Title: Crown structure, growth performance, nutritional characteristics, and their genetic parameter estimates in juvenile loblolly and slash pine
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00100776/00001
 Material Information
Title: Crown structure, growth performance, nutritional characteristics, and their genetic parameter estimates in juvenile loblolly and slash pine
Physical Description: Book
Language: English
Creator: Xiao, Yu, 1962-
Publisher: University of Florida
Place of Publication: Gainesville Fla
Gainesville, Fla
Publication Date: 2000
Copyright Date: 2000
 Subjects
Subject: Forest Resources and Conservation thesis, Ph. D   ( lcsh )
Dissertations, Academic -- Forest Resources and Conservation -- UF   ( lcsh )
Genre: government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )
 Notes
Summary: ABSTRACT: An understanding of growth, crown structure, nutritional attributes, and their interrelationships can provide valuable information regarding future opportunities for improving forest productivity. This dissertation focused on production ecology, genetics, and nutrition of two important and widely planted commercial timber species in the southeastern United States, loblolly pine and slash pine, as a basis to investigate the interspecific and intraspecific differences in growth strategies. Genetically improved loblolly pine, improved slash pine, and unimproved slash pine were managed under two levels of silvicultural treatments at two locations in north central Florida. Comparisons and contrasts were made at ages 3 and 4 years among the three taxa, while genetic parameters were estimated from 16 loblolly pine and 32 slash pine open-pollinated families to enhance the understanding of genetic architecture of the two species. Loblolly pine trees had more branches, wider crowns, higher amounts of foliage biomass and leaf area, and overall growth than slash pine at both ages, but produced less volume per unit leaf area (804 cm3 wood/m2 leaf area) than improved slash pine (1,106) and unimproved slash pine (1,173). Differences in growth were associated with crown structural and nutritional attributes among taxa.
Summary: ABSTRACT (cont.): Loblolly pine consistently had higher foliage N and P concentrations over the life cycle of a needle cohort, higher N, K, Mg, and Ca use efficiency for leaf area production, higher crown (foliage) nutrient content, and higher nutrient retranslocation efficiency for N, P, and K than slash pine. Narrow-sense heritability estimates for most attributes for the two species were low to moderate. Both species had moderate heritabilities in leaf area (h2 = 0.25 and 0.28, respectively). Loblolly pine had higher heritability (maximum h2 =0.83) for foliar N concentration, but lower heritabilities for foliar Ca and Mg concentrations than slash pine throughout an entire leaf life cycle. Loblolly pine also had higher heritabilities in N and P use efficiency (h2 = 0.41 and 0.27, respectively), but was lower for Ca and Mg use efficiency than slash pine (h2 = 0.32 and 0.26, respectively). Genotype x environment interactions were not important for most traits except those for crown structure in loblolly pine. Genetic and environmental correlations between growth and crown structural attributes in loblolly pine and between growth and nutritional attributes in slash pine were all positive and low to moderate. Results from this study have provided a comparison of growth strategies that can be used to select species suitable for plantation establishment at different locations and management intensities and to evaluate potential traits for tree improvement programs.
Summary: KEYWORDS: crown structure, growth, nutrition, genetic parameters, loblolly pine, slash pine
Thesis: Thesis (Ph. D.)--University of Florida, 2000.
Bibliography: Includes bibliographical references (p. 203-220).
System Details: System requirements: World Wide Web browser and PDF reader.
System Details: Mode of access: World Wide Web.
Statement of Responsibility: by Yu Xiao.
General Note: Title from first page of PDF file.
General Note: Document formatted into pages; contains xv, 221 p.; also contains graphics.
General Note: Vita.
 Record Information
Bibliographic ID: UF00100776
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 50744203
alephbibnum - 002678774
notis - ANE6001

Downloads

This item has the following downloads:

Final_Thesis ( PDF )


Full Text











CROWN STRUCTURE, GROWTH PERFORMANCE, NUTRITIONAL
CHARACTERISTICS, AND THEIR GENETIC PARAMETER ESTIMATES IN
JUVENILE LOBLOLLY AND SLASH PINE














By

YU XIAO













A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2000

































Copyright 2000

by

Yu Xiao















ACKNOWLEDGMENTS

I would like to thank my major advisor and the chairman of my supervisory

committee, Dr. Eric J. Jokela, for his guidance, support, and encouragement throughout my

study and research period, which are crucial to the completion of this study. I am also

grateful to Dr. Timothy L. White for taking much time in discussion of research subjects and

results. Their helpful and constructive comments, along with careful editing, improved this

dissertation. Special thanks go to my other supervisory committee members, Drs. Henry L.

Gholz, Tim A. Martin, and Jerry B. Sartain, for their invaluable guidance throughout this

program. I am also indebted to Dr. Dudley A. Huber for his valuable help in statistical

analyses and enlightening discussion of results.

Special thanks go to Mr. Richard Jeff English, the only person that accompanied me

through all the field data collection and sampling work that spanned more than two years.

His friendship, field working experience, endurance to sweltering weather, insect bites, and

long hours of driving and carrying a 12 foot ladder, coupled with chemical analysis and

laboratory maintenance skills, have largely shortened the physical work phase of this proj ect.

Many thanks are also extended to numerous undergraduate or graduate students from within

and outside the School of Forest Resources and Conservation, who helped me either in the

field, in the laboratory, or both. A long list includes Jennifer Gagnon, Kelly Chinners,

Wayne Hogan, Shawn Stewart, Steve King, Jennifer Hinckly, Yael Teutsch, Annie

Hermansen, and Javier Lopez-Upton.









I would like to thank the Forest Biology Research Cooperative (FBRC), which

provided funding for this project, and the Cooperative Forest Genetics Research Program

(CFGRP), which allowed me to use their established field sites to conduct the experiment.

I would also like to acknowledge the help and convenience given by The Timber Company

and Jefferson-Smurfit Corporation during the field work stage.

Finally, but not least important, I would especially like to thank my wife, Lianmei,

and my son, Siyao, for their support, understanding, love, and endurance to long hours of

boring time during this research work. In particular, I appreciate constant encouragement

from my parents in China, who have always motivated me to study and work more

diligently.















TABLE OF CONTENTS

Page

ACKNOW LEDGM ENTS ........................................ ... iii

LIST OF TABLES .......................................... ......... x

LIST OF FIGURES ................................................. xii

ABSTRACT .............................................. xiv

CHAPTERS

1 INTRODUCTION

Background .................................... ........ 1
Interspecific and Intraspecific Differences in Crown Attributes ....... 2
Nutrient Dynamics and Genetic Variation of Forest Trees ........... 6

2 EARLY GROWTH PERFORMANCE, CROWN STRUCTURE, AND THEIR
RELATIONSHIPS IN JUVENILE LOBLOLLY AND SLASH PINE

Introduction ........ ................................. 12
M materials and M ethods ..................................... 14
Study Sites .............................. ...... . 14
Experimental Design and Treatments .................... 14
Sampling Procedures and Inventory ................. 15
Biomass and Leaf Area Estimation ...................... 17
Statistical Analyses ................. ............... 18
Biomass estimation ............................ 19
Analyses of mensurational and crown structural attributes
. . . . . . . . . . . . . . . . . . 2 1
Analyses of crown biomass and leaf area attributes .... 25
Relationships between volume increment and leaf area
(growth efficiency) ...................... 25
R results ....................................... ....... 26
Biomass Estimation Equations ......................... 26
Selection of Models for ANOVA ....................... 26









Mensurational and Crown Structural Characteristics ......... 28
Biomass and Leaf Area Characteristics of Tree Crowns ...... 32
Specific leaf area ............................ 33
Leaf biomass, branch biomass, and total crown biomass 36
Leaf area characteristics ........................ 37
Vertical Distribution of Total Leaf Area .................. 38
Relationships between Volume Increment and Total Leaf Area per
Tree (Growth Efficiency) ..................... 39

Discussion ............................. ....... ....... 44
Comparisons of Foliage and Branch Biomass Estimation Equations
.......................................... 44
Mensurational and Crown Structural Characteristics ......... 44
Crown Biomass, Leaf Area, and Their Vertical Distribution ... 47
Leaf Area and Growth Efficiency ....................... 49
Sum m ary .......... .................................. 52

3 SEASONAL DYNAMICS OF FOLIAR NUTRIENTS, NUTRIENT USE
EFFICIENCY, AND RETRANSLOCATION IN JUVENILE LOBLOLLY
AND SLASH PINE

Introduction ........ ................................. 54
M materials and M methods ..................................... 57
Plant M materials and Field Sites ......................... 57
Experimental Layout ............................... 58
Sampling Procedures ................................ 58
Nutrient Analyses ................................... 60
M ethod I: Kjeldahl digestion ................... 61
Method II: wet acid digestion .................... 61
Nutritional Variables ................................ 62
Statistical Analyses ......................... ....... 63
R results ...................... ........ ..... ... ........ 64
Variation of Leaf Nutrient Concentrations and Fascicle Weight
. . . . . . . . . . . . . . . . . . . . . 6 4
Crown (Leaf) Nutrient Content ......................... 75
Nutrient Use Efficiency of Leaf Area Production ........... 78
Fascicle Nutrient Retranslocation Efficiency ............ 80
Discussion ................. ......................... 86
Dynamics of Macronutrients in the Foliage Life Cycle ....... 86
Fascicle Nutrient Content and Crown Nutrient Content ...... 87
Nutrient Use Efficiency of Leaf Area Production ........... 90
Nutrient Retranslocation Efficiency .................. .. 91
Sum m ary .................................. ......... 93









4 GENETIC PARAMETER ESTIMATES FOR CROWN STRUCTURAL AND
GROWTH CHARACTERISTICS IN JUVENILE LOBLOLLY AND
SLASH PINE

Introduction ..................................... ..... 95
M materials and M methods ..................................... 98
Experiments and Sampling Description ................. 98
Statistical Analyses ................. .............. 100
Estimation of variance components ............... 100
Estimation of genetic parameters ................. 102
Results and Discussion .................................... 104
Heritability Estimates ............................ 104
Genotype x Environment Interactions ................... 108
Genetic and Environmental Correlation Analyses .......... 110
Sum m ary ........................................... 114

5 GENETIC AND ENVIRONMENTAL CONTROLS ON NUTRITIONAL
CHARACTERISTICS AND CORRELATIONS BETWEEN GROWTH
AND NUTRIENT TRAITS IN LOBLOLLY AND SLASH PINE

Introduction ........ ................................. 117
M materials and M ethods .................................... 120
Field Sites and Experimental Layout ................... 120
Sam pling Procedures ............................... 121
Nutrient Analyses and Variables ....................... 122
Statistical Analyses ................ .............. 124
Data standardization and statistical model .......... 124
Estimation of genetic parameters ................. 126
R results ............................... ................. 128
Variation of Heritability Estimates over An Entire Leaf Life Cycle
......................................... 128
Heritability Estimates for Other Selected Nutrient Attributes 133
Genotype x Environment Interactions ................. 136
Genetic and Environmental Correlations ................. 138
L oblolly pine ............................... 138
Slash pine ................................. 140
Discussion ............................................. 142
Variation of Genetic Control on Mineral Nutrients over A Complete
Foliage Life Cycle ............................ 142
Heritabilities of Nutrient Attributes .................... 144
Genotype x Environment Interactions in Nutritional Attributes
......................................... 146
Genetic and Environmental Correlations Between Growth and
N utrient Traits ............................ 148









Sum m ary ..................................... .. .... 150

6 SUMMARY AND CONCLUSIONS

Interspecific Level Conclusions ............................. 153
Intraspecific Level Conclusions ............................. 156

APPENDICES

A RELATIONSHIPS BETWEEN BRANCH DIAMETER AND PREDICTED
FOLIAGE BIOMASS, AND RESPONSES OF SELECTED CROWN
ATTRIBUTES TO SILVICULTURAL TREATMENTS FOR IMPROVED
LOBLOLLY, IMPROVED SLASH AND UNIMPROVED SLASH PINE
AT TWO LOCATIONS IN NORTH CENTRAL FLORIDA ....... 162

B ANOVA FOR MENSURATIONAL AND CROWN STRUCTURAL
ATTRIBUTES FOR 3- AND 4-YEAR-OLD LOBLOLLY AND SLASH
PINE PLANTED AT TWO LOCATIONS IN NORTH CENTRAL
FL O R ID A ...................................... ... 165

C ANOVAFOR CROWN BIOMASS AND LEAF AREA ATTRIBUTES FOR
3- AND 4-YEAR-OLD LOBLOLLY AND SLASH PINE PLANTED AT
TWO LOCATIONS IN NORTH CENTRAL FLORIDA .......... 167

D ANOVA FOR NUTRIENT CONCENTRATION, NUTRIENT CONTENT
PER FASCICLE, AND AVERAGE FASCICLE WEIGHT IN LOBLOLLY
AND SLASH PINE AT AGE 3 AND 4 YEARS. EIGHT SAMPLING
PERIODS CORRESPONDED TO THE LIFE CYCLE OF THE SAME
NEEDLE COHORT. ALL EXPERIMENTAL TREES WERE
SUBJECTED TO TWO LEVELS OF SILVICULTURAL TREATMENTS
AND PLANTED AT TWO LOCATIONS IN NORTH CENTRAL
FL O R ID A ...................................... ... 171

E TAXA MEANS BY EXPERIMENTAL LOCATION, SILVICULTURAL
TREATMENT AND SAMPLING PERIOD FOR FASCICLE NUTRIENT
CONCENTRATION, CONTENT, AND WEIGHT IN LOBLOLLY AND
SLASH PINE IN NORTH CENTRAL FLORIDA ............... 183

F DYNAMICS IN FASCICLE NUTRIENT CONTENT OVER THE
COURSE OF A NEEDLE COHORT LIFE CYCLE FOR IMPROVED
LOBLOLLY, IMPROVED SLASH AND UNIMPROVED SLASH PINE
FROM AGES 3 TO AGE 4 YEARS. ALL INDIVIDUAL TREES WERE
SUBJECTED TO TWO LEVELS OF SILVICULTURAL TREATMENTS
AND PLANTED AT TWO LOCATIONS IN NORTH CENTRAL









FL O R ID A ......................................... . 192

G FAMILY x SITE (rB-site), AND FAMILY x TREATMENT
INTERACTIONS (rBtreat) FOR AVERAGE FASCICLE WEIGHT,
NUTRIENT CONCENTRATIONS AND NUTRIENT CONTENT OVER
AN ENTIRE LIFE CYCLE OF A NEEDLE COHORT FROM AGES 3 TO
4 YEARS IN LOBLOLLY AND SLASH PINE PLANTED AT TWO
LOCATIONS IN NORTH CENTRAL FLORIDA ............... 198

REFERENCES ........................................... ........ 203

BIOGRAPHICAL SKETCH .................................... ..... 221















LIST OF TABLES


Table Page

S 2-1: Geographic locations, climatic and site conditions of the two experimental sites
in north central Florida .............. ................... ....... 15

* 2-2: Treatment regimes for intensively-managed and non-intensively-managed
blocks at the two research locations ................................. 16

S 2-3: Model parameter estimates and summary statistics for estimating foliage and
branch biomass in loblolly and slash pine ............................. 27

* 2-4: Individual tree growth and crown characteristics for 3- and 4-year-old loblolly
and slash pine planted at two locations in north central Florida ............ 30

* 2-5: Specific leaf area (SLA), leaf biomass, branch biomass, and leaf area for 3- and
4-year-old loblolly and slash pine planted at two locations in north central Florida
.......................................................... 34

* 2-6: A comparison (r2) between estimation models for foliage and branch biomass
based on branch diameter alone and multi-factor variables ............... 45

S 2-7: Variation ofleaf:branch biomass ratio as influenced by location, treatment, and
tree age in loblolly and slash pine in north central Florida ............... 48

* 3-1: ANOVA for crown nutrient content (g/tree) of individual trees of loblolly and
slash pine at age 3 years Experimental trees were subjected to two levels of
silvicultural treatments and planted at two locations in north central Florida ... 76

* 3-2: Nutrient content (g/tree) in the crowns (foliage) of 3-year-old loblolly and slash
pine managed under two silvicultural treatments and planted at two locations in north
central Florida ................................... ........... 77

* 3-3: ANOVA for LA, (cm2 leaf area / mmol element) for loblolly and slash pine at
age 3 years. Experimental trees were subjected to two levels of silvicultural
treatments and planted at two locations in north central Florida ............ 79









3-4: ANOVA for nutrient retranslocation efficiency (%) and amount retranslocated
(mg/fascicle) prior to senescence for a single cohort of needles in loblolly and slash
pine at ages 3 to 4 years. Experimental trees were subjected to two levels of
silvicultural treatments and planted at two locations in north central Florida ... 84

4-1: Narrow sense heritability (h2) estimates and standard errors for growth and
crown structural attributes in 3-year-old loblolly and slash pine planted at two
locations in north central Florida ................................ 106

4-2: Family x site interaction (rBsite), and family x treatment interaction (rBtreat) for
growth and crown structural attributes in 3-year-old loblolly and slash pine planted
at two locations in north central Florida ............................. 109

4-3: Estimates of genetic (upper triangle) and environmental (lower triangle)
correlations among growth and crown structural attributes in 3-year-old loblolly
pine planted at two locations in north central Florida ................... 111

