<%BANNER%>

Interactions between Xylem Structure and Water Relations of Southern Pines

Permanent Link: http://ufdc.ufl.edu/UFE0024152/00001

Material Information

Title: Interactions between Xylem Structure and Water Relations of Southern Pines
Physical Description: 1 online resource (165 p.)
Language: english
Creator: Gonzalez, Carlos
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: canopy, cavitation, conductivity, elliottii, hydraulics, irrigation, latewood, palustris, punis, soil, taeda, tracheid, transpiration, water, wood
Forest Resources and Conservation -- Dissertations, Academic -- UF
Genre: Forest Resources and Conservation thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: This dissertation focused on the study of xylem structure and water relations of three major species of Southern pines: loblolly (LO), longleaf (LL) and slash pine (SL). The study was divided in two principal areas: (1) assessment of water availability and genetic family effects on water relation traits, growth and wood properties of mid-rotation LO; and (2) characterize water relation traits and secondary xylem structure of mature LL and SL. For the first study, water availability was controlled by irrigation into two fast growing families, one from Atlantic Coastal Plain (SC) and the other composed by a mix of Florida families (FL). The second study was carried out in a naturally-regenerated mixed stand of mature LL and SL on a flatwood site in north-central Florida. For LO, increasing water availability via irrigation increased transpiration and stomatal conductance scaled at canopy level; whole-tree water conduction efficiency was maintained at high levels due to avoidance of xylem embolism. LO tends to maintain constant maximum water potential gradient from roots to shoots at a cost of loss of conductivity under water-limited conditions. The two genetic families evaluated showed differences in canopy conductance response to water-limited conditions: SC adjusted their overall canopy conductance in response to drought, while FL did not. At age 11, irrigation increased specific gravity and latewood percentage and the mechanism of this response was an extension of the basal area growing season by 24.6 days in irrigated trees. The main effect of irrigation was an increase in latewood growth. Before canopy closure irrigation caused null or negative effect on specific gravity and latewood percentage due to large effect on earlywood growth associated with fast leaf area index development. After year 7, earlywood growth was similar between control and irrigated trees but latewood growth was larger on irrigated plots, increasing the overall year-ring specific gravity and latewood percentage. Trees from SC family had more desirable wood properties than trees from FL family, independent of irrigation; this effect was associated with greater yearly latewood growth in SC. In mature LL and SL, mean daily transpiration rate was higher for SL than LL trees and there were no significant differences between species in daily transpiration rate per unit leaf area. Species differences in transpiration rate were principally determined by differences in leaf area per tree; SL had 60% more leaf area per unit basal sapwood area than LL (p=0.086). LL had larger crown conductance than SL on days with high soil moisture and reduced to similar values than SL on days with low soil moisture. In terms of hydraulic architecture and tracheid anatomy, root sapwood-specific hydraulic conductivity of LL was larger than SL, but there were no species differences for any other organ tested. There were no differences in vulnerability to cavitation between species in any of organ evaluated and there was a weak trade-off between water conduction efficiency and safety. Tracheid hydraulic diameter was strongly correlated with sapwood-specific hydraulic conductivity across all organs. Tracheid allometry changed markedly between sapwood of roots, trunk and branches, reflecting different mechanical reinforcement needs. Higher sapwood to leaf area ratio and higher ks in roots of LL are anatomical traits that may allow LL to survive and dominate in drier soil microsites.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Carlos Gonzalez.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Martin, Timothy A.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024152:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024152/00001

Material Information

Title: Interactions between Xylem Structure and Water Relations of Southern Pines
Physical Description: 1 online resource (165 p.)
Language: english
Creator: Gonzalez, Carlos
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: canopy, cavitation, conductivity, elliottii, hydraulics, irrigation, latewood, palustris, punis, soil, taeda, tracheid, transpiration, water, wood
Forest Resources and Conservation -- Dissertations, Academic -- UF
Genre: Forest Resources and Conservation thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: This dissertation focused on the study of xylem structure and water relations of three major species of Southern pines: loblolly (LO), longleaf (LL) and slash pine (SL). The study was divided in two principal areas: (1) assessment of water availability and genetic family effects on water relation traits, growth and wood properties of mid-rotation LO; and (2) characterize water relation traits and secondary xylem structure of mature LL and SL. For the first study, water availability was controlled by irrigation into two fast growing families, one from Atlantic Coastal Plain (SC) and the other composed by a mix of Florida families (FL). The second study was carried out in a naturally-regenerated mixed stand of mature LL and SL on a flatwood site in north-central Florida. For LO, increasing water availability via irrigation increased transpiration and stomatal conductance scaled at canopy level; whole-tree water conduction efficiency was maintained at high levels due to avoidance of xylem embolism. LO tends to maintain constant maximum water potential gradient from roots to shoots at a cost of loss of conductivity under water-limited conditions. The two genetic families evaluated showed differences in canopy conductance response to water-limited conditions: SC adjusted their overall canopy conductance in response to drought, while FL did not. At age 11, irrigation increased specific gravity and latewood percentage and the mechanism of this response was an extension of the basal area growing season by 24.6 days in irrigated trees. The main effect of irrigation was an increase in latewood growth. Before canopy closure irrigation caused null or negative effect on specific gravity and latewood percentage due to large effect on earlywood growth associated with fast leaf area index development. After year 7, earlywood growth was similar between control and irrigated trees but latewood growth was larger on irrigated plots, increasing the overall year-ring specific gravity and latewood percentage. Trees from SC family had more desirable wood properties than trees from FL family, independent of irrigation; this effect was associated with greater yearly latewood growth in SC. In mature LL and SL, mean daily transpiration rate was higher for SL than LL trees and there were no significant differences between species in daily transpiration rate per unit leaf area. Species differences in transpiration rate were principally determined by differences in leaf area per tree; SL had 60% more leaf area per unit basal sapwood area than LL (p=0.086). LL had larger crown conductance than SL on days with high soil moisture and reduced to similar values than SL on days with low soil moisture. In terms of hydraulic architecture and tracheid anatomy, root sapwood-specific hydraulic conductivity of LL was larger than SL, but there were no species differences for any other organ tested. There were no differences in vulnerability to cavitation between species in any of organ evaluated and there was a weak trade-off between water conduction efficiency and safety. Tracheid hydraulic diameter was strongly correlated with sapwood-specific hydraulic conductivity across all organs. Tracheid allometry changed markedly between sapwood of roots, trunk and branches, reflecting different mechanical reinforcement needs. Higher sapwood to leaf area ratio and higher ks in roots of LL are anatomical traits that may allow LL to survive and dominate in drier soil microsites.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Carlos Gonzalez.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Martin, Timothy A.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024152:00001


This item has the following downloads:


Full Text

PAGE 1

INTERACTIONS BETWEEN XYLEM STRU CTURE AND WATER RELATIONS OF SOUTHERN PINES By CARLOS A. GONZALEZ A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009 1

PAGE 2

Carlos A. Gonzlez 2

PAGE 3

To Claudia, Cristbal, Gabriel and Sa ntiago. You are the reason for everything. 3

PAGE 4

ACKNOWLEDGMENTS Many people come to mind at this time. I al ways will remember the guidance, direction and friendship of my major advi sor, Dr. Tim Martin. Thanks go to Dr. Eric Jokela for his continuous encouragement and support; he is a gr eat teacher and at the same time a very kind person with his door always opened to help. I ha ve to thank Dr. Gary Peter for his excellent suggestions and great disposition to help. Thanks also to Dr. Matias Kirst, for his participation as supervisory committee member. Thank to Dr. Th omas Sinclair, who with his comments and observations directed important parts of the anal ysis. I would also like to thank Dr. Dudley Huber, who, even though he was not a memb er of my committee, generously provided enlightening help with statistical analysis. Its b een a great honor to work together with all of them. Thanks also go to Dr. Phillip Dougherty, w ho encouraged me to start a PhD program and put me in contact with Dr. Martin. This project could not have been done w ithout the financial s upport of USDA-Forest Service, the National Science Foundation and the Forest Biology Research Cooperative. I would like to thank International Paper for providing access to the study site, and Dr. Mike Kane and Mark Register for providing logi stical support and for maintaini ng the irrigation system through the study period. My thanks go to many people who provided tech nical and logistical assistance. Mr. Dan Schultz and Mr. Lamar Courtney assisted with field activities at the Austin Cary Memorial Forest. Dave Nolletti was a big help in the ea rly stages of my time at UF, which was a time when lots of help was needed. Dr. J.C Domec generously provided information on the vacuum soaking method. Dr. Lisa Samuelson shared Fi eld of Dreams data, Dr. Thomas Teklemariam taught me the secrets of building sap flow probes. Dr. Rosana Higa helped with field work, Mr. Fabian Hergenreder was an invaluable field a nd lab assistant, and Mr Kenneth Carabantes 4

PAGE 5

helped with tracheid measurements. I am gratef ul to Gregory Gorman for his friendship and valuable help in the field. To my friends and fe llow graduate students: Andres Susaeta, thanks for his help in biomass sampling, Patricio Muoz, thanks for help me in measuring trees, and Xiaobo Li, thanks for his help in wood-co re and soil sampling in the field. Thanks go to Francisco Flores and Christian Mora, for their support, friendship and help in the field and data analysis. My friends Pablo, Rodrigo, Bernardo, Federico, Jose and Cynnamon, were great company in after work activ ities (sorry if Im missing somebody). Thanks go to all the guys at the soccer team; nobody but us knows the importance of each game. Thanks go to my family in Chile, for their l ove and support. I am indebted to Claudia, Cristbal, Gabriel and Santiago, for holding on a nd supporting the dream. Im grateful for this life and this great opportunity; the journe y never ends. 5

PAGE 6

TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ........10 LIST OF FIGURES.......................................................................................................................11 ABSTRACT...................................................................................................................................14 CHAPTER 1 INTRODUCTION................................................................................................................. .16 2 WATER AVAILABILITY AND GEN ETIC FAMILY EFFECTS ON WATER RELATIONS OF AN 11 YEAR-OLD LOBLOLLY PINE ( Pinus taeda L.) PLANTATION.......................................................................................................................20 Introduction................................................................................................................... ..........20 Materials and Methods...........................................................................................................21 Site and Stand Description..............................................................................................21 Meteorological Measurements........................................................................................22 Soil Moisture...................................................................................................................23 Soil Water Retention Curves...........................................................................................24 Soil Particle Size Distribution, Bulk Density and Hydraulic Conductivity....................24 Water Extraction at Different Soil Depths......................................................................25 Tree Selection..................................................................................................................26 Sap Flow Measurements..................................................................................................26 Leaf Area Index...............................................................................................................29 Canopy Conductance.......................................................................................................29 Whole-Tree Hydraulic Conductance...............................................................................31 Foliar Analysis.................................................................................................................31 Statistical Analysis..........................................................................................................3 2 Results.....................................................................................................................................32 Environmental Conditions...............................................................................................32 Radial Profile in Sap Flow..............................................................................................33 Sap Flow..........................................................................................................................34 Leaf Area Index...............................................................................................................35 Soil Moisture...................................................................................................................35 Canopy Conductance.......................................................................................................36 Whole-Tree Hydraulic Conductance...............................................................................37 Foliar Analysis.................................................................................................................37 Discussion...............................................................................................................................37 Conclusion..............................................................................................................................44 6

PAGE 7

3 WATER AVAILABILITY AND FAMILY EFFECTS ON WOOD PROPERTIES OF LOBLOLLY PINE ( Pinus taeda L.)......................................................................................59 Introduction................................................................................................................... ..........59 Materials and Methods...........................................................................................................61 Site and Stand Description..............................................................................................61 Meteorological Measurements........................................................................................61 Soil Moisture...................................................................................................................62 Diameter and Basal Area Growth....................................................................................62 Wood Properties..............................................................................................................63 Statistical Analysis..........................................................................................................6 4 Results.....................................................................................................................................64 Environmental Conditions...............................................................................................64 Soil Moisture...................................................................................................................65 Diameter and Basal Area Growth and Day of Growth Cessation...................................65 Wood Properties..............................................................................................................65 Annual Ring Growth.......................................................................................................67 Statistical Analysis..........................................................................................................6 8 Discussion...............................................................................................................................68 Conclusion..............................................................................................................................71 4 WATER USE, WHOLE-TREE HYDRAUL IC CONDUCTANCE AND CANOPY CONDUCTANCE DYNAMICS IN MATURE Pinus palustris AND Pinus elliottii TREES....................................................................................................................................79 Introduction................................................................................................................... ..........79 Materials and Methods...........................................................................................................80 Site and Stand Description..............................................................................................80 Meteorological Measurements........................................................................................81 Soil Moisture...................................................................................................................81 Tree Selection..................................................................................................................82 Sap Flow Measurements..................................................................................................83 Leaf and Sapwood Area..................................................................................................85 Crown Conductance........................................................................................................86 Whole-Tree Hydraulic Conductance...............................................................................87 Statistical Analysis..........................................................................................................8 8 Results.....................................................................................................................................88 Environmental Conditions...............................................................................................88 Sap Flux Density, Transpiration, Wate r Storage Use and Soil Moisture........................89 Crown Conductance........................................................................................................91 Whole-Tree Hydraulic Conductance...............................................................................92 Discussion...............................................................................................................................92 Conclusion..............................................................................................................................95 7

PAGE 8

5 HYDRAULIC ARCHITECTURE AND TR ACHEID ALLOMETRY IN MATURE Pinus palustris AND Pinus elliottii TREES.........................................................................104 Introduction................................................................................................................... ........104 Materials and Methods.........................................................................................................105 Site and Stand Description............................................................................................105 Environmental Measurements.......................................................................................106 Tree Selection................................................................................................................106 Hydraulic Conductivity and Vulner ability to Cavitation Curves..................................106 Leaf to Sapwood Area Ratio.........................................................................................112 Tracheid Length, Diameter and Cell-Wall Thickness...................................................113 Specific Gravity.............................................................................................................11 5 Statistical Analysis........................................................................................................116 Results...................................................................................................................................116 Hydraulic Conductivity and Vuln erability to Cavitation..............................................116 Radial Profile in Hydraulic Conductivity......................................................................118 Tracheid Anatomy.........................................................................................................118 Specific Gravity.............................................................................................................12 0 Pit and Lumen Conductivity..........................................................................................121 Discussion.............................................................................................................................121 Conclusion............................................................................................................................126 6 SUMMARY AND CONCLUSIONS...................................................................................139 Water Availability and Family Effects on Wate r Relations on 11-year -old Loblolly Pine..140 Water Availability and Family Effects on Basal Area Growth and Wood Properties of Loblolly Pine.....................................................................................................................140 Water Relations of Longleaf and Slash Pine........................................................................141 Hydraulic Architecture and Tracheid A llometry of Longleaf and Slash Pine......................141 Comparison across Experiments...........................................................................................142 Future Research Areas Should Focus on..............................................................................143 APPENDIX A ANOVA MODEL FOR IRRIGATION X FA MILY ANALYSIS FOR LOBLOLLY PINE.....................................................................................................................................147 B ANOVA MODEL FOR ORGAN X SPECIES ANALYSIS FOR LONGLEAF AND SLASH PINE..................................................................................................................... ...148 C FOLIAR BIOMASS AND BIOMETRI C VALUES OF MEASURED TREES.................149 D PRESSURE ACHAMBER AND PRESSURE-SLEEVE APPARATUS............................150 E DAILY LAI FOR LOBLOLLY PINE AT SILVER FOREST............................................151 F PARTICLE SIZE DISTRIBUTION AND WA TER RELEASE CURVES FOR SILVER FOREST...............................................................................................................................152 8

PAGE 9

LIST OF REFERENCES.............................................................................................................153 BIOGRAPHICAL SKETCH.......................................................................................................165 9

PAGE 10

LIST OF TABLES Table page 2-1. Mean diameter at breast height (DBH), height, basal area (BA) and leaf area index (LAI) for Irrigation x Family treatments...........................................................................46 2-2. Diameter at breast height (DBH) and number of variab le length sapflow probes (in parenthesis) installed on FL family trees...........................................................................47 2-3. Least-square means of measured traits fo r irrigation (control and irrigated) and family (FL and SC) treatments......................................................................................................48 3-1. Mean diameter at breast height (DBH), height, basal area (BA) and leaf area index (LAI) for Irrigation x Family treatments...........................................................................72 4-1. Least-square means of measured traits for longleaf and slash pines.....................................97 5-1. Least-square means of diameter at breast height (DBH), total hei ght, total stem volume without bark (Volume), leaf area (LA) a nd all-sided specific leaf area (SLA)...............127 5-2. Least square means of hydraulic and tracheid anatomy traits..............................................128 A-1. Biometric values of measured trees....................................................................................149 A-2. Foliar biomass, specific leaf area and leaf area of measured trees.....................................149 E-1. LAI curve parameters for control and irrigated plots..........................................................151 F-1. Soil particle size distribu tion and water release curve (WRC ) applied at different depth for each replicate............................................................................................................. .152 F-2. Water release curve parameters...........................................................................................152 10

PAGE 11

LIST OF FIGURES Figure page 2-1. Environmental conditions through study period...................................................................49 2-2. Diurnal sap flux density ( Js) patterns at different radial positions for six selected trees of FL family on control and irrigated plots........................................................................50 2-3. Diurnal variation in weighted sapflow and radial profile in sap flux density at noon for six selected trees of FL family...........................................................................................51 2-4. Seasonal pattern in tran spiration, soil water content and rainfall and available soil water...................................................................................................................................52 2-5. Proportion of soil water extraction used for transpiration from the 0-200 cm depth on nonirrigated plots and from the 0-75 cm depth on control and irrigated plots for FL family plots........................................................................................................................53 2-6. Monthly average transpiration rate (E) and mean daily canopy conductance ( GCday) for control and irriga ted treatments over the study period......................................................54 2-7. Examples of relationship between GC and D for FL family under control and irrigated conditions. .................................................................................................................. ......55 2-8. Mean daily canopy conductance ( GCday), normalized canopy conductance ( GC%), available soil water (ASW) for 0-75 cm depth..................................................................56 2-9. Relationships between mean daily canopy conductance ( GCla) and pre-dawn water potential (pred), maximum daily water potential gradient ( ) and whole-tree sapwood-specific hydraulic conductance ( KS-wt)...............................................................57 2-10. Relationships between m ean daily canopy conductance ( GCla) and foliar nitrogen concentration (N) with 13C for control and irrigated plots...............................................58 3-1. Daily average precipitation and soil wate r content (mm) in 2005 for the 0-35 cm depth for a loblolly pine stand unde r an irrigation treatment......................................................73 3-2. Average monthly basal area growth and cumulative basal area accretion for irrigated (open circle) and control (fille d circle) plots during 2005.................................................74 3-3. Relationship between the cessation of ba sal area growth and late wood percentage and specific gravity for irrigated (triangl e) and control (circle) plots......................................75 3-4. Annual latewood specific gravity, latewood percentage, ring specif ic gravity, and total rainfall and water adde d by drip irrigation........................................................................76 11

PAGE 12

3-5. Annual growth in ring area, latewood area, earlywood area, and total rainfall and water added by drip irrigation......................................................................................................77 3-6. Relationship between mean latewood area difference between irrigated and non irrigated tree and the amount of water added through irrigation.......................................78 4-1. Environmental conditions through study period...................................................................98 4-2. Diurnal sap flux density ( JS) patterns at different radial positions for two longleaf and slash pine trees on one selected day...................................................................................99 4-3. Example of diurnal courses of PAR and D and sap flow....................................................100 4-4. Average daily transpiration rate per tree, transpiration pe r unit leaf area, water storage use and soil matric potential at 50 and 150 cm depth......................................................101 4-5. Average daily crown conductance and volum etric soil water content at 50 cm depth for longleaf (filled circle) and slash (open circle) pine.........................................................102 4-6. Average daily water storage use a nd relationship between sapwood volume and average water storage use dur ing the measurement period ............................................103 5-1. Percentage lo ss of conductivity ( PLC ) versus the applied air pressure in longleaf pine (filled circle) and slash pine (open circle)........................................................................130 5-2. Radial profile in maximum sa pwood-specific hydrau lic conductivity ( ks-max) and percentage loss of conductivity ( PLC ) for stem at 1.8 m height and at crown base .......131 5-3. Estimated using VC-curves v/s measured PLC on roots and branches of longleaf and slash pine trees.................................................................................................................132 5-4. Xylem tension that causes 50% of loss of conductivity ( 50) versus maximum sapwood specific hydraulic conductivity ( ks-max) for branch, root, Sbase and Scrown...................133 5-5. Maximum sapwood specifi c hydraulic conductivity ( ks-max) and xylem tension that causes 50% of loss of conductivity ( 50) versus tracheid length ( Lt) and hydraulic tracheid lumen diameter ( Dh)...........................................................................................134 5-6. Allometry of tracheids of longleaf and slash pine trees.......................................................135 5-7. Relationships between cell-wall to hydraulic diameter ratio ( C ) and specific gravity (SG), maximum sapwood specifi c hydraulic conductivity ( ks-max) and vulnerability to cavitation ( 50)................................................................................................................136 5-8. Relationship betw een tracheid length ( Lt) and sapwood-specific hydraulic conductivity of pits ( ks-pit).....................................................................................................................137 12

PAGE 13

5-9. Maximum sapwood specifi c hydraulic conductivity ( ks-max) versus sapwood-specific hydraulic conductivity of lumen ( ks-lumen) for branch, root, Sbase and Scrown...............138 6-1. Relationship between sensitivity of stom atal conductance at crow n (longleaf and slash) and canopy (loblolly) scales to D (-d GC/dln D ) and canopy conductance at D =1 kPa ( GCref)...............................................................................................................................145 6-2. Relationship between average daily stom atal conductance per uni t leaf area at crown (longleaf and slash) or canopy (loblolly) basis and volumetric soil water content ( v) at 50 cm soil depth...........................................................................................................1 46 D-1. Pressure chamber for vulnerability to embolism measurements.........................................150 D-2. Pressure-sleeve apparatus for hydraulic conductivity measurements on trunk segments...150 E-1. Daily LAI relative to yearly maximum for control and irrigated plots................................151 13

PAGE 14

Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy INTERACTIONS BETWEEN XYLEM STRU CTURE AND WATER RELATIONS OF SOUTHERN PINES By Carlos A. Gonzlez May 2009 Chair: Timothy A. Martin Major: Forest Resources and Conservation This dissertation focused on the study of xylem structure and water rela tions of three major species of Southern pines: loblolly (LO), longleaf (LL) and slash pine (SL). The study was divided in two principal areas: (1 ) assessment of water availability and genetic family effects on water relation traits, growth a nd wood properties of mid-rotation LO; and (2) characterize water relation traits and secondary xylem structure of mature LL and SL. For the first study, water availability was controlled by irrigation into two fa st growing families, one from Atlantic Coastal Plain (SC) and the other composed by a mix of Florida families (FL). The second study was carried out in a naturally-regene rated mixed stand of mature LL and SL on a flatwood site in north-central Florida. For LO, increasing water availability via irri gation increased transpiration and stomatal conductance scaled at canopy level; whole-tree water conduction e fficiency was maintained at high levels due to avoidance of xylem embolism. LO tends to maintain constant maximum water potential gradient from roots to shoots at a cost of loss of conduc tivity under water-limited conditions. The two genetic families evaluate d showed differences in canopy conductance response to water-limited conditi ons: SC adjusted their overa ll canopy conductance in response to drought, while FL did not. At age 11, irriga tion increased specific gravity and latewood 14

PAGE 15

percentage and the mechanism of this respons e was an extension of the basal area growing season by 24.6 days in irrigated trees. The main effect of irriga tion was an increase in latewood growth. Before canopy closure irri gation caused null or negative e ffect on specific gravity and latewood percentage due to large effect on earlywood growth associat ed with fast leaf area index development. After year 7, earlywood growth was similar between control and irrigated trees but latewood growth was larger on irrigated plots, in creasing the overall year -ring specific gravity and latewood percentage. Trees from SC family had more desirable wood properties than trees from FL family, independent of irrigation; this effect was associated with greater yearly latewood growth in SC. In mature LL and SL, mean daily transpirati on rate was higher for SL than LL trees and there were no significant differences between specie s in daily transpiration rate per unit leaf area. Species differences in transpiration rate were principally determined by differences in leaf area per tree; SL had 60% more leaf area per unit basal sapwood area than LL (p=0.086). LL had larger crown conductance than SL on days with high soil moisture and reduced to similar values than SL on days with low soil moisture. In term s of hydraulic architecture and tracheid anatomy, root sapwood-specific hydrauli c conductivity of LL was larger than SL, but there were no species differences for any other organ tested. Th ere were no differences in vulnerability to cavitation between species in any of organ eval uated and there was a weak trade-off between water conduction efficiency and safety. Tracheid hydraulic diameter was strongly correlated with sapwood-specific hydraulic conductiv ity across all organs. Tracheid allometry changed markedly between sapwood of roots, trunk and branches, re flecting different mechanical reinforcement needs. Higher sapwood to leaf area ratio and higher ks in roots of LL are anatomical traits that may allow LL to survive and domin ate in drier soil microsites. 15

PAGE 16

CHAPTER 1 INTRODUCTION Plant primary production requires a substa ntial amount of wate r and its loss through transpiration is an unavoidable consequence of the need fo r stomata to open for carbon assimilation (Gartner and Meinzer, 2005; Lamb ers et al., 1998). Water plays a crucial role, directly or indirectly, in all physiological processes (Lambers et al., 1998), being a major factor in plant growth regulation (Kramer and Boyer, 1995) Plants can regulate excessive water loss in response to water stress by reducing stomatal conductance (Tyree, 2003; Sperry, 2000). The water potential at which stomatal closure occurs has been shown to be closely related to the water potential level at which xylem cavitation becomes significan t (Sperry and Ikeda, 1997). In other words, regulation of stomat al conductance and water loss app ears to have evolved so that catastrophic xylem embolism is avoided (Sperry, 2000). Hydraulic architect ure is the structure of water transport system in plants (Tyree and Ew ers, 1991); Cochard (1994), cited by Cruiziat et al. (2002), defined hydraulic arch itecture as the set of hydr aulic characteristics of the conducting tissue of a plant which qualify and quan tify the sap flux from roots to leaves. This structure can limit plant water relations, gas ex change through the leaves and probably the maximum height that particular tree species can achieve (Spe rry, 2000; Tyree and Ewers 1991). Of thirteen native pine species in the south eastern United States, eight are classified as southern pines (Sternitzke and Nelson, 1970), th e common name given to the species members of the subsection Australes of the Pinus genus (Schultz, 1997). Longleaf pine ( Pinus palustris Mill.), slash pine ( Pinus elliottii Engelm. var. elliottii) and loblolly pine ( Pinus taeda L.) are the three major species within this classification. The United States is one of the largest producers of industrial timber in the world, accounting for 28% of total timber production. Of countrys total wood fiber output, 58% originates from southern pines (Prestem on and Abt, 2002). In the US, 16

PAGE 17

forest trees are the ecologically predominant life-form on approximate ly 111 million hectares, and in southern states, southern pines are found on more than 40 million hectares, either as pure or mixed stands (Sternitzke and Nelson, 1970). The natural range of longleaf pine includes most of the Atlantic and Gulf Coasta l Plains from southeastern Virg inia to eastern Texas and south through the northern two-thirds of peninsular Florida (Burns and Honkala, 1990). Slash pine grows throughout the flatwoods of north Florida and south Georgia; its native range includes the lower Coastal Plain, part of the middle Coastal Pl ain, and the hills of south Georgia; within its natural range, slash pines pref erred sites include poorly drained flatwoods and stream edges, as well as seasonally flooded areas such as bays and swamps (Barnett and Sheffield, 2002). Lohrey and Kossuth, 1990 indicates that 69% of slash pine st ands are plantations, and that stand area that corresponds to approximately 4.2 million hectares in the year 2000 is largely concentrated in Florida and Georgia. Loblolly pi ne is one of the fastest grow ing pines and thrives on various sites from east Texas to southern Missour i to north Florida to south New Jersey. Flatwoods are the most extensive type of te rrestrial ecosystems in Florida, covering approximately 50% of its land area (Ewel, 1990). In north-Central Florida, two dominant tree species are characteristic of pine flatwoods: longleaf pine and slas h pine. Longleaf pine forest at one time the may have occupied as much as 24 million ha, although today less than 1.6 million ha remained (Burns and Honkala, 1990). More th an 97% of the original land area of longleaf pine has been lost to other uses (Johnson and Gjerstad, 2006) and some of the most species-rich areas and highest concentrations of endangered and threatened sp ecies in the Southeast are found on mesic Flatwood Sites (Cohen et al., 2004). On the other hand, loblolly pine is the most commercially important timber species in the southeastern United States, predominating 17

PAGE 18

southern pines lands with land area of approximate ly 7.2 millions hectares (Schultz, 1997). From total seedlings planted in southe rn US more than 84% are lobl olly pine (McKeand et al., 2003). To explain natural patterns of productivity and ha bitat preferences of southern pines, or to increase productivity of plante d stands, it is important to und erstand the controls of water availability over tree water relati ons and the consequences for tr ee growth (Lambers et al., 1998). The overall goal of this disserta tion was to obtain a better know ledge about water relations of three major species of southern pines: loblolly, longleaf and slas h pine. The study was divided in two principal areas: (1) assessment of water availability and genetic family effects on water relation traits, growth and wood properties of mid-rotation loblo lly pine; and (2) characterize water relation traits a nd secondary xylem struct ure of mature longleaf and slash pine trees cohabiting the same site and the relationship of th ese traits with microsit e habitat preference of each species. For the first study, water availability was controlled by irrigation in two fast growing families that received the extra water input since plantation establishment. This study was divided into two main areas: (a) water availability and family effects on water use, whole-tree hydraulic conductance and canopy conductance respons es to varying environmental conditions; and (b) water availability and fa mily effects on wood properties, examining the responses at age eleven on date of diameter growth cessation a nd its relationship with specific gravity and latewood percentage. For long term responses, the effects of irrigation on wood properties since plantation establishment were examined using complete wood cores. The second study, which was carried out in a naturally-regenerated mixed stand of longleaf and slash pine on a flat wood site in north-central Florid a, was also divided into two main areas of research: (a) species characteristic s in water relation traits such as total daily 18

PAGE 19

transpiration, water storage use, whole-tree hydraulic conductance, leaf to sapwood area and hydraulic adjustments in crown conductance to varying environmental conditions; and (b) a comparison of species characteristics in hydrau lic conductivity, vulnera bility to cavitation and tracheid allometry of roots, stem and branches. The results of this study should deepen our knowledge of the water relations of southern pines, impacting future tree growth modeli ng and management decision-making related to species (for loblolly, long leaf, and slash) and seed sources (for loblolly). 19

