [mm d−1] EU [mm d−1] Eeco [mm d−1] EU/EO (EO + EU)/Eeco Rainfall [mm mon−1] Apr 0.64 (± 0.4) 0.14 (± 0.1) 0.87 (± 0.3) 0.22 0.89 24.0 May 1.20 (± 0.4) 0.17 (± 0.1) 1.71 (± 0.7) 0.14 0.80 82.0 June 1.39 (± 0.4) 0.16 (± 0.1) 2.27 (± 1.1) 0.11 0.68 136.5 July 0.80 (± 0.5) 0.07 (± 0.1) 1.50 (± 0.8) 0.09 0.58 630.0 Aug 0.99 (± 0.3) 0.11 (± 0.1) 2.60 (± 1.0) 0.11 0.42 291.5 Sept 0.98 (± 0.4) 0.10 (± 0.1) 1.45 (± 0.9) 0.10 0.74 121.0 Chapter 3 – Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use ! 78 Figure 3.5 Relationships between transpiration and photosynthetic active radiation (PAR) and between transpiration and vapor pressure deficit (VPD). Numbers of 40, 4, and 2 indicate 40-m, 4-m, and 2-m heights, respectively. Note that PAR2 was multiplied by a factor 10 for the convenience of plotting with PAR40 except for April. Closed and open circles indicate overstory and understory, respectively. All the regressions are in polynomial functions. Chapter 3 – Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use ! 79 3.3.5 Canopy conductance Mean daytime GC of the overstory was significantly higher before Changma than after (t-test, df = 42, t = 5.02, P <0.0001). Thus, mean daytime GC of the overstory was 8.5 ± 4.5 mm s−1 before (May 24 to June 16) Changma compared to 4.5 ± 1.4 mm s−1 after (August) Changma (Table 3). Similarly, mean daytime GC of the understory decreased from 2.6 ± 1.0 mm s−1 to 1.9 ± 1.2 mm s−1 after Changma. The difference was, however, not statistically significant. Declining trend of GCref was also found after Changma, showing GCref of the overstory at 12.8 mm s−1 before Changma compared to 11.9 mm s−1 after Changma and GCref of the understory at 2.2 mm s−1 and 1.0 mm s−1 (Table 3.3). Daytime GC of the overstory exponentially and significantly declined with increasing daytime VPD, (R2 = 0.46, P = 0.0012 before Changma; R2 = 0.66, P <0.0001 after Changma), while the relationship between GC of the understory and VPD4 was not significant (Figure 3.6). The ratio of −m/GCref, which indicates the variations of −m (the sensitivity of GC response to VPD; see Eq. 9) to GCref, for the overstory was 0.66 and 0.68 before and after Changma, respectively, and that for the understory was 0.81 and 0.91. The observed – m/GCref ratio for the overstory was close to the theoretical ratio of 0.6 (Oren et al. 1999), while that for the understory was higher than the theoretical ratio (Table 3.3). Table 3.3 Mean values of daytime vapor pressure deficit (VPD, kPa), daily sum of photosynthetic active radiation (PAR, mol m−2 d−1), canopy conductance (GC, mm s−1), canopy conductance at VPD = 1 kPa (GCref, mm s−1) and the ratio of the sensitivity of GC response to VPD to GCref (−m/GCref) at the overstory and understory for the period before and after Changma. Numbers in paranthesis indicate standard deviation. Before Changma (May 24−June 16) After Changma (August 1−31) Over Under Over Under VPD 1.4 (± 0.5) 0.7 (± 0.4) 2.1 (± 0.5) 0.7 (± 0.3) PAR 42.4 (± 15.4) 2.3 (± 0.8) 39.9 (± 12.7) 2.7 (± 1.2) GC 8.5 (± 4.5) 2.6 (± 1.0) 4.5 (± 1.5) 1.9 (± 1.2) GCref 12.8 2.2 11.9 1.0 –m/GCref 0.66 0.68 0.81 0.91 Chapter 3 – Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use ! 80 Figure 3.6 Relationships between daytime canopy conductance (GC) and daytime vapor pressure deficit (VPD) for the overstory (closed) and the understory (opened) before Changma (June 1−June 16) (triangles) and after Changma (August 1 to August 30) (circles). 3.4 Discussion 3.4.1 Regulation of the overstory and understory transpiration The seasonal pattern of EO was synchronized to the seasonal development of leaf area until July, at the onset of the monsoon period. Similarly, patterns of EU early in the season reflected the pattern of LAIU development. However, this relationship became distorted with increasing development and closure of the overstory canopy. After full overstory canopy development, EU was mainly influenced by the prevailing environmental conditions in the understory. The understory canopy had an earlier bud break than the overstory canopy and the former attained its maximum leaf area earlier than the latter. As a result, maximum EU occurred in late April, coinciding with the highest leaf area, highest PAR and VPD in the understory, at a time when the overstory canopy was not yet fully developed. Maximum EO occurred in early June, coinciding with the peak overstory leaf area. Differences in timing of the canopy development of the understory and overstory have been reported for a similar forest type in Japan (Nasahara et al. 2008). In this temperate forest stand, the understory, therefore, accounts for most of the total leaf area early in the growing season, since overstory leaf development is delayed. These phenological differences between the two layers result into temporal shifts in functionalities, in terms of forest transpiration. Similar observations have been Chapter 3 – Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use ! 81 made by Unsworth et al. (2004), who reported shifts in the EO to EU ratio during the growing season due to the temporal differentiation in phenology and canopy development of the dominant species of the two layers. Responses of EO and EU to the corresponding PAR and VPD were similar throughout the vegetative period, except in April when the leaves in both layers were in their developing stages. The behavior observed in April is likely due to premature leaves and less developed stomata at this stage (Hiyama et al. 2005). Throughout most of the growing season, except in July, EO saturated on the days when PAR reached 40 to 60 mol m−2 d−1, which at most times corresponded to a daily VPDmax of 1.8 to 2.0 kPa. EU, however, never reached saturation likely due to the low light and VPD environment in the understory. Although PAR at the understory was slightly higher in August and September, due to canopy opening (decreased PAI) caused by the monsoon and typhoons, this change in PAR did not influence the magnitude of EU since this period was characterized by low VPD, low stomatal conductance and aging leaves. According to Kang et al. (2009b; 2010), decoupling coefficient (Ω), defined as a degree of decoupling between vegetation and the atmosphere, was, on average, equal to 0.39 for the overstory and 0.15 for the understory during the growing season at Gwangneung. This was comparable with Ω of 0.35 for the overstory and 0.23 for the understory in the temperate deciduous oak forest in Tennessee whose total contribution of transpiration to Eeco was similar with that of Gwangneung. Low Ω at both layers demonstrated a greater dependence of EO and EU on VPD than PAR. Another notable observation was the significant drop in the magnitudes of EO and EU during and after Changma. This was attributed to the strong decline in PAR and VPD (Figure 3.2), the decrease in PAI and the advanced leaf age. A decrease in stomatal conductance with increasing leaf age has been reported in other studies (Field and Mooney 1983; Radoglou 1996; Rey and Jarvis 1998). Consistent with this, after Changma decreased GC and GCref for both the overstoy and understory were observed in our study. Moreover, the ratio of –m/GCref of both layers was slightly increased after Changma, which suggests an increased stomatal sensitivity to VPD due to aging leaves. Greater –m/GCref of the understory, compared to the overstory and the theoretical ratio of 0.6 (Oren et al. 1999) might be because of smaller VPD range in the understory (0.6–1.5 kPa) (Herbst et al. 2008). Herbst et al. (2008) reported that there is a tendency of increasing the value of –m/GCref for low ranges of VPD. These changes in GC and stomatal sensitivity together with the altered environmental factors after Changma, may explain the decline in EO and EU after Changma. 