4-4: Estimates of genetic (upper triangle) and environmental (lower triangle)
correlations among growth and crown structural attributes in 3-year-old slash pine
planted at two locations in north central Florida ....................... 113

5-1: Narrow sense heritability (h2) estimates and standard errors for crown nutrient
attributes in 3-year-old loblolly and slash pine planted at two locations in north
central Florida ................................... .......... 135

5-2: Family x site interaction (rBsite), and family x treatment interaction (rBtreat) for
crown nutrient attributes in 3-year-old loblolly and slash pine planted at two
locations in north central Florida .................................. 137

5-3: Estimates of genetic (upper triangle) and environmental (lower triangle)
correlations among growth and nutrient attributes in 3-year-old loblolly pine planted
at two locations in north central Florida ............................. 139

5-4: Estimates of genetic (upper triangle) and environmental (lower triangle)
correlations among growth and nutrient attributes in 3-year-old slash pine planted
at two locations in north central Florida ............................. 141

6-1: A comparison of growth characteristics in crown structure and nutritional
attributes for loblolly and slash pine at ages 3 and 4 years old ............ 160















LIST OF FIGURES


Figure Page

* 2-1: Vertical distribution of total leaf area (right) and branch biomass (left) by
crown positions at age 3 years for loblolly and slash pine managed under two
silvicultural treatm ent regim es ..................................... 40

* 2-2: Relationship between individual tree volume increment for ages 3 to 4 years
and total leaf area (all-sided) for loblolly and slash pine managed under two
silvicultural treatment regimes in north central Florida .................. 43

* 3-1: Variation in needle N concentration for improved loblolly, improved slash and
unimproved slash pine throughout a life cycle of a needle cohort managed under two
silvicultural treatments and two locations in north central Florida .......... 65

* 3-2: Variation in needle P concentration for improved loblolly, improved slash and
unimproved slash pine throughout a life cycle of a needle cohort managed under two
silvicultural treatments and two locations in north central Florida .......... 66

* 3-3: Variation in needle K concentration for improved loblolly, improved slash and
unimproved slash pine throughout a life cycle of a needle cohort managed under two
silvicultural treatments and two locations in north central Florida .......... 67

* 3-4: Variation in needle Ca concentration for improved loblolly, improved slash and
unimproved slash pine throughout a life cycle of a needle cohort managed under two
silvicultural treatments and two locations in north central Florida .......... 68

* 3-5: Variation in needle Mg concentration for improved loblolly, improved slash
and unimproved slash pine throughout a life cycle of a needle cohort managed under
two silvicultural treatments and two locations in north central Florida ....... 69

* 3-6: Variation in average fascicle weight for improved loblolly, improved slash and
unimproved slash pine throughout a life cycle of a needle cohort managed under
two silvicultural treatments and two locations in north central Florida ....... 73

* 3-7: Relationships between total leaf are at age 3 years and September Mg
concentration for genetically improved loblolly pine (PTA), improved slash (PEE)
and unimproved slash pine (PEU) when managed under two silvicultural treatments









at two locations in north central Florida. Symbols with higher Mg concentrations
(i.e., lower leaf area) were from the non-intensively managed treatment, while those
with lower Mg concentrations (i.e., higher leaf area) were from the intensively-
managed treatment. The horizontal dashed lines represent critical (minimum) foliar
Mg concentrations for loblolly pine (0.07%) and slash pine (0.05%) ........ 74

3-8: Nutrient use efficiency for leaf area development for genetically improved
loblolly pine (PTA), improved slash (PEE) and unimproved slash pine (PEU) when
managed under two silvicultural treatments at two locations in north central Florida.
Means among taxa for a given nutrient followed by the same letter were not
statistically significant at the 95% confidence level using the LSMEANS test of the
MIXED procedure. Note the different scales between N and other elements . 81

3-9: Nutrient retranslocation efficiency from fascicles of a single needle cohort prior
to senescence for genetically improved loblolly pine (PTA), improved slash (PEE)
and unimproved slash pine (PEU) when managed under two silvicultural treatments
at two locations in north central Florida. Means among taxa for a given nutrient
followed by the same letter were not statistically significant at the 95% confidence
level using the LSMEANS test of the MIXED procedure ................ 83

5-1: Narrow sense heritability for average fascicle weight over an entire lifespan of
a needle cohort in loblolly and slash pine planted at two locations in north central
Florida ................. ................. ......... ....... 129

5-2: Narrow sense heritabilities for macronutrient concentrations and fascicle
nutrient content over an entire lifespan of a needle cohort in loblolly and slash pine
planted at two locations in north central Florida ....................... 130















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

CROWN STRUCTURE, GROWTH PERFORMANCE, NUTRITIONAL
CHARACTERISTICS, AND THEIR GENETIC PARAMETER ESTIMATES IN
JUVENILE LOBLOLLY AND SLASH PINE

By

Yu Xiao

December 2000

Chairperson: Eric J. Jokela
Major Department: Forest Resources and Conservation

An understanding of growth, crown structure, nutritional attributes, and their

interrelationships can provide valuable information regarding future opportunities for

improving forest productivity. This dissertation focused on production ecology, genetics,

and nutrition of two important and widely planted commercial timber species in the

southeastern United States, loblolly pine and slash pine, as a basis to investigate the

interspecific and intraspecific differences in growth strategies. Genetically improved

loblolly pine, improved slash pine, and unimproved slash pine were managed under two

levels of silvicultural treatments at two locations in north central Florida. Comparisons and

contrasts were made at ages 3 and 4 years among the three taxa, while genetic parameters

were estimated from 16 loblolly pine and 32 slash pine open-pollinated families to enhance

the understanding of genetic architecture of the two species.









Loblolly pine trees had more branches, wider crowns, higher amounts of foliage

biomass and leaf area, and overall growth than slash pine at both ages, but produced less

volume per unit leaf area (804 cm3 wood/m2 leaf area) than improved slash pine (1,106) and

unimproved slash pine (1,173). Differences in growth were associated with crown structural

and nutritional attributes among taxa. Loblolly pine consistently had higher foliage N and

P concentrations over the life cycle of a needle cohort, higher N, K, Mg, and Ca use

efficiency for leaf area production, higher crown (foliage) nutrient content, and higher

nutrient retranslocation efficiency for N, P, and K than slash pine.

Narrow-sense heritability estimates for most attributes for the two species were low

to moderate. Both species had moderate heritabilities in leaf area (h2 = 0.25 and 0.28,

respectively). Loblolly pine had higher heritability (maximum h2 =0.83) for foliar N

concentration, but lower heritabilities for foliar Ca and Mg concentrations than slash pine

throughout an entire leaf life cycle. Loblolly pine also had higher heritabilities in N and P

use efficiency (loblolly pine h2 = 0.41 and 0.27, respectively), but was lower for Ca and Mg

use efficiency than slash pine (slash pine h2 = 0.32 and 0.26, respectively). Genotype x

environment interactions were not important for most traits except those for crown structure

in loblolly pine. Genetic and environmental correlations between growth and crown

structural attributes in loblolly pine and between growth and nutritional attributes in slash

pine were all positive and low to moderate. Results from this study have provided a

comparison of growth strategies that can be used to select species suitable for plantation

establishment at different locations and management intensities and to evaluate potential

traits for tree improvement programs.















CHAPTER 1
INTRODUCTION

Background

With increasing societal demands for timber products and decreasing access to older

timber, modern forestry has begun shifting its harvest to short-rotation plantations. To

improve the efficiency of management systems, a better understanding of genetic

characteristics and growth strategies of forest trees is necessary. Such information will

prove useful in the development of advanced protocols for improving yield and timber

quality. Selection of genetically superior trees, efficient utilization of fertilizers, and

reduction of competing understory and pests have significantly contributed to the

improvement of forest yields. Nevertheless, like many related fields in agriculture, the

practical success in forestry far exceeds the theoretical progress. Therefore, an

understanding of the mechanisms controlling tree growth and its adaptation to the

environmental complexity is critical not only for the future advancement of forestry, but also

for the conservation of forest resources and the protection of environmental quality.

Forests are primarily composed of woody plants that vary in size. To increase the

productivity of forest stands steadily, the mechanisms involved in stand growth must be fully

examined. A better understanding of stand growth as influenced by biological factors and

the surrounding environment is essential for continued improvement in growth rates. Many

studies have been conducted to address the various issues related to growth, in which close









2

attention has long been paid to the growth characteristics of forest trees and their interaction

with the environment for (1) improving the yield and quality of timber and bettering the

resistance of trees to pathogens and insects (Ross and Berisford 1990; Ross et al. 1990); (2)

increasing the knowledge on general linkages between biological processes and

environmental influences (Gholz et al. 1994); and (3) clarifying the growth dynamics of

stands by examining the performance of single trees, especially during the seedling and

sapling growth stage (Kinerson et al. 1974; Madgwick 1983; Ceulemans et al. 1990; Gower

et al. 1993).

In comparison with studies conducted on growth at the stand level, few attempts have

been made to comprehensively investigate relationships among growth, nutritional

physiology, and genetics in individual trees of a species. In some studies, although

individual trees were sampled, analyzed, and scaled to the stand level to estimate overall

productivity, the characteristics of the sampled individuals and their relations to one another

were mostly unknown (Forrest and Ovington 1971; Kira 1975). One of the difficulties in

conducting research at the single tree level is related to selecting representative trees from

a spectrum of well-recognized good and poor families of a species and to implement them

experimentally in the field while maintaining the uniqueness of family structures.


Interspecific and Intraspecific Differences in Crown Attributes

Canopy structure is one of the most important factors affecting stand growth. A

well-developed canopy can efficiently intercept solar energy. Consequently, crown

architecture is closely related to forest productivity (Cannell et al. 1987; Wang and Jarvis

1990; Dalla-Tea and Jokela 1991; McCrady and Jokela 1996). For many annual crops,









3

certain morphological traits have been successfully incorporated into breeding programs

since the ideotype concept was first proposed by Donald (1968). For example, new and high

yield cultivars of wheat (Triticum aestivum), barley (Hordeum vulgare), maize (Zea mays),

bean (Phaseolus vulgaris) and the others have been bred using ideotype selection techniques

(Donald 1968; Mock and Pearce 1975; Adams et al. 1986; Rasmusson 1987). The breeding

of crop ideotypes usually assumes significant genetic relationships between crop yield and

morphological or physiological traits used as indirect selection criteria (Fakorede and Mock

1978).

Three kinds of ideotypes have been proposed: isolation ideotypes, competition

ideotypes, and crop ideotypes (Donald and Hamblin 1976), which could also be applied to

forest trees. Isolation ideotypes are expressed in free standing trees that can exploit their

surroundings to almost the fullest extent. Such trees have tall, dense and well-developed

crowns that spread foliage over a broad area both horizontally and vertically. They grow

best when spacing is wide, and will be nonetheless strong competitors if they appear in a

forest. Competition ideotypes tend to exceed neighboring trees in height growth and

intercept more light at their neighbors' expense. Stands comprised of this ideotype soon

differentiate in crown and stem diameter classes (e.g., dominant, codominant, intermediate,

and suppressed trees). Although the dry-matter production of individual trees is higher, the

mortality rate of stands is also high, thus making the stand biomass production rates low.

Crop ideotypes are individuals that are not strong competitors, and can make efficient use

of the limited site resources to which they have access. Such trees have dense and narrow

crowns, and produce stands with a minimal differentiation in stem diameters. The biomass









4

production rates are potentially high. Clearly, crop ideotypes are ideal for intensively

managed production systems (Donald and Hamblin 1976; Dickmann 1985).

Tree crowns are much more complicated than canopies of annual crops; therefore,

it is much more difficult to establish accurate relations between stand productivity and crown

structures (Dickmann 1985). Tree canopies not only include structural traits, but also

involve phenological characters that can shift between years. Some studies have

demonstrated that trees having long narrow crowns (i.e., a higher crown length/crown width

ratio) with high leaf area, and relatively few but thin, short branches that are borne at acute

angles, will produce higher yields and have higher stem wood allocation percentages. For

example, a genotype ofPinus sylvestris with a narrow-crowned habit had higher production

efficiency, and was likely determined by a single, dominant gene, as shown by the

segregation in F, and F2 progenies (Karki and Tigerstedt 1985).

Some crown traits, such as crown width, branch angle and numbers of branches per

whorl, were found to have high heritabilities and significant genetic correlations in Pinus

sylvestris and Populus clones (Karki and Tigerstedt 1985; Ceulemans et al. 1990). These

crown traits may be greatly influenced by environmental factors such as density and

competition, and the relations become less useful as the stands age. Structural

characteristics, such as number of branches, number of clusters, and branch basal areas, were

highly correlated with tree height and diameter at breast height (Forrest and Ovington 1971;

Madgwick 1983).

In a series of studies on Douglas-fir (Pseudotsuga menziesii var. menziesii), it was

found that equations predicting component biomass and leaf area differed among open-









5

pollinated families (St. Clair 1993). Narrow-sense heritability estimates were high for

biomass components, several biomass partitioning ratios, and stem growth increment. These

estimates could be used to select families with favorable partitioning to the stem wood

component and, thus, improve stand productivity (St. Clair 1994a). Further, substantial

genetic variation was detected for some crown traits such as relative crown size, branch

diameter and length, needle size and leaf area.

All of these studies have suggested that it is promising to select crop ideotypes with

high yield traits which have high heritabilities that favor more biomass accumulation in pure

and closed stands. Moreover, the ideal crown structure may change with environmental

conditions and management intensities, or even stages of stand development. Therefore, an

understanding of ideotype and environment interactions is a prerequisite for combining the

ideotype concept into tree breeding programs. With this knowledge, we can recognize

whether crown structure is an adaptive strategy to environments and cultural treatments, or

originates from genetic or physiological controls, or alternatively, from genotypex

environment interaction.

Soil nutrient supply is also important in determining biomass allocation patterns

among different tissue components. More biomass was significantly allocated to roots at

high irradiances and low nitrate supplies, causing a lower leaf area ratio and leaf mass ratio

in Mycelis muralis (Clabby and Osborne 1997). The regulation of canopy nutrients in the

production efficiency (stem wood production / unit leaf area) was also determined for some

hardwood forest communities (Jose and Gillespie 1996). Canopy nutrient contents showed

a strong correlation with production efficiency on a unit leaf area basis rather than on unit









6

leaf mass basis. Further, the authors found that specific leaf area was negatively correlated

with the canopy nutrient content per unit leaf area, in contrast to some other studies reported

above. Because the study was conducted along a moisture gradient, water was also involved

in the relations between nutrient content and specific leaf area. Although all of the above

studies have demonstrated that leaf area and nutrients were closely related to growth, few

studies have considered the importance of genetics and the interaction between genetics and

nutrition on these traits for improving overall growth performance of trees at the

intraspecific level.


Nutrient Dynamics and Genetic Variation of Forest Trees

Nutrition of forest trees has been extensively studied in relation to the physiological,

ecological and silvicultural aspects influencing the enhancement of growth. Most studies

pertinent to the genetic aspects of tree nutrition have only a recent origin, however. In

contrast, genetic screening to detect nutritional deficiencies and factors related to abiotic

stress, and practices to breed low input cultivars and nutrient use efficient cultivars, have

achieved great success in many agronomic plants (Sari* 1981; 1983; Gabelman and

Loughman 1987; Bassam et al. 1990). At the intraspecific level, the objective is often to

characterize the influence of ancestry on the nutritional status of progeny plants (Rosen and

Luby 1987), with the superior parental materials being utilized in the breeding program.

However, tree improvement programs have generally paid less attention to genetic

differences in nutritional attributes, and have concentrated more on factors relevant to

growth, form and pest resistance (Zobel and Talbert 1984). The research in this area has

lacked direction, depth and specific goals (Nambiar 1984), although certain outcomes can









7

be found in the genus Pinus or Populus for intensively managed plantations (Forrest and

Ovington 1971; Ceulemans et al. 1990; Li et al. 1991a). To meet the potential needs of tree

improvement programs in the future, the amount, cause, and nature of the variation in

nutritional traits must be better understood.

About 50% of the yield increase in agriculture in the last few decades has been due

to the utilization of fertilizers (Sari* 1987) and, thus, the response functions of different

crops to nutrients have received widespread attention. Extensive studies have revealed that

N, K, and Ca are under strong genetic control; however, exceptions can also be found. For

example, P was only genetically controlled in some crop species (Sari* 1987). In several

wheat cultivars, P accumulation, translocation and utilization efficiency were contingent on

the genotype in relation to environmental conditions (Papakosta 1994). Differences were

also detected for sorghum (Sorghum bicolor) hybrids in response to P levels, with largest

differences appearing at low soil P levels (Furlani et al. 1987). Other crops, which included

barley and oats (Avena sativa), have also varied among cultivars in their grain accumulation

of N, P, Ca, Zn, Mn, and Co (Nambiar 1976). These findings suggested that genetic

improvement of nutritional traits could lead to the more efficient use of nutrients by crops,

which may decrease investments in fertilizer application.

With forest trees, several studies have previously quantified genetic variation in

nutritional traits. Full-sib and half-sib families, open-pollinated families from seed origin,

and clones from vegetative propagation have been the most common experimental materials

used in this research. For example, variability among and within a series of full-sib loblolly

pine seedlings from intra-provenance and inter-provenance crosses was found to differ by









8

genotype in nutrient content (P, K, Ca, Mg) of aboveground components, especially in their

ability to absorb Ca and Mg (Woessner et al. 1975). Under three levels of nitrogen fertility,

genotype x environment interaction was examined in 40 full-sib black spruce (Picea

mariana) families in greenhouse conditions (Mullin 1985; Mullin and Park 1994).