PAGE 20

CHAPTER 2 WATER AVAILABILITY AND GENETIC FA MILY EFFECTS ON WATER RELATIONS OF AN 11-YEAR-OLD LOBLOLLY PINE (PINUS TAEDA L.) PLANTATION Introduction The United States is the larges t producer of indust rial timber in the world, accounting for about 28% of the production (Prestemon and Abt, 2002) and around 58% of total timber harvested and close to 75% of tota l tree planting each year in the country is made in the southeast industrial plantations (McKeand et al., 2003). Loblolly pine ( Pinus taeda L.) is the most commercially important timber species in the S outheastern United States, accounting for more than 84% of seedlings planted are loblolly pine (McKeand et al., 2003). Loblolly pine is also one of the fastest growing pines and thrives on various sites from east Texas to southern Missouri to north Florida to south New Jersey. Water availability is one of the main factors controlling tree growth and species distribution (Lambers et al., 1998) Water supply to leaves is n eeded to sustain photosynthesis and can influence the characteristics of the secondary xylem (w ood) formed during the growing season (Panshin and de Zeeuw, 1980). Plant pr imary production requires substantial amounts of water and this water loss is an unavoidable consequence of photosynthesis when stomata are open for carbon assimilation (Lambers et al., 1998; Gartner and Meinzer, 2005). Plants regulate excessive water loss in response to water stre ss by reducing stomatal conductance and this reduction is related to avoiding excessive cav itation (Sperry and Iked a, 1997; Sperry, 2000; Tyree, 2003). Hydraulic conductance is the chan ge in flow rate of liquid water through the system per change in hydraulic pressure dr iving the flow (Sperr y, 2000). It can limit the maximum rate of gas exchange and carbon ga in (Tyree, 2003). Plants with high hydraulic conductance tend to have higher growth rates and are more vulnerable to embolism than plants with lower hydraulic conduc tance (Tyree and Zimmerman, 2002). Low values of hydraulic 20

PAGE 21

conductance reduce the ability to maintain stomatal conductance as evaporative demands increases at low soil moisture content (Wullschleger et al., 1998; Wang et al., 2003). The water potential at which stomatal clos ure occurs has been shown to be closely related to the water potential level at which xylem cavitation becomes significant (Sperr y and Ikeda, 1997). In other words, regulation of stomatal conductance and water loss appear s to have evolved so that catastrophic xylem embolism is avoided (Sperry, 2000). In this study we examined the effect of water availability and genetic family on water use, canopy conductance and whole-tree hydraulic conductance in a mid-rotation loblolly pine plantation. We hypothesize that i) due to water availability lim itations whole-tree hydraulic conductance ( KS-wt) will be decreased because of cavitati on of water-conducting tracheids and ii) due to natural selection on sites with differences in native mois ture conditions, trees from wet climate will present a smaller degr ee of stomatal control in respons e to water stress and this can be reflected in smaller adjustments in canopy conductance ( GC) under water stressed conditions. Water availability was controlled through an irrigation treatment applied to two fast growing families that received the extra water input sin ce plantation establishment. This study was carried out with the aim to understand th e degree that varying seed sour ces differ in water relations behavior. Materials and Methods Site and Stand Description The study took place in an irriga tion and genetics experiment established in January 1995 by International Paper, Inc. in the Upper Coasta l Plain 22 km west of Bainbridge, GA (30 N latitude and 84 W longitude). Soils at th is location were classi fied as well-drained Grossarenic Paleudults, with 50 cm sandy lo am over sandy clay loam (Samuelson, 1998). 21

PAGE 22

The study consisted of four treatments: two ge netic entries plus two water availability entries repeated three times in a split-plot, ra ndomized complete block design (12 plots total), with irrigation as the whole-plot and genetic entry (family) as the sub-plot. This experiment represented a subset of the whole experiment, which also included two more families, fertilization and pest control treatments. Each measurement plot (excluding 2 treated buffer rows) had an area of 0.026 ha, containing 28 sa mple trees planted at a 2.4 m x 3.7 m spacing. More details about the site have been described by Samuelson (1998) and Samuelson et al. (2001 and 2008). The genetic treatments consisted of two open-pollinated second generation improved families: one originating from a mixture of north ern Florida half-sib fam ilies (FL), and a second originating from a single half-sib family from the South Carolina Coastal Plain (SC), widely used by forest companies. The water tr eatments included and irrigated (6.25 mm day-1 on drip irrigation from July to November in 2005 a nd from March to October in 2006) and a nonirrigated control treatment. Meteorological Measurements Meteorological data were recorded from J une 2005 to January 2007 with an automatic weather station located in an open area adjacent to the stand. Relative hu midity and temperature were measured with a relative humidity and temperature probe (HMP45C-L, Vaisala, Inc., Helsinki, Finland). Photosynthe tic photon flux density (PPFD, mol m-2 s-1) was measured with a quantum sensor (Li-190, Li-Cor Inc., Lincoln, NE, USA). Wind direction and velocity were measured with an anemometer and vane (03001L, Campbell Scientific, Logan, UT, USA) and precipitation was recorded with a tipping bucket rain gage (TR525-I, Texas Electronics, Dallas, TX, USA). Data for all sensors were collected on a 30 s cycle and stored on an automatic datalogger (CR10X, Campbe ll Scientific, Logan, UT, USA) as 30 min averages. 22

PAGE 23

Soil Moisture Vertically oriented, 20 cm long capacitive so il moisture probes (EC-20, Decagon Devices, Pullman, WA, USA) were used to measure volumetric soil water content ( v, cm3 cm-3). In June 2005 probes were installed at 0.2 m depth into the bed and interbed positions of each plot. In March 2006 one additional probe wa s installed at 0.5 m depth in the center of each plot to measure changes in volumetric water content in deep er portions of the soil profile. To evaluate the dynamics of soil water extraction under rain fed conditions, probes were installed at 1.0, 1.5 and 2.0 m depth in each replicate of the non-irrigated north Florida family plots. The estimations of water content of each probe were extrapolat ed to the mid-point distance between sensors. Specific calibrations were developed from soil samples taken 0.5 m away from each probe from each plot containing FL family trees (control a nd irrigated plots on all three replicates); soil samples were collected from the same depth where the probes were installed. Soil-specific calibrations had no difference with manufacturer default calibrat ion model to transform sensors outputs (mV) to volumetric water content: v = -0.24508 + 0.0007958 mV. There was no difference between samples taken at 10-30, 40-60 and 90-110 cm depth, and the manufacturers calibration (p=0.78 and 0.78 for inter cept and slope, respectively). Available soil water (ASW, %) was calculated for each plot and day. By analyzing the limits of wetting and drying of the soil thr ough the entire study period, drained upper limits (DUL, mm) and lower limits of wa ter extraction (LL, mm) were de termined for each plot and depth; ASW was calculated using th e following formula (Ritchie, 1981): ASW = 1 LLDUL WCDUL where WC is the water content (mm) at any given measurement day and depth. 23

PAGE 24

Soil Water Retention Curves The relationship between v and soil matric potential ( s ,MPa) was quantified by constructing soil water release curves using th e filter paper method (Deka et al., 1995; Ophori and Maharjan, 2000; Marinho and Oliveira, 2006) on soils sampled from 0.1-0.3, 0.4-0.6, 0.91.1, 1.4-1.6 and 1.9-2.1 m depth in each control FL plots The relationship between v and s was modeled as (Warren et al. 2005): s = c vba 1 where a, b and c are parameters determined by th e non-linear regression. Ot her models were also tested (e.g., s = a + (b/ v) + (c/ v 2), s = 1/ (a+b v c) or s = a v b), but the model proposed by Warren et al. (2005) provided the best fit for all data. Soil Particle Size Distribution, Bulk Density and Hydraulic Conductivity For the same samples used to develop the wa ter release curves, part icle size distribution was determined using the hydrometer method (Gee and Bauder, 1986). Soil bulk density ( s, g cm-3) was determined using a soil core sampler (0200, Soil Moisture Equipment Corp, Santa Barbara, CA, USA) on samples at < 100 cm depth, b ecause of difficulty in extracting intact cores below that depth. Three samples per plot and dept h were extracted, and the average was used for further analysis. Particle size di stribution changed with soil depth, with increasing clay content found with depth; for 10-30, 40-60, 90-110, 140-160 and 190-210 cm depth, the average sand and clay content were 80.7, 76.2, 63.3, 68.4 and 67.4%, and 6.8, 11.8, 22.7 23.8 and 23.8%, respectively. For s estimations below 100 cm depth, data were obtained from the NRCS-USDA soil survey for the study area ( http://soildatamart.nrcs.usda.gov April 2008). Saturated soil hydraulic conductivity ( ksat, mm day-1) and unsaturated hydraulic c onductivity as a function of v ( k, mm day-1) were determined for each plot and depth using a computer program based on 24

PAGE 25

pedotransfer functions (Schaap et al., 2001; ROSETTA versi on 1.2, US Salinity Laboratory ARS-UDA, Riverside, CA, USA). This model us ed the equation presented by Schaap and Leij (2000): k = ( k0 Se L)*(1-(1-Se (n/(n-1)))(1-1/n) ) 2 where k0 is the matching point at saturation, L is an empirical pore tortuosity/connectivity parameter, n is a curve shape parameter and Se is the relative saturation, which can be determined as: Se = rs rv where r and s (cm3 cm-3) are residual and saturated water contents, respectively. Water Extraction at Different Soil Depths The local water balance method (Oren et al., 1998; Ewers et al., 1999) was used to estimate the proportion of daily water use associated with changes in soil water content for each plot and depth. An expression for infiltration rate (I, cm day-1) was estimated by the equation (Hillel, 1998): I = k z where z is the soil layer depth (cm). Soil layer depth was assumed as the distance from the mid point between two contiguous soil moisture probes (e.g., for a soil moisture probe installed at 50 cm, a soil layer depth of 40 cm was assumed, representing a soil layer between 35 and 75 cm, because above and below this probe there were sensors installed at 20 and 100 cm depth). For each depth and for days without rain (Ewers et al., 1999), changes in water content ( S, mm) were determined using the formula presented by Oren et al. (1998): S = z 25

PAGE 26

where is daily change in volumetric water co ntent. Daily water uptake by trees from each layer was computed as the difference between S and net infiltration (Inet), which was calculated as the difference between I from the layer above and I to the layer below). No lateral water movement was assumed (terrain slope < 5%). Tree Selection Sap flux density in the stem xylem was measured on a subset of eight trees per plot in all evaluated plots (96 tree in tota l) using Granier-type heat di ssipation sap flow measurement probes (Granier 1985, 1987) from June 2005 to January 2007. The measurement trees were chosen from across the range of tree sizes using "quantiles of total", a stratification scheme which weights the selection of large trees more heavily (Hatton et al., 19 95; Martin et al. 1997; ermk et al., 2004). By selecting measurement tr ees using this method, a sampling size of 8 trees per plot was sufficient to control estimati on error on sapflow determinations (Hatton et al., 1995). Sap Flow Measurements Twenty mm long Granier type he at dissipation probes were installed on the north side of the stem in all 96 sample trees at 1.8 to 2.0 m above ground. This method has been used widely and described elsewhere (Braun and Schmid, 1999; Clearwater et al., 1999 ; Lu, 1997; Ewers and Oren, 2000; Lu et al., 2004) Briefly, the system consists of tw o probes inserted radially into the stem, one above another about 10-15 cm apart. Th e upper probe contains a heater and a T-type thermocouple and the lower probe contains only the thermocouple. The upper probe was heated at constant power while the lower one was us ed as a reference for measuring the ambient temperature. The temperature difference between the heated and reference probes ( T) was recorded, and by comparing the difference to the maximum occurring at predawn ( Tm) when there was assumed to be no flow (Granier 1985, 1987; Lu, 1997; Jimenez et al., 2000; Oren and 26

PAGE 27

Pataki, 2001; Ford et al ., 2004). Sap flux density ( JS, g water m sapwood s) was determined using the empirical calibration developed by Granie r (1985) and confirmed by Braun and Schmid (1999) and Clearwate r et al. (1999): JS = 119 t1.231 where t = ( Tm T) / T. Radial variations in sap flux density and the proportion of active conducting sapwood must be considered to avoid erro rs at extrapolating wa ter use (Hatton et al., 1990; Zang, Beadle and White, 1996; Lu, 1997; Clearwater et al., 1999). In all measured trees, at the same point as the sapflow probes, bark depth was measured with ba rk caliper, and sap wood thickness was measured using 0.5 cm diameter w ood cores. Color differences and changes in water content were used to identify sapwood a nd heartwood boundaries. Diameter with bark at breast height (DBH) was measured with diam eter tape. Cross sectional sapwood area was calculated by fitting the equation: ln(As)= 0 + 1 ln(DBH) where As is sapwood area at probe installation height (m2), DBH is diameter with bark at breast height (mm) and 0 and 1 are the fitted parameters, which were estimated from the regression intercept and slope, resp ectively, with the regressing of logtransformed As and DBH. Along the two seasons of measurements, changes in As we re determined with monthly measurements of DBH in all sampling trees. Daily As for each tree wa s determined with linear interpolation of As between 2 consecutive monthly measurements. To account for radial patterns in sap flux dens ity (Hatton et al., 1990; Lu, 2000; Jimenez et al., 2000; Wullschleger and King, 2000; Nadezhdina et al., 2002; James et al., 2003; Ford et al., 2004, 2007), variable length sap flow probes (VLP), as proposed by James et al. (2002), were installed in a subset of 3 trees (small, medium and large diameter) on both irrigated and control 27

PAGE 28

plots on the Florida mix family in one replicate. Sample trees were also selected using the stratification method, chosen sampling trees from acr oss the range of tree diameters. Probes were installed at 1.8 2.0 m height and inserted at 1 cm depth interv als, with the tip of the 10 mm probes reaching 1, 2, 3, 4, 6, 8, 10 and 12 cm depths (assuming to measure sap flux density in discrete depths, e.g., 0-1, 1-2, 11-12 cm). The probe for the outermost position was installed on the north side of the stem, and subsequent pr obes were installed clockwise around the stem at 45 intervals until the depth of the stem pith was reached, which was at 8, 10 and 12 cm for small, medium and large trees, respectively. For the radial profile sample trees, total tree sapflow was calculated following Hatton et al. (1990), as the sum of all individual sapflow estimations determined for each sensor along the xylem radius, where each 30-min Js was converted to sapflow (Q, 10-3m3min-1) by weighting each sap flux density measured at each depth by the sapwood area containing the annulus corresponding to each probe. For each tim e step, sapflow for the whole tree (QT, 10-3m3min-1) was calculated by summing all individual flows for all n individual annuli in each tree following Nadezhdina et al. (2002): QT = n i iiAJ1 where Ji and Ai corresponds to sap flux density and area for each individual annuli, respectively. Single point sensor sap flow wa s estimated by calculating sapflow for the outermost two rings (Q20), assuming that this estimation corresponded with regular 20 mm Gr anier-type sapflow probes: Q20 = J1A1 + J2A2 28

PAGE 29

where J1 and J2 correspond to sap flux density measured at 1 and 2 cm depth, and A1 and A2 are the areas of corresponding annuli. For each day and tree, a relationship between QT and Q20 was fitted. With a chisel, bark and cambium were remove d at the probe installation point to insert the sensors entirely into the xylem. Probes we re coated with thermally-conductive silicone grease before placement in the tr ees. All the sensors (regular 20 mm and VLP) were protected against radiation, thermal gradients and precipitation by reflective insu lation. Sensors were routinely replaced if probes were physically da maged or if null or negative readings were recorded. Leaf Area Index Using a Canopy Area Analyzer (LAI2000, Li-Cor Inc., Lincoln, NE, USA), projected leaf area index (LAI, m2 m-2) was measured in October 2005. To determine daily LAI for each plot, seasonal patterns were developed using information collected previously from periodic measurements from January-2002 to January -2005 (Samuelson, L., unpublished data) (See APPENDIX E). Canopy Conductance Canopy stomatal conductance for water vapor ( GC, m s-1) was calculated as in Granier and Loustau (1994): GC = )( sGDcRs EGa p La where Ga is the aerodynamic conductance (m s-1), EL is the stand transpira tion per unit leaf area (kg m-2 s-1), is the latent heat of water vaporization (J kg-1), is the psychometric constant (Pa K-1), s is the rate of change of saturati ng vapor pressure with temperature (Pa K-1), R is the radiation absorbed by the canopy (W m-2), is the density of dry air (kg m-3), cp is the specific 29

PAGE 30

heat of air (J K-1 kg-1) and D is the vapor pressure defic it (Pa). Aerodynamic conductance was calculated from the wind profile equa tion (Monteith and Unsworth, 2007): Ga = 2 0 2))/)(ln(( zdz wk where k is Von Karmans constant (0.41), w is wind speed (m s-1), z is anemometer height (m), d is roughness length (m) and z0 is displacement height (m). Values of d and z0 were set as 2/3 and 1/10 of canopy height, respectively (Phillips and Oren, 1998). Afte r correcting for temperature changes in water density, GC was transformed to molar units (mmolwater m-2 leaf area s-1). To reduce error due to instrument limitations on relative humidity measurements, GC was calculated only when D 0.6 kPa (Ewers and Oren, 2000). A reference GC ( GC ref) was calculated at D =1 kPa (Granier et al, 1996; Oren a nd Pataki, 2001). The response of GC to D was quantified using boundary line analysis (Ewers et al., 2001; Sc hfer et al., 2000). The upper boundary line for each plot was derived by binning GC data into 0.2 kPa D intervals (from 0.6 kPa to 4.6 kPa) and then selecting the highest 95% GC for any interval. For each plot, all upper GC values in each D interval were related to the natural logarithm of D (Granier et al, 1996): GC = GC ref m ln D where m is the slope of the regression f it, representing stomatal sensitivity to D (i.e. d GC /dln D ). Using diurnal values of GC (Phillips and Oren, 1998) diurnal average GC ( GC day) was calculated. Leaf-specific GC day ( GCla) was calculated dividing GC day by daily leaf area of each plot. Relative canopy conductance ( GC%) was calculated for each fa mily as the ratio between GCla of control over irrigated plots for each day. 30

PAGE 31

Whole-Tree Hydraulic Conductance Whole-tree sapwood-speci fic hydraulic conductance (KS-wt, molwater m-2 sapwood s-1 MPa-1 m1) was computed following the regression technique (Wullschleger et al., 1998) and calculated as (Phillips et al., 2002; Franks, 2004): KS-wt = hg Ewsl S where ES is transpiration rate per unit sapwood area (molwater m-2 sapwood s-1), l and s are leaf and soil water potential (MPa), respectively, w is the density of liquid water (Kg m3), g is the acceleration due to gravity (m s-2) and h is tree crown height (m). Tree height was measured on December 2005, one month after water pot ential measurements (Nov-15, 2005). As h is the vertical distance from soil to sites of evapora tion in crown needles and sampling tissue for water potential measurements was extracted from th e upper fifth of the crown in branches fully exposed to light, h was assumed to be 4/5 of total tree height. l was measured using a portable pressure chamber (PMS 1000, MPS Instrument Co., Corvallis, OR, USA) on one shoot tip from pre-dawn (5:30-6:00) to late af ternoon at 2 hour intervals, comp leting 4 to 5 measurements per tree throughout the day. Samples were collected on a subset of 4 tr ees per plot (on which sapflow was measured). Each measurement was comple ted within 3 minutes after shoot excision, covering the sample with wet towels inside a plas tic bag and stored inside an insulated box to minimize desiccation. Shoot sampling was carried out using a self propell ed telescopic lifting machine and pole pruner. Foliar Analysis Foliar sampling was carried out in December 2005 on the same 48 trees measured for water potential in November 2005. Needles from the la st flush with fully elongated needles were collected by excising tips of branches from the t op fifth of the crown. Once the branch was cut, 31

PAGE 32

needles were extracted and kept in labeled plastic bags on ice. In the laboratory, each sample was divided into two subsamples, one for chemi cal analysis of Carbon isotope composition ( 13C, ) and nitrogen concentration (N, %), a nd one for specific leaf area (SLA, cm2 g-1) determination. 13C on needle tissue was measured using a mass spectrometer and N was measured using a Europa Scientific ANCA-SL Stable Isotope Analys is System (Europa Scientific, Crewe, U.K.) at the Cornell Boyce Thompson Institu te Stable Isotope Laboratory. All-sided SLA was determined using the ratio between surface area and dry weight of needles. Individual needle surface area was calculated according to Murthy and Dougherty (1997) and Niinemets et al. (2001) from needle radius and length measured with a 10 x scaled magnifier and a digital caliper (CD-6, Mitutoyo, Kawasaki, Japan), respectively, on 10 needles per tree. After surface area was determined, the needles were oven-dried for 48h at 75C and weighed to the nearest 0.0001g (XA-100, Denver Instruments, Denver, CO, USA). Statistical Analysis Analysis of variance (ANOVA) was used to an alyze effects of irrigation and family in water relation traits, including B onferroni adjustments for differe nces in least square means (PROC MIXED, SAS Inc., Cary, NC, USA). The ANOV A model for the analysis is described in APPENDIX A. Repeated measures analysis was used to analyze time series data. Results Environmental Conditions Precipitation in both measurement years was below historic averages, but evenly distributed throughout the year. Total prec ipitation from July-December 2005 was 463 mm, 28.4% lower than the 30-year average for Bainbridge ( http://cdo.ncdc.noaa.gov/cgibin/climatenormals/climatenormals.pl May 2008). In 2006, total precipitation was 902 mm, 35.9% lower than the long-term average. For 2006, mean daily temperature during summer was 32

PAGE 33

very uniform, averaging 24.3 C; mean daily te mperature during winter was 13.6 C (Figure 21). Total daily PAR during winter was only 43.8 % of that received during spring. Average daylight vapor pressure deficit was similar during spring and summer (1.4 kPa), decreasing to mean values of 0.81 and 0.67 kPa during fall and winter, respectively (Figure 2-1). Radial Profile in Sap Flow The relationship between sapwood-related sa p flow and distance from bark had no distinctive shape, changing th e pattern of variation in Js during the year, depending on environmental conditions and tree size (Figures 22 and 2-3). For example, on summer days with high D and v (as August 17th, Figure 21), diurnal variation in Js for maximum conductivity xylem zones was similar between different tree si zes, with and without ir rigation. However, late in the season, on days when D was low and v on control plots was also low (as November 15th, Figure 2-1), this pattern varied depending on tree size and irriga tion treatment (Figure 2-2). In terms of radial variation in Js, for the same example days, la rge trees reached maximum midday Js at 4 and 2 cm xylem depth, for non-irrigated tr ees, and 8 and 10 cm xylem depth, for irrigated trees, on August 17th and November 15th, respectively (Figure 2-3). A strong relationship between Q20 and QT (p<0.0001, r2>0.93) was observed in all trees and days analyzed. Changes in the slope of th at relationship for diffe rent days were highly correlated with changes in environmental conditio ns affecting sapflow and radial profile in sap flux density. Using multiple linear regressions, D PAR and v of top 10-30 cm were incorporated into the model, in order to account for changes in the slope of the relationship between Q20 and QT. The final model used to transform sa pflow values estimated with regular 20 mm Granier-style probes to whole sapflow us ing radial profile information was: QT = 1.09186 Q20 + 1.54759 As + 0.00089165 D + 0.00001125 PAR 0.28217 v. The model explained 97% of the variation in QT (r2 = 0.968). Using this model, sapflow was estimated for 33

PAGE 34

each time step in all 96 trees measured only with probes inserted into the outer 20 mm of xylem. Sapwood area was determined for each tree using linear regression between DBH and As after log-transformation of both variables (p<0.0001, r2=0.981). The same linear model was used for all treatments (p>0.05): ln( As ) = 15.347 + 2.185 ln(DBH). Sap Flow After correcting sapflow estimations using Granier-type probes w ith the model that accounts for radial profile and environmental cond itions, sapflow was determined for each plot for all the measurement period (June-2005 to January-2007). After the irrigation treatment started (June-2005), soil water cont ent and the transpiration rate ti me series clearly separated for control and irrigated treatments in both gene tic families (Figure 2-5). The maximum daily transpiration rate was similar for both seasons, reaching a maximum of 4.3 mm day-1 in both seasons for irrigated plots. In 2006 on control plot s, the maximum transpiration rate only reached 2.6 mm day-1. On rainfed control plots, transpiration rate was highly dependent on rainfall events (Figure 2-4 and 2-6); monthly average da ily transpiration rate was 1.8 and 3.2 mm day-1 for summer months on control and irrigated plots, respectively. The average annual transpiration during the 2005 measurement season (163 days) was 248 and 369 mm, for c ontrol and irrigated plots, respectively, a 49% difference (p = 0.03 fo r irrigation effect). Th ere was no effect of families (p=0.29) and no family by irrigation interaction (p=0.09). In 2006, irrigated plots transpired 89% more than control plots (490 mm vs. 930 mm, for control and irrigated plots, respectively); there was no effect of family (p=0 .68) or interaction of family with irrigation treatment (p=0.41). As total rainfall dur ing measurement seasons 2005 and 2006 was 463 and 902 mm, respectively. Average tota l transpiration for control plots represented 53.5 and 54.3% of total rainfall for each season. Assuming an average daily irri gation input of 6.25 mm day-1, and discounting a total of 26 and 29 days of malfunctioning system during seasons 2005 and 34

PAGE 35

2006, respectively, average total transpiration fo r irrigated plots represented 28.0 and 31.0% of total water input (rainfall plus irriga tion) for seasons 2005 and 2006, respectively. Leaf Area Index No differences between treatments were observed in LAI measured on October 15th 2005 (Table 2-1). Average projected LAI for ir rigated and control plots was 4.35 and 4.5 m2 m-2, respectively (p=0.35); at the family level, LAI for FL and SC families was 4.4 and 4.5 m2 m-2, respectively (p=0.59). Soil Moisture Soil water content was highly affected by irriga tion treatment (drip-irrigation treatment in 2005, which ran from June to November; during season 2006, irrigation started in March and finished in October). On non-ir rigated plots, the 0-35 cm soil layer remained at the minimum extractable soil water values for long periods in both seasons (Figure 2-4b). Seasonal v average for control and irrigated plot s, including non-system functi oning periods, was 10.9 and 17.7% in 2005 and 11.3 and 17.1% in 2006, respectively. Relative water use from individual layers in the upper 2 m was determined from late February to late August 2006 for the FL fam ily only on non-irrigated plots (Figure 2-4c). Analysis of soil water movement and transpirat ion rate indicates that, across the season, daily transpiration rate matched with soil water extraction up to 200 cm depth (p=0.11; paired t-test) on non-irrigated plots. Average da ily water uptake from the 0 to 75 cm depth corresponded to 92 and 73% of the average daily transpiration rate for ir rigated and control plots, respectively. Dayto-day variation in transpiration rate followed a similar pattern to changes in ASW in the upper 35 cm of soil (Figure 2-5). A little more than half of the water used for transpiration was taken from the upper 35 cm depth, averaging across the season 58, 15, 17, 5 and 5% for the 0-35, 3575, 75-125, 125-175 and 175-225 cm depths, respectively (Figure 2-6a). 35

PAGE 36

Interaction between relative uptake from each layer and time (month) was significant only for water uptake at 0-35 cm depth between April and June (p=0.012). In or der to investigate the effect of irrigation, the same analysis was perfor med for irrigated and control plots, but only for 10-30 and 40-60 cm depth probes, because no deeper sensors were installed on irrigated plots. From the amount of water extracted from the first 75 cm, 67 and 86% were taken from the upper 35 cm for control and irrigated pl ots, respectively (Figure 2-6b). There was no significant time x irrigation interaction (p>0.05), i ndicating that th e average extraction pa ttern did not change during the evaluation period. Canopy Conductance Canopy conductance to water vapor was highl y affected by the irrigation treatment (Figure 2-6, p=0.0001 and 0.0099 for seasons 200 5 and 2006, respectively). There was no irrigation x family interaction in average GC (p>0.28 in all cases and seasons). For each year there was an irrigation x time in teraction (p<0.0001) associated with interruptions in the irrigation treatment. There was irrigation x family in teraction in the response of GC to vapor pressure deficit (Table 2-3). Under water-de ficit conditions (control) GC sensitivity to D and GC at D =1 kPa for the South Carolina Coastal Plain family (SC) were smaller than the North Florida source (FL) (p=0.01 and 0.03, respectively). Under non-water limited conditions (irrigated) both families presented no differences in d GC /dln D and GCref (p=0.08 and 0.44, respectively). The FL family had the same d GC /dln D and GCref under control and irrigated conditions (p=0.97 and 0.81, respectively). There were no si gnificant differences between genetic families and irrigation treatment in Dmax (D when GC equals zero). GC min, the minimum GC at any given D above 2 kPa, was higher on irrigated than on control plots (p=0.0002). At high D values (above 2 kPa, Figure 2-7), the minimum GC observed was always higher on irri gated than on control plots (p<0.001). 36

PAGE 37

There was no difference between irrigated and control plots in the relationship between sensitivity of canopy conductance to D (d GC /dln D ) and GCref (p= 0.471, Figure 2-7c). Whole-Tree Hydraulic Conductance There was a strong effect of Irrigation and no effect of family or their interaction on average KS-wt, pred and min (Table 2-3). Control and i rrigated treatments had average KS-wt of 0.83 and 2.87 mol m-2 s-1 MPa-1, respectively. Average pred for control and irrigated plots was 1.15 and -0.67 MPa, respectively. Average min on the control plots reached as low as -1.92 MPa, while on irrigated plots min was -1.39 MPa. Maximum leaf water potential gradient ( MPa ), calculated as the difference between min and pred, was not different between irrigation and family treatments (Table 2-3). Foliar Analysis There was a strong effect of irrigation and genetic family, but no interaction, on foliar N concentration and 13C (Table 2-3). There was no difference in all-sided specific leaf area (SLA) between treatments. Irrigated trees had higher foliar N concentration (p=0.005) and higher discrimination against 13C (more negative 13C; p=0.041) than non-irrigate d trees. At a genetic family level, the SC source showed higher N concentration and less discrimination against 13C than FL source (p=0.024 and 0.036, respectively). Discussion Depending on environmental conditi ons and tree size, radial pr ofile in sap flux density had no distinctive shape, with a cha nging pattern during the year. For the same species, Ford et al. (2004) found a Gaussian shape in the JS radial profile, with maximum midday JS occurring in the outer 4 cm of sapwood and decreasing towards the heartwood. In contrast, Jimenez et al. (2000) also reported that radial prof iles of sapflow were highly variab le in all measured trees, not finding distinguishable groups based on quantitative differenc es, and Wullschleger and King 37

PAGE 38

(2000) found no relationship between sapwood thic kness and fraction of functional sapwood in yellow-poplar trees. The non-uniform changes in radi al profile in sap flux density described in Figure 2-2 and 2-3 were similar to those presented by Nadezhdina and ermk (2000). They reported that after changes in v, radial profile in JS changed non-homogeneously across radius, decreasing more drastically, as soil drought, in zones of higher JS (see Figure 2-3 on Nadezhdina and ermk, 2000). This is an indicator of interacti ons between environment and radial profile in JS, related to differences in vulnerability to cavitation (D omec and Gartner, 2003) or to differences in water potential gradient (Domec et al., 2005) at different depth into conductive xylem. This evidence supports our approach not to use a single model that only accounts for differences in radial profile at midday at a particular time. The maximum daily transpiration rate was similar for both seasons, reaching a maximum as high as 4.3 mm day-1 in both seasons for irrigated plots; this value is concordant with Samuelson and Stokes (2006) who re ported for the same plots at age 5 a maximum transpiration rate of 3.9 mm day-1. Monthly daily average transpiration rate ( Eday) was negatively correlated to GCday in control plots (p<0.0001, r=0.719) but not correlated to GCday in irrigated plots (Figure 26). Using multiple linear regression, includi ng total daily radiation, average daylight D and daily average soil water content up to 35 cm, variation in Eday in control plots was explained principally by total dail y radiation (partial r2=0.859; p<0.0001) and to a much lesser extent by GCday, soil water content and D (partial r2=0.070, 0.014 and 0.008, respectively); under irrigated conditions, variation in Eday was explained only by changes in radiation (r2=0.987; p<0.0001). After pooling all monthly averages of cont rol and irrigated plots, variation in Eday was explained by changes in soil water content (partial r2=0.914; p<0.0001) and to a lesser extent by radiation (partial r2=0.0254; p<0.0001); the model also indicates that there was a slight, but significant, 38