3.4.2 Partitioning of ecosystem water use The maximum rate of EO observed in this forest stand was 1.9 mm d−1 (Figure 3.4). This value is within the range of maximum EO rates reported for similar forest types located within the same latitudinal range. For Chapter 3 – Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use ! 82 example, the maximum EO reported for a temperate deciduous forest stand in Duke, North Carolina, with PAI of 5.4, was 2.0 mm d−1 (Oren and Pataki 2001). Wullschleger et al. (2001) reported a value of EO of 2.2 mm d−1 in a temperate deciduous oak forest stand in Tennessee with a PAI of 6.2. On the other hand, the average EU was 0.1 ± 0.1 mm d−1, which contributed approximately 13% of the total stand transpiration amounting to 196 mm p.a. This proportion of EU to total stand transpiration is similar to those reported for other temperate forest stands. EU of a similar temperate oak forest stand accounted for 17% of total stand transpiration (Wullschleger et al. 2001). Baldocchi and Vogel (1996) reported 5 to 25% contribution of the understory to the total stand transpiration of a similar deciduous forest stand in North America. Our total stand transpiration was, however, approximately 30% lower compared to values measured in the temperate deciduous forests of North America (i.e., Wullschleger et al. 2001). Despite similarities in magnitudes of the annual mean RN and Ta and annual precipitation between the two sites, total stand transpiration is lower, in comparison, at our Asian study site because of differences in distribution patterns of these environmental drivers. In total, EO and EU contributed 47% and 8% of the annual Eeco, respectively. Consequently, the remaining 45% of Eeco originated from Eg and Ew. Previous studies conducted at the same forest stand estimated 16% contribution from Eg (Kang et al. 2009b) and 13~32% from Ew to annual Eeco (Kang et al. 2012). Kang et al. (2009b) reported that Eg was high in March and November when light availability at the forest floor was high, but was negligible during the growing season when light availability at the forest floor and VPD were low. Even though Ew reported by Kang et al. (2012) was highly variable, its relatively high contribution to Eeco reveals that Ew is a major component of the partitioning of ecosystem water use, especially during the growing season. The contribution of stand transpiration (EO + EU) to Eeco varied seasonally. Between April and May, stand transpiration accounted for 80 to 89% of the total water loss from the forest. This value declined to 58% in July and 42% in August (Table 3.2). The decline in July was primarily due to the decrease in EO in response to the declining PAR and VPD associated with Changma and PAI reduction due to heavy rainfall and strong winds. However, the reduced contribution of transpiration to Eeco in July and August may be due to the unaccounted amount of Ew, which was found to be highest in these months, corresponding with intensified rainfall (Kang et al. 2012). In September, the proportion increased to 74% as transpiration became the dominant contributor to Eeco. The partitioned components of Eeco in our study indicate the importance of the unaccounted contributions of Ew to Eeco of Gwangneung forest stand during the growing season. Chapter 3 – Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use ! 83 3.5 Conclusions The understory contributes significantly to the total annual forest water budget. Since bud break in the understory occurs almost one month earlier than the overstory canopy, the understory becomes the dominant source of transpiration water loss early in the season. Its dominance is, however, subdued with the development and closure of the overstory canopy, even though it still accounts for a significant proportion of the total forest PAI most of the year. The seasonal pattern and contribution of EO to Eeco, however, are strongly synchronized to the overstory canopy development. Despite the understory comprising of different plant species, its role in total forest water budget seems to be governed by processes of the overstory. Considering the daily maximum and mean rates of EO and EU and their contributions to Eeco, the warm Asian temperate deciduous forest is functionally (in terms of water use) similar to the temperate oak forests in North America. Total stand transpiration in the temperate forests in Asia, however, is lower in comparison, likely due to the depression in transpiration during the monsoon as a result of decline in canopy PAI, PAR and VPD, suggesting that ecosystem process (i.e., EO and EU) is decoupled from rainfall input. Independent estimates of EO and EU and their partitioning to Eeco in this study can improve our understanding of the seasonal dynamics of forest water use, with a holistic assessment and enhancement of the performance of forest hydrology models, by highlighting the need of incorporating separate canopy development stage of the overstory and understory, in the case, regions under the influence of monsoon climates. 3.6 Acknowledgements This study was carried out as part of the International Research Training Group TERRECO (GRK 1565/1) funded by the Deutsche Forschungsgemeinschaft (DFG) in cooperation with the University of Bayreuth, Germany and the Korean Research Foundation (KRF) at Kangwon National University, Chuncheon, S. Korea and a grant (code: 1-8-3) from Sustainable Water Resources Research Center of 21st Century Frontier Research Program. We would like to acknowledge the input from Ms. Margarete Wartinger of Plant Ecology Department, University of Bayreuth and students of Seoul National University and Kangwon National University for their support during fieldwork. 3.7 References Augusto L, Dupouey JL, Ranger J (2003) Effects of tree species on understory vegetation and environmental conditions in temperate forests. Ann For Sci 60:823−831 Chapter 3 – Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use ! 84 Baldocchi DD, Vogel C (1996) A comparative study of water vapor, energy and CO2 flux densities above and below a temperate broadleaf and a boreal pine forest, Tree Physiol 16:5−16 Barbier S, Gosselin F, Balandier P (2008) Influence of tree species on understory vegetation diversity and mechanisms involved–A critical review for temperate and boreal forests. For Ecol Manage 254:1−15 Barbour MM, Hunt JE, Walcroft AS, Rogers GND, McSeveny TM, Whitehead D (2005) Components of ecosystem evaporation in a temperate coniferous rainforest, with canopy transpiration scaled using sapwood density. New Phytol 165:549−558 Black TA, Kelliher FM (1989) Processes controlling understorey evapotranspiration. Philos. Trans R Soc London B 324:207−229 Campbell GS, Norman JM (1998) An introduction to environmental biophysics. Springer, New York Canham DC, Finzi AC, Pacala SW, Burbank DH (1994) Causes and consequences of resource heterogeneity in forests: interspecific variation in light transmission by canopy trees. Can J For Res 24:337−349 Cho YC, Shin HC, Kim SS, Lee CS (2007) Dynamics and Conservation of the Gwangneung National Forest in Central Korea: A National Model for Forest Restoration. J Plant Biol 50:615−625 Ewers BE, Oren R (2000) Analyses of assumptions and errors in the calculation of stomatal conductance from sap flux measurements. Tree Physiol 20:579−589 Field C, Mooney HA (1983) Leaf age and seasonal effects on light, water and nitrogen use efficiency in a California shrub. Oecologia 56:348−355 Granier A (1987) Evaluation of transpiration in a Douglas-fir stand by means of sap flow measurements. Tree Physiol 3:309−320 Grime VL, Sinclair FL (1999) Sources of error in stem heat balance sap flow measurements. Agric For Meteorol 94:103−121 Ham JM, Heilman JL, Lascano RJ (1990) Determination of sol water evaporation and transpiration from energy balance and stem flow measurements. Agric For Meteorol 52:287−301 Hatton TJ, Ruprecht J, George RJ (2003) Preclearing hydrology of the Western Australia wheatbelt: Target for the future. Plant Soil 257:341–356 Herbst M, Rosier PTW, Morecroft MD, Gowing DJ (2008) Comparative measurements of transpiration and canopy conductance in two mixed deciduous woodland differing in structure and species composition. Tree Physiol 28:959−970 Hiyama T, Kochi K, Kobayashi N, Sirisampan S (2005) Seasonal variation in stomatal conductance and physiological factors observed in a secondary warm-temperate forest. Ecol Res 20:333−346 Hong J, Kwon H, Lim JH, Byun YH, Lee J, Kim J (2009) Standardization of KoFlux eddy-covariance data processing. Kor J Agric For Meteorol 2:12−26 Chapter 3 – Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use ! 