Significant family x nitrogen interactions were found. For other nutrient characteristics,

such as nutrient use efficiency, significant variation was detected among 23 open-pollinated

loblolly pine families grown under two levels of nitrogen treatment (Li et al. 1991b).

Narrow-sense heritability estimates for nutrient use efficiency were 0.84 and 0.69,

respectively, under the low and high N levels, suggesting that this trait was under strong

genetic control.

Several studies, using clones as experimental materials, have also shown that some

nutritional traits are under strong genetic control. Forrest and Ovington (1971) reported

large differences in foliar nutrient levels (P, Ca, K, Mg, Mn, and Zn) among six clones of

radiata pine (Pinus radiata). Broad-sense heritabilities among radiata pine clones for foliar

nutrients were higher for K, Mg and Ca than those for N, P, B, Mn, Zn, and Cu (Beets and

Jokela 1994). The authors inferred that foliar nutrient levels were controlled by genetic

factors, and that nutritional differences were genotype specific. Raupach and Nicholls

(1982) observed that few nutrients (N, K, Mg, Zn) were significantly different among radiata

pine clones in their study. For nutrient use efficiency (amount of dry weight produced per

unit weight of nutrients absorbed), Sheppard and Cannell (1985) found 10 30% differences

among 8-year-old clones of Picea sitchensis and Pinus contorta. These differences were

closely related to the nutrient concentration of foliage, and an ideotype for high nutrient use









9

efficiency was proposed as trees having an inherently low nutrient concentrations in needles.

Such trees might be well-suited to grow on nutrient poor sites.

Genotype (e.g., species, clones or families) x environment interactions will occur

whenever genotypes do not achieve consistent performance across a range of environmental

conditions. Although many studies have detected significant genotype x environment

interactions in the growth characteristics of trees (Sato 1994; Ronnbergwastljung et al. 1994;

Isik et al. 1995; Khasa et al. 1995; Johnsen and Major 1995), opposite results have also been

reported (Danj on 1995). However, nearly all studies that have detected significant genotype

x environment interactions have provided little further information on the underlying

environmental factors causing these interactions. Such information is critical for

maximizing gains from genetic selection trials (Jackson et al. 1995).

Saulescu and Kronstad (1995) designed a simple approach to describe the specificity

of each genotype's response to environmental factors. Environmental variables (e.g., water

deficit, minimum temperature of the winter) were directly computed or derived from a

simulation model. Simple correlation coefficients between deviations of each genotype from

a check (actual yield or simulated yield) and environmental indices were then calculated to

describe relations between environments and the performance of each genotype. When

genotype x nutrition interactions were found, their patterns often differed from genotype x

environment interactions for growth characteristics because significant variation in nutrient

traits occurred within a year. New and efficient approaches to deal with nutritional

specificity of genotypes under different environmental conditions have not yet been well

developed.









10

Silvicultural treatments can significantly influence growth performance of forest

trees, primarily by improving the nutrient conditions for growth. Colbert et al. (1990)

reported that fertilizer treatments produced almost the same effects on aboveground biomass

production as weed control treatments in juvenile loblolly pine (Pinus taeda L.) and slash

pine (P. elliottii Engelm. var. elliottii) plantations. In a study ofradiata pine from age 6 to

11 years, Fife and Nambiar (1995) reported that nitrogen fertilization increased foliar

nitrogen concentration and significantly affected two physiological indices, predawn foliage

water potential and water stress integral (an index of cumulative water stress over time).

However, the two indices were not significantly influenced by family or family x nitrogen

interactions. Schmidtling (1995) found that when fertilizer rates increased, foliar Mn and

B concentrations also increased, while those of Mg and Zn decreased, and other foliar

macro- or micro-nutrients were almost not affected in loblolly pine ramets. The effects of

genotype are incomparable among most of these studies because of different plantation ages.

More important, the impact of silvicultural treatments on nutritional characteristics and

nutrient interactions of different families within a species is still unknown, let alone the

response features of families in growth to the treatments. If nutrient traits are to be utilized

as selection criteria in tree breeding programs, genetic heritabilities of each mineral nutrient

must first be identified. Genotype x nutrition effects and genotype x fertilizer interactions

should also be taken into consideration and clearly understood for those nutrients with high

heritabilities under field conditions.

This dissertation concentrates on the genetics, nutrition and production ecology of

loblolly and slash pine, two commercially important and widely planted pine species in the









11

southeastern United States. The focus is on (1) evaluating the relations among growth

strategies, crown structure, and nutrient characteristics; (2) examining patterns of genetic

interaction of each taxon with growth, crown structure, foliar chemistry and the extent to

which they are subjected to genetic control and environmental influence; and (3) comparing

growth strategies of individual trees on the basis of families or species. Expected results will

positively impact future tree growth modeling and in refining of management prescriptions

that involve genotype deployment and silvicultural treatments. The results will also aid in

examining genetic and environmental controls on several biological characteristics of trees

as a basis for understanding growth strategies and the nutritional physiology of genetically

improved loblolly, improved and unimproved slash pine.















CHAPTER 2
EARLY GROWTH PERFORMANCE, CROWN STRUCTURE, AND THEIR
RELATIONSHIPS IN JUVENILE LOBLOLLY AND SLASH PINE

Introduction

Crown structure represents an important factor affecting individual tree and stand

level growth (Forrest and Ovington 1971; Madgwick 1983; Cannell et al. 1987; Dalla-Tea

and Jokela 1991). Many factors, such as inclination and orientation of leaves and geometric

properties of twigs and branches, can contribute to variation in crown characteristics and

growth performance (Dickmann 1985; Wang and Jarvis 1990). Previous research has

suggested that one of the most important factors influencing growth is the amount and

distribution of leaf area, as it affects the interception of photosynthetically active radiation

(Stenberg et al. 1994; Vose et al. 1994; McCrady and Jokela 1996, 1998). Trees that have

long narrow crowns (i.e., a higher crown length/crown width ratio) with high leaf area, and

relatively few but thin, short branches borne at acute angles have been reported to produce

both high yields and stem wood allocation percentages (Karki and Tigerstedt 1985). The

growth "efficiency" (stem wood production/leaf area) of these trees may be high because

they maintain a large crown surface area per unit of growing space (Ford 1985). It follows

that as crown width increases, stem wood growth efficiency may decline because the central

portion of the crown becomes dominated by supporting branches that produce little

photosynthate relative to growth and maintenance respiration demands.









13

In the southeastern United States, the two most important and widely planted

commercial species are loblolly (Pinus taeda L.) and slash pine (P. elliottii Engelm. var.

elliottii). Both species in this region are commonly managed under a regime of intensive

silvicultural practices that include mechanical and chemical site preparation (Shiver et al.

1990), woody and herbaceous competition control (Miller et al. 1991), genetic tree

improvement and fertilization (Neary et al. 1990; Jokela et al. 2000). From a management

perspective, species deployment decisions are most often based on estimates of potential site

productivity and value accrued at the end of the rotation.

Few comparative studies exist with southern pines that examine species variation in

crown structure in relation to growth performance for a range of silvicultural treatments and

site types. In one study, loblolly pine demonstrated greater sensitivity than slash pine to

fertilizer applications, especially in allocating more carbon to branches and foliage (Jokela

and Martin 2000). The crown structure of loblolly pine facilitated greater retention of leaf

area than slash pine on those plots receiving fertilizer additions. Establishing a more

thorough understanding of the relationships between crown structure and growth efficiency,

especially at the interspecific level, will be essential for improving our understanding of

growth strategies, development of crop ideotypes and species-site deployment decisions.

The current study utilized two genetics experiments to (1) determine the magnitude

of the effects of silvicultural treatments, locations, taxa (genetically improved loblolly pine,

improved slash pine, and unimproved slash pine) and their interactions on crown structural

characteristics and overall growth performance; (2) ascertain whether significant differences

in crown attributes, especially the vertical distribution of leaf area, existed among pine taxa









14

when managed under different silvicultural treatments and site locations; and (3) clarify

whether the general relationship between stem wood production and leaf area (growth

efficiency) varied among pine taxa and silvicultural treatments.


Materials and Methods

Study Sites

Two locations in north central Florida (Dunnellon in Levy County, and Palatka in

Putnam County) were chosen as the experiment sites (Table 2-1). The two sites are part of

a larger series of eleven experiments being conducted by the University of Florida's

Cooperative Forest Genetics Research Program for genetically testing several pine taxa and

their hybrids (Lopez-Upton 1999). Genetically improved loblolly pine, unimproved slash

pine, and improved slash pine were selected as experimental materials. Sixteen open-

pollinated half-sib families for each of the three taxa were planted across the two sites.

Climatic conditions between locations were similar, but the soil types did differ (Table 2-1).

The soils at Dunnellon were classified as the Smyrna series (sandy, siliceous, hyperthermic

Aeric Alaquods), while the Adamsville series (hyperthermic, uncoated Aquic

Quartzipsamments) was dominant at Palatka (Soil Survey Staff 1998).


Experimental Design and Treatments

Within each field site, the experimental design was a randomized complete block

split-split plot design, with three complete blocks within each of two silvicultural treatments

(intensively-managed treatment, including fertilizer, insecticide, and herbicide utilization;

and non- intensively-managed treatment). Each taxon was randomly assigned in each block,









15

and sixteen families were nested within each taxon plot, with five seedlings being planted

per family in a row plot.



Table 2-1. Geographic locations, climatic and site conditions of the two experimental sites
in north central Florida.
Yearly Yearly
average average Site
Site Location County average average Site Soil series
temperature precipitation index
(C) (mm)
29020' N
Dunnellon o0 Levy 21 1332 21 Smyrna
82050 W

2940' N
Palatka 290Putnam 22 1368 22 Adamsville
81042' W
a Site index is expressed as tree height in meters at age 25 years.


All trees were grown in greenhouses before being transplanted to field sites in

December, 1994. Site preparation for both locations included bedding and chopping. The

seedlings were planted at a 1.5 m x 3.4 m spacing at Palatka, and a 1.8 m x 3.0 m spacing

at Dunnellon. Fertilizers, herbicides, and insecticides were applied in the intensively-

managed treatment blocks only (Table 2-2).


Sampling Procedures and Inventory

For each location, two sample trees within a family from each five-tree plot were

randomly chosen by a SAS procedure (SAS Institute 1990), and then a systematic sampling

method was applied to all other families and taxa. All sample trees were healthy and free of

disease, and 192 sample trees (2 treatments x 3 blocks x 16 families x 2 trees) were selected









16

for each of the three taxa at each site. Overall, 1,152 trees (2 sites x 2 treatments x 3 blocks

x 3 taxa x 16 families x 2 trees) were sampled across the two experimental locations.

In August 1997, an inventory of all 1,152 trees was made. Measurements included

DBH, total height, crown height, and crown width. In addition, branch position, branch

diameter, and branch angle were measured along the entire stem of each tree. Other factors

derived from these records included total branch number per tree, crown shape ratio (CSR

= crown height/ crown width), and branch-free stem height (McCrady and Jokela 1996).



Table 2-2. Treatment regimes for intensively-managed and non-intensively-managed blocks
at the two research locations.
Culture Non-intensive management Intensive management

Bedding Double (Dunnellon) Double (Dunnellon)
Single (Palatka) Single (Palatka)

Fertilization None 280 kg/ha DAP + 224 kg/ha KCl
600 kg/ha 10-10-10 + Micronutrients

Herbicide None Year 1: Roundup and Atrazine

Insecticide None Year 1: Asana, Diomethorate or Pyridine
3 x standard (Dunnellon)
4 x standard (Palatka)
Note: 280 DAP + 224 KC1 kg/ha = 50 N, 56 P, 112 K kg/ha, respectively.
macronutrient and micronutrient application rates-
N = 60, P = 24, K = 50, Ca = 20, Mg = 10, S = 13, Fe = 0.5, Zn = 0.06, Mn = 0.5,
Cu = 0.06, B = 0.06 (kg/ha)


Similar measurements were made during the fourth growing season in 1998.

However, in contrast to sampling 16 families within a taxon, six families were selected

based on the first year growth data and long-term breeding values. Within each taxon, three

good and three below average families were selected to contrast differences between crown









17

structure and growth performance. A subset of eighteen families totaling 432 trees was

measured in 1998.


Biomass and Leaf Area Estimation

In July 1997 and 1998 (ages 3 and 4 years), 1,080 branches (2 locations x 2

treatments x 3 blocks x 3 taxa x 3 crown positions x 5 trees x 2 years) were destructively

harvested to develop biomass prediction equations. Individual branches, randomly selected

from each of three equally-divided crown positions (upper, middle, and lower), were cut

after insertion angle on the stem was measured. All needles from a branch were removed

separately by age class (current-year, and one-year-old), and both the needles and branches

were weighed after being oven-dried at 70 'C for at least 48 hours.

In August 1997 and 1998, foliage samples were collected to determine specific leaf

area for use in estimating total leaf area per tree. About 20 fascicles in each of the samples

by age class and crown position were randomly chosen from one tagged sample tree in each

five-tree row plot. Because many trees did not have old foliage in the upper crown, only

two positions for old foliage (upper-middle and lower) were sampled. A total of 5,760

foliage samples (2 locations x 2 treatments x 3 blocks x 3 taxa x 16 families x 2 ages x 1

trees x 3 positions [current year foliage] + 2 locations x 2 treatments x 3 blocks x 3 taxa x

16 families x 2 ages x 1 trees x 2 positions [old foliage]) were collected from the trees in

1997, and 2,160 samples were collected from 6 families of each taxa (18 families in total)

by the same procedure in 1998.

Fifteen samples from each of the three taxa were collected by location and age class

to determine specific leaf area. All-sided leaf surface area was measured in the laboratory









18

using a volume displacement method (Johnson 1984), in which we found a high correlation

between fresh foliage weight (g) and volume (cm3) of the needle samples. Regression

methods were used to estimate the volume from fresh needle weight, and then the formula

developed by Johnson (1984) was used to estimate leaf surface area of the needle samples.

Needle volume estimation equations were combined for taxa, location or age class where

statistical results indicated non-significant differences among these effects at =0.10. The

final equations for the two growing seasons were as follows:

Needle volume (cm3) = 0.067 + 1.093 Weight (g)

(r2 = 0.99, n = 180) Year 1997

Needle volume (cm3) = 0.111 + 1.129 Weight (g)

(r2 = 0.99, n = 180) Year 1998

Leaf area at each crown position was calculated as the product of leaf biomass and

specific leaf area at that position. Total leaf area per tree was a summation of leaf area at

the three different crown positions.


Statistical Analyses

According to the nature and the amount of data collected in this study, biomass

equation construction, growth performance evaluation, and crown attribute analyses were

divided into two major phases: (1) apriori estimation and hypothesis formulation; and (2)

a posteriori testing and interpretation. In phase 1, main effects and the interactions that

could potentially influence tree growth and crown structure were taken into consideration,

and a full model was established for each variable (e.g, height, total branch number per tree,

total leaf area per tree). At this phase, the likely behavior of the variable was also









19

hypothesized. In phase 2, full models were tested, and main effect and interaction terms that

were not significant (5% level by F test) were removed from the models. Final models were

developed for each variable that (1) were parsimonious in number of main effects and their

interactions; (2) contained only effects that were biologically meaningful and interpretable;

and (3) accounted for at least 60% (R2 0.60) of the variation. Predicted results were

subsequently justified against the hypothesized outcomes for each variable and all models

were checked graphically. If discrepancies existed, the final models were re-analyzed to

detect which effect or a set of effects caused unusual behaviors. In such cases, we either re-

formulated theory or decomposed the reduced model to develop a new set of functionally

efficient models that offered stronger biological interpretations.


Biomass estimation

Biomass prediction equations were developed to estimate branch and foliage biomass

using data collected from destructively harvesting 1,080 branches. The prediction equations

were applied to the inventory data (branch diameter by crown position) to estimate branch

and foliage biomass at the tree level. A full model was formulated for each biomass

component (model 2-1). All main effects except branch diameter were discrete variables in

the model.

Yijklm = + i + j +O k + 1 + m + b D + (' )ij + (' )ik ( )im (0 )il (0 )j k + ()jm

+ (* *)km + D i+ D*j + Do k + D + D m + (0 * )ijk + (0 )ijl + ( )kml + ijklm

(2-1)

where Yijklm is the estimates of leaf or branch biomass per branch at position m of taxa 1

in the treatment k and location j in year i,











* is the overall mean of the model,

* i is the effects of year (Y97 or Y98),

j is the effects of location (Dunnellon or Palatka),

Sk is the effects of treatment (non-intensive or intensive),

Sm is the effects of taxa (improved loblolly, improved slash, or unimproved slash

pine),

*, is the effects of crown position (lower, middle, or upper),

D is the branch base diameter (mm),

(* )j is the year x location interaction,

(* )k is the year x treatment interaction,

(* )im is the year x taxa interaction,

(* )i is the year x position interaction,

(* )jk is the location x treatment interaction,

(* )jm is the location x taxa interaction,

(* )km is the treatment x taxa interaction,

D* i, D j, D* k, D 1, D*m are Diameter x year, diameter x location, diameter x

treatment, diameter x taxa, and diameter x position interactions, respectively,

(* )ijk is the year x location x treatment interaction,

(* )ij is the year x location x taxa interaction,

(* )km is the treatment x taxa x position interaction,

Sijklm is the error term.