PAGE 39

effect of genetic family, decreasing Eday by 0.119 mm day-1 for the SC trees under the same soil moisture and radiation conditions. For seasons 2005 and 2006 most of the water used for transpiration (58%) was taken from the upper 35 cm soil depth on rainfed conditions and, on average across th e season, only 10% of total transpiration was sustained from water be low 1 m depth. This resu lt is highly variable, depending on site characteristics, but several auth ors confirmed the trend that most of the water extraction in loblolly pine plan tations comes from top soil layers (Oren et al., 1998; Ewers et al., 1999 and Retzlaff et al., 2001a). Using water releas e curves developed for each soil layer, soil water potential for soil below 1 m was always below -1.2 MPa, reaching values below -1.5 MPa from April until late August (data not shown). Hacke et al. (2000), measured vulnerability to cavitation in loblolly pine roots indicating that at -1.5 MPa, 50 and 75% of loss in conductiv ity due to cavitation is reached for loam and sandy soils, respectively. Ewers et al. (2000) re ported 80% loss of conductivity in roots of 14year-old loblolly pine when xylem water poten tial reached -1.5 MPa. Assuming similar patterns of cavitation vulnerability on our site, we can su rmise that loss of conductivity was much higher than 50% during most of the season for roots found below 100 cm depth. Several authors have indicated that most of fine root s biomass is confined to top 20 to 50 cm soil (Ewers et al., 2000; Hacke et al., 2000; Retzlaff et al., 2001b), so the sm all amount of roots present in loblolly pine stands below 1 m, together with the low water potential presen t during most of the season, can help to explain the soil water extraction pattern at this site. Average rela tive water uptake from 035 cm depth showed interaction with time only between April and June (p=0.012) and for the rest of soil layers and periods evaluated there was no time effect on relative water uptake, 39

PAGE 40

meaning that the proportion of water uptake used for transpiration was constant across the season. Canopy conductance to water vapor was highly affected by soil water availability (Figures 2-6 and 2-8). This response has been widely repo rted for loblolly pine at the leaf-level under greenhouse (Bongarten and Teskey, 1986; Teskey et al., 1986) and field-growing conditions (Samuelson, 1998), and also at the canopy level (Oren et al., 1998). Ther e was a significant interaction between water availability and genetic family (GxE) in the response of GC to D This GxE interaction implies that under water-limite d conditions, the SC family trees had stronger stomatal control than FL trees, but this difference was not pres ent when water was not limiting. In an analysis of 4 provenances of loblolly pine, including Atlantic Coastal Plain (ACP, where family SC comes from) and Central Florid a (CF) sources, Sierra-Lucero et al. (2002) reported that although CF yielded 10% more aver age volume per land area than ACP sources at age 10, there was a large GxE interaction of fam ilies within the Florid a provenance, and that ACP families were more stable across sites than CF On a study of physiological traits of loblolly pine seedlings comparing ACP families with xe ric families from East Texas carried out under dry field conditions, Grissom and McKeand (2001) found no differences in midday light saturated net photosynthesis, but significant differences in midday stomatal conductance (gs); ACP families had lower gs and higher intrinsic water use efficiency than xeric provenance. Differences in GC sensitivity to D between SC and FL under water-limited conditions can be related to the fact th at FL trees originate in an environm ent with more summer rainfall, less summer moisture deficit and exte nded growing season than SC trees but when they are exposed to low soil moisture conditions they express less stomatal regulation to D compared with wellwatered conditions; as FL trees evolved in a much less water-limited environment, they show in 40

PAGE 41

a lesser way the drought avoidance ch aracteristic of reducing stomat al aperture. The slope of the relationship between d GC /dln D and GCref (0.633) was similar to the mean slope obtained from a large number of species (~ 0.6; Oren et al., 1999) and GCref explains 92% of variation in canopy conductance sensitivity to D At D values above 2 kPa, the minimum GC observed was always higher on irrigated than on contro l plots (p<0.001). This implies that, while water availability does not affect GC sensitivity to D or GCref, when PAR and temperature restrict GC (below the boundary line on Figure 2-7b), stom ata remain slightly open if v doesnt reach a critical value. For the same six-month period used for so il water extraction an alysis, half-hourly GC data was conditionally selected to minimize PAR or D constrains in GC (PAR >500 mol m-2 s-1 and D < 1.4 kPa). After expressing GC per unit leaf area ( GCla), relative conductance ( GC%) was calculated as daily means of control plots normalized by daily means of irrigated (well-watered) GCla for each genetic family within each replicate (F igure 2-8) and usi ng a logistic model GC% was fitted to available soil water (Lecoeur and Sinclair, 1996). The sensitivity of GCday to soil water content was not different between families and the same response curve can be used for both families (Figure 2-8d), because all parameters of the logistic model were not significantly different (p>0.40 for all 3-parameters). The response of canopy conductance to soil moisture was expressed in relative terms using a logistic model GC%=(0.834)/(1+(4.935)*exp((16.147)*ASW)). In our study, GC response to v presents a sigmoid shape with two characteristic response zones: When ASW > 30%, there was no effect on GC; when ASW < 30%, GC decreased linearly until it reached minimum valu es between 2-18% of well-watered plots and when ASW was zero (Figure 2-8c). This non-zero y-axis inter cept was explained because ASW was calculated from the 0-75 cm soil depth, from where 73 and 90% of transpiration water on 41

PAGE 42

control and irrigated plots, resp ectively, so water uptake could theo retically continue if ASW in the upper 75 cm reaches zero. Whole-tree sapwood-specifi c hydraulic conductivity ( KS-wt) was similar to those reported by Samuelson and Stokes (2006) for the same study site at age 4.5 years. No differences in KS-wt, pred and were found at that time between control and irrigated plots, suggesting that the treatment differences found in our study were caused by in-situ xylem cavitation rather than tracheid morphological changes resulting from higher water availabilit y. After expressing the difference between KS-wt of irrigated and cont rol plots relative to KS-wt of irrigated plots as a surrogate of percentage loss of conductivity ( PLC ,%) (Tyree et al., 1992), and using pred as an index of s of rooting zone, a strong rela tionship (p=0.008) appears between PLC and pred (Figure 2-9). For the range of data, reductions in pred explained 86% of the variability in PLC When pred for control plots reached -1.3 MPa, KS-wt was only 20% of well-watered trees; this result is comparable with vulnera bility curves measured on roots of the same species by Ewers et al. (2000), where PLC was 80% at -1.5 MPa. The maximum transpiration-induced water pot ential gradient from roots to shoots ( ), was relatively constant across treatments (p= 0.52) and genetic family (p=0.72), averaging 0.75 MPa. This relatively constant is similar to that reported by Samuelson and Stokes (2006), which was maintained across treatments duri ng summer months on an average value of 0.86 MPa. On the other hand, minimum daily xylem water potential ( min) was different between control and irrigated tr ees (p<0.0001), being, on average, 0.5 MP a lower for control trees. This response in leaf water potentia l corresponds to that presente d by Franks et al. (2007) as isohydrodynamic, where a strong stomatal control maintains re latively constant (Table 23; p=0.522), but at the same time allowed leaf to fluctuate dramati cally in synchrony with s 42

PAGE 43

(Table 2-3; between control and irrigated plots, p=0.021 and <0.0001 for pred and min, respectively). For the day of leaf water poten tial measurement, average canopy conductance per unit leaf area ( GCla) was highly correlated with pred and KS-wt, following a sigmoid curve-shape (Figure 2-9). Teskey et al (1986) reported that gs declined slightly until xylem pressure potential reached threshold point of approxi mately -1.0 MPa, after which gs declines rapidly and total stomatal closure occurs at -2.0 MP a; visual analysis of their results indicate that approximately 50% of loss of gs is reached around -1.2 MPa. In our study, above pred ~ -1.1 MPa, GCla increased linearly until plateauing above -0.6 MPa. The relationship between GCla and KS-wt follows the same pattern, with negligible increment in GCla if KS-wt is higher than, approximately, 3.0 mol m-2 s-1 MPa-1. Canopy conductance per unit leaf area ( GCla) was very stable across treatments and was not correlated with (Figure 2-9; p=0.61). Carbon isotope composition ( 13C) is an indicator of the relationship between stomatal conductance and photosynthesis rate (Farquhar et al., 1989). Increases in 13C (less negative value) can be a result of lower stomatal condu ctance or higher photosynthetic capacity, or both. 13C was highly affected for both irri gation and genetic family; this response to water availability was comparable to that reported for the sa me species by Yang et al. (2002) and for Pinus radiata (Korol et al., 1999) and Eucalypt us globulus (Macfarlane et al ., 2004). In our study, at family level, higher foliar N and less negative 13C on the SC trees were indi cators of higher water use efficiency of this family in comparison to the FL source. This is cons istent with the reduced stomatal control we observed in the FL family compared with SC. 13C was strongly negatively correlated with mean daily canopy conductance (calculated as the mean value of daylight canopy conductance for the period between irrigation tr eatment start, on June 2005, to foliar sampling day, in December 2005) on control plots but not on irrigated plots (p=0.007 and 0.58, 43

PAGE 44

respectively; Figure 2-10a). 13C was strongly positively correlat ed with foliar N concentration on both, the control and irrigated plots (Figure 2-10b) and the slope of this relationship was not significantly different between treatments (p=0.13), but trees under water stressed conditions showed, for the same 13C, lower foliar N concentration comp ared with irrigated; the difference in intercept of this relationship (Figur e 2-10b) is presumably caused by reduced GC in control plots. As foliar N was higher on irrigated compared with the control plots (p=0.005) (similar to Yang et al., 2002 and Albaugh et al., 2004), for irrigated trees variability in 13C could be explained by changes in photosynthetic capacity while for water stressed trees, changes in 13C could be explained by both, vari ation in stomatal conductance a nd photosynthetic capacity. As larger foliar N levels were associated with increased photosynthetic capacity, the positive correlation between 13C and foliar N could indicate higher water use efficiency in plots with both higher, foliar N and 13C. Yang et al. (2002) argued that the higher foliar N concentration on irrigated plots could be an effect of reductions in fine root system growth due to drought; in our study site Samuelson et al. (2008) reported that total root biom ass was significantly higher in irrigated plots and Fabio et al. (1995) also reported si gnificant increment in fine roots of 6 yearold irrigated Eucalyptus globulus but additional research is needed in order to establish a causal relationship. Conclusion For 11 year-old loblolly pine plantation, increasing water availability via irrigation increased transpiration rate and GC and whole-tree water conduction efficiency (KS-wt) was maintained at high levels due to avoidance of xyl em embolism. Loblolly pine tends to maintain constant water potential gradient from roots to shoots at a co st of loss of conductivity under water-limited conditions. The two genetic fa milies evaluated showed differences in GC 44

PAGE 45

sensitivity to D under water-limited conditions, foliar N and 13C, maybe reflecting differences in adaptive traits related with ambient hum idity and water availability. These different environments maybe drove different evolving pathways expressed in higher water use efficiency in trees from the family of Atlantic Coastal Plain than trees from the mix of north Florida families. 45

PAGE 46

Table 2-1. Mean diameter at br east height (DBH), height, basal area (BA) and leaf area index (LAI) for Irrigation x Family treatments. Control Irrigated p > F FL SC FL SC Irrig Fam IrrigxFam DBH (cm) 18.5 18.9 20.3 20.7 0.026 0.018 0.527 Height (m) 13.8 14.5 15.8 16.4 < 0.01 0.019 0.895 BA (m2ha-1) 31.2 29.4 35.8 36.1 0.0452 0.525 0.406 LAI (m2m-2) 4.2 4.5 4.6 4.4 0.348 0.597 0.083 Means at age 10 when the study started (Jun-2005), LAI was measured on Oct-2005. FL: Florida family; SC: South Carolina Coastal Plain family. p-values using mixed model procedure for split-plot design. 46

PAGE 47

Table 2-2. Diameter at breast height (DBH) and number of variable le ngth sapflow probes (in parenthesis) installe d on FL family trees. DBH (cm) Small Medium Large Control 16.7 (6) 19.6 (7) 23.0 (8) Irrigated 17.2 (6) 20.9 (7) 27.1 (8) Maximum depths reached with 6, 7 and 8 probes were 8, 10 and 12 cm, for small, medium and large trees, respectively. 47

PAGE 48

Table 2-3. Mean d GC/dln D GCref, GC min, Dmax, KS-wt, pred, min, SLA, N% and 13C for irrigation (control and irrigated) a nd family (FL and SC) treatments. Control Irrigated p > F FL SC FL SC Irrig Fam IrrigxFam GCref 136.57 111.59 135.98 148.64 0.309 0.188 0.008 -d Gc/dln D 82.70 66.62 79.67 83.80 0.572 0.158 0.042 Dmax 5.24 5.34 5.69 5.96 0.230 0.484 0.737 GCmin 0.72 1.14 15.89 15.97 0.0002 0.772 0.845 KS-wt 0.843 0.824 2.918 2.825 0.029 0.779 0.855 pred -1.172 -1.122 -0.677 -0.660 0.021 0.111 0.366 min -1.939 -1.916 -1.404 -1.389 <.0001 0.537 0.889 0.766 0.793 0.727 0.729 0.522 0.727 0.764 SLA 109.48 100.62 108.20 100.97 0.108 0.786 0.635 N% 1.36 1.42 1.45 1.52 0.005 0.024 0.836 13C -29.57 -29.07 -30.12 -29.62 0.041 0.036 0.997 FL: Florida family; SC: South Carolina Coastal Plain family. p-values using mixed model procedure for split-plot design. 48

PAGE 49

-10 -5 0 5 10 15 20 25 30 35 Temperature (oC) 0 10 20 30 40 50 60 70 Precipitation (mm d-1) 0 5 10 15 20 25 30 PAR (mol m-2 d-1) 0.0 0.5 1.0 1.5 2.0 2.5 3.0 D (kPa) Tmean Tmin Precipitation PAR VPDJul Ago Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Ago Sep Oct Nov Dec b a Figure 2-1. Environmental conditions thr ough study period. (a) Daily mean (Tmean) and minimum (Tmin) temperatures and to tal precipitation; (b) daily sum of photosynthetically active radiation (PAR) and daily average of vapor pressure deficit ( D ) during daylight hours over the study period. 49

PAGE 50

0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 1 2 3 4 6 8 10 12 a b c d e f g h i j k l 6 12 18 24 6 12 18 24 6 12 18 246 12 18 246 12 18 246 12 18 24 Time of Day (h) Control Small Medium Large Irri g ated Small Medium Large Js (gm2 s1 ) Js (gm2 s1 ) August 17th November 15th Figure 2-2. Diurnal sap flux density ( Js) patterns at different radial positions for six selected trees of FL family on control and irrigated plots at two contrasting days during season 2006, August 17 (a f) and November 15 (g l). 50

PAGE 51

Sapflow (10-3 m3 30m-1) 0.0 0.2 0.4 0.6 0.8 Sapflow (10-3 m3 30m-1) 0.0 0.2 0.4 0.6 0.8 Control Small Control Medium Control Big Irrigated Small Irrigated Medium Irrigated BigJs (g cm-2 s-1) 0 4 8 12 16 20 Js (g cm-2 s-1) 0 4 8 12 16 20 6 10 14 18 22 Time of da y ( hour ) 0 2 4 6 8 10 12 Distance From Bark a b c d Figure 2-3. Diurnal variation in we ighted sapflow (a, b) and radial profile in sap flux density at noon (c, d) for six selected trees of FL fa mily at two contrasting days during season 2006, August 17 (a, c) and November 15 (b, d). 51

PAGE 52

Jul Sep Nov Jan Mar May Jul Sep Nov 2005 2006Transpiration (mm) 0 1 2 3 4 5 Soil Water Content 0 75 cm depth (mm) / Rainfall (mm) 0 10 20 30 40 50 60 70 80 90 Control FL Control SC Irrigated FL Irrigated SCa b Figure 2-4. Seasonal pattern in tr anspiration (a), soil water conten t and rainfall (b) for loblolly pine stands under irrigation treatment (contr ol and irrigated) in cluding two different genetic families (FL and SC) and available soil water for five different soil layers on non-irrigated FL family plots on a sixmonth period during season 2006 (c). Each symbol is the mean of three replicate plots. 290 5 10 25 35 40 0 35 cm Available Soil Water (mm) 35 75 cm Feb30-Mar29-Apr29-May28-Jun28-Jul27-Aug 15 20 30 75 125 cm 125 175 cm 175 225 cmc 52

PAGE 53

% Soil Water Extraction from 0 200 cmMarch 20 50 100 150 200 0.0 0.2 0.4 0.6 0.8 1.0 April 20 50 100 150 200 May 20 50 100 150 200 June 20 50 100 150 200 0.0 0.2 0.4 0.6 0.8 1.0 July 20 50 100 150 200 August 20 50 100 150 200 % Soil Water Extraction from 0-75 cm0 35 cm 0.0 0.2 0.4 0.6 0.8 1.0 Control35 75 cm March April May June July August 0.0 0.2 0.4 0.6 0.8 1.0 Irrigated March April May June July August a b Figure 2-5. Proportion of soil wate r extraction used for transpira tion from the 0-200 cm depth on nonirrigated plots (a) and from the 0-75 cm depth on control and irrigated plots (b) for FL family plots. 53

PAGE 54

E (mm day-1) 0 1 2 3 4 GCday (mmol m-2 s-1) 0 20 40 60 80 100 120 A S O N D J F M A M J J A S O N D 2005 2006 A S O N D J F M A M J J A S O N D 2005 2006ac bd Figure 2-6. Monthly averag e transpiration rate ( E ) (a, c) and mean daily canopy conductance ( GCday)(b, d) for control and irrigated treatm ents over the study period. Error bars indicate standard error. Letters in xaxis indicate month fr om August (A) 2005 to December (D) 2006. 54

PAGE 55

0 20 40 60 80 100 120 140 160 Gc (mmol m-2 s-1) 0 20 40 60 80 100 120 140 160 Gc (mmol m-2 s-1) -dGc/dlnD 50 60 70 80 90 100 110 120 Control Irrigated0 1 2 3 4 5 D (kPa) 100 120 140 160 180 GCref (mmol m-2 s-1) a b c Figure 2-7. Examples of relationship between GC and D for FL family under control (a) and irrigated (b) conditions (upper line represents boundary line analysis fitting) along the study period. Relationship between se nsitivity of canopy conductance to D (d GC/dln D ) and canopy conductance at D =1 kPa ( GCref) for all plots (c). 55

PAGE 56

Gcday (mmol m-2 s-1) 0 20 40 60 80 100 120 Gc% ( %, relative to well watered) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Gc% (%) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 ASW (mm) 0 20 40 60 80 FL SC FL SCIrrigated Controld a b c Figure 2-8. (a) Mean daily canopy conductance ( GCday), (b) normalized canopy conductance ( GC%), (c) available soil water (ASW) for 0-75 cm depth and (d) the relationship between ASW and normalized canopy conductance, for a six-month period on year 2006 for loblolly pine under irrigation trea tment (control and irrigated) including two different genetic families (FL and SC). 56

PAGE 57

pred (MPa)GCla (mmol m-2 s-1) -1.4-1.2-1.0-0.8-0.6-0.4 0 5 10 15 20 Control FL Control SC Irrigated FL Irrigated SC (MPa)GCla (mmol m-2 s-1) Figure 2-9. Relationships between mean daily canopy conductance (GCla) and pre-dawn water potential (pred) (a), maximum daily wa ter potential gradient ( ) (b) and whole-tree sapwood-specific hydraulic conductance ( KS-wt) (d) for loblolly pine under control and irrigation treatments, including two different genetic families (FL and SC). The loss of conductivity of control in relation to irrigated plots ( PLC ) is presented including both genetic families (c). All data were calculated with data collected the same day (Nov-15th, 2005). 0 0 5 10 15 20 .50.60.70.80.91.0 pred (MPa)PLC (%) -1.4-1.3-1.2-1.1-1.0-0.9 0.5 0.6 0.7 0.8 0.9 1.0 FL SCKS-wt (mol m-2 s-1 MPa-1)GCla (mmol m-2 s-1) 0.00.51.01.52.02.53.03.54.04.5 0 5 10 15 20 a b c d 57

PAGE 58

GCday (mol m-2 day-1)13C () 2030405060708090 -30.5 -30.0 -29.5 -29.0 -28.5 N (%)13C () 1.31.41.51.61.7 -30.5 -30.0 -29.5 -29.0 -28.5 Control r = -0.93, p = 0.007 Irrigated r = 0.28, p = 0.581 Control r = 0.90, p = 0.012 Irrigated r = 0.95, p = 0.003ab Figure 2-10. Relationships betw een mean daily canopy conductance ( GCla) (a) and foliar nitrogen concentration (N) (b) with 13C for control and irrigated plots. 58

PAGE 59

CHAPTER 3 WATER AVAILABILITY AND FAMILY EFFECTS ON WOOD PROPERTIES OF LOBLOLLY PINE (PINUS TAEDA L.) Introduction In the Southeast United States there are more than 13 million hectares of southern pine plantations (Fox et al., 2007). Around 58% of total timber harvested and close to 75 % of the total tree planting each year in the country is made in the Southeast industrial plantations (Jordan et al., 2007; McKeand et al., 2003), constituting the la rgest timber production zone in the world (McKeand et al., 2003; Allen et al., 2005). Loblolly pine ( Pinus taeda L.) is the most important commercial timber species in the Southeastern United States (more than 84% of seedlings planted are loblolly pine, McKea nd et al., 2003); it is one of the fastest growing pines and thrives on various sites from east Texas to southern Mi ssouri to north Florida to south New Jersey. Loblolly pine wood is used to produce both fiber-based and solid wood products and the knowledge and understanding of factor s controlling wood properties va riability is fundamental in wood production industry, where increases in productivity are re quired due to rising production costs and competition with other wood-producing countries. Tree improvement programs are key components of the success of the loblolly pine forest industry (McKeand et al., 2003). These improve ment programs have focused primarily on volume growth, disease resistance, tree form and wood properties. Specific gravity (SG) is the most important wood property in determining th e performance of products manufactured from loblolly pine (Jordan et al., 2008); SG is positively correlated with stiffness and strength and also with pulp yield (Panshin and de Zeeuw, 1980) a nd is also an important determinant of paper quality (Jordan et al., 2008). Small increases in SG due to tree improvement (genotype) or silviculture (environment) can have a large impact on wood production and value (Panshin and de Zeeuw, 1980). 59

PAGE 60

On a geographic scale, Megraw (1985) indi cated that there was a tendency for SG to increase with decreasing latitude ; this is probably due to length of growing season, with a longer season producing more latewood. When comparing wood properties on loblolly pine provenance studies (Jett et al., 1991; Jayawi ckrama et al., 1997), in general, genotypes from Florida (e.g. Gulf Hammock and Marion County) ha d lower SG compared with Atlantic Coastal provenances. Water availability is one of the main f actors controlling tree growth and species distribution (Lambers et al., 2006) and can influence the characteristics of the secondary xylem (Panshin and de Zeeuw, 1980). Dougherty et al. ( 1994) indicated that diameter growth cessation date was dependent on soil moisture conditions a nd transpiration rate a nd Cregg et al. (1988) concluded that among-year differences in SG a nd latewood percentage (LW%) were associated with the amount of rainfall during the summer (years with high summer rainfall resulted in higher SG and LW% compared to dry summers). In this study we examined the effect of water availability and ge netic family on growth and wood properties on a mid-rotatio n loblolly pine plantation. Wate r availability was controlled by an irrigation treatment on two fast growing fam ilies that received the extra water input since plantation establishment. We hypot hesize that i) Diameter growth cessation date is dependent on soil moisture conditions and transpiration rate, ii) Irrigation treatments will extend the period of xylem growth, resulting in higher LW% and higher SG and iii) Florida source material will have different wood quality characteristics in response to irrigation than Atlantic coastal plain material. Specifically, the objectives of this study were to determine, at age 11 years: i) the effect of water availability and genetics (families) on date of diameter growth cessation and it relationship with SG and LW%, and ii) for long te rm responses, the effect of water availability and genetic families on properties of the complete wood core. 60

PAGE 61

Materials and Methods Site and Stand Description The study took place in an irriga tion and genetics experiment established in January 1995 by International Paper, Inc. in the Upper Coasta l Plain 22 km west of Bainbridge, GA (30 N latitude and 84 W longitude). Soils at th is location were classi fied as well-drained Grossarenic Paleudults, with 50 cm sandy lo am over sandy clay loam (Samuelson, 1998). The study consisted of four treatments: two ge netic entries plus two water availability entries repeated three times in a split-plot, ra ndomized complete block design (12 plots total), with irrigation as the whole-plot and genetic family as the sub-pl ot. These treatments represented a subset of the whole experiment, which also included two more families, fertilization and pest control treatments. Each measur ement plot (excluding 2 treated buffer rows) had an area of 0.026 ha, containing 28 sample trees planted on a 2.4 m x 3.7 m spacing. More details about the study site have been described by Samuelson (1998) and Samuelson et al. (2004 and 2008). The genetic treatments consisted of two open polli nated second generation improved families: one originating from a mixture of northern Florida half-sib families (FL), and a second originating from a single half-sib family from the Sout h Carolina Coastal Plain (SC). The irrigation treatments were irrigated (dai ly drip irrigation) and a non irrigated control treatment. The irrigation scheme has been operating since1995 a nd is described in detail in Samuelson (1998) and Samuelson et al. (2008). During 2005 a da ily drip irrigation rate of 6.25 mmday-1 was applied from June to November (total seasonal water addition was 631 mm). Meteorological Measurements Meteorological data were recorded form June 2005 to January 2007 with an automatic weather station located in an open area adjacent to the stand. Precipitation was recorded with a tipping bucket rain gage (TR525 -I, Texas Electronics, Dallas TX, USA). The sensors were 61

PAGE 62

measured each 30 s with an automatic datalogge r (CR10, Campbell Scientific, Logan, UT, USA) and were stored as 30 min averages. Long te rm meteorological measurements from stand establishment (January 1995) until the beginni ng of our measurements (June 2005) were obtained from a weather station in stalled by International Paper in an area close to the study site (30 N latitude and 84 W longitude) ( http://cdo.ncdc.noaa.gov/ cgi-bin/cdo/cdoprod.pl October 2008 ). Soil Moisture In order to assess the effectiv eness of irrigation applicati on, vertically oriented, 20 cm long capacitive soil moisture probes (EC-20, D ecagon Devices, Pullman, WA, USA) were used to measure volumetric soil water content ( v, cm3 cm-3). In June 2005 probes were installed at 0.1-0.3 m depth on the bed and interbed locations of each plot. Specific calibrations were developed from soil samples taken 0.5 m away fr om each probe; soil samples were collected from the same depth where the probes were installed. Soil-specific calib rations had no difference with the manufacturers defau lt calibration model to transform sensors outputs (mV) to v. Diameter and Basal Area Growth During the 2005 growing season, monthly change s in diameter at breast height (DBH, mm) and basal area (BA, m2 ha-1) were measured on eight trees pe r plot, previously selected for sapflow measurements (See Chapter 2). To estimate BA, plot-specific expansion factors were determined by measuring all the trees of all plots at the beginning and at the end of the study. For each plot, cumulative basal area growth was plotted against time, and basal area cessation day was determined as the date (day of year) when 95% of the total growth was attained (Hanover, 1963; Jayawickrama et al., 1998) using a non-lin ear fitting with Chapman-Richards model (PROC NLIN, SAS Inc., Cary, NC, USA) : 62

PAGE 63

BA = cX)(-b)e (1a where a, b and c are curve fitting parameters, and X corresponds to day of year of Julian calendar. Wood Properties In April 2006, 12 mm increment wood cores that crossed from ba rk to pith were extracted at breast height on the same trees used for DBH measurements. Wood co res were labeled, dried and sent to USDA-Forest Serv ice Forestry Sciences Laboratory in Athens, GA, for x-ray densitometry analysis (Hoag and Krahmer, 1991; Clark et al., 2006; QTRS-01X, Quintek Measurement Systems, Knoxville, TN). Latewood specific gravity (LWSG), latewood percentage (LW%, %), ring specific grav ity (SG), latewood width (LWW, mm) and ring width (RW, mm) were determined for each ring on all 96 sample trees. In the laboratory the increment cores were glued to core holders and sawn into 2 mm th ick strips. X-ray densitometry is based on the relationship between x-ray attenuation and density (Hoag and Krahmer, 1991): I = I0 t -e where I is the intensity of the ra diation beam after passing though the sample (attenuated), I0 is the intensity of the radiation beam without passing though the sample (unattenuated), is the sample linear attenuation coefficient (cm-1) and t is the sample thickness (cm). As I and I0 are measured for each point in the profile, the is calculated for each point in the profile. After estimation, sample density ( g cm-3) is calculated following the relationship: = m where m is the sample mass attenuation coefficient (cm2 g-1). SG is dimensionless and corresponds to the ratio of the we ight of a wood sample in relation to the weight of an equa l volume of water at a standard temperature (Larson et al., 2001). 63

PAGE 64

SG of earlywood and latewood from each annual ri ng for each radial strip was determined at 0.06 mm intervals. A SG value of 0.48 was used to distinguish earlywood from latewood (Clark et al., 2006; Jordan et al., 2007) LW% was calculated for each ring as the percentage of latewood area to total growth ring area (Jayawickrama et al., 1997). Weighted whole-core specific gravity (WCSG) and latewood percentage (WCLW%) were calculated as the sum of the product of SG or LW% of each ring by the corresponding ring basal area (Jordan et al., 2007). Statistical Analysis Analysis of variance (ANOVA) was used to an alyze effects of irrigation and family in wood properties, includi ng Bonferroni adjustments for differe nces in least square means (PROC MIXED, SAS Inc., Cary, NC, USA). The ANOVA model for the analysis is described in APPENDIX A. Repeated measures analysis was used to analyze time series data. Results Environmental Conditions Total yearly precipitation in 2005 (1490 mm) wa s evenly distributed along the year and was 5.8% higher than historic averages (1409 mm). Total precipitation from July to December 2005 was 574.8 mm, 4.3% higher than the 30year average for Bainbridge, GA ( http://cdo.ncdc.noaa.gov/cgi-bin/ climatenormals/climatenormals.pl May 2008 ). However, from September 1st and November 19th (79 days span) only 36 mm of precipitation occurred as rain (Figure 3-1). Yearly rainfall records from 1995 to 2006 indicate that prec ipitation was in most cases lower than normal, being between 15 to 22% lower from 1997 to 2001, and slightly higher, between 3.5 to 5.8% in 1995, 1996 and 2005 (Figure 3-4). 64