85 Im ES, Ahn JB, Kwon WT, Giorgi F (2008) Multi-decadal scenario simulation over Korea using a on-way doubled-nested regional climate model system. Part 2: future climate projection (2021-2050). Clim Dyn 30:239−254 Jackson PC, Cavelier J, Goldstein G, Meinzer FC, Holbrook NM (1995) Partitioning of water resources among plants of a lowland tropical forest. Oecologia 101:197−203 Jung EY, Otieno D, Lee B, Lim JH, Kang SK, Schmidt MWT, Tenhunen J (2011) Up-scaling to stand transpiration of an Asian temperate mixed-deciduous forest from sing tree sapflow measurements. Plant Ecol 212:383−395 Kang M, Park S, Kwon H, Choi HT, Choi YJ, Kim J (2009a) Evapotranspiration from a deciduous forest in a complex terrain and a heterogeneous farmland under monsoon climate. Asia-Pacific J Atmos Sci 45:175−191 Kang M, Kwon H, Lim JH, Kim J (2009b) Understory evapotranspiration measured by Eddy-Covariance in Gwangneung deciduous and coniferous forests. Kor J Agric For Meteorol 11:233−246 Kang M, Kwon H, Cheon JH, Kim J (2012) On estimating wet canopy evaporation from deciduous and coniferous forests from Asian monsoon climate, J Hydrometeorl 13:950−965 Kim J, Lee D, Hong JY (2006) HydroKorea and CarboKorea: cross-scale studies of ecohydrology and biogeochemistry in a heterogeneous and complex forest catchment of Korea. Ecol Res 21:881−889 Kim JH (2002) Community ecological view of the natural deciduous forest in Korea. In: Ecology of Korea, Bumwoo, Seoul Korea Meteorological Administration (2011) A white paper on Changma. Kosugi Y, Takanashi S, Tanaka H, Ohkubo S, Tani M, Yano M, Katayama T (2007) Evapotranspiration over a Japanese cypress forest. I. Eddy covariance fluxes and surface conductance characteristics for 3 years. J Hydrol 337:269−283 Köstner BMM, Schulze ED, Kelliher FM, Hollinger DY, Byers JN, Hunt JE, McSeveny TM, Meserth R, Weir PL (1992) Transpiration and canopy conductance in a pristine broad-leaved forest of Nothofagus: an analysis of xylem sap flow and eddy correlation measurements. Oecologia 91:350−359 Köstner B, Granier A, Cermák J (1998) Sapflow measurements in forest stands: methods and uncertainties. Ann Sci For 55:13−27 Kumagai T, Aoki S, Nagasawa H, Mabuchi T, Kubota K, Inoue S, Utsumi Y, Otsuki K (2005) Effects of tree-to-tree and radial variations on sap flow estimates of transpiration in Japanese cedar. Agric For Meteorol 135:110−116 Kwon H, Kim J, Hong J, Lim JH (2010) Influence of the Asian monsoon on net ecosystem carbon exchange in two major ecosystems in Korea. Biogeoscience 7:1493−1504 Chapter 3 – Water use by a warm-temperate deciduous forest under the influence of the Asian monsoon: Contributions of the overstory and understory to forest water use ! 86 Landsberg JJ, Gower ST (1997) Applications of Physiological Ecology to Forest Management. Academic Press, San Diego Latham RE, Ricklefs RE (1993) Global patterns of tree species richness in moist forests: energy-diversity theory does not account for variation in species richness. Oikos 67:325−333 Lau K-M, Li M-T (1984) The monsoon of E. Asia and its global associations−A survey. Amer Met Soc 65:112−125 Lei H, Zhang ZS, Li XR (2010) Sap flow of Artemisia ordosica and the influence of environmental factors in a revegetated desert area: Tengger Desert, China. Hydrol Proc 24:1248−1253 Lieffers VJ, Messier C, Stadt KJ, Gendron F, Comeau PG (1999) Predicting and managing light in the understory of boreal forests. Can J For Res 29:796−811 Lim JH, Shin JH, Jin GJ, Chun JH, Oh JS (2003) Forest stand structure, site characteristics and carbon budget of the Kwangneung natural forest in Korea. Kor J Agric For Meteorol 5:101−109 Magnani F, Leonardi S, Tognetti R, Grace J, Borghetti M (1998) Modeling the surface conductance of a broad-leaf canopy: effects of partial decoupling from the atmosphere. Plant Cell Environ 21:867−879 Meinzer FC (2003) Functional convergence in plant responses to the environment. Oecologia 134:1−11 Meinzer FC, Bond BJ, Warren JM, Woodruff DR (2005) Does water transport scale universally with tree size? Funct Ecol 19:558−565 Monteith JL (1965) Evaporation and environment. Symp Soc Expt Biol 19:205–234 MOST (1999) Geologic map of Seoul-Namchon report. Ministry of Science and Technology, Korea Murray FW (1967) On the computation of saturation vapor pressure. J Appl Meteor 6:203−204 Nakashizuka T, Iida S (1995) Composition, dynamics and disturbance regime of temperate deciduous forests in Monsoon Asia. Vegetatio 121:23−30 Nasahara KN, Muraoka H, Nagai S, Mikami H (2008) Vertical integration of leaf area index in Japanese deciduous broad-leaved forest. Agric For Meteorol 148:1136−1146 Oh JS, Shin JH, Lim JH (2000) Long-term ecological research programme in Korea Forest Institute. Kor J Ecol 23:131−134 Oren R, Phillips N, Katul G, Ewers B, Pataki E (1998) Scaling xylem sap flux and soil water balance and calculating variance: A method for partitioning water flux in forests. Ann Sci For 55:191−216 Oren R, Sperry JS, Katul GG, Pataki DE, Ewers BE, Phillips N, Schäfer KVR (1999) Survey and synthesis of intraand interspecific variation in stomatal sensitivity to vapour pressure deficit. Plant Cell Environ 22:1515−1526 Oren R, Pataki DE (2001) Transpiration in response to variation in microclimate and soil moisture in southeastern deciduous forests. Oecologia 127:549−559 Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 93 Table 4.1 Locations, sizes, geological traits, soil characteristics and structural characteristics of the study sites. Trees with diameter at breast height (DBH) ≥1.0 cm were considered for calculating basal area (BA), tree density, mean DBH, and stand sapwood area (AS). ± indicates standard deviation (SD). 450N 650N 650S 950N Coordinates 128°7’50.09’’E 38°17’18.64’’N 128°8’26.07’’E 38°18’56.83’’N 128°8’27.13’’E 38°18’57.07’’N 128°6’0.86’’E 38°14’43.37’’N Elevation [m a.s.l.] 450 650 650 950 Plot area [m2] 575 750 325 200 Inclination [°] 20 23 15 21 Exposure Southeast Southeast Northwest Southeast Bedrock Granite Granite Granite Granitic gneiss Soil texture Loam Sandy-loam Sandy-loam Sandy (surface) Loam Soil depth [cm] 8–100 18–68 19–65 22–100 Soil N contents [mg g-1] 1.6 ± 0.9 5.0 ± 2.3 5.0 ± 2.3 1.7 ± 0.7 Stand age [yrs] ca. 30 ca. 30 ca. 30 ca. 20 BA [m2 ha-1] 21.2 20.7 25.9 22.3 Tree density [Tree m-2] 0.4 0.2 0.5 1.4 Mean DBH [cm] 5.6 ± 5.8 9.4 ± 5.3 6.6 ± 4.4 4.2 ± 1.6 Stand AS [m2 ha-1] 13.7 13.3 17 15.4 Canopy height [m] 10 12 10 5 Species composition (BA cover [%]) Quercus mongolica (24) Alnus sibirica (18) Q. aliena (15) Q. serrata (15) Ulmus laciniata (10) Q. dentate (8) Tilia mandshurica (6) Q. dentate (65) Betula davurica (19) Q. mongolica (14) Q. mongolica (50) T. mandshurica (25) Q. dentate (14) Fraxinus rhynchophylla (4) Q. serrata (4) Q. mongolica (72) F. rhynchophylla (13) Euonymus hamiltonianus (9) Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 94 4.2.2 Measurement of abiotic factors 4.2.2.1 Micrometeorology Solar radiation (RS), air temperature (Ta), rainfall, relative humidity, and wind speed were measured every 30 seconds, averaged and logged every 30 minutes by automatic weather stations (AWS; WS-GP1, Delta-T Devices, Cambridge, UK) installed at 2 m above the ground. AWS at 450 m (450 AWS) and 950 m (950 AWS) were installed in the open space next to the respective. AWS at 650 m (650 AWS) was located in an open space between 650N and 650S. Additional air humidity and temperature sensors (Funky Clima, ESYS GMBH, Berlin, Germany) were installed within the crowns at each site and half-hourly averages of fiveminute data recorded. Vapor pressure deficit (D) of each site was derived from measured TA and relative humidity (Murray 1967). 