21

where i = 1, 2 for years; j = 1, 2 for locations; k = 1, 2 for treatments; 1 = 1, 2, 3 for crown

positions; m = 1, 2, 3 for taxa.


The above model was further tested for the homogeneity of error variances and

where necessary logarithmic transformations were performed. After testing, a simplified

final model was determined (model 2-2).

log( ijk ) = + i+ j + k + (* )ik + bllog(D) (2-2)


where "ijk is the estimated leaf or branch biomass at crown position k of taxa j in year i,


is the overall mean of the model,

i is the effect of year (Y97 or Y98),

j is the effect of taxa (improved loblolly, improved slash, or unimproved slash

pine),

k is the effect of crown position (lower, middle, or upper),

(* )k is the year x position interaction,

D is the branch base diameter (mm).


Analyses of mensurational and crown structural attributes

Analysis of variance (ANOVA) was used as the primary method for analyzing the

mensurational and crown structure data. Characteristics included diameter at breast height

(DBH), total tree height, total branch number per tree, live crown length, crown width,

crown shape ratio, and branch-free stem length. All of these attributes were measured on

a sample of 1,152 trees at age 3 years and 432 trees at age 4 years. Analyses were conducted









22

separately by year because of sample size differences. A full model with main effects and

their interactions was used to test each attribute within a year:

Yijklm = + i + k j + k + fl(k) + bm,) + (0 )ij + (0 )ik + ft il(k) + (0 0 )jk+ fo jl(k) + b* jkm + bfjklm

+ (* * )ijk + f( )ijkl + ijklm (2-3)

where Yijklm is the mean of two sample trees at family 1 of taxa k in block m of treatment

j of location i,

is the population mean,

i is the random variable of location NID (0, 2.),

Sis the fixed effect of treatment (non-intensive or intensive),

k is the fixed effect of taxa (improved loblolly, improved slash, or unimproved

slash pine),

fl(k) is the random variable for family nested within taxa NID (0, 2),

bmi) is the random variable for block nested within treatment NID (0, 2b),

(* )j is the random variable for location x treatment interaction NID (0, 0 2..),

( )ik is the random variable for location x taxa interaction NID (0, 2..),

f* il(k) is the random variable for location x family(taxa) interaction NID (0, 2 ),

(* )jk is the fixed effect for treatment x taxa interaction,

f* jl(k) is the random variable for treatment x family(taxa) interaction NID (0, 2 ),

b* jkm is the random variable for taxa x block (treatment) interaction NID (0, 0 2.),

bfjklm is the random variable for family(taxa) x block(treatment) interaction NID

(0, 2bf),









23

(* )ijk is the random variable for location x treatment x taxa interaction NID (0,

S .2 ),

f(* )ijk is the random variable for location x treatment x family(taxa) interaction *

NID (0, 2.),

*i ijkm s the error term NID (0, 2.).

where i = 1, 2 for locations; j = 1, 2 for treatments; k = 1, 2, 3 for taxa; 1 = 1, 2, ..., 16 for

families per taxa; and m = 1, 2, 3 for blocks.


Variance homogeneity for each variable was examined to ensure appropriate analyses

and data transformations (logarithmic transformation) were performed where necessary. If

either location x taxa or treatment x taxa interaction was not significant in the full model

(i.e., variation in location environments and treatment levels did not translate into significant

differences among taxa), but evidence suggested that differences among taxa should exist

in these attributes (Nemeth 1973; Vose and Allen 1988; Colbert et al. 1990; Dalla-Tea and

Jokela 1991; Zhang et al. 1997; Albaugh et al. 1998; Samuelson 1998; Lopez-Upton 1999),

then probable causing effects were examined. To make meaningful biological

interpretations, the full model was decomposed into separate models by location, treatment,

or both. If the analysis was done within a location but across treatments, the corresponding

final model was of the following form for each location:

Yijkm = i j fk(j) + bm(i) + (' )ij + f ijk + b ijm + t ijklm (2-4)

where Yijkm is the mean of two sample trees at family k of taxa j in block m of treatment i,

is the population mean,

i is the fixed effect of treatment (non-intensive or intensive),









24

j is the fixed effect of taxa (improved loblolly, improved slash, or unimproved

slash pine),

fkg) is the random variable for family nested within taxa NID (0, 2),

bm(i) is the random variable for block nested within treatment NID (0, 2b),

(* )j is the fixed effect for treatment x taxa interaction,

f* ijk is the random variable for treatment x family(taxa) interaction NID (0, 2.),

b* ijm is the random variable for taxa x block (treatment) interaction NID (0, 2b.),

ijkm is the error term NID (0, 2.).


When the analysis was performed within a combination of locations and treatments,

the same principles were used to obtain the final model (i.e., eliminate all terms related to

subscript i and keep other terms). The resultant model could be viewed as the full model for

analyses by location and treatment.

For pooled or separate analyses, PROC GLM in the SAS System was utilized to test

for significance of random effects, while PROC MIXED was used to test the fixed effects

and to perform linear single-degree-of freedom contrasts among taxa (Littell et al. 1996;

SAS Institute 1996). The two linear contrasts used to separate taxa differences were (1)

loblolly vs. improved slash pine (PTA vs. PEE); and (2) improved slash vs. unimproved

slash pine (PEE vs. PEU). A default level of = 0.05 was used to declare significance

unless otherwise specified.











Analyses of crown biomass and leaf area attributes

Analytical procedures for crown biomass and leaf area attributes were identical to

those described above for mensurational and crown structural variables. The attributes

examined included current-year specific leaf area (SLA), one-year-old SLA, current-year

leaf biomass, one-year-old leaf biomass, total leaf biomass, branch biomass, total crown

biomass, current-year leaf area, one-year-old leaf area, and total leaf area per tree. All

attributes were based on measurements from individual trees.

Additional analyses were performed to test for differences in the vertical distribution

patterns of leaf area within the crown. Means among the three crown positions were

compared to describe vertical crown structure. Similarly, linear contrasts were performed

to compare crown position means at = 0.05.


Relationships between volume increment and leaf area (growth efficiency)

Growth efficiency at the individual tree level was expressed as volume increment

(D2H, age 4 minus age 3 years) per unit leaf area. A similar scheme as employed for the

development of biomass estimation equations was also adopted here: a full model including

main effects, leaf area, stem volume increment, and their interactions was first proposed, and

then tested for variance homogeneity and the significance of each effect. Certain

modifications of the model were performed where necessary.

A subset of data at age 3 years corresponding to those families (18 in total for all

three taxa) selected for age 4 years were used to calculate volume increment. Leaf area data

at age 3 were used as one of the independent variables. A total of 432 sample trees was used

in the analysis. Each effect that remained in the final model was significant at = 0.05.











Results

Biomass Estimation Equations

Separate equations were developed for estimating current-year foliage, one-year-old

foliage, total foliage, and branch biomass of individual branches (Table 2-3). Treatment and

location effects were not significant and were excluded from all models. Crown position

and year x position had a larger influence on leaf biomass than year and taxa, while taxa

explained more of the variation in branch biomass than other factors. Significant differences

in prediction equations were found among taxa for all variables except one-year-old foliage;

however, differences only existed in the intercept, and not in the slope (Appendix A --

Figure A-i). Crown position had the most significant influence on foliage biomass among

the discrete variables (Table 2-3). Although only total foliage biomass per branch for

loblolly and improved slash pine were reported in Appendix A -- Figure A-1, similar results

were found for unimproved slash pine and other biomass components. In general, biomass

differences between improved and unimproved slash pine were minor (Table 2-3).


Selection of Models for ANOVA

The full model (2-3) that combined all main effects produced irregular behavior and

contradictory results, thus making it necessary to conduct separate analyses by location and

treatment. Results also indicated that differences among locations rather than treatments

were the main contributor to variation in growth patterns among taxa (Appendix A -- Figure

A-2). Two types of differences were found in all attributes across locations: (1) significant

treatment x taxa interactions in one location, but not in the other (e.g., crown width); and

(2) significant differences in absolute values of taxa (e.g., branch biomass) between











Table 2-3. Model parameter estimates and summary statistics for estimating foliage and
branch biomass in loblolly and slash pine.
Model form: log(biomass (g) ) = + (year + taxa + position + year x position) +
b,(log(diameter) (cm))
Factors Parameters Current year One-year-old Total Branch
Factors Parameters Branch
foliage foliage foliage

R2 0.65 0.70 0.72 0.91

-1.714 -2.161 -0.862 -3.735

97 -0.082 -0.844 -0.172 -0.203
Year
98 0 0 0 0

PTAa 0.260 b 0.140 0.418
Taxa PEU -0.032 -0.002 -0.016

PEE 0 0 0

Lower -2.160 1.720 -0.429 0.213
Crown
n Middle -0.840 1.776 0.030 0.189
position
Upper 0 0 0 0

97 Lower 1.027 1.169 0.693 c

97 Middle 0.789 0.130 0.214

Yearx 97 Upper 0 0 0
Position 98 Lower 0 0 0

98 Middle 0 0 0
98 Upper 0 0 0

b, Diameter 2.175 1.500 1.911 2.708
a PTA = improved loblolly pine
PEE = improved slash pine
PEU = unimproved slash pine
b taxa was not statistically significant and, therefore, was not included in the model.
c yearxposition was not statistically significant and, therefore, was not included in the
model.









28

locations. Hence, all ANOVA analyses were separately conducted by location using model

(2-4) to investigate treatment effects and their interactions with taxa and families.

Genetic effects on most variables were better expressed under the intensive

treatment for all taxa at both locations (Appendix A -- Figure A-2). Differences between

improved and unimproved slash pine for most variables were not statistically significant

under the non-intensive treatment (p 0.05), but significant differences were detected under

the intensive treatment (p 0.05). Lopez-Upton (1999) showed similar results for volume

when making comparisons among the three taxa.


Mensurational and Crown Structural Characteristics

Interspecific differences existed in some mensurational and crown structural

characteristics (Table 2-4). Loblolly pine was more responsive than the two slash pine taxa

to the silvicultural treatments at Dunnellon, with DBH responses averaging 103% and 73%

at age 3 and age 4 years, respectively. Slash pine also showed a similar trend, but at a

diminished level, i.e., the increase in DBH at age 3 and age 4 years was 63% and 44% for

improved slash pine, and 47% and 30% for unimproved slash pine, respectively. At Palatka,

the same trend followed at age 3 years, but the most noticeable difference was observed in

DBH growth at age 4 years. Treatment effects only increased DBH about 17%, 23%, and

28% for loblolly, improved slash, and unimproved slash pine, respectively. Therefore,

treatment x taxa interactions for DBH were significant at Dunnellon, but not at Palatka for

the years examined (Appendix B).

Height growth was consistently greatest in loblolly pine, but not significantly

different from slash pine under the same treatments (Table 2-4). Differences in tree height









29

did not exist between improved and unimproved slash pine. Treatment x taxa interactions

were significant at Dunnellon for both ages, but not at Palatka (Appendix B). Loblolly and

improved slash pine were most responsive to the intensive treatment at Dunnellon, especially

at age 4 years, with height growth responses averaging 48% and 37%, respectively.

Corresponding height growth responses at Palatka were 17% and 27%. Height growth

responses to the intensive treatment were smaller at age 4 years at Palatka. For example,

loblolly pine gained only 17% in height at age 4 years, compared to 41% at age 3 years.

Loblolly pine maintained about 44 and 39 branches per tree at age 3 and 4 years,

respectively. Slash pine retained about 10 and 8 branches fewer than loblolly pine (Table

2-4). All three taxa showed decreases in total branch numbers between ages 3 to 4 years

under the intensive treatment at both locations, while trees grown under the non-intensive

treatment maintained the same number of branches between years.

Live crown length and branch-free stem length showed similar responses to the

treatments in all taxa (Appendix B). At this stage of development, almost 90% of the stem

contained branches and, therefore, live crown length reflected the same trend as tree height

for the respective treatments. An almost constant live crown length between ages 3 and 4

years was observed under the intensive treatment at Palatka, implying that crown closure

occurred there at age 4 years.

Crown width followed the same trend as live crown length between ages 3 and 4

years under the intensive treatment at Palatka, presumably due to crown closure (Table 2-4).












Table 2-4. Individual tree growth and crown characteristics for 3- and 4-year-old loblolly and slash pine planted at two locations
in north central Floridad.
Location Dunnellon Palatka

Treatment Non-intensively Intensively managed Non-intensively Intensively managed

managed managed

Taxa PTAb PEE PEU PTA PEE PEU PTA PEE PEU PTA PEE PEU

Variables:

DBH (cm) Year 3 3.0a 3.8b 3.8b 6.1a 6.2a 5.6a 4.9a 5.la 4.8a 8.la 8.la 7.2a

Year 4 5.2a 6.1b 6.4b 9.0a 8.8a 8.3a 9.2a 8.3ab 7.6b 10.8a 10.3a 9.6a


Height (m) Year 3 2.9a 2.8a 2.9a 4.3a 3.9ab 3.7b 3.4a 3.1b 3.0b 4.8a 4.4b 4.2c

Year 4 4.0a 3.8a 4.2a 5.9a 5.2a 5.la 5.2a 4.4b 4.3b 6.1a 5.7a 5.5a

Branch no. Year 3 36a 31b 30b 46a 37b 36b 41a 31b 29b 51a 39b 39b
per tree
Year 4 35a 29a 29a 38a 28b 34ab 42a 34b 32b 39a 30b 31b


Live crown Year 3 2.6a 2.5a 2.6a 4.0a 3.5ab 3.4b 3.la 2.8b 2.7b 4.5a 4.0ab 3.9b
length (m)
Year 4 3.5a 3.2a 3.6a 4.9a 4.0b 4.1b 4.4a 3.7b 3.4b 4.7a 3.9b 4.lab













Table 2-4--Continued.
Location Dunnellon Palatka

Treatment Non-intensively Intensively managed Non-intensively Intensively managed
managed managed

Taxa PTAb PEE PEU PTA PEE PEU PTA PEE PEU PTA PEE PEU

Variables:


Crown Year 3 1.4a 1.2b 1.3b 2.2a 1.8b 1.6b 1.8a 1.4b 1.3b 2.6a 2.2b 1.9b
width (m)
Year 4 1.7a 1.6a 1.6a 2.2a 1.9a 2.0a 2.2a 1.9b 1.7b 2.2a 2.3a 2.2a


Crown Year 3 1.9a 2.lb 2.lb 1.9a 2.0a 2.2a 1.8a 2.0ab 2.1b 1.8a 1.9a 2.lb

shape ratio Year 4 2.1a 2.3b 2.3b 2.3a 2.2a 2.1a 2.1a 2.0a 2.0a 2.2a 1.7b 1.9ab

Branch- Year 3 0.2a 0.3b 0.3b 0.3a 0.4b 0.4b 0.3a 0.3a 0.3a 0.3a 0.3a 0.3a
free stem
length(m) Year 4 0.5a 0.6a 0.6a 0.9a 1.2a 1.0a 0.7a 0.8a 0.9a 1.4a 1.9a 1.4a


a Taxa means were tested by year, location and treatment separately for each age group. Means among the three taxa for a given variable
and year followed by the same letter within a treatment are not statistically significant at the 95% confidence level using linear
contrasts of the MIXED procedure.
b PTA = improved loblolly pine
PEE = improved slash pine
PEU = unimproved slash pine









32

Trees at Dunnellon had smaller crowns (crown length and width) than those at Palatka.

Loblolly pine consistently had greater crown width than slash pine, and was more responsive

and unimproved slash pine (33% vs. 46%). A similar trend was observed at age 4 years, but

it was not as significant as at age 3 years, and likely reflected the advent of crown closure.

Crown shape potentially influences the light distribution patterns within the crown

and can be viewed as an inherent species' characteristic that reflects adaptation to the

prevailing environment. Silvicultural treatments had no significant effects on crown shape

ratio (=crown length/crown width), but taxa differences were significant at both ages and

locations (Appendix B). At Dunnellon, crown shape ratio steadily increased between years

under both silvicultural treatments, except for unimproved slash pine in the intensive

treatment. Trees at Palatka exhibited a reverse trend in crown shape ratio between loblolly

and slash pine; crown shape ratio for loblolly pine increased from ages 3 to 4 years under

both treatments, while this ratio declined in slash pine.


Biomass and Leaf Area Characteristics of Tree Crowns

Tree crowns are most dynamic during early stages of development. Environmental

effects (including treatments) can exert significant influence on crown characteristics. As

revealed in this study, the intensive treatment significantly affected all crown characteristics

at both sites (Appendix C). Significant interspecific (taxa) and intraspecific (families)

differences (p 0.01) were observed in almost all crown characteristics, except at age 4

years at Dunnellon, where significant differences in many crown characteristics diminished.