PAGE 65

Soil Moisture Soil water content was highly affected by the irrigation treatment. Seasonal v average for the control and irrigated plot s, including non-system func tioning periods, was 10.9 and 17.7% during the 2005 growing season (Figure 3-1). Diameter and Basal Area Growth and Day of Growth Cessation There was a significant effect of irrigation on cumulated BA at age 11 (p=0.002; Table 31), but no family or treatment interaction effects. On an individual tree basis, using both DBH (outside bark) measured on December 2005 or cumulative ring width using whole wood core (tree radius without bark), there was a significant effect of irrigation (p=0.0003 and p=0.0002, respectively) and family (p=0.039 and p=0.039, resp ectively); the interaction effect was not significant in either case (p=0.577 and p=0.757, respectively). This difference between ground area and individual tree basis is explained by the higher mortality observed on SC (14.9%) compared with FL (5.38%) (p=0.004). There was a significant effect of irrigation for the day of growth cessation (p=0.036); the BA growing period on irrigated pl ots was 24.6 days longer than th e control plots (Figure 3-2). There was no difference between families (p=0 .87) and no irrigation by family interaction (p=0.69) on growth cessation date. On average, ir rigated plots terminated basal area growth at day 275 (October 1). LW% and SG where correlated with BA gr owth cessation day for 2005 (Figure 3-3a and 3b). Changes in the date of BA growth cessati on explained 33.5 and 39.6% of the variability in LW% and SG, respectively. Wood Properties The first two growth years were eliminated fr om the analysis for tw o reasons: (1) because wood cores were sampled at breast height and only 29% of the trees reached that height in year 65

PAGE 66

1; and (2) even though around 94% of the trees showed a growth ring in the sampled wood core in 2006 (for 65% of the trees that was the first ring), those ri ngs were mostly pith-associated wood (Larson et al., 2001). There was a signif icant effect on SG and LW% of irrigation (p=0.027 and 0.017, respectively) and family (p=0.019 and 0.050, respectively), with no irrigation x family interactions (p=0.324 a nd 0.142, respectively). Wood properties were also dependent on time (p<0.0001), with a significant time x irrigation inte raction (p=0.055 and 0.086 for SG and LW%, respectively); family x year interaction was not sign ificant (p=0.219 and 0.594 for SG and LW%, respectively). During the 2003 (when irrigation system was working only intermittently) and 2004 growing season (when the system was not functioning) (Samuelson L., personal communication), there were no differences between irrigated and control plots in SG (p=0.574 and 0.459) or LW% (p=0.522 and 0.179). For 2005, when irrigation was re-activated in June, there was a significant effect of irrigation on SG and LW% (p=0.0047 and 0.0275, respectively). Average SG for the control and irrigate d plots was 0.511 and 0.547 respectively. LW% was increased from 50.8% on control plot s to 57.8% on irrigated plots. At a family level, SC showed higher SG and LW% than FL after age 5, and th ose differences were maintained even when irrigation was suspended (p<0.05; Figure 3-4). Latewood specific gravity (LWSG) was not affected by irrigation (p=0.708) or family (p=0.421) and no irrigation x family interaction was detected (p=0.919). There was a tendency to steadily increase LWSG from initial values of about 0.43 in 1995 (similar to ring SG at the same ag e) to a maximum of approximately 0.696 in 2004 (Figure 3-4). There was a positive and strong relationship between the amount of irrigation applied between July and November and the difference in latewood area between irrigated and control 66

PAGE 67

plots (p=0.009; r2=0.65; Figure 36). In contrast, family differe nces in latewood production were not associated with precipitation during each year; Figure 3-5f shows a relatively constant difference in latewood area, indepe ndent of rainfa ll or irrigation. In terms of whole-core weighted SG and LW%, irrigation showed a marginal nonsignificant effect on WCSG and WCLW% (p=0.052 and 0.112, resp ectively), aver aging across families WCSG values 0.417 and 0.438 and WCLW% values of 34.80 and 37.29%, for control and irrigated trees, respectivel y. The SC family had higher WCSG and WCLW% than the FL source (p=0.008 and 0.029, respectiv ely), averaging across irriga tion treatments WCSG values 0.415 and 0.439 and WCLW% values of 33.56 and 38.53% for FL and SC trees, respectively; no irrigation by family interaction was detected for both variables (p>0.629). Annual Ring Growth After excluding rings from 1995 and 1996, there wa s a significant effect of irrigation on annual ring area (Aring, cm2), earlywood ring area (AEW, cm2) and latewood width (ALW, cm2) (p=0.0001, 0.002 and <0.0001, respectively; Figure 3-5); however the response to water availability was also dependent on year (p<0.05 fo r year by irrigation inte raction). Irrigated and control plots did not differ in Aring growth on years 2002, 2003 and 2004 (these two last years received scarce or null irri gation). In the case of AEW the irrigation treatmen ts were only different between years 1997 to 1999 (p<0.0001). The effects of irrigation on ALW was maintained until year 2002 (p<0.05), three years longer than Aring, then during insufficien t irrigation years (2003 and 2004) there was no difference in latewood pr oduction, but during year 2005, when irrigation was re-started, the effect water supply was significant again (p=0.0016; Figure 3-5e). At a family level, even though SC showed a tendency to grow more in AEW than FL (Figure 3-5b), only in years 1998 and 2005 was that difference significant (p<0.025). AEW was only different between 67

PAGE 68

SC and FL in 1997 (p=0.0004), but ALW was significantly higher in SC during most of irrigated years since year 1999, only during season 2003 there were no significant differences (p=0.134). Statistical Analysis Analysis of variance (ANOVA) was used to an alyze effects of irrigation and family in wood properties traits, in cluding Bonferroni adjustments for di fferences in least square means (PROC MIXED, SAS Inc., Cary, NC, USA). Repeat ed measures analysis was used to analyze time series data. The ANOVA model for the an alysis is described in APPENDIX A. Discussion Increasing summer and fall water availability (via irrigation from June to November) increased SG and LW% by increasing the growin g season. The date of basal area cessation was shifted 24.6 days and SG and LW% were incr eased 0.036 and 7%, respectively, on irrigated trees. SG of latewood was not different between the control and irrigate d trees (p=0.122), but latewood area was increased signif icantly (p=0.0016) on irrigated co mpared with control trees. From the different phases of stem growth de scribed by Dougherty et al (1994) (cell production by cambium, radial growth of cells and seconda ry cell wall thickening and lignification), it appears from this study that the two firsts were the stages that were affected by water limitation. Dougherty et al (1994) concluded that diameter growth cessation depended on soil moisture and evaporation demand and, in the ab sence of drought, diameter growth can continue until late in the season when other factors, as temperature and/or photoperiod, trigger cessation. Changes in the date of BA cessation through irrigation were well correlated with changes in wood properties. Long-term effects of irrigation on SG, discarding data from years 1 and 2, showed null effect on years 3 and 5 and negativ e effect on year 4 (lower SG on irrigated plots, p=0.042); this 68

PAGE 69

response was associated with maximum response in cross section area growth, where earlywood growth was remarkably greater on irrigated trees but not different from year 7 and on. Early responses to irrigation observed in ring ar ea and earlywood area can be related to LAI deployment response to irrigation. Allen et al. (2 005) indicated that wa ter availability was thought to have more effect on growth efficiency than leaf area deployment in southern pines; nevertheless Samuelson et al. (2004) reported, for the same st and where our study was carried out, that peak annual projected LA I on irrigated plots was around 3.8 m2 m-2 at age 4 (year 1998), in contrast, the control plots had an annual LAI peak of 2.0 m2 m-2. After that age control plots continued to increase steadi ly in LAI reducing the differences but still significant until year 6. Between ages 8-10 year, Samuelson et al. (2008) reported no differences in peak LAI between control and irrigated plots, re sults concordant with our LAI measurements during October 2005 (Table 3-1). After this initial effect of water availability until the sta nds reached maximum LAI, growth in earlywood was not different between irrigated and control trees (Figure 3-5c). Independent of this effect on earlywood forma tion, more latewood was produced each year when irrigation was applied. As Larson et al. ( 2001) pointed out, earlywood and latewood are independently related to SG, so annual growth in earlywood and latewood areas determinations (Figure 3-5) can help to unders tand SG patterns: between age 3 (1997) and 8 (2002) and again at age 11 (2005) latewood ring area production of irrigated trees was always bigger than nonirrigated trees and that difference was fairly homogeneous across years (between 3.5 to 5.9 cm2; Figure 3-5e). When irrigation was not a pplied properly (year s 2003 and 2004), latewood production was not different betwee n control and irrigated trees. This dynamic in earlywood and latewood production response to water availability is reflected in the marginal effect of irrigation in whole-core weighted SG. Al baugh et al. (2004) reported no e ffect of irrigation on SG and 69

PAGE 70

LW% of 12 year-old loblolly pine after 5 years of irrigation. Larson et al (2001) indicate that irrigation through out the whole season will promote more earlywood and latewood with negligible final effects on SG, but as soil mois ture is most commonly limiting during late growth season, irrigation would promote latewood forma tion increasing SG (in our site, in average between 1995-2005, 35% of total yearly rain felled betw een January and March). The positive and strong relationship between amount of irrigation applied between July and November and the gain in latewood area was concordant with the results of Cregg et al. (1988), where higher SG and LW% in years with high summer rainfall compared to years with low summer rainfall. Several authors indicate that l oblolly pine latewood initiation can be controlled genetically, as Cregg et al. (1988) who pointed out, for 10 yea r-old loblolly pine, that 70% of the trees that initiated latewood formation early in a dry year also initiated la tewood formation earlier in a wet year, or Jayawackrima et al. (1997), who repor ted for 5-6 year-old loblolly pine trees provenances from Atlantic coastal plain (ACP, where family SC comes from) presented 10 to 20 days earlier transition date from earlywood to la tewood than North Florida provenance trees. The same authors also reported larger average SG and LW% on ACP than Florida source trees. In their report of 1998, the same au thors (Jayawackrima et al., 1998) pointed out that bud break started during the same week in all the families and provenances, and concluded that ACP and FL did not vary in the day of height growth ini tiation. Family differences in SG were related to the fact that the SC family presented larg er latewood area (SG of latewood and earlywood growth were similar between families). This di fference in latewood growth was not related, at least during 2005 season, to differences in gr owing period cessation (both families averaged similar BA growth cessation day). So, differences between both families can be the effect of 70

PAGE 71

higher growth rate of SC during latewood formati on and/or the effect of earlier transition from earlywood to latewood production for the SC family. Conclusion After the fast initial response to irrigation a ssociated with response in LAI development, long term responses are more related to an extended growing season, benefiting both, stem growth and wood properties, b ecause of larger latewood pr oduction. Trees from the South Carolina Coastal Plain family evaluated in this study produced more latewood than trees from the mix of the fast growing families from Florida. Soil moisture limitations constrained latewood growth on loblolly pine plantation studied, reducing both SG and LW%. The two genetic families evaluated showed differences in th e amount on carbon allocated to latewood, maybe reflecting differences in adaptive tr aits related with ambient humid ity and water availability that promote differences in growth period length. All these founding in juvenile wood be evaluated in further years in order to confirm the trends of water availability and family effects on mature wood properties of loblolly pine. 71

PAGE 72

Table 3-1. Mean diameter at br east height (DBH), he ight, basal area (BA) and leaf area index (LAI) for Irrigation x Family treatments. Control Irrigated p > F FL SC FL SC Irrig Fam IrrigxFam DBH (cm) 20.08 20.97 21.31 22.82 0.0003 0.0039 0.577 Height (m) 15.4 16.0 17.2 17.8 0.0002 0.1860 0. 939 BA (m2ha-1) 34.0 32.9 39.7 39.8 0.0030 0.7073 0.654 LAI (m2m-2) 4.2 4.5 4.6 4.4 0.348 0.597 0.083 Means at age 11 (Dec-2005), LA I was measured on Oct-2005. FL: Florida family; SC: South Carolina Coastal Plain family. p-values usi ng mixed model procedure for split-plot design. 72

PAGE 73

Day of year 200220240260280300320340360 Soil water content 0-35 cm depth / Precipitation (mm) 0 20 40 60 80 Control FL Control SC Irrigated FL Irrigated SC Precipitation Figure 3-1. Daily average precipitation and so il water content (mm) in 2005 for the 0-35 cm depth for a loblolly pine stand under an ir rigation treatment (cont rol and irrigated) that included two different genetic families (FL and SC). 73

PAGE 74

Day of year 200220240260280300320340360 Basal Area (m2 ha-1) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 Basal Area Increment (m2 ha-1 month-1) 0.0 0.2 0.4 0.6 0.8 Control Irrigated Figure 3-2. Average monthly basa l area growth (a) and cumulative basal area accretion (b) for irrigated (open circle) and control (filled circle) plots during 2005. Measurements started in early June and ended in later December. Non-linear fitting for cumulative basal area corresponds to the Chapman-Richards model. 74

PAGE 75

Day of year of BA growth cessation 230240250260270280290300 Specific gravity 0.46 0.48 0.50 0.52 0.54 0.56 0.58 Latewood percentage 45 50 55 60 65 Control FL Control SC Irrigated FL Irrigated SC Figure 3-3. Relationship between the cessation of basal area growth and (a) latewood percentage and (b) specific gravity for irrigated (triangl e) and control (circle) plots that included families from FL (filled) and SC (open) in 2005. A linear fit was significant for both cases (p=0.0283 and p=0.048, for latewood percentage and specific gravity, respectively). 75

PAGE 76

Latewood percentage 10 20 30 40 50 FL SC Specific gravity 0.35 0.40 0.45 0.50 0.55 FL SC Latewood percentage 10 20 30 40 50 Control Irrigated Specific gravity 0.35 0.40 0.45 0.50 0.55 Control Irrigated Year 19951996199719981999200020012002200320042005 Rainfall Irrigation (mm) 0 500 1000 1500 2000 2500 3000 Rainfall (mm) Irrigation (mm) Normal year rainfall (mm) Year 19951996199719981999200020012002200320042005 Rainfall Irrigation (mm) 0 500 1000 1500 2000 2500 3000 Rainfall (mm) Irrigation (mm) Normal year rainfall (mm) Specifig gravity of latewood 0.3 0.4 0.5 0.6 0.7 Control Irrigated Specifig gravity of latewood 0.3 0.4 0.5 0.6 0.7 FL SC a c b d ef gh Figure 3-4. Annual latewood specifi c gravity (a, b), latewood per centage (c, d), ring specific gravity (e, f), and total rainfall (open bar) and water added by drip irrigation (dashed bar) during each year from 1995 through 2005 (g, g) for water availability (a, c, e, g) and genetic family (b, d, f, h) treatments. Thirty-year rainfall mean for the site (1408 mm) is indicated with a strai ght line on panels e and f. 76

PAGE 77

0 10 20 30 40 50 Year 19951996199719981999200020012002200320 042005 Latewood ring area (cm2) 0 5 10 15 20 Control Irrigated Ring Area (cm2) Control Irrigated Year 19951996199719981999200020012002200320042005 Latewood ring area (cm2) 0 5 10 15 20 FL SC 0 10 20 30 40 50 FL SC Ring Area (cm2) Earlywood ring area (cm2) 0 10 20 30 Control Irrigated 0 10 20 30 FL SC Earlywood ring area (cm2) Year 19951996199719981999200020012002200320042005 Rainfall Irrigation (mm) 0 500 1000 1500 2000 2500 3000 Rainfall (mm) Irrigation (mm) Normal year rainfall (mm) Year 0 500 1000 1500 2000 2500 3000 19951996199719981999200020012002200320042005 Rainfall Irrigation (mm) Rainfall (mm) Irrigation (mm) Normal year rainfall (mm) a c b d ef gh Figure 3-5. Annual growth in ri ng area (a, b), latewood area (c, d), earlywood area (e, f), and total rainfall (open bar) a nd water added by drip irriga tion (dashed bar) during each year from 1995 through 2005 (g, g) for water availability (a, c, e, g) and genetic family (b, d, f, h) treatments. Thirty-year rainfall mean for the site (1408 mm) is indicated with a straight line on panels e and f. 77

PAGE 78

Irrigation July-November (mm) 0 200 400 600 800 Latewood Area Gain (Irrigated Control) (cm2) 0 1 2 3 4 5 6 7 linear fit; p=0.009; r2=0.65 Figure 3-6. Relationship between mean late wood area difference between irrigated and non irrigated tree and the amount of water a dded through irrigation during July-November for years 1997 to 2005. 78

PAGE 79

CHAPTER 4 WATER USE, WHOLE-TREE HYDRAUL IC CONDUCTANCE AND CANOPY CONDUCTANCE DYNAMICS IN MATURE P. PALUSTRIS AND P. ELLIOTTII TREES Introduction Flatwoods are the most extensive type of terrestrial ecosystem in Florida, covering approximately 50% of its land area (Ewel, 1990). In north-Central Florida, two dominant species are characteristic of pine flatwoods: longleaf pine ( Pinus palustris Mill.) and slash pine ( Pinus elliottii Engelm. var. elliottii). More than 97% of the original land area of longleaf pine has been lost to other uses (Johnson and Gjerstad, 2006) and some of the most species-rich areas and highest concentrations of enda ngered and threatened species in the Southeast are found on mesic flatwood sites (Cohen et al., 2004). The densities and proportions of both speci es depend on geographic location, climate, edaphic conditions, fire history and human influences (Ewel, 1990). Longleaf pine (LL) and slash pine (SL) can overlap on mesic flatwoods sites, LL typically dominates on higher ground (drier) and better-drained sites while SL dominates on lower (wetter) sites where seasonal ponding occurs, SL are primarily located around or even inside the seasonal ponds and LL dominate the higher areas between the ponds (E wel, 1990; Peet, 2006). Ba rnett and Sheffield (2002) indicate that the native habitats of SL ar e poorly drained flatwoods and stream edges, as well as seasonally flooded area s such as bays and swamps. At global scale water availability is one of the most important factors controlling productivity and species distri butions (Kramer and Boyer, 1995; Lambers et al., 1998) and interactions between water availability and wa ter relations within species are important in determining habitat associations (Baraloto et al., 2007) Few studies exist that have examined water relations of mature LL and SL under fiel d conditions (Teskey et al., 1994; Martin, 2000; Vose et al., 2003; Ford et al., 2003, 2004; Addington et al., 2004, 2006;), and even fewer exist in 79

PAGE 80

determining whole-tree sapwood-sp ecific hydraulic conductivity ( KS-wt) (Addington et al., 2004, 2006). For that reason, and due to the ecological importance of this two species, this study was carried out. The objective of this study was to characterize wate r use behavior of both species cohabiting the same site, in term s of total daily tran spiration, water storage use, whole-tree hydraulic conductance and hydraulic adjustments in crown conductance and leaf to sapwood area to varying vapor pressure deficit ( D ) and volumetric water content ( v) conditions. We hypothesize that water relations tr aits of the two species are important in determining the microsite habitat "preference" of each species. Materials and Methods Site and Stand Description The study was carried out at the University of Floridas, the Austin Cary Memorial Forest (ACMF), located 15 km northeast of Gainesvi lle, FL (29 N latitude and 82 W longitude). Soils are classified as poorly-drained Pomona sands (sandy, siliceous, hyperthermic Ultic Aplaquods) with a discontinuous spodic hori zon at 30-60 cm depth and deeper argillic horizon at 100-140 cm depth (Gaston et al., 1990) The study stand cons ists of a naturally regenerated mixed LL and SL sta nd with tree ages ranging between 25 and 85 years, with a mean age of 65 years. Within the stand, SL tended to be clumped in the lower lying areas and along pond margins, all within a matrix of LL. Sta nd basal area at the time of measurement was 16.9 m2ha-1, distributed as 73% and 17% LL and SL respectively. The understory consisted on native species, dominated by gallberry ( Ilex glabra (L.) Gray), saw palmetto ( Serenoa repens (Bartr.) Small), wax myrtle ( Myrica cerifera L.) and wiregrass ( Aristida stricta Michx) (Powell et al., 2005). 80

PAGE 81

Meteorological Measurements Environmental conditions were monitored from January 2007 to August 2007 using an automatic weather station located on the top of a 30 m scaffolding tower, extended approximately 5 m above the mean canopy height (Powell et al., 2005). Relative humidity and temperature were measured using a relative humidity and temperature probe (HMP45C-L, Vaisala, Inc., Helsinki, Finland), photosynthetic photon flux density (PPFD, molm-2s-1) was measured with a quantum sensor (Li-190, Li-Cor Inc., Lincoln, NE, USA), wind direction and velocity were measured using an anemometer and vane (03001-L, Campbell Scientific, Logan, UT, USA) and precipitation was recorded with a tipping bucket rain gage (TR525-I, Texas Electronics, Dallas, TX, USA). All sensor data were measured on 30 s intervals using an automatic datalogger (CR10X, Campbell Scientific Logan, UT, USA) and were stored as 30 min averages. Water table depth was measured usi ng a Stevens water depth gauge (F-68, Leupold and Stevens, Inc., Beaverton, OR, USA) located approximately at the midpoint between the LL and SL sectors. Soil Moisture Vertically oriented, 20 cm long capacitive soil moisture probes (EC-20, Decagon Devices, Pullman, WA, USA) were used to measure volumetric soil water content ( v, cm3cm-3). In April 2007, probes were installed at 50 and 150 cm depths into two sectors representing the SL and LL pine micro sites. Estimations of wa ter content for each probe were extrapolated to the mid-point distance between sensors. Specific calibrations were developed from soil samples taken 50 cm away from each probe; soil samples were collected from the same depth where the probes were installed. Soil-speci fic calibrations had no differen ce with manufacturers default calibration model to transform sensors out puts (mV) to volumetric water content: v = -0.24508 + 0.0007958mV. There was no difference between samples taken at 40-60 and 140-160 cm 81

PAGE 82

depth, and the manufacturers calibration (p=0.6 5 and 0.82 for intercept and slope, respectively). Soil matric potential ( s, MPa), was estimated for the 50 cm depth measurements using previous soil water retention curves generated for local Spodosols (H.L. Gholz, unpublished data) and for the 150 cm depth measurements using a computer program based on pedotransfer functions and parameterized with soil texture and bulk densit y data obtained at each location (Schaap et al., 2001; ROSETTA version 1.2, US Salinity Labor atory ARS-UDA, Riverside, CA, USA). Available soil water (ASW, %) was calculated by an alyzing the limits of wetting and drying of the soil through the entire study period, drained upper limit (DUL, mm) and lower limit of water extraction (LL, mm) where determined for each plot and depth; ASW wa s calculated using the following formula (Ritchie, 1981): ASW = 1 LLDUL WCDUL where WC is the water content (mm) at any given measurement day and depth. Tree Selection The measurement trees were chosen from across the range of tree diameters using "quantiles of total", a stratific ation scheme which weights the selection of large trees more heavily (Hatton et al., 1995; Martin et al. 1997, ermk et al., 2004). Four trees per species were selected for leaf water potential, sapflow, a nd leaf area measurements. Average diameter of selected trees averaged 32.6 and 34.8 cm, ranging between 19.8 to 41.8 cm and from 20.4 to 49.3 cm, for LL and SL, respectively (see APPEND IX C). Average sapwood thickness at a 1.8 m height (Sbase) was 5.2 and 6.9 cm and ranged be tween 3.9 to 6.6 cm and from 4.2 to 9.1 cm, for LL and SL, respectively. 82

PAGE 83

Sap Flow Measurements To determine radial patterns in sap flux density ( JS, g waterm sapwoods) (Hatton et al., 1990; Lu, 2000; Ford et al., 2004, 2007), variable length sap flow probes (VLP), as proposed by James et al. (2002), were installed at Sbase. Probes were inserted at 1 cm depth intervals, with the tip of the 10 mm probes reaching 1, 2, 3, 4, 6, 8 and 10 cm depths (assuming to measure sap flux density in discrete depths, e.g., 0-1, 1-2, 9-10 cm). The probe for the outermost position was installed on the north side of the stem, a nd subsequent probes were installed clockwise around the stem at 45 intervals. Weighted average JS at Sbase ( Js-b ) was calculated at each time step as the sum of the product of JS and annulus area corresponding at each depth, divided by total sapwood area, in or der to determine a unique and integrated value of JS for each tree and time step. Twenty mm long Granier type heat dissi pation probes (Granier, 1985, 1987) were installed approximately 20 cm below the base of the crown (Scrown) of each measurement tree (between 15.3 to 19.35 m). Weighted average JS at Scrown ( Js-c ) was calculated using sap flux density values measured at crown base at 0-20 mm sapwood depth corrected by the radial profile in native sapwood-specifi c hydraulic conductivity ( ks-nat, Kgs-1m-2 sapwoodMPa-1m) at Scrown determined for each tree (see Chap ter 5). As radial profile in ks-nat showed no significant differences between 1 to 4 cm depth (p>0.14, data not shown) and sapwood area of the outer 20 mm corresponds, in average, to 50.3 and 46.7% of total sapwood for LL and SL, respectively, we are confident that JS estimation at crown bas ( Js-c ) estimations are unbiased. Sapflow probes installation at Scrown was carri ed out using a self propelled telescopic lifting machine. Js-c estimations were carried out with the objective to minimize water storage effects on time lags between transpiration and sapflow measurements at lower portions of the stem, as has been reported for large trees when sap flow was measured at breast height (Phillips 83

PAGE 84

et al., 2003). This method has been used widely and described elsewhere (Braun and Schmid, 1999; Lu et al., 2004). Briefly, the system consists of two probes in serted radially into the stem, one above another about 10-15 cm apart. The up per probe contains a heater and a T-type thermocouple and the lower probe contains only the thermocouple. The upper probe was heated at constant power while the lower one was used as reference measuring the ambient temperature of the wood. The temperature difference be tween the heated and reference probes ( T) was recorded, and by comparing the difference to the maximum occurring at predawn ( Tm) when there was assumed no flow (G ranier 1985, 1987; Ford et al., 2004). Sap flux density was determined using the empirical calibration de veloped by Granier (1985) and confirmed by Braun and Schmid (1999) and Clearwater et al. (1999): JS = 119 t1.231 where t = ( Tm T) / T. Sap flux density was converted to sapflow ( F gs-1) after multiplying JS at Sbase and Scrown by the corresponding sapwood area of each measurement point. With a chisel, bark and cambium were removed at the probe installation point to insert the sensors entirely into the xylem. Probes were coated with thermally-conductive silicone grease before placement in the trees. All the sensors were prot ected against radiation, thermal gradients and precipitation by reflective insulation. Daily stored water use (S WU) was computed following Goldstein et al. (1998), as the difference between sa p flow at crown base and sap flow at tree base when basal sapflow was less than crown flow (water storage withdrawal). The percent of stored water used for daily transpiration (Phillips et al., 2003) wa s calculated as the propo rtion of daily water stored to total sap flow at crown base. Daily wa ter use statistical analys is was performed using 84

PAGE 85

repeated measures analysis including Bonferroni adjustments for differences in least square means (PROC MIXED, SAS Inc., Cary, NC, USA). Leaf and Sapwood Area At the end of the study all four sap flow measurement trees from each species were felled to determine total tree leaf area (LA, m2), using a destructive determin ation of leaf mass (LW, g) in conjunction with tree-specific estimations of all-sided specific leaf area (SLA, cm2g-1) (see APPENDIX C). Trees were felle d at a stump height of approximately 20 cm and all branches were cut at the point where live needles started. All branches tips were weighed green in the field and summed to determine whole-tree branch tip fresh mass. For each tree, 20 randomly selected subsamples of branch tips were weighed, ba gged and brought back to the laboratory for determining moisture contents and dry masse s of wood and needles of each branch tip separately. Dry mass of all livi ng needles of the whole tree cr own (LW) was calculated as the product of field-determined whole-tree branch tip fresh mass and the average dry needle to fresh branch tip ratios derived from all 20 branch ti p subsamples. Needles were weighed separately into two age classes: current year and older needles, so dry needle to fresh branch tip ratios were also calculated separately for each needle type. SLA was determined using the ratio between surface area and dry weight of needles for each needle age class. Individual needle surf ace area was calculated according to Murthy and Dougherty (1997) and Niinemets et al. (2001) from needle radius and length measured with a 10 x scaled magnifier and a digital caliper (CD6, Mitutoyo, Kawasaki, Japan), respectively, on 10 needles per tree. After surface area determina tion, the needles were oven-dried for 48h at 75C and weighed to the nearest 0.0001 g (XA-100, De nver Instruments, Denver, CO, USA). A weighted average SLA (Table 4-1) was determin ed for each tree after including the proportion of leaf mass of each needle age. To determine daily LA for each tree, seasonal patterns were 85

PAGE 86

developed using information collected previous ly from periodic measurements of LAI from January-2000 to January-2006 (Powell et al. 2008). LA was estimated backward in time from LA measurement day (between June-21 to July-9) until sapflow installa tion day (April 17). In all measured trees, at the same point where sapflow probes were installed, sapwood area (SA, m2) was estimated at the end of the stu dy by measuring sapwood depth and stem diameter directly in 20 cm-thick disks extracted after the trees were felled for foliar biomass and leaf area determination at the point were sapflo w probes were installed (at 1.8 m and crown base heights). The disks were transported back to th e laboratory covered with wet towels in plastic bags and stored at 5C for further hydraulic cond uctivity measurements (described in Chapter 5). Above ground sapwood volume below crown base was calculated for each tree by using sapwood area measured at stump and crown base and distance from stump to crown base. Stem diameter and bark thickness were measured eac h 2 m to determine stem volume. Huber value (HV, or sapwood to leaf area ratio, m2m-2) was calculated for each tree as SA at Sbase or Scrown, divided by total LA (m2). Crown Conductance Stomatal conductance for water vapor at the individual tree crown scale ( Gcrown, ms-1) was calculated as in Granier and Loustau (1994): Gcrown = )( sGDcRs EGa p la where Ga is the aerodynamic conductance (ms-1), El is tree transpirati on per unit leaf area (kg waterm-2 leaf areas-1), is the latent heat of water vaporization (Jkg-1), is the psychometric constant (PaK-1), s is the rate of change of saturating vapor pressure with temperature (PaK-1), R is the radiation absorbed by the canopy (Wm-2), is the density of dry air (kgm-3), cp is the specific 86