4.2.2.2 Soil water content and soil water retention Soil water content (θ) in 30 cm depth was measured at each site using soil moisture sensors (5TE, Decagon Devices, Washington, USA) and data logged every 30 minutes (EM50 Data logger, Decagon Devices) over the period of sap flow measurements. Relative θ was determined at each site as daily mean θ divided by maximum value of daily mean θ, allowing for a better comparison of soils with different textures. To analyze soil texture and bulk density, three soil samples were collected from each study site using a soil corer and a bulk density sampler, respecttively. A and B horizons of the soils were separately analyzed. Humus was eliminated by H2O2. Sand, silt and clay contents were determined by wet sieving for sand and laser particle analyzer (Mastersizer S MAM5004, Malvern Instruments, Herrenberg, Germany) for silt and clay in the Soil Physics laboratory, University of Bayreuth. Based on the bulk density and soil texture data, soil hydraulic parameters for each site, such as residual water content (θr), saturated water content (θs) and empirical shape parameters (α and n) were estimated using a computer program RETC−Retention Curve Program (PC−Progress, Prague, Czech Republic) and then soil water retention curves were determined following van Genuchten function (van Genuchten 1980, Schaap et al. 2001). From these retention curves soil water retention levels, corresponding to measured daily mean θ, was read out and also wilting point (θw) and field capacity (θc) for each site determined, and then applied to gap−filling of canopy transpiration (EC; see the section 4.2.4). Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 95 4.2.3 Measurement of biotic factors 4.2.3.1 Biometric data Monthly leaf area index (LAI) at each site was measured once a month with a plant canopy analyzer (LAI2000, LI-COR Inc., Lincoln, USA) under diffuse light conditions at fixed 9 sampling points with 10 m × 10 m grid interval in order to determine the seasonal changes of LAI. The maximum leaf area index (LAImax) was estimated from leaf litter collected with 0.5 m × 0.5 m litter traps (five in each site) randomly placed at ca. 1 m height above the forest floor in order to calibrate the values of LAI measured by LAI-2000. Leaf litter was collected monthly, transferred to the laboratory and sorted according to species. A sub-sample of leaf litter was measured for leaf area (LI-3100, LI-COR Inc.), dried for 48 hours at 75°C and weighed. Specific leaf area was determined from the ratio of leaf area/dry mass (cm2 g−1). L of each species (AL) from each site was computed from dry leaf weight multiplied by specific leaf area, summed over the season and divided by the area of the litter trap. In July 2010, a survey of all stems larger than 2 cm in diameter at 1 m height was carried out on 25 m x 25 m grids. Based on this survey, mean DBH, basal area (BA, m2 ha−1) and tree density (Tree m−2) were calculated for each plot. Sapwood area (AS, m2) of sample trees was calculated from the sapwood depths at sap flow sensor heights, determined from tree cores extracted at the end of the sap flow measurements. Sapwood was identified by dying the core samples, using bromocresol green (Sigma Chemicals, Germany) (Burrows 1980). Data from the sample trees were used to build an allometric function from which AS of all the trees in the stand were estimated (Vertessy et al. 1995, Meinzer et al. 2005): € AS= α DBH β (1) where α is a constant and β is the allometric scaling exponent. Using this empirical relationship of each site, stand AS (m2 ha−1), i.e., total AS of all trees in the plot (Ast) per ground area (AG), was estimated on the basis of the stem survey. Allometric relationships for the studied species (Querucs spp., A. sibirica, B. davurica, and T. mandshurica) are shown in Figure 4.1. Three respective regressions in the form of power function were built separately for Quercus spp. alone (ring porous; AS = 0.7642DBH1.8057, n = 31, R2 = 0.89, P <0.0001), for the rest of the species without Quercus spp. (diffuse porous; As = 0.8838DBH1.8905, n = 13, R2 = 0.98, P <0.0001), and for all the species together (AS = 0.5974DBH1.9374, n = 44, R2 = 0.84, P <0.0001). Total stand AS estimated from separate regressions (Quercus spp. alone vs. the rest of the species) were similar or slightly larger than estimates from a combined species regression (i.e., the difference was less than 1% at 450N and 650N, 5% at 650S, and 10% at 950N). We, therefore, chose the general regression for further analyses related to AS. Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 96 After determination of total AL and total AS of each species from each site, species-specific ALAS −1 was calculated for all studied species from each site. The tree cores extracted for the determination of the AS were brought to the lab and ring widths were visually measured using a stereo microscope. Figure 4.1 Relationship between stem diameter at breast height (DBH, cm) and sapwood area (AS, cm2) for Quercus species (open symbols) and for the rest of the studied species (close symbols) from all study sites. Black solid line is the regression for Quercus spp.: AS = 0.7642DBH1.8057 (n = 31, R2 = 0.89, P <0.0001), gray solid line is the regression for the rest of the species: AS = 0.8838DBH1.8905 (n = 13, R2 = 0.98, P <0.0001), and broken line is a regression of all species: AS = 0.5974DBH1.9374 (n = 44, R2 = 0.84, P <0.0001). Letters of Q, As, Bd, Tm indicate Quercus spp., Alnus sibirica, Betula davurica, Tilia mandshurica, respectively. Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 97 4.2.3.2 Sap flux density and transpiration To estimate canopy transpiration (EC), dominant and sub-dominant tree species covering more than 80% of BA at each plot were selected as sample trees of sap flow measurements. Sample size per species was decided based on the species dominance in each plot. Trees with larger than 7 cm diameter at breast height (DBH), were chosen in order to apply sap flow probes. Detailed information of sample trees including species and number of individuals are indicated in Table 4.2. Sap flux density was monitored with 20 mm long thermal dissipation probes heated at constant power supply. The probes were constructed at the technical laboratory, Department of Plant Ecology, University of Bayreuth, based on the original design of Granier (1987). The probes were inserted into 0−20 mm of the sapwood in all the sample trees. Depending on the determined sapwood depth (d), additional sensors were installed deeper into 20−40 mm or 40−60 mm of the sapwood in order to cover the most of the sapwood. Sensors were installed at ca. 1.3 m height above the ground and on the north-facing side (azimuth) of the trees to minimize direct solar heating (Wilson et al. 2001; Wullschleger et al. 2001). Further, each probe was covered with Styrofoam sheaths and aluminum foil to minimize direct thermal load from the sun. The temperature differences (ΔT) between the heated and reference probes aligned vertically 