Specific leaf area

Inconsistent treatment effects were observed across locations for current-year specific

leaf area (SLA). At Palatka, current-year SLA increased significantly (p 0.0004 at age

3, and p 0.0046 at age 4) in response to the intensive silvicultural treatment. In contrast,

current-year SLA decreased at Dunnellon on the intensive plots at age 3 years (p 0.0161)

but increased significantly at age 4 years (p 0.0295) (Table 2-5). Overall, loblolly pine

appeared to have higher current-year SLA (176.6 cm2/g) than slash pine (146.4 cm2/g).

Improved slash pine consistently had higher current year SLA than unimproved slash pine

(147.8 cm2/g vs. 144.9 cm2/g).

Among all taxa, one-year-old needles generally had lower SLA than current-year

needles (Table 2-5). Loblolly pine had higher SLA (135.6 cm2/g) in one-year-old needles

than slash pine (117.7 cm2/g).Treatment effects on SLA for older needles were non-

significant. For example, treatment effects on this variable across the three taxa diminished

at Dunnellon (p 0.2438 at age 3 years, and p 0.2000 at age 4 years). Further, treatment

x taxa interactions were not significant for SLA at either age (p > 0.10), indicating that all

taxa responded similarly to the silvicultural treatments.













Table 2-5. Specific leaf area (SLA), leaf biomass, branch biomass, and leaf area for 3- and 4-year-old loblolly and slash pine planted at
two locations in north central Floridaa.

Location Dunnellon Palatka

Treatment Non-intensively Intensively managed Non-intensively Intensively managed
managed managed

Taxa PTAb PEE PEU PTA PEE PEU PTA PEE PEU PTA PEE PEU


Variables:

Current-year Year 3
SLA (cm2/g) Year 4

One-year-old Year 3
SLA (cm2/g) Year 4

Current year
Year 3
leaf biomass
Year 4
(kg/tree)

One-year-old
Year 3
leaf biomass
Year 4
(kg/tree)


Year 3
Year 4


181.la 148.lb
173.2a 151.5b

135.7a 121.9b
131.5a 115.7b


0.93a 0.81a
1.61a 1.35a



0.05a 0.04a
0.15a 0.14a



0.98a 0.85a
1.76a 1.49a


147.8b
145.7b

119.4b
114.7b


0.80a
1.50a



0.04a
0.15a



0.84a
1.65a


172.5a
183.3a

139.2a
127.3a


2.38a
3.46a



0.10a
0.27a



2.48a
3.73a


143.3b
158.4b

122.6b
117.4b


1.69b
2.45b



0.08ab
0.22a



1.77ab
2.67b


140.8b
152.3b

119.9b
109.9b


1.39b
2.40b



0.07b
0.23a



1.46b
2.63b


171.4a
170.0a

143.6a
127.0a


1.74a
4.04a



0.07a
0.29a



1.81a
4.35a


134.2b
145.4b

117.3b
111.7b


1.26b
2.68b



0.06ab
0.25b



1.32b
2.93b


130.0c
147. lb

114.5b
110.6b


1.07b
2.16b



0.04b
0.21b



1.13b
2.37b


179.4a
181.6a

144.9a
135.7a


3.87a
4.52a



0.15a
0.33a



4.02a
4.85a


149.2b
152.2b

128.6b
120.9b


2.92b
3.44b



0.13a
0.29a



3.05b
3.73ab


147.8b
148.0b

125.6b
113.2b


2.15b
3.11b



0.lOab
0.27a



2.25b
3.38b


Total leaf
biomass
(kg/tree)













Table 2-5--Continued.
Location Dunnellon Palatka

Treatment Non-intensively managed Intensively managed Non-intensively managed Intensively managed

taxa PTAb PEE PEU PTA PEE PEU PTA PEE PEU PTA PEE PEU

Variables:

Branch
Branch Year3 0.48a 0.33b 0.33b 1.62a 0.85b 0.65b 1.14a 0.63b 0.51b 3.17a 1.87b 1.22b
biomass
bioms Year 4 1.04a 0.70b 0.81ab 2.99a 1.62b 1.44b 3.58a 1.73b 1.28b 4.26a 2.61b 2.25b
(kg/tree)

Total crown
Totalro Year 3 1.45a 1.17a 1.17a 4.10a 2.62b 2.11b 2.94a 1.95b 1.65b 7.18a 4.91b 3.46b
biomass
Yiomass ear 4 2.79a 2.18a 2.45a 6.72a 4.30b 4.06b 7.93a 4.66b 3.64b 9.11a 6.34b 5.63b
(kg/tree)

Currentyear Year3 15.2a l .b 10.8b 38.2a 22.6b 18.lb 27.4a 15.9b 13.2b 64.5a 41.2b 29.6b
leaf area
(M/treea Year 4 25.la 18.lb 19.7b 54.1a 33.8b 31.7b 60.5a 34.8b 28.3b 71.0a 46.4b 40.8b
(m2/tree)

One-year- Year 3 0.6a 0.5a 0.5a 1.4a 1.b 1.0b 1.la 0.7b 0.6b 2.2a 1.6ab 1.3b
old leaf area
(M /tree) Year 4 1.9a 1.6a 1.7a 3.4a 2.5a 2.5a 3.9a 2.7b 2.3b 4.4a 3.4b 3.0b
(m2/tree)

Total leaf
Total leaf Year 3 15.8a 11.5b 11.3b 39.6a 23.6b 19.1b 28.5a 16.6b 13.8b 66.7a 42.8b 30.9b
area
(2/tree) Year 4 27.0a 19.7b 21.4b 57.5a 36.3b 34.2b 64.4a 37.5b 30.6b 75.4a 49.8b 43.8b
(m /tree)
a Taxa means were tested by year, location and treatment separately for each age group. Means among the three taxa for a given
variable and year followed by the same letter within a treatment were not statistically significant at the 95% confidence level by linear
contrasts using the MIXED procedure.
b PTA = improved loblolly pine PEE = improved slash pine PEU = unimproved slash pine











Leaf biomass, branch biomass, and total crown biomass

Leaf biomass is a major component of crown biomass at early growth stages. The

intensive silvicultural treatment significantly influenced leaf biomass accumulation

(Appendix C). At age 3 years, leaf biomass (current year, one-year-old, and total) under

the intensive treatment was more than double that under the non-intensive treatment at both

sites, except for unimproved slash pine at Dunnellon (Table 2-5). Loblolly pine tended to

accumulate more leaf biomass than improved slash pine, regardless of treatments and

locations (e.g., 15% and 40% greater under non-intensive and intensive treatment at

Dunnellon, respectively). Unimproved slash pine generally accumulated less leaf biomass

than improved slash pine. Treatment x taxa interactions were significant in current-year and

total leaf biomass, except at age 4 years at Palatka (Appendix C). Differential responses

between the three taxa to the silvicultural treatments were the primary cause for this

interaction. Further analyses revealed that scale effects rather than rank changes contributed

to the interaction (Table 2-5).

Branch biomass generally accounted for less than 50% of the total crown biomass.

When branch biomass was compared across locations for a given treatment, trees grown at

Palatka had higher amounts than at Dunnellon (Table 2-5). For example, branch biomass

at age 3 years under the non-intensive treatment at Palatka was 138% greater than at

Dunnellon. In general, the order of branch biomass accumulation among taxa was loblolly

pine > improved slash pine > unimproved slash pine. Treatment x taxa interactions were

inconsistent across locations and years (Appendix C).









37

Crown biomass (total leaf and branch biomass) showed similar results to the

individual components across locations and treatments (Table 2-5). At age 4 years, crown

biomass for loblolly and slash pine grown under the non-intensive treatment at Palatka

exceeded that of the intensive treatment at Dunnellon. Loblolly pine increased crown

biomass by 170% under the non-intensive treatment at Palatka, but only 27% under the

intensive treatment from ages 3 to 4 years. Similar results were found for slash pine

(increases of 139% and 29% for improved slash pine, 121% and 63% for unimproved slash

pine for the corresponding treatments at ages 3 and 4 years, respectively). Increases in

crown biomass over the same period were also observed at Dunnellon, but not as markedly

as at Palatka. Interspecific differences were significant, with loblolly pine accumulating

more crown biomass than slash pine. Crown biomass for improved and unimproved slash

pine showed different trends for the two treatments across locations. Unimproved slash pine

accumulated more crown biomass than improved slash pine under the non-intensive

treatment, but the reverse was true under intensive treatment at Dunnellon. At Palatka,

improved slash pine accumulated more crown biomass than unimproved slash pine for both

treatments (Table 2-5). Hence, treatment x taxa interactions represented a rank change at

Dunnellon, but a scale effect at Palatka.


Leaf area characteristics

Leaf area is one of the most important variables to influence biomass accumulation

and productivity in forest stands. Significant interspecific and intraspecific differences in

leaf area were found across locations (Appendix C). Current-year leaf area generally









38

accounted for more than 90% of the total leaf area, indicating that leaf life span was not

significantly different among loblolly and slash pine.

The intensive silvicultural treatment significantly increased total leaf area per tree

at both locations. For example, leaf area for loblolly pine at age 3 increased from 15.8

m2/tree under the non-intensive treatment to 39.6 m2/tree under the intensive treatment at

Dunnellon. Similar results were found at Palatka, where total leaf area for loblolly pine

increased from 28.5 m2/tree to 66.7 m2/tree due to treatment. Slash pine accumulated less

leaf area than loblolly pine at both locations, but differences between unimproved and

improved slash pine were not the same across locations. Improved slash pine accumulated

more leaf area than unimproved slash pine under both treatments at Palatka, but the

differences were not significant under most instances at Dunnellon (Table 2-5). On average,

trees grown under the non-intensive treatment increased leaf area by 70% at Dunnellon

andl22% at Palatka between ages 3 and 4 years. In contrast, trees gained 45 79% more

leaf area due to the intensive management at Dunnellon, and 13 42% at Palatka,

respectively. Treatment x taxa interactions for other attributes (e.g., total leaf area) were

significant (p 0.05) in the two years at Dunnellon, but only significant (p 0.10) at age

3 years at Palatka (Appendix C).


Vertical Distribution of Total Leaf Area

The vertical distribution of leaf area can be important in affecting the interception

of light energy by leaves within the crown. Previous research has shown that the vertical

distribution of leaf area can exert significant influence on light extinction patterns within the

crown (Waring and Schlesinger 1985). Vertical differences in total leaf area per tree among









39

taxa were shown in both relative distribution patterns and the absolute amount of leaf area

along crown profiles (Figure 2-1). Trees planted at both locations had the most leaf area in

the lower and middle crown positions at age 3 years. Loblolly pine tended to have 50% of

the total leaf area in the lower crown at Dunnellon, while at Palatka it either had the highest

amount of leaf area in the middle crown (non-intensive treatment) or an equal amount

between the middle and lower crown positions (intensive treatment). On average, loblolly

pine partitioned about 8% of its leaf area in the upper crown, while slash pine partitioned

about 13%, as reflected in the significant taxa x position interactions (p 0.0008).

Location x treatment x position interactions were statistically significant (p *

0.0030), as was the location x treatment x taxa x position interactions (p 0.0001) for total

leaf area. These results demonstrated the complex and significant interactions that occurred

among locations, treatments, taxa, and crown positions on the vertical distribution of leaf

area. In addition, the vertical distribution of leaf area generally corresponded to the vertical

distribution of branch biomass.


Relationships between Volume Increment and Total Leaf Area per Tree (Growth
Efficiency)

The rapid growth of forest stands depends on the accumulation of leaf area to

intercept light energy for photosynthesis. Thus, the amount of leaf area in a stand can be a

direct measure of potential production. A set of linear models with both quantitative and

qualitative variables was developed to examine the relationships between leaf area and

volume increment between ages 3 and 4 years (r2 = 0.77, p 0.0001). Based on predicted

results, improved and unimproved slash pine were not significantly different in volume

















Dunnellon
Intensive management

PEU





PEE


b
b


PTA


0.5
Branch biomass (kg/tree)


a 14%0
b 50%
c 36%o


a 14%
b 48%
c 38%


Upper

Middle

Lower


15% a
47% b
38% c


160 a
44% b
40% b


9% 10 a
43% 40%


b


48% 50%


0


2 4 6 8 10 12 14 16
Total leaf area (m2/tree)


Figure 2-1. Vertical distribution of total leaf area (right) and branch biomass
(left) by crown positions at age 3 years for loblolly and slash pine managed
under two silvicultural treatment regimes in north central Florida.


1 0 5 10 15 20 25 30 35

Dunnellon
Non-intensive management
a 10% 11% a Upper
PEU b 45% 43% b Middle
b 45% 46% b Lower


a 13% 13% a
b 45% 45% b
PEE
b 42% 42% b

a 8% 10% a
b 42% 41% b

PTA b 50% 49% c


I I I I I I I I I I I I I I


I




















PEU


a 130/ 16% a
52% 48%


c 35% 36% c

a 112% 14% a
52% 50%
36% 36%


b


Palatka
Intensive management


Upper

Middle

Lower


b


PEE




b
PTA
b


2 1 0 5 10 15 20 25 30 35

Palatka
Non-intensive management
a 19 21% a Upper
PEU b 59% 55% b Middle
a 22 24 a Lower


a 18% 22% a
b 53% 49% b
c 29% 29% c


a 18% 11% a
b 55% 51% b

PTA c 37% 38% c




.0 0.5 0 2 4 6 8 10 12 14 16
Branch biomass (kg/tree) Total leaf area (m2/tree)

Figure 2-1-- Continued.
Note: PEU = unimproved slash pine, PEE = improved slash pine, PTA = improved loblolly pine
Position means for a given variable of a species followed by the same letter are not significantly
different at the 95% confidence level by linear contrasts using the MIXED procedure of SAS.
Percentage numbers around the middle of the bars are the relative partition of leaf area or branch
biomass by crown positions.


a 7% 1 9% a
49% 45%


44% 46%


I


i I












produced per unit leaf area (1,106 and 1,173 cm3 stemwood/m2 leaf area, respectively,

averaged across locations and treatments), but differences did exist between loblolly and

slash pine (Figure 2-2). Loblolly pine consistently produced less volume per unit leaf area

(804 cm3 wood/m2 leaf area) than slash pine under the same locations and treatments.

Similar results were also reported in a stand-level comparison of growth efficiency between

4-year-old loblolly and slash pine (Colbert et al. 1990).

The intensive silvicultural treatment significantly increased growth efficiency ofboth

taxa at Dunnellon, while the opposite was true at Palatka (p 0.0001). Additionally,

location x treatment effects were significant (p 0.0001), which indicated that the Palatka

site provided pine trees with more efficient growth per unit leaf area than the Dunnellon site

for the non-intensive treatment (856 vs. 1,357 cm3 stemwood/m2 leaf area, averaged across

taxa for Dunnellon and Palatka, respectively). In contrast, the Dunnellon site was more

favorable for growth efficiency than the Palatka site when intensive silvicultural treatments

were applied (1,089 vs. 809 cm3 stemwood/m2 leaf area, averaged across taxa for Dunnellon

and Palatka, respectively). Further, volume growth rates converged among taxa as leaf area

approached 54 m2/tree under either silvicultural treatment at Dunnellon, and 40 and 55

m2/tree under the non-intensive and intensive treatments, respectively, at Palatka.


















Dunnellon


U


-~
v7
7,


-0- PEU low
-0O... PEE low
-- PTAlow
--.. PEU_high
-- PEEhigh
-- PTA_high


;


I I I I I I I
12 3 4 51 2 3 4
Total leaf area (m2/tree) (log scale)
Figure 2-2. Relationship between individual tree volume increment from ages 3 to 4 and total leaf area (all-sided)
for loblolly and slash pine managed under two silvicultural treatment regimes in north central Florida.


unimproved slash pine
improved slash pine
improved loblolly pine


low = non-intensive management
high = intensive management


PEU
PEE
PTA


Palatka











Discussion

Comparisons of Foliage and Branch Biomass Estimation Equations

The foliage-carrying capacity of branches varied significantly among the three crown

positions. Leaf biomass per branch, when expressed separately by needle age class, could

not be accurately predicted by branch diameter alone (r2 = 0.20 for current year leaves, and

r2 = 0.11 for one-year-old leaves). Similar results were found in western hemlock (Tsuga

heterophylla), Douglas-fir (Pseudotsuga menziesii), and grand fir (Abies grandis) (Kershaw

and Maguire 1995). Therefore, equations unique to different crown positions were needed

to satisfactorily predict leaf biomass and leaf area distribution within the crown (Gilmore

and Seymour 1997). However, total leaf biomass and branch biomass could be predicted

satisfactorily using only branch diameter (Table 2-6), though other factors were still

important. Model evaluation revealed that leaf biomass was significantly variable under

diverse conditions, but branch biomass was relatively independent of external conditions.


Mensurational and Crown Structural Characteristics

Significant differences were observed among taxa for the many mensurational and

crown characteristics examined. Loblolly pine generally had greater DBH and height than

slash pine. A related study showed that these taxa differences could be partly attributed to

lower fusiform rust incidence in loblolly pine (Lopez-Upton 1999). However, studies

conducted in rust-free loblolly and slash pine stands suggested that other factors such as

canopy structure and growth habit may be more responsible for interspecific differences

(Nemeth 1973; McCrady and Jokela 1998). As revealed in this study, branch numbers per









45

tree, crown length and width, and crown shape ratio all showed interspecific variation at this

stage of development.