PAGE 87

heat of air (JK-1kg-1) and D is the vapor pressure defic it (Pa). Aerodynamic conductance was calculated from the wind profile equa tion (Monteith and Unsworth, 2007): Ga = 2 0 2))/)(ln(( zdz wk where k is Von Karmans constant (0.41), w is wind speed (ms-1), z is anemometer height (m), d is roughness length (m) and z0 is displacement height (m). Values of d and z0 were set as 2/3 and 1/10 of canopy height, respectively (Phillips and Oren, 1998). Afte r correcting for temperature changes in water density, Gcrown was transformed to molar units (mmolwaterm-2 leaf areas-1). To reduce error due to instru ment limitations on relative humidity measurements, Gcrown was calculated only when D 0.6 kPa (Ewers and Oren, 2000). A reference Gcrown ( GCref) was calculated at D =1 kPa (Granier et al, 1996; Oren and Pataki, 2001). The response of Gcrown to D was quantified using boundary line analysis (Ewe rs et al., 2001; Schfe r et al., 2000). The upper boundary line for each plot was derived by binning Gcrown data into 0.2 kPa D intervals (from 0.6 kPa to 4.6 kPa) and then selecting the highest 95% Gcrown for any interval. For each plot, all upper Gcrown values in each D interval were related to the natural logarithm of D (Granier et al, 1996): Gcrown = GCref mlnD where m is the slope of the regression f it, representing stomatal sensitivity to D (i.e. d Gcrown /dln D ). Whole-Tree Hydraulic Conductance Whole-tree sapwood-specifi c hydraulic conductance ( KS-wt, molwaterm-2 sapwoods-1MPa-1) was computed following the regression technique (Wullschleger et al., 19 98) and calculated as (Phillips et al., 2002; Franks, 2004): 87

PAGE 88

KS-wt = sl SE where ES is transpiration rate per unit sapwood area (molm-2s-1), l and s are leaf and soil water potential (MPa), respectively. l was measured using a portable pressure chamber (PMS 1000, MPS Instrument Co., Corvallis, OR, USA) on one shoot tip from pre-dawn (5:00-5:30) to late afternoon at intervals of 2 hours approximately, completing 4 to 5 measurements per tree along the day. Each measurement was completed w ithin 3 minutes after sh oot excision, covering the sample with wet towels inside a plastic bag and maintained inside an insulated box to minimize desiccation. Shoot sampling was carried out using a self-propell ed telescopic lifting machine and pole pruner. For each tree, s was determined by using pre-dawn leaf water potential (pred) as a surrogate. Maximum leaf water potential gradient ( MPa was calculated as the difference between pred and midday leaf water potential ( midday, the minimum l measured close to midday). Statistical Analysis Analysis of variance (ANOVA) was used to an alyze effects of organ and species in water relation traits, including Bonfe rroni adjustments for differences in least square means (PROC MIXED, SAS Inc., Cary, NC, USA). The ANOVA model for the analysis is described in APPENDIX B. Repeated measures analysis was used to analyze time series data. Results Environmental Conditions During the sapflow measurement period (from Ap ril 17th to July 9th), total precipitation was 152.9 mm, corresponding to 45.7 % of historic av erages for the site (during a "normal" year, 334.2 mm of rainfall would o ccur during that period) (http://cdo.ncdc.noaa.gov/cgibin/climatenormals/climatenormals.pl August 2008). Mean daily te mperature (Tmean) increased 88

PAGE 89

gradually from values close to 17C during the thir d week of April to 29C during the first week of August (Figure 4-1b). Total da ily PAR ranged from about 20 molm-2d-1 to 37.8 molm-2d-1 when the study was completed. Daily average D followed a similar pattern than PAR, but peaking approximately one month earlier. Both PAR and D increased in variability during the summer in comparison to winter daily values (F igure 4-1a). Water tabl e depth ranged from 180 cm at the beginning of the growing season to 250 cm in mid-July (Figure 4-1b). As a difference in altitude of approximately 100 cm was measured between the midpoint of the LL and SL trees, and assuming constant slope between both sectors, water table depth fluctuated between 150 to 200 in SL and 250 to 300 cm in LL dur ing the sapflow measurement period. Sap Flux Density, Transpiration, Water Storage Use and Soil Moisture At 1.8 m height (Sbase), the relationship between Js and distance from cambium had no distinctive shape for any species, changi ng patterns during the season, depending on environmental conditions and tree size (age) (Figur e 4-2). In general the out er probes (installed from 1 to 4 cm depth) had higher values and foll owed the daily PAR cycle, compared with inner probes (installed at 6 and 8 cm depth) which tended to peak in late afte rnoon, even in dark when Js for outer probes decreased to near -zero values; this can be an i ndicator of use of water stored in deeper zones of the xylem at the stem base. After comparing F at 1.8 m height ( Fb ) and at the crown base ( Fc ) with daily courses of D and PAR, it is possible to observe that Fc fits much better with D (Figure 4-3), and the use of sap flux density measurements at the tree base ( Fb ) can result in large errors when the objective is to match Js with other simultaneous meteorological or physiological measurements such as crown conductance or whole-tree hydrau lic conductivity. Because of that, Fc was used for further analysis of tree water use, crown c onductance and whole-tree hy draulic conductivity. 89

PAGE 90

For the period analyzed from April 28th to June 18th, 2007, overall mean daily transpiration rates (Eday, lday-1tree-1) for LL and SL were 33.2 and 50.0 lday-1tree-1, respectively, these species level estimates were marginally not statisti cally different (p=0.113; Figure 4-4a). There was a str ong (p=0.0045) interaction between species and time (day). From April 28th to May 30th (i.e. day 118 to 151), Eday of SL was significantly greater than Eday of LL (p<0.05); this period corresponds to the time when s of LL site reached values below -1.0 MPa from 50 to 150 cm soil depth (Figure 4-4c). After a 45 mm rainfall on June 6th, s at the 50 cm depth reached values close to -0.2 MPa and tr anspiration increased fo r both species, and Eday for the two species was not statistically different (p>0.14) for the rest of the measurement period. When Eday was expressed per unit leaf area ( El, lday-1m-2) there were no significant differences in mean El between LL and SL trees across whole pe riod (p=0.243; Figure 44b), but there was a strong (p<0.0001) interacti on between species and time. Between days 140 and 153, El of SL was larger than LL (p<0.044), and this period corresponded with large differences in soil moisture between SL and LL sites (Figure 4-4c). Before and after that time period, El of SL and LL were not statistically different (p>0.14). It is interesting to note that when the soil was drying (during the two weeks before the June 6th rainfall), and s declined far below -1.0 MPa, El for LL was lower than SL (p=0.0002), averaging values of 0.296 and 0.366 lday-1m-2 for LL and SL, respectively. After rainfall this trend changed, with El for LL becoming higher than SL (p=0.0002), with average valu es of 0.501 and 0.414 lday-1m-2 for LL and SL, respectively. During the measurement period, daily stored water use (WSU) was not different between species (p=0.5512), averaging 8.35 and 10.24 lday-1tree-1, for LL and SL, respectively (Figure 4-6a). WSU was highly correlated with tree size; daily average WSU (lday-1) had a strong linear relationship (p=0.001; r2=0.854) with above ground sapwood volume (Figure 4-6b). After 90

PAGE 91

expressing WSU in m3 it is po ssible to extrapolate from th e relationship between WSU and sapwood volume that, on average, 1.06% of the sapwood volume is used for water storage. The percent of stored water used for daily transpira tion averaged 39.0 and 23.2%, for LL and SL, respectively, but the difference between spec ies varied depending on time (p<0.0001 for time x species interaction ). During a dry soil period (e.g. day 139 to 151; Figure 4-6c), mean relative WSU was 60.2 and 26.5% for LL and SL, respectiv ely (p=0.01). During wet soil periods (e.g. day 153 to 168; Figure 4-4c), mean relative WSU decreased to 21.7 and 16.9% for LL and SL, respectively (p=0.63). Within sp ecies relative WSU was statistically different for LL (p<0.0001) but not for SL (p=0.245) across soil moisture peri ods. These two subsets of time were selected based on soil moisture differences between them (p<0.0001) but similar among speciesmicrosites (p=0.12). Across species, v and s for dry and wet periods at 50 cm soil depth reached values of 8.6 and 16.9%, and -1.39 and -0.1 MPa, respectively; daily average D and total daily PAR were similar between select ed periods (p=0.266 and 0.112, respectively). Crown Conductance Average daily crown conduc tance to water vapor ( GCday, mmolm-2s-1) ranged between 11.4 to 58.8 and 12.8 to 42.5 mmolm-2s-1, for LL and SL, respectively. Overall mean GCday for LL and SL was 28.89 and 23.89 mmolm-2s-1, respectively, but these values were not statistically different (p=0.4732). There was a strong (p<0.00 01) interaction between species and time; during the dry period GCday was not different for LL and SL, but during periods of high soil moisture (after rain) LL had larger GCday than SL (p<0.05). GC response to vapor pressure deficit was similar for both species (Table 4-1). Average Gcrown sensitivity to D (-d Gcrown /dln D ) and Gcrown at D =1 kPa (GCref) were not different between both species (p=0.127 and 0.203, respectively). There were no significa nt differences between species in Dmax ( D when Gcrown equals zero) neither in the relationship between d Gcrown /dln D and GCref (p=0.181 and 0342, 91

PAGE 92

respectively). The slope of this rela tionship after pooling data was 0.68 and GCref explained 96% of variation in canopy conducta nce sensitivity to D. The relationship between Gcrown and v was different between LL and SL. (Figure 4-5). Under well-watered conditions (e.g. v >0.12 cm3cm-3), GCday of LL was larger than SL (p=0.001), averaging 35.8 and 25.5 mmolm-2s-1, respectively; on the other hand, when ASW was lower than 35% (e.g. v >0.11 cm3cm-3), there was no difference in average GCday between LL and SL (p=0.45), averaging 25.8 and 23.6 mmolm-2s-1, respectively. A similar decline in GCday was measured for LL and SL below this v threshold value (c orresponding to 28% available soil water, approximate ly). Lower minimum values of v at 50 cm depth were observed in LL compared with SL; this can be related to lower extraction rate at that depth on SL due to higher water availability at deeper profiles in SL due to the proximity of water table (Figure 44c). Whole-Tree Hydraulic Conductance There were no differences in average KS-wt, midday and between species (Table 4-1). Average KS-wt, midday and for LL and SL were 3.37 and 4.20 molm-2s-1MPa-1, -1.83 and 1.67 MPa and 0.99 and 0.95 MPa, respectively. Average pred was statistically different between species (p=0.012), averaging values of -0.84 and -0.72 MPa, for LL and SL, respectively. Discussion Daily transpiration rates were comparable to similar size LL and SL evaluated in other studies. Vose et al. (2003) reported mean sap flow between 30 to 80 ltree-1day-1 for LL and SL trees ranging from 35 to 45 cm DBH. In a 14 year-old SL plantation, Martin (2000) determined average and maximum winter daily tr anspiration rate of 28.3 and 71.6 ltree-1day-1, respectively, for a 24 cm DBH trees. On a natural regenerate d longleaf pine savanna, Ford et al. (2004) measured sapflow on LL and SL, reporting av erage daily water use from 36.8 to 218.3 lday-1, 92

PAGE 93

for trees ranging between 19 and 118 cm DBH for LL trees; and 99.3 lday-1, for a single SL tree of 70 cm DBH. Differences in average daily transpiration rate between species, although non significant (p=0.11) reflected, on average, 50% higher wate r use by SL than LL (Figure 4-4a); that difference can be explained, at least in part, by th e fact that SL had more leaf area than LL for trees of the same sapwood area; Huber value (HV, m2 sapwoodm-2 leaf area) of LL was 60% larger than SL (p=0.086; Table 4-1). Under well-watered conditions the transpira tion rate per unit leaf area (El) of LL was higher than SL, but as the soil dried in the LL zone ( s < -1 MPa), this trend was reversed, with El of SL higher than LL. This response ca n be related to species differences in Gcrown sensitivity to v (Figure 4-5) and D (Table 4-1). At soil mo isture conditions above 35% ASW, average GCday of LL was 40% greater than SL, and even though GCref and sensitivity of Gcrown to changes in D (-dGcrown /dln D ) were not different between species (p=0.20 and 0.127, respectively), the overall mean of LL was 37.4% and 54.8% larger than SL, respectively. All these differences (some not significant) in water relations traits per unit leaf area are largely compensated for by the larger leaf area per unit sapwood area of SL. The relationship between GCref and d Gcrown /dln D was not different between species (p=0.58), following a strong linear relation (p=0.0001, r2=0.982) with a slope of 0.68, concordant with that proposed by Oren et al. (1999). This rela tionship implies that trees with overall high stomatal (crown) conductan ce will tend to be more sensitive to D showing higher stomatal closure as D increases. Our estimates of water storage use are base d on the lag between sapflow at the crown base and sapflow at 1.8 m height; this rationale is in agreement with ermk et al. (2007), who concluded that the xylem sapwood below the most active portion of th e crown is the most 93

PAGE 94

important source of free stored water, accounting for 91.1% of the total free water volume. These authors carried out that study in old-growth Dougl as-fir, with a large liv ing crown of 31 m. In contrast, the pine trees in our study had crown lengths that av eraged 8.9 and 8.4 m, for LL and SL, respectively. Stem volume without bark a bove the Scrown probes averaged only 10.4 and 11.7% of the total stem volume, for LL and SL, respectively, ther efore it is highly unlikely that the sapwood above the Scrown probes had a significant impact on total daily water storage compared with the sapwood volume below the live crown. The use of water stored in stem was not different between the species, and averag ed across time and species, 1.06% of sapwood volume; so the larger the tree, the larger the stored water use (Figure 4-6b). Meinzer et al. (2004, 2006), working with different species of angiosperms and gymnosperms, concluded that storage characte ristics were species-inde pendent and linearly related to tree size. In our study, average WSU was 8.35 and 10.25 ltree-1day-1. Phillips et al. (2003) reported mean WSU of 6.5 and 16.7 ltree-1day-1 for 25 and 36 m tall Quercus garryana and P. ponderosa, respectively, averaging up to 20% of da ily sapflow. In angiosperm tropical tress, Goldstein et al. (1998) reported aver age daily water storage between 4.0 and 8.7 ltree1day-1 for trees between 16 to 27 cm DBH, pointing out a common relationship between sapwood area or tree height and diurnal storage capacity independent of taxa. ermk et al. (2007) also found that about 20% of daily sapf low came from stored water in old-growth Douglas-fir trees, rangin g from 25 to 55 ltree-1day-1. The same authors also indicated that that deeper zone in the sapwood can serve as a s ource of stored water, concurring with our observations of time lags with depth in the radial profile in sap flux dens ity (Figure 4-2). This high reliance of mature LL and SL on short-term stem water stor age is a confirmation of the 94

PAGE 95

important role of sapwood water storage in ma intaining high transpiration rates and limiting diurnal fluctuations in leaf wate r potential (Meinzer et al., 2004). In order to compare our estimates of KS-wt with the results obtained by Addington et al. (2004, 2006) on mature longleaf pine, KS-wt was expressed in leaf area and an hour basis (KL-wt, kgm-2 leaf areah-1 MPa-1). Our average estimate of KL-wt, for LL and SL were 0.088 and 0.071 kgm-2h-1MPa-1, respectively, when v at 50 cm depth was 9.5% (d ata not shown). These values were similar to the observations of Addington et al. (maximum KL-wt, close to 0.075 kgm-2h1MPa-1), but v was close to 7% at the time of thei r measurement. Longleaf pine and SL measured in this study both have higher KS-wt than values reported for loblolly pine ( Pinus taeda ). Samuelson and Stokes (2006) and Gonzlez and Martin (unpublished data) reported, for non soil water-limited conditions, an average KS-wt of 3.0 and 2.9 molm-2s-1MPa-1, respectively. Maximum root-to-leaf water potential gradient ( ) was not different between species, and concordant with values reported by Teskey et al. (1994) for 23-year -old SL; visual estimations of the difference between their estimations of midday and pred, indicate that average was 0.94 MPa across the whole season, showing a strong si milarity between both species in stomatal control, resulting in relatively constant allowing l to fluctuate in synchrony with s. Conclusion We found remarkable similarities between LL and SL in use of stored water and KS-wt, but differences in other important traits including leaf to sapwood area ratio, stomatal response to v and, to a lesser extent, stomatal response to D Differences in leaf area to sapwood area ratio, where SL produces much more leaf area per unit basal sapwood than LL, are partially compensated for by the large plasticity of Gcrown response to v and to some extent to D in LL. The tendency of SL to occupy wetter zones, close or in the fringe of small ponds may be related also to the development of aerenchyma tissues (Fisher and Stone, 1990), which enables taproot 95

PAGE 96

and sinkers to be submerged for weeks or even months due to periodic rais ing of the water table, and perhaps to differences in root vulnerability to cavitation, ra ther than stomatal regulation differences between species. More research is needed in order to be tter understand microsite habitat preferences of each species, as deta iled characteriza tion of hydraulic architecture, tracheid anatomy, vulnerability to cavitation and the re lationship of these traits with more integrated traits such as crown conductance and whole-tree hydraulic conductance of each species. 96

PAGE 97

Table 4-1. Least-square m eans of measured traits for longleaf and slash pines. Trait Unit Species p > F Longleaf Slash species DBH (cm) 32.35 34.18 0.807 Height (m) 23.93 27.28 0.137 Volume (m3) 0.91 1.21 0.545 STbase (cm) 5.21 6.93 0.193 STcrown (cm) 5.61 5.86 0.822 SA-base (m2) 0.040 0.052 0.438 SA-crown (m2) 0.020 0.027 0.474 LA (m2) 61.22 104.84 0.127 SLA (cm2g-1) 96.29 98.06 0.812 HVbase (m2m-2) x10-4 8.25 5.15 0.086 HVcrown (m2m-2) x10-4 3.85 2.67 0.189 GCref (mmolm-2s-1) 93.43 67.99 0.203 -dGc/dlnD (mmolm-2s-1ln(kPa)-1) 58.16 37.56 0.127 Dmax (MPa) 4.99 6.75 0.181 midday (MPa) -1.83 -1.67 0.111 pred (MPa) -0.84 -0.72 0.015 (MPa) 0.99 0.95 0.602 KS-wt (molm-2s-1MPa-1) 3.37 4.20 0.254 Diameter at breast height (DBH), total heig ht, stem volume without bark (Volume), sapwood thickness at 1.8 m height (STbase), sapwood thic kness at crown base (STcrown), sapwood area at 1.8 m height (SAbase), sapwood area at crown base (SAcrown), leaf area (LA), all-sided specific leaf area (SLA), Huber value at 1.8 m height (HVbase), Huber value at crown base (HVbase), crown conductance at D =1 kPa (GCref), crown conductance sensitivity to changes in D (-d Gcrown/dln D ), maximum D at which Gcrown=0 ( Dmax), midday leaf water potential ( midday), predawn water potential ( pred), maximum daily water potential gradient ( ) and whole-tree sapwood-specific hydraulic conductance ( KS-wt) for longleaf and slash pi nes. p-values using mixed model procedure (n=4). 97

PAGE 98

D (kPa) 0.0 0.5 1.0 1.5 2.0 2.5 3.0 PAR (mol m-2 day-1) 0 5 10 15 20 25 30 35 40 D PAR Time (day) 0 50 100 150 200 Temperature (C) 0 5 10 15 20 25 Precipitation (mm day-1) 0 20 40 60 80 100 Water table depth (cm) 0 100 150 200 250 Tm pp Water table depth a b Figure 4-1. Daily average of vapor pressure deficit (D ) during daylight hours and daily sum of photosynthetically active radiation (PAR) (a ) and daily mean temperature (Tmean), precipitation and depth to water table at the midpoint between LL and SL over the study period (b). 98

PAGE 99

Longleaf Js (g m-2 s-1) 0 10 20 30 40 50 1 cm 2 cm 3 cm 4 cm 6 cm 8 cm time (h) 8001000120014001600180020002200 Slash Js (g m-2 s-1) 0 10 20 30 40 50 time (h) 8001000120014001600180020002200 a b cd Figure 4-2. Diurnal sap flux density ( JS) patterns at different radial positions for two longleaf (a, b) and slash (c, d) pine trees on one selected day (May 3, 2007). Trees shown correspond to the smallest (a, c) and larges t (b, d) sampled trees of each species. 99

PAGE 100

D (kPa) 0.0 0.5 1.0 1.5 2.0 2.5 PAR ( mol m-2 s-1) 0 200 400 600 800 1000 D PAR Sap flow (g s-1) 0.0 0.5 1.0 1.5 2.0 Js at 1.8 m Js at crown base Time (h) 0 500 1000 1500 2000 Sap flow (g s-1) 0.0 0.5 1.0 1.5 2.0 Js at 1.8 m Js at crown base c b a Figure 4-3. Example of diurnal courses of PAR and D (a), sap flow (gs-1) measured at 1.8 m height (filled circle) and at the crown base (open circle) of (b) the biggest longleaf pine (DBH=41.8 cm) and (c) the smallest slash pine (20.4 cm DBH), during leaf water potential measurement day (April 17, 2007). 100

PAGE 101

Day of the year 100 120 140 160 180 200Soil 0 10 20 30 40 50 60 70 Figure 4-4. Average daily transpiration rate per tree (a), transpirati on per unit leaf area (b), water storage use (c) and soil matric potential at 50 and 150 cm depth, rainfall and water table depth (d) for longleaf (filled ci rcle) and slash (open circle) pine. Matr ic Pa) Pptentia l (M -2.0 -1.5 -1.0 -0.5 0.0 Rainfal l (mm) 0 20 40 60 Water table depth (cm) 200 220 240 260 LL 50 cm Sl 50 cm LL 150 cm SL 150 cm Rainfall Water table depth c b a LL SL Transpiration (l day-1) 0.0 0.2 0.4 0.6 0.8 1.0 Transpiration per unit leaf area (l day-1m) -2 LL SL 101

PAGE 102

v (m3 m-3) 0.060.080.100.120.140.160.18 Gcday (mmol m-2 s-1) 0 10 20 30 40 50 60 70 LL SL LL regression SL regression Figure 4-5. Average daily crown conductance and volumetric soil wa ter content at 50 cm depth for longleaf (filled circle) a nd slash (open circle) pine, including Gompertz function fitted for both species. Error bars represents SE (n=4). 102

PAGE 103

Sapwood Volume (m3) 0.00.20.40.60.81.01.2 Average daily water storage use (liter day-1) 0 5 10 15 20 LL SL Time (day) 100 120 140 160 Water Storage Use (liter day-1) 0 5 10 15 20 a b Figure 4-6. Average daily water storage use (a) and relationship be tween sapwood volume and average water storage (r2=0.85) use during the measurem ent period (b) for longleaf (filled circle) and slash (open circle) pine. 103

PAGE 104

CHAPTER 5 HYDRAULIC ARCHITECTURE AND TRAC HEID ALLOMETRY IN MATURE P. PALUSTRIS AND P. ELLIOTTII TREES Introduction Flatwoods are the most extensive type of terrestrial ecosystem in Florida, covering approximately 50% of its land area (Ewel, 1990). In north-central Florida, two dominant tree species are characteristic of pi ne flatwoods: longleaf pine ( Pinus palustris Mill.) and slash pine ( Pinus elliottii Engelm. var. elliottii ). Longleaf pine (LL) and slas h pine (SL) overlap on mesic flatwoods that occupy sandy, seldom inundated flatlands. LL typically dominates on higher ground (drier) and better-drained sites and SL domin ates on lower (wetter) sites (Ewel, 1990); so on flatwood sites with seasonal ponding with a saw grass fringe, SL are primarily located surrounding or even inside the seasonal ponds and LL dominate the higher areas between ponds (Peet, 2006). Barnett and Sheffield (2002) indicate that characteri stics SL native habitats are poorly drained flatwoods and stream edges, as we ll as seasonally flooded areas such as bays and swamps According to Tyree and Ewers (1991), to look at trees as a whole functional organism it is necessary to study their hydraulic architecture. Trees have evolved as large and hydraulically complex organisms and that within-tree variati on in hydraulic architecture ultimately controls whole-tree water relations (T yree and Zimmermann, 2002). Duri ng severe drought stomatal conductance declines and the water potential of th e tree will tend to follow that of soil. Under water-stress, plants also reduce their water supply to the leaves when their xylem conduits are cavitated, reducing plant gas exchange (Hacke et al., 2000b), so vulnerability to cavitation puts an evolutionary limit on the water potential at which stomata close (Tyree and Ewers, 1991). Xylem resistance to cavitation is an important tr ait that influences dr ought resistance (Maherali et al., 2006), and a useful tool for assessing resi stance to cavitation is the vulnerability curve 104

PAGE 105

(Sperry and Tyree, 1988). The mechanisms contro lling within-tree variation in hydraulics are not clear and this within tree vari ation in hydraulic archite cture traits as vulne rability to cavitation and sapwood-specific hydraulic co nductivity and their linkage with xylem anatomy tracheid are questions that remain unsolved. Based on the fact that SL tends to inhabit wetter zones and LL tends to dominate in higher and well-drained soils we hypothesize that inter-specific difference in vulnerability to cavitation can be associat ed with this land distribution pattern, with SL being more vulnerable to cavitation than LL. Our second hypothesis states th at within tree variati on in sapwood-specific hydraulic conductivity and vulnerability to cavitation are associated with changes in tracheid anatomy Materials and Methods Site and Stand Description The study was carried out at the University of Floridas the Au stin Cary Memorial Forest (ACMF), located 15 km northeast of Gainesvi lle, FL (29 N latitude and 82 W longitude). Soils are classified as poorly-drained Pomona sands (sandy, siliceous, hyperthermic Ultic Aplaquods), with a discontinuous spodic horizon at 30-60 cm depth and deeper argillic horizon at 100-140 cm depth (Gaston et al., 1990) The study stand cons ists of a naturally regenerated mixed LL and SL sta nd with tree ages ranging between 25 to 85 years, with a mean age of 65 years. Within the stand, SL tended to be clumped in the lower lying areas and along pond margins, all within a matrix of LL. Sta nd basal area at the time of measurement was 16.9 m2ha-1, distributed as 73% and 17% LL and SL, respectively. The understory consisted of native species, dominated by gallberry ( Ilex glabra (L.) Gray), saw palmetto ( Serenoa repens (Bartr.) Small), wax myrtle ( Myrica cerifera L.) and wiregrass ( Aristida stricta Michx) (Powell et al., 2005). 105

PAGE 106

Environmental Measurements Environmental information was recorded from April 2007 to August 2007, using an automatic weather station located on the top of a 30 m scaffolding tower, extended approximately 5 m above mean canopy height (Powell et al., 2005). Relative humidity and temperature were measured using a relative humidity and temperature probe (HMP45C-L, Vaisala, Inc., Helsinki, Finland), photosynthetic photon flux density (PPFD, molm-2s-1) was measured using a quantum sensor (Li-190, Li-Cor Inc., Lincoln, NE, USA), wind direction and velocity were measured using an anemometer and vane (03001-L, Campbell Scientific, Logan, UT, USA) and precipitation was recorded with a tipping bucket rain gage (TR525-I, Texas Electronics, Dallas, TX, USA). All sensors were measured each 30 s with an automatic datalogger (CR10X, Campbell Scie ntific, Logan, UT, USA) and were stored as 30 min averages. Tree Selection The measurement trees were chosen from across the range of tree basal area using "quantiles of total", a stratific ation scheme which weights the selection of large trees more heavily (Hatton et al., 19 95; Martin et al. 1997; ermk et al., 2004). Four trees per species were selected for leaf water potential, sapflow, hydr aulic conductivity (on tr unk, roots and branches) and leaf area measurements; an additional th ree individuals were selected for hydraulic conductivity measurements on branches and roots. Average diameter of selected trees averaged 32.6 and 34.8 cm, ranging between 19.8 to 41.8 cm and from 20.4 to 49.3 cm, for LL and SL, respectively. Average sapwood thickness was 5.2 and 6.9 cm, ranging between 3.9 to 6.6 cm and from 4.2 to 9.1 cm, for LL and SL, respectively. Hydraulic Conductivity and Vulnera bility to Cavitation Curves Between April 18th and July 6th 2007 a sample harvest was carried out on branches, roots and trunk in all selected trees. Hydraulic conductivity and air in jection protocol to determine 106

PAGE 107

vulnerability curves on branches and roots segments has been desc ribed elsewhere, as by Sperry et al. (1988), Sperry and Salie ndra (1994), Spicer and Gartne r (2001) and Domec and Gartner (2001). Hydraulic conductivity ( k Kg waters-1MPa-1m) was calculated accord ing to Darcys law as the flow rate of water (Kgs-1) for a given pressure gradie nt through a segment of known length (MPam-1) (Tyree and Zimmermann, 2002). Sapwood-specific ( ks, Kg waters-1m2 sapwoodMPa-1m) or leaf-specific ( kl, Kg waters-1m-2 leaf area MPa-1m) hydraulic conductivity were determined as k divided by corresponding sapwood or leaf ar ea distal to the segment. Leaf area distal to the branch stem segment was determin ed using needle mass and SLA measured for each sample. Sapwood area was determined by measuri ng directly the area on scanned sections cut with razor blade on both ends of the sample us ing image analysis software (ImagePro, Media Cybernetics, Carlsbad, CA, USA). Branches and roots were sampled early in the morning and immediately wrapped in plastic ba gs with wet towels to preven t desiccation and stored in a cooler until measured; both types of samples we re at least 1 m long. Branches were collected from the mid-crown using a self -propelled telescopic lifting m achine and pole pruner the same day of leaf water potential measurements (Apr il 18, 2007); roots were extracted from the upper 20-40 cm depth layer immediately after all bran ch measurements were completed (starting on May 14 and ending on June 6, 2007). For each tree, one sample was collected for branches and three samples were collected for roots, because of high variability in k and vulnerability to cavitation on roots reported by Hacke et al. (2000a) associated with differen ces in root diameter. In the lab, segments were cut free under water (0.15-0.20 m in length and with xylem diameters of 2-12 mm) and fitted with both ends protruding into a Sperry tubing apparatus. One end of the segment was attached to tubing filled with dilute solution of degassed 20 mM KCl and filtered (0.2 m); no pH adjustments were performed (Pockman and Sperry, 2000). The 107

PAGE 108

other end of the segment was attached to a tube filled with the same solution, with a 1 ml micropipette at the end, that conduit the effl ux to the water reservoir on a balance (XA-100, Denver Instruments, Denver, CO, USA) connected to a computer. Water volume changes in the reservoir were determined using changes in weigh measured to the nearest 0.0001 g every 10 seconds over a 5 to 10 minutes interval. As a solution reservoir and to control the hydraulic pressure difference across the segment, a Mariotte tube was used (SMS, Tucson, AZ, USA), maintaining a hydraulic pressure head of 0.0075 MPa for all measurements. After the sample connection was made to the apparatus, flow into the segments without a pressure head (background flow) was measured before and after each gravimetric measurement (Davis et al., 1999; Pittermann and Sperry, 2003). These background measurements were averaged and subtracted from regular gravimetric m easurements to obtain a definitive k value. Initial native conductivity ( knat) was determined as the first measurem ent immediately after sample collection (no more than 4 hours after excision). After knat determination, segments were soaked under vacuum for 48 h to refill embolized tracheids (Domec and Gartner, 2001; Domec et al., 2005) with the same degassed and filtered 20 mM KCl solution in order to estimate maximum conductivity ( kmax). At the same time of leaf ar ea harvesting (between June 21st and July 9th 2007), bole xylem segments sampling was carried out. Hydr aulic conductivity was measured using the pressure sleeve apparatus as described by Spicer and Gartner (1998a, 1998b and 2001). From four trees selected for each species harvest was conducted early in the morning to minimize tensions in the xylem. From each harvested tree two 20-25 cm thick discs were extracted using a chain saw at the sapflow probe measurement height, one disc at approximately 1.8 m height (Sbase) and the other disc at the crown base (S crown). After the disc was cut, the sample was 108