10 cm apart in the sapwood was recorded, and by comparing ΔT to the maximum temperature difference (ΔTmax) occurring at predawn when there is no sap flow, sap flow density (Fd, g m−2 s−1) was calculated according to Granier (1987): Fd=119 (ΔTmax − ΔT) ΔT # $ %& ' ( 1.231 (2) Natural ∆T without heating was negligible since all measurements in the present study were carried out in forest stands with relatively closed canopies. The Clearwater-correction (Clearwater et al. 1999) was applied to Js of ring-porous Quercus with less than 20 mm sapwood depth. Azimuth variation in Js was not considered during EC estimates, based on the findings of our previous study (Jung et al. 2011), which showed no significant differences in Js in different azimuths of the tree trunk. Tree water use (TWU, kg h−1) of the individual trees was computed as: TWU =Fdt As (3) Fdt =(Fdi Asi ) ∑ Asi ∑ (4) where Fdt is sapwood area weighted sap flow density, Fdi is sap flow density of annulus i, and Asi is sapwood area of annulus i. Canopy transpiration (EC, mm h−1) was calculated as: Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 98 EC=Fdt Ast AG −1 (5) where Fdt is the mean Fdt of the sample trees. Measurements were conducted between May and October 2010, during the vegetation period. Fd data were sampled every 30 seconds, averaged and recorded every 30 minutes (DL2, Delta-T Devices, Cambridge, UK). The period with active transpiration was regarded as the vegetative/growth period. 4.2.3.3 Canopy conductance Canopy conductance (GC, mm s−1) is estimated from transpiration per unit leaf area (EL, mm s−1) scaled from Fdt (Monteith and Unsworth 1990) as: GC=KGEL VPD (6) where KG is the conductance coefficient as a function of Ta (115.8 + 0.4236Ta, kPa m3 kg−1) accounting for temperature effects on the psychrometric constant, latent heat of vaporization, specific heat of air at constant pressure and the density of air (Phillips and Oren 1998). EL (mm s−1) is transpiration per unit leaf area determined by: EL=Fdt AS AL (7) This simplification of the Penman-Monteith equation is based on the assumption that forests are well coupled aerodynamically, when leaves are exposed to sufficiently high wind speeds. Thus, VPD can be used as an approximation of the total driving force for transpiration. We tested the assumption of strong coupling in all the studied sites by comparing aerodynamic conductance (GA) and GC for 8 days from 1–8 June. GC reached ca. 1% of GA, agreeing with the assumption of Eq. 6. To assess stomatal sensitivity to VPD at each site, a modified Lohammar’s function was applied. GC(VPD)=GCref −mlnVPD (8) where GCref is canopy conductance at VPD = 1 kPa and −m (i.e., –ΔGC/ΔlnVPD) is the sensitivity of GC response to VPD (Oren et al. 1999). Stomatal sensitivity analysis was limited to GC under conditions in which VPD ≥0.6 kPa in order to minimize the uncertainties of GC estimates (Ewers and Oren 2000). Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 99 Table 4.2 Species, number (n), ranges of diameter at breast height (DBH, cm), sapwood depth (d, mm), maximum sap flow density weighted by sapwood area of each tree (max Fdt, g m−2 s−1) and maximum tree water use (max TWU, kg d−1) of all the sample trees with sap flow sensors at each site. Quercus mongolica Q. dentata Q. serrata Q. aliena Alnus sibirica Betula davurica Tilia mandshurica 450N n 5 3 1 1 3 - - DBH 14.7–25.3 13.6–16.3 12.0 26.0 11.5–25.5 - - d 22–40 23–33 20 36 40–70 max Fdt 14.7–28.5 13.3–22.0 48.3 45.2 40.7–88.2 - - max TWU 8.6–40.9 6.6–8.5 13.3 34.5 18.1–59.1 - - 650N n 4 5 - - - 5 1 DBH 8.9–14.4 15.8–18.2 - - - 14.0–22.9 23.0 d 16–24 18–25 42–60 58 max Fdt 15.2–21.8 11.2–15.7 - - - 33.1–64.2 36.1 max TWU 2.7–6.6 3.9–5.5 - - - 17.6–42.6 10.4 650S n 5 1 - - - - 4 DBH 9.9–17.8 14.3 - - - - 10.4–22.7 d 20–37 20 40–60 max Fdt 19.4–26.5 13.6 - - - - 20.4–55.6 max TWU 2.8–9.1 3.3 - - - - 4.3–58.6 950N n 6 - - - - - - DBH 7.3–9.8 - - - - - - d 15–23 max Fdt 15.7–41.2 - - - - - - max TWU 2.4–8.2 - - - - - - Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 100 4.2.3.4 Gap-filling A modified Jarvis-Stewart model as defined by Whitley et al. (2009) was used to estimate EC for the period when data gap occurred. Whitley et al. (2008) and Whitley et al. (2009) expressed EC in the same way as GC, as defined by Jarvis (1976) and Stewart (1988). EC=ECmax f1(RS)f2(VPD)f3( θ )f4(LAI) (9) The functions € fi , which take on values between 0 and 1, are a set of scaling terms reducing a maximum stand transpiration (ECmax, mm d−1) in response to changes in RS, VPD and θ. Daily estimates of EC were determined by the functions fi using the optimal estimates of parameters. The functions of RS, VPD and θ were taken from Whitley et al. (2008) based on those of Stewart (1988), Wright et al. (1995) and Harris et al. (2004). A function describing a radiation response is: € f1(RS)=RS 30 " # $ % & ' 30 +k1 RS+k1 " # $ % & ' (10) where k1 is an empirical coefficient describing the curvature of the relationship. It shows an asymptotic function saturating at approximately 30 MJ m−2 d−1. A functional response of EC to VPD was expressed as: f2(VPD)=k2VPDexp(−k3VPD) (11) where k2 and k3 are the parameters describing the rate of change at low and high VPD. This function of VPD for Ec follows Boltzmann distribution. A function of soil moisture response was described to be a three-phase relationship as: € f3( θ )= # $ % & % 0, θ < θ w θ − θ w θ c− θ w , θ w< θ < θ c 1, θ > θ c (12) where θw and θc are wilting point and field capacity of each site, respectively. A function of LAI response to EC was described as: f4(LAI)=LAI LAImax (13) A model parameterization was performed using measured data on daily basis via nonlinear least squares analysis. To avoid errors of division by zero and conditions of wet canopy, only the data between 8:00h and 18:00h were included, and the data on the rainy days were excluded. The data from each study site were partitioned into two separate sets of random days in order not to use same data for both parameterizations and validation of the model. Root-mean-square-error (RMSE) and an agreement index (d), developed by Willmott (1984) were used to evaluate the agreement between the predicted EC and the observed EC. The Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 101 ideal model would give RMSE = 0 and d = 1. Total gap-filled period of each site was 27 days for 450N, 54 days for 650N, 68 days for 650S, and 8 days for 950N. The gap-filled data were only used for quantification of annual EC. Annual EC was estimated by summing up the measured and simulated EC as Eq. 9–Eq. 13. Table 3 contains the best estimates of parameters (k1, k2, and k3) along with their respective standard errors. Parameters of k2 and k3 were found to be statistically significant (P <0.001), but k1 was not significant (P >0.1), with large standard errors. RMSE was less than 0.2 and d was larger than 0.9 for all study sites (Table 4.3). Table 4.3 Optimal estimates of Jarvis-Stewart model parameters (k1, k2, k3) and statistical parameters for error assessment such as root-mean square-error (RMSE) and index of agreement (d) for all sites. Standard errors are given brackets next to each value. 450N 650N 650S 950N k1 0.57 (1.63) 2.08 (3.81) 0.56 (1.24) 0.92 (3.92) k2 1.92 (0.21) 3.75 (0.65) 2.43 (0.24) 2.75 (0.45) k3 0.68 (0.07) 0.94 (0.12) 0.79 (0.08) 0.99 (0.14) RMSE 0.16 0.21 0.12 0.18 d 0.96 0.91 0.97 0.96 4.2.3.5 Water use efficiency and leaf nitrogen contents From each site, five sunand five shade-leaves each from five Q. mongolica canopy trees were collected on 24 June 2010, during mid season. The samples were oven-dried at 75°C for 48 hours and then ball-milled before subjected to 13C/12C isotopic ratio analysis at BayCEER – Laboratory of Isotope Biogeochemistry, Germany. Analyses were conducted with an elemental analyzer NA 1108 (CE Instruments, Milan, Italy) coupled to an isotope ratio mass spectrometer delta S (Finnigan MAT, Bremen, Germany) via an open split interface ConFlo III (Finnigan MAT, Brement, Germany) as described by Bidartondo et al. (2004). Standard CO2 gas was calibrated with respect to international standard (CO2 in Pee Dee Belemnite) by use of the reference substance NBS 16 to 20 for carbon isotopic ratio provided by the international Atomic Energy Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 102 Agency IAEA, Vienna, Austria. The 13C/12C isotopic ratios, denoted as delta values were calculated according to the equation δ 13C=Rsample Rstd −1 " # $% & '×1000 (14) where δ13C is the isotope ratio of carbon in delta units relative to the PDB standard. Rsample and Rstd are the 13C/12C of the samples and the PDB standard, respectively. δ13C was used as an index of seasonally integrated water use efficiency (WUE) (Tieszman and Archer 1990). 4.2.3.6 Statistical analysis Statistical analyses including linear regression, ANOVA, and Tukey HSD, as well as model parameterizations were conducted with R (R development Core Team, 2010). Nonlinear curve fits were performed using Sigma Plot (Version 11, SPSS, San Rafael, CA). 4.3 Results 4.3.1 Elevation effects on microclimate Annual mean solar radiation (RS), daily air temperature (TA), air humidity and rainfall differed among sites (Table 4). The annual mean RS was higher (ANOVA, P <0.0001) at 450 m > 950 m > 650 m. The annual mean TA and the mean annual daytime D were higher at 450 m > 650 m > 950 m. Thus, TA and daytime D decreased with increasing elevation. Differences in TA and daytime D among the sites were significant (ANOVA, P <0.001). Increasing annual rainfall was observed with increasing elevation. In other words, 950m received higher amount of rainfall than lower elevations, corresponding to higher rainfall intensities and frequencies through the year. Mean wind speed at 950 m was ca. 4.5 m s−1, which was about twice as high as at 450 m (2.6 m s−1). Although absolute values of RS, Ta, VPD, and rainfall were different among sites, their seasonal trends were similar across the sites (Figure 4.2). RS increased during spring and reached its maximum in June. A strong decline in RS occurred in July and August, during with the monsoon period. A second peak in RS occurred in September, after the monsoon, but this peak was lower than the pre-monsoon value. RS significantly dropped during fall and winter. The minimum and maximum Ta were recorded in January and August, respectively. VPD increased steadily during spring and reached its maximum in June, but declined to 20% of its maximum between July and August, during the monsoon. A slight increase in VPD occurred in Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 109 Figure 4.5 Seasonal patterns of daily canopy transpiration (EC, mm d−1) at each study site in 2010. Closed circles are measured and open circles are simulated values using Jarvis-Stewart model (Eq. 11). The shaded area indicates the period of the monsoon in Korea, 2010. Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 110 Figure 4.6 Relationship between vapor pressure deficit (VPD, kPa) and daily canopy transpiration (EC, mm d−1) at each site from June to September 2010. Fitted functions were (a) EC = 1.70VPD(1−e(−2.21VPD)) for 450N, (b) EC = 1.30VPD(1−e(−2.28VPD)) for 650N, (c) EC = 1.48VPD(1−e(−1.58VPD)) for 650S, (d) EC = 1.48VPD(1−e(−5.71VPD)) for 950N. All the estimated parameters of each regression were statistically significant (P <0.0001). 4.3.4 Elevation effects on canopy conductance Canopy conductance (GC) is a measure of the intensity of EC regulation among forests. Mean daytime GC on clear, sunny days between June (mature canopy) was similar at lower sites, i.e., 450N, 650N, and 650S, which was about 4 mm s−1 (TukeyHSD, Padj >0.8), while GC at the 950N was significantly higher than the other sites (ANOVA, F = 4.568, P = 0.003; Table 4.6). GCref, which denotes GC at VPD = 1 kPa, was 5.4 ± 0.8 mm s−1 at 450N, 3.6 ± 0.4 mm s−1 at 650N, 3.7 ± 0.5 mm s−1 at 650S, and 4.5 ± 1.5 mm s−1 at 950N. GCref differed significantly (ANOVA, F = 24.52, P <0.0001) among the sites except between 650N and 650S (TukeyHSD, Padj = 0.71). Thus, at the same range of VPD, GC was higher at 450 > 950 > 650 m elevations. The response of GC to VPD under saturating RS (>400 W m−2) at each site is plotted in Figure 4.7. The Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 111 gradient of the exponential regression curve relating GC to VPD was significantly steeper for the 950N compared to the other sites. Figure 4.7 Relationships between canopy conductance (GC, mm s−1) and vapor pressure deficit (VPD, kPa) under saturating global radiation (RS >400 W m−2) at each study site. All the estimated parameters of each regression were statistically significant (P <0.0001). Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 112 Table 4.6 Canopy conductance (GC, mm s−1) on clear days in June, its respective vapor pressure deficit (VPD, kPa) for daytime (from 8:00h to 18:00h) and GCref (GC at VPD = 1 kPa) at each site. ± are standard deviation (SD). 450N 650N 650S 950N GC 4.4 ± 3.0 4.1 ± 4.1 4.3 ± 3.1 5.3 ± 6.2 VPD 1.4 ± 0.7 1.4 ± 0.8 1.3 ± 0.7 1.1 ± 0.5 GCref 5.4 ± 0.8 3.6 ± 0.4 3.7 ± 0.5 4.5 ± 1.5 In Figure 4.8, the values of sensitivity of GC response to VPD (−m = –ΔGC/ΔlnVPD, see Eq. 8) were plotted against GCref at each site. The average of the empirical slopes (GCref to –ΔGC/ΔlnVPD) to for all the sample trees at 450N, 650N, 650S, and 950N were 0.64, 0.65, 0.67, and 0.84, respectively. The slope for the 950N site was significantly higher than those at the other sites (ANOVA, F = 24.33, P <0.0001; TukeyHSD, Padj <0.0001). Thus, the slopes for the 450N, 650N, and 650S were close to the universal slope of 0.6 (Oren et al. 1999), while it was 25% higher at the 950N, which indicates higher stomatal sensitivity to VPD at this site/elevation. Figure 4.8 The response of canopy conductance to vapor pressure deficit (–ΔGC/ΔlnVPD, see Eq. 10) plotted against the reference conductance (GCref) for all studied trees at each site, i.e., Quercus spp. (Q), Alnus sibirica (As), Betula davurica (Bd), and Tilia mandshurica (Tm). The universal ratio of 0.6 suggested by Oren et al. (1999) is indicated by a broken line. Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 113 4.4 Discussion 4.4.1 Interactions among elevation, abiotic factors and tree growth parameters A gradient along the elevation was found for Ta, rainfall, and VPD, while RS, wind speed and relative θ showed no tendency. Low elevations are likely to have higher Ta and less rainfall (low humidity), which result in higher VPD (Tromp-van Meerveld and McDonnell 2006). Kubota et al. (2005) reported a decreasing VPD at higher elevations due to height-dependant decrease in Ta. In our study, the annual mean daytime VPD at the 450N was 0.5 kPa >0.4 kPa at 650N >0.2 kPa at 950N, corresponding to higher annual mean Ta and lower annual rainfall at lower elevations. We observed significant and negative linear relationships between elevation and annual daytime mean D, annual mean Ta, and annual rainfall (R2 = 0.98 for VPD, 0.99 for elevation, and Ta, and 0.95 for rainfall; P <0.0001). The amount of rainfall, soil characteristics (mainly infiltration rates), and soil depths interact to determine the amount of water available for plants (Chapin et al. 2002). Studies in rugged mountainous terrains report contrasting patterns of soil moisture availability along the elevational gradient. For example, Kumagai et al. (2008) observed that the upper slope soils were constantly drier than soils down slope in a Japanese cedar forest. On the other hand, Kubota et al. (2005) reported increasing soil water content with increasing elevation as a result of increasing rainfall at higher elevations. In our study, despite differences in soil depths and precipitation amounts at the three different elevations, there were no significant differences in soil water retention among the sites. In most cases, soil moisture was close to field capacity at all sites (Figure 4.1), as a result of high rainfall amounts in Haean region. The forests studied, therefore, never experienced water stress during the study period and any differences in water use among the forest sites cannot be attributed to soil moisture availability. A correlation between TWU and DBH (Figure 4.3) and that between AS and DBH (Figure 4.2) among tree species and across sites in different elevations provides a link between tree allometry and tree water use (Vertassy et al. 1995; Bucchi et al. 2004; Meinzer et al. 2001; Meinzer et al. 2005; Gebauer et al. 2008; Jung et al. 2011), demonstrating the deterministic role of xylem water transport on overall forest water use as well as the dependence of both parameters on tree size. We observed a nonlinear increase in AS with increasing DBH, which was detached from elevation since DBH sizes were not defined by elevation. Similar universal functional relationships between DBH and AS have been reported for 24 co-occurring canopy tree species in tropical forests (Meinzer et al. 2001; Meinzer et al. 2005) and for five tree species growing in an Asian temperate forest (Jung et al. 2011). However, Gebauer et al. (2008) reported considerably higher AS (about 80% of stem cross-sectional area) in diffuse porous trees compared to 20% in ring porous trees. In our case, although ring and diffuse porous trees were distinguishable according to DBH:AS ratios, a universal regression comparing all the trees together was statistically significant (Figure 4.1), suggesting that the Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 114 species basically function in a similar manner in terms of water use (see Figure 4.3). Our results compare favorably with those of Meinzer et al. (2005), who compared 18 angiosperm species. Based on these results, it was clear that comparisons of water use and its regulation among the respective forest stands at different elevations could be performed without considering species composition. This was a departure from our initial hypothesis that differences in species composition may mask differences arising from elevation. In temperate forests, the time during which active leaf transpiration occurs corresponds to the productive period of the forest (Körner 2007) and has a strong influence on the total water use by forest ecosystems. The total period of transpiration decreased with increasing elevation and decreasing Ta, i.e., 4 days 100 m−1 and 6 days °C−1, which was comparable with the results from the similar type of temperate deciduous forests in Europe and North America (3–5 days 100 m−1 and 7–13 days °C−1) (Rotzer and Chmielewski 2001; Dittermar and Elling 2006; Richardson et al. 2006; Vitasse et al. 2009), even though the range of elevational gradient was relatively narrow (450 m to 950 m) in our case. This gradient in duration of active tree transpiration resulted from differences in timing of leaf onset and senescence at the different elevations, which is likely as a result of Ta variations among the sites. The vegetation at higher elevations experienced delayed leaf flush due to extended low winter temperatures and an earlier onset of leaf senescence as a result of rapid cooling in autumn, at higher elevations. These temperature changes along the elevational gradient also influence tree productivity and growth, as demonstrated by the relationship between annual tree ring width and mean annual temperature of the respective forest sites. This qualifies Ta as an important determinant of tree growth and functioning on mountain slopes. 4.4.2 Elevation effects on canopy transpiration and its regulation The annual EC declined from 175 mm year−1 at 450N to 90 mm year−1 at 950N. This trend of forest canopy transpiration is consistent with previous findings of McDowell et al. (2008) who reported a decline in EC with increasing elevation in forests dominated by different coniferous species in the southern Rocky Mountains. Matyssek et al. (2009), however, found no relationship between the magnitude of EC and elevation in mixed or coniferous forest stands at the collinear, mountainous, and subalpine elevations. Both McDowell et al. (2008) and Matyssek et al. (2009) compared EC from different forest types or forest stands dominated by different species, which makes it difficult to separate the impact of elevation. Our study addressed similar forest types dominated by oaks. This made it possible to compare EC from the different forest stands and key out elevation effects on forest water use. Although Kubota et al. (2005) studied elevation effects on JS of mountainous beech forests in Japan, to the best of our knowledge, no such study in annual EC regarding to elevational gradients has been conducted in similar forest types in Asia. Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 115 We established a functional relationship between EC and Ta as EC = 29.56Ta – 85.19 with R2 = 0.90 and P <0.0001 for the annual mean Ta and EC = 35.67Ta – 462.7 with R2 = 0.97 and P <0.0001 for the growing season Ta. Based on these regressions, transpiration is zero when the annual mean Ta and a growing season’s mean Ta drop down to ~3°C and ~13°C, respectively. Körner (2003; 2007) suggested that a strong relationship between growing season length and mean Ta contributes to the reduction in annual EC at high elevations especially in humid, temperate forests where soil moisture is not limiting. Pfautsch et al. (2010) reported a strong linear relationship between daily mean Fdt and daily Ta, indicating a temperaturedependence of water use in Eucalyptus regnans forests in southeastern Australia. In our study, however, we did not observe any direct temperature-dependence of EC on a daily scale, but on an annual scale. On a daily basis, VPD was the dominant determinant of transpiration (Figure 4.6). Similar observation was reported by Jung et al. (2011) for the deciduous forest in Korea. Our daily Ta and VPD were not correlated, likely due to the rapidly changing humidity at our study sites. This may explain the lack of a relationship between Ta and EC on a daily basis, since Ta lags behind. On an annual basis, however, a significant linear relationship between Ta and VPD (R2 = 0.99, P <0.0001) was found. Water transpired by forests is determined by the stomata. Stomatal functioning is influenced by VPD (Kelliher et al. 1997; Saugier et al. 1997), wind speeds (Campbell-Clause 1998; Schulze et al. 2005) and soil moisture (Sala and Tenhunen 1996; Kelliher et al. 1997; Tognetti et al. 2009). In our study, VPD was the dominant determinant of GC and also EC at all the sites. The similarities in the responses of daily EC to VPD (Figure 4.6) as well as GC to VPD (Figure 4.7) among the 450N, 650N, and 650S suggest similarities in stomatal functioning among the sites. Differences in EC at 950N can be attributed to differences in the response of GC to prevailing VPD at the highest elevation (Figure 4.7). A departure from this trend observed for trees at the 950N, which was characterized by a conspicuous depression in EC between 10:00h and 13:00h (Figure 4.4). An early drop of EC at relatively low VPD (Figure 4.6) and a steeper slope of the curve between VPD and GC (Figure 4.7) were likely due to high wind speeds, since this site was more exposed. Campbell-Clause (1998) reported decreasing transpiration rates at wind speeds more than 4 m s−1. Moreover, significant differences