Table 2-6. A comparison (r2) between estimation models for foliage and branch biomass
based on branch diameter alone and multi-factor variables.
Biomass components Diameter based modela Multiple factor model

Total foliage 0.62 0.72

Current year foliage 0.20 0.65

One-year-old foliage 0.11 0.70

Branch 0.82 0.91
a model constructed using only branch base diameter,
i.e., log(biomass) = bo + b,(log(diameter)).
b model form was identical to that in Table 2-3.



Crown structure was difficult to model since much variation occurred among

individual trees (Doruska and Burkhart 1994). As found in Scots pine (Pinus sylvestris),

structural attributes related to higher biomass accumulation include higher branch numbers

per whorl and longer crowns (Kuuluvainen et al. 1988). Total branch number accounted for

42% of the variation in volume accumulation among Douglas-fir families (King et al. 1992).

In the current study, differences in branch numbers could also be related to the superiority

of loblolly pine growth relative to slash pine. When both total branch number per tree and

crown length were significantly greater than slash pine, loblolly pine generally had greater

height growth (Table 2-4). However, when loblolly pine had higher branch numbers per tree

but shorter crown length (e.g., under the non-intensive treatment at Dunnellon), its DBH or

height growth was less than slash pine.









46

The distribution of foliage and branches within a crown can influence light

penetration and, ultimately, growth potential. Crowns that have a tight branching pattern

could reduce light penetration, especially to the lower foliage elements. Sparser crowns

could be an important adaptive mechanism that allow trees to optimize growth performance

in varying environments (Makela and Vanninen 1998). A simulation study demonstrated

that asymmetrical crown development, an expression ofphenotypic plasticity of crowns, was

advantageous to productivity (Sorrensen et al. 1993). In contrast, Kellomaki et al. (1985)

reported that crown shape had little influence on light interception, and that narrower,

symmetrical crowns were most efficient in affecting growth potential. Other theories have

also been proposed relative to the importance of crown structure on adaptation to harsh

environments (Sprugel 1989; Smith and Brewer 1994). In the current study, narrower

crowns did not facilitate more rapid growth, as loblolly pine was more productive than slash

pine and it also had wider crowns.

Crown dimensional differences have been shown to significantly influence stem

biomass partitioning in Picea abies and Picea abies f. pendula (Pulkkinen 1991),

aboveground biomass in Scots pine (Kuuluvainen and Kanninen 1992), and height increment

in loblolly pine (McCrady and Jokela 1996). Although CSR combined two important crown

parameters and was statistically significant among taxa, it appeared to have little ecological

significance in this study because the ratio tended to stabilize around 2 (Table 2-4). In a

similar study, McCrady (1993) observed significant intraspecific variation in crown shape

ratio in young loblolly pine plantations, but did not find an advantage of narrower crowns

over wider crowns in height growth. Crown shape ratio may be more of an indication of









47

environment-induced adaptation rather than a significant characteristic that can be used to

differentiate functional groups. Although long-term breeding programs have selected

progenies of Norway spruce and Scots pine with high crown shape ratios in the cold

temperate region (Kellomaki et al. 1985), rapid crown development at early growth stages

is one characteristic that distinguishes subtropical pine taxa from other coniferous taxa in the

north, temperate regions. As shown in this study and a related stand level study in loblolly

pine plantations (McCrady 1993), higher crown shape ratio did not translate into growth

advantages at early stages of stand development (r = -0.17, p 0.0005 between crown shape

ratio at age 3 and volume increment).


Crown Biomass, Leaf Area, and Their Vertical Distribution

Crown (branch and leaf) biomass, total leaf area, and their vertical distribution

patterns have been closely associated with stand structure, forest productivity, and

microclimate of the habitat (Maguire and Bennett 1996). Total leaf area at the tree level and

leaf area index at the stand level were both positively associated with the annual productivity

of many species (Waring and Schlesinger 1985). Silvicultural treatments increased total

leaf area per tree by primarily augmenting leaf biomass rather than changes in specific leaf

area (Table 2-5). For example, specific leaf area consistently decreased in all taxa at age 3

years at Dunnellon, while leaf area doubled due to the intensive silvicultural treatment.

Similar results were previously reported in loblolly pine (McCrady and Jokela 1996). These

results contrast those reported for sweetgum (Liquidambar styraciflua), where leaf area

increases due to fertilization were largely attributable to leaf size increases (Kuers and

Steinbeck 1998).









48

Biomass allocation between leaves and branches varied markedly among taxa across

locations and treatments (Table 2-7). Slash pine allocated more crown biomass to foliage

than loblolly pine. Trees at Dunnellon had a higher leaf:branch biomass ratio than at Palatka

for all taxa and treatments. Loblolly pine allocated more biomass to branches, which may

offer growth advantages by building larger crowns during the early stages of stand

development. Consequently, its crown carrying capacity could increase (more leaf area per

tree was attained), and mutual shading of leaves could also be avoided. The advantage of

this biomass allocation pattern was more pronounced when leaf area was large, suggesting

that growth efficiency of loblolly pine, although lower than slash pine at low levels of leaf

area, could eventually exceed that of slash pine after crown closure (Figure 2-2). Intensive

management tended to favor relative biomass allocation to branches in developing spacious

crowns at early growth stages (Table 2-7).



Table 2-7. Variation of leaf:branch biomass ratio as influenced by location, treatment, and
tree age in loblolly and slash pine in north central Florida.
Non-intensive management Intensive Management
Location Age
PTAa PEE PEU PTA PEE PEU

3 2.04 2.58 2.55 1.53 2.08 2.25
Dunnellon
4 1.69 2.13 2.04 1.25 1.65 1.83

3 1.60 2.09 2.22 1.27 1.63 1.84
Palatka
4 1.22 1.69 1.85 1.14 1.43 1.50
a PTA = improved loblolly pine
PEE = improved slash pine
PEU = unimproved slash pine









49

Treatment effects improved growth performance by increasing crown size (Table 2-

4), but the vertical distribution of branches and leaves were not significantly affected. For

example, intensive management significantly increased total leaf area, foliage and branch

biomass, but their relative vertical distribution within the crown largely remained unchanged

(Figure 2-1). Gillespie et al. (1994) reported similar results in young loblolly pine, and

further indicated that for a given branch size, fertilized plots could carry more leaf biomass

than untreated plots. In the current study, treatment effects on leaf carrying capacity of

branches were not statistically significant (Table 2-3). Joggi et al. (1983) argued that the

vertical distribution of leaf area was less important in determining net photosynthetic rate

than LAI and position of leaf age in canopies of red clover (Trifoliumpratense). However,

forest trees are much larger than herbaceous plants and should have developed optimum

vertical distribution patterns that could intercept more light energy.


Leaf Area and Growth Efficiency

Variation in leaf area is probably one of the most prominent and dynamic

characteristics of forest stands in corresponding to seasonal or yearly changes in

environmental conditions. Positive relationships can be found between leaf area and growth

rates or total biomass accumulation in many species (Gholz et al. 1991; Gower et al. 1993;

McCrady and Jokela 1998). Although annual wood formation per unit leaf area has been

reported independent of cultural treatments and species in some studies (Norby 1996), highly

significant differences in growth efficiency were detected in relation to silvicultural

treatments between loblolly and slash pine (Colbert et al. 1990). Relationships between leaf

area and growth efficiency (basal area growth per unit leaf area) also varied in jack pine









50

(Pinus banksiana) and red pine (P. resinosa) with different stand origins (Penner and

Deblonde 1996). Variability in aboveground net primary production (ANPP) could be

largely explained by specific leaf area and leaf area index in some conifer and hardwood

stands (Fassnacht and Gower 1997). As shown in this study, growth efficiency differed

significantly in relation to silvicultural treatments, taxa, and locations (Figure 2-2),

suggesting that growth efficiency is highly variable and it reflects the growing conditions

of forest stands.

During the juvenile stages of stand development, slash pine was more efficient in

dry-matter production per unit leaf area than loblolly pine (Colbert et al. 1990). However,

following peak leaf area accretion, stem wood growth efficiencies of loblolly pine can match

that of slash pine (Jokela and Martin 2000). The issue that remains unanswered is the reason

behind the changes in growth efficiency between loblolly and slash pine. Because of the

enhanced growth impacts caused by intensive management, crown closure was accelerated,

especially at Palatka. Under the non-intensive treatment at both Palatka and Dunnellon,

crown closure would not be achieved for at least two or more years. These findings, along

with similar results from other studies, suggests that crown closure is the period when

loblolly pine meets or exceeds slash pine in growth efficiency. Under similar conditions,

loblolly pine, having more branches, higher SLA, and larger crown biomass than slash pine,

may have more efficient light penetration and interception by foliage after crown closure.

However, before crown closure, when light is not a limiting factor, these crown

characteristics could not fully provide loblolly pine with higher light use efficiency.

Although slash pine had fewer branches, lower SLA, and smaller leaf area in comparison









51

to loblolly pine, its sparser crowns could allow more efficient light penetration in the crown

and, therefore, could offset any disadvantage associated with lower leaf amount to gain a

more efficient level of stemwood growth. Ford (1985) came to the same conclusion by

presenting differences in foliage display schemes and branching patterns among contrasting

conifer species.

Biomass allocation to various tree components between loblolly and slash pine trees

may also contribute to differences in stemwood growth efficiency for the two species. At

early growth stages, loblolly pine preferably allocated more photosynthate to crown (foliage

and branch) development than slash pine. On the contrary, slash pine allocated more

photosynthate to the stem for storage. For example, slash pine allocated more biomass to

the bole (stem + bark) at age 4 years than loblolly pine (58% vs. 44%, respectively) (Colbert

et al. 1990). Therefore, slash pine showed more efficient stemwood growth than loblolly

pine. After crown closure, loblolly pine started to allocate more photosynthate to stemwood

because foliage development peaked, which led to the convergence of growth efficiency for

the two species (Figure 2-2). For example, Jokela and Martin (2000) found non-significant

differences in stemwood growth efficiency for the two species at age 14 years. Stemwood

allocation (percentage of total aboveground biomass) averaged 65.3% for loblolly pine and

62.8% for slash pine. However, more detailed studies should include physiological

responses of needles to environment or treatment induced changes to better understand

growth strategies for the two species in the future.











Summary

Comparisons and contrasts were made on juvenile growth performance and crown

structural characteristics among genetically improved loblolly, unimproved and improved

slash pine planted at two locations and managed under two levels of silvicultural intensity

in north central Florida. Loblolly pine accumulated more volume and crown biomass than

slash pine at both ages 3 and 4 years. Improved slash pine generally grew faster than

unimproved slash pine, but a significant treatment x taxa interaction was detected as

unimproved slash pine outperformed improved slash pine when the silvicultural treatment

intensity was low. Significant differences in growth were associated with variation in crown

structure or biomass characteristics among the taxa.

Crown position and branch diameter were the most significant factors influencing

foliage biomass per branch. However, location and treatment effects were not statistically

significant (p 0.10) in determining the foliage biomass carrying capacity of branches.

Significant differences in crown structural traits (total branch number per tree, crown width

and length) were related to the growth performance between loblolly and slash pine. At

early growth stages, loblolly pine had more branches per tree and allocated more biomass

to branches than slash pine for crown development. Branch:leafbiomass ratios were closely

related to the growth performance among taxa. A greater branch:leaf biomass ratio could

represent a growth strategy important for developing spacious crowns that facilitate faster

growth due to increased leaf area carrying capacity within the crown.

Treatments significantly increased total leaf area accumulation, but had little impact

on the relative distribution of leaf area along the crown profile. Corresponding crown









53

structural changes and biomass accumulation patterns under the intensive treatment led to

significant differences in overall growth performance. Specific leaf area (SLA) was one of

the adaptation variables sensitive to location, treatment, taxa, crown position, and leaf age.

Evidence from this study showed that leaf area increases associated with the intensive

silvicultural treatment were primarily attributed to increases in leaf biomass, rather than

large changes in SLA. However, the importance of SLA in differentiating interspecific

characteristics should not be neglected.

Significant differences in growth efficiency (volume produced per unit leaf area per

year), mediated by location and treatment, were detected between loblolly and slash pine.

Loblolly pine generally had lower growth efficiency than slash pine, although a convergence

among taxa was achieved when leaf area levels became large, and possibly resulted from

crown structural changes that facilitated more effective light interception by loblolly pine.















CHAPTER 3
SEASONAL DYNAMICS OF FOLIAR NUTRIENTS, NUTRIENT USE
EFFICIENCY, AND RETRANSLOCATION IN JUVENILE LOBLOLLY AND
SLASH PINE

Introduction

Nutrient levels in the soil and plant are primary determinants of biological

productivity in forest stands. Differences in nutritional physiology (i.e., efficiencies of

nutrient uptake, nutrient utilization, and nutrient retranslocation) contribute to the

contrasting relative growth rates among species and their ecological responses to

environments (Boerner 1984; McGraw and Chapin 1989). A better understanding of

nutritional characteristics and their relations is central to the improvement of forest

productivity and awareness of ecosystem functioning (Baruah and Ramakrishnan 1988;

Knops et al. 1997). In practice, recognition of nutrients in relation to growth performance

among species or cultivars can help design conceptual ideotype models for different

objectives in breeding programs (Forrest and Ovington 1971; Mehall et al. 1983; Nambiar

1984).

Nutrient levels vary temporally in response to growth requirements and annual

physiological cycles (Drossopoulos et al. 1996; Santa et al. 1997). They also vary spatially

in adaptation to environmental conditions (e.g., soil fertility, weather) (Miller 1966; Boerner

1985; 1986). Our knowledge of nutrient dynamics for species having multiple cohorts of

leaves primarily comes from nutrient comparisons among different leaf age groups sampled









55

in the same year. However, patterns of major leaf nutrients (N and P) are not always closely

related in different years (Insley et al. 1981), suggesting that non-standardized sampling

procedures may provide inaccurate information or varying interpretations of nutrient

dynamics. Therefore, more attention should be paid to the dynamic patterns of nutrients

throughout a complete life cycle of the same cohort of leaves.

Nutrient-use efficiency has been most commonly defined in terms of biomass

production per unit of nutrient uptake (Gholz et al. 1985; Day 1987; Elliott and White 1993),

litterfall production per unit of litterfall nutrient content (Vitousek 1982; Knops et al. 1997;

Fassnacht and Gower 1999), or litter produced per unit of nutrient uptake (Garkoti and

Singh 1995). The primary consideration in the above definitions is to evaluate ecosystem

functioning. Other definitions of nutrient use efficiency encompass physiological aspects

of individual tree growth (Kost and Boerner 1985; Reich and Schoettle 1988; Li et al.

1991b). From the standpoint of nutritional and production physiology, absorbed nutrients

should meet the demand for foliage development, which in turn affects interception of light

energy and overall growth. Therefore, nutrient use efficiency can be defined as the

cumulative nutrientuse efficiency of dry-matter production, i.e., biomass production per unit

of incorporated (uptake + retranslocation) nutrient content (Larcher 1995). Nutrient-use

efficiency in terms of leaf area production per unit of nutrient content can serve as the basis

for evaluating both nutrient utilization and production efficiency. At present, little

information is available about genetic controls on nutrient use efficiency in relation to

strategies used by trees to achieve site dominance in growth.









56

Nutrient retranslocation has been interpreted as mechanisms ranging from increasing

plant adaptation to environments to more efficient utilization of nutrients (Nambiar and Fife

1991). Many studies have reported nutrient withdrawals from senescing leaves to young

tissues; however, retranslocation efficiency is not related to soil fertility in many species

(Chapin and Kedrowski 1983). Nutrient retranslocation not only occurs in senescing leaves,

but substantial amounts can also be retranslocated from young leaves throughout the year

for meeting growth requirements within the plant (Nambiar 1990). For example, foliar N

and P content retranslocated above 40% prior to senescence in many woody species (Reich

et al. 1995; Zhang and Allen 1996). Retranslocation of nutrients among growing leaves is

probably driven by growth requirements, i.e., nutrient retranslocation from young foliage

is closely associated with foliage production (Fife and Nambiar 1984). Strong evidence

indicated that N retranslocation was significant in the spring when active shoot elongation

was occurring. However, N, P, and K retranslocation was not significant in the fall in Acer

freemanii (Rose and Biernacka 1999). On the other hand, nutrient retranslocation during

senescence is a characteristic of many woody plants, and is controlled by many factors

(Nambiar and Fife 1991). Therefore, nutrient retranslocation efficiency and amount could

contribute to differences in overall growth performance of trees.

In this study, we consider nutrient issues related to growth strategies of individual

trees. Our objectives are to (1) determine patterns of leaf macronutrient (N, P, K, Mg, and

Ca) concentrations and content throughout a complete leaf life cycle as influenced by

locations and silvicultural treatments; (2) evaluate nutrient use efficiency of leaf area

production among several pine taxa; and (3) ascertain variation and significance of nutrient









57

retranslocation efficiency prior to leaf senescence with location, fertilizer treatment, and

taxa.


Materials and Methods

Plant Materials and Field Sites

Loblolly pine (Pinus taeda L.) and slash pine (P. elliottii Engelm. var. elliottii)

represent two of the most important commercial timber species in the southeastern United

States. Long-term studies of the two species have established clear interspecific relations

and intraspecific family structures. Three pine taxa (genetically improved loblolly,

improved and unimproved slash pine) were chosen for this study.