PAGE 109

immediately wrapped in plastic ba gs with wet towels to prevent desiccation, stored in a cooler and transported to the laboratory for measurement. Native conductivity meas urements were conducted within 8 hours of harvest. For each disc, at eight selected depths along stem ra dius, wood xylem samples were extracted to determine the hydraulic c onductivity radial profile (at 1, 2, 3, 4, 6, 8 and 10 cm depth, from outer to inner sapwood). In the case of the crown-base sample from small trees, only samples until a 6 cm depth were extracted). Discs were first split with an axe, keeping the sample wet. Once the disc was split, the parts (with previously mark ed distal and proximal faces) were placed under water and radial and tangential surfaces were obtained with corner chisels. Samples of approximately 1 cm2 size at the cross-sectional surface were extracted at the same depth as the sapflow measurements, more or less 5 cm apart of the sensor position. Still under water, about 57 cm of each end of the sample was cut with ch isel and razor blade, obtaining a final sample approximately 10 cm long. A pressure-sleeve apparatus was construc ted following Spicer and Gartner (1998a) directions in order to seal the sides of the sample during hydraulic conductivity measurements (see APPENDIX D). Samples were fitted inside the latex sleeve of the apparatus and air inside the chamber was pressurized to 0.1 MPa to hold the membrane against the sides to prevent leakage. Sealing was improved by adding silicon gr ease along radial walls of the sample before insertion into the latex sleeve. With pressure held constant inside the chamber and checked constantly by a pressure gauge c onnected to the air inlet, the c onnector of either ends of the sample were flushed using a syringe filled with the filtered solution of KCl (20 mM) to remove air bubbles inside the tubing. Af ter this, the connectors of th e coupling device chamber were attached to the Mariotte tube reservoir of dilute solution (distal end) and to a tube with a 1 ml 109

PAGE 110

pipette (proximal end). Water flows for na tive and maximum hydraulic conductivity were measured following the same procedure as was desc ribed for branches and roots. From each disc, ks weighted value was calculated using the formula: ks = a aksii where corresponds to the ks at depth i ai is the area of the corresponding annulus at depth i and a is total sapwood area of the disc. Radial prof ile in hydraulic conductivity was analyzed after normalizing axial distance of the cambium relativ e to sapwood thickness for each tree (expressed into classes each 20% of sapwood thickness), and using repeated measures analysis including Bonferroni adjustments for differences in least square means (Littell et al., 2006). The vulnerability to cavitation curve (VCcurve) was determined on all roots and branches; in the case of stem bole, VC-curve wa s determined only on sample extracted at a 2 cm depth, because most of sapflow radial profiles me asured in the same trees showed that midday sap flux density peaked close to that depth. The segments were pl aced, both ends protruding, in a double-ended pressure chamber, constructed with portable pressure chamber caps with compression gland cover (PMS Instrument Co., Corvallis, OR, USA) and a custom designed aluminum body (see APPENDIX D). This design improved sample sealing because it was possible to use different gasket s and inserts adapting the chambe r to different sample size and shape. In the case of branches, two shallow ( 0.5 mm deep) notches were cut into opposite sides of the xylem about 5 cm apart in the center of the segment in order to ensure entry of air into the xylem inside the pressure chamber; for roots a nd trunk notching was not necessary. Both ends of the segment were connected to the apparatus in the same way as was done previous for k measurements. 110

PAGE 111

To obtain a VC-curve, the chamber was firs t pressurized to 0.1 MPa for 10 min. Flow through the segment was stopped by closing the input valve. Afte r pressurization time (10 min) conductivity was measured until it st abilized (5 to 10 minutes). Af ter this initial measurement, flow through the segment was reduced by closing th e valve, and air pressure inside the chamber was increased to a prescribed va lue (0.5 MPa steps) and held fo r 10 minutes. Then air pressure was lowered back (depressurized) to 0.1 MPa fo r 3 minutes. Flow was re-introduced by opening the valve and the hydraulic conductance re-measured. Exposure of the segment to progressively higher air pressures continued until hydraulic conductance measurements were at least 95% below the initial value. Before and after each k measurement, background flow rates were also determined. For the trunk samples, segments were moved alternatively from cavitation chamber (where air-injection was carried out) to the pressure sleeve apparatus (where k was measured); in the case of branches and roots k was measured inside the cavitation chamber, the sample place where cavitation was induced. A VC-curve was later constructed for each segment showing the cumulative percentage decrease in h ydraulic conductance versus the negative of airinjection pressure applied. Percentage loss of conductivity (PLC) at a given pressure was calculated using the equation given by Sperry and Tyree (1988 and 1990): PLC( ) = 100 1max )( k k where PLC( ) is the percentage loss of conductivity at pressure k( ) is the hydraulic conductivity measured af ter apply pressure and kmax is the maximum hydraulic conductivity measured previously after vacuum soaking. The pl ot of these data is the VC-curve (Sperry and Tyree, 1988). Using Pammenter and Vander Will inger (1998) approach, a VC-curve can be 111

PAGE 112

described by the following sigmoidal equation: PLC( ) = bae 1 100 where a is an indicator of the slope and b represents the pressure a pplied at which 50% loss of conductivity occurred. As described by Domec a nd Gartner (2001), several parameters can be calculated in order to compare different curves: 50 = b where 50 is the xylem tension (MPa) at which 50% of loss of conductivity occurs. air = a2 + b where air is the air entry point, an estimate of xylem tension (MPa) when cavitation starts and pit membrane is overcome. max = a2 + b where max is the full embolism point, an estimate of the maximum tension (MPa) in the xylem before failing and becoming non-conductive. s = 25a where s is the slope of the linear portion of the VC-curve (% loss of k MPa-1), an estimate of rate of change in loss of k per unit change in xylem tension. Leaf to Sapwood Area Ratio After determination of leaf area (LA, m2) and sapwood area (SA, m2) at Sbase or Scrown (Chapter 4), Huber value (HV, or sapwood to leaf area ratio, m2m-2) was calculated for each tree as SA at Sbase or Scrown, divided by total LA. 112

PAGE 113

Tracheid Length, Diameter and Cell-Wall Thickness Tracheid diameter, double cell-wall thickness an d length were measured in all segments used for the VC-curves. For tracheid lumen diam eter and double cell-wall thickness estimations, cross sections made with a vibratome (Leica VT1000 S; Leica Microsystems, Wetzlar, Germany) of approximately 50 m thickness, were extracted at approx imately 1 cm apart of the proximal end of each sample and mounted in deionized di stilled water over microscope slides. Using a digital camera (Retiga 1300; QIMAGING, Surrey, BC) attached to a light microscope (Olympus Ix70; Olympus, Tokyo, Japan) images of the xylem were captured with a QCapture Suite V2.60 (QIMAGING, Surrey, BC, Canada) with a magnifi cation of 20 for roots and 40 for trunk and branch samples. Ten to 15 images were randomly selected across the radi al distribution of the whole cross section, giving a minimum number of 400 total cells to measure conduit size on each sample. Conduit lumen area and perimeter of all tr acheids contained in each image were measured, encompassing both latewood and earlyw ood using the image analysis software ImagePro Plus 4.0 (Media Cybernetics, Bethesda, MD, USA). Double cell-wall thickness was measured directly in 20 pairs of cell on all images using the same image analysis software; cell wall thickness (cwt, m) was computed as double cell-wa ll thickness divided by two. Using tracheid lumen area and perimeter measurements and assuming an elliptical shape for all conduits, major (a) and minor (b) semi-axis for th e ellipse-equivalent were solved for each tracheid. Equivalent diameter (d) for elliptical conduit was calculated using the formula given by White (1991): d = 4 1 22 332 ba ba 113

PAGE 114

where a and b are major and minor semi-axis for the ellipse-equivalent. For each image, mean total tracheid major and minor semi-axis ( a and b, respectively) were computed as average a and b plus average cwt, respec tively. Mean tracheid area (atr, m2) was computed as atr = ab. The average number of tracheids per unit area ( Tn, m-2) was computed as Tn = atr -1. The mean hydraulic diameter ( Dh, m), was calculated according to Tyree and Zimmermann (2002) as: Dh = 4 1 4 N d where N is the total number of conduits measured for each sample after pooling all 10 to 15 images per cross section. Using Dh in Haggen-Poiseulle equation, an estimate of theoretical whole-wood lumen specific conductivity ( ks-lumen Kgs-1MPa-1m-1) was determined according to Domec et al. (2006): ks-lumen = 1284DhTn where Tn is the average number of tracheids per unit area (m-2) and is the viscosity of water at 25C (8.9x10-10 MPas). ks-lumen value was converted to the same units as ks (Kgs-1mMPa-1) by multiplying by water density equals to 1000 kgm-3 H2O. Whole-wood specif ic conductivity of the pits (Ks-pits Kgs-1MPa-1m-1), which represents the parallel conductivity of all pits on a tissue basis, was computed as the difference betw een the inverses sapwood specific maximum conductivity (e.g. ks-max) and ks-lumen following Domec et al. (2006) rationale, where ks-max represents the conductivities of both, pits and lumen toge ther at full saturation: ks-pit = 111 lumens maxskk Tracheid length was determined with a fibe r quality analyzer (FQA) (OpTest Equipment, Inc., Hawkesbury, Ontario) on previously macerat ed samples (Robertson et al., 1999). Twenty 114

PAGE 115

mm longitudinal strips made with a razor blad e were placed in 10 ml glass containers. A maceration procedure using acetic acid and peroxide was performed following the Ruzin (1999) protocol. Into each vial containing the wood strip, a 1:4:5 maceration solution consisting of 30% acetic peroxide, deionized water and glacial acetic acid was added, but not filling completely the container. Then the vials, loosely covered, we re placed into an oven at 60C for 72-120 h, until the wood segments were whitish to translucent. After this, the samples were washed with deionized water and left overnight in water. Finally the samples were transferred to 50% ethanol and sent to the Wood and Fiber Quality Labora tory at the University of British Columbia, Canada, for FQA analysis. As tracheid length dist ribution is biased toward s shorter tracheids an estimate of tracheid length, the le ngth-weighted tracheid length ( Lt, m), was calculated, as: Lt = ii iiln ln2 where li is the measured fiber length of the ith length class and ni is the number of tracheids measured of the ith length class (Schimleck et al., 2004). Specific Gravity For roots and branch samples that were te sted hydraulically, speci fic gravity (SG) was determined using the water displacement met hod (Hacke et al., 2000b; Pittermann et al., 2006b). One cm length sections long were cut out of the sa mples and SG, the ratio of the density of wood to the density of water, was calculated following the equation given by Simpson (1993): SG = w gdVW/ where Wd is the oven-dry weight of wood (g), Vg is the volume of wood at moisture condition (cm3) and w is the density of water (1 gcm-3). On each sample, the pith as well as the bark were removed with a razor blade. Vg was determined by volume displacement (Archimedes principle) 115

PAGE 116

on the same precision balance used for ks determinations; th e displacement wei ght was converted to sample volume by dividing it with 0.998 gcm3, which is the density of water at 20 C. Afterwards Vg determination, the sections were oven-dried at 60 C for 3 days and Wd was determined. Statistical Analysis Analysis of variance (ANOVA) was used to analyze effects of organ and species in k vulnerability to cavitation and tracheid size, incl uding Bonferroni adjustments for differences in least square means (PROC MIXED, SAS Inc ., Cary, NC, USA). The ANOVA model for the analysis is described in APPENDIX B. Repeated measures analysis was used to analyze radial profile in k data. Results Hydraulic Conductivity and Vu lnerability to Cavitation As sampling was carried out on different days with variable soil moisture conditions ( s at 50 cm depth varied between -0.05 and -1.51 MPa on LL), no comparisons were possible in PLC between organs or species; ho wever, under the assumption that early in the morning, when root sampling was done, s could be used as a proxy of root water potential, there was a good agreement between PLC determined with measurements of ks-nat and ks-max and the estimation with VC -curves using s as root water potential (p<0.0001, r2=0.80); the slope of this relationship was 0.99 when the intercept was forced through zero (F igure 5-3). In general, a low PLC (between 0 to 5%) was obtai ned on days with wet soil ( s ~ -0.1 MPa), and a high PLC (between 20 to 57%) was obtai ned on days with dry soil ( s between -0.7 MPa to -1.5MPa). This result confirms the robustness of the method and gives a solid base for forthcoming analysis and discussion. 116

PAGE 117

There were no significant diffe rences between species in weighted values of native sapwood specific hydraulic conductivity (e.g. ks-nat) for any tissue evaluated (Table 5-2). For LL, roots presented the highest ks-nat (10.39 Kgs-1MPa-1m-1), followed by the stem at the crown base (Scrown), stem at 1.8 m height (Sbase) and branches (8.61, 5.12 and 2.11 Kgs-1MPa-1m-1, respectively). No significant differences in ks-nat between root and Scrown were observed For SL, there was no difference in ks-nat among roots, Sbase and Scrown (7.25, 6.87 and 10.36 Kgs-1MPa1m-1, respectively); branches, on the ot her hand, had significantly lower ks-nat than the other organs (1.49 Kgs-1MPa-1m-1), similar to LL. After removing embolism by vacuum soaking, ksnat increased between 7 to 24%, obtaining mean we ighted values of maximum sapwood specific hydraulic conductivity (e.g. ks-max) for roots, Sbase, Scrown and branches, of 11.86, 6.81, 9.29 and 2.31 Kgs-1MPa-1m-1, respectively, for LL. Corresponding values for SL were 7.92, 7.69, 10.94 and 1.70 Kgs-1MPa-1m-1, respectively. There were differences in ks-max between species only for roots (p=0.023). The same pattern of statistical differences acro ss organs observed for ks-nat was detected for ks-max for both species. Vulnerability to embolism parameters, that were estimated after fitting sigmoidal function for each measured segment (Figure 5-1), indicated that 50 was lower (more negative) in branches than in roots, but not different to Sbase and Scrown for both species. Average 50 of branch, root, Sbase and Scrown was -1.77, -1.32, -1.25 and -1.49 MPa, respectively. A species contrast showed no differences in vulnerability to embolism (e.g. 50) for any organ tested (p>0.11; Table 5-2). The slope of the linear portion of the VC-curve ( s ) represents an index of the sensitivity of ks to changes in xylem wate r potential. The higher the s the higher the reduction in ks ( PLC ) per unit reduction in xylem water potential. No differences between species were found across all organs (p>0.07; Table 5-2). For LL s of branches was smaller than root and 117

PAGE 118

trunk xylem. SL, in contrast, only showed differences in s between roots and branches. (Table 52). The air entry point ( air) and the xylem water potential causing full embolism ( max) were not different between species for each organ measured. In general, max of branches was significantly more negative than Sbase, Scrown and roots. In LL, air was not different across organs, but for SL, air of Sbase was higher than Scrown and roots (Table 5-2). Radial Profile in Hydraulic Conductivity There were no differences in ks-nat and ks-max along radius containing sapwood for both species, either at 1.8 m height or at the crown base (only ks-max is showed in Figure 5-2). Even though average ks-max at 1.8 m height increased from 7.27 to 9.30 Kgs-1MPa-1m-1 from 0-20% to 20-40% sapwood thickness ( swt ) for LL, and from 7.07 to 8.76 Kgs-1MPa-1m-1 at the same swt for SL, this 18% and 24% increase in ks-max was not statistically significant. Close to the fringe of sapwood, at 80-100% of swt average ks-max was 6.55 and 6.76 Kgs-1MPa-1m-1, for LL and SL, respectively, corres ponding to 9.9 and 2.4% lower ks-max than outermost rings (Figure 52a). At the crown base ks-max of LL and SL averaged 8.75, 10.69 and 7.41; and 12.40, 12.37 and 11.68 Kgs-1MPa-1m-1at 0-20%, 40-60% and 80-100% of swt respectively. Within each species there were no statistical differences in ks-max along radius (Figure 5-2b). Tracheid Anatomy There were no significant diffe rences between species in we ighted tracheid length ( Lt) for any tissue evaluated (Table 5-2). Average Lt for roots, Scrown, Sbase and branches was 2.61, 2.37, 1.33 and 1.89 mm for LL; and 2.24, 2.19, 1.11, and 1.16 mm for SL, respectively. For LL, there were no differences in Lt between roots and Scrown, and between Sbase and branches; roots and Scrown Lt was statistically larger than the Sbase and branches. For SL, root Lt was significant larger than branches, but not different to Sbase and Scrown. 118

PAGE 119

Tracheid length in Sbase was not statistically larger than in branches (Table 5-2). Tracheid hydraulic diameter ( Dh) was not different betwee n species for roots, Sbase and branches (p=0.112, 0.845 and 0.689, respectively), but slightly larger for SL at the Scrown (p=0.047). Average Dh for roots, Scrown, Sbase and branches were 37.4 and 32.8, 28.4 and 27.8, 26.8 and 31.6, and 14.1 and 13.7 m, for LL and SL, respectively (Table 5-2). After plotting ks-max and 50 against Lt and Dh for both species pooled (Figure 5-5), there was a significant positive correlation between Lt and Dh with ks-max (p<0.0001); variation in Lt and Dh accounted, separately, for 46 % and 73% of variation in ks-max; no differences between species were observed for thes e relationships (p>0.38). While Dh was linearly correlated with ksmax (Figure 5-5b), Lt was non-linearly correlated with ks-max, independently of the organ and species tested (Figure 5-5a). For Lt smaller than approximately 1.5-2.0 mm, resistance to cavitation increased sharply as Lt decreased; beyond that threshold value, 50 was relatively constant around -1.0 to -1.5 MPa independent of the organ and species te sted, not being related to changes in Lt. When all organs were considered, th ere was a significant (p=0.0003) but weak (r2=0.29) relationship between Dh and 50, but within roots, Scrown, Sbase and branches, there was no relationship between Dh and 50. Between species, cell-wall thickness ( cwt ) was only different for roots (p<0.0001), averaging 3.83 and 2.76 for LL and SL, respectively. For both species, tracheid cell walls were thicker at Sbase and Scrown than in roots and branches, with cwt of roots also larger than branches (average cwt across species was 2.45, 5.97 and 5.81 mm, for branches, Sbase and Scrown, respectively; Table 5-2). Afte r calculating the ratio between cwt and Dh ( C ), there were no significant differences between species in C for any tissue evaluated, averaging 0.09, 0.18, 0.21 and 0.20, for roots, branches, Sbase and Scrown, respectively. For tracheids of similar Dh 119

PAGE 120

(averaging, approximately, 30 m in Figure 5-6a), root xylem ce lls were much less reinforced (smaller C ) than Sbase and Scrown. On the other hand, branch trachei ds were also more reinforced than roots, but a reduction in diameter of branch tracheids, for similar cwt compared with roots, was the cause of the difference in C (Figure 5-6a). The number of tracheids per unit area (Tn) was not significantly different between species for branches, Sbase and Scrown, but was different for roots (p=0.0467, Table 5-2). Average Tn was significantly higher in branch es than the other tissue s, averaging 4136, 1095, 1091 and 1061 tracheids per mm2, for branches, roots, Sbase and Scrown, respectively. When all organs were considered, there was a positive correlation between Lt and Dh (p<0.0001, r=0.67; Figure 5-6b), but if branches were not include d, there was no relationship between Lt and Dh (p=0.29, r=0.20), for tracheids in branches, Sbase and Scrown. In order to get a better estimate of the imp act of each tracheid anatomy trait measured into ks-max and 50, multiple linear regression was performed; cwt Tn and C as well as Lt and Dh. For ks-max, Dh, Lt and C were all significant variables, ac counting with 73.4, 2.4 and 1.6% of total ks-max variability (whole-model r2=0.767). For 50, only Tn was significant, accounting with 30.2% of total 50 variability. For the same Dh, tracheids in Sbase were longer than tracheids at the Scrown (Table 5-2) and this difference in tracheid length (included as Dh to Lt ratio in covariance analysis) accounted for most of the variation in ks-max between Sbase and Scrown (p=0.033). Specific Gravity There were no differences between species in mean SG for roots and branches (p=0.909 and 0.678, respectively), but significant differe nces between organs within each species (p<0.0001 for both species). Average SG for root s and branches across species was 0.226 and 0.429, respectively. SG was strongly correlated with cell-wall to hydraulic lumen diameter ratio 120

PAGE 121

( C ) (p<0.0001; r2=0.88) of roots and branches. SG was not correlated with ks-max, either in roots or branches (p= 0.69 and 0.75, respectively). Pit and Lumen Conductivity There was no difference between species in ks-lumen or ks-pit, for any organ measured (Table 5-2). For roots, pit resistance (1/ ks-pit) accounted for 72% of to tal hydraulic resistances, while for branch, Sbase and Scrown, ks-pit accounted for 53.7, 36.6 and 49.0% of the total resistances, respectively. Tracheid length was positively correlated with ks-pit when branches, Sbase and Scrown were pooled (p<0.0001, r2=0.71; Figure 5-8), but not for roots (p=0.39). The relationship between ks-max and ks-lumen can be used as an index of the relative efficiency of water transport of the xylem in relation to the maxi mum theoretical conductivity according to the Haggen-Poiseulle equation for capillaries. Ther e was a significant and positive relationship between ks-max and the theoretical calculated ks (e.g. ks-lumen); when all organs were pooled (Figure 5-9), an average ratio between ks-max and ks-lumen was 0.208, but this relationship is improved when roots were excluded, increasing the r2 from 0.60 to 0.84, with a new slope of 0.471. Stem xylem of branches and Scrown were the organs which deviated the least from the theoretical conductivity, averagi ng, across species, ks-max to ks-lumen ratio of 49.4 and 59.4%, respectively. Conversely, for Sbase and roots, this ratio was signif icantly lower than roots and the Scrown (p<0.05, averaging 36.5 and 25.5%, respectively .Between species only roots were different in ksmax to ks-lumen ratio (p=0.019), averaging 0.295 and 0.201 for LL and SL, respectively. Discussion Root xylem had higher water transport efficiency and also was the organ least resistant to cavitation (as compared with br anches and trunk sapwood). Similar results have been reported elsewhere (Domec et al., 2006; Hacke et al., 2000 a and b; Kavanagh et al ., 1999; Maherali et al., 2006; Martinez-Villalta, 2002). Our mean values of 50 for roots were similar to those reported 121

PAGE 122

for Pinus caribaea M., and Pinus taeda L., species also member of the subsection Australes of the Pinus genus that have been hybridized with longleaf and slash pine trees in breeding programs (Duncan et al., 1996). For P. caribaea root 50 averaged -1.91 MPa (Pittermann et al., 2006b), for P. taeda root 50 average ranged between -1.05 to -2.3 MPa (Hacke et al., 2000a) and -1.74 MPa (Maherali et al., 2006). It has been hypothesized that r oots, which tend to be more vulnerable to cavitation than other tree organs, act as a hydraulic "fuse" for the plant, limiting cavitation during soil drying to "expendable" roots and protecting the aboveground parts of the hydraulic system (Sperry et al. 2002). Our data, when considered in the context of within-tree water potential gradients, do not support this hypothesis. In this study, the average difference in 50 between roots and Scrown was -0.26 MPa, and the gravitational po tential at that height averaged cross species was -0.16 MPa (p=0.84). Therefore, in real terms (discountin g gravitational potential) the difference in 50 between roots and Scrown was only 0.1 MPa higher in the roots. When additional gradients due to friction were considered, the difference in 50 between roots and upper stem is compensatory and likely results in eq ual vulnerability to cavita tion. It is important to consider that most hydraulic conductivity studies in pines have been carried out with coarse roots (>2 mm diameter) and the construction cost of that tissue in term of carbon, even though it is cheaper than branches or trunk, is high and directly correlated with loss of ab sorption of nutrients a nd water. Johnsen et al. (2005) and Matamala et al. (200 3) reported, for loblolly pine, that fine root mean life span was between 4.5 and 5.7 years, with even higher replacement time for coarse roots. Therefore, loss of the root absorbing surface due to xylem cavitation was not a trivial issue for pine trees, and lateral roots (as those evaluated in this study) must be kept in place in order to ensure water and nutrient transport to carbonfixing sites. 122

PAGE 123

The fact that Dh and Lt were correlated (r=0.67) explains the lack of importance of Lt on total ks when the analysis was done across different organs. Nevertheless, for stem xylem at different heights (at Sbase and Scrown, approximately 1.8 m and 20 m height, respectively), the difference in ks-max between these points was corre lated with the difference in Lt between them ( Dh and cwt were not significantl y different between Sbase and Scrown, and ks-max at Scrown was 39.5% higher than at Sbase). Domec et al., 2005 also reported a positive correlation between ks and Lt in the stem. The same study also reported a positive correlation between trach eid length and pit density and a negative correlation between tracheid diameter and pit chamber size. Th e positive relationship between ks-pit and Lt (Figure 5-8) likely resulted because longer tracheids are wider and therefore have a greater total area of pit membrane connection, and larger tracheids have more porous pit membranes (Choat et al., 2005). Furthermore, Comstock (1970) indicated that the number of pits transversed in series per uni t length was invers ely proportional to Lt, so the larger the tracheid, the fewer the pits that have to be cros sed as water moves longitudinally. Pit resistance in roots accounted for 72% of the total hydrau lic resistance, while at the crown base pit resistance accounted for only 36.6%, even though Dh and Lt of roots were similar or greater than at Scrown (Table 5-2). Differences in pit permeability density and distribution between roots and branches and stem tracheids may explain why roots showed no relationship with ks-pit. Domec et al. (2006) concluded that even though tracheid Dh was correlated with pit functioning, pit aperture and pit pore size were the tracheid stru cture characteristics that ultimately constrained hydraulic conductivity. The relationship between ks-max and 50 indicated a significant but weak trade-off between water conduction efficiency and safety (p=0.009, r2=0.16; Figure 5-4) Several authors 123

PAGE 124

have reported similar results for gymnosperms. Fo r example, Burguess et al. (2006) for branches of Sequoia sempervirens and Maherali and De Lucia (2000) for branches of Pinus ponderosa ; in contrast, Hacke et al. (2000b) reported no co rrelation between root tracheid diameter and cavitation resistance. Conduit diameter was related to ks but was not necessarily important in determining 50 (Figure 5-5b and d), because the mech anism of water stress-induced embolism (Tyree and Zimmermann, 2002; Hacke et al., 2004; Domec et al., 2006) and embolism repair (Zwieniecki and Holbrook, 2000) are related to pit characteristics. As we did not measure SG on trunk sapwood, C was used as an index of mechanical strength for inclusion in the analysis. C was not correlated with ks-max, for any of the organs examined (p>0.253). After pooling roots and branches, a negative exponential relationship (p<0.0001; r=-0.768) was evidenced acr oss roots and branches (circles in Figure 5-7b), similar to that reported across species (Santiago et al., 2004; Christensen-Dalsgaard et al., 2007; Pittermann et al., 2006 a and b). However, when trunk data were included (triangles in Figure 5-7) the relationship was less evident (p=0.018; r=-0.36). Similar patterns emerged for the relationship between C and 50; without trunk xylem, p=0.0002 and r= -0.668, but after including trunk xylem there was no relationship (p=0.094 and r=-0.268). Sperry et al. (2006) also reported a weak relationship between em bolism resistance and wood density (r2=0.16 for branches and 0.38 for roots) for different conifer species. Lateral root tracheids have evolved for differe nt functions than tr acheids of trunk or branches, and that was reflected in differences in tracheid allometry re lated with differential mechanical reinforcement needs. For example, C was almost 50% lower in roots than in branches or tr unk tracheids (Table 5-2), with no significant differences between these two last organs In contrast, branch tracheids had cwt and Dh values smaller than trunk tracheids. Th ese differences in tracheid size were 124

PAGE 125

finally reflected in the number of cells per unit ar ea, where branches had three to four times more cells per mm2 than roots and trunk. These results suggest that when tracheid size was related to ks or 50 across organs, caution must be taken because re sults must depend on organ evaluated. Both species presented similar mean valu es on most evaluated hydraulic traits. No differences in either, ks-nat and 50 were concordant with other water relation traits measured on the same trees, such as transpiration rate per unit leaf area, water storage use, stomatal regulation and whole-tree hydraulic conductance (Chapter 3). Nevertheless, root maximum sapwoodspecific hydraulic conductivity was significant larger in LL (p =0.023; Table 5-2). This finding, associated with Huber value differences (even not statistica lly significant at Scrown, but more than 43% higher on LL than SL), may have important eco logical implications associated with habitat characteristics of both species, where LL has the tendency to inhabit drier soil microsites than SL. Differences in HV can be better assesse d if we examine the model suggested by Whitehead et al. (1984), that combines Penm an-Monteith model and Darcys Law, where stomatal (crown) conductance was directly proportional to ks, and HV, and inversely proportional to D and tree height: to maintain high stomatal conductance under water stressed conditions, as low soil moisture, without the risk of excessive em bolism by decreasing (because vulnerability to embolism is similar be tween both species, Table 5-2), LL trends to produce roots with higher ks and at the same time allocate less carbon to foliage per unit SA. This suite of traits as consistent with an adaptation to avoid soil water deficits, because higher root ks allows LL to maintain high water supply to foliage at similar (Table 5-2), and lower LA in LL results in reduced transpiration, and avoidance of excessive desi ccation. De Lucia et al. (2000) points out that for Pinus ponderosa higher HV was associated to acclimations to dry conditions. 125

PAGE 126

Other difference in hydraulic architecture be tween LL and SL, but not assessed directly in this study, could be species differences in ro ot distribution. Gholz et al. (1986) reported that, in a 27 year-old SL stand, ar ound 80% of total root biomass was found in the top 10 cm, and only 0% was present below a 50 cm soil depth. On the other hand, Addington et al. (2006) reported, for 44 year-old LL, root area index distribution with so il depth. Approximately, 32% of the total roots accumulated in the top 20 cm soil depth, and about 44% below 60 cm soil depth. This more even root distribution of LL with dept h could be an adaptation for more xeric habitats, allowing LL to reach deeper zones along soil profile More research is needed to validate this assumption, by measuring root lengt h distribution of both species c ohabiting the same site. Conclusion The analysis of hydraulic archit ecture and tracheid allometry revealed several similarities between mature LL and SL trees. Tracheid allometry changed mark edly between roots, trunk and branches sapwood, and tracheid lumen diameter was highly correlated with sapwood hydraulic conductivity. Vulnerability to cavita tion was not different between species for sapwood of roots, branches or trunk. On the other hand, higher sapwood to leaf area ratio and higher maximum sapwood-specific hydraulic conductivity in roots of LL were anatomi cal traits that may allow LL to survive and dominate in drie r soil microsites of the flatwoods of north Florida. SL, on the other hand, with its ability to produce aeren chyma under low oxygen condition when the water table rises, appears most suitable to survive and dominate at the fr inge of ponds and wetter zones. 126

PAGE 127

Table 5-1. Least-square means of diameter at br east height (DBH), to tal height, total stem volume without bark (Volume), leaf area (L A) and all-sided specific leaf area (SLA) for long leaf and slash pines. Trait Unit Species p > F Longleaf Slash species DBH (cm) 32.61 34.85 0.664 Height (m) 24.51 26.67 0.161 Volume (m3) 0.91 1.21 0.545 LA (m2) 61.22 104.84 0.127 SLA (cm2g-1) 96.29 98.06 0.812 p-values using mixed model procedure (n=7 for DBH and Height ; n=4 for all other variables). 127

PAGE 128

Table 5-2. Least square means of hydraulic and tr acheid anatomy traits evaluated for roots, stem bole at 1.8 m height (Sbase), stem bole at crown base (Scrown) and branches for longleaf and slash pine trees. Trait Tissue Species p > F Longleaf Slash species ks-nat Root 10.39 a 7.25 a 0.257 Sbase 5.12 b 6.87 a 0.099 Scrown 8.61 a 10.36 a 0.415 Branch 2.11 c 1.49 c 0.251 ks-max Root 11.86 a 7.92 a 0.023 Sbase 6.81 b 7.69 a 0.178 Scrown 9.29 b 10.94 a 0.422 Branch 2.31 c 1.70 c 0.322 SA Sbase 0.040 a 0.052 a 0.438 Scrown 0.020 b 0.027 b 0.474 HV Sbase 8.25 a 5.15 a 0.086 Scrown 3.86 b 2.68 b 0.189 ST Sbase 5.21 a 6.93 a 0.078 Scrown 5.61 a 5.86 b 0.822 k Sbase 0.26 a 0.41 a 0.297 Scrown 0.19 a 0.32 a 0.330 50 Root -1.31 b -1.33 b 0.878 Sbase -1.35 ab -1.16 ab 0.111 Scrown -1.47 ab -1.52 ab 0.718 Branch -1.81 a -1.74 a 0.758 s Root 87.66 b 96.9 b 0.262 Sbase 61.61 b 60.75 ab 0.998 Scrown 67.31 b 74.93 ab 0.606 Branch 41.73 a 51.84 a 0.115 air Root -0.69 a -0.69 a 0.721 Sbase -0.52 a -0.32 b 0.074 Scrown -0.72 a -0.82 a 0.537 Branch -0.55 a -0.72 ab 0.393 midday Root -1.96 b -1.95 b 0.944 Sbase -2.18 b -1.99 b 0.217 Scrown -2.22 b -2.22 b 0.366 Branch -3.04 a -2.75 a 0.984 SG Root 0.222 a 0.229 a 0.678 Branch 0.428 b 0.429 b 0.909 Lt Root 2.61 a 2.44 a 0.448 Sbase 1.33 b 1.37 b 0.230 Scrown 2.24 a 2.20 a 0.848 Branch 1.11 b 1.16 b 0.374 Dh Root 37.42 a 32.81 a 0.238 Sbase 28.37 b 28.72 a 0.656 128

PAGE 129

Table 5.2. Continued. Trait Tissue Species p > F Longleaf Slash species Scrown 26.78 b 31.60 a 0.047 Branch 14.14 c 13.66 b 0.689 cwt Root 3.83 b 2.77 b 0.014 Sbase 5.75 a 6.19 a 0.470 Scrown 5.73 a 5.89 a 0.822 Branch 2.41 c 2.49 b 0.536 cwt / Dh Root 0.105 a 0.093 a 0.194 Sbase 0.207 b 0.217 b 0.173 Scrown 0.214 b 0.186 b 0.151 Branch 0.176 b 0.184 b 0.640 Tn Root 905 b 1285 b 0.047 Sbase 1153 b 1029 b 0.665 Scrown 1208 b 916 b 0.274 Branch 4224 a 4048 a 0.714 ks-lumen Root 44.08 a 34.84 a 0.018 Sbase 16.15 b 17.14 b 0.758 Scrown 14.83 b 22.59 ab 0.093 Branch 4.39 c 3.47 c 0.430 ks-pit Root 18.70 ab 11.05 ab 0.037 Sbase 9.40 b 12.87 ab 0.825 Scrown 25.14 a 28.13 a 0.623 Branch 5.71 b 4.89 b 0.765 Significant differences (p<0.05) in parameters be tween organs are indicated by different letters. p-values using mixed model proced ure (n=7 for roots and branches; n=4 for Sbase and Scrown ). 129

PAGE 130

PLC (%) 0 20 40 60 80 100 LL SL Pressure Applied (MPa) 01234 PLC (%) 0 20 40 60 80 100 Pressure Applied (MPa) 01234 PLC (%) 0 20 40 60 80 100 0 20 40 60 80 100 PLC (%) Figure 5-1. Percentage loss of conductivity (PLC ) versus the applied air pr essure in longleaf pine (filled circle) and slash pine (open circle) (m ean SE) for roots (a; n=7), branches (b; n=7), stem at 1.8 m (c; n=4) and stem at crown base (d; n=4). a b c d 130

PAGE 131

k smax (Kg s -1 MPa -1 m -1 ) 0 2 4 6 8 10 12 14 k smax (Kg s -1 MPa -1 m -1 ) 0 2 4 6 8 10 12 14 SL LL Radial Relative axial distance 0.10.30.50.70.91.11.31.5 PLC (%) 0 10 20 30 40 Radial Relative axial distance 0.10.30.50.70.91.11.31.5 PLC (%) 0 10 20 30 40 a b c d Figure 5-2. Radial profile in maximum sapwood-specific hydraulic conductivity ( ks-max) (a, b) and percentage lo ss of conductivity ( PLC ) (c, d) for stem at 1.8 m height (a, c) and at crown base (b, d) for longleaf and slash pi ne. Relative axial distance is defined as the distance from bark to measurement point divided by distance from bark to sapwood/heartwood boundary. 131

PAGE 132

PLC estimated 0102030405060 PLC measured 0 10 20 30 40 50 60 LL SL linear fit p < 0.0001; r2=0.69; slope=0.67 Figure 5-3. Estimated using VC-curves v/s measured PLC on roots and branches of longleaf and slash pine trees. As samples were taken early in the morning, soil matric potential was used to estimate xylem water potential. The diagonal line represen ts a 1:1 relationship and the dashed line represents a linear fit. 132

PAGE 133

k s-max (Kg s -1 MPa -1 m -1 ) 0 5 10 15 20 50 (MPa) -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 Branch Root S base S crown Figure 5-4. Xylem tension that caus es 50% of loss of conductivity ( 50) versus maximum sapwood specific hydraulic conductivity ( ks-max) for branch, root, Sbase and Scrown. Both species included. 133

PAGE 134

0 5 10 15 20 0 5 10 15 20 Branch k s-max (Kg s -1 MPa -1 m -1 ) Root K s-max (Kg s -1 MPa -1 m -1 ) S base S crown L t (mm) 0.51.01.52.02.5 3.0 50 (MPa) -3.0 -2.5 -2.0 -1.5 -1.0 D h ( m) 1020304050 50 (MPa) -3.0 -2.5 -2.0 -1.5 -1.0 Figure 5-5. Maximum sapwood sp ecific hydraulic conductivity ( ks-max) (a, b) and xylem tension that causes 50% of loss of conductivity ( 50) (c, d) versus tracheid length ( Lt) (a, c) and hydraulic tracheid lumen diameter ( Dh) (b, d) for longleaf and slash pine trees pooled. All organs measured (root, Sbase, Scrown and branch) are included. a b c d 134

PAGE 135

cwt ( m) 234567 D h ( m) 10 20 30 40 50 Branch Root S base S crown L t (mm) 1.01.52.02.53.0 D h ( m) 0 10 20 30 40 50 a b Figure 5-6. Allometry of tracheids of longleaf and slash pine trees. Relationship of mean hydraulic diameter ( Dh) with cell wall thickness ( cwt ) (a) and tracheid length ( Lt) (b) for different organs evalua ted. Both species included. 135

PAGE 136

Specific Gravity (%) 0.1 0.2 0.3 0.4 0.5 Branch Root linear fit p<0.0001, r 2 =0.80 k s-max (Kg s -1 MPa -1 m -1 ) 0 5 10 15 Branch Root Sbase Scrown cell wall to lumen diameter ratio (C) 0.050.100.150.200.250.30 -3.0 -2.5 -2.0 -1.5 -1.0 (MPa) 50 Branch Root Sbase Scrown Figure 5-7. Relationships between cellwall to hydraulic diameter ratio ( C ) and (d) specific gravity (SG), (e) maximum sapwood specific hydraulic conductivity ( ks-max) and (f) vulnerability to cavitation ( 50), for roots, branches, Sbase and Scrown of longleaf and slash pine (see symbol legend on each figure). 136

PAGE 137

L t (mm) 0.51.01.52.02.53.0 k s-pit (Kg s -1 MPa -1 m -1 ) 0 10 20 30 40 Branch Root Sbase Scrown linear fit for Branch, Sbase and Scrown p < 0.0001; r2=0.71 linear fit for Root p=0.39; r2=0.06 Figure 5-8. Relationship be tween tracheid length ( Lt) and sapwood-specific hydraulic conductivity of pits (ks-pit) for different organs evaluated in longleaf and slash pine trees. Lines represent linear fit for branch es, Sbase and Scrown (solid) and roots (dashed). 137

PAGE 138

k s-lumen 0102030405060 k s-max 0 10 20 30 40 50 60 Branch Root Sbase Scrown linear fit for all organs p < 0.0001; r2=0.60; slope=0.208 inear fit excuding roots p < 0.0001; r2=0.84; slope=0.471 Figure 5-9. Maximum sapwood sp ecific hydraulic conductivity ( ks-max) versus sapwood-specific hydraulic conductivity of lumen ( ks-lumen) for branch, root, Sbase and Scrown. The solid diagonal line represents a 1:1 relationship and the dashed lin es represent a linear fit. Both species included. 138

PAGE 139

CHAPTER 6 SUMMARY AND CONCLUSIONS The overall goal of this di ssertation was to obtain a d eepen knowledge about water relations of three major species of southern pines: loblolly, longleaf and slash pine. The study was divided in two principal areas: (1) assessment of water availabi lity and genetic family effects on water relation traits, growth and wood properties of mid-rotation loblolly pine; and (2) characterize water relation traits of mature longleaf and slash pine trees cohabiting the same site and the relationship of these traits with microsite habita t preference of each species. For the first study, water availability was c ontrolled by irrigation for two fast growing families that received the extra water input since plantation establishment. This study was divided in two main areas: (a) water availability and family effects on water use, whole-tree hydraulic conductance and canopy conductance respons es to varying environmental conditions; and (b) water availability and fa mily effects on wood properties, examining the responses at age eleven on date of diameter growth cessation a nd its relationship with specific gravity and latewood percentage. For long term responses, the effect of irrigation on wood properties, since plantation establishment, was examin ed using the complete wood core. The second study, which was carried out in a naturally-regenerated mixed stand of longleaf and slash pine on a flatw ood site in north-central Florida was also divided into two main areas of research: (a) species characteristics in water relation traits such as total daily transpiration, water storage use, whole-tree hydraulic conductance, leaf to sapwood area and hydraulic adjustments in crown conductance to varying environmental conditions; and (b) compare species characteristics in hydraulic cond uctivity, vulnerability to cavitation and tracheid allometry of roots, stem and branches. 139

PAGE 140

Based on the results of the previous chapters the main conclusions and implications can be summarized as follows: Water Availability and Family Effects on Water Relations on 11-year-old Loblolly Pine Under rainfed conditions, the water used for transpiration (58%) was taken from the upper 35 cm depth of the solum and, on average across the season, only 10% of the total transpiration was sustained from water below 1 m depth. The transpiration-induced water potent ial gradient from roots to shoots ( ) was relatively constant across treatments and genetic family, av eraging 0.75 MPa. This re flects stomatal control that maintains relatively constant but at the same time allows leaf to fluctuate dramatically in synchrony with soil. Increasing soil water availability via irrigation affected water use mainly by maintaining high whole-tree hydraulic conductance ( KS-wt) rather than increasing LAI or stomata behavior. Under non-irrigated conditions KS-wt was reduced to less than a third of what irrigated trees showed, a response perhaps related to avoiding xylem embolism under non-water stressed conditions. There was an irrigation x family interacti on in the response of canopy conductance to water vapor ( GC) to vapor pressure deficit ( D ). This GxE interaction im plies that under water-limited conditions, South Carolina Coastal Plain family trees (SC) had stronger stomatal control than Florida families trees (FL), although this differe nce was not present when water was not limiting. There was no family difference in the response of canopy conductance to water vapor to changes in volumetric water content ( v). The two genetic families evaluated also showed differences in foliar N and 13C, reflecting higher water use efficiency for SC compared with FL. Family differences in water re lation traits listed above reflect differences in adaptive traits related to ambient humidity and water availabili ty in their provenance ranges. These different environments may have driven different evol ving pathways expresse d in higher water use efficiency in trees from the Atlantic Coastal Plai n family compared to trees from the mix of north Florida families. Water Availability and Family Effects on Basal Area Growth and Wood Properties of Loblolly Pine Increasing soil water availability via irrigati on from June to November on 11 year-old trees increased specific gravity (SG) and latewood percentage (LW%) by an average of 0.036 and 6.93%, respectively. The mechanism of this response was an extension of diameter growth by 25 days when trees wee irrigated. The main effect of eliminating soil water st ress though irrigation was an increase in latewood growth. Before canopy closure, irrigation caused null or negative effects on SG and LW% due to 140

PAGE 141

the large effect on earlywood growth associated with fast LAI development. After year 7, earlywood growth was similar between the contro l and irrigated trees but latewood growth was larger on irrigated plots, increa sing the overall year-ring SG and LW %. No effect of irrigation on latewood specific gravity was detected. Trees from SC had larger SG and LW% than tr ees from FL, independent of irrigation; this effect was associated with greate r yearly latewood growth in SC. Water Relations of Longleaf and Slash Pine Mean daily transpiration rate ( Eday) was higher for slash pine (SL) than longleaf pine (LL) trees, averaging 50.0 and 33.2 lday-1tree-1, respectively. Stored water use was an important component of daily water use, being similar for both species and linearly related to tree size. Average daily water storage use, across species and measurement period, was 9.3 lday-1tree-1, corresponding, approximately, to 31% of daily transpiration rate. Even though there were no significant differences between species in daily transpiration rate per unit leaf area ( El), there was a species difference in th e response of stomatal conductance at crown scale ( Gcrown) to changes in soil volumetric water content ( v); LL had larger Gcrown on days with high v and reduced to similar or even lowe r values than SL on days with low v. Both species showed no differences in maxi mum root-to-leaf water potential gradient ( ) and sensitivity of Gcrown to increasing D evidencing similar stomatal control, presumably maintaining high rates of carbon ga in and at the same time avoiding excessive xylem cavitation. Species differences in transpiration rate were principally determin ed by differences in leaf area per tree (LA). SL had 60% more LA per unit basal sapwood area than LL (p=0.086), but larger plasticity of LL stomatal regulation partia lly compensates for leaf area differences. Hydraulic Architecture and Tracheid A llometry of Longleaf and Slash Pine Root xylem had higher sapwood-specific hydraulic conductivity ( ks) and also was the organ least resistant to cavitation (as compared with branches and trunk sapwood). Root ks of LL was larger than SL, but there were no species differences for any other organ tested. There were no differences in vulnerability to cavitation between species in any of the organs evaluated. Across all organs, there was a significant but weak trade-off between water conduction efficiency and safety. Tracheid hydraulic diameter ( Dh) was strongly correlated with ks across all organs, explaining more than 73% of the variation in ks; in contrast tracheid length ( Lt) only contributed 2.4% of the variability. Nevertheless, for stem xylem at different heights (at approximately 1.8 m and 20 m height), the difference in ks between these points (39.5% higher ks at 20 m height) was correlated 141

PAGE 142

with the difference in Lt between them ( Dh and cwt were not signifi cantly different between those points). Tracheid allometry changed markedly between sapwood of roots, stem and branch sapwood, reflecting different mechanical reinforcement needs. Higher sapwood to leaf area ratio, higher maximum ks in roots and higher Gcrown under high v conditions are anatomical and physiological featur es that may allow LL to survive and dominate in drier soil microsites of flatwoods of north Flor ida. SL, on the other hand, with its ability to produce aerenchyma under low oxygen conditions when the water table rises, is most suitable to survive and dominate at the fr inge of ponds and wetter zones. Comparison across Experiments Even though the experimental un its are different (plot for LO and individual tree for LL and SL), it is possible to qualitatively compar e the mean values of some hydraulic traits evaluated among the three species: The three species showed the same relationshi p between crown (or can opy) sensitivity to increasing D and crown (or canopy) conductance at D =1 kPa (Figure 6-1); the same trade off was observed for all three species and the general slope includi ng all species was 0.619. Loblolly pine (LO) had GCref 42% and 95% higher than LL and SL, re spectively. However the sensitivity of GCref to D was also higher in the same proportion, maintaining the same relationship across species. It is important to remark that even t hough the relationship described in Figure 3 is a useful summary graph, the high correlation between both variables could be an artifact of the regression method used to derive the parameters estimates, but several authors such as Monteith (1995), Granier et al. (1996) and Oren et al. (1999), support this approach The relationship between volumetric water co ntent at a 50 cm soil depth and average daily crown (LL and SL) or canopy (LO) conductance showed similar response at v below approximately 0.11, but at high soil moisture conditions LO presented much higher GCday (Figure 6-2a). Under favorable soil moisture conditions GCday of LO overcomes LL and SL; under dry soil conditions all thre e species reduces stomatal conductan ce at similar rates. For example, 142

PAGE 143

under wet soil conditions ( v between 0.12 and 0.14), GCday averaged 57.3, 37.2 and 27.1 mmolm-2s-1 for LO, LL and SL, respectively, while under dry soil conditions ( v below 0.10), GCday averaged 19.2, 18.2 and 19.5 mmolm-2s-1 for LO, LL and SL, respectively. For LO, family effect was observed at low v, when FL maintains higher GCday than SC; under wet soil conditions no difference occurred between species (Figure 6-2b). Average maximum water potential gradient ( ) was 24% lower in LO compared with LL and SL (0.75, 0.99 and 0.95 MPa, respectively). Similar was the behavior of whole-tree sapwood-specific hydraulic conductance ( KS-wt), where even irrigated LO had lower mean values than LL and SL, averaging 2.87, 3.37 and 4.20 molm-2s-1MPa-1, respectively. After including both parameters in order to estimate tr anspiration rate per unit sapwood area ( ES, molm-2s-1), it was also lower on LO compared wi th LL and SL (2.09, 3.33 and 3.99 molm-2s-1, respectively). The results of this study should deepen our knowledge of the water relations of southern pines, impacting future tree growth modeli ng and management decision-making related to species (for loblolly, long leaf, and slash) and seed source (for loblolly). Future Research Areas Should Focus on 1. For water availability and genetic family effects. Investigate mechanism behind GxE interaction in GC response to D Incorporate season 2006 wood properties into a dataset in or der to corrobor ate results. With 2006 ring data, analyze th e effects of irrigation and family on juvenile/mature wood transition. Evaluate the effect of nutrition and family (and interactions with water availability) on water relations and wood propert ies in loblolly pine. Does nutrition affects GC response to D or KS-wt on both families? Does nutrition affects wood properties? Evaluate the effect of irri gation on latewood initiation date. 143

PAGE 144

Evaluate genetic associations between high LW% in the SC family and known genes related to wood properties, as cad (linked to lignin biosynthesis) or lp3-1 ( linked to water stress inducible protein 1). Evaluate family x irrigation effects on hydraulic architecture ( ks and vulnerability to cavitation). Incorporate family-based information obtaine d into physiological-based growth models and water transport models. 2. For water relations and hydrau lic architecture of LL and SL. Associate hydraulic architecture and tracheid allometry with water use, Gcrown and KS-wt (integrate chapters 4 and 5). Measure bordered pit characteristics (densi ty, size and permeability) on roots, stem and branches of both species in order to a ccount for part of the variability in ks and vulnerability to cavitation not explained by tracheid diameter and length, and also better e xplain the lean trade off between water transport efficiency and safety. Integrate water use estimations of this study wi th eddy covariance measurements at the same site. Determine understo ry evapotranspiration. Incorporate information collected into physiolo gical-based growth models and water transport models. 144

PAGE 145

GCref (mmol m-2 s-1) 20406080100120140160180200 -d GC/dlnD 0 20 40 60 80 100 120 Loblolly Control Loblolly Irrigated Longleaf Slash linear fit; p<0.0001; r2=0.97 slope=0.619 Figure 6-1. Relationship between sensitivity of stomatal conductance at crown (longleaf and slash) and canopy (loblolly) scales to D (-d GC/dln D ) and canopy conductance at D =1 kPa ( GCref) for loblolly rainfed (open triangle) and irrigate d (open circle), longleaf (filled triangle) and slash (filled circle) pines. 145

PAGE 146

v (cm3cm-3) 0.080.100.120.140.160.180.20 GCday (mmol m-2s-1) 0 20 40 60 80 100 120 Loblolly-Control Loblolly-Irrigated Longleaf Slash v (cm3cm-3) 0.080.100.120.140.160.180.20 GCday ( mmol m-2s-1 ) 0 20 40 60 80 100 120 Control-FL Control-SC Irrigated-FL Irrigated-SC a b Figure 6-2. Relationship between average daily st omatal conductance per unit leaf area at (a) crown (longleaf and slash) or canopy (loblolly) basis and vo lumetric soil water content (v) at 50 cm soil depth for loblolly rain fed (open triangle) and irrigated (open circle), longleaf (filled triangle) and slash (filled ci rcle) pines. Panel (b) represents the same relationship but only for loblolly pine (at canopy basis), under irrigation treatment (control and irrigated) including tw o different genetic families (FL and SC). 146

PAGE 147

APPENDIX A ANOVA MODEL FOR IRRIGATIO N X FAMILY ANALYSIS FOR LOBLOLLY PINE Yijk = + bi + Ij + Fk(j) + ( IF )k(j) + ( bI )ij + ( bF )ik(j) + ( bIF ) ik(j) + ijk where Yijk is the parameter value of the plot of the kth family nested in the jth irrigation treatment in the ith replicate; i=1,2 and 3 for re plications; j=control and ir rigated; k=FL and SC; and = population mean, bi = random variable of replication ~ NID (0, ), 2 b Ij = fixed effect of irrigati on (control or irrigated), Fk(j) = fixed effect of family (FL or SC) nested within irrigation, ( IF )k(j) = fixed effect of irrigation x family(irrigation) interaction, ( bI )ij = random variable for replicati on x irrigation interaction ~ NID (0, ), 2 bI ( bF )ik(j) = random variable for replication x family(irrigation) interaction ~ NID (0, ), 2 bF ( bIF ) ik(j) = random variable for replication x irriga tion x family(irrigatio n) interaction ~ NID (0, ), 2 bIF ijk = error term ~ NID (0, ) 2 147

PAGE 148

APPENDIX B ANOVA MODEL FOR ORGAN X SPECIES A NALYSIS FOR LONGLEAF AND SLASH PINE Yijk = + Si + tk(i) + Oj(i) + ijk where Yijk is the parameter value of the jth organ nested in the ith species in the kth tree; i=LL and SL; j=root, Sb, Sc and branch; k=1,2,3 and 4 for trees (replications); and = population mean, Si = fixed effect of species (LL or SL), tk(i) = random effect of tree nested within species. Oj(i) = fixed effect of organ (root, Sb, Sc and branch) nested within species, ijk = error term ~ NID (0, ) 2 For roots and branches k=1,2,3,4,5,6 and 7. 148

PAGE 149

APPENDIX C FOLIAR BIOMASS AND BIOMETRIC VALUES OF MEASURED TREES Table A-1. Biometric values of measured trees. Species Tree DBH (cm) Height (m) Volume (m3) Volume sapwood (m3) BA (m2) SWA_Sb (m2) SWA_Sc (m2) Longleaf 1 20.6 22.76 0. 399 0.191 0.02516 0.01915 0.00664 Longleaf 2 31.5 25.03 0. 955 0.565 0.06312 0.03921 0.01917 Longleaf 3 38.3 23.04 1. 204 0.528 0.08321 0.04328 0.01940 Longleaf 4 40.9 24.87 1. 841 0.723 0.10986 0.05899 0.03541 Slash 1 21.1 22.64 0. 347 0.147 0.02602 0.01815 0.01003 Slash 2 35.0 26.01 1. 151 0.694 0.06290 0.05505 0.02481 Slash 3 41.0 29.50 1. 845 0.802 0.09676 0.06079 0.03089 Slash 4 45.5 30.95 2. 453 1.093 0.12285 0.07532 0.04125 DBH: Diameter with bark at breast height ; SAW_Sb: sapwood area at 1.8 m height; SAW_Sc: sapwood area at crown base; Table A-2. Foliar biomass, sp ecific leaf area an d leaf area of measured trees. Species Tree Needle (new) dry weight (g) Needle (old) dry weight (g) SLA (old) (cm2 g-1) SLA (new) (cm2 g-1) leaf area (old) (m2) leaf area (new) (m2) LA total (m2) Longleaf 1 7376.2 1078.6 119. 29 105.92 24.87 4.10 28.96 Longleaf 2 12941.0 3026.1 105. 17 84.89 34.97 10.13 45.10 Longleaf 3 11352.7 1442.4 133. 41 87.28 31.54 6.13 37.66 Longleaf 4 27269.4 3758.1 119. 29 93.18 80.88 14.27 95.15 Slash 1 9731.6 1856.1 110. 20 113.94 35.29 6.88 42.18 Slash 2 28417.3 7220.1 110. 20 93.75 84.80 25.33 110.13 Slash 3 45963.6 10529.1 114. 68 89.53 130.99 38.43 169.43 Slash 4 30704.2 4705.2 105. 72 83.68 81.78 15.83 97.61 SLA: specific leaf area; LA: leaf area 149

PAGE 150

APPENDIX D PRESSURE ACHAMBER AND PRESSURE-SLEEVE APPARATUS igure D-1. Pressure chamber for vulnerability to embolism measurements. segments. F Figure D-2. Pressure-sleeve apparatus for hydraulic conductivity measurements on trunk 150

PAGE 151

APPENDIX E DAILY LAI FOR LOBLOLLY PINE AT SILVER FOREST igure E-1. Daily LAI relative to yearly maximum for control and irrigated plots. AI expressed relative to maximum LAI. eak day: Control: July 3rd; Irrigated: July 14th. odel: 3rd degree Polynomial Fit: y=a + b*x + c*x2 + d*x3, Table E-1. LAI curve parameters for control and irrigated plots. eter Control Irrigateda 0.685208 0.646364 F L P M where: x= Day of the year Param b 0.005668 0.005861 c -3E-05 -2.9E-05 d 4.08E-08 3.56E-08 Daily LAI0.50 0.60 0.70 0.80 0.90 1.00 0100200300400 Day of the YearLAI (relative to maximum) Control Irrigated 151

PAGE 152

APPENDIX F PARTICLE SIZE DISTRIBUTION AND WATE R RELEASE CURVES FOR t di depth for each replicate. Replicate Depth Sand Silt Clay WRC Model SILVER FOREST fferent Table F-1. Soil particle size di stribution and water re lease curve (WRC) applied a (cm) % % % 1 20 81.56 12.815.631 55 77.50 13.139.381 75 68.13 13.1318.752 115 63.44 10.0026.563 160 69.69 10.0020.314 200 69.69 8.4421.883 2 25 79.06 11.569.381 55 72.81 13.1314.061 75 72.81 11.5615.631 180 75.94 6.8817.194 200 69.69 8.4421.884 3 25 81.56 12.815.631 55 78.44 9.6911.881 65 78.44 9.6911.881 75 62.81 11.2525.942 185 59.69 6.5633.754 200 62.81 9.6927.504 Model: s = 1 c c vba Table F-2. Water release curve parameters. Model a b 1 0.7816 2.1587 11.0782 2 0.9997 0.00207 6899.4 -0.24476901 3 4 2 10.5 2.80 0.74 5 2 44 1.85 7 10.975 152

PAGE 153

LIST ES Abrahamson, W.G. and D.C. Ha rtnett. 1990. Pine flatwoods and dry prairies. In: Ronald L. Myers and John J. Ewel, eds. Ecosystems of Flor ida. Orlando, FL: University of Central Florida R., andDono matal sensitivity to vapor pressure deits rel nshipydraulonductance in Tree Physiol. 24:561-569. Addinton, R.NA. D, R che. Vose, S.Decot. S.B. Jack, U.G. Hacke, J.S. Srry and en. dj Pinus palustris maintain similar stomatal conductance in xeric and meitats. Plan ell Environ. 29:535-545. Albaugh, T.J., H. L. Allen, P. M. Dougherty, and K. H. Johnsen. 2004. Long term growth responses of loblolly pine to optimient ater resouravailability. For. Ecol. Manage. 192:3-19. Allen, H.L., T.R. Fox, and R.G. Campbell. 2005. Wh at is ahead for intensive pine plantation silviculture in the South? South. J. Appl. For. 29:62-69. Baraloto C., F. Morneau, D. Bonal, L. Blanc Ferry. 200Seasonal water stress tolerance and habitat associations within four neotropical tree ge nera. Ecology. 88:478-489. Barne, J.P. and. Sh. 20las status and trends In: Dicke, E.D., arn. Hd an Jokela, edslash Pine: still growing and rowing! Proceedings of the slash pine symposium: 1-6 pp. C. Bongrten, and R. O. Teskey. 1986. Seasonal patterns of net photosynthesis diverse origins. Can. J. For. Res. 16:1063-1068. ter relations of loblolly pine seedlings from diverse ree Physiol. 1:265-276. Bos D.D.o of tance-based soil water probes in coastal plain soils. Vad Zonl. 3 raun P. and J. Schmid. 1999. Sap flow measurements in grapevines ( Vitis vinifera L.) 2. ranier measurements. Plant and Soil. 215: 47-55. urgess S.S.O., J. Pittermann and T.E. Dawson. 2006. Hydraulic efficiency and safety of branch ylem increases with height in Sequoia sempervirens (D. Don) crowns. Plant Cell Environ. 9:229-239. ermk J., J. Ku era and N. Nadezhdina 2004. Sap flow measurements with two thermodynamic methods, flow integration within trees and scaling up from sample trees to entire forest stands. Trees, Structure and Function. 18: 529-546. OF REFERENC Press: 103-149 pp. Addington, R.N., R.J. M itchell Oren L.A. van. 2004. Sto fic and it ati o to h ic c Pinus palustris g ., L. onovan .J. Mit ll, J.M P pe R. Or 2006. A ustments in hydraulic architecture of sic hab t C al nutr a nd w ce and B. 7. tt R.M effield 05. S h pine : characteristics, history, ns J.P. B ett, W.G ubbar d E.L S g Boltz, B. A., Ba of loblolly pine from Bongarten, B.C., and R.O. Teskey. 1986. Wa geographic origins. T ch, e 2004. C mparison ca paci ose Journa 3 :1380-1 89. B G B x 2 153