in the response of GC to VPD occurred among sites at different elevations with the 950N showing higher (0.83) stomatal sensitivity to changes in VPD compared to the other study sites (0.63–0.65) (Figure 4.8). The greater sensitivity caused the higher GC at 950N around VPD = 0.6–1.0 kPa, resulting in higher Fdt and hence higher EC (in 30min), but the lower GC at 950N at VPD >1.0 kPa, driving a drop in Fdt and EC, consequently. Such adjustments to local conditions may have implications for the overall forest stand water use. Furthermore, several studies reported that the intensity of exponentially decreasing GC in relation to increasing VPD might vary among species (McNaughton and Jarvis 1991; Oren et al. 1999; Oren and Pataki 2001; Herbst et al. 2008), masking differences arising from elevation. For example, higher stomatal Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 116 sensitivity to VPD has been reported for diffuse porous (Acer rubrum) compared to ring porous species (Quercus alba) (Oren and Pataki 2001). A dissimilar result was found by Herbst et al. (2008) that the forest with higher proportion of ring-porous trees showed higher stomatal sensitivity. In this study, we did not find any significant difference in stomatal sensitivity to changing VPD between diffuse and ring porous tree species growing together at the same elevation (Figure 4.8). Carbon isotope composition has been used as an indirect measure of stomatal conductance and water use efficiency (WUE) of leaves (Hubick et al. 1986; Körner et al. 1988, 1991) and can be employed to describe the variations in plant water use strategies (Marshall and Zhang 1994; Sun et al. 1996; Osorio et al. 1998; Li et al. 2006). δ13C concentration in plants provides an integrated measure of WUE during plant growth since the δ13C concentration of newly fixed carbon increases under conditions of low internal CO2 concentration (Ehleringer 1993). In our study, δ13C of Q. mongolica increased with elevation (Figure 4.9), and the difference between 950N and the other sites was significant (Padj <0.01, TukeyHSD). Higher δ13C at higher elevations suggests that WUE decreased with increasing elevation as observed by Körner et al. (1988; 1991), Sparks and Ehleringer (1997), and Cordell et al. (1998). Figure 4.9 Stable carbon isotope (13C) compositions in the leaves of Quercus mongolica distributed along an elevation gradient in the Haean catchment. Error bars indicate standard deviation from the means of individual trees along this gradient. Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 117 Although our study considered a relatively narrow elevational gradient (i.e., 450−950 m), significant reduction in total EC occurred. This was attributed to changes in Ta, VPD, and the length of growing season along the elevation, which influence tree transpiration. We demonstrated that the maximum daily water use of individual trees was universally defined by tree size regardless of species and elevation. Stem diameter is, therefore, a good indicator of TWU irrespective of species and elevation. Differences in species composition, therefore, do not influence water use by forest stands in these temperate mountain forests. Differences in stomatal sensitivity as observed along the elevational gradient, however, impacted EC and should be considered when conducting forest water budget in the mountain regions. 4.5 Acknowledgements This study was carried out as part of the International Research Training Group, TERRain and ECOlogical Heterogeneity (TERRECO; GRK 1565/1) funded by the Deutsche Forschungsgemeinschaft (DFG) at University of Bayreuth, Germany and the Korean Research Foundation (KRF) at Kangwon National University, Chuncheon, Korea. The isotope abundance analyses by the BayCEER – Laboratory of Isotope Biogeochemistry are kindly acknowledged. 4.6 References Barry RG (1981) Mountain weather and climate. Methuen, London Beniston M (2003) Climatic change in mountain regions: a review of possible impacts. Climatic Change 59: 5–31 Bidartondo MI, Burghardt B, Gebauer G, Bruns TD, Read DJ (2004) Changing partners in the dark: isotopic and molecular evidence of ectomycorrhizal liaisons between forest orchids and trees. Proc R Soc Lond B 271:1799−1806 Bucci SJ, Goldstein G, Meinzer FC, Scholz FG, Franco AC, Bustamante M (2004) Functional convergence in hydraulic architecture and water relations of tropical savanna trees: from leaf to whole plant. Tree Physiol 24:891−899 Burrows LE (1980) Differentiating sapwood, heartwood and pathological wood in live mountain beech. New Zealand Forest Service, Forest Research Institute, Protection Forestry Report 172 Campbell-Clause M (1998) Stomatal response of grapevines to wind. Aust J Exp Agr 38:77−82 Chapin FS III, Matson PA, Mooney HA (2002) Principles of Terrestrial Ecosystem Ecology. Springer, New York Clearwater MJ, Meinzer FC, Andrade JL, Goldstein G, Holbrook NM (1999) Potential errors in Chapter 4 – Influence of elevation on canopy transpiration of temperate deciduous forests in a complex mountainous terrain of South Korea ! 118 measurement of nonuniform sap flow using heat dissipation probes. Tree Physiol 19:681–687 Dittmar C, Elling W (2006) Phenological phases of common beech (Fagus sylvatica L.) and their dependence on region and altitude in Southern Germany. Eur J For Res 125:181−188 Ehleringer JR (1993) Carbon and water relations in desert plants: an isotopic perspective. In Ehleringer JR, Hall AE, Farquhar GD (Eds.), Stable isotopes and plant carbon-water relations, Academic Press, San Diego, California, USA Engelhardt S, Matyssek R, Huwe B (2009) Complexity and information propagation in hydrological time series of mountain forest catchments. Eur J For Res 128:621−631 Ewers BE, Oren R (2000) Analyses of assumptions and errors in the calculation of stomatal conductance from sap flux measurements. Tree Physiol 20:579−589 Gebauer T, Horna V, Leuschner C (2008) Variability in radial sap flux density patterns and sapwood area among seven co-occurring temperate broad-leaved tree species. Tree Physiol 28:1821−1830 Granier A (1987) Evaluation of transpiration in a Douglas-fir stand by means of sap flow measurements. Tree Physiol 3:309−320 Granier A, Biron P, Bréda N, Pontailler JY, Saugier B (1996) Transpiration of trees and forest stands: short and long-term monitoring using sapflow methods. Global Change Biol 2:265−274 Harris PP, Huntingford C, Coxb PM, Gasha JHC, Malhi Y (2004) Effect of soil moisture on canopy conductance of Amazonian rainforest. Agric For Meteorol 122:215−227 Hirobe M, Tokuchi N, Iwatsubo G (1998) Spatial variability of soil nitrogen transformation patterns along a forest slope in a Cryptomeria japonica D. Don plantation. Eur J Soil Biol 34:123−131 Hoch G, Körner C (2005) Growth, demography and carbon relations of Polylepis trees at the world’s highest tree line. Funct Ecol 19:941−951 Hubick KT, Farquhar GD, Shorter R (1986) Correlation between water-use efficiency and carbon isotope discrimination in diverse peanut (Arachis) germplasm. Australian J Plant Physiol 13:803−816 Jarvis PG (1976) The interpretation of the variations in leaf water potential and stomatal conductance found in canopies in the field. Phil Trans Roy Soc Lond B 273:593−610 Jung EY, Otieno D, Lee B, Lim JH, Kang SK, Schmidt MWT, Tenhunen J (2011) Up-scaling to stand transpiration of an Asian temperate mixed-deciduous forest from sing tree sapflow measurements. Plant Ecol 212:383−395 Kelliher FM, Hollinger DY, Schulze ED, Vygodskaya NN, Beyers JN, Hunt JE, McSeveny TM, Milukova I, Sogatchev A, Varlargin A, Ziegler W, Arneth A, Bauer G (1997) Evaporation for an eastern Siberian larch forest. Agric For Meteorol 85:135−147 Korea Forest Service (2009) National Report on Sustainable Forest Management in Korea 2009