This study was part of an existing larger series of experiments designed by the

University of Florida's Cooperative Forest Genetics Research Program to test growth

performance of several pine taxa and their hybrids in relation to location and intensity of

silvicultural treatments (Lopez-Upton 1999). The two field sites utilized were in north

central Florida (Dunnellon, Levy County, 29020' N, 82050' W and Palatka, Putnam County,

29040' N, 81 42' W). The climate for each location is humid, temperate with a mean annual

temperature of 21C. Annual precipitation averages 1,332 mm at Dunnellon and 1,368 mm

at Palatka. The nearly level landscape is underlain by soils classified as sandy, siliceous,

hyperthermic Aeric Alaquods (somewhat poorly drained, Smyrna series) at Dunnellon and

hyperthermic, uncoated Aquic Quartzipsamments (moderately well drained, Adamsville

series) at Palatka (Soil Survey Staff 1998). The site indices for the Smyrna and Adamsville

series were 19 m and 20 m (base age 25 years), respectively.












Experimental Layout

Sixteen open-pollinated families from each of three pine taxa (genetically improved

loblolly pine, and improved and unimproved slash pine) were planted at both sites in a five-

tree row plot in each of three complete blocks using a split-split plot experimental design.

Two levels of silvicultural treatments (intensive vs. non-intensive) were applied. Prior to

study establishment, each site was chopped and bedded. Understory vegetation in the

intensive silvicultural treatment blocks was controlled during the first growing season using

a combination of mechanical and pre- and post-plant directed spot spray applications of

glyphosate applied at labeled rates. Containerized seedlings were planted in December 1994

at a 1.5 m x 3.4 m spacing at Palatka, and a 1.8 m x 3.0 m spacing at Dunnellon. Fertilizers

were broadcast applied in the high intensity treatment during years 1 and 3 as a balanced mix

of macro- and micronutrients. Total elemental application rates for plots receiving fertilizer

additions at both locations were approximately (kg ha-1): N (110), P (80), K (162), Ca (20),

Mg (10), S (13), Fe (0.5), Zn (0.06), Mn (0.5), Cu (0.06), and B (0.06). Insecticides (Asana,

Diomethorate or Pyridine) were applied 3-4 times during the first growing season to control

tip moth (Rhyacionia spp.) on the high intensity treatment. The low intensity treatment did

not receive herbicide, fertilizer or insecticide applications. An untreated buffer of at least

21 m separated the high and low intensity treatment.


Sampling Procedures

Two sample trees within a 5-tree row-plot in each family from each block were

randomly chosen by a SAS procedure, and then a systematic sampling method was applied









59

to all other families and taxa at the two sites. Sample trees were healthy and free of disease.

In total, 192 sample trees (2 treatments x 3 blocks x 16 families x 2 trees) were chosen for

each taxa and site. Overall, 1,152 trees (2 locations x 2 treatments x 3 blocks x 3 taxa x 16

families x 2 trees) were sampled across the two sites.

The specific tissues sampled in this study and the timing of collections were

consistent among sample trees to avoid likely variation in nutrient concentrations (Bates

1971). Recommended tissue sampling procedures entailed collection of full-length and

complete fascicles from the upper third of the crown from representative sample trees

(Madgwick and Mead 1990). Because foliage at different ages can vary in nutrient

concentration, sample tissues should be age specific (Hom and Oechel 1983; Ernst 1995;

Zhang and Allen 1996). Generally, current-year foliage has higher nutrient concentration

and lower tree-to-tree variability than older needles and, therefore, it is widely accepted as

being most useful for diagnostic purposes (Mead 1984).

Previous research with loblolly and slash pine has shown that foliage nutrients

exhibit distinct temporal patterns over the course of a year (CRIFF 1987). To overcome

problems with leaf age and season of year, needle samples were collected eight times over

a two year period from the same branch of every sample tree through the life cycle of the

same needle cohort. Specifically, needle samples were collected from both sites in: June

(the first month current year foliage attains full length), September (when needle N

concentration is generally lowest during the year), November (a critical stage in

retranslocation of some mineral nutrients) in 1997, and February (when needle N

concentration is generally at the highest level of a year), April (initial new growth may affect









60

nutrient status of 1-year-old needles), June, September, and December (last month for 1-

year-old needles to stay on the tree) in 1998. At each sampling interval, about 50 fascicles

were collected from each sample tree. Sample branches were randomly selected from the

upper third of the crown in 1997 (the crown position for needle samples became the middle

crown in 1998) from a uniform aspect (south). Approximately 9,216 total leaf samples (2

locations x 2 treatment x 3 blocks x 3 taxa x 16 families x 2 trees x 8 times) were processed

for chemical analyses.

All tissues were oven-dried at 70 C for 48 hours or until dry. About 20 complete

and full-length fascicles were randomly chosen from each sample to determine total dry

weight and the number of needles per fascicle. All dried tissues were ground in a Wiley mill

to pass a 2 mm stainless steel screen. The ground dry samples were stored in sealed plastic

vials until nutrient analyses were performed.


Nutrient Analyses

Selection of an efficient nutrient digestion method from the many established

procedures depends upon the nutrient status of plant materials, which is critical with respect

to N. Conifers naturally grow on acid soils where ammonification is the dominant N

conversion process (Sarigumba et al. 1977; Pritchett and Smith 1970), and NH4+ can be as

high as 90% in the mineral soil N pool (Carlyle 1995). Pine trees prefer NH4+ as the primary

N source from soils (Durzan and Steward 1967; McFee and Stone 1968), largely due to their

long-term adaptation to acidic soil environments. Two methods for determining total N in

pine foliage were analyzed and compared on a sample subset to determine accuracy and N









61

recovery. A total of 200 samples that included improved loblolly, improved slash and

unimproved slash pine that were equally and randomly chosen from the sampling periods

were used in the analysis.


Method I: Kjeldahl digestion

To determine foliar N concentrations, a 200 mg sample was weighed into a 50 ml

Pyrex test tube, and then 3.2 g of salt catalyst (9:1 K2SO4: CuSO4), 2 glass beads, and 5 ml

of concentrated H2SO4 were vortexed in the tube under a hood. Two ml of 30% H202 was

added to reduce frothing. Tubes were digested in an aluminum block digester at 380 C for

240 minutes (Bremner 1965; Gallaher et al., 1975; Jones et al. 1991). The tubes were

capped with small Pyrex funnels which allowed for evolving gases to escape while

preserving refluxing action. Cool digested solutions were vortexed with approximately 20

ml of deionized water and allowed to cool to room temperature. Samples were then brought

to a 50 ml volume, transferred to 20 ml square Nalgene storage bottles (glass beads were

filtered out), sealed, mixed, and stored. Nitrogen that was trapped as (NH4)2 SO4 was

analyzed. Eight standard pine materials with known N concentration values from National

Standard Institute (NSI) were subjected to the same procedures and used as checks.


Method II: wet acid digestion

Needle N concentrations were measured using the method as outlined in Thomas et

al. (1967) and Jones et al. (1991), which was similar to the Kjeldahl procedure except that

a catalyst was not added. In brief, 100 mg of homogenous tissue was weighed and placed

in a 50 ml Pyrex test tube, and then 2 ml of concentrated H2SO4 was added. The samples









62

were placed in a digestion block at 380 C for 30 minutes. All tubes were then removed

from the block and allowed to cool. Small amount of 30% H202 was added into the tubes.

Repeated heating and cooling was conducted several times until the solution became clear.

In this method, H2SO4 was added to raise the temperature of the mixture, while H202 was

used to speed and complete the digestion procedure (Jones et al. 1991). The other

macronutrients (P, K, Mg, and Ca) were analyzed using the same digestion procedure.

Nutrient concentrations were determined using an inductively-coupled plasma emission

spectrometer (ICP or ICAP).

Pairwise sample comparisons for N concentrations indicated that Method I was about

7% lower than Method II in estimating total N concentration, but no significant differences

were found between the two methods. To process samples more efficiently, Method II was

selected for analyzing all plant tissues. A detailed discussion on advantages and

disadvantages of both methods was given by Jones et al. (1991).


Nutritional Variables

Concentrations and fascicle content of N, P, K, Mg, and Ca, and fascicle dry weight

over the eight sample periods were included in the statistical analysis. Fascicle nutrient

content was calculated as the product of nutrient concentration and average fascicle weight.

Nutrient use efficiency of leaf area production (LAm) was defined as peak leaf area

production per unit nutrient accumulated in current-year foliage and expressed as leaf area

(cm2 ) / nutrient (mmol). LA, was determined using foliar nutrient concentration and

specific leaf area estimates from September,1997, with a sample size of 1,152 for that









63

period. Because only current-year needles were analyzed for nutrient concentrations in

1997, LAE was computed for current-year foliage only. This index measures peak leaf

area production (generally from late August to early September) in terms of total amount of

nutrients incorporated in current-year foliage that are available for metabolism.

Nutrient retranslocation efficiency (NRE) was calculated using the following formula

(Saur et al. 2000):

FCi FC2
NRE (%) = x 100
FC1


where FC, was the maximum fascicle nutrient content during the needle cohort life cycle,

and FC2is nutrient content of green fascicles in early December, 1998, prior to abscission.

The term (FC, FC2) represented the amount of nutrients that were retranslocated. The

sample size used in the analyses for nutrient retranslocation totaled 1,152.


Statistical Analyses

All analyses of variance and comparisons of means were conducted using individual

tree data. Model selection procedures and criteria were similar to those described in Chapter

2 for growth analyses. Main effects under investigation included location, treatment, taxa,

and family, in which all effects except family were regarded as fixed effects. In brief, a full

model including all main effects and their interactions (significant at =0.25) was chosen

for preliminary analysis. Non-significant effects and interactions were deleted from the full

model, and a final model was then developed. Interaction effects involving family were

always kept in the model for appropriate selection of error terms in ANOVA tests, even









64

though they were not significant. SAS procedures GLM and MIXED were used for

ANOVA (SAS Institute 1996). Where variance homogeneity was not satisfied, ANOVA

was separately performed by locations or treatments. For fascicle nutrient concentrations

and content, ANOVA was performed separately by sampling periods.

Means for the various nutrient characteristics (i.e., concentration, content, LAu,

and nutrient retranslocation efficiency) among the three taxa were compared using the

LSMEANS statement in PROC MIXED. A default level of = 0.05 was used to test

significance among the means unless otherwise specified. In presenting the data means

were combined across locations or treatments if scale effects (i.e., no rank changes among

the taxa) were detected.


Results

Variation of Leaf Nutrient Concentrations and Fascicle Weight

Nitrogen, P, and K concentrations generally decreased over a complete leaf life cycle

among the three pine taxa (Figures 3-1 to Figure 3-3). The decrease was less pronounced

for P at the Dunnellon site, while consistent decreases occurred in N and K for both

locations and treatments. In contrast, concentrations of less mobile elements, Ca and Mg,

generally increased from the beginning to the end of the leaf life cycle (Figure 3-4 and

Figure 3-5).

Differences in N and P concentrations were consistent among taxa across locations

and treatments, with loblolly pine having significantly higher concentrations than slash pine.

For example, loblolly pine had significantly higher N concentrations than either slash pine

taxa in 7 out of 8 sampling periods at Palatka, regardless of silvicultural treatment (Figure
















2.0


^0
S1.5 -


-<-
1.0
o -


0.5

-

0.0




2.0



S1.5-



o 1.0
o


0
.0 0.5



0.0


Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998


I I I I I I I I
Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998
Time


Dunnellon
Non-intensive
6


Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998


Time Time
Figure 3-1. Variation in needle N concentration for improved loblolly, improved slash
and unimproved slash pine throughout a life cycle of a needle cohort managed under
two silvicultural treatment and two locations in north central Florida.
PTA = improved loblolly pine, PEE = improved slash pine, PEU = unimproved slash pine.
Integers listed below locations and treatments indicate the number of sampling periods out
of eight where significant taxa differences were detected at the 95% confidence level using
LSMEANS test of the MIXED procedure.


Palatka
Intensive
7


Palatka
Non-intensive
7


Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998
Time


Dunnellon
Intensive
6

















0.20


^0

= 0.15

')
U-
5 0.10 -
U


-cr

-c


0.00




0.20


^0

= 0.15

')
U-
5 0.10
U


-c
S0.05



0.00
0.00 -


I I I I I I I I
Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998
Time


I I I I I I I I
Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998
Time


Dunnellon
Non-intensive
8


Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998
Time


Figure 3-2. Variation in needle P concentration for improved loblolly, improved slash
and unimproved slash pine throughout a life cycle of a needle cohort managed under
two silvicultural treatment and two locations in north central Florida.
PTA = improved loblolly pine, PEE = improved slash pine, PEU = unimproved slash pine.
Integers listed below locations and treatments indicate the number of sampling periods out
of eight where significant taxa differences were detected at the 95% confidence level using
LSMEANS test of the MIXED procedure.


-.- PTA
--- PEU
-y- PEE


Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998
Time


Dunnellon
Intensive
7














1.2
Palatka Palatka
1.0 Intensive Non-intensive
2 7

0.8 -


8 0.6


2 0.4
S- PTA
o -o-- PEU
S0.2 PEE


0.0
Jun Sep Nov Feb Apr Jun Sep Dec Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998 1997 1998
Time Time
1.2
Dunnellon Dunnellon
1.0 Intensive Non-intensive
0 3

0.8 -


S0.6

2 0.4


0.2


0.0
Jun Sep Nov Feb Apr Jun Sep Dec Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998 1997 1998
Time Time
Figure 3-3. Variation in needle K concentration for improved loblolly, improved slash
and unimproved slash pine throughout a life cycle of a needle cohort managed under
two silvicultural treatment and two locations in north central Florida.
PTA = improved loblolly pine, PEE = improved slash pine, PEU = unimproved slash pine.
Integers listed below locations and treatments indicate the number of sampling periods out
of eight where significant taxa differences were detected at the 95% confidence level using
LSMEANS test of the MIXED procedure.















0.8
Palatka Palatka
Intensive Non-intensive
0.6 0
-e- PTA
-- PEU
PEE
0.4







0.0
Jun Sep Nov Feb Apr Jun Sep Dec Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998 1997 1998
Time Time
0.8
Dunnellon Dunnellon
Intensive Non-intensive
6 6
S0.6



S0.4



I 0.2



0 .0 I... I I I I
Jun Sep Nov Feb Apr Jun Sep Dec Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998 1997 1998
Time Time
Figure 3-4. Variation in needle Ca concentration for improved loblolly, improved slash
and unimproved slash pine throughout a life cycle of a needle cohort managed under
two silvicultural treatment and two locations in north central Florida.
PTA = improved loblolly pine, PEE = improved slash pine, PEU = unimproved slash pine.
Integers listed below locations and treatments indicate the number of sampling periods out
of eight where significant taxa differences were detected at the 95% confidence level using
LSMEANS test of the MIXED procedure.
















0.12

-
S0.10 -

-
0.08

o
-
S0.06

S0.04
Ca
-












0.12

0.02
0.10

2 0.08 -


o 0.06

o

S0.02 -




0.00


Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998


Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998
Time


Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998


Time Time
Figure 3-5. Variation in needle Mg concentration for improved loblolly, improved slash
and unimproved slash pine throughout a life cycle of a needle cohort managed under
two silvicultural treatment and two locations in north central Florida.
PTA = improved loblolly pine, PEE = improved slash pine, PEU = unimproved slash pine.
Integers listed below locations and treatments indicate the number of sampling periods out
of eight where significant taxa differences were detected at the 95% confidence level using
LSMEANS test of the MIXED procedure.


-.- PTA
-o- PEU
--- PEE


Palatka
Intensive
3


I I I I I I I I
Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998
Time


Palatka
Non-intensive


Dunnellon
Intensive
6


Dunnellon
Non-intensive









70

3-1). Differences in nutrient concentrations for N and P between improved and unimproved

slash pine were generally non-significant (Appendix D). Less significant differences among

the three taxa occurred in K concentrations, except for the non-intensive treatment at the

Palatka site (Figure 3-3). Differences among the three taxa and within a taxon were also

highly variable for Ca and Mg concentrations. Each taxon had high concentrations in Ca

and Mg at some period over the course of the experiment (Appendix E), although loblolly

pine tended to have consistently lower Ca concentrations than slash pine at Dunnellon for

both the intensive and non-intensive silvicultural treatments.

Both treatment and location generally affected the foliar concentrations of all the

macronutrients except for P, where only minor treatment differences were found in 2 out of

8 sampling periods (Appendix D). Location x treatment interactions for foliar

concentrations were significant under most sampling periods, showing differential responses

among taxa to treatments across locations. Treatments generally did not significantly

influence nutrient concentration differences between loblolly and slash pine. The significant

treatment x taxa interactions were caused by differential treatment responses between

improved and unimproved slash pine, with improved slash pine having lower nutrient

concentrations under the non-intensive treatment, but higher concentrations under the

intensive treatment compared to unimproved slash pine (Appendix E).

Differences in environmental factors at the two locations also significantly influenced

nutrient concentrations in all taxa. Loblolly pine had consistently higher N and P

concentrations than slash pine; thus, the significant location x taxa interactions for N and P

were primarily caused by rank changes between improved and unimproved slash pine.









71

However, significant location x taxa interactions in K, Ca, and Mg were involved with rank

changes among all three taxa. Loblolly pine had lower concentrations than slash pine for

these nutrients during certain sampling periods, and was especially pronounced for Ca

(Figure 3-3 to Figure 3-5).