PAGE 154

ermk J., J. Ku era, W.L. Bauerle kley. 2007. Tree water storage and its diurnal dynamics related to sap flow and changes of stem volume in old-growth Douglasr, P.J. Melcher, M.A. Zweiniecki and N.M. Holbrook. 2005. The spatial pattern of air-seeding th resholds in mature sugar maple trees. Plant Cell Environ. 28:1082-1089. mechanical properties, and density reflecting the fall in strain along the lateral roots rnal. easurement of non uniform sap flow using heat dissipation probes. Tree Physiol. turbed longleaf pine sites. Comstock, G.L. 1970. Directional permeab ility of softwood. Wood Fiber 1:283-289. tors. Ca n. J. For. Res. 18:851-858. for measurement of soil matric potential. Eur. J. oil. Sci. 46:233-238. Carey 2000. Climate-driven changes in biomass allocation pines. Global Change Biology 6:587-593. water storage capacity in bole xylem gments of mature and young D ouglas-fir trees. Trees. 15:204-214. and native embolism in trunks of young and old growth ponderosa pine trees. Plant ell Environ. 128:1103-1113. N. Phillips and T.M. Hinc fir trees. Tree Physiol. 27:181-198. Choat, B., E.C. Lah Christensen-Dalsgaard, K.K., A,R. Ennos and M. Fournier. 2007. Changes in hydraulic conductivity, of two species of tropical trees. J. Exp. Bot. 58:4095-4105. Clark, A., R.F. Daniels, J.H. Miller and H. Ja mes. 2006. Effect of controlling herbaceous and woody competing vegetation on wood quality of plan ted loblolly pine. Forest Products Jou 56: 40-46. Clearwater, M.J., F.C. Meinzer, J.L. Andrade, G. Goldstein and N.M. Holbrook. 1999. Potential errors in m 19:681-687. Cohen, S., R. Braham, and F. Sanchez. 2004. Seed bank viability in dis Restoration Ecology. 12:503-515. Cregg, B.M., P.M. Dougherty and T.C. Hennessey. 1988. Growth and wood quality of young loblolly pine trees in relation to stand density and climate fac Davis S.D., J.S. Sperry and U.G Hacke. 1999. The relationship between xylem conduit diameter and cavitation caused by freezing. Am. J. Bot. 86:1367-1372. Deka, R.N., M. Wairiu, P.W. Mtakwa, C.E. Mullins, E.M. Veenendaal and J. Townend. 1995. Use and accuracy of the filter-paper technique S De Lucia E.H., H. Maherali and E.V. in Domec, J.C. and B.L. Gartner. 2001. Cavita tion and se Domec, J.C., M.L. Pruyn, and B.L. Gartner. 2005. Axial and radial profiles in conductivities water storage C 154

PAGE 155

Domec J.C., Scholz F.G., Bucci S.J., Meinzer F.C., Goldstein G. and Villalobos-Vega R. 2006 Diurnal and seasonal variation in root xylem embolism in tropical savanna woody species: impact on stomatal control of plant w a ter status. Plant Cell Environ. 29:26-35. ies and wel, J.J. 1990. Introduction. In Ronald L. Myers and John J. Ewel, eds. Ecosystems of Florida. ., R. Oren, T.J. Albaugh, and P.M. Dougherty. 1999. Carry-over effects of water and utrient supply on water use of Pinus taeda Ecol. Appl. 9:513-525. atal rom sap flux measurements. Tree Physiol. 20:579-589. se in models. Can. J. For. Res. 31:198-207. and hotosynthesis. Ann. Rev. Plant Physiol. Mol. Biol. 40:503-537. ot biomass in an Eucalyptus globulus planta tion under different wate r and nutrient regimes. isher, H.M. and E.L. Stone. 1990. Air-conducting por osity in slash pine roots from saturated ire, R.J. Mitchell a nd R.O. Teskey. 2004. Assessing variation in the dial profile of sap flux density in Pinus species and its effect on daily water use. Tree Physiol. Ford, C.R., and J.M. Vose. 2007. Tsuga Canadensis (L.) Carr. mortality will impact hydrologic ox, T.R., E.J. Jokela, and H.L. Allen. 2007. The de velopment of pine plan tation silviculture in t rol and hydraulic conductance, w ith special reference to tall ees. Tree Physiol. 24:865-878. Dougherty P.M., D. Whitehead and J.M. Vose. 1994. Environmental influences on the phenology of pine. Ecol. Bull. 43:64-75. Duncan P.D., T.L. White and G.R. Hodge. 1996. Fi rst-year freeze hardiness of pure spec hybrid taxa of Pinus elliottii (Engelman) and Pinus caribaea (Morelet). New Forest. 12:223-241. E Orlando, FL: University of Ce ntral Florida Press. 3-10 pp. Ewers, B.E n Ewers, B.E. and R. Oren. 2000. Analysis of assump tions and errors in the calculation of stom conductance f Ewers, B.E, R. Oren, K.H. Johnsen, and J. J. Landsberg. 2001. Estimating maximum mean canopy stomatal conductance for u Farquhar, G.D., J.R. Ehleringer and K.T. Hubick. 1989. Carbon isotope discrimination p Fabio, A., M. Madeira, E. Steen, T. Ktterer, C. Ribeiro and C. Arajo. 1995. Development of ro Plant and Soil. 168, 215. F soils. Forest Science. 36:18-33. Ford, C.R., M.A. McGu ra 24:241-249. processes in southern Appalachian fore st ecosystems. Ecol. Appl. 17: 1156-1167. F the southern United States. J. For. 105: 337-347. Franks, P.J. 2004. Stomatal con tr 155

PAGE 156

Franks P.J., P.L. Drake and R.H. Froe nd. 2007. Anisohydric but isohydrodynamic: seasonally constant plant water potential gradient explaine d by a stomatal control mechanism incorpora variable plant hydraulic conductan ce. Plant Cell Environ. 30:19-30. ting r rt in Plants (eds M. Zwieniecki & N.M. Holbrook). lsevier/Academic Press, Oxford, UK. pp. 307-318. orphological properties at a Florida flatwoods site. So il Sci. Soc. Am. J. 54:527. 383. In A. Klute et al. (ed.) ethods of soil analysis. Part 1. 2 ed. Agron. Mongr. 9. ASA and SSSA, Madison, WI. namics of fine roots in lantations of slash pine ( Pinus elliottii ) in north Florida. Can. J. For. Res. 16:529-538. P. Jackson, A. Celis. 998. Stem water storage and diurna l patterns of water use in tr opical forest canopy trees. Plant race, J.C. 1987. Theoretical ratio between "one-sid ed" and total surface area for pine needles. ranier, A. 1985. Une nouvelle mthode pour la mesu re du flux de sve brut e dans le tronc des spiration in a Douglas-fir stand by means of sap flow easurements. Tree Physiol. 3:309-320. Measuring and m odeling the transpiration of a maritime pine anopy from sap-flow data. Ag ric. For. Meteorol. 71:61-81. and long-term monitoring usi ng sapflow methods. Globa l Change Biology 2:26574. notypes of blolly pine under dry field conditions. In: Pr oceedings, 26th Biennial SFTIC, Athens, GA. 122acke U.G., J.S. Sperry and J. Pittermann. 2000 b. Drought experience and cavitation resistance acke U.G., Sperry J.S. and Pittermann J. 2004. Analysis of circular bo rdered pit function. II. Gymnosperm tracheids with torus-mar go pit membranes. Am. J. Bot. 91:386-400. Gartner B.L. and F. C. Meinzer. 2005. Stru cture-function relationships in sapwood wate transport and storage. In Vascul ar Transpo E Gaston, L., P. Nkedi-Kizza, G. Sawka, and P.S.C. Rao. 1990. Spatial variability of m Gee, G.W., and J.W. Bauder. 1986. Particle-si ze analysis. pnd M Gholz H.L., L.C. Hendry and W.P. Cropper. 1986 Organic-matter dy p Goldstein G., J.L. Andrade, F.C. Meinzer, N.M. Holbrook, J. Cavelier 1 Cell Environ. 21:397-406. G New Zealand Journal of Fo restry Science. 17:292-296. G arbres. Ann. Sci. For. 42:193-200. Granier, A. 1987. Evaluation of tran m Granier A. and D. Loustau. 1994 c Granier, A., P. Biron, N. Breda, J.-Y. Pontailler, and B. Saugier. 1996. Transpiration of trees and forests: short 2 Grissom, J.E. and S.E. McKeand. 2001. Comparat ive physiology of contrasting ge lo 124 pp. H in six desert shrubs of the Great Basin, Utah. Basic and Applied Ecol. 1:31-41. H 156

PAGE 157

Hatton, T.J., E.A. Catchpole and R.A. Vertessy. 1990. Integration of sapflow velocity to estimate plant water use. Tree Physiol. 6:201-209. Hatton, T.J., S.J. Moore and P.H. Reece. 1995. Estimating stand transpiration in a Eucalyp populnea woodland with the heat pulse method: m easurement errors and sampling strategies. Tree Physiol. 15:219-227. tus oag, M.L. and R.L. Kramer. 1991. Polychromatic x-ray attenuation characteristics and wood ck S.B., MC. P. Sheffield and D.J. McConv ille. 2002. Comparison of growth efficiency of enth echnical eport SRS-48: 81-85 pp. llo. 2003. Axial and radi al water transport and internal water storage tropical forest canopy trees. Oecologia 134: 37-45. dstein. 2002. Heat dissipation sensors of ariable length for the measurement of sap flow in trees with deep sapwood. Tree Physiol. 22: menez, M.S., N. Nedezhdina, J. ermak and D. Morales. 2000. Radial variation in sap flow in hnson, J.D. 1984. A rapid technique for estimating to tal surface area of pine needles. Forest hnson K., C. Maier and L. Kress. 2005. Quantifying root lateral distribu tion and turnover using on, R. and D. Gjerstad. 2006. Restoring the overstory of longleaf pine ecosystems. In: Jose ., E.J. Jokela and D.L. Miller eds. The L ongleaf Pine Ecosystem: Ecology, Silviculture, and rama, K.J.S., S.E. McKeand, J.B. Jett and E.A. Wheeler. 1997. Date of earlywoodtewood transition in provenances and families of loblolly pine, and its relationship to growth yawickrama, K.J.S., S.E. McKeand, and J.B. Jett. 1998. Phenological vari ation in height and Hillel D. 1998. Environmental Soil Physics. Academic Press, San Diego. 771 p. H densitometry applications. Wood Fiber Sci. 23:23-31. Ja mature longleaf and slash pine trees. In: Oucalt, K.W., editor. Proceedings of the Elev Biennial Southern Silvicultural Research Conference. USDA Forest Service General T R James, S.A., F.C. Meinzer, G. Goldstein, D. Woodruff, T. Jones, T. Restom, M. Mejia, M. Clearwater and P. Campane in James, S.A., M.J. Clearwater, F.C. Meinzer, an d G. Gol v 277-283. Ji five laurel forest tree species in Tenerife, Canary Is lands. Tree Physiol. 20:1149-1156. Jo Science. 30: 913-921. Jo pine trees with a distinct stable carbon isotope signature. Funct. Ecol. 19:81-87. Johns S Restoration. Springer, New York, NY: 271-295 pp. Jayawick la phenology and juvenile wood specific grav ity. Can. J. For. Res. 27: 1245-1253. Ja diameter growth in provenances and fam ilies of loblolly pine. New Forests 16:11-25. 157

PAGE 158

Je tt, J.B., McKeand, S.E., and We ir, R.J. 1991. Stability of juve nile wood specific gravity of rdan, L., Schimleck, L.R., Clark, A., Hall, D.B. and Daniels, R.F. 2007. Estimating optimum or. rdan, L., Clark, A., Schimleck, L.R., Hall, D. B. and Daniels, R.F. 2008. Regional variation in es. 38: 698-710. Douglas-fir s eedlings. Tree Physiol. 9:31-38. arson, P.R., Kretschmann, D.E., Clark III, A., Isen-brands J.G. 2001. Juvenile wood formation tment ecoeur, J. and Sinclair, T.R., 1996. Field pea transpiration and leaf growth in response to soil G. Milliken, W. Stroup, R. Wolfinger and O. Schabenberger .2006. SAS system for ixed models. Cary, NC: SAS Institute. 814 pp. dified ranier Measuring System. Aust. J. Plant Physiol. 24:701-705. E.K. Ch acko. 2000. Spatial variations in xylem sap flux density in the unk of orchard-grown, mature mango trees under changing soil water conditions. Tree Physiol. u P., L, Urban and P. Zhao. 2004. Granier's Thermal Dissipation Probre (TDP) method for vity and vulnerability to cavitation of onderosa pine growing in contrasting climates. Tree Physiol. 20:859-867. temperate forest ees. Plant Cell Environ. 29:571-583. est. J. vol 9. 9 pp. loblolly pine in diverse geographi c areas. Can. J. For. Res. 21:1080-1085. Jo sampling size to determine weighted core specific gravity of planted loblolly pine. Can. J. F Res. 37:2242-2249. Jo wood specific gravity of planted loblolly pine in the United States. Can. J. For. R Kavanagh K.L., B.J Bond, S.N. Aitken, B.L. Gartner and S. Knowe. 1999. Shoot and root vulnerability to xylem cavitation in four populat ions of 1 Lambers, H., F.S. Chapin III and T.L. Pons. 1998. Plant Physiological Ecology. Springer-Verlag, New York. 540 p. L and proper-ties in southern pine. Gen. Tech. Rep. FPL-GTR-129, Madison, WI, U.S. Depar of Agriculture, Forest Service, Forest Products Laboratory. 42 pp. L water deficits. Crop Sci. 36, pp. 331. Littell, R., m Lu, P. 1997. A Direct Method for Estimating the Average Sap Flux Density Using a Mo G Lu, P., W.J Muller and tr 20:683-692. L measuring sap flow in trees: theory a nd practice. Acta bot anica sinica. 46:631-646. Maherali H. and E. De Lucia. 2000. Xylem conduc ti p Maherali H., C.F Moura, M.C. Caldeira, C.J. Willson and R.B. Jackson. 2006. Functional coordination between leaf gas exchange and vulnerability to xylem cavitation in tr Marinho A.M. and O.M. Oliveira. 2006. The filter paper method re visited. Geotech. T 3 158

PAGE 159

Martin, T.A., K.J. Brown, J. Cermak, R. Ceulemans, J. Kucera, F.C. Meinz er, J.S. Rombold, .G. Sprugel and T.M. Hinckley. 1997. Crown conduc tance and tree and sta nd transpiration in a Prat, I. Oliveras and J. Pinol. 2002. Xylem hydraulic. properties of roots nd stems of nine Mediterranean woody species. Oecologia. 133:19-29. pacts of fine root turnover on forest N PP and soil C sequestration potential. Science. S., T. Mullin, T. Byram and T. Wh ite. 2003. Deployment of genetically improved blolly and slash. pine in the S outh. Journal of Forestry. 101: 32-37. Brooks, J.C Domec, B.L. Ga rtner, J.L Warren, D. Woodruff, K. Bible and .C. Shaw. 2006. Dynamics of wate r transport and storage in coni fers studied with deuterium nspiration, sap flow and use f stored water in tropical forest canopy trees. Tree Physiol. 24:901-909. in loblolly pine. The influence of tree age, position in ee, and cultural practice on wood specific gravity, fiber length, and fibril angle. Tappi Press, onteith. J.L. 1995. A reinterpreta tion of stomatal responses to humidity. Plant Cell Environ. onteith J.L. and M. Unsworth. 2007. Pr inciples of Environmental Physics, 3rd edn. Academic d P.M. Dougherty. 1997. Estimating foliage area of loblolly pine shoots. For. Sci. 3:299-303. coarse root dii to changes of water supply. In: Stokes A (ed.) The supporting r oots of trees and woody wer, stems D second-growth Abies amabilis forest. Can. J. For. Res. 27:797-808. Martinez-Vilalta J., E a Matamala R., M.A. Gonzalez-Meler, J.D. Jast row, R. Norby and W.H. Schlesinger, 2003. Im 302:1385-1387. McKeand, lo Meinzer F.C., J.R. D and heat tracing techniques. Plant Cell Environ. 29: 105-114. Meinzer F.C., S.A. James and G. Goldstein. 2004. Dynamics of tra o Megraw, R.A. 1985. Wood quality factors tr Atlanta, GA. 88 pp. M 18: 357-364. M Press, Ney York, 440 pp. Murthy, R. an 4 Nadezhdina N. and J. ermk. 2000. Responses of sap flow rate along tree stem and ra plants: form, function and physiolo gy. Developments in plant and soil sciences, vol 87. Klu Dordrecht. 227 pp. Nadezhdina, N., J. ermk and R. Ceulemans. 2002. Radial patterns of sap flow in woody of dominant and understory species: scaling errors associated with posit ioning of sensors. Tree Physiol. 22:907. 159

PAGE 160

N iinemets U., D.S. Ellsworth, A. Lukjanova and M. Tobias. 2001. Site fertility and the l. phori D.U. and B. Maharjan. 2000. First approxima tions of soil moisture retention curves using ren R., B.E. Ewers, P. Todd, N. Phillips and G.Katul. 1998. Water balance delineates the soil rry, C.G. Katul, D.E. Pataki, B.E. Ewers, N. Phillips and K.V.R. Schfer. 1999. urvey and synthesis of intraand interspecific variation in st omatal sensitivity to vapour ren, R. and D.E. Pataki. 2001. Transpiration in response to variation in microclimate and soil tical and statistical analysis of the urves illustrating vulnerability of xy lem to cavitation. Tree Physiol. 18:589-593. n, McGraw-Hill, New ork, 722 pp. management, and restoration. S. Jose, E. Jokela and D. Miller. Eds. pringer, NY. 51-94 pp. R. Oren. 1998. A comparison of two daily representations of canopy conductance ased on two conditional time-averaging methods and the dependence of daily conductance on hillips N., B.J. Bond, N.G McDowell and M. G Ryan. 2002. Canopy and hydraulic conductance hillips, N.G., M.G. Ryan, B.J. Bond, N.G. McDowell, T.M. Hinckley and J. ermk. 2003. rry. 2003. Tracheid diamet er determines the extent of freeze-thaw duced cavitation in conifers. Tree Physiol. 23:907-914. ic efficiency of conifer w ood: the role of tracheid allometry and cavitation rotection. Am. J. Bot. 93:1105-1113. morphological and photosynthetic acclimation of Pinus sylvestris needles to light. Tree Physio 21:1231. O the filter-paper method. H ydrol. processes. 14:1131-1138. O layer in which moisture affects can opy conductance. Ecol. Appl. 8:990-1002. Oren R., J.S. Spe S pressure deficit. Plant Cell Environ. 22:1515-1526. O moisture in southeastern deciduous forests. Oecologia. 127:549-559. Pammenter N.W. and C. Vander Willigen. 1998. A math ema c Panshin A.J. and C. de Zeeuw. 1980. Textbook of wood technology, 4th ed Y Peet, R.K. 2006. Ecological classification of l ongleaf pine woodlands. In: Longleaf pine ecosystems: ecology S Phillips N. and b environmental factors. Ann. des Sci. Forest. 55:217-235. P in young, mature and old Douglas-fir trees. Tree Physiol. 22:205-211. P Reliance on stored water increases with tree size in three species in the Pacific Northwest. Tree Physiol. 23:237-245. Pittermann J. and J.S Spe in Pittermann J., J.S. Sperry, U.G. Hacke, J.K. Wheeler and E. Sikkema. 2006a. Inter-tracheid pitting and the hydraul p 160

PAGE 161

Pittermann J., J.S. Sperry, J.K. Wheeler, U.G. Hacke, E. Sikkema. 2006b. Mechanical reinforcement against tracheid implosion compromi ses the hydraulic efficiency of conifer xylem. Plant Cell Envir on. 29:1618-1628. flatwoods forest north Florida. Can. J. For. Res. 35:1568-1580. arbon st in north Florida. Global Change Biol. 4:2523-2538. and. P. 299-325 (Chapter 3) in Wear, D.N. and J.G. Greis, Editors. Southern Forest Resource Assessment. General olly pine ( Pinus taeda L.). Trees Structure and Function. 15: 154-162. le-tree of Can. J. For. Res. 31: 960-970. 99. Measurement of Fiber Length, oarseness, and Shape with the Fiber Quality Analyzer. TAPPI Journal. 82: 93-98. rsity Press, New York, 36 pp. pine seedlings. Forest Science. 44:308-316. f tensive management. Tree Physiol. 1:369-376. sponse to six years of intensive management. For. Ecol. anage. 192:59-70. Pockman W.T., J.S. Sperry. 2000. Vulnerability to cavitation and the distribution of Sonoran Desert vegetation. Am. J. Bot. 87:1287-1299. Powell, T.L., G. Starr, K.L. Clark, T.A. Martin, and H.L. Gholz. 2005. Ecosystem and understory water and energy exchan ge for a mature, naturally regenerated pine in Powell, T.L., H.L. Gholz, K.L. Clark, G. Starr, W.P. Cropper, Jr., and T.A. Martin. 2008. C exchange of a mature naturally-regenerated pine fore 1 Prestemon, J.P. and R.C. Abt. 2002. Timber products supply and dem 1 Technical Report SRS-53. Ashevill e, NC: U.S. Department of Agriculture Forest Service, Southern Research Station. Retzlaff W.A., Blaisdell G.K. a nd Topa M.A. 2001a. Seasonal changes in water source of four families of lobl Retzlaff W.A., Handest J.A., OMalley D.M., McKeand S.E. and Topa M.A. 2001b. Who biomass and carbon allocation of juve nile trees of loblolly pine ( Pinus taeda L.): influence genetics and fertilization. Ritchie, J.T. 1981. Soil water ava ilability. Plant and Soil. 58:327-338. Robertson, G., J. Olson, P. Allen, B. Chan and R. Seth. 19 C Ruzin, S.E. 1999. Plant microtechnique and micros copy. Oxford Unive 3 Samuelson, L.J. 1998. Influence of intensive cult ure on leaf net photosynth esis and growth of sweetgum and loblolly Samuelson, L.J., T. Stokes, T. Cooksey and P. McLemore III. 2001. Production efficiency o loblolly pine and sweetgum in response to four years of in 2 Samuelson, L.J., K. Johnsen, and T. Stokes. 2004. Production, allocation, and stemwood growth efficiency of Pinus taeda L. stands in re M 161

PAGE 162

Samuelson, L.J. and T.A. Stokes. 2006. Transp iration and canopy stomat al conductance year-old loblolly pine in response to inte nsive management. Forest Science. 52:313-323. of 5amuelson, L.J., T.A. Stokes, and M.D. Colema n. 2007. Influence of irrigation and fertilization amuelson, L.J., J. Butnor, C. Maier, T. St okes, K. Johnsen and M. Kane. 2008. Growth and wth antiago, L.S., G. Goldstein, F.C. Meinzer, F.C. Fisher J., K. Machado, D. Woodrduf and T. chaap M.G. and F.J. Leij. 2000. Improved predicti on of unsaturated hydraul ic conductivity with ameters with hierar chical pedotransfer functions, J. Hydrol. 25163-176. level omatal conductance. Plant Cell Environ. 23:365-377. strips by near infrared spectroscopy. olzforschung. 58:375-381. ierra-Lucero, V., S.E. McKeand, D.A. H uber, D.L. Rockwood, and T.L. White. 2002. in e by environment interaction and deployment considerations for families from Florida rovenances of loblolly pine. Forest Genetics. 10:85-92. re lationship for wood. Gen. ech. Rep. FPL-GTR-76. USDA Forest Servic e, Forest Prod. Lab., Madison, WI. 13 pp. 81-587. S on transpiration and hydraulic properties of Populus deltoides Tree Physiol. 27:765-774. S physiology of loblolly pine in response to l ong-term resource management: defining gro potential in the southern United St ates. Can. J. For. Res. 38:721-732. S Jones. 2004. Leaf photosynthetic traits scale with hydraulic conductivity and wood density in Panamanian forest canopy trees. Oecologia. 140: 543-550. S the Mualem-Van Genuchten model. Soil Sci. Soc. Am. J. 64:843. Schaap M.G., F.J. Leij and M.Th. Van Genuchten. 2001. Rosetta: a computer program for estimating soil hydraulic par 1 Schfer K.V.R., R. Oren and J.D. Tenhunen. 2000. The effect of tree height on crownst Schimleck, L.R., R.D. Jones, G.F. Peter, R. F. Daniels and A. Clark. 2004. Nondestructive estimation of tracheid length from sections of ra dial wood H S Performance differences and genetic parameters for four coastal provenances of loblolly pine the southeastern United States. For. Sci. 48:732-742. Sierra Lucero, V., D.A. Huber, S.E. McK eand, T.L. White, and D.L. Rockwood, D.L. 2003. Genotyp p Simpson, W.T. 1993. Specific gravity, moisture c ontent, and density T Sperry, J.S. and M.T. Tyree. 1988. Mechanism of water-stress induced xylem embolism. Plant Physiol. 88:5 Sperry, J.S., M.T. Tyree and J.R. Donnelly. 1988 Vulnerability of xylem to embolism in a mangrove vs. an inland species of Rhizophor aceae. Physiologia Plantarum. 74:276-283. 162

PAGE 163

Sperry, J.S. and M.T. Tyree. 1990. Water-stressinduced xylem embolism in three species of conifers. Plant Cell Environ.13:427-436. Sperry, J.S. and N.Z. Saliendra. 1994. Intraa nd inter-plant variation in xylem cavitation in Betula occidentalis Plant Cell Environ. 17:1233-1241. Sperry J.S. and T. Ikeda. 1997. Xylem cavitation in roots and stems of Douglas-fir and white f Tree Physiol. 17:275-280. ir. perry, J.S. 2000. Hydraulic constraints on plan t gas exchange. Agr. For. Meteor. 104:13-23. s tatus picer R. and B.L. Gartner. 1998b. Hydr aulic properties of Douglas-fir ( Pseudotsuga menziesii ) picer, R. and B.L. Gartner. 2001. The effects of cambial age and positio n within the stem on Samuelson, a nd B.C. Bongarten. 1986. Stomatal and nonstomatal mitations to net photosynthesis in Pinus taeda L. under different environmental conditions. gherty, and T.C. Hennessey. 1987. hysiology and genetics of tree growth respons e to moisture and temperature stress: an luence of climate and nutrient vailability on net photosynthesis of mature slash pine. Tree Physiol. 14:1215-1227. ater ress in intact juveniles of Quercus rubra and Populus deltoides Tree Physiol. 10:411-415. architecture of trees and other woody plants. ew. Phytol. 119:345-360. dition. Springer, Berlin He idelberg. New York. 283 pp. S Sperry, J.S., U.G. Hacke, R. Oren and J.P. Comstock. 2002. Water defici ts and hydraulic limit to leaf water supply. Plant Cell and Environ. 25:251-263. Sperry J.S., U.G. Hacke and J. Pittermann. 2006. Size and function in conifer tracheids and angiosperm vessels. Am. J. Bot. 93:1490-1500. Spicer R. and B.L. Gartner. 1998a. How does a gy mnosperm branch assume the hydraulic s of a main stem when it takes over as leader? Plant Cell Environ. 21:1063. S branches and branch halves with referen ce to compression wood. Tree Physiol. 18:777-784. S specific conductivity in Dougl as-fir sapwood. Trees. 15:222. Teskey, R.O., J.A. Fites, L.J. li Tree Physiol. 2:131-142. Teskey, R.O., B.C. Bongarten, B.M. Cre gg, P.M. Dou P examination of the characteri stics of loblolly pine ( Pinus taeda L.). Tree Physiol. 3:41-61. Teskey, R.O., H.L. Gholz and W.P. Jr. Cropper. 1994. Inf a Tyree M.T., J. Alexander and J.L. Machado. 1992. Loss of hydraulic conductivity due to w st Tyree M.T. and F.W. Ewers. 1991. The hydraulic N Tyree M.T. and M.H. Zimmermann. 2002. Xylem structure and the ascent of sap. Second E 163

PAGE 164

Tyree, M.T. 2003. Hydraulic limits on tree performa nce: transpiration, carbon gain and growt of trees. Trees. 17:95-100. h in a young mounta in ash forest. Tree hysiol. 15:559-567. .J. Elliott, and B. D. Clinton. 2003. Measuring and modeling tree and and level transpiration. In P hytoremediation: Transformation a nd Control of Contaminants, ed. architect ure and growth of seedlings. rees. 17:269-277. ill, New York. 736 pp. or. Res. 4:940-947. e udies in trees. Tree Physiol. 18:499-512. elocity as a function of stem iameter and sapwood thickness in yellowpoplar trees. Tree Physiol. 20:511-518. apwood and use of a co rrection coefficient. Tree Physiol. 16:697-703. Vertessy R.A, R.G. Benyon, S.K. O'Sullivan and P.R. Gribben. 1995 Relationships between stem diameter, sapwood area, leaf area and transpiration P Vose, J.M., G.J. Harvey, K st S.C. McCutcheon and J.L. Schnoor, 263. New York: Wiley. Wang, T., S.N. Aitken and K.L. Kavanagh. 2003. Selection for improved growth and wood quality in lodgepole pine: eff ects on phenology, hydraulic T White F.M. 1991. Viscous fluid flow. McGraw-H Whitehead, D., W.R.N. Edwards and P.G. Ja rvis. 1984. Conducting sapwood area, foliage area, and permeability in mature trees of Picea sitchensis and Pinus contorta Can. J. F 1 Wullschleger, S.D., F.C. Meinzer and R.A. Ve rtessy. 1998. A review of whole-plant water us st Wullschleger, S.D. and A.W. King. 2000. Radial vari ation in sap v d Zang, D., C.L. Beadle and D.A. White. 1996. Vari ation of sapflow velocity in Eucalyptus globulus with position in s Zwieniecki M.A. and N.M. Holbrook. 2000. Bordered pit structure and vessel wall surface properties implications for embolis m repair. Plant Physiol. 123:1015-1020. 164

PAGE 165

BIOGRAPHICAL SKETCH Carlos Gonzlez Benecke, the middle s on of Blanca Benecke Gallegos and Carlos e stituto de Humanidades Luis Campino, he demonstrated his interest in biology, ecology and t Universidad de Chile, which was the perfect am algam among his scientific interests. At that hattin. Together they grew up and raised a family. After finishing his studies at Un iversidad de Chile he started to work at Forestal Celco S.A. .A., here he worked as project leader of the Nutritional and Silvicultural Management research in the Ecophysiology research project for the same ngineering at the Universidad de Concepcin. ecided to change their life and start a new adventure: Moving to Gainesville, USA, to begin a PhD program at the School of Forest Resources and Conservation at the Un iversity of Florida. The rest is history. Gonzlez Vargas, was born in Santiago, Chile, in 1969. Starting in middle and high school at th In math. When he had to decide what to study at un iversity he had no hesitation: Forest Engineering a time he also met the person who became his partner for the rest of his life: Claudia Costagliola C as a field manager in site productivity studies. Tw o years after that he moved to Bioforest S w project; three years later he star ted to work company. During this period he earned his fi rst graduate degree: Diploma in Industrial E In the Fall of 2004, when everything was going well and smoothly, the GonzlezCostagliola family, now including three wonderful sons, Cristbal, Gabriel and Santiago, d 165