A noticeable change in nutrient concentrations occurred during the final stage of the

leaf life cycle, where concentrations of each element among the three taxa tended to

converge within a location and treatment. ANOVA also showed that location x taxa

interactions and treatment x taxa interactions for all elements (except in Mg) generally

became non-significant during the later portion of the leaf life cycle among the three taxa

(Appendix D and E).

Fascicle weight of loblolly pine was significantly lower than slash pine, regardless

of locations and treatments; an exception occurred during the second sampling period for

the intensive treatment at Dunnellon (Figure 3-6). Differences between improved and

unimproved slash pine in fascicle weight were not significant (Appendix E). The intensive

treatment significantly increased fascicle weight in all taxa, but did not change the rank

between loblolly and slash pine. The significant treatment x taxa interactions found for

some sampling periods were caused by interactions between improved and unimproved slash

pine. Locations also exerted significant influence on fascicle weight, with 7 out of 8 periods

showing statistically significant differences between the two experimental locations

(Appendix D). Variation in local environments between locations also significantly induced

significant location x taxa interactions in 5 out of 8 sampling periods. Additionally,

treatment effects on fascicle growth were significantly different across locations in 7









72

sampling periods (location x treatment interactions were significant). Another noticeable

and irregular change over the course of fascicle development occurred in the non-intensive

treatment at the Palatka site, where fascicle weight decreased markedly in June 1998, more

or less corresponding to the decrease in Ca and Mg concentrations during the same time

period (Figures 3-4 to Figure 3-6). Severe drought conditions at Palatka from March to

June, 1998 could have contributed to decreased fascicle growth under the non-intensive

treatment where the crowns had not yet closed. Total rainfall during that period was only

2.5 mm, or about 6% of the precipitation typically received in a normal year. Trees grown

under the intensive treatment may have avoided declines in fascicle growth by shedding

lower branches (shaded) to compensate for low soil water availability.

Nutrient dilution can occur in plants when nutrient supply rates cannot match overall

biomass accumulation rates. The lowest levels of foliar Mg and Ca concentrations were

detected in the intensive treatment in September, following the end of the major portion of

the growing season. Especially apparent were large differences in Mg concentrations

between the intensive and non-intensive treatments for loblolly pine (Figure 3-7).

Treatments obviously decreased needle Mg concentrations for all taxa, but were more severe

in loblolly pine than in slash pine. For example, Mg concentrations for loblolly pine for the

intensive treatment were well below the minimum critical level (0.07%) (Jokela et al. 1991),

while Mg concentrations for slash pine were above the minimum levels. Loblolly and

improved slash pine had lower Mg concentrations under both treatments at Palatka than at

Dunnellon, while the reverse was true for unimproved slash pine. The dynamics of fascicle

nutrient content over the leaf life cycle showed mixed patterns (Appendix F). Loblolly pine















250


200


150


100


50


Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998
Time


Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998
Time


Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998
Time


Figure 3-6. Variation in average fascicle weight for improved loblolly, improved slash
and unimproved slash pine throughout a life cycle of a needle cohort managed under
two silvicultural treatment and two locations in north central Florida.
PTA = improved loblolly pine, PEE = improved slash pine, PEU = unimproved slash pine.
Integers listed below locations and treatments indicate the number of sampling periods out
of eight where significant taxa differences were detected at the 95% confidence level using
LSMEANS test of the MIXED procedure.


Palatka
Intensive
8


Palatka
Non-intensive
8


-e- PTA
-o- PEU
--- PEE


Jun Sep Nov Feb Apr Jun Sep Dec
1997 1998
Time


0




250
-

-
c 200 -


150
-


100
5r
c-

^ 50-
<


Dunnellon
Intensive
7


Dunnellon
Non-intensive
8














0.10

-.- P PTA
0.09 -o- P PEE
-- PPEU
-v- D PTA
O 008 -- D PEE
-.- D PEU

I 0.07 Loblolly


1 0.06 -


0.05- Slash


0.04
0 20 40 60 80
Total leaf area per tree (m2/tree)
Figure 3-7. Relationships between total leaf area at age 3 years and September
Mg concentration for genetically improved loblolly pine (PTA), improved slash (PEE)
and unimproved slash pine (PEU) when managed under two silvicultural treatments
at two locations in north central Florida. Symbols with higher Mg concentrations (i.e.,
lower leaf area) were from the non-intensively managed treatment, while those with
lower Mg concentrations (i.e., higher leaf area) were from the intensively-managed
treatment. The horizontal dashed lines represent critical (minimum) foliar Mg
concentrations for loblolly pine (0.07%) and slash pine (0.05%). P = Palatka site,
D = Dunnellon site.









75

generally had significantly lower fascicle nutrient content than slash pine for all elements

at different sampling stages. Fascicle content for N, P, and K generally increased overtime,

peaked, and then decreased (N) or was maintained (P, K) at a steady state in all taxa.

Magnesium content showed a similar pattern but peaked much later (after April of second

growing season). On the contrary, Ca steadily increased in the fascicles, with the content

reaching a maximum before leaf abscission (Appendix E). Analysis of variance was also

performed on fascicle nutrient content, and results were similar to those revealed for nutrient

concentrations and fascicle weight (Appendix D).


Crown (Leaf) Nutrient Content

Crown nutrient content depends on the accumulation of foliage biomass and leaf

nutrient concentration. The amount of nutrients stored in the crown can affect crown

development and growth performance. Significant effects of locations and treatments on

crown nutrient content were found for all elements among the three taxa (Table 3-1). Trees

grown at Palatka tended to accumulate significantly greater amounts of nutrients in the

crown than those at Dunnellon when managed under the same silvicultural treatment (Table

3-2).

Differences among taxa in crown nutrient content were statistically significant under

most circumstances (Table 3-2). Loblolly pine accumulated significantly greater quantities

of all nutrients than slash pine, except under the non-intensive treatment at Dunnellon.

Although differences in crown nutrient content between improved and unimproved slash

pine were not significant, improved slash pine consistently had higher absolute amounts

compared to unimproved slash pine.









76

Table 3-1. ANOVA for crown nutrient content (g/tree) of loblolly and slash pine at age 3
years. Experimental trees were subjected to two levels of silvicultural treatments and
planted at two locations in north central Florida".
Source of Variation N P K Mg Ca

Location *** *** *** *** ***

Treatment *** *** *** *** ***

Taxab:

PTA vs. PEE *** *** *** *** **

PEE vs. PEU ** ** ** *** **

Location*treatment *** *** *** *** ***

Location*taxa *** *** *** NS NS

Treatment*taxa ** ** **

Family(taxa) NS NS NS NS NS

Block(treatment) NS NS NS NS **

Location*family(taxa) NS ** NS NS NS

Treatment*family(taxa) NS NS NS NS NS

Taxa*block(treatment) *** *** *** *** NS

Location*treatment*taxa *** *** *** ** NS

Location*treatment* NS
family(taxa)
a For a given source of variation, main effects and interactions were significant at ***
p 0.01, ** p 0.05, p 0.10. NS= not significant.
b PTA = improved loblolly pine
PEE = improved slash pine
PEU = unimproved slash pine












Table 3-2. Nutrient content (g/tree) in the crowns (foliage) of 3-year-old loblolly and slash pine managed under two silvicultural
treatments and planted at two locations in north central Floridad.
Location Dunnellon Palatka

Treatment Non-intensively Intensively managed Non-intensively Intensively managed
managed managed
Taxa PTAb PEE PEU PTA PEE PEU PTA PEE PEU PTA PEE PEU

ElementC:

N 7.9a 5.7a 5.7a 30.4a 17.lb 12.4b 22.6a 11.6b 10.2b 57.0a 39.5b 23.0c

P 1.la 0.6b 0.5b 3.6a 1.7b 1.lb 3.4a 1.6b 1.2b 8.la 4.0b 2.4b

K 2.6a 2.5ab 2.0b 19.6a 10.2b 7.7b 8.1a 5.9ab 4.1b 35.3a 24.6b 14.7b

Mg 0.9a 0.7ab 0.5b 1.6a 1.lb 0.8b 1.4a 1.0b 0.9b 2.3a 2.0a 1.4b

Ca 1.6a 1.4a 1.2a 4.3a 3.7ab 2.8b 4.5a 4.0a 3.5a 8.2a 6.9ab 6.0b
a Taxa means were analyzed by location and treatment separately. Means among the three taxa for a given variable
followed by the same letter were not statistically significant at the 95% confidence level by comparing least square means
using the MIXED procedure (SAS Institute 1996).
b PTA = improved loblolly pine
PEE = improved slash pine
PEU = unimproved slash pine
c Estimates of crown nutrient content were made in September, 1997.









78

Differences among taxa in crown nutrient content tended to be smaller under the

non-intensive treatment than under the intensive treatment at both sites. Therefore, genetic

differences among taxa were better expressed under the intensive treatment than under the

non-intensive treatment, which led to the significant treatment x taxa interactions for all

nutrients (Table 3-1). For example, crown N content at Dunnellon quadrupled, tripled, and

doubled under the intensive treatment in comparison to the non-intensive treatment for

loblolly, improved and unimproved slash pine, respectively (Table 3-2). Loblolly pine

accumulated larger quantities of nutrients in the crown than slash pine under the intensive

treatment, while differences among taxa were less pronounced under the non-intensive

treatment at both sites for all elements.


Nutrient Use Efficiency of Leaf Area Production

Significant interspecific (species) and intraspecific (family) differences in LANU

were detected among the three taxa (Table 3-3). This index was calculated based on the leaf

area development per unit nutrient accumulated from ages 3 to 4 years. Loblolly pine was

significantly more efficient in using N, K, Mg, and Ca to develop leaf area than slash pine,

while differences between improved and unimproved slash pine were generally not

significant except for K use efficiency. Unimproved slash pine was most efficient in P use

efficiency (Figure 3-8).

The intensive silvicultural treatment significantly decreased LANu for N, P, and K

in all taxa, regardless of differences in locations (Table 3-3). The most significant decrease

was found in K use efficiency. When averaged across locations and taxa, K use efficiency

decreased markedly from 19.2 cm2 /mmol under the non-intensive treatment to 9.6









79

Table 3-3. ANOVA for LAu (cm2 leaf area / mmol element) for loblolly and slash pine
at age 3 years. Experimental trees were subjected to two levels of silvicultural treatments
and planted at two locations in north central Floridad.

Source of Variation N P K Mg Ca

Location

Treatment *** *** *** NS

Taxab:

PTA vs. PEE *** *** *** ***

PEE vs. PEU NS *** *** NS NS

Location*treatment *** *** *** *** ***

Location*taxa NS *** NS *** ***

Treatment*taxa NS NS *** NS

Family(taxa) ** NS ** NS

Block(treatment) NS ** NS NS

Location*family(taxa) NS NS NS NS NS

Treatment*family(taxa) NS NS NS NS NS

Taxa*block(treatment) *** *** ** *** ***

Location*treatment*taxa *** **** NS *

Location*treatment* *
family(taxa)
a For a given source of variation, main effects and interactions were significant at ***
p 0.01, ** p 0.05, p 0.10. NS= not significant.
b PTA = improved loblolly pine
PEE = improved slash pine
PEU = unimproved slash pine









80

cm2 /mmol under the intensive treatment. However, taxa responses to the silvicultural

treatments were similar in N use efficiency, resulting in non-significant treatment x taxa

interactions. Significant treatment x taxa interactions were detected in K use efficiency, but

further analysis showed a scale effect instead of a rank change among taxa. For less mobile

elements, the intensive silvicultural treatment significantly increased LAU in Mg, but not

in Ca.

LAuE was significantly influenced by location and location x treatment interactions

(Table 3-3). The Dunnellon site had higher nutrient use efficiencies for all elements except

Mg. For example, trees across all three taxa at Dunnellon were 30%, 47%, 45%, and 39%

higher than those at Palatka for N, P, K, and Ca use efficiency for leaf area production,

respectively. Treatments also showed varied influence on LAUE, at different locations, as

location x treatment interactions were highly significant for all elements. Apparently,

nutrient use efficiencies for all elements among the three taxa were sensitive to

environments, as also shown by the significant location x treatment x taxa and location x

treatment x family(taxa) interactions.


Fascicle Nutrient Retranslocation Efficiency

Significant differences occurred in N, P, and K retranslocation efficiencies prior to

senescence among taxa, while no differences were found in Mg and Ca retranslocation

efficiencies (Figure 3-9). For instance, loblolly pine retranslocated about 45% of the fascicle

N prior to abscission, while slash pine only retranslocated 28% during the same period.

Differences between improved and unimproved slash pine in nutrient retranslocation were

















^3 70
a

I a 60

b b b b b
--- 50
E 2 a a
U a
I 40

b
30 b
0 30 b


20 a
b c

10


S0 0
PTA PEE PEU PTA PEE PEU PTA PEE PEU PTA PEE PEU PTA PEE PEU
N P K Mg Ca
Taxa and elements
Figure 3-8. Nutrient use efficiency for leaf area development for genetically improved loblolly pine (PTA), improved slash
(PEE) and unimproved slash pine (PEU) when managed under two silvicultural treatments at two locations in north central
Florida. Means among taxa for a given nutrient followed by the same letter were not statistically significant at the 95%
confidence level using the LSMEANS test of PROC MIXED. Note the different scales between N and other elements.












82

and unimproved slash pine in nutrient retranslocation were not significant, except for N.

Large differences among taxa were found in K retranslocation efficiency, where loblolly

pine retranslocated 48% of K, in comparison to 21% for improved and unimproved slash

pine. Some loss of K from leaching could occur, which would lead to higher K

retranslocation estimates.

The silvicultural treatments had no significant effect on nutrient retranslocation

efficiencies for all taxa (Table 3-4). Differences among taxa for retranslocation efficiencies

were primarily affected by location. The Palatka site had higher levels of nutrient

retranslocation efficiency for N, P, K, and Ca than Dunnellon, while the opposite was true

for Mg. Significant location x treatment interactions further indicated that location effects

on nutrient retranslocation efficiencies (except K) were different across treatments. For

example, N retranslocation efficiency in loblolly pine was higher under the intensive

treatment (43%) than under the non-intensive treatment (35%) at Dunnellon, while it was

higher under the non-intensive treatment (54%) than under the intensive treatment (48%)

at Palatka. Slash pine showed a similar trend. In addition, although significant location x

taxa interactions were found in retranslocation efficiencies for N, P, and Mg, the mode of

influence was different. Scale effects were found in N retranslocation, with loblolly pine

having the highest retranslocation efficiency and unimproved slash pine having the lowest.

For P retranslocation, loblolly pine consistently had higher efficiencies than slash pine, while

rank changes occurred between improved and unimproved slash pine. In Mg

retranslocation, the highest efficiencies were found in each taxon at different locations and















50 a
-, a


40
5 40 -





ca
(U

a bb
2 30 a
a a


20 b b
aa a









0
PTA PEE PEU PTA PEE PEU PTA PEE PEU PTA PEE PEU PTA PEE PEU
N P K Mg Ca
Taxa and elements
Figure 3-9. Nutrient retranslocation efficiency from fascicles of a single needle cohort prior to senescence for
genetically improved loblolly pine (PTA), improved slash (PEE) and unimproved slash pine (PEU) when managed
under two silvicultural treatments at two locations in north central Florida. Means among taxa for a given nutrient
followed by the same letter were not statistically significant at the the 95% confidence level using the LSMEANS
test of the MIXED procedure.












Table 3-4. ANOVA for nutrient retranslocation efficiency (%) and amount retranslocated (mg / fascicle) prior to
senescence for a single cohort of needles in loblolly and slash pine at ages 3 to 4 years. Experimental trees were
subjected to two levels of silvicultural treatments and planted at two locations in north central Floridaa.

Retranslocation efficiency (%) Retranslocation amount (mg/fascicle)
Source of variation
N P K Mg Ca N P K Mg Ca

Location *** *** NS *** ** *** *** *** *** ***

Treatment NS NS NS NS NS ** ** *** ** NS

Taxab: *** *** *** NS NS ** NS ** *

PTA vs. PEE *** *** *** NS NS NS NS NS ** *

PEE vs. PEU ** NS NS NS NS ** NS NS NS

Location*treatment *** ** NS *** *** *** NS NS *** NS

Location*taxa *** NS *** NS ** *** ** NS

Treatment*taxa NS NS NS *** NS NS NS NS *

Family(taxa) NS NS NS NS NS NS NS NS NS NS












Table 3-4 -- Continued.

Retranslocation efficiency (%) Retranslocation amount (mg/fascicle)
Source of variation
N P K Mg Ca N P K Mg Ca

Block(treatment) NS NS NS NS NS NS NS NS NS NS

Loc*family(taxa) NS NS NS NS NS NS NS NS NS NS

Treatment*family(taxa) NS NS NS NS NS NS NS NS NS NS

Taxa*block(treatment) *** *** *** *** *** *** *** ** *** ***

Location*treatment*taxa *** NS NS NS NS *** NS NS NS

Location*treatment ** NS NS NS NS
family(taxa)
a For a given source of variation, main effects and interactions were significant at *** p 0.01, ** p 0.05,
p 0.10. NS = not significant.
b PTA = improved loblolly pine
PEE = improved slash pine
PEU = unimproved slash pine




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs