soil water depletion, growth, physiology, and yield of carambola trees in krome soil

SOIL WATER DEPLETION, GROWTH, PHYSIOLOGY, AND YIELD OF

CARAMBOLA TREES IN KROME SOIL

By

RASHID AL-YAHYAI

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

2004

Copyright 2004

by

Rashid Al-Yahyai

To Vina, Amir, & AJ

ACKNOWLEDGMENTS

I would like to acknowledge the help and support of many individuals without

whom this work would not have been possible. My sincere gratitude goes to my major

professors, Drs. Bruce Schaffer and Frederick S. Davies, for their guidance and direction

of matters related to my academic performance and professional development. I deeply

appreciate the help and support of members of my supervisory committee Drs. Jonathan

H. Crane, Yuncong Li, and Thomas Obreza. I also thank the following individuals for

their assistance with the research project: Angel Colls, Osvany Rodriguez, Mark Kohout,

and Dr. Rafael Muñoz-Carpena. I am thankful to many friends who made my stay in

Florida most enjoyable.

urhasanah, Simon Raharjo, Pam Moon, Rita Duncan, Alicia Ally, and Maritza

Ojeda. I am grateful to my mother and siblings for their encouragement and support. Last

but not least, my love and sincere appreciation go to my wife, Dr. Divina M. Amalin, for

her endless patience, care, love, and support.

iv

TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................................................................................. iv

LIST OF TABLES............................................................................................................ vii

LIST OF FIGURES ........................................................................................................... ix

ABSTRACT..................................................................................................................... xiii

1 INTRODUCTION ........................................................................................................1

2 REVIEW OF LITERATURE.......................................................................................8

Carambola.....................................................................................................................8 Origin and Distribution..........................................................................................8 Economic Importance............................................................................................9 Botany and Uses ..................................................................................................10 Tree Phenology....................................................................................................14 Cultivation ...........................................................................................................15 Environmental Stress Physiology of Carambola.................................................23

Temperature .................................................................................................23 Wind .............................................................................................................25 Flooding .......................................................................................................26 Drought and water stress ..............................................................................26

Water Stress ................................................................................................................29 Effect of Water Stress on Vegetative Growth .....................................................32

Root growth and distribution........................................................................32 Leaf area .......................................................................................................32 Trunk growth................................................................................................33 Shoot growth ................................................................................................34

Effect of Water Stress on Reproductive Growth and Yield ................................34 Flowering and fruiting..................................................................................34 Yield .............................................................................................................36 Fruit quality ..................................................................................................36

Effect of Water Stress on Physiological Processes .............................................37 Net CO2 assimilation and transpiration........................................................37 Stomatal conductance...................................................................................38 Leaf and stem water potential ......................................................................39

Irrigation Scheduling of Fruit Trees ...........................................................................42

v

Irrigation scheduling based on plant water status ........................................43 Irrigation scheduling based on soil water status...........................................45

Conclusions.................................................................................................................50 �3 CALIBRATION OF THE NEUTRON PROBE AND MULTISENSOR

CAPACITANCE PROBES IN A GRAVELLY LOAM CALCAREOUS SOIL ......53 �

��Materials and Methods ...............................................................................................55 ��Results and Discussion ...............................................................................................59 ��Conclusions .................................................................. ................................ .............61

��4 MONITORING SOIL WATER CONTENT FOR IRRIGATION SCHEDULING IN

��A CARAMBOLA ORCHARD IN A GRAVELLY LIMESTONE SOIL .................68 �

��Materials and Methods ...............................................................................................71 ��Results and Discussion ...............................................................................................73

��5 ��HYSIOLOGICAL RESPONSES OF CARAMBOLA TREES TO WATER

��DEPLETION IN KROME VERY GRAVELLY LOAM SOIL.................................83 ��

��Materials and Methods ............................................................................... ................85 ��Results and Discussion .................................................................................... ...........88 ��Conclusions....................................................................................................... ..........91

��6 GROWTH, YIELD, AND FRUIT QUALITY OF CARAMBOLA AS AFFECTED

��BY SOIL WATER DEPLETION................................................................... ..........101 �

��Materials and Methods ............................................................................. ................102 ��Results and Discussion ........................................................................... ..................105 ��Conclusions............................................................................................ ...................112

��7��HENOLOGICAL CYCLES OF CARAMBOLA AT FOUR LEVELS OF SOIL

W��ATER DEPLETION IN SOUTH FLORIDA ...................................... .................127 �

��Materials and Methods .............................................................................................130 ��Results.......................................................................................................................133 ��Discussion.................................................................................................................137 C��onclusions.................................................................................................. ............141

��8 SUMMARY AND CONCLUSIONS.......................................................................149 ��APPENDIX �SOIL WATER DEPLETION, PHYSIOLOGY, GROWTH, AND YIELD,

��OF CARAMBOLA TREES IN KROME SOIL...............�......................152 ��LIST OF REFERENCES................................................................................................158 ��BIOGRAPHICAL SKETCH ..........................................................................................173

vi

LIST OF TABLES

Table page 2-1. Major carambola cultivars of economic importance in several major carambola

production regions....................................................................................................21

3-1. Fitted parameters of van Genuchten’s model used to describe soil water retention curves of Krome very gravelly loam soil where soil water content was measured in a carambola orchard using multisensor capacitance probes (EnviroSCAN) at soil depths of 10, 20, and 30 cm. ....................................................................................62

3-2. Fitted parameters of van Genuchten’s model used to describe soil water retention curves of Krome very gravelly loam soil where soil water content was measured in a carambola orchard using a neutron probe at soil depths of 10, 20, and 30 cm. ....63

4-1. Fitted parameters of van Genuchten’s model used to create a soil water retention curves for Krome soil. Soil water content was measured in a carambola orchard with a neutron probe and multisensor capacitance probes. ......................................78

6-1. Effect of soil water depletion (SWD) on fresh and dry weights of leaves, stem, and roots of carambola trees in containers....................................................................114

6-2. Effect of soil water depletion (SWD) on total fruit number and fruit weight of carambola trees in an orchard.................................................................................115

6-3. Effect of soil water depletion (SWD) on fresh weight of mature fruit (108 ohue) of carambola trees in an orchard.................................................................................116

6-4. Effect of soil water depletion (SWD) on dry weight of mature fruit (108 ohue) of carambola trees in an orchard.................................................................................117

6-5. Effect of soil water depletion (SWD) on fresh weight of ripe fruit (tree-ripened, 87

ohue) of carambola trees in an orchard...................................................................118

6-6. Effect of soil water depletion (SWD) on dry weight of ripe fruit (tree-ripened, 87

ohue) of carambola trees in an orchard...................................................................119

6-7. Effect of soil water depletion (SWD) on fresh and dry weights of mature fruit (108 ohue) of carambola trees in containers. ..................................................................120

vii

6-8. Effect of soil water depletion (SWD) on fresh and dry weights of ripe fruit (tree-ripened, 87 ohue) of carambola trees in containers.................................................121

6-9. Effect of soil water depletion (SWD) on total soluble solid (TSS) content of ripe fruit (tree-ripened, 87 ohue) of carambola trees in an orchard........................................122

6-10. Effect of soil water depletion (SWD) on total soluble solid (TSS) content of ripe fruit (tree-ripened, 87 ohue) of carambola trees in containers. ...............................123

7-1. The effect of soil water depletion (SWD) on shoot flushing, axillary shoot growth, flowering, and fruit load of orchard-grown carambola trees in 2003. ...................142

7-2. The effect of soil water depletion (SWD) on shoot flushing, axillary shoot growth, and flowering of container-grown carambola trees in 2003...................................143

A-1. Effect of soil water depletion (SWD) on total number and weight of carambola trees grown in containers. ...............................................................................................152

A-2. Average soil water depletion (SWD) prior to irrigation of carambola trees in containers in the field and in an orchard in 2002 and 2003. ..................................154

viii

LIST OF FIGURES

Figure page 3-1. Relationship between scaled frequency (SF) from a multisensor capacitance probe

(EnviroSCAN) and volumetric water content (%) fitted to obtain calibration coefficients for Krome very gravelly loam soil in containers, y = 0.011x + 0.5206; r2 = 0.98. ...................................................................................................................64

3-2. Relationship between count ratio from the neutron probe and volumetric water content (%) fitted to obtain a calibration equation for Krome very gravelly loam soil in containers, y = 29.05x - 6.4; r2 = 0.95. ..........................................................65

3-3. Soil water retention curves for Krome very gravelly loam soil at 10-, 20-, and 30-cm depths based on van Genuchten’s model calculated from volumetric water content (θv) measured with capacitance probes and soil water tension measured with tensiometers..............................................................................................................66

3-4. Soil water retention curves for Krome very gravelly loam soil at 10-, 20-, and 30-cm depths based on van Genuchten’s model calculated from volumetric water content (θv) measured with a neutron probe and soil water tension measured with tensiometers..............................................................................................................67

4-1. Total monthly rainfall and evapotranspiration (ET) during 2001 and 2002 in Homestead, Fla. Source: Florida Automated Weather Network, IFAS, Univ. of Fla., Gainesville................................................................................................................79

4-2. Relationship between volumetric soil water content measured with a neutron probe and soil water tension measured with a tensiometer in a carambola orchard in Krome very gravelly loam soil. The response line was fitted using the van Genuchten equation (r2 = 0.42). ...............................................................................80

4-3. Relationship between volumetric soil water content measured with multisensor capacitance probes and soil water tension measured with a tensiometer in a carambola orchard in Krome very gravelly loam soil. The response line was fitted using the van Genuchten equation (r2 = 0.35). .........................................................81

4-4. Relationship between soil water content (%) as measured by multisensor capacitance probes and a neutron probe in Krome very gravelly loam soil (y = 0.67 x + 12.19, r2 = 0.35). ................................................................................82

ix

5-1. Soil water depletion (SWD) and stem water potential (Ψs) of carambola trees in containers in a greenhouse (y = -0.44 - (3.59e-06) x3, r2 = 0.82). ............................93

5-2. Soil water depletion (SWD) and stem water potential (Ψs) of carambola trees in containers in the field (y = -0.48 - (2.79e-06) x3, r2 = 0.56).....................................93

5-3. Soil water depletion (SWD) and stomatal conductance to water vapor (gs) of carambola trees in containers in a greenhouse (y = 95.46 – 0.0002 x3, r2 = 0.46). ..94

5-4. Stem water potential (Ψs) and stomatal conductance to water vapor (gs) of carambola trees in containers in a greenhouse (y = 11.88 + 132.94 ex , r2 = 0.54)....................95

5-5. Stem water potential (Ψs) and stomatal conductance to water vapor (gs) of carambola trees in containers in the field (y = 44.76 – 41.57 / x, r2 = 0.32). .............................95

5-6. Stem water potential (Ψs) and transpiration (E) of carambola trees in containers in a greenhouse (y = 0.30 + 2.94 ex, r2 = 0.55)................................................................96

5-7. Stem water potential (Ψs) and transpiration (E) of carambola trees in containers in the field (y = 1.03 + 2.11 ex, r2 = 0.28).....................................................................96

5-8. Stomatal conductance to water vapor (gs) and transpiration (E) of carambola trees in containers in a greenhouse (y = 0.33 + 0.02 x, r2 = 0.76).........................................97

5-9. Stomatal conductance to water vapor (gs) and transpiration (E) of carambola trees in containers in the field (y = 0.34 + 0.02 x, r2 = 0.75). ...............................................97

5-10. Stem water potential (Ψs) and net CO2 assimilation (A) of carambola trees in containers in a greenhouse (y = -0.09 + 9.76 ex, r2 = 0.70). .....................................98

5-11. Stem water potential (Ψs) and net CO2 assimilation (A) of carambola trees in containers in the field (y = 1.19 - 2.95/x, r2 = 0.31). ................................................98

5-12. Stomatal conductance (gs) and net CO2 assimilation (A) of carambola trees in containers in a greenhouse (y = -2.46 + 0.39 (lnx)2, r2 = 0.75). ...............................99

5-13. Stomatal conductance (gs) and net CO2 assimilation (A) of carambola trees in containers in the field (y = -2.24 + 0.40 (lnx)2, r2 = 0.45). .......................................99

5-14. Leaf (ΨL) and stem water potentials (ΨS) of carambola trees in containers in the field (y = 0.10 + 1.00x, r2 = 0.97) and in a greenhouse (y = 0.33 + 0.91x, r2= 0.97)....................................................................................100

5-15. Leaf (ΨL) and stem water potentials (ΨS) of carambola trees in an orchard in the field (y = 0.73 + 1.38x, r2 = 0.85)...........................................................................100

x

6-1. Effect of soil water depletion (SWD) on shoot length of orchard-grown carambola trees during 2002 (A) and 2003 (B). Symbols with vertical bars represent means ± SE of 4 shoots per tree for 6 trees. .........................................................................124

6-2. Effect of soil water depletion (SWD) on trunk diameter of young carambola trees grown in containers. Symbols with vertical bars represent means ± SE of 6 trees. ................................................................................................................125

6-3. Effect of soil water depletion (SWD) on fruit length of orchard-grown carambola trees during 2003 summer harvest (August) and winter harvest (December). Symbols with vertical bars represent means ± SE of 4 fruit per tree for 6 trees. ...126

7-1. Average 41-year and 2003 monthly precipitation and average 2003 monthly temperature in Homestead, Fla. Source: Florida Automatic Weather Network (FAWN) (2003) and The National Oceanic and Atmospheric Administration (NOAA) (2003). .....................................................................................................144

7-2. Shoot flush ratings of carambola trees in Krome very gravelly loam soil of South Florida in 2003 irrigated at four levels of soil water depletion (SWD) in an orchard (A, n = 51) and in containers (B, n = 22). ..............................................................145

7-3. Shoot growth ratings of carambola trees in Krome very gravelly loam soil of South Florida in 2003 irrigated at four levels of soil water depletion (SWD) in an orchard (A, n = 51) and in containers (B, n = 22). ..............................................................146

7-4. Flowering ratings of carambola trees in Krome very gravelly loam soil of South Florida in 2003 irrigated at four levels of soil water depletion (SWD) in an orchard (A, n = 51) and in containers (B, n = 22). ..............................................................147

7-5. Fruit load rating of carambola trees in Krome very gravelly loam soil of South Florida in 2003 irrigated at four levels of soil water depletion (SWD) in an orchard (A, n = 51) and in containers (B, n = 22). ..............................................................148

A-1. Soil water depletion treatments (SWD) as determined from measurements of soil water content using multisensor capacitance probes (EnviroSCAN) in an 8-year-old carambola orchard in Krome soil in South Florida. Field capacity was determined based on the pattern of soil water depletion over a period of 7 d. Visual stress symptoms (VSS) occurred when the leaves became wilted and chlorotic.............153

A-2. Vapor pressure (VP) difference between saturated and ambient air vapor pressure in the greenhouse and the orchard. .............................................................................155

A-3. Soil water depletion (SWD) as measured with capacitance probes (EnviroSCAN) in an 8-year-old carambola orchard in Krome soils. Points of ‘field capacity’ and ‘onset of water stress’ were determined based on the pattern of soil water depletion. Measurements of stem water potential (ΨS) and net CO2 assimilation (A) indicated by the vertical double-line on the x-axis were chosen randomly at a point beyond the theoretical ‘onset of water stress’ point............................................................156

xi

A-4. Ratings of the phenological cycles of carambola trees in Krome very gravelly loam soil in an orchard in South Florida in 2003 Trees were irrigated at four levels of soil water depletion (SWD) in an orchard (n = 51).......................................................157

xii

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

SOIL WATER DEPLETION, GROWTH, PHYSIOLOGY, AND YIELD OF CARAMBOLA TREES IN KROME SOIL

By

Rashid Al-Yahyai

December 2004

Chair: Bruce Schaffer Cochair: Frederick S. Davies Major Department: Horticultural Science

Carambola (Averrhoa carambola) is a commercially important tropical fruit tree in

South Florida. Little is known about carambola water requirements or its response to

limited soil water supply under subtropical conditions such as those of southern Florida.

The objective of this study was to determine the effects of four levels of soil water

depletion (SWD) that ranged from field capacity (FC) to the first visual signs of water

stress (leaf yellowing and abscission), on leaf gas exchange, stem water potential,

phenology, growth, and yield of 2-year-old, container-grown carambola trees, and 8-year-

old 'Arkin' carambola trees in an orchard.

Soil water depletion treatments were determined from continuous measurement of

soil water content using multisensor capacitance probes (EnviroSCAN, Sentek PTY Ltd.,

Kent Town, Australia). Irrigation of orchard trees was initiated when SWD reached one

of the following levels (where 0% SWD = FC): 0-8% SWD, 9-11% SWD, 12-14% SWD,

or 15-17% SWD. Trees in 95 L containers filled with Krome very gravelly loam soil

xiii

were irrigated when SWD reached one of the following levels: 0-21% SWD, 22-31%

SWD, 32-50% SWD, or 51-60% SWD.

An increase in SWD resulted in a reduction in stem water potential (ΨS), net CO2

assimilation (A), transpiration (E), and stomatal conductance to water vapor (gs) of

carambola trees grown in containers in a Krome very gravelly loam soil. In the orchard,

SWD never reached a point where there was a decrease on ΨS, A, E, or gs. For carambola

trees in containers, when SWD levels were above 50% there was a reduction in ΨS that

subsequently reduced gs. A reduction in gs resulted in a linear decrease in E and a sharp

decline in A when gs fell below 50 mmol·m-2·s-1. Leaf gas exchange was better correlated

with ΨS than with SWD level. Therefore, ΨS may be a more accurate than SWD (as

determined from measurements of soil water content) for irrigation scheduling of

carambola in gravelly soils.

In the orchard, precipitation and possibly capillarity from the shallow water table

resulted in sufficient soil water content to obtain adequate vegetative growth and yields.

Adequate soil water content also resulted in no significant phenological differences

among treatments. In containers where lateral water movement and capillary rise were

prevented, tree trunk diameter and total dry weight were lower for trees irrigated at 32-

50% SWD and 51-60% SWD than for trees irrigated at 0-21% SWD, 22-31% SWD

treatments.

Based on the results of this study, irrigation of carambola trees in Krome soil at

17% SWD in an orchard and at 32% SWD in containers does not reduce tree growth,

yield and fruit quality.

xiv

CHAPTER 1 INTRODUCTION

Carambola (Averrhoa carambola L.), also called star fruit, is a member of the

Oxalidaceae and a native of Southeast Asia. This tropical fruit tree is cultivated between

30o latitude (Watson et al., 1988), primarily in Southeast Asia and also in China, India

(Nand, 1970; Popenoe, 1948; Tidbury, 1976), Guyana, Australia, Israel, and the United

States (mainly Hawaii and Florida) (Crane, 1994). There are approximately 100 ha of

carambola trees in Florida (J. H. Crane, personal communication), of which 46 ha are in

Miami-Dade County (Degner et al., 2002). A sweet-type, ‘Arkin’, has been the leading

commercial carambola cultivar in Florida (Campbell, 1989; Knight and Crane, 2002;

Lamberts and Crane, 1990; Núñez-Elisea and Crane, 1998).

Carambola trees are adapted to various types of well-drained soils (Crane, 1994);

however, deep, rich, and fertile soils of loamy-sand or clay-loam texture are generally

favorable for optimum growth and production (Galán Saúco et al., 1993).

Environmental factors such as low temperature (Campbell and Malo, 1981;

Campbell et al. 1985; George et al., 2002a, 2002b; Watson et al., 1988), drought (Ismail

and Awang, 1992; Ismail and Noor, 1996a; Ismail et al., 1994, 1996; Salakpetch et al.,

1990), flooding (Ismail and Noor, 1996b; Joyner and Schaffer, 1989), sunlight, (Marler et

al., 1994) and wind (Campbell, 1985; Crane, 1993; Crane et al., 1994a; Marler and

Zozor, 1992; Morton, 1987), as well as cultural practices such as cultivar selection

(Salakpetch et al., 1990), mulching (George et al., 2000, 2001), fertilization (Campbell,

1985, 1989; Crane, 1994; Knight, 1982), and pruning (Núñez-Elisea and Crane, 1998,

1

2

2000) influence vegetative and reproductive growth of carambola trees. In their native

tropical habitat, trees grow and bear fruit year-round (Núñez-Elisea and Crane, 2000).

Trees may bear new flushes, flowers, and immature and mature fruit simultaneously

(Galán Saúco et al., 1993). However, this phenological cycle can be interrupted due to

unfavorable environmental conditions during the winter months that include low

temperatures, dry winds, and high light intensity (Núñez-Elisea and Crane, 1998).

Tree responses to water stress depend on the stage of growth and duration and

severity of the stress. Vegetative growth of carambola was reduced with increased water

stress (Ismail and Awang, 1992; Ismail et al., 1994, 1996). The response of carambola

tree flowering to water stress depends on the time of stress. Flowering of carambola is

more affected by water stress and cultivar than by diurnal temperature variation and

photoperiod (Salakpetch et al., 1990). Even when tree growth is not affected by drought,

yield was significantly reduced in rain-fed trees compared to irrigated trees, mainly due

to a decrease in fruit size (Bookeri, 1996).

Leaf water potential (ΨL) decreases when water loss via transpiration (E) is greater

than water uptake (Salisbury and Ross, 1985). Studies on carambola have shown that ΨL

declines with increasing water stress (Ismail and Noor, 1996b; Ismail et al., 1994, 1996).

Salakpetch et al. (1990) observed that leaves of stressed carambola trees wilted when

midday ΨL reached -2.0 MPa and became yellow and abscised 5 d later. However, trees

appeared normal and rapidly recovered to their original ΨL levels of -0.6 MPa 2 to 3 d

after rewatering. Ismail et al. (1994) found that net CO2 assimilation (A) of B17

carambola clones grafted onto B2 clones in a greenhouse declined as a result of water

3

stress. Stomatal conductance (gs) was two to three times lower in stressed trees compared

to control trees.

Irrigation scheduling involves regulating the amount and frequency of irrigation to

meet the water demand of a particular crop. Application of water may vary depending on

plant needs which may vary depending on growth stage, climatic conditions, and soil

type. Regulation of irrigation is necessary to avoid over-watering fruit orchards that may

result in reduced tree growth and production from waterlogging, loss of water through

runoff or deep percolation, waste of energy resources, and potential agrochemical

leaching. Efficient irrigation management helps reduce the effects of insufficient water

supply that results in plant stress and a subsequent reduction in growth, yield and fruit

quality. The amount of water needed depends mainly on soil evaporation and plant

transpiration (evapotranspiration, ET).

Drought (Ismail and Noor, 1996a; Ismail et al., 1996; Salakpetch et al., 1990) and

excessive soil water (Ismail and Noor, 1996b; Joyner and Schaffer, 1989) reduce

carambola growth and yield. In some areas, scheduling irrigation is vital for commercial

carambola production. However, little is known about irrigation scheduling practices in

carambola producing countries.

In south Florida, Crane (1994) recommended applying 33 mm of water per ha

twice a week during dry periods throughout the year. However, irrigation rates and

frequencies needed for optimum growth and yield have not been established. Most of the

annual rainfall in Florida occurs during the summer months. Historical climatic records

for south Florida show that about 80% of rainfall occurred between May and October in

2001 to 2003. During winter, irrigation is used to compensate for the low rainfall and for

4

freeze protection. Irrigation is also used to compensate for unevenly distributed rainfall

within a month. A survey by Li et al. (2000) in 1998 showed that 73% of tropical fruit

growers of Miami-Dade County schedule irrigation based on the frequency and quantity

of rain. The percentage declined to 64.3% in 2002 according to a more recent survey by

Muñoz-Carpena et al. (2003). The number of respondents who monitored soil water

content for irrigation scheduling increased to 48.8% in 2002 (Muñoz-Carpena et al.,

2003) compared to only 15% in 1998 (Li et al., 2000). Problems concerning variability in

irrigation duration and frequency were rated high among tropical fruit growers (Li et al.,

2000), which highlights the need for a better understanding of irrigation requirements of

these crops.

Tensiometers, neutron probes, and capacitance probes directly or indirectly monitor

soil water content and are often used for irrigation scheduling. Tensiometers measure soil

suction or matric water potential rather than soil water content (Richards, 1942; Smajstrla

and Harrison, 1998). A soil water retention curve must be established to determine the

soil water content that corresponds to the water matric potential in order to estimate soil

water content with a tensiometer. The neutron probe is a portable compact unit that is

easy to operate, and volumetric soil water content can be obtained instantly at different

depths in the soil profile. Neutron probes are considered more accurate than many other

soil water status monitoring devices for irrigation scheduling (Evett and Steiner, 1995;

Mostert and Hoffman, 1996). A “count rate” obtained from a neutron probe inserted into

access tubes in the field is converted to volumetric water content using a calibration

curve. Neutron probes measure soil volumetric water content as a percentage of water per

volume of soil. A neutron probe contains radioactive materials which can cause health

5

and regulatory concerns; it is also relatively costly to purchase and maintain.

Capacitance probes measure the soil water content based on the dielectric constant of the

soil (Paltineanu and Starr, 1997; Phene et al., 1990; Wu, 1998), which includes dielectric

constants of water (80.4), soil particles (3 to 7) and air (1) (Robinson and Dean, 1993;

Paltineanu and Starr, 1997; Starr and Paltineanu, 1998a, 1998b; Wu, 1998). Since the

dielectric constant of the soil particles and air are small and relatively constant compared

to that of water, changes in the dielectric constant of the soil are a measure of the change

in soil water content. The volumetric water content can be expressed either as a

percentage at specific depths or total amount of water per volume of soil (Núñez-Elisea et

al., 2001; Paltineanu and Starr, 1997; Starr and Paltineanu, 1998a). An advantage of

capacitance probes system for measuring soil water content is that it provides continuous

monitoring. However, initial capital investment and maintenance are costly for this

system.

Soil water content should be correlated with physiological responses, growth, and

fruit production of tropical fruit crops, such as carambola, to determine the appropriate

amount of water to apply to the crop. The response of fruit trees to soil water deficit is

different from that of annual herbaceous plants. Trees are able to tolerate longer periods

of low soil water content through specialized short-term and long-term physiological,

phenological, anatomical, and morphological responses (Ludlow, 1989). Growth

reduction, stomatal closure, and osmotic adjustment are among the physiological

responses to a lower water potential (Hsiao, 1973; Turner, 1979). The correlation of soil

water depletion to tree physiological processes, growth and fruit production provides a

tool for accurate determination of tree water requirements. This aids in developing

6

irrigation scheduling strategies such as regulated deficit irrigation to obtain optimum

yield and enhance fruit quality. Regulated deficit irrigation has been used for several fruit

crops including peach (Chalmers et al., 1984), citrus (Castel and Buj, 1990; Domingo and

Ruiz-Sanchez, 1996), and apple (Mills et al., 1997).

Carambola trees respond to water deficit by drought avoidance (Neuhaus, 2003).

This strategy involves minimizing water loss by stomatal control, leaf movement, and

reduced leaf area (Ludlow, 1989). The lethal level of ΨL in carambola occurs at about

-2.9 MPa and the lethal leaf relative water content (RWC) at about 56% (Ismail et al.,

1994). The response of carambola trees to soil water deficit can be divided into four

phases (Neuhaus, 2003): (1) reduction in osmotic adjustment and low osmotic

adjustment; (2) increased root:shoot ratio by shedding leaves and fruit and increased leaf

osmotic adjustment; (3) irreversible damage caused by chlorophyll degradation, root

growth inhibition, and further leaf and fruit shedding; and (4) attainment of the lethal

physiological state is reached. Rewatering the trees at the second phase results in the

recovery of the plant to the original physiological state in a relatively short time

(Salakpetch et al., 1990). Thus, correlating physiological changes to soil water depletion

is essential to optimize irrigation amount and frequency.

The objective of this research project was to relate soil water contents in a gravelly

loam soil to physiology, growth, and yield of carambola trees. Studies were designed to

determine the effects of different soil water depletion levels on physiology, growth and

yield of young, container-grown, and mature, field-grown 'Arkin' carambola trees. The

main hypothesis was that physiology, growth, and yield responses of carambola trees are

closely related to soil water content.

7

Specific project hypotheses and questions addressed in this dissertation include the

following:

�Hypothesis 1: Soil moisture measurements using multisensor capacitance probes

and a neutron probe are site specific. The following question is addressed in chapter 3:

What are the calibration equations for a neutron probe and multisensor capacitance

probes (EnviroScan) in Krome very gravelly loam soil?

�Hypothesis 2: Tensiometers, multisensor capacitance probes, and neutron probes

differ in their accuracy and effectiveness for assessing soil water content. The following

question is addressed in chapter 4: What is the most accurate and practical technique for

determining soil water content in a Krome very gravelly loam soil?

�Hypothesis 3: Water stress due to low soil water availability decreases leaf gas

exchange and water potential of carambola trees. The following question is addressed in

chapter 5: What is the effect of soil water depletion on A, gs, E and ΨS of carambola

trees?

�Hypothesis 4: Reduced soil water availability reduces carambola growth, yield,

and fruit quality. The following question is addressed in chapter 6: What is the effect of

four soil water depletion levels on carambola vegetative growth, fruit growth, yield, and

fruit quality?

Hypothesis 5: Reduced soil water availability affects carambola phenology. The

following question is addressed in chapter 7: What is the effect of four soil water

depletion levels on carambola tree phenology?

CHAPTER 2 REVIEW OF LITERATURE

Carambola

Origin and Distribution

Carambola (Averrhoa carambola L.) is one of two species of the two genera of

woody plants in the Oxalidaceae, a family composed of primarily herbaceous species.

Another woody species in the same genus, Averrhoa bilimbi L., also produces edible fruit

commonly called bilimbi (Nakasone and Paull, 1998).

The exact origin of carambola is not known. It is believed to be native to

southeastern Asia and probably originated in Indochina or Indonesia (Galán Saúco et al.,

1993; George and Nissen, 1994), or Sri Lanka and Moluccas (Crane, 1994 and Morton,

1987). Carambola is cultivated within 30o north and south of the equator (Watson et al.,

1988; Galàn Saúco et al., 1993), primarily in Southeast Asia, Indonesia, Malaysia, and

Thailand, but also in China, Taiwan, India (Nand, 1970; Popenoe, 1948; Tidbury, 1976),

Brazil, Mexico, Guyana, Australia, Israel, and the United States (mainly Hawaii and

Florida) (Crane, 1994; Mauro Pace, 2004, personal communication). Carambola grows

best in tropical lowland areas with uniform medium to high rainfall throughout the year

(Campbell and Malo, 1981; Martin et al., 1987; Popenoe, 1948; Samson, 1980; Tidbury,

1976). However, it can also grow well in humid, subtropical areas (Campbell and Malo,

1981) as well as in areas with seasonal periods of drought (Galán Saúco et al., 1993;

Martin et al., 1987; Tidbury, 1976).

8

9

Economic Importance

The spread of carambola as a commercial tropical fruit tree was due to the

improvement in postharvest and cultivation practices, such as the control of fruit fly and

the selection of sweet cultivars with high productivity and good quality (Galàn Saúco et

al., 1993). No collective data on worldwide carambola production exist, but estimates

from individual countries have been reported. According to Galàn Saúco et al. (1993) and

Mauro Pace (2004, personal communication), major carambola producing countries are

Indonesia, Malaysia, Taiwan, Thailand; emerging producers are Brazil, Mexico and,

more generally, Caribbean countries. The current estimate of world production is 150,000

to 200,000 metric tons (MT) (Mauro Pace, 2004, personal communication). Carambola

production in Indonesia, which includes wild trees, fluctuated from 50,000 to 55,000 MT

between 1995 and 2000. Commercial production in Malaysia was 8,600 MT in 2000-

2002, declining from 29,000 MT in 1990-91. Malaysian production estimates, however,

may differ according to sources, in some cases reaching 40,000 MT, including production

from wild trees and non-commercial production. In Taiwan, production was estimated to

be 25,000 MT on 1,000 ha in 2002, declining from an area of 3,100 ha with 38,270 MT

produced in 1986 (Galàn Saúco et al., 1993). In 2001, production in Thailand was

estimated to be about 20,000 MT, Peru produced 1,766 MT, Israel produced 200 MT, and

commercial production from the USA (Florida and Hawaii) was estimated at about 4,000

MT (Mauro Pace, 2004, personal communication).

In the United States, the greatest hectarage of carambola occurs in Florida with an

estimated 100 ha in 2002 (Jonathan H. Crane, 2003, Personal communication) with

estimated annual production of 36.5 MT per ha. Hawaii has an estimated 8 ha of

carambola with an annual production of 43.5 MT per ha (H.A.S.S., 2001). Carambola

10

production in south Florida increased in the late 1980s and early 1990s as improved

cultivars such as ‘Arkin’ and aggressive consumer marketing created interest among

growers (Degner et al., 2002; Knight et al., 1984). In addition, an increase in immigration

from Southeast Asia and the recognition by growers and packers of the productivity and

market potential of this fruit led to increased planting of carambola (Knight et al., 1984).

By 1990, an estimated 242 ha had been planted in Miami-Dade County in Florida. In the

years following Hurricane Andrew in 1992, hectarage in Miami-Dade County began to

decline until there were only about 46 ha in 2001 (Degner et al., 2002). Despite the

decline in hectarage, carambola continued to be one of the highest value crops in south

Florida making it an economically viable fruit crop (Degner et al., 2002).

Botany and Uses

Growth habit. Carambola is a short-trunk evergreen tree with a broad, rounded

canopy (Morton, 1987; and Crane, 1994; Nakasone and Paull, 1998). Branches of

carambola are plagiotropic and tree height increases by continual superposition of major

limbs one on top of the other in a growth pattern that conforms to Troll’s model (Crane et

al., 1991; Crane et al., 1994a). According to Troll’s growth model, non-trained trees have

multiple leaders with an initial vertical orientation. Upon fruiting, the upright branches

become laterally oriented and new shoots develop along these laterally oriented limbs

(Halle et al., 1978). This results in overlapping branches and a spreading growth pattern

with the interior canopy becoming defoliated (Crane et al., 1991; Crane et al., 1994a).

Carambola trees possess a deep taproot and extensive lateral roots that develop a very

dense mass of thin roots with profuse absorbent hairs (Galàn Saúco et al., 1993). Due to

the gravelly texture of South Florida soils, which prevents deep penetration of the root

system, roots in South Florida orchards are concentrated in the top 10 to 20 cm of the soil

11

profile (Núñez-Elisea et al., 2001). Due to the hardness of the soil in southern Florida,

tropical trees including carambola are grown at intersections of perpendicular trenches

that are 40 cm wide and 50 to 60 cm deep (Bünemann and Crane, 2000; Colburn and

Goldweber, 1961; Núñez-Elisea et al., 2001). Most of the roots in Krome soil are

confined to the trenches and no taproot is evident; instead, numerous short vertically-

oriented roots grow to the bottom of the trench (Crane et al., 1994a). However, roots

radiate laterally in all directions in the “plow” layer (10 cm deep). Root restriction can

lead to reduced leaf and root growth especially when trees are water stressed (Ismail and

Noor, 1996a).

In Florida, vegetative and reproductive growth of carambola is vigorous from April

to October. After the October harvest, trees bloom and vegetative growth slows down

until it eventually ceases from November to February (Núñez-Elisea and Crane, 1998).

During this period, leaves often exhibit interveinal chlorosis, which is a typical symptom

of iron deficiency, and the rate of leaf senescence increases resulting in partial or full

defoliation (Núñez-Elisea and Crane, 1998).

Leaves. Leaves of carambola trees are compound, spirally arranged and alternate,

15 to 30 cm long, with 5-11 nearly opposite, ovate-oblong leaflets that are 1.5 to 9.0 cm

long and 1.0 to 4.5 cm wide (Campbell et al., 1985; Galán Saúco et al., 1993). They are

medium-green, and smooth on the upper surface, and slightly pubescent and whitish on

the lower surface (Crane, 1994; Morton, 1987).

Flowers. Carambola flowers are 5 to 11 mm long, 5 mm in diameter, with pink to

lavender petals, perfect and borne in clusters in axils of leaves on young branches or on

older branches without leaves. Most flowering and fruit production occur at the middle

12

canopy level (Crane et al., 1991). Carambola trees are heterostylous. Some cultivars bear

flowers with long styles and short stamens, while others have short styles and long

stamens (Campbell et al., 1985; Crane, 1994; Galán Saúco et al., 1993; Morton, 1987). In

tropical areas, carambola trees have the potential to flower year round (Samson, 1980;

Tidbury, 1976). Núñez-Elisea and Crane (1998) reported that pruning could stimulate

year-round flowering. They found that buds older than 10 weeks produced panicles,

while 6-week-old buds produced vegetative growth. Carambola flowers open in the

morning between 0800 and 1000 HR and close in the afternoon between 1400 and 1800

HR the same day (Salakpetch et al., 1990).

Due to their bright color, availability of nectar, and pollen viscosity, carambola

flowers are attractive to bees (Apis spp.), which are essential for pollination. However,

flowers may also be partially wind pollinated since some pollen is released outwards at

the moment when anther dehiscence occurs (Galán Saúco et al., 1993). Some flowers

may require cross-pollination to attain adequate fruit set and yield because of flower

heterostyly (Knight, 1965). However, cross-pollination by cultivars with different stylar

types is not always necessary in ‘Arkin’ and ‘Golden Star’ because they produce an

abundant fruit crop when planted in solid blocks (Crane, 1994).

Fruit. Carambola fruit are oval to ellipsoid, 5-15 cm long, and 3-10 cm in

diameter, with 4-6 (usually 5) prominent longitudinal ribs. Fruit slices cut in cross-section

are star-shaped, hence the name star fruit. The skin is thin, yellow to dark yellow and

smooth with a waxy cuticle. Immature fruit are green, turning yellow or orange when

ripe. The flesh is light-yellow to yellow, translucent, crisp, very juicy, and is not fibrous.

Depending on the cultivar, the fruit vary from very sour with little sugar, to very sweet

13

with little acidity. It takes about 60 to 75 d from fruit set for the fruit to reach maturity

(Campbell et al., 1985; Crane, 1994; Galán Saúco et al., 1993). Most of the fruit (>75%)

is produced on the periphery of the tree; therefore, it is subject to ‘wind-scar’ from dead

twigs under windy conditions (Núñez-Elisea and Crane, 1998, 2000).

The potential annual fruit yield of carambola ranges from 4.5-18 kg/tree in the 2nd

to 3rd years, 45 to 68 kg/tree in the 5th to 6th year, and 112-160 kg/tree at maturity

between the 7th and 12th year (Campbell 1989; Crane, 1993,1994; Núñez-Elisea and

Crane, 1998).

Carambola seeds are small, 0.6 to 1.3 cm long, thin, brown, edible, and there are

usually 10 to 12 seeds per fruit (Morton, 1987). In subtropical climates, some fruit have 1

to 2 seeds (Galán Saúco et al., 1993), while some fruit are seedless (Crane, 1994). Seeds

are light brown and enclosed by a gelatinous aril. Seeds lose their viability in a few days

after removal from the fruit (Crane, 1994).

Uses. Carambola fruit from sweet cultivars are mainly consumed fresh; however,

the fruit may be frozen, pickled, preserved, or used in various culinary preparations

(Crane 1994; Cooper et al., 1998). Fruit of acidic cultivars are consumed cooked (Galàn

Saúco et al., 1993). Most of Florida carambola fruit are marketed for fresh consumption.

Approximately 20% to 40% of the fruit produced do not meet the fresh market standards,

but can be utilized for processing (Matthews et al., 1988).

Carambola fruit are a good source of potassium and a moderate source of vitamin C

(Cooper et al., 1998). One hundred grams of the edible portion contains 35.7 calories, 89-

91g water, 0.38g protein, 0.08g fat, 9.38g carbohydrates, 0.8-0.9g fiber, 4.4-6mg calcium,

14

15.5-21mg phosphorus, 0.32-1.65mg iron, 26-53.1mg ascorbic acid, and trace amounts of

carotene, thiamine, riboflavin, and niacin (Morton, 1987).

Tree Phenology

Cull (1986) divided the phenological cycle of tropical evergreen trees into three

phases: the vegetative phase, the flower development phase, and the root development

phase. Each phase has an effect on the other phases. For example, vegetative growth is

essential for carbohydrate accumulation that is required for the development of young,

photosynthetically-efficient leaves, root growth, vegetative growth, and fruit growth and

development.

Environmental factors such as low temperatures (Campbell and Malo, 1981;

Campbell et al., 1985; Crane, 1994; George et al., 2002a, 2002b; Watson et al., 1988),

drought (Ismail and Noor, 1996a; Ismail and Awang, 1992; Ismail et al., 1994, 1996;

Salakpetch et al., 1990) , flooding (Ismail and Noor, 1996b; Joyner and Schaffer, 1989),

sunlight, (Marler et al., 1994) and wind (Campbell, 1985; Crane, 1993; Crane et al.,

1994a; Marler and Zozor, 1992; Morton, 1987), as well as cultural practices such as

cultivar selection (Salakpetch et al., 1990), fertilization (Campbell, 1985,1989; Crane,

1994; Knight, 1982), mulching (George et al., 2000; 2001), and pruning (Núñez-Elisea

and Crane, 1998, 2000), influence vegetative and reproductive growth of carambola.

In their native tropical habitat, carambola trees grow and bear fruit year-round

(Núñez-Elisea and Crane, 2000). Trees may produce new flushes, flowers, and bear

immature and mature fruit simultaneously (Galán Saúco et al., 1993). Under southern

Florida’s winter conditions that include low temperatures, dry winds, and high light

intensities (above 1000 µmol·m-2·s-1), canopies become chlorotic and defoliate from

February to March (Núñez-Elisea and Crane, 1998). Trees remain visually quiescent with

15

complete or partial defoliation during cool weather and begin to re-foliate from March to

mid-May and may continue to produce new leaves to mid-October (Núñez-Elisea and

Crane, 1998). Shoot growth alternates with periods of quiescence and apical growth

results in an extension of the shoot axis. Each growth flush contains between eight and 12

leaves (Núñez-Elisea and Crane, 1998). Shoots that are formed during the spring have

four to five flushes but those formed in the late summer normally flush once before shoot

extension growth ceases in the fall (Núñez-Elisea and Crane, 1998).

Carambola trees have several flowering periods and may flower continuously

during the year under favorable cultural and environmental conditions (Núñez-Elisea and

Crane, 1998, 2000). Two reproductive growth phases were observed in South Florida.

The first phase is characterized by peaks of flowering in May and fruit maturity from July

to October. The second phase is characterized by a peak bloom in September that results

in fruit maturity from December to February (Campbell et al., 1985; Núñez-Elisea and

Crane, 1998). In south Florida, the flowering phase of carambola extends from March to

early November, and production season extends from mid-July to mid-February

(Campbell et al., 1985; Núñez-Elisea and Crane, 1999, 2000). Núñez-Elisea and Crane

(1998) concluded that growth promoting environmental conditions such as water

availability or cultural practices such as pruning, defoliation, and nitrogen application

impact floral initiation more directly than tree carbohydrate status.

Cultivation

Soil requirements. Carambola trees are adapted to various types of well-drained

soils (Crane 1994; Popenoe, 1948; Nakasone and Paull, 1998). However, deep, rich, and

fertile soils with loamy-sand or clay-loam texture are generally favorable for optimum

growth and production (Galán Saúco et al., 1993). Soil pH between 5.5 and 6.5 is

16

generally ideal for tree growth (Campbell and Malo, 1981; Crane, 1994; Galán Saúco et

al., 1993; Green, 1987). In acid soils in tropical climates, carambola trees grow

vigorously and produce large deep-green leaves with few micronutrient deficiency

symptoms. Symptoms of micronutrient deficiency are more prevalent in calcareous soils

(Green, 1987). Carambola trees can tolerate soils with a pH below 5.5 much better than

avocado, mango or lytchee (Galán Saúco et al., 1993). In fact, carambola trees are grown

in pH 4.5 soils in Selangor State, Malaysia (Green, 1987) and in oolitic-limestone soils

with a pH between 7.5 and 7.8 in south Florida (Knight, 1982). However, due to the high

alkalinity of limestone soils, application of chelated iron, manganese, and zinc are

necessary to prevent micronutrient deficiencies (Crane, 1994; Knight, 1982).

In Miami-Dade County, Florida, carambola trees are generally grown in Krome

very gravelly loam soils. This mineral soil is very shallow, with a pH of 7.4 to 8.4 (Noble

et al., 1996). The soil is extremely low in organic matter and commercial farming largely

depends on fertilizer applications (Degner et al., 2002). Krome soil is very porous with a

gravelly loam surface layer between 8 and 23 cm thick, below which is a hard but porous

layer of limestone bedrock (Degner et al., 1997; Degner et al,, 2002; Moseley et al.,

1990).

Planting density. Spacing varies among cultivars, soil types and pruning practices.

A minimum distance of 5-7 m between trees and 6-9 m between rows is recommended

(Nakasone and Paull, 1998). In South Florida, carambola trees are usually planted at

moderate densities. Trees are planted at 4.6 to 6.1 m within and 6.1 to 7.6 m between

rows and planting density ranges from 286 to 356 trees/ha (Crane, 1993). Interplanting

with fruit crops such as papaya and banana is a common practice because these crops

17

provide an early income for growers and wind protection for young carambola trees

(Crane, 1989).

Shading. Carambola trees can survive under a wide range of light intensities

(Marler et al., 1994). Shading is favorable for tree growth, and therefore, carambola trees

can be successfully grown under shade cloth (allowing approximately 75% of sunlight

through) used as a windbreak (Crane, 1994; Watson et al., 1988). Marler et al. (1994)

exposed ‘Arkin’ seedling tree to approximately 25, 50, or 100% sunlight for 39 weeks

and found that shading increased rachis length and leaf area but decreased leaflet

thickness. Shading also reduced dark respiration, light compensation, and saturation

points, but increased chlorophyll concentration and nitrogen-use efficiency. Under 100%

sunlight, branches were more vertically orientated and stomatal density was higher than

in shaded trees.

Pruning. Pruning is an important cultural management practice used by carambola

growers to improve light penetration, shape the canopy, limit tree size, and remove dead

branches (Campbell, 1989; Crane, 1994; Green, 1987). Optimum tree height is about 3.5

to 4.5 m which facilitates manual harvesting (Campbell, 1989; Crane et al., 1991),

improves equipment movement between rows, improves efficacy and penetration of

foliar sprays, helps retain a bearing canopy throughout the tree, and reduces the potential

for wind damage to the canopy (Crane et al., 1991).

Pruning may stimulate flowering year-round (Núñez-Elisea and Crane, 1998,

2000), although Crane et al. (1991) found no consistent trend in flowering between

pruned and non-pruned trees. Núñez-Elisea and Crane (2000) reported that early-season

fruit production was increased by selective pruning and crop removal. They found that

18

pruning of 3- to 4-year-old branches to their main axes in early March resulted in

flowering by mid-April and a crop by the end of June compared to non-pruned trees that

produced mature fruit 4 to 5 weeks later. Pruning also reduced fruit wind-scar damage by

100%, whereas only 20% of non-pruned branches produced fruit free from wind damage.

Núñez-Elisea and Crane (2000) reported that fruit removal from ‘Arkin’ carambola trees

in November resulted in an average yield of 48 kg per tree on July, compared to an

average yield of 5 kg per non-defruited trees.

Irrigation. It is necessary to supply sufficient amounts of water to the soil to

maintain optimum growth and yield of carambola especially during dry periods when

rainfall does not meet tree water requirements. Drought before and after flowering can

result in poor flowering, early fruit abscission, reduced yield and small fruit (Nakasone

and Paull, 1998).

Scheduling irrigation can be beneficial for commercial carambola production.

However, little information is available about irrigation scheduling practices in

carambola producing countries. General recommendations for irrigation scheduling were

proposed in some regions of the world. During dry periods in northern Queensland and

the Northern Territory of Australia, irrigation of 30 to 75 mm of water/tree per week was

recommended for mature trees (Galán Saúco et al., 1993; Lim, 1996). In Malaysia where

annual rainfall exceeds 2000 mm, Bookeri (1996) reported an increase in carambola yield

(c.v. B17) with increasing irrigation rates of 4, 12, or 30 L/plant per day, but no water

amount or irrigation frequency was recommended. In southern Florida, Crane (1994)

recommended applying 33 mm of water per ha twice a week during dry periods

19

throughout the year. However, the optimum water application rates and frequencies have

not been established.

Microsprinklers are widely used to irrigate and fertigate carambola trees. High-

volume sprinklers are also used for irrigation and freeze protection (Campbell, 1989;

Crane, 1993, 1994). Carambola irrigation in Florida is generally based on visual

observations, since there are no established irrigation scheduling programs or water

requirement data specifically for carambola trees in this area.

Fertilization. A fertilizer application program for carambola trees should begin

with soil analysis prior to planting (Galán Saúco et al., 1993). Soil and leaf sampling can

be used to determine nutrient requirements of carambola trees (Galán Saúco et al., 1993).

Young bearing carambola trees should receive 0.4 to 0.8 kg/ha per tree per year of NPK

(11:12:17 to 15:15:15), while older trees (over 8 years old) may require 6 to 25 kg per

tree per year (Nakasone and Paull, 1998). Fertilizer should be applied approximately

every 3 months when trees are bearing fruit (Campbell, 1989).

The Krome very gravelly loam soil in which carambola trees are grown in South

Florida has low fertility and is often deficient in soluble iron, zinc, magnesium, and

manganese due to the high pH (Campbell 1985; Campbell et al., 1985; Crane, 1993).

Therefore, for adequate growth and yield, applications of chelated-iron, zinc, magnesium,

and manganese are necessary to prevent deficiencies of these elements (Campbell, 1985,

1989; Crane, 1994; Knight, 1982). Fertilizer mixtures with 6-8% nitrogen, 4%

phosphorous, 6-8% potash, and 3-4% magnesium are satisfactory for carambola trees

(Crane, 1994). Young trees require light applications of NPK fertilizer (20-20-20) every

60-90 d until they are established. Two fertilizer applications per year are needed in

20

deeper soils and three applications per year in shallow soils. For young trees, fertilizer

application rates are 400-500 g every 4-6 weeks during the rainy season or every 8 weeks

during the dry season (Campbell, 1989). Growers in South Florida apply 500-600 kg/ha

mixed fertilizers containing nitrogen, phosphorus, potassium and magnesium divided into

4-6 applications per year to mature orchards (Campbell, 1989). Chelated iron is applied

annually at 50-100 g/tree, increasing to 300-400 g/tree as trees grow. Zinc and

manganese are applied as inorganic salts, chelated compounds, or foliar sprays at a rate of

1.3-1.8 kg·l-1 of water (Campbell, 1989).

Cultivars. Programs for selecting improved carambola cultivars have been

established in Malaysia, Taiwan, Thailand, Florida and Hawaii (Nakasone and Paull,

1998). The Malaysian Agricultural Research and Development Institute (MARDI) was

one of the most active centers for carambola selection and breeding. Some of the

desirable characteristics for selections include: early season fruit maturity, abundant and

regular production, ease of tree training, plasticity of growth, desirable fruit quality, and

tolerance or resistance to biotic and abiotic stresses (Galàn Saúco et al., 1993). Several

carambola cultivars have been selected and cultivated in different parts of the world

(Table 2-1) (Galàn Saúco et al., 1993; Nakasone and Paull, 1998).

21

Table 2-1. Major carambola cultivars of economic importance in several major carambola production regions (Galàn Saúco et al., 1993; Nakasone and Paull, 1998).

Major Cultivars Growing regions

Arkin, Golden Star, B10 Florida, US

Sri Kembangsaan, Kary Hawaii, US

Kaput, Ting Go, Demak Indonesia

Lang Bak, Juron Singapore

Hong Hug, Far Dee China

Cheng Tsey Taiwan

B-2, B-10, B-17 Malaysia

Arkin, B-1, B-6, B-10, Jungle Gold Australia

Fwang Tung, Thai Knight Thailand

22

Several carambola cultivars have been selected for their adaptation to the South

Florida climate. In addition, cultivars were screened for high fruit yields, good fruit

quality, medium fruit size, yellow fruit color, ability of fruit to withstand damage during

handling, length of fruit storage time, and marketing potential (Wagner et al., 1975). A

study conducted by Crane (1989) showed that 89% of the carambola trees surveyed in

southern Florida were 4-years-old or younger; 97% of these were cultivars with sweet

fruit and 3% were cultivars with tart fruit. The dominant sweet cultivar has been ‘Arkin’

and the main tart type has been ‘Golden Star’ (Campbell, 1989; Crane, 1989; Knight and

Crane, 2002). Other cultivars with very few commercial plantings in South Florida

include ‘Kary’, from Hawaii, ‘B-10’, from Malaysia, ‘Fwang Tung’, from Thailand, and

‘Maha’ from Malaysia (Campbell, 1989; Crane et al., 1998; Knight and Crane, 2002).

‘Arkin’ is the most commonly planted commercial cultivar in South Florida to date

because of its resistance to picking, packing and shipping damage, as well as its good

marketability (Knight and Crane, 2002). The fruit are medium-sized, weigh between 90

and 200 g and are golden yellow at the early maturation stage, turning to yellow-orange

when ripe. The ribs are thick and compact with a relatively large angle, and the edges of

the wings are slightly rounded. The fruit has crisp texture and is juicy with a sweet flavor

and relatively low acidity. Total soluble solid (TSS) content from July-harvested fruit

was 4 to 8% (mean 6.8%) and that from November-harvested fruit was found to be 6 to

8% (mean 7.1%) (Crane et al., 1998). The fruit are also relatively insensitive to chilling

injury during storage (Campbell 1989; Galán Saúco et al., 1993). Potential annual yield

of carambola ranges between 4.5 to18 kg/tree when trees are 2 to 3 years old; 45 to 68

23

kg/tree at 5 to 6 years old; and 112 to 160 kg/tree at maturity between 7 to 12 years old

(Campbell 1989; Crane 1994; Núñez-Elisea and Crane, 1998).

Campbell (1985) selected and named ‘Golden Star’ in Florida in 1985. This

cultivar has tart fruit that can attain a sweet flavor if allowed to ripen on the tree (Crane,

1994). Open pollinated seedlings of Golden Star are used primarily as rootstocks in

Florida. Under south Florida conditions, Golden Star rootstocks are preferred to Arkin

seedling rootstocks because they are well adapted to calcareous soils (Crane, 1994;

Knight, 1982), moderately tolerant of flooded conditions (Joyner and Schaffer, 1989),

and provide good tree anchorage (Crane, 1994).

Environmental Stress Physiology of Carambola

Information on the effects of environmental factors including temperature, drought,

and flooding on tropical fruit crops such as carambola is less available in comparison to

that for temperate fruit species. This is due in part on limited resources available for

research and development in developing countries where tropical species are largely

cultivated (Schaffer and Andersen, 1994). Among the adverse environmental conditions

that affect carambola growth and yield are temperature extremes, flooding, wind, and

water stress.

Temperature

Carambola trees are best adapted to hot lowland tropical climates but grow well in

warm subtropical areas if not subjected to severe freezes (Campbell et al., 1985; Crane,

1994). Ideal temperatures for optimum growth and production of carambola are between

21 and 32oC (Nakasone and Paull, 1998; Ngah et al., 1989).

High temperatures reduced yields by reducing fruit set due to lack of pollination.

Temperatures of more than 30 to 35oC during flowering reduced pollen germination and

24

pollen tube growth (Salakpetch, 1987). Carambola leaflet petiolules are equipped with

pulvini that allow leaflet movement. Under direct full sunlight, vertical leaflet movement

helps reduce leaf temperature and increase photochemical efficiency of carambola leaves

(Marler and Lawton, 1995). Marler and Lawton (1995) reported that maintaining a leaflet

in a horizontal position for 3.5 h during midday under full sunlight reduced chlorophyll

fluorescence and increased leaf temperature. Leaflet seismonastic movement also occurs

in response to shaking of the tree canopy (Marler and Zozor, 1992).

Low temperatures decrease carambola growth and freezing temperatures can lead

to death of branches or trees (George and Nissen, 1994). Young trees can be killed if

exposed to temperatures of -2.8 to -1.7 oC, while temperatures of -6.7 to -4.4 oC may kill

large branches and mature trees (Campbell and Malo, 1981; Campbell et al., 1985;

Watson et al., 1988; Crane, 1994).

In Florida, carambola trees can be grown in warm locations along the southeastern

and southwestern coasts (Campbell et al., 1985; Crane, 1994). The temperatures in south

Florida are suitable for growing carambola about 9 to 10 months of the year, during

which carambola trees grow vigorously and may grow vegetatively and reproductively at

the same time. Carambola leaf production in Florida declines after mid-October with a

heavy fruit load until December causing a sharp decline in tree carbohydrate reserves

during the winter (Núñez-Elisea and Crane, 1998). The trees remain quiescent and

partially defoliated until mid-March when refoliation occurs due to warm temperatures.

George et al. (2000) reported that root mulching using organic and polyethylene mulches

decreased defoliation during the winter months by increasing root temperature.

25

Wind

Carambola is severely damaged by strong winds that can cause desiccation,

dieback, defoliation and fruit scaring (Crane, 1993; Crane et al., 1994a; Campbell, 1985).

Wind may reduce tree growth by reducing leaf gas exchange or by mechanical stress or

damage (Marler and Zozor, 1992). Increasing wind speed from 0 to 2 m s-1 reduced net

CO2 assimilation (A) of container-grown carambola trees by 22% and leaf area and dry

weight production by about 50%. Wind-stressed carambola seedlings had less plant

height, leaf area, dry weight, trunk cross-sectional area, a lower relative growth rate and

leaf-area ratio, and lower A, and gs values than plants that were not wind-stressed (Marler

and Zozor, 1992).

In South Florida, carambola trees are frequently subjected to wind stress caused by

strong winds, storms and hurricanes. Planting carambola trees in wind-protected orchards

with man-made or natural windbreaks is a common practice in Florida. Commercially,

carambola trees in Florida are grown within windbreaks that are either composed of trees

with a rapid growth rate or fabric screens supported by posts and wires (Crane, 1993).

After Hurricane Andrew in 1992, 93% of the carambola trees survived. Of the surviving

trees, 13% were toppled, 4% were stumped (reduced to trunk), whereas 76% of the trees

remained intact (Crane et al., 1993). Tree decline continued for 14 to 15 months after the

hurricane. This decline was due to detachment of major roots and bark at the soil line

(Crane et al., 1994a). Crane et al. (1994b) proposed several pre-hurricane and post-

hurricane practices to reduce the damage from hurricanes. Pre-hurricane practices include

the use of grafted material and preparation of planting sites to increase tree rooting depth.

After the hurricane, clearing the debris, resetting toppled trees, protecting trees from

sunburn, and irrigating and fertilizing trees frequently hastened tree recovery.

26

Flooding

Carambola is considered a moderately flood-tolerant tree (Crane, 1994; Joyner and

Schaffer, 1989). However, continuous flooding for up to 18 weeks resulted in reduced A,

transpiration (E), and gs (Joyner and Schaffer, 1989). Productivity of mature trees was

reduced under flooded conditions (Campbell, 1989) because of reduced leaf, shoot, and

root dry weights (Joyner and Schaffer, 1989). However, after flooded trees were

unflooded, trees were able to recover from continuous and intermittent flooding and A, E,

and gs increased to pre-stress levels (Joyner and Schaffer, 1989; Ismail and Noor, 1996b).

Drought and water stress

Carambola trees grow best in tropical lowlands with uniform rainfall distribution

throughout the year (Campbell and Malo, 1981; Samson, 1980). However, trees can also

grow well in humid, subtropical areas (Campbell and Malo, 1981). In some dry areas,

carambola trees are very tolerant of seasonal drought periods once they are well

established (Galán Saúco et al., 1993). The optimum rainfall for maximum production in

Malaysia is 2500 mm annually (Ngah et al., 1989), while rainfall between 1500 to 3000

mm is suitable for production in Australia (Galán Saúco et al., 1993). Locations with

annual rainfall above 1800 mm in Australia produced fruit with better flavor and quality

compared to trees in areas with less rainfall (Morton, 1987). Good quality fruit have been

produced in the Canary Islands with annual rainfall and irrigation of 800 mm (Nakasone

and Paull, 1998). Average annual rainfall in south Florida where carambola trees are

grown is 1500 mm, two-third of which occurs from May to November.

Vegetative growth. Shoot expansion in plants is greatly affected by water stress

and therefore can be an indicator of water stress (Hsiao, 1973; Hsiao and Bradford,

1983). Vegetative growth of carambola decreases with increased water stress (Ismail and

27

Awang 1992; Ismail et al., 1994, 1996), particularly when root growth is restricted

(Ismail and Noor, 1996a). An adaptive mechanism to water stress is greater allocation of

dry matter to the roots which results in a high root to shoot ratio in water-stressed plants

(Ismail and Awang 1992; Ismail et al., 1996). Stressed plants, however, continue to

produce new leaves, although at a slower rate than nonstressed control trees (Ismail et al.,

1996). Leaves wilted 10 d after withholding water from 2-year-old ‘Fwang Tung’ and

‘Thai Knight’ carambola trees (Salakpetch et al., 1990). Leaf wilting coincided with

midday leaf water potentials (ΨL) of approximately -2.0 MPa. Five days later, some

leaves became yellow and abscised. Upon rewatering, leaves appeared normal (non-

wilted) and ΨL returned to previous levels (-0.6 MPa) 2 to 3 d later.

Trunk diameter or tree girth can be used to evaluate the effect of long-term and

short-term stress on fruit trees. Bookeri (1996) found no significant difference in girth

between irrigated and non-irrigated (rainfed) young carambola trees because sufficient

rainfall occurred following periods of drought.

Reproductive growth and yield. The response of carambola tree flowering to

water stress depends on the time of stress. Flowering of carambola trees is more strongly

influenced by water stress and cultivar than by diurnal temperature variation and

photoperiod (Salakpetch et al., 1990). Ismail et al. (1996) reported that water stress

induced early flowering in container-grown young carambola trees by approximately 6

weeks. Production of new flowers continued in water-stressed trees and there were more

flowers in water-stressed than in well-watered trees. However, flowering was reduced

when water stress occurred during flowering or fruit set. In addition, prolonged severe

water stress may reduce flowering of carambola trees. In a heated glasshouse experiment,

28

Salakpetch et al. (1990) found that cyclical or continuous (-2.0 MPa midday ΨL) water

stress for 3 to 4 weeks inhibited flowering of ‘Thai Knight’ carambola. Flowering of

carambola trees is also influenced by root growth. Flower initiation was hastened by

restricting root growth even under well-watered conditions (Ismail and Noor, 1996a).

Although tree growth was not affected by drought, yield was significantly

increased in irrigated compared to rainfed trees (Bookeri, 1996). The increase in yield

was attributed to a 17% increase in fruit size, although fruit number was not affected by

irrigation (Bookeri, 1996).

Physiological processes. Leaf water potential decreases when water lost via E is

greater than water uptake (Salisbury and Ross, 1985). Leaf water potential of carambola

trees declined with increasing water stress (Ismail and Noor, 1996a; Ismail et al., 1994,

1996). Ismail et al. (1994) carried out an experiment on 6-month-old, container-grown

‘B17’ carambola trees grafted onto B2 rootstock in a greenhouse and observed that the

mean midday ΨL of water stressed trees was -2.20 to -0.35 MPa lower than that of non-

stressed trees. Ismail and Noor (1996b) found that ΨL (measured 4 h after sunrise) of

flooded carambola trees declined progressively to -1.3 MPa 21 d after trees were flooded

compared to -0.80 MPa for non-flooded trees.

Ismail et al. (1994) studied leaf gas exchange of 6-month-old, container-grown

carambola trees grown in a greenhouse at mean ambient temperatures of 29.4 ± 5.6 oC

subjected to water stress. The mean A of non-stressed trees was 3.13 µmol.m-2.s-1

compared to 0.24 to 1.83 µmol.m-2.s-1 for the water stressed trees. Stomatal conductance

was two to three times lower in water-stressed trees than in control trees. In another study

of carambola trees in a glasshouse, Ismail and Awang (1992) observed that A was only

29

reduced when ΨL fell below -0.85 MPa and that plants were photosynthesizing 14 d after

withholding water. They also found that gs, respiration, and chlorophyll content were

only reduced 7 d after withholding water. Reduction in A could be attributed to reduced

gs. Ismail et al. (1994) reported that A was reduced when stomatal resistance exceeded

3.5 s cm-1. They concluded that a photosynthetic photon flux (PPF) higher than 1000

µmol.m-2.s-1 reduced A of stressed and non-stressed plants due to photoinhibition. Care

must be taken when measuring leaf gas exchange of carambola trees because leaves are

sensitive to mechanical stress caused by the leaf cuvette clamping onto the leaf during

gas exchange measurements. This mechanical damage reduced gs of carambola leaves

(Marler and Mickelbart, 1992).

Water Stress

Lack of water is one of the most limiting environmental factors for crop production

in the world. Plants require large amount of water to produce their biomass, ranging from

several hundred to 2000 g of water per g of dry matter produced (Hsiao, 1993). Thus,

plant water status is perhaps the most important factor that must be controlled to obtain

high yields of good quality horticultural crops (Jones, 1990). Plant responses to water

stress depend on the quantity of water lost, the rate of loss, and the duration of stress

(Bray, 1997). Water stress can reduce vegetative growth, flowering and fruit growth.

Hanan (1972) defined water stress as the difference between water supply and demand. A

water deficit occurs when a plant’s E exceeds its water uptake and is a component of

other stresses including drought, salinity, and low temperature (Bray, 1997). Water stress

is therefore caused by any restriction in supply or resistance to flow through the plant

when demand for water increases. Halevy (1972) stated that water stress is a result of an

imbalance between three components that determine plant water status: water absorption,

30

translocation and loss. Lakso (1985) argued that the relative sparsity of roots in fruit

crops leads to low root hydraulic conductivity and thus under high evaporative demand,

low plant Ψ develop. Therefore, in fruit trees, daily plant Ψ is controlled more by the

evaporative demand than by soil water tension. Thus, for fruit trees, plant water status

may be a better determinant than soil water content for scheduling irrigation to obtain the

desired growth and yield.

Water deficits impact many physiological and developmental processes that affect

fruit production, including growth (cell division and cell expansion), and gas exchange

(stomatal aperture, and photosynthetic and respiratory enzymes) (Jones et al., 1985). The

response of fruit trees to water stress varies between species and varieties within the same

species. Germanà and Sardo (1996) reported variable responses to water deficit for 41

citrus cultivars grown under nearly identical vapor pressure deficits (VPD) and PPF.

They cited morphological and functional differences among the leaves as factors

influencing water use efficiency (WUE) where larger leaves with higher stomatal

densities, had higher A with no significant difference in E compared to trees with smaller

leaves. Mild water stress increased water use efficiency of apple (Davies and Lakso,

1979; Flore et al., 1985), cherry (Flore et al., 1985) and peach (Chalmers et al., 1984;

Girona et al., 1993).

Turner (1979; 1986) separated the mechanisms of plant adaptation to water deficits

into three categories:

1) Drought escape: the ability of a plant to complete its life cycle before serious soil

and plant water deficits develop. This can be achieved by rapid phenological

development and developmental plasticity.

31

2) Drought tolerance with low plant water potential: the ability of a plant to endure

periods without significant rainfall and endure low tissue water status, i.e.

dehydration tolerance (Kramer, 1980). The mechanisms involved in this process

include maintenance of turgor by osmotic adjustment, increase in cell elasticity, a

decrease in cell size, and desiccation tolerance through protoplasmic tolerance.

3) Drought tolerance with high plant water potential: the ability of a plant to endure

periods without significant rainfall while maintaining a high plant water status,

i.e. dehydration postponement (Kramer, 1980). This can be achieved by reduction

of water loss through an increase in stomatal and cuticular resistance, a reduction

in radiation absorbed, and/or a reduction in leaf area, and maintenance of water

uptake through increased root density and depth and increased hydraulic

conductance.

The root system of fruit crops affects the rate of water stress development. A slow

rate of onset of stress may allow for development of adaptive mechanisms such as

osmotic adjustment, decreasing leaf area, abscission of leaves, leaf folding, rolling or

reorientation of leaves, and an increased root growth rate (Flore and Lakso, 1989).

Lankes (1985) found that drought tolerance of apple leaves and fruit varied depending on

the time of season or stage of development. In the early stage of development, leaf A and

E responded in a pattern of ‘drought tolerance at high water potential’ and later in the

season, the response pattern was ‘drought tolerance at low water potential’. Osmotic

adjustment is an example of avoidance at the cellular level as a low rate of water loss

may allow acclimation to water deficit (Bray, 1997).

32

Effect of Water Stress on Vegetative Growth

Root growth and distribution

Plants often respond to water stress by a shift in new assimilate distribution to

support the growth of roots at the expense of the shoot (Mansfield and Atkinson, 1990).

Enhanced root growth under water stress is due to osmotic adjustment in root cells

(Hsiao, 1973). Shallow roots of avocado can grow deeper than 1.2 to 1.5 m in well-

drained sandy soils or under restricted water supply compared with other fruit trees

(Lahav and Whiley, 2002). The root to shoot ratio was 3.5 times higher in water-stressed

kiwifruit vines compared to well-watered vines (Chartzoulakis et al., 1993). In carambola

trees, the root to shoot ratio was proportional to the duration of water stress (Ismail et al.,

1994).

Leaf area

Leaf growth is known to be very sensitive to water stress. Leaf area is often

reduced when trees are subjected to water stress. In deciduous fruit trees, when the

canopy is incomplete at the beginning of the growing season and intercepts only part of

the incident radiation, even mild water stress can slow leaf growth (Hsiao, 1993). A

reduction of 30% (70% of total ET) of irrigation water reduced leaf area in walnut

(Cohen et al., 1997). A reduction in leaf area reduced A leading to reduced biomass

accumulation. Chartzoulakis et al. (1993) reported that decreased growth of kiwifruit

induced by water stress was a result of a reduction in A and preferential partitioning of

photosynthate to the roots and thus affected leaf area development. Elongation of leaves

of container-grown, 6-month-old carambola trees was reduced as a result of water stress

(Ismail et al., 1994). Leaf area of trees was reduced by 60% when water was withheld for

33

1 week and 90% when water withheld for either 2 or 4 weeks before restoring soil to field

capacity (Ismail et al., 1994).

Trunk growth

Responses of stem diameter to changes in plant water status and the development

of instruments that automatically measure these changes have lead to the use of trunk

diameter as a method for irrigation scheduling for several fruit crops (Hilgeman, 1962;

Higgs and Jones, 1984; Huguet et al., 1992; Simonneau et al., 1993). A decrease or

cessation of daily stem diameter growth was the first indication of water stress in field-

grown apple and peach trees. If trees received a sufficient amount of water, maximum

daily stem shrinkage was a versatile indicator of E (Huguet et al., 1992). The change in

diameter of a living stem over 24 h can be described by measurements of maximum daily

shrinkage (MDS) which is the maximum diameter usually observed in the early morning

and the minimum diameter usually observed in mid-afternoon, and daily evolution (DE)

which is the total change in stem diameter over a 24-h period beginning at dawn (Huguet

et al., 1992; Katerji et al., 1994). Maximum daily stem shrinkage is affected by

environmental factors affecting plant E, e.g. soil water content and potential

evapotranspiration.

Trunk diameter fluctuations can be measured with linear variable differential

transformers (LVDT) (Huguet et al., 1992). Trunk diameter variations were a good

indicator of water stress in walnut (Cohen et al., 1997), peach (Garnier and Berger, 1985;

Huguet et al., 1992; Girona et al., 1993), avocado (Lahav and Kalmar, 1977), cherry

(Kozlowski, 1968), citrus (Hilgeman, 1962), and apple (Huguet et al., 1992). Huguet et

al. (1992) observed two types of specific daily shrinkage patterns: (1) a ‘peach-tree’

pattern, where maximum daily shrinkage increased as the severity of water stress

34

increased; and (2) an ‘apple-tree’ pattern, where maximum daily shrinkage was smaller

under severe water stress than that of well-watered trees.

Shoot growth

Reductions in extension shoot growth and leaf expansion are the first visible

responses of crop plants to water stress (Hsiao and Acevedo, 1974; Lakso, 1985). In

grapevines, shoot growth was found to be more sensitive to water stress than fruit yield

(Nagarajah, 1989). In crop plants, defoliation and restricted leaf growth are water stress

adaptations aimed at reducing E and conserving water (Nagarajah, 1989). A decrease in

shoot growth caused by a water deficit is a result of a reduction in both A and

photosynthate partitioning to the shoot in favor of root growth (Chartzoulakis et al.,

1993). Carambola (Ismail et al., 1994) and avocado (Lahav and Whiley, 2002) shoot

growth was significantly reduced under water stress conditions.

Effect of Water Stress on Reproductive Growth and Yield

Flowering and fruiting

Water status of flowers and their response to water stress depends on their stage of

development. Halevy (1972) stated that young flowers of gladioli, up to about 2 d prior

to anthesis of the first floret, were more sensitive to drought than old leaves because they

were the first to wilt. Water relocation was reversed just before anthesis when water was

translocated to the flowers from other organs. Hsiao (1993) suggested that during severe

water stress at the time of anthesis, pollination can be prevented and the number of

reproductive sinks for assimilates is reduced. However, mild to moderate water stress

before and during anthesis can enhance the partitioning of assimilates toward

reproductive sinks. Increased partitioning of assimilates to the reproductive sinks then

promotes embryo development which in some species can result in early maturation of a

35

portion of the fruit. Menzel et al. (1995) found that before anthesis in lychee trees, tree

water status was not related to extension growth of floral panicles or leafy shoots. In

contrast, they found that no vegetative shoots were initiated after fruit set in drought-

stressed trees when ΨL declined to –2.5 MPa.

A reduction in the number of fruit as a result of moderate to severe water stress is

often attributed to an inhibition of pollination. However, more likely causes of fewer fruit

as a result of water stress are a reduced number of fruiting sites associated with the

smaller plant size and abortion of younger fruit (Hsiao, 1993). Water deficits reduced

initial fruit set of lychee by 30% and final yield by 70% as a result of fruit splitting

(Menzel et al., 1995). ‘Valencia’ orange had delayed bloom and lower fruit set when

irrigated at soil water tension of 70 kPa compared to trees irrigated at 45 kPa or below 20

kPa (Davies and Bower, 1994). Predawn leaf water potentials of –1.7 to –3.5 MPa

resulted in reduced growth but did not induce flowering in avocado and lychee compared

to trees with leaf water potentials of –0.4 to –0.7 MPa (Chaikiattiyos et al., 1994). Similar

responses were observed for mango when predawn relative water content was 90-93%

compared with 97% or above in control trees. Schaffer et al. (1994) stated that a dry

period preceding flowering is essential for reliable mango production in the tropics. In

lemon trees, flowering is mainly determined by water stress over a range of daytime

temperatures from 18oC to 30oC. More flowers were produced by water-stressed lemon

trees in comparison to well-watered trees (Barbera and Cascio, 1985; Chaikiattiyos et al.,

1994). Similar observations were made for ‘Tahiti’ lime by Southwick and Davenport

(1986, 1987).

36

Yield

Yield losses in water stressed plants can also be attributed to a reduction in

inflorescence initiation, fruit set, and increased fruit abscission (Nagarajah, 1989). The

effect of water stress on yield varies among tree species and cultivars. Reduction in yields

due to water stress has been extensively studied in several fruit crops in temperate,

tropical and subtropical climates. Water stress reduced fruit size and yield of temperate,

tropical, and subtropical fruit trees including apples (Failla et al., 1992), tangerines

(Huang et al., 2000), walnut (Cohen et al., 1997), grapes (Nagarajah, 1989), apricot

(Torrecillas et al., 2000), lychee (Batten et al., 1994; Menzel et al., 1995), and avocado

(Lahav and Whiley, 2002). In an irrigation experiment in Miami-Dade County, Florida,

fruit of ‘Tahiti’ lime matured earlier (before 1 June) when one third (high) or two thirds

(moderate) of available soil moisture had been depleted compared to trees irrigated 2 d

following ‘midday wilt’ with 22.86 mm of irrigation water (Krome et al., 1970).

However, total annual yield was not significantly different among lime trees subjected to

these three irrigation treatments.

Fruit quality

Water stress affects several fruit quality factors. Nagarajah (1989) reported that

mild water stress during the ripening stage of grapevines increased the concentration of

sugars and anthocyanins and decreased acid and pH levels. Moderate water stress

improved fruit quality of prunes (Shackel et al., 2000). Total soluble solids (TSS) of

satsuma mandarin trees increased with deficit irrigation of -30, -50, -70, or -100 kPa as

determined with tensiometers (Peng and Rabe, 1998). Frequent microsprinkler irrigation

at low soil tension (10-15 kPa) also resulted in increased fruit size and increased TSS of

Florida grapefruit when compared to high soil tension (34-45 kPa) with less frequent

37

irrigation (Boman, 1996). Domingo et al. (1996) subjected ‘Fino’ lemon trees to three

irrigation levels: 100% ETc all year, 25% ETc all year but 100% ETc during fruit growth,

and 100% ETc all year but 70% ETc during fruit growth. They found no significant

differences among treatments in TSS, acidity, ascorbic acid glucose, fructose, and

sucrose levels. Avocado oil content was 10% higher in trees irrigated at 7-d intervals with

889 mm per year than in trees irrigated every 28 d with 594 mm of water per year (Lahav

and Kalmar, 1977).

Effect of Water Stress on Physiological Processes

Net CO2 assimilation and transpiration

Net CO2 assimilation and E are processes that involve gas and water vapor

exchange through the stomata. To study the effect of water stress on these processes, it is

recommended that they be examined together (Lakso, 1985). The relationship between A

and gs is different than that between E and gs (Lakso, 1985; Schulze, 1986). Under a

constant vapor pressure gradient, E is linearly correlated with gs, whereas A is

hyperbolically correlated. Under low water stress, gs may be very high and partial

stomatal closure may reduce E with little effect on A (Lakso, 1985). Both A and E are

reduced equally, however, under low humidity or progressive water stress that induces

stomatal closure (Lakso, 1985). The difference in vapor pressure between leaf and air, i.e.

vapor pressure gradient, is very important at both the leaf and canopy levels, because it

drives E and directly affects gs and A (Schulze, 1986).

Water uptake plays an important role in plant A. There is less resistance to water

efflux than CO2 influx into leaves through the stomata (Hsiao, 1973, 1993). Therefore,

large amounts of water are lost through E as a consequence of stomates opening to allow

entry of carbon dioxide into the leaves. Transpiration varies in the morning from day to

38

day depending on variability in evaporative demand. Thus, early afternoon

measurements, when E is low but stable, are usually recommended (Braun et al., 1989).

During the development of water stress, stomatal closure is the dominant

mechanism restricting water loss in mesophytes (Hsiao, 1973). This affects A directly by

restricting the CO2 supply. The decrease in gs is therefore an indicator of plant water

stress (Braun, et al., 1989; Nagarajah, 1989). However, it should only be measured on

clear days, to exclude effects of rapidly changing light and heat conditions (Braun et al.,

1989). Severe water stress can affect the photochemical and enzymatic reactions of A

(Nagarajah,1989). Reduction in A in response to water stress has been reported in many

fruit crops such as satsuma mandarin (Peng and Rabe, 1998), kiwifruit (Chartzoulakis et

al., 1993), lychee (Menzel et al., 1995), ‘Valencia’ orange (Davies and Bower, 1994),

peach (Girona et al., 1993), pistachio (Palma et al., 1997), apple (Sritharan and Lenz,

1989), carambola (Ismail et al., 1994), lemon, and lychee (Chaikiattiyos et al., 1994).

Stomatal conductance

Plants exhibit two basic stomatal-related responses to water stress which help to

alleviate the stress (Mansfield and Atkinson, 1990). The first response is a shift in

distribution of new assimilates from shoots to roots to support root growth at the expense

of shoot growth. This increases root area and the capacity of the root to obtain water and

restricts leaf expansion to minimize water loss through E. The second response of plants

to water stress is an adjustment in the degree of stomatal opening to inhibit water loss by

E. A reduction of A and E of grapevines (Nagarajah, 1989), and kiwifruit (Chartzoulakis

et al., 1993) among other crops was attributed to stomatal closure caused by reduced ΨL.

Flore et al. (1985) found that stomata of well-watered apple, cherry, and blueberry plants

were more responsive to vapor pressure deficit and less responsive to the CO2 gradient

39

than stressed plants. In peach, leaf gs was reduced in water-stressed trees compared to

well-watered trees (Tan and Layne, 1991).

Stomata of most plants close in response to an increase in the VPD between the leaf

and the surrounding air, which influences the turgor of stomatal guard cells (Lakso 1985;

Mansfield and Atkinson, 1990). Other factors that play a role in stomatal closure under

water stress conditions include carbon dioxide (CO2) and leaf abscisic acid (ABA)

concentration (Flore et al., 1985).

The level of ABA in the leaf increases during water stress in many fruit crops

including apples (Davies and Lakso, 1978) and grapes (Nagarajah, 1989). Abscisic acid

is a major signaling phytohormone for water stress that is synthesized in leaves and in

roots (Bray, 1997; Jackson, 1997; Mansfield and Atkinson, 1990; Schulze, 1986; Seeley,

1990). Abscisic acid concentration increased linearly with decreased leaf turgor in apple

seedlings (Davies and Lakso, 1978). Turgor-controlled ABA synthesis is essential for

regulating water loss under water stress. At the early stage of water stress, ABA is

translocated in the E stream through xylem vessels from the roots to the leaves to bring

about stomatal closure, and more ABA is synthesized as water stress progresses

(Mansfield and Atkinson, 1990; Comstock, 2002). Stomatal closure in response to water

stress is vital in perennial fruit crops because plant Ψ is dependent on E (Lakso, 1985).

Leaf and stem water potential

Water potential is a measure of the chemical potential of water and is essential in

determining the direction of water flow (from higher to a lower potential), as in the E

stream (Jones et al., 1985; Jones, 1990; Plaut and Meiri, 1994). Water potential has been

used as a measure of plant water status and consists of osmotic potential (Ψo) pressure or

turgor potential (Ψp), matric potential (Ψm), and gravitational potential (Ψg) (Jones et al.,

40

1985). Gravitational potential (Ψg) increases by 0.01 MPa per 1 m of plant height,

therefore it is less than 0.05 MPa for most trees and is usually ignored when assessing a

tree’s water status (Jones et al., 1985). A reduction in osmotic potential (Ψo) due to

increased solute content or decreased water content (dehydration) is a result of water

stress (Plaut and Meiri, 1994). Turgor potential (Ψp) is considered a better measure of

plant water status than osmotic potential (Ψo) (Hsiao and Bradford, 1983) and its

reduction as a result of water stress can be detrimental to physiological processes related

to growth. However, direct measurement of plant turgor pressure is difficult under field

conditions (Zimmermann, 1978). Leaf xylem water potential as an indicator of plant

water status can be determined using a pressure chamber (Scholander et al., 1965).

Different physiological processes vary in their sensitivity to reduced ΨL caused by

water stress. Cell expansion can be inhibited at a ΨL of approximately –0.4 to –0.8 MPa

(Plaut and Meiri, 1994), while photosynthetic CO2 fixation is much less sensitive to

changes in ΨL and can be reduced at ΨL of –1 to –2 MPa (Hsiao and Acevedo, 1974). In

citrus, midday Ψ is stabilized by effective stomatal control of leaf water status under a

wide range of evaporative demands (Levy, 1980). For plants that have been moderately

water stressed, midday ΨL can be higher than that of well-watered plants due to stomatal

closure (Levy, 1980). Similar effective stomatal control of ΨL has been reported in apple

(Jones et al., 1983). In grapevines, Nagarajah (1989) indicated that water stress reduced

both water and turgor potential. The reduction in ΨL induced stomatal closure and

increased ABA levels in the leaves. Stomatal closure reduced A and E (Nagarajah, 1989).

Water potential of lychee trees declined to –2.6 to –2.8 MPa under water stress where soil

water content reached minimum of 7.6% compared to 15.5% for well watered soils

41

(Menzel et al., 1995). Predawn ΨL differences were more significant among water deficit

treatments than midday ΨL in peach trees (Girona et al., 1993). In carambola, Ismail et al.

(1994) reported a decrease in ΨL and relative water content with increased water stress.

They also found that ΨL at predawn was – 0.65 MPa and – 0.9 MPa at midday, while ΨL

after 4 weeks of withholding water was – 2.52 MPa at midday. With decreasing soil

water content, water uptake in trees during the afternoon and night becomes increasingly

less resulting in low pre-dawn ΨL (Braun et al., 1989).

Since Ψp = Ψ – Ψo, during stress, as Ψ decreases, a positive Ψp is maintained by a

reduction in Ψo (Nagarajah, 1989). This adaptive response by which Ψo is reduced in

response to water stress is called osmotic adjustment (Begg and Turner, 1976; Hsiao and

Bradford, 1983). Maintaining positive Ψp is essential for the continuation of many

processes in the plant such as cell expansion, stomatal movement, and enzymatic

reactions (Nagarajah, 1989). Osmotic adjustment may occur in different parts of the plant

and varies depending on the level of stress, species and variety, season and growth stage.

In grapevines for example, osmotic adjustment takes place in leaves or shoot apices but

not in fruit or root apices (Nagarajah, 1989). In apple fruit, osmotic adjustment occurred

during the enlargement phase but not in the cell division phase of fruit growth and

development (Failla et al., 1992). As stress developed in apple, sorbitol, glucose, and

fructose concentrations in the leaves increased, while sucrose and starch levels decreased

significantly, strongly suggesting that sugar alcohol and monosaccharides are the most

important osmotica for osmotic adjustment in apple (Wang and Stutte, 1992).

42

Irrigation Scheduling of Fruit Trees

Optimizing water applications to fruit orchards may increase water conservation,

reduce production costs, and increase plant growth and yield. Irrigation scheduling is

especially important in horticultural crops because net returns are normally higher than

those of other crops (Fereres, 1997). According to Hillel (1998), the main issue with

irrigation management is to determine the frequency, quantity, and timing of irrigation to

optimize crop growth and productivity. Fereres (1997) stated that irrigation scheduling of

horticultural crops is essentially reduced to predicting the amount of water required to

meet the crop demand. The basis for predicting a crop’s water requirements is an

estimation of crop ET (Fereres, 1997).

Measurement of soil water content and soil matric potential provide an index of the

rate at which water is taken up by the plant or lost from the root zone. Soil water content

or potential is therefore most useful in conjunction with other information about the soil-

plant-atmosphere system (Campbell and Campbell, 1982). Water extraction from the soil

is dependant on plant properties that determine the plant Ψ at which a particular plant

species can continue to grow and extract water from the soil (Jones, 1990).

The majority of subtropical and tropical fruit crops in South Florida are irrigated at

rates and frequencies based on experience and observations of crop growth and yield

rather than on quantitative scientific information (Schaffer, 1998). Li (2000) suggested

several methods of irrigation scheduling for tropical fruit orchards in South Florida.

These included irrigation during flowering at soil tension (determined with tensiometers)

of 10-15 kPa and at 15-20 kPa at other times. He also suggested irrigating midway

between field capacity and the onset of stress as determined with an EnviroSCAN

43

multisensor capacitance probe system and associated software. Li (2000) also suggested

indirect methods of irrigation scheduling based on the following equation:

A = ET·L·W·15.8, where: A is the amount of water required (L/tree per day), ET is

evapotranspiration (mm·d-1), L is the spacing between trees within the row (m), and W is

the spacing between rows (m).

Although the previously suggested methods provide a means for estimating

irrigation amount and timing, they do not take into account the variability between fruit

tree species and cultivars, growth stage, or the response of trees to soil moisture deficit.

The water refill point, which is the lowest possible soil water content with no decrease in

yield or fruit quality, varies among different tree varieties, rootstocks, soils and seasons

(Braun et al., 1989). It is often beneficial to use both soil and plant factors for irrigation

scheduling. An integrated approach utilizing soil and plant factors suggested by Buss

(1989) included soil surveying to determine soil properties and available soil moisture

content and crop factors that included crop type, canopy size, rooting width and depth,

and crop density. Irrigation scheduling was then planned from a combination of available

water data, crop water extraction rates, and the irrigation system layout.

Irrigation scheduling based on plant water status

The effect of several factors that control water status were summarized by Jones

(1990) and divided into four groups: (a) the soil Ψ and the factors involved in its control.

These include rainfall and water use, soil texture and depth, and root distribution and

activity; (b) crop evapotranspiration (ET), historic and current; (c) hydraulic

conductances within the plant that are important for controlling water; (d) the

relationships between plant Ψ and other measures of plant water status (i.e. turgor

44

pressure, osmotic potential, or water content) which can be affected by processes such as

osmotic adjustment, cell wall elasticity and mesophyll structure.

Several techniques used to schedule irrigation including destructive and non-

destructive, and discrete and continuous water status measurements in plant populations,

individual plants, or plant organs. Methods of irrigation scheduling based on plant

responses to soil water include the use of remote sensors such as radiation sensors and

infrared sensors, and measurements of trunk diameter, ΨL, and stem water potential (ΨS),

and sap flow. Irrigation scheduling based on atmospheric or soil variables include basing

irrigation on ET or soil water status monitored with devices such as tensiometers, neutron

probes, or capacitance probes. Several physiological variables are used as indicators of

plant water status. Among the most frequently used is ΨL (Hsiao, 1990).

Plant water potential. Plaut and Meiri (1994) defined Ψ as the quotient of the

chemical potential of water and its partial molal value (18 cm3 mol-1). The chemical

potential of water is a qualitative expression of the free energy of water and is influenced

by three factors: concentration, pressure, and gravity, all expressed as the difference

between the chemical potential at a given state and its value at a standard state (Plaut and

Meiri, 1994). The Ψ concept satisfies two main functions (Plaut and Meiri, 1994): (a) it is

the driving force of water movement from the soil to the root and it controls the direction

of water flow across cell membranes; and (b) it is a measure of plant water status that can

be used as a plant stress indicator and for irrigation scheduling. Naor et al. (1999) found

that total yield and fruit size of nectarines grown in a semiarid zone were more highly

correlated with midday ΨS than with soil Ψ. Similar results were reported for apple (Naor

et al., 1995) and pear (Shackel et al., 1997). A low correlation between yield of nectarine

45

and soil Ψ was attributed to the typically high variability in tensiometer readings (Naor et

al., 1999).

Several techniques have been developed to determine plant Ψ. Among the most

commonly used are the pressure chamber and thermocouple psychrometer (Turner, 1981;

Phene et al., 1990). A pressure chamber is used to determine plant ΨL (Scholander et al.,

1965). With this technique, the leaf or branch is cut and enclosed in a steel pressure

chamber with the cut end protruding through a rubber stopper which is used to seal the

chamber. The pressure in the chamber is gradually increased until the sap appears at the

end of the xylem vessels. After the pressure is recorded, it is released through an outlet

valve and the sample is removed (Turner, 1981; Phene et al., 1990; Shackel et al., 1997).

It is often necessary to use a microscope or hand lens to observe the point at which sap

first starts to exude from the xylem. The entire procedure can be completed in 2 min for

each leaf, depending on the dehydration level of the tissue. This allows for the

measurement of about 20 to 30 samples in an hour (Turner, 1981). To avoid water loss,

leaves should be enclosed in a plastic bag prior to excision (Turner and Long, 1980).

Stem water potential can also be measured with the pressure chamber by covering the

leaves prior to measurement for a minimum of 30 min (Begg and Turner, 1970). Shackel

et al. (1997) found that in prune trees, ΨS was more sensitive than ΨL as an indicator of

progressive water stress and suggested its use as a guide for making irrigation decisions.

Irrigation scheduling based on soil water status

Irrigation scheduling based on grower observations of soil moisture is an ancient

practice and is still the major method of irrigation timing used by fruit growers. Soil

water is usually measured as water content (mass or volume fraction) or as soil Ψ (i.e.,

the energy required to remove water from the soil). Measurements of soil water content

46

can be direct (gravimetric or volumetric methods) and indirect that involve installation of

sensors that measure soil water content, such as capacitance probes and time domain

reflectometry, or portable equipment such as the neutron scattering probe, or measuring

soil matric potential using tensiometers.

Gravimetric or volumetric method. The gravimetric or volumetric method is a

direct measurement of soil water content. The percent of water is measured per mass or

volume of soil using the ratio of dry weight to wet weight of the soil sample (Phene et al.,

1990). This is considered a destructive method since it involves soil sampling from the

field (Campbell and Campbell, 1982). Volumetric or gravimetric methods require drying

the soil, and therefore may take too long for making decisions about when to irrigate

which often need to be made instantaneously.

Soil matric potential (tensiometers). Tensiometers were developed in the 1920s

(Richards, 1942) and measure soil suction or matric Ψ rather than soil water content

(Muñoz-Carpena et al., 2002; Phene et al., 1990; Smajstrla and Harrison, 1998). Soil

suction or matric potential is defined as the amount of water held by capillarity against

the force of gravity and is often measured in units of pressure such as kPa (Muñoz-

Carpena et al., 2002). Soil water tension is also a measure of the energy required for

plants to extract water from the soil (Smajstrla and Harrison, 1998). A soil water

retention curve must be established to determine the soil water content that corresponds

to the water matric potential in order to estimate soil water content with a tensiometer.

The main components of a tensiometer include a porous ceramic cup, a water column,

and a pressure-measuring device (usually a vacuum gauge). Tensiometers are filled with

water then calibrated with a manual pump or vacuum chamber (Smajstrla and Pitts, 1997)

47

to remove air from the water column, to check for leakage, and to ensure accurate gauge

readings. There must be good contact between the ceramic cup and the surrounding soil

(Phene et al., 1990) to insure the accuracy of the reading. Therefore, for accurate

tensiometer reading in porous, rocky soil such as Krome very gravelly loam, it is

essential to apply soil slurry around the ceramic cup when tensiometers are installed

(Núñez-Elisea et al., 2001).

Tensiometers are relatively simple to use and inexpensive (Campbell and

Campbell, 1982). A disadvantage of using tensiometers for determining soil water

content is that they are limited to measuring soil matric potential up to 40 kPa for low-

tension tensiometers and 80 kPa for high-tension tensiometers (Campbell and Campbell,

1982; Phene et al., 1990). At higher tensions, air enters the tensiometer cup (Phene et al.,

1990) and makes the readings inaccurate. Moreover, tensiometers require frequent

servicing to function properly. Also, they measure soil matric potential within a very

limited volume; therefore, several tensiometers are needed for a reliable measurement

(Campbell and Campbell, 1982) of soil water matric potential.

In Krome very gravelly loam soils in southern Florida, soil water tension readings

from tensiometers were more highly correlated with volumetric water content when

tensiometers were installed at a depth of 10 than at 30 cm (Núñez-Elisea et al., 2001).

Lower correlation at 30 cm depth was due to the presence of more rock fragments larger

than 2 mm at the 30 than at the 10 cm depth (Núñez-Elisea et al., 2001). However, air

entry into tensiometers at tensions greater than 20 kPa was observed in Krome very

gravelly loam soils (Al-Yahyai et al., 2003), where tensiometer readings did not reflect

changes in soil water content (Muñoz-Carpena et al., 2002).

48

Neutron probe. Neutron probes measure soil volumetric water content. The main

component of a neutron probe is a fast neutron source encased in a protective shield and

an electronic counting scaler which are connected by an electric cable that is also used to

lower the probe into an access tube to determine soil moisture in the soil profile

(Chanasyk and Naeth, 1996). Neutrons with high energy are emitted by a radioactive

source, such as americium 241/beryllium, into the soil and are slowed down by elastic

collisions with nuclei of atoms, primarily hydrogen in the soil (Gardner and Kirkham,

1952). The density of the resultant cloud of slow neutrons is a function of the soil

moisture content (Chanasyk and Naeth, 1996). The number of fast neutrons that are

slowed is measured as a ‘count rate’ per unit time (Gardner et al., 1991). The count rate is

converted to volumetric water content using a calibration curve. A polyvinyl chloride

(PVC) or aluminum access tube is inserted vertically into the soil and the neutron probe

is inserted into the tube at various depths to determine water content at different depths in

the soil profile (Burman et al., 1983). Determination of soil water content using the

neutron probe is calculated from the ratio of the count taken at any depth to the mean

standard count (Phene et al., 1990). However, measurements close to the soil surface are

unreliable (Cole, 1989, and Gardner et al., 1991) because fast neutrons can escape to the

atmosphere (Cuenca, 1989). Neutron probes have been widely used for measuring soil

moisture content, but the radioactive source requires careful handling and is concern to

some potential users.

Time Domain Reflectometry (TDR). Time domain reflectometry (TDR) is used

to measure soil water content based on the propagation of an electric signal which is

dependent on the dielectric constant of the soil (Phene et al., 1990; Paltineanu and Starr,

49

1997). Parallel transmission lines are installed in the soil and serve as conductors, while

the soil serves as the dielectric medium between them (Phene et al., 1990). A pair of rods

acts as wave-guides and the signal propagates as a plane wave in the soil (Topp and

Davis, 1985). A high-frequency signal is delivered from the TDR receiver to the

transmission lines. The signal is reflected from the end of a transmission line (Phene et

al., 1990; Paltineanu and Starr, 1997) back to the TDR receiver. A timing device

measures the time between sending and receiving the reflected signal. This time interval

relates directly to the velocity of the signal in the soil (Topp and Davis, 1985). The

velocity and amplitude of the reflected signal is used to calculate the dielectric constant.

There is a strong relationship between the dielectric constant and volumetric soil water

content (θV) (Topp and Davis, 1985). Topp et al. (1980) stated that the determination of

θV by TDR is accurate without having to calibrate the system in each soil type. On the

contrary, Werkhoven (1993) suggested calibrating TDR against volumetric soil water

content for each soil in order to get accurate determinations of soil water content. Time

domain reflectometry is useful in studies of real-time soil water content dynamics

(Platineanu and Starr, 1997). However, this technique works well only in small areas due

to the limitation of cable length, which should not be more than 25 m (Platineanu and

Starr, 1997) in order not to interfere with the reflected signal.

Capacitance probes. Multisensor capacitance probes measure soil moisture

content based on the dielectric constant of the soil (Phene et al., 1990; Paltineanu and

Starr, 1997; Wu, 1998). The dielectric constant of the soil is composed of the dielectric

constants of water (80.4), soil particles (3-7) and air (1) (Robinson and Dean, 1993;

Paltineanu and Starr, 1997; Starr and Paltineanu, 1998a, 1998b; Wu, 1998). Since the

50

dielectric constant of the soil particles and air are small and relatively constant compared

to that of water, changes in the dielectric constant of the soil are primarily a measure of

the change in soil water content. A commercially available capacitance probe system

called the EnviroSCAN (Sentek PTY Ltd., Kent Town, Australia) is composed of

multisensor capacitance probes with capacitance sensors, a data-logger and an internal

power supply charged by a solar panel or by an external battery power (Paltineanu and

Starr, 1997). The probes are inserted in a PVC tube at the desirable depth for measuring

soil water content and then connected to data-logger via a cable (Paltineanu and Starr,

1997). Prior to installation, the data-logger is configured to the number of probes and

sensors that will be installed, and normalized for air and water counts by inserting the

probe in a tube surrounded by water. The system can be set up to record the sensor

reading in intervals, ranging from 1 to 9,999 min and provides real-time measurements of

the change in volumetric soil water content. The data can be downloaded to a computer

(Paltineanu and Starr, 1997; Starr and Paltineanu, 1998a, 1998b; Mead et al., 1994) and

sensor readings can be converted to volumetric soil water content using calibration

equations stored in the data-logger (Paltineanu and Starr, 1997). The system comes with

software that is designed for irrigation scheduling.

Conclusions

The goal of irrigation scheduling of fruit crops is to meet tree water requirements

for optimum yield by determining the water quantity and frequency of irrigation. Lack of

irrigation scheduling may lead to excessive or inadequate water application. Excessive

irrigation can result in waterlogging, potential chemical leaching from the root zone into

groundwater, and increased production costs. On the other hand, a poor estimation of tree

water requirements may result in water stress resulting in adverse effects on tree

51

physiological processes, water relations, vegetative growth, flower and fruit production

and quality.

Measurements of soil water content, plant water status or climatic data are all

practical and feasible methods of irrigation scheduling of fruit crops. Irrigation

scheduling based on soil water content has been widely adopted due to the availability

and ease of use of various devices that monitor soil water status either discretely (e.g.

tensiometers, neutron probes) or continuously (e.g. multisensor capacitance probes).

Irrigation scheduling based on tree water status has gained interest in recent

decades. Several studies have determined that tree stem water potential, gas exchange

measurements, sap flow, and stem daily shrinkage, among other tree measurements, can

be utilized for making decisions about irrigation of fruit crops. For several fruit crops,

yield is more highly correlated with ΨS than with soil water potential, thus validating the

use of basing irrigation decisions on plant water status rather than soil water content.

There have been a few studies comparing soil water content to plant water physiological

variables as bases for irrigation scheduling.

Carambola is cultivated in many tropical and subtropical regions of the word. It

has among the highest economic returns of any fruit crop in southern Florida. The tree is

well adapted to Krome very gravelly loam soils, and produces two crops per year. Tree

water requirements have not been determined; neither has the long-term response to

various levels of soil moisture under subtropical climatic conditions of South Florida.

Therefore, carambola can serve as a good model crop for comparing soil water content to

plant water status determinations for making irrigation decisions for subtropical fruit tree

species. Determination of soil water content can be achieved by evaluating and

52

calibrating various devices for use in Krome soils, as well as determining the impact of

various soil moisture levels on carambola tree phenology, physiology, growth, yield, and

fruit quality.

CHAPTER 3 CALIBRATION OF THE NEUTRON PROBE AND MULTISENSOR CAPACITANCE

PROBES IN A GRAVELLY LOAM CALCAREOUS SOIL

There is a lack of quantitative information about water requirements of tropical

fruit crops in calcareous soils. Several methods and devices have been tested in these

soils, including, tensiometers, neutron thermalization probes, and capacitance probes.

Gravimetric determination of water content is generally not practical for irrigation

scheduling in these very hard soils because sampling at different depths in the soil profile

is labor intensive, time consuming, and requires special digging tools. In calcareous soils

in South Florida, tensiometers (Richards, 1942) have been used for irrigation

management of vegetable (Li et al., 1998) and fruit crops (Núñez-Elisea et al, 2000) due

to their low cost and simple installation. However, because of the gravelly texture of soil,

tensiometers were not effective with tension greater than 20 kPa due to discharge from

the water column or air entering into the suction cup (Al-Yahyai et al., 2003; Núñez-

Elisea et al, 2001). Fruit trees can tolerate higher soil tensions than herbaceous annual

crops where irrigation at soil tensions less than 20 kPa are essential to avoid water stress

and reduction in growth and yield (Li et al., 1998). Therefore, tensiometers may not be

practical tools for irrigation scheduling of fruit crops in oolitic limestone soils.

Monitoring soil water content using a neutron probe is a reliable method of irrigation

scheduling for fruit crops.

Monitoring soil water content based on the dielectric constant of the soil has long

been used in time domain reflectometry (TDR) (Topp et al., 1980). Technology utilizing

53

54

this concept is employed in multisensor capacitance probes, whereby sensors installed at

various depths in the soil profile can measure soil water content continuously (Paltineanu

and Starr, 1997; Phene et al., 1990; Wu, 1998). A multisensor capacitance probe system

(EnviroSCAN, Sentek PTY Ltd., Kent Town, Australia) was found useful for irrigation

scheduling of lime, avocado and carambola trees (Zekri et al., 1999; Núñez-Elisea et al,

2001) in calcareous soils in South Florida. Due to the variability in soil properties,

calibration equations for each soil type are essential to accurately assess soil water

content (Baumhardt et al., 2000; Hanson and Peters, 2000; Morgan et al., 1999;

Paltineanu and Starr, 1997). Depending on the nature of the soil, calibration of the

EnviroSCAN system can be carried out in the field (Morgan et al., 1999) or in containers

(Paltineanu and Starr, 1997). A calibration curve has been developed for the fine sandy

soils of Florida (Morgan et al., 1999), but the system has not been calibrated for use in

oolitic limestone soils in the tropical fruit production areas of southern Florida.

The relationship between soil water tension as measured with tensiometers and

volumetric water content is known as a soil water characteristic curve or soil water

retention curve. This curve changes with soil texture and it is unique to each soil. Soil

water retention curves for a calcareous soil in South Florida have been described by

Muñoz-Carpena et al. (2002) and Núñez-Elisea et al. (2001) in an orchard using

capacitance probes to directly measure volumetric water content. Núñez-Elisea et al.

(2001) found that a soil water retention curve was better correlated with soil water

tension at 10 than at 30 cm due to the difference in soil texture between the two depths.

In a carambola orchard in a calcareous soil in South Florida, Al-Yahyai et al. (2003)

observed that tensiometer readings did not reflect volumetric soil water content at a 10-

55

cm depth when soil water tension reached greater than 10 kPa. The large changes in

tension that correspond to small changes in soil water content is a characteristic soil water

retention pattern of rock-plowed calcareous soils of South Florida (Muñoz-Carpena et al.,

2002). The objective of this study was to calibrate a neutron probe and a multisensor

capacitance probe system (EnviroSCAN) in Krome very gravelly loam soil, a calcareous

soil classified as a loamy-skeletal, carbonatic, hyperthermic, Lithic Udorthents (Noble at

al., 1996).

Materials and Methods

Capacitance probe calibration. Calibration of multisensor capacitance probes was

conducted in 15 L containers of Krome gravelly loam soil [loamy-skeletal, carbonatic,

hyperthermic, Lithic Udorthents (Noble et al., 1996)] and placed above ground on a

perforated plastic pad in Homestead, Fla. The soil was obtained from a location adjacent

to a tropical fruit orchard, oven dried at 105 oC for 5 d and then placed in the containers.

The soil was packed every 5 cm as it was added to the container to maintain equal bulk

density of approximately 1.4 g·cm-3 throughout the soil profile to a total volume of 11.44

L. One probe was installed in a polyvinyl chloride (PVC) access tube. The capacitance

sensor was centered at 5 cm below the soil surface. The radius of the soil in the

containers was approximately the same radius of the soil water detection sphere of the

capacitance sensor (10 cm). At the beginning of the experiment, the containers were

placed inside larger containers and gradually flooded from the bottom to displace air

from the soil. The soil was saturated for 3 h and then the containers were weighed every 2

h, three times on the first day and twice a day (at 1000 and 1600 HRS) thereafter for 13 d.

The capacitance probe system was set to record data every 10 min for the duration of the

experiment.

56

Gravimetric soil water content was determined by weighing the containers.

Volumetric water content (θv) was calculated as the percentage of depleted soil water per

time interval divided by the soil volume of the container. The calibration equation for the

multisensor capacitance probe was obtained from the relationship between volumetric

soil water content and capacitance probe scaled frequency in order to obtain the

coefficients for the system to correctly calculate the volumetric soil water content. The

scaled frequency was calculated from the following equation [SF = (Fa-Fs)/(Fa-Fw)],

where SF is the scaled frequency of oscillation of the EnviroSCAN capacitance sensor,

Fs is the sensor frequency reading in the soil, Fa is the sensor frequency reading in the

air, and Fw is the sensor frequency reading in the water (Paltineanu and Starr, 1997).

Neutron probe calibration. A neutron probe (Model 503DR, Campbell Pacific

Nuclear, Inc., Martinez, Calif.) was calibrated in 95 L containers in Krome very gravelly

loam soil. The soil volume in the container was within the soil water detection sphere of

the neutron probe. The radius (cm) of the detection sphere was 21.4 cm as calculated with

the Kristensen (1973) equation (R = 100 / (1.4 + 0.1W), where W is the volumetric water

content which was 32.71%. One PVC access tube was placed in the center of each of the

three containers for neutron probe measurements. The containers were filled with Krome

very gravelly loam soil that was collected from a site adjacent to a tropical fruit orchard

in Homestead, Fla. Prior to filling the containers, the soil was oven dried at 105 oC for 5

d. The soil was packed from the bottom of the container every 10 cm to maintain a bulk

density of about 1.4 g·cm3 throughout the soil profile. The total soil volume was 88.45 L

per container. The containers were placed above ground on a perforated plastic pad. The

bottom of the containers was flat with multiple holes to allow free water drainage. The

57

containers were filled with water and the soil was allowed to become saturated overnight.

Prior to installing the probe into the soil, 10 standard counts were taken 1 m above the

soil surface. Raw counts were recorded during 32-s intervals at a depth of 20 cm once a

day (at 1600 HR) and the containers were weighed simultaneously with a heavy-duty

scale (Model CD-11, OHAUS Corp., Florham Park, N.J.). The calibration equation for

the neutron probe was obtained from the relationship between the soil water content and

the neutron probe count ratio. Count ratios were calculated by dividing the raw counts

taken at a 20-cm depth by the average standard count taken at 1 m above ground level

while the probe was in the shield (Hanson and Dickey, 1993).

Field soil water retention curves. Soil water retention curves were determined in

an 8-year-old carambola orchard in Krome very gravelly loam soil in Homestead, Fla.

Soil matric potential was determined with using low-tension tensiometers (0 to 40 kPa)

(Model LT, Irrometer Co., Inc., Riverside, Calif.), and soil volumetric water content

measured with multisensor capacitance probes (EnviroSCAN, Sentek PTY Ltd., Kent

Town, Australia) and a neutron probe (Model 503DR, Campbell Pacific Nuclear, Inc.,

Martinez, Calif.). Tensiometers were calibrated prior to installation using a vacuum

chamber (Smajstrla and Pitts, 1997) to ensure that the water column in the tensiometer

was free of air, that there were no leaks, and to standardize gauge readings among

tensiometers. Three tensiometers were installed 60 cm from the trunk of each of four

trees at 10, 20 and 30 cm depths.

Prior to installation, a hole was made in the soil, slightly larger than the diameter of

the tensiometer. A slurry prepared with sieved Krome very gravelly loam soil mixed with

water was poured into the tensiometer hole to ensure that the ceramic cup was in contact

58

with the soil (Núñez-Elisea et al., 2001). Tensiometers were regularly maintained in the

field using a hand-pump and water was refilled whenever it drained from the tensiometer

tube.

One 85-cm long polyvinyl chloride (PVC) access tube was installed at 60 cm from

the trunk of each of the four replicate trees. The neutron probe was placed in the access

tubes and lowered to 10, 20, and 30 cm depths. Neutron probe readings were recorded for

32 s at each depth.

Multisensor capacitance probes were installed in the orchard following the

procedure described by Paltineanu and Starr (1997). One probe was installed inside a 74

cm-long PVC access tube at 60 cm from the trunk of each of four replicate trees. Three

sensors were placed in each probe at 10-, 20-, and 30-cm depths. The PVC tubes were

installed with a motorized drill and slurry was then added to the hole. The slurry

consisted of 2:1:1 by volume of calcareous rock, cement, and water to prevent air gaps

from forming between the tubes and the surrounding soil (Núñez-Elisea et al., 2001).

The sensors in each probe were connected to a data-logger powered by a 12-volt battery

charged by a solar panel. Data were recorded every 30 min and downloaded from the

data-logger to a portable laptop computer where graphs of soil water depletion rates at

each soil depth and location were created with EnviroSCAN software.

Soil water retention curves were calculated based on van Genuchten’s model

(1980):

θ(Ψ) = θr + (θs – θr) [1 + (α|Ψ|n)]-m

59

Where Ψ is the soil matric potential, θr is the residual water content, θs is the

saturated water content, and α, n, and m are fitting parameters directly dependent on the

shape of the θ(Ψ) curve.

Results and Discussion

Capacitance probe calibration. The relationship between capacitance probe

scaled frequency (SF) and volumetric soil water content (θv) was linear (SF = 0.011 θv +

0.5206; r2 = 0.98) (Fig. 3-1). The coefficients from this equation were A = 0.011, B = 1,

C = 0.5206. These coefficients can be used instead of the default EnviroSCAN equation

for an accurate determination of volumetric water content in Krome soil. Volumetric soil

water content (θv) is calculated by the EnviroSCAN based on the following equation

(Paltineanu and Starr, 1997):

B

ACSFv

1

⎟⎠⎞

⎜⎝⎛ −

=θ , where θv is volumetric water content and A, B, and C are the

coefficients determined above.

Neutron probe calibration. The relationship between volumetric soil water

content and neutron probe count ratio (x) was linear (θv = 29.05 x - 6.4; r2 = 0.95) (Fig. 3-

2). Calibration relationships of neutron probes are generally linear through the common

range of soil moisture (Merriam and Knoerr, 1961).

Field soil water retention curves. In Krome soil, soil water tension measured with

a tensiometer was more closely correlated with capacitance probe measurements of

volumetric water content at a depth of 10 cm (r2 = 0.86) than at 20 cm (r2 = 0.62 ), with

the lowest coefficient of determination at 30 cm (r2 = 0.38) (Fig. 3-3, Table 3-1). This is

in agreement with data obtained for Krome soil by Núñez-Elisea et al. (2001), who found

that the volumetric soil water content determined with multisensor capacitance probes

60

was less strongly correlated with soil water tension at a 30-cm than at 10-cm depth due to

the larger percentage of gravel at the 30-cm depth. Soil water retention curves for Krome

soil can be divided into two tension ranges: from 0-10 kPa for gravel and 10-40 kPa for

loam soils, where volumetric water content did not correspond to a change in soil tension

(Muñoz-Carpena et al., 2002).

Soil water retention curves were obtained from fitting volumetric soil water content

measurements using a neutron probe with tensiometer readings based on van

Genuchten’s model (1980) (Fig. 3-4). Coefficients of determination between observed

and fitted data of the soil water retention curves were higher than those obtained from the

capacitance probes. Coefficients of determination at different depths in the soil profile

were 0.94 at 10 cm, 0.68 at 20 cm, and 0.88 at 30 cm (Table 3-2). Volumetric soil water

content measurements obtained with the neutron probe were more closely correlated to

soil water tension than those obtained by capacitance probes. The capacitance sensors

may have been less accurate than the neutron probe due to an increased amount of air

from coarse textured soil which was greater at lower depth in the soil profile (Núñez-

Elisea et al., 2001). Differences in precision of the neutron probe and capacitance probes

can be caused by differences in interference, measurement methods, and sphere area of

the soil water detection zone (Paltineanu and Starr, 1997). Nonetheless, several studies

suggested that the neutron probe was more consistent than capacitance probes under field

conditions (Evett, 2000; Evett and Steiner, 1995; Evett et al., 2002; Heng et al., 2002).

This is especially true in coarse textured soil such as Krome very gravelly loam where the

potential for air gaps is high (Al-Yahyai et al., 2003; Tomer and Anderson, 1995).

61

Conclusions

The calibration equation obtained for the capacitance probe was y = 0.011x +

0.5206; r2=0.98, and that for the neutron probe was y = 29.05x-6.4; r2=0.95.

In Krome very gravelly loam soil, soil water retention curves of tension vs.

volumetric water content using capacitance probes or a neutron probe were fitted with

van Genuchten’s model. The r2 values indicated that tensiometers are more highly

correlated with neutron probe measurements of volumetric soil water content than with

multisensor capacitance probe measurements. This suggests that capacitance probe

measurements of soil water content were more variable in response to changes in soil

texture throughout the soil profile.

62

Table 3-1. Fitted parameters of van Genuchten’s model used to describe soil water retention curves of Krome very gravelly loam soil where soil water content was measured in a carambola orchard using multisensor capacitance probes (EnviroSCAN) at soil depths of 10, 20, and 30 cm.

Parametersy 10 cm 20 cm 30 cm

θs 0.47 0.47 0.47

θr 0.20 0.29 0.28

α 0.72 0.87 0.10

n 1.56 1.98 22.51

m 0.36 0.50 0.96

r2 z 0.86 0.62 0.38 zr2 values for observed vs. fitted values, n = 4. yθs is saturated water content, θr is residual water content, and α, n, and m, are model fitting parameters.

63

Table 3-2. Fitted parameters of van Genuchten’s model used to describe soil water retention curves of Krome very gravelly loam soil where soil water content was measured in a carambola orchard using a neutron probe at soil depths of 10, 20, and 30 cm.

Parametersy 10 cm 20 cm 30 cm

θs 0.47 0.47 0.470

θr 0.06 0.19 0.08

α 0.44 4.10 4.63

n 1.22 1.22 1.14

m 0.18 0.18 0.13

r2 z 0.94 0.69 0.88 zr2 values for observed vs. fitted values, n = 4. yθs is saturated water content, θr is residual water content, and α, n, and m, are model fitting parameters.

64

θv (%)

6 8 10 12 14 16 18 20 22 24

Sca

led

Freq

uenc

y (S

F)

0.56

0.58

0.60

0.62

0.64

0.66

0.68

0.70

0.72

0.74

0.76

0.78

Fig. 3-1. Relationship between scaled frequency (SF) from a multisensor capacitance probe (EnviroSCAN) and volumetric water content (%) fitted to obtain calibration coefficients for Krome very gravelly loam soil in containers, y = 0.011x + 0.5206; r2 = 0.98.

65

Count Ratio

0.9 1.0 1.1 1.2 1.3 1.4

θ v (%

)

20

22

24

26

28

30

32

34

Fig. 3-2. Relationship between count ratio from the neutron probe and volumetric water content (%) fitted to obtain a calibration equation for Krome very gravelly loam soil in containers, y = 29.05x - 6.4; r2 = 0.95.

66

10 cmr2 = 86

θ v (%

)

0

10

20

30

40

T(cm) vs ES-10 h=(T*10) vs ES-10-k

20 cmr2 = 62

θ v (%

)

0

10

20

30

40

30 cmr2 = 38

Soil water tension (cm)

0 20 40 60 80 100 120 140

θ v (%

)

0

10

20

30

40

Fig. 3-3. Soil water retention curves for Krome very gravelly loam soil at 10-, 20-, and 30-cm depths based on van Genuchten’s model calculated from volumetric water content (θv) measured with capacitance probes and soil water tension measured with tensiometers.

67

10 cmr2 = 94

θ v (%

)

0

10

20

30

40

20 cmr2 = 68

θ v (%

)

0

10

20

30

40

30 cmr2 = 88

Soil water tension (cm)

0 20 40 60 80 100 120 140

θ v (%

)

0

10

20

30

40

Fig. 3-4. Soil water retention curves for Krome very gravelly loam soil at 10-, 20-, and 30-cm depths based on van Genuchten’s model calculated from volumetric water content (θv) measured with a neutron probe and soil water tension measured with tensiometers.

CHAPTER 4 MONITORING SOIL WATER CONTENT FOR IRRIGATION SCHEDULING IN A

CARAMBOLA ORCHARD IN A GRAVELLY LIMESTONE SOIL

There are approximately 100 ha of carambola trees in Florida (J. H. Crane,

University of Florida, personal communication), of which 46 ha are in Miami-Dade

County (Degner et al., 2002). A sweet-type, ‘Arkin’, is the leading commercial

carambola cultivar in Florida (Knight and Crane, 2002).

Excessive soil water content (Joyner and Schaffer, 1989) and drought (Ismail and

Noor, 1996a; Ismail et al., 1996; Salakpetch et al., 1990) decrease carambola growth and

yield. In South Florida, 86% of the total rainfall in 2001 and 79% of the total rainfall in

2002 occurred between May and October (Fig. 4-1). During the winter and drought

periods, irrigation may be needed to compensate for the lack of water from precipitation.

Therefore, irrigation based on quantitative information is important for commercial

carambola production in south Florida.

The soil of the Miami-Dade County, where carambola is cultivated, is composed

primarily of calcium carbonate (Degner et al., 1997) and classified as Krome very

gravelly loam. This is a very shallow, mineral soil with a pH of 7.4 - 8.4 (Noble et al.,

1996). This soil is low in organic matter and commercial farming largely depends on

fertilizer applications (Degner et al., 2002). The high demand for fertilizer coupled with

potentially excessive irrigation creates a potential for agrochemical leaching into the

groundwater (Muñoz-Carpena et al., 2002; Zekri et al., 1999). In addition to reducing

potential agrochemical leaching, scheduling irrigation to apply only the amount of water

68

69

required by the plant should increase grower economic returns by reducing fertilizer and

water inputs, and improving plant growth and yields.

A survey by Li et al. (2000) in 1998 showed that 73% of tropical fruit growers in

Miami-Dade County schedule irrigation based on the frequency and quantity of rain. The

percentage declined to 64.3% in 2002 according to a more recent survey by Muñoz-

Carpena et al. (2003). The number of growers monitoring of soil water content for

irrigation scheduling has increased to 48.8% of the respondents to the 2002 water-use

survey (Muñoz-Carpena et al., 2003) compared to only 15% in 1998 (Li et al., 2000).

According to the 1998 survey, methods of determining soil water included tensiometers,

capacitance probes, digging and squeezing soil, and the feel and appearance of the soil

(Li et al., 2000). Duration and frequency of irrigation were highly variable among

tropical fruit growers, which highlight the need for a better understanding of irrigation

requirements of these crops. With overhead, high-volume sprinklers, irrigation was

applied from one to three times per week for one to 12 h per application (Li et al., 2000).

For microsprinklers, the frequency of operation ranged from 0.5 to 7.5 h per application

and from one to seven per week. The amount of water applied per tree ranged from 110

to 2302 L with overhead sprinklers, 19 to 341 L per tree with microsprinklers, and from

7.6 to 45 L per tree for drip irrigation (Li et al., 2000). The variability in responses to

irrigation quantities and frequencies used by growers can be attributed to a lack of basic

quantitative information.

Tensiometers, neutron probes, and capacitance probes monitor soil water content

and are often used for irrigation scheduling. Tensiometers measure soil suction or matric

water potential rather than soil water content (Richards, 1942; Smajstrla and Harrison,

70

1998). A soil water retention curve must be established to determine the soil water

content that corresponds to the water matric potential in order to estimate soil water

content with a tensiometer.

The neutron probe is a portable compact unit that is easy to operate and volumetric

soil water content can be obtained instantly at different depths of the soil profile. Neutron

probes are considered more accurate than other soil water content monitoring devices for

irrigation scheduling (Evett and Steiner, 1995; Mostert and Hoffman, 1996). The main

component of a neutron probe is a fast neutron source encased in a protective shield and

an electronic counting scaler which are connected by an electric cable that is also used to

lower the probe into an access tube to determine soil water content throughout the soil

profile (Chanasyk and Naeth, 1996). Neutrons with a high energy are emitted by a

radioactive source, such as americium 241/beryllium, into the soil and are slowed down

by collisions with nuclei, primarily hydrogen atoms (Gardner and Kirkham, 1952). The

density of the resultant cloud of slow neutrons is a function of the soil water content

(Chanasyk and Naeth, 1996). The number of fast neutrons that are slowed is detected and

measured as a ‘count rate’ per unit time (Gardner et al., 1991). The count rate is

converted to volumetric water content using a calibration curve. Neutron probes measure

soil volumetric water content as the percentage of water per volume of soil.

Capacitance probes measure the soil water content based on the dielectric constant

of the soil (Paltineanu and Starr, 1997; Phene et al., 1990; Wu, 1998) a concept that was

first proposed for soil monitoring by time domain reflectometry (TDR) (Topp et al.,

1980). The dielectric constant of the soil is composed of the dielectric constants of water

(80.4), soil particles (3-7) and air (1) (Robinson and Dean, 1993; Paltineanu and Starr,

71

1997; Starr and Paltineanu, 1998a; Wu, 1998). The dielectric constant of the soil particles

and air are small and relatively constant compared to that of water. Thus changes in the

dielectric constant of the soil are a measure of the change in soil water content. The

volumetric water content can be expressed either as a percentage or a depth of water (mm

of water / 10 cm of soil) (Núñez-Elisea et al., 2001; Paltineanu and Starr, 1997; Starr and

Paltineanu, 1998a).

The objective of this study was to evaluate and compare tensiometers, multisensor

capacitance probes, and a neutron probe for accurately assessing soil water content in a

carambola orchard in Krome very gravelly loam soil.


The experiment was conducted in an orchard of 8-year-old ‘Arkin’ carambola trees

grafted onto open-pollinated Golden Star rootstock at the Tropical Research and

Education Center in Homestead, Fla. Trees were spaced at 4.5 m within rows and 6.1 m

between rows.

Low-tension tensiometers (0 to 40 kPa) (Model LT; Irrometer Co., Inc., Riverside,

Calif.) were calibrated prior to installation using a calibration vacuum chamber (Smajstrla

and Pitts, 1997) to ensure that the water column in the tensiometer was air free and that

there were no leaks. One tensiometer was installed 60 cm from the trunk of each of the

three replicate trees in each treatment. Tensiometers were installed at a depth of 10 cm.

Prior to installation, a hole was made in the soil, slightly larger in diameter than the

tensiometer. A slurry, prepared with sieved soil mixed with water, was applied to the

tensiometer hole to ensure that the ceramic cup of the tensiometer was in contact with the

soil (Núñez-Elisea et al., 2001). Tensiometers were maintained regularly in the field

using a hand-pump and water was refilled whenever it drained from the tensiometer tube.

72

For neutron probe measurements, one 85-cm long polyvinyl chloride (PVC) access

tube was installed at 60 cm from the trunk of each of three replicate trees in each of the

four treatments. The neutron probe (Model 503DR, Campbell Pacific Nuclear, Inc.

Martinez, Calif.) was placed in the access tubes and lowered to 10, 20, 30, and 50 cm

depths below the soil surface. Neutron probe counts were recorded during 16-s intervals

were obtained at each depth.

A multisensor capacitance probe system (EnviroSCAN, Sentek PTY Ltd., Kent

Town, Australia) was used to automatically and continuously measure soil water content

in each treatment. Prior to installation, dataloggers were configured in the laboratory

following the procedure described by Paltineanu and Starr (1997). Sensors were

normalized to air and water counts by placing the probe in a tube surrounded by water.

One probe was installed inside a 74 cm-long PVC access tube at 60 cm from the trunk of

each of the three replicate trees in each of the four treatments. Four sensors were placed

in each probe at 10, 20, 30, and 50 cm below the soil surface. The PVC tubes were

installed with a motorized drill and slurry was then poured into the hole. The slurry

consisted of 2:1:1 by volume of calcareous rock, cement, and water to prevent air gaps

from forming between the tubes and the surrounding soil (Núñez-Elisea et al., 2001).

The sensors in each probe were connected to a datalogger powered by a 12-volt battery

charged with a solar panel. Data were recorded every 30 min and downloaded from the

datalogger to a portable laptop computer and graphs of soil water depletion rates at each

soil depth and location were created with the EnviroScan software.

Trees were separated into four irrigation treatments based on the rate of soil water

depletion (SWD) between field capacity and the first visual sign of stress (i.e., leaf

73

yellowing and abscission) of carambola trees (as measured with multisensor capacitance

probes in a preliminary experiment). The SWD levels were: 0-8% SWD, 9-11% SWD,

12-14% SWD, or 15-17% SWD. Treatments were arranged in a completely randomized

design with three replications per treatment. When the soil water content reached the

treatment range, trees were irrigated with microsprinklers (at 89 L·h-1) to restore the soil

to field capacity.

Data from the three instruments were compared and analyzed by linear and

nonlinear regression and correlation tests.


A soil water retention curve was developed using van Genuchten’a (1980) model:

θ(Ψ)= θr + (θs – θr) [1+(α|Ψ|)n]-m

where Ψ is the matric potential (suction or water tension), θr is the residual water content,

and θs is the saturated water content, and α, n, and m are fitting parameters directly

dependent on the shape of the θ(Ψ) curve. Parameters (i.e. θr, θs, α, n, and m) for van

Genuchten’s model for the soil water retention curve of Krome soil were obtained from

the relationship between soil matric potential (Ψ) measured by tensiometers and

volumetric water content (θ) determined by neutron and capacitance probes (Table 4-1).

Soil matric potential measured with tensiometer fit van Genuchten’s model better when

volumetric soil water content was measured with a neutron probe (r2 = 0.42) than with

multisensor capacitance probes (r2 = 0.35) (Fig. 4-2 and 4-3, respectively). The fairly

weak relationship between soil water tension and volumetric soil water content in Krome

very gravelly loam soils can be attributed to the inaccuracy of tensiometer readings above

a suction of 20 kPa and heterogeneity of very gravelly Krome soils (Núñez-Elisea et al.,

2001). Tensiometers installed at 10 cm depth were not effective at a tension of above 20

74

kPa because air entered into the suction cup through the large pores in the gravel and

produced inaccurate measurements or water discharged completely from the tensiometer.

In a previous study, tensiometers installed at a 10-cm depth in Krome very gravelly loam

soil in south Florida were successfully used to schedule irrigation of tomato (Li et al.,

1998). In contrast, a previous study in tropical fruit orchards with the same soil showed

that tensiometers were not effective in accurately estimating soil water potential at a

depth of 30 cm (Núñez-Elisea et al., 2001). The difference between the usefulness of

tensiometers in the vegetable field and fruit orchard may have been a result of

significantly larger soil particles in the fruit orchards. In south Florida, vegetable fields

are rock-plowed and repeatedly disked to break up the top layer of the soil (Colburn and

Goldweber, 1961). Núñez-Elisea et al. (2001) reported that the lack of effectiveness of

tensiometers in tropical fruit orchards at a depth of 30 cm was attributed to the rockiness

of Krome soil at that depth where 71 to 73% of the soil was gravel compared to 26% to

38% at in top 10 cm. Similar relationship between tensiometer readings and capacitance

probes readings were also observed in the present study. Muñoz-Carpena et al. (2002)

measured gravimetric water content and soil suction in a laboratory experiment. They

reported that at soil suction above 10 kPa, soil water content in Krome very gravelly loam

soil is relatively insensitive to tension changes, thus large changes in soil water tension

reflect small changes in actual soil water content. In the present study, soil suction above

10 kPa resulted in variable tensiometer readings and discharge from the water column

above 20 kPa. Thus, tensiometers are not very useful for monitoring soil water content

for irrigation scheduling in carambola orchards in Krome very gravelly loam soil.

75

Volumetric soil water content determined with a neutron probe and capacitance

probes were positively correlated at all depths (Fig. 4-4). However, the correlation was

not very high, probably due to differences in the principles of operation of the devices,

variability in soil microclimate around the access tubes, and the larger sphere of influence

(volume of soil measured by the probe) measured by the neutron probe than each

capacitance sensor. Neutron probe measurements close to the soil surface are not reliable

(Chanasyk and Naeth, 1996; Cole, 1989; Gardner, et al., 1991;) due to escape of fast

neutrons to the atmosphere (Cuenca, 1989) as a result of the radius of the sphere of

influence of neutron probes (approximately 20 cm; Chanasyk and Naeth, 1988). This

explains the overestimation of soil water content observed in this study (Fig. 4-2)

compared to values estimated by Muñoz-Carpena et al. (2002) for the same soil using

pressure plate apparatus in the laboratory. Of the three soil water monitoring devices we

tested, variability among readings at different locations at the same depth were the lowest

with the neutron probe [coefficient of variation (CV) = 11%] compared to tensiometers

(CV = 17%) or capacitance probes (CV =33%). Although neutron probe readings were less

variable than those obtained from capacitance probes and tensiometers, its use is less

practical than that of the other two devices for irrigation scheduling due to the radioactive

source of fast neutrons which can cause health hazards if the neutron probe is not used

properly. Moreover, neutron probe operation and transportation require permits and

continuous monitoring for radiation leakage. Specific calibration is required for

accurately assessing actual volumetric soil water content. The use of a neutron probe to

monitor soil water content in Krome soil is labor intensive because it requires installation

of several access tubes using a motorized-drill and the manually obtaining of readings

76

over a large area at different depths in the soil profile. The initial cost of the instrument is

very high compared to that of tensiometers.

Studies have shown that volumetric soil water content determined with multisensor

capacitance probes was highly correlated with volumetric soil water content determined

gravimetrically (Paltineanu and Starr, 1997) in Mattapex silt loam fine soil, in silty clay

loam soils (Ould Mohamed et al., 1997), in fine sand soils of Florida (Morgan et al.,

1999), and in weathered heterogeneous soils (Wu, 1998). Multisensor capacitance probes

accurately measured soil water content for citrus orchards in Candler fine sand (Fares and

Alva, 2000). They also proved reliable for irrigation scheduling in avocado and ‘Tahiti’

lime orchards in south Florida (Núñez-Elisea et al., 2001; Núñez-Elisea et al., 2000;

Zekri et al., 1999). However, in the present study, there were often variable readings

among sensors at the same depth receiving the same irrigation treatment. Inaccurate

measurements of volumetric water content by capacitance probes were reported by

Hanson and Peters (2000) in silty loam and silty clay soils in California, and by Evett and

Steiner (1995) in Amarillo fine sandy loam soil. Tomer and Anderson (1995) concluded

that for coarse textured soils, changes in water content are difficult to detect using

capacitance probes. Although not proven, we speculate that increased capacitance probe

variability in Krome soil over time could be attributed to temperature variations from the

winter to the summer, soil wetness and drying cycles, long-term changes in soil chemical

and physical properties, and the initial soil disturbance caused by trenching. In addition,

air gaps and disturbance of soil around the probe may lead to a change in soil bulk

density that can produce measurement errors (Paltineanu and Starr, 1997). Throughout

this study, lightning had a devastating effect on electrical connections and capacitance

77

sensors even when the system was surge protected and a grounding rod was installed. At

times, capacitance sensors briefly produced erroneous readings, such as sudden decreases

in soil water readings, spikes, or discontinuous readings. However, the system generally

recovered from these temporary disturbances. Variability in capacitance probe readings

in this study can also be attributed to a low battery or a damaged solar panel and

condensation of water vapor in the probes. To reduce water vapor condensation in access

tubes, silica gel desiccants were frequently replaced but due to the large tube size, this

may not have been sufficient to maintain a dry climate inside each probe.

A major advantage of irrigation scheduling using capacitance probes is that

scheduling is based on the rate of soil water depletion rather than on the absolute water

content. Therefore, capacitance probe measurements reflect the change in soil water

content and evapotranspiration over time (Zekri et al., 1999). Using Sentek EnviroSCAN

software, the data are presented graphically and options are provided for the user to

divide the display into separate zones to manage soil water content. Multisensor

capacitance probe system can be a useful tool for irrigation scheduling if the cost of

operation and maintenance is financially justified.

In addition to tensiometers, neutron probes, and multisensor capacitance probes,

there are several other methods of monitoring soil water status that include dielectric

sensors, volumetric water content measurement devices, and water suction devices.

Several of these devices are currently being tested for use in Krome very gravelly loam

soils at the Tropical Research and Education Center in Homestead, Fla.

78

Table 4-1. Fitted parameters of van Genuchten’s model used to create a soil water retention curves for Krome soil. Soil water content was measured in a carambola orchard with a neutron probe and multisensor capacitance probes.

Parametersz Neutron probe Capacitance probes

θs θr α n m

0.4758 0.226 0.029 2.67 0.625

0.4758 0.229 0.0598 2.41 0.585

zθs is saturated water content, θr is residual water content, and α, n, and m, are model fitting parameters.

79

0

50

100

150

200

250

300

350

Jan-0

1

Feb-01

Mar-01

Apr-01

May-01

Jun-0

1Ju

l-01

Aug-01

Sep-01

Oct-01

Nov-01

Dec-01

Jan-0

2

Feb-02

Mar-02

Apr-02

May-02

Jun-0

2Ju

l-02

Aug-02

Sep-02

Oct-02

Nov-02

Dec-02

Amou

nt o

f rai

nfal

l and

ET

(mm

)RainET

Fig. 4-1. Total monthly rainfall and evapotranspiration (ET) during 2001 and 2002 in Homestead, Fla. Source: Florida Automated Weather Network, IFAS, Univ. of Fla., Gainesville.

80

0

5

10

15

20

25

30

35

40

45

50

0 10 20 30 40 5

Soil water tension (kPa)

Neu

tron

pro

be v

olum

etric

wat

er c

onte

nt

(%)

0

Fig. 4-2. Relationship between volumetric soil water content measured with a neutron probe and soil water tension measured with a tensiometer in a carambola orchard in Krome very gravelly loam soil. The response line was fitted using the van Genuchten equation (r2 = 0.42).

81

0

5

10

15

20

25

30

35

40

45

50

0 10 20 30 40 5

Soil water tension (kPa)

Cap

acita

nce

prob

e vo

lum

etric

wat

er

cont

ent (

%)

0

Fig. 4-3. Relationship between volumetric soil water content measured with multisensor capacitance probes and soil water tension measured with a tensiometer in a carambola orchard in Krome very gravelly loam soil. The response line was fitted using the van Genuchten equation (r2 = 0.35).

82

15

20

25

30

35

40

45

15 20 25 30 35 40

Capacitance probe soil water content (%)

Neu

tron

pro

be s

oil w

ater

con

tent

(%)

Fig. 4-4. Relationship between soil water content (%) as measured by multisensor capacitance probes and a neutron probe in Krome very gravelly loam soil (y = 0.67 x + 12.19, r2 = 0.35).

CHAPTER 5 PHYSIOLOGICAL RESPONSES OF CARAMBOLA TREES TO WATER

DEPLETION IN KROME VERY GRAVELLY LOAM SOIL

Physiological processes in fruit trees such as water potential and gas exchange are

sensitive to changes in soil water content. These physiological variables, growth, and

fruit production should be correlated with soil water content prior to determining the

appropriate amount of water to apply to an orchard. Continuous monitoring of soil water

content in tropical fruit orchards in Krome very gravelly loam soils has been done with

multisensor capacitance probes (Al-Yahyai et al., 2003; Núñez-Elisea et al., 2001;

Schaffer, 1998; Zekri et al., 1999). Proprietary software (EnviroSCAN 4.0, Sentek PTY

Ltd., Kent Town, Australia) is available for plotting soil water depletion (SWD) over

time and determining of the ‘full point’ (field capacity), the ‘refill point’ (irrigation

point), and the theoretical ‘onset of water stress’, the point at which the

evapotranspiration rate is significantly reduced (Schaffer, 1998; Zekri et al., 1999). Field

capacity (FC) has been determined with multisensor capacitance probes (Zekri and

Parsons, 1999; Fares and Alva, 2000) based on Veihmeyer and Hendrickson’s concept

which defines FC as the amount of water in the soil after excess water had drained and

the rate of downward movement has decreased (Hillel, 1998). Irrigation decisions using

SWD measurements are based on maintaining soil water content between the ‘full point’

and the ‘refill point’ before the occurrence of the ‘onset of water stress’. This approach to

irrigation may be more applicable to annual, herbaceous crops than to fruit trees. Trees

can tolerate longer periods of low soil water content than herbaceous plants through

83

84

specialized long-term and short-term physiological, phenological, anatomical, and

morphological responses (Ludlow, 1989), such as growth reduction, stomatal closure, and

osmotic adjustment (Hsiao, 1973).

Carambola (Averrhoa carambola L.), a tropical fruit tree native to S.E. Asia,

responds to water deficits by the ‘avoidance’ strategy (Neuhaus, 2003). This strategy

involves minimizing water loss by stomatal control, leaf movement, reducing leaf area,

and leaf abscission (Ludlow, 1989). The lethal leaf water potential (ΨL) level for

carambola is about -2.9 MPa and the lethal leaf relative water content (RWC) is about

56% (Ismail et al., 1994). Carambola responses to soil water deficit can be divided into

four phases (Neuhaus, 2003): (1) minimizing water loss by stomatal closure, initial

osmotic adjustment, and leaflet movement from a horizontal to vertical orientation; (2)

increasing root:shoot ratio via leaf and fruit abscission, and increasing proline levels to

increase leaf osmotic adjustment; (3) reaching a stage of irreversible damage caused by

chlorophyll degradation, root growth inhibition, and further leaf and fruit shedding; and

(4) tree death. Rewatering the trees during the second phase results in plant recovery

within a few days (Salakpetch et al., 1990).

Plant ΨL decreases when water loss via transpiration (E) is greater than water

uptake (Salisbury and Ross, 1985). Leaf water potential of field and container-grown

carambola trees declined with increasing water stress (Ismail and Noor, 1996a; Ismail et

al., 1994, 1996). Ismail et al. (1994) found a difference in ΨL of up to -2.20 MPa between

stressed and well-watered carambola trees. Salakpetch et al. (1990) found that ΨL of

water-stressed carambola trees (-2.0 MPa) returned to the previous level (-0.6 MPa) upon

rewatering. Net CO2 assimilation (A) of carambola decreased with increased water

85

stress. Net CO2 assimilation of 6-month-old, container-grown carambola trees was 3.13

µmol·m-2·s-1 for non-stressed trees compared to 0.24 to 1.83 µmol·m-2·s-1 for water

stressed trees (Ismail et al., 1994). Net CO2 assimilation was reduced when ΨL fell below

-0.80 MPa (Ismail et al., 1994; Razi et al., 1992). The decrease in A was due to reduced

stomatal conductance. Ismail et al. (1994) reported that the leaf photosynthetic rate was

reduced when stomatal resistance exceeded 3.5 s·cm-1.

Most of the studies of the responses of carambola tree to water stress in tropical

climates were conducted on young trees grown in containers in greenhouses. Little is

known about the response of mature and young carambola trees to soil water depletion

under field conditions especially in gravelly soils in subtropical climates such as that of

South Florida. The objective of this study was to test the effect of soil water depletion on

stem water potential (ΨS) and leaf gas exchange of carambola in a Krome very gravelly

loam soil in South Florida.


Experimental site and plant material. An orchard of 8-year-old ‘Arkin’

carambola trees grafted onto open pollinated Golden Star rootstock were used for this

study. The trees were planted with 4.5 m within and 6.0 m between rows with an artificial

(Polypropylene ribbon shade cloth) windbreak on the northern, eastern, and western

perimeters and a line of sapodilla [Manilkara zapota (L.) von Royen] trees as a

windbreak on the southern perimeter of the orchard. Adjacent to the orchard trees, 2-

year-old ‘Arkin’ carambola trees grafted onto Golden Star rootstock were planted in 95 L

containers. Six 2-year-old ‘Arkin’ carambola trees grafted on Golden Star rootstock were

also planted in 95 L containers and placed in a greenhouse. The containers were filled

with Krome very gravelly loam soil [loamy-skeletal, carbonatic, hyperthermic, Lithic

86

Udorthents (Noble et al., 1996)] that was excavated from a site next to the orchard.

Orchard trees were grown in the same soil. Fertilizer application, pest management, and

fertilization followed standard practice for South Florida’s commercial carambola

production (Crane, 1994).

Trees were irrigated with 89 L·h-1 microsprinklers with a 360o wetting area

(Maxijet, Dundee, Fla.). In the orchard and in field containers, trees were irrigated at

various times and rates to provide a range of SWD over the course of the experiments.

The ranges were determined from a preliminary study (Al-Yahyai, unpublished data) and

based on the SWD from soil field capacity (FC) to leaf yellowing, which was the first

visual sign of water stress. For orchard trees, irrigation was withheld toward the end of

the experiment for 121 d from 1 Aug. to 30 Nov. 2003 to obtain high SWD values. Trees

in containers in the greenhouse were divided into two treatments: (1) soil water

maintained at FC and (2) no irrigation water applied.

Soil water measurements. Soil water depletion was determined by continuously

measuring soil water content using multisensor capacitance probes (EnviroSCAN, Sentek

PTY Ltd., Kent Town, Australia). Capacitance probes were installed 60 cm north of tree

trunks in the orchard and 20 cm from tree trunks in the containers. The sensors recorded

soil moisture content every 30 min at depths of 10, 20, 30 and 50 cm. The data were

stored in a datalogger and later downloaded to a computer for analysis. Soil water content

measurements from sensors placed at the 4 depths were summed and plotted using the

EnviroSCAN proprietary system software. Installation of the system in Krome soils was

previously described by Al-Yahyai et al. (2003) and Núñez-Elisea et al. (2001), and

87

technical multisensor capacitance probe specifications were described by Paltineanu and

Starr (1997).

Leaf gas exchange and water potential measurements. Net CO2 assimilation

(A), stomatal conductance to water vapor (gs), transpiration (E), and stem water potential

(ΨS) were determined periodically for trees in the orchard and in containers in the field.

Net CO2 assimilation, gs, E, and ΨS for the container-grown trees in the greenhouse were

determined three times per week for 6 weeks from 18 Oct. to 2 Dec. 2002. Four mature

leaves per tree from the orchard and three mature leaves per tree from trees in containers

in the field were randomly selected for leaf gas exchange and water potential

measurements from 1100 to 1400 HRS. Net CO2 assimilation, E, and gs were determined

with a portable infrared gas analyzer (CIRAS II, PPSystems Ltd., Herts, UK).

Measurements were made on the terminal leaflet of fully matured, sunlit leaves at a

photosynthetic photon flux (PPF) of 1000 µmol·m-2·s-1 which is above the saturation

point for photosynthesis of carambola (Marler et al., 1994), and reference CO2 of 375

µmol·mol-1 at a flow rate of 200 ml·min-1. Leaf and stem water potential (ΨS) was

measured between 1100 and 1400 HRS using a pressure chamber (Soil Moisture

Equipment Corp., Santa Barbara, Calif.). Stem water potential was determined using a

fully matured leaf enclosed in a reflective plastic bag for 1 h to suppress E and allow ΨS

to equilibrate with leaf water potential (Begg and Turner, 1970; McCutchan and Shackel

1992; Shackel et al., 1997).

Statistical analysis. Linear regression analysis was performed with SigmaPlot

software (SYSTAT Software Inc., Richmond, Cal.) and nonlinear regression analysis was

performed with TableCurve 2D software (SYSTAT Software Inc., Richmond, Cal.).

88


Stem water potential of trees in containers in the greenhouse was above -1.0 MPa

when SWD increased from 0% to 60% above which there was a sharp decrease in ΨS

(Fig. 5-1). A similar trend was observed in field-container trees with a sharp decrease in

ΨS when SWD reached 50% (Fig. 5-2). For the orchard trees, SWD never reached above

30%. Within the range of SWD from 0% to 30%, ΨS remained above -1.0 MPa and did

not significantly correlate with SWD (data not shown). Therefore, within a 0 to 30%

SWD range, there was no significant effect of ΨS on leaf gas exchange, presumably due

to sufficient soil water content in the orchard from frequent precipitation and capillary

rise from the water table which was 1-2 m below the soil surface. Similarly, in a study of

mango trees in Krome soil, predawn water potential ranged from -0.3 to -0.5 MPa and

there were no significant differences in A or E were observed between irrigated and non-

irrigated trees (Larson et al., 1989). For 6-month-old carambola trees in containers, Ismail

et al. (1994, 1996) observed that withholding water for 1, 2, or 4 weeks resulted in

midday ΨL of -1.3 to -1.5 MPa, -2.0 to -2.5 MPa, or -2.9 MPa, respectively, in

comparison to trees irrigated at field capacity that had a ΨL of -0.9 MPa. Similarly,

Salakpetch et al. (1990) found that leaves of 2-year-old carambola trees in containers

wilted when midday ΨL reached -2.0 MPa, 10 d after withholding water. Leaf water

potential returned to its original level (-0.6 MPa) 2-3 d after rewatering (Salakpetch et al.,

1990). The ability of carambola trees to regulate their water potential was attributed to

osmotic adjustment through increased levels of proline (Ismail et al., 1994) and a

reduction in gs (Ismail et al., 1994, 1996; Razi et al. 1992).

Leaf gas exchange of container-grown carambola trees in the field and in a

greenhouse was more closely correlated with ΨS than with SWD. For example, gs

89

relationship with SWD coefficient of determination (r2) was 0.46 (Fig. 5-3) compared to

r2 = 0.54 (Fig. 5-4) with ΨS. This is in agreement with studies that found ΨS as a useful

indicator of water stress and thus can be used for irrigation management (Naor, 2000;

Shackel et al., 1997; Stern et al., 1998).

Stem water potential effects on stomatal conductance. A reduction in ΨS below

-1.0 MPa resulted in reduced gs for trees in containers in the greenhouse (Fig. 5-4) and in

the field (Fig. 5-5). Similarly, Razi et al. (1992) reported a decline in gs in carambola as a

consequent of a decrease in ΨL. Differences between greenhouse and field trees in the

response of physiological variables to SWD were presumably due to greater climatic

variability in the field than in the greenhouse. For example, vapor pressure was more

variable in the field [coefficient of variation (cv) = 29%] than in the greenhouse

[coefficient of variation (cv) = 17%].

Stem water potential effects on transpiration. Reductions in E were related to

reductions in ΨS in carambola trees in containers in the greenhouse (Fig. 5-6) and in the

field (Fig. 5-7). As SWD increased, transpiration of trees in containers in the greenhouse

(Fig. 5-8) and the orchard (Fig. 5-9) decreased linearly as gs decreased. This response of

carambola trees is in agreement with results from Razi et al. (1992) who found a linear

relationship between ΨL and E in carambola trees.

Stem water potential effect on net CO2 assimilation. Net CO2 assimilation

decreased as ΨS declined below about -0.5 MPa in trees grown in containers in the

greenhouse (Fig. 5-10) and in the field (Fig. 5-11). Ismail et al. (1994) found that mean A

of young carambola trees in containers was 3.13 µmol·m-2·s-1 for well-watered trees

compared to 0.24 to 1.83 µmol·m-2·s-1 for water-stressed trees. They concluded that a ΨL

90

lower than -0.80 MPa significantly reduced A. Similarly, Razi et al. (1992) observed that

A of carambola trees was reduced when water potential fell below -0.85 MPa. These

results suggest that A of carambola trees can be reduced when ΨS fall below -0.5 MPa.

The reduction in A with decreasing ΨS was presumably due to a decrease in gs for

trees in containers in the greenhouse (Fig. 5-12) and in the field (Fig. 5-13). Ismail et al.

(1994) reported that increased stomatal resistance in young carambola trees limited CO2

uptake during water stress. Reduction in A of carambola trees in containers in the

greenhouse and the field in Krome soils occurred when gs was below 50 mmol·m-2·s-1

(Figs. 5-12 and 5-13).

Stomatal conductance effects on transpiration and net CO2 assimilation. The

effect of gs on E (Fig. 5-8 and 5-9) and A (Fig. 5-12 and 5-13) suggested that partial

stomatal closure was essential to conserve water by reducing E. This partial stomatal

closure resulted in a gradual decline in A as soil water initially depleted. However, as soil

water depletion increased and water stress progressed, a further decline in gs (below 50

mmol·m-2·s-1) resulted in sharp decline in A. This perhaps was an avoidance mechanism

by carambola trees to conserve water by reducing E, while maintaining A during periods

of occasional or short-term water stress. Thus a SWD lower than 50% in Krome soils did

not affect tree gas exchange. In the orchard, the SWD never reached greater than 30%

(data not shown). Thus, the range of SWD in the orchard also remained below the point

at which leaf gas exchange was decreased.

Stem water potential relationship to leaf water potential. It has been suggested

that ΨL is a less reliable measure when making irrigation scheduling decisions for fruit

trees than ΨS due to its inherent variability under field conditions (Garnier and Berger,

91

1985; McCutchan and Shackel, 1992; Naor, 2000). Enclosing leaves in a reflective, dark

plastic bag for 1-2 h prior to water potential measurements will reduce E until an

equilibrium is achieved between the ΨL and the ΨS (Begg and Turner, 1970), thus

providing a more precise method of determining tree water status (Garnier and Berger,

1985; McCutchan and Shackel, 1992; Naor, 2000; Shackel et al., 1997). In this study, ΨL

and ΨS from the same branch were measured within 2 min of each other to determine

their relationship. There was a strong relationship (r2 = 0.97) between ΨL and ΨS in field

and greenhouse container trees (Fig. 5-14). For orchard trees (Fig. 5-15), ΨL and ΨS were

also highly correlated (r2 = 0.85) despite the narrow water potential range of -0.3 to -1.1

MPa compared to that of container trees. Jones (1997) suggested that either ΨS or ΨL may

be used to assess the effect of water stress on a particular tree response that is assumed

being affected by tree water status. Our results with carambola trees appeared to confirm

this.

Conclusions

An increase in SWD resulted in reduction in ΨS, A, E, and gs of carambola trees

grown in containers in Krome very gravelly loam soil. In a carambola orchard, SWD

never reached a point where there was a decrease in ΨS, A, E, or gs. Correlations between

physiological variables were higher for container-grown trees in a greenhouse than in the

field, which may have been due to greater fluctuations in vapor pressure in the field. In

the orchard, SWD did reach greater than 30% even if trees were not irrigated for up to

121 d. This may have been due to frequent precipitation and capillary rise to the root

zone from the water table, located at 1 to 2 m below the soil surface, was sufficient to

maintain adequate soil water even after several weeks of no irrigation.

92

For trees in containers in Krome very gravelly loam soil, SWD levels above 50%

caused a reduction in ΨS that subsequently reduced gs. This correlated with a linear

reduction in E and a sharp decline in A when gs fell below 50 mmol·m-2·s-1. Leaf gas

exchange was better correlated with ΨS than with SWD level. Therefore, ΨS may be

better indicator than SWD as determined from measurements of soil water content for

irrigation scheduling of carambola trees in Krome soils.

93

SWD (%)

0 20 40 60 80

ψ

S (

MPa

)

100

-4

-3

-2

-1

0

Fig. 5-1. Soil water depletion (SWD) and stem water potential (Ψs) of carambola trees in containers in a greenhouse (y = -0.44 - (3.59e-06) x3, r2 = 0.82).

SWD (%)

0 20 40 60 80 1

ψs

(MPa

)

00

-4

-3

-2

-1

0

Fig. 5-2. Soil water depletion (SWD) and stem water potential (Ψs) of carambola trees in containers in the field (y = -0.48 - (2.79e-06) x3, r2 = 0.56).

94

SWD (%)

0 20 40 60 80 100

g s (m

mol

m-2

s-1

)

0

20

40

60

80

100

120

140

160

180

200

Fig. 5-3. Soil water depletion (SWD) and stomatal conductance to water vapor (gs) of carambola trees in containers in a greenhouse (y = 95.46 – 0.0002 x3, r2 = 0.46).

95

Ψs (MPa)

-4 -3 -2 -1 0

gs

(mm

ol m

-2 s

-1)

0

20

40

60

80

100

120

140

160

180

200

Fig. 5-4. Stem water potential (Ψs) and stomatal conductance to water vapor (gs) of carambola trees in containers in a greenhouse (y = 11.88 + 132.94 ex , r2 = 0.54).

ψs (MPa)

-4 -3 -2 -1 0

gs (m

mol m

-2 s

-1)

0

20

40

60

80

100

120

140

160

180

200

220

Fig. 5-5. Stem water potential (Ψs) and stomatal conductance to water vapor (gs) of carambola trees in containers in the field (y = 44.76 – 41.57 / x, r2 = 0.32).

96

ψs (MPa)

-4 -3 -2 -1 0

E (

mm

ol m

-2 s

-1)

0

1

2

3

4

Fig. 5-6. Stem water potential (Ψs) and transpiration (E) of carambola trees in containers in a greenhouse (y = 0.30 + 2.94 ex, r2 = 0.55).

ψs (MPa)

-4 -3 -2 -1 0

E (

mm

ol m

-2 s

-1)

0

1

2

3

4

Fig. 5-7. Stem water potential (Ψs) and transpiration (E) of carambola trees in containers in the field (y = 1.03 + 2.11 ex, r2 = 0.28).

97

gs (mmol m-2 s-1)

0 20 40 60 80 100 120 140 160 180 200

E (

mm

ol m

- 2 s

-1)

0

1

2

3

4

Fig. 5-8. Stomatal conductance to water vapor (gs) and transpiration (E) of carambola trees in containers in a greenhouse (y = 0.33 + 0.02 x, r2 = 0.76).

gs (mmol m-2 s-1)

0 20 40 60 80 100 120 140 160

E (

mm

ol m

- 2 s

-1)

0

1

2

3

4

Fig. 5-9. Stomatal conductance to water vapor (gs) and transpiration (E) of carambola trees in containers in the field (y = 0.34 + 0.02 x, r2 = 0.75).

98

ψs (MPa)

-4 -3 -2 -1 0

A (η m

ol m

- 2 s

-1)

-2

0

2

4

6

8

10

Fig. 5-10. Stem water potential (Ψs) and net CO2 assimilation (A) of carambola trees in containers in a greenhouse (y = -0.09 + 9.76 ex, r2 = 0.70).

ψs (MPa)

-4 -3 -2 -1 0

A (η m

ol m

- 2 s

-1)

-2

0

2

4

6

8

10

12

Fig. 5-11. Stem water potential (Ψs) and net CO2 assimilation (A) of carambola trees in containers in the field (y = 1.19 - 2.95/x, r2 = 0.31).

99

gs (mmol m-2

s-1

)

0 20 40 60 80 100 120 140 160 180 200

A (η m

ol m

- 2 s

-1)

-2

0

2

4

6

8

10

Fig. 5-12. Stomatal conductance (gs) and net CO2 assimilation (A) of carambola trees in containers in a greenhouse (y = -2.46 + 0.39 (lnx)2, r2 = 0.75).

gs (mmol m-2

s-1

)

0 20 40 60 80 100 120 140 160

A (η m

ol m

- 2 s

-1)

-2

0

2

4

6

8

10

Fig. 5-13. Stomatal conductance (gs) and net CO2 assimilation (A) of carambola trees in containers in the field (y = -2.24 + 0.40 (lnx)2, r2 = 0.45).

100

ψL (MPa)

-4 -3 -2 -1 0

ψS (

MPa

)

-4

-3

-2

-1

0

Field-container trees

Greenhouse-container trees

Fig. 5-14. Leaf (ΨL) and stem water potentials (ΨS) of carambola trees in containers in the field (y = 0.10 + 1.00x, r2 = 0.97) and in a greenhouse (y = 0.33 + 0.91x, r2= 0.97).

ψS (MPa)

-0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3

ψL

(MPa

)

-1.2

-1.0

-0.8

-0.6

-0.4

Fig. 5-15. Leaf (ΨL) and stem water potentials (ΨS) of carambola trees in an orchard in the field (y = 0.73 + 1.38x, r2 = 0.85).

CHAPTER 6 GROWTH, YIELD, AND FRUIT QUALITY OF CARAMBOLA AS AFFECTED BY

SOIL WATER DEPLETION

Carambola (Averrhoa carambola L.) is an evergreen, tropical fruit tree native to

S.E. Asia where it is primarily cultivated. The tree is also cultivated in regions of

Australia, India, Israel, the Caribbean Islands, and the United States (mainly Hawaii and

Florida) (Crane, 1994; Mauro Pace, 2004, personal communication). A sweet type,

‘Arkin’, is the leading commercial carambola cultivar in Florida (Knight and Crane,

2002). In South Florida, carambola is grown in Krome very gravelly loam soil that has a

high pH (7.4-8.4), low fertility, and low soil water holding capacity (Noble et al., 1996).

Due to the hardness of the soil and to facilitate root growth, carambola trees are planted

in perpendicular intersections of trenches that are approximately 50 to 75 cm wide and 50

cm deep (Colburn and Goldweber, 1961). In southern Florida, carambola growth ceases

during winter and the tree remains quiescent until it refoliates in mid-March (Núñez-

Elisea and Crane, 1998). Low temperatures, dry winds, and high light intensity (above

1000 µmol·m-2·s-1) cause carambola leaves to become chlorotic and trees partially

defoliate from February to March (Núñez-Elisea and Crane, 1998; George et al., 2002a,

2002b).

Fulfilling tree water requirements is important for economically profitable fruit

production of carambola. Insufficient (Ismail and Noor, 1996a; Salakpetch et al., 1990) or

excessive (Ismail and Noor, 1996b; Joyner and Schaffer, 1989) soil moisture can

decrease tree growth and production. Increased water stress reduced vegetative growth

101

102

(Ismail et al., 1994, 1996) and flowering (Salakpetch et al., 1990) of carambola. Even

when tree growth was not affected by drought, yield was significantly reduced in rainfed

trees compared to irrigated trees in Malaysia, mainly due to a decrease in fruit size of

non-irrigated trees (Bookeri, 1996).

Little is known about water requirements and irrigation scheduling practices for

carambola trees. In Australia, 30-75 mm of water per week was recommended for mature

carambola trees grown in the Northern Territory (Lim, 1996) and northern Queensland

(Galán Saúco et al., 1993). In southern Florida, Crane (1994) recommended 33 mm per

ha twice a week during dry periods throughout the year. However, water application rates

and frequencies that meet tree water requirements for optimum growth and yield have not

been established. Understanding tree water requirements is important for proper irrigation

management of carambola trees in subtropical climates such as that of South Florida. The

objective of this study was to determine the effects of four levels of soil water depletion

(SWD) on growth, yield and fruit quality of mature orchard-grown and young container-

grown carambola trees in a Krome very gravelly loam soil.


Plant material and soil properties. The study was conducted in an orchard with

8-year-old ‘Arkin’ carambola trees grafted on open-pollinated Golden Star rootstock in

Krome very gravelly loam soil in Homestead, Fla. at 25.5 oN Lat. and 80.5 oW Long. The

trees were spaced at 4.5 m within rows and 6 m between rows and surrounded by an

artificial (Polypropylene ribbon shade cloth) windbreak on the northern, eastern, and

western perimeters and sapodilla [Manilkara zapota (L.) von Royen] trees as windbreaks

on the southern perimeter of the orchard. Adjacent to the orchard, 2-year-old ‘Arkin’

carambola trees grafted on Golden Star rootstock were planted in 95 L plastic containers.

103

The containers were filled with Krome very gravelly loam soil which was excavated from

a site next to the orchard. Fertilizer applications and pest management practices for the

orchard and trees in containers followed recommendations for South Florida’s

commercial carambola production (Crane, 1994).

Soil water depletion treatments. Soil water depletion was determined by

monitoring soil water content with multisensor capacitance probes (EnviroSCAN, Sentek

PTY Ltd., Kent Town, Australia). Capacitance probes were installed 60 cm north of the

trunk in the orchard and 20 cm from the trunk in containers of 3 trees per treatment. Each

probe in the orchard consisted of four capacitance sensors located at soil depths of 10, 20,

30, and 50 cm. Probe sensors in containers were placed at of 10-, 20-, and 30-cm depths

below the soil surface. The sensors recorded soil water content every 30 min and the data

were stored in a datalogger and later downloaded to a portable computer for analysis.

Description of the installation of the capacitance probe system in Krome soils was

previously described by Al-Yahyai et al. (2003) and Núñez-Elisea et al. (2001), and

technical specifications of the multisensor capacitance sensors were described by

Paltineanu and Starr (1997). Data from the 10-, 20- and 30-cm depths sensors were

summed and plotted using EnviroSCAN software (EnviroSCAN 4.0, Sentek PTY Ltd.,

Kent Town, Australia). Continuous measurements of soil water content using capacitance

probes allowed for the determination of Field capacity (FC) (Hillel, 1998; Zekri and

Parsons, 1999; Fares and Alva, 2000). Field capacity values were selected based on

Veihmeyer and Hendrickson’s concept that define FC as the amount of water in the soil

after excess water had drained and the rate of downward movement had decreased (Hillel

1998). After the treatment SWD levels were reached, irrigation was applied for each

104

treatment to restore soil moisture to FC in the orchard and to container capacity in

containers using microsprinklers with a 360o wetting pattern (Maxijet, Dundee, Fla.,

USA) at 89 L·h-1.

Soil water depletion levels were predetermined in the orchard and in containers

based on a preliminary study (Al-Yahyai, unpublished data) whereby irrigation water was

withheld and soil water content was monitored via capacitance probes until visual water

stress symptoms (leaf yellowing and abscission) appeared on the trees. Irrigation of the

orchard trees was initiated when SWD reached one of the following four levels (where

0% SWD = FC): 0-8% SWD, 9-11% SWD, 12-14% SWD, or 15-17% SWD. Irrigation of

container trees was initiated when SWD reached one of the following four levels: 0-21%

SWD, 22-31% SWD, 32-50% SWD, and 51-60% SWD. Containers were placed on metal

bases above the ground. Soil water depleted faster in containers than in the orchard

because lateral and capillary water movement into the container soils were prevented.

Growth and yield measurements. Shoot growth was measured on four randomly

selected, actively growing branches per tree in the orchard and three shoots per tree in the

container study. The shoots were tagged at 10 cm below the shoot apex and measured

using a metric ruler. Trunk caliper of trees in containers was measured from 27 July 2002

until 18 Dec. 2003 at 10 cm above the graft union. Fruit from carambola trees have an

irregular shape and are composed of five to seven longitudinal ridges, thus fruit length

was used to determine differences in fruit growth among treatments. Fruit length of four

randomly selected fruit in the field and three fruit in containers were measured with a

caliper. At the end of the container experiment, trees were removed from the containers

105

and the fresh and dry weights of the leaves, stems and roots of each tree were recorded to

determine tree biomass.

In the orchard, fruit from six trees per treatment were harvested during August

(summer harvest) and December (winter harvest) of 2002 and 2003. The harvest dates

coincided with commercial peak harvesting months for carambola trees in South Florida.

Fruit from six trees per treatment were harvested on Dec. 2002 and Dec. 2003. Fruit were

sorted based on the stage of maturity and color from 120 ohue (green) to 80 ohue (yellow)

to immature (115 ohue), mature (108 ohue), and ripe (tree-ripened; 87 ohue) as determined

with a colorimeter (Minolta Chroma Meter, Minolta, Inc., Ramsey, N.J.). Fruit were then

counted and weighed to determine the total yield. Total soluble solids (TSS), fresh and

dry weights of 20 randomly selected mature (108 ohue) and 20 ripe (87 ohue) fruit per tree

were determined from six trees per treatment.

Experimental design and statistical analysis. Treatments in the orchard and in

containers were arranged in a completely randomized design. In the orchard, each

treatment consisted of three rows with five trees per row. Two trees were randomly

selected from each row for growth and yield measurements. In containers, treatments

were replicated three times with two trees per replication. Data were analyzed by

repeated measures analysis (SAS V.8.2, SAS Institute, Cary, NC) and treatment means

were compared with Duncan’s multiple range test.


Shoot length. Shoot elongation from March to September in 2002 (Fig. 6-1A) and

from April to December in 2003 (Fig. 6-1B) did not significantly differ among SWD

treatments for carambola trees in the orchard. In orchard trees, this lack of shoot

elongation in response to treatments could be attributed to sufficient soil water content

106

throughout the experiment. Frequent precipitation (1,355 mm in 2002 and 1,606 mm in

2003) and capillary rise from the water table located 1-2 m below soil surface may have

provided an adequate water supply. In Malaysia, sufficient rainfall also resulted in no

significant difference in shoot growth rate between irrigated and rain-fed field-grown

carambola trees (Bookeri, 1996). Carambola trees can tolerate periods of low soil water

content once they are established (Galán Saúco et al., 1993). Therefore, irrigation of

carambola trees in an orchard at 17% SWD in Krome soils of South Florida did not

reduce shoot length.

Trunk diameter. Trunk diameter of 2-year-old carambola trees in containers in the

0-21% SWD treatment was higher than that of trees in the other treatments from July

2002 until Dec. 2003 (Fig. 6-2). Trees in the 0-21% SWD treatment had significantly

greater trunk diameter than trees in the 32-50% SWD treatment throughout the

experiment. Trunk diameter of trees in the 0-21% SWD treatment was significantly

greater than that of trees in the 51-60% SWD treatments from September to December

2003. Trunk diameter of trees in the 22-31% SWD treatment was not significantly

different from that of trees in the 32-50% SWD and 51-60% SWD treatments throughout

the experiment.

Trunk diameter of 2-year-old container-grown trees decreased when SWD was

greater than 21%. Similarly, in a previous study, a smaller stem diameter was observed

for water-stressed carambola trees in comparison to well-watered trees in containers

(Ismail et al., 1994, 1996). Trunk diameter of container-grown carambola trees was

smaller in treatments where SWD was allowed to drop below 32% compared to SWD of

0-21%, which indicated that whole-tree growth was reduced. Therefore, reduction in

107

trunk diameter was a better indicator of tree response to SWD than individual

measurements of shoot length. Soil water depletion in the orchard, however, may not be

high enough to cause a significant reduction in tree trunk diameter.

Tree dry and fresh weight. There were no significant differences among

treatments in leaf and root fresh and dry weights (Table 6-1) for container-grown trees at

the end of the experiment. Trees in the 0-21% SWD treatment had significantly greater

stem fresh weight than trees in the 22-31% SWD and 51-60% SWD treatments. Stem dry

weight of trees in the 0-21% SWD treatment was also significantly greater than that of

trees in the 51-60% SWD treatment. Total tree dry weight of trees irrigated at 0-21%

SWD was significantly greater than for trees irrigated at the 51-60% SWD (Table 6-1).

Stem and total dry weight reduction at high SWD levels are in agreement with studies

from tropical regions such as of Malaysia in which leaf, stem, and root dry weights of

carambola were reduced with increased water stress (Ismail et al., 1994,1996; Ismail and

Noor, 1996a; Razi et al., 1992).

There were no significant effect of SWD in the root:shoot ratio (on a dry weight

basis) of trees in containers among SWD treatments (Table 6-1), due to the lack of SWD

treatment effects on leaf and root dry weight. In contrast, Razi et al. (1992) found that

root:shoot ratio of carambola was significantly higher in trees irrigated at 95% SWD in

comparison to those irrigated at FC or 50% SWD. Ismail et al. (1996), however, reported

a higher root:shoot ratio for carambola trees subjected to moderate water stress compared

to well-watered trees or trees subjected to severe water stress. Severe water stress in the

latter study may have decreased root growth due to reduced stomatal conductance and

photosynthesis per unit leaf area (Hsiao, 1993; Chartzoulakis et al., 1993). An increased

108

root:shoot ratio in water-stressed trees is an adaptive mechanism to water stress resulting

from greater allocation of dry matter to the roots than the shoots (Begg and Turner, 1976;

Ismail et al., 1994; 1996; Razi et al., 1992). Young carambola trees grown in Krome soils

outdoors in containers may not have reached water stress levels that could have resulted

in a significant increase in the root: shoot ratio even in trees irrigated at 51-60% SWD.

This was presumably due to the frequent precipitation during months when trees were

actively growing (March to November).

Fruit length and yield. Fruit length of carambola trees in the orchard did not differ

significantly among SWD treatments (Fig. 6-3). Adequate soil water content during fruit

growth periods in 2003 may have contributed to the lack of significant differences in fruit

growth. Irrigation of carambola trees in Krome soil at 17% SWD did not reduce fruit

length.

There were no significant statistical interactions between SWD treatment and year

or SWD treatment and harvest date for orchard or container-grown trees with respect to

fruit number or weight. Number and weight of mature (108 ohue) and ripe (tree-ripened;

87 ohue) fruit from harvest 1 (Aug. 2002), harvest 2 (Dec. 2002), harvest 3 (Aug. 2003),

and harvest 4 (Dec. 2003) of orchard trees were not significantly different among SWD

treatments (Table 6-2). The total fruit number and weight were significantly greater in

2002 than in 2003. In addition, fruit number and weight from orchard trees in December

were significantly greater than those harvested in August (data not shown). However, no

significant differences in total fruit weight or number among treatments were observed in

container-grown carambola trees.

109

In a study of carambola trees in Malaysia, Bookeri (1996) found that irrigation

treatments did not affect fruit number but irrigated trees had fruit that were 17% larger

than non-irrigated trees. In the present study, soil water content did not reach sufficiently

low levels to cause water stress in orchard-grown trees and resulted in no significant

differences in fruit length, number or weight among treatments. In 2-year-old container-

grown carambola trees, the lack of yield differences among irrigation treatments may

have been due to variability in fruit production among individual trees. Within the same

treatment, several carambola trees in containers did not set fruit, while others produced

36 fruit per tree.

Fruit fresh and dry weight of orchard-grown trees. There was a significant

effect of year on fruit fresh and dry weights. Fruit from 2003 had higher fresh and dry

weights than those from 2002. Fresh weight of mature fruit (108 ohue) of trees in the

orchard was significantly lower in the 12-14% SWD treatment than that in 9-11% SWD

treatment for fruit harvested in Aug. 2002 and significantly lower than 0-8% SWD

treatment for fruit harvested in Dec. 2002 (Table 6-3). In Dec. 2003, fruit from trees in

the 15-17% SWD had significantly greater fruit fresh weight than all other treatments,

whereas fruit from trees in the 9-11% SWD had the lowest fresh weight.

Dry weight of mature fruit from the Aug. 2002 harvest was significantly greater in

the 9-11% and 15-17% SWD treatments than in the 0-8% and 12-14% SWD treatments

(Table 6-4). In Dec. 2002, fruit of trees irrigated at 0-8% and 9-11% SWD had

significantly greater fruit dry weight than those of trees in the 12-14% and 15-17% SWD

treatments. In Aug. 2003, trees in 12-14% SWD had greater fruit dry weight than trees in

110

all other treatments. In Dec. 2003, fruit fresh weight was highest in the 0-8% SWD

treatment and lowest in 12-14% SWD treatment.

Fresh weight of ripe (tree-ripened, 87 ohue) fruit from the Aug. 2002 harvest was

significantly greater in the 0-8% SWD treatment than in the 9-11% and 12-14% SWD

treatments (Table 6-5). There were no significant differences in fruit fresh weight among

treatments in Dec. 2002. In Aug. 2003, fresh weight of fruit from of trees in the 0-8% and

9-11% SWD treatments was significantly greater than that of the 12-14% SWD

treatment. In Dec. 2003, fresh weight of ripe fruit from the 15-17% SWD treatment was

significantly greater than that of trees in all other treatments, whereas, fruit from trees in

the 9-11% SWD treatment had the lowest fruit fresh weight.

In Aug. 2002, dry weight of ripe fruit from trees in the 15-17% SWD was

significantly greater than that from all other treatments (Table 6-6). In Dec. 2002, fruit of

trees in 15-17% SWD treatment had a lower fruit dry weight than fruit of trees in the 0-

8% and 12-14% SWD treatments. In Aug. 2003, dry weight of ripe fruit was greatest in

the 12-14% SWD treatment. Fruit dry weight in the 0-8% SWD was significantly greater

than that of all other treatments. Fruit from 9-11% SWD and 12-14% SWD had the

lowest dry weight in Dec. 2003. There was no effect of SWD on fruit fresh or dry

weights of carambola trees in the orchard.

Fruit fresh and dry weight of container-grown trees. Mature fruit had greater

fresh weight than those in 2002 than in 2003. However, no significant statistical

interaction between SWD treatment and year was observed for dry weight of ripe and

mature fruit and for fresh and dry weights of ripe fruit. In Dec. 2002, mature fruit of trees

in the 0-21% SWD treatment had a significantly greater fresh weight than those of trees

111

in all other treatments (Table 6-7). In Dec. 2003, fresh weight of mature fruit of trees in

the 32-50% SWD treatment was significantly greater than that of trees in the 0-21% and

51-60% SWD treatments. Mature fruit dry weight was not significantly different among

irrigation treatments in Dec. 2002. In Dec. 2002, dry weight of fruit from trees in the 22-

31% and 32-50% SWD treatments was significantly greater than that from trees in the 0-

21% SWD treatment.

Ripe fruit from container-grown trees in the 0-21% and 22-31% SWD treatments

had greater fruit fresh weight than those in the 32-50% and 51-60% SWD treatments in

Dec. 2002 (Table 6-8). In Dec. 2003, fresh weight of ripe fruit in the 22-31% and 32-50%

SWD treatments was significantly greater than that in the 0-21% and 51-60% SWD

treatments. In Dec. 2002, dry weight of ripe fruit was greater in the 32-50% and 51-60%

SWD treatments than in the 0-21% and 22-31% SWD treatments (Table 6-8). In Dec.

2003, fruit dry weight was significantly greater in the 32-50% SWD than all other

treatments.

Total soluble solid (TSS) content. Total soluble solid (TSS) content of ripe (tree-

ripened, 87 ohue) fruit from orchard trees varied among irrigation treatments from one

harvest to another and significant interactions were found between treatments and year

and treatments and harvest date for TSS. Total soluble solids of fruit harvested in 2003

were higher than those harvested in 2002 and TSS of fruit harvested in December was

higher than in fruit harvested in August (data not shown). There was no clear trend of the

effect of SWD treatments on fruit TSS. For example, fruit harvested in Dec. 2002 had the

highest TSS in the 0-8% SWD treatment and lowest in the 15-17% SWD treatment,

whereas, fruit harvested in Dec. 2003 had the highest TSS at the 15-17% SWD and

112

lowest TSS at 0-8% SWD (Table 6-9). Total soluble solid content in ripe fruit of trees in

containers was not significantly different among treatments in Dec. 2002 (Table 6-10).

However, fruit from Dec. 2003 had a significantly higher TSS in the 51-60% SWD

treatment than in the 0-21% and 22-31% SWD treatments. Total soluble solid content

was higher than the reported average TSS of 6.8-7.2% for carambola in southern Florida

(Campbell and Koch, 1989; Crane et al., 1998). Increased TSS in fruit was perhaps due to

reduced irrigation water applied in this study. Irrigation at 32-50% SWD and 51-60%

SWD in containers resulted in an increase in TSS in Dec. 2003. Irrigation at high levels

of SWD may be beneficial for increasing fruit TSS in containers. However, due to

variability of TSS within harvest dates, no conclusion can be made on the effect of SWD

treatments on ripe fruit TSS in orchard-grown carambola trees in this study.

Conclusions

Shoot growth of carambola trees grown in an orchard and in containers was not

affected by SWD. In the orchard, sufficient soil water content from precipitation and the

shallow water table possibly resulted in sufficient soil water content to obtain adequate

vegetative growth and yields. For container-grown trees where lateral water movement

and capillary rise were prevented, trunk diameter decreased in trees in the highest SWD

treatment. Soil water depletion of 51-60% reduced total dry weight of trees in containers.

The root:shoot ratio was also not significantly different among SWD treatments because

there were no significant differences in leaf and root dry weights among SWD treatments.

Reduction in dry weight of young carambola trees did not occur until soil water content

decreased to 50% of FC. Fruit length, number and weight did not differ significantly

among SWD treatments in orchard-grown trees. Total soluble solid content did not

follow a specific trend in response to irrigation at various levels of SWD. Differences

113

among SWD treatments in mature and ripe fruit fresh and dry fruit weight were highly

variable in container- and in orchard- grown trees. Thus it appears that irrigation of

carambola trees in Krome very gravelly loam soil in an orchard can be applied at 17%

SWD without a significant effect on fruit yield or quality.

114

Table 6-1. Effect of soil water depletion (SWD) on fresh and dry weights of leaves, stem, and roots of carambola trees in containers.

Treatments Leaf fresh wt. (g)

Leaf dry wt. (g)

Stem fresh wt. (g)

Stem dry wt. (g)

Root freshwt. (g)

Root dry wt. (g)

Total dry wt. (g)z

Root:shoot ratioy

0-21% SWD 318.55 ax 98.89 a 2400.83 a 1032.79 a 2158.33 a 1005.29 a 2137.00 a 0.89 a 22-31% SWD 350.17 a 111.76 a 1824.00 b 820.29 ab 1947.00 a 909.29 a 1841.30 ab 0.97 a 32-50% SWD 433.20 a 153.22 a 1965.00 ab 873.29 ab 1767.00 a 805.29 a 1831.80 ab 0.82 a 51-60% SWD 229.36 a 67.89 a 1610.83 b 631.96 b 1552.50 a 728.62 a 1428.50 b 1.11 a

zTotal dry weight of leaves, stems and roots. yRoot:shoot ratio based on dry weight. xMeans within columns followed by the same letter are not significantly different according to Duncan’s multiple range test, P ≤ 0.05.

114

115

Table 6-2. Effect of soil water depletion (SWD) on total fruit number and fruit weight of carambola trees in an orchard.

Treatments Totalz fruit no. Totalz fruit wt. (kg)0-8% SWD 1107.3 ay 103.02 a 9-11% SWD 881.7 a 76.21 a 12-14% SWD 1123.3 a 112.33 a 15-17% SWD 1261.7 a 116.05 a

zTotal of four harvests per tree (Aug. 2002, Dec. 2002, Aug. 2003, and Dec. 2003) of fruit excluding immature fruit. yMeans within columns followed by the same letter are not significantly different according to Duncan’s multiple range test, P ≤ 0.05.

116

Table 6-3. Effect of soil water depletion (SWD) on fresh weight of mature fruit (108 ohue) of carambola trees in an orchard.

Fruit fresh wt. (g) Treatments Aug. 2002 Dec. 2002 Aug. 2003 Dec. 2003 0-8% SWD 121.63 abz 127.21 a 125.59 a 147.42 b 9-11% SWD 129.83 a 120.40 ab 122.73 a 81.02 c 12-14% SWD 117.60 b 111.48 b 116.78 a 147.79 b 15-17% SWD 123.74 ab 119.90 ab 115.54 a 162.72 a

zMeans within columns followed by the same letter are not significantly different according to Duncan’s multiple range test, P ≤ 0.05.

117

Table 6-4. Effect of soil water depletion (SWD) on dry weight of mature fruit (108 ohue) of carambola trees in an orchard.

Fruit dry wt. (g) Treatments Aug. 2002 Dec. 2002 Aug. 2003 Dec. 2003 0-8% SWD 11.89 bz 11.23 a 17.86 b 15.21 a 9-11% SWD 15.45 a 10.56 a 16.64 b 13.20 bc 12-14% SWD 12.34 b 9.11 b 20.22 a 12.31 c 15-17% SWD 15.56 a 9.49 b 17.64 b 14.33 ab


118

Table 6-5. Effect of soil water depletion (SWD) on fresh weight of ripe fruit (tree-ripened, 87 ohue) of carambola trees in an orchard.

Fruit fresh weight (g) Treatments Aug. 2002 Dec. 2002 Aug. 2003 Dec. 2003 0-8% SWD 163.17 az 129.44 a 154.46 a 182.84 b 9-11% SWD 141.55 b 124.78 a 155.64 a 96.61 c 12-14% SWD 143.95 b 131.09 a 137.96 b 180.19 b 15-17% SWD 151.00 ab 124.07 a 146.13 ab 202.85 a


119

Table 6-6. Effect of soil water depletion (SWD) on dry weight of ripe fruit (tree-ripened, 87 ohue) of carambola trees in an orchard.

Fruit dry weight (g)

Treatments Aug. 2002 Dec. 2002 Aug. 2003 Dec. 2003 0-8% SWD 14.70 bz 11.97 a 21.00 b 22.85 a 9-11% SWD 14.02 b 10.93 ab 21.30 b 17.24 c 12-14% SWD 14.65 b 11.41 a 23.77 a 16.37 c 15-17% SWD 17.02 a 10.17 b 21.33 b 19.75 b


120

Table 6-7. Effect of soil water depletion (SWD) on fresh and dry weights of mature fruit (108 ohue) of carambola trees in containers.

Fruit fresh wt.(g) Fruit dry wt.(g) Treatments Dec. 2002 Dec. 2003 Dec. 2002 Dec. 2003 0-21% SWD 105.61 az 52.58 b 9.39 a 4.20 b 22-31% SWD 87.40 b 67.95 ab 9.81 a 7.67 a 32-50% SWD 84.19 b 87.50 a 9.67 a 7.51 a 51-60% SWD 80.84 b 53.85 b 9.64 a 5.60 ab


121

Table 6-8. Effect of soil water depletion (SWD) on fresh and dry weights of ripe fruit (tree-ripened, 87 ohue) of carambola trees in containers.

Fruit fresh wt.(g) Fruit dry wt.(g)

Treatments Dec. 2002 Dec. 2003 Dec. 2002 Dec. 2003 0-21% SWD 108.69 az 51.55 b 9.62 b 4.46 b 22-31% SWD 113.56 a 93.72 a 9.89 b 5.65 b 32-50% SWD 53.34 b 97.68 a 11.55 a 9.47 a 51-60% SWD 56.05 b 62.25 b 11.23 a 5.98 b


122

Table 6-9. Effect of soil water depletion (SWD) on total soluble solid (TSS) content of

ripe fruit (tree-ripened, 87 ohue) of carambola trees in an orchard. TSS (%) Treatments Aug. 2002 Dec. 2002 Aug. 2003 Dec. 2003 0-8% SWD 7.45 cz 9.13 a 9.31 a 10.43 c 9-11% SWD 8.36 b 8.73 b 9.33 a 10.68 bc 12-14% SWD 9.00 a 8.34 c 9.04 a 11.04 b 15-17% SWD 8.08 b 8.13 d 8.99 a 11.97 a


123

Table 6-10. Effect of soil water depletion (SWD) on total soluble solid (TSS) content of ripe fruit (tree-ripened, 87 ohue) of carambola trees in containers.

TSS (%) Treatments Dec. 2002 Dec. 20030-21% SWD 8.11 az 10.61 c 22-31% SWD 8.14 a 10.83 bc 32-50% SWD 7.86 a 11.28 ab 51-60% SWD 7.45 a 11.44 a


124

Shoot

length

(cm

)

0

10

20

30

40

50

0-8% SWD 9-11% SWD12-14% SWD15-17% SWD

Date01

-Mar

01-A

pr

01-M

ay

01-Jun

01-Jul

01-A

ug

01-S

ep

01-O

ct

01-N

ov

01-D

ec

Shoot

length

(cm

)

0

10

20

30

40

50

A

B

Fig. 6-1. Effect of soil water depletion (SWD) on shoot length of orchard-grown carambola trees during 2002 (A) and 2003 (B). Symbols with vertical bars represent means ± SE of 4 shoots per tree for 6 trees.

125

Date

01-Jul-0

2

01-S

ep-0

2

01-N

ov-0

2

01-Jan

-03

01-M

ar-0

3

01-M

ay-0

3

01-Jul-0

3

01-S

ep-0

3

01-N

ov-0

3

Tru

nk

dia

met

er (

mm

)

20

25

30

35

40

45

50

0-21% SWD22-31% SWD32-50% SWD51-60% SWD

Fig. 6-2. Effect of soil water depletion (SWD) on trunk diameter of young carambola

trees grown in containers. Symbols with vertical bars represent means ± SE of 6 trees.

126

Date

01-Jun

-03

01-Jul-0

3

01-A

ug-0

3

01-S

ep-0

3

01-O

ct-0

3

01-N

ov-0

3

01-D

ec-0

3

Fruit len

gth

(m

m)

30

40

50

60

70

80

90

100 0-8% SWD 9-11% SWD12-14% SWD15-17% SWD

Summer harvest Winter harvest

Fig. 6-3. Effect of soil water depletion (SWD) on fruit length of orchard-grown

carambola trees during 2003 summer harvest (August) and winter harvest (December). Symbols with vertical bars represent means ± SE of 4 fruit per tree for 6 trees.

CHAPTER 7 PHENOLOGICAL CYCLES OF CARAMBOLA AT FOUR LEVELS OF SOIL

WATER DEPLETION IN SOUTH FLORIDA

The phenological cycle describes the morphological developmental pattern of tree

crops (Cull, 1986). In perennial evergreen trees such as carambola, the annual

phenological cycle consists of multiple factors operating in synchrony, producing multi-

order interactions that influence productivity (Cull, 1986). This pattern depends on

species, cultivar, management, and environment. Environmental factors, however, are

considered the most important factors that affect tree performance and management

strategies (Whiley et al., 1988).

The phenological cycle of tropical evergreen trees can be divided into three phases:

the vegetative phase, the flower development phase, and the root development phase

(Cull, 1986). Each phase has an effect on the other phases. For example, vegetative

growth is essential for carbohydrate accumulation that is required for the development of

young, photosynthetically efficient leaves, root growth, vegetative flushes, and fruit

growth and development (Cull, 1986). The phases of the phenology cycle of tropical fruit

trees are influenced by climatic conditions, mainly temperature and precipitation (Cull,

1986). The effects of climatic and environmental conditions on annual phenological

cycles have been studied in several tropical and subtropical fruit crops including avocado

(Cull, 1986; Salomon, 1984; Whiley et al., 1988), citrus (Salomon, 1984), mango

(Whiley, 1993), and olive (Cimato et al., 1990). Just as the phenology of deciduous

temperate trees is mainly determined by seasonal variations in temperature and

127

128

photoperiod, the phenology of tropical fruit is affected by seasonal variability in water

availability (Reich and Borchert, 1984). Phenological cycles of tropical evergreen trees

are less obvious than those of temperate fruit trees. Tropical fruit growers must recognize

annual critical growth changes in order to utilize growth cycles as orchard management

tools (Whiley et al., 1988). Environment and orchard cultural practices affect vegetative

and reproductive development throughout the annual phenological cycle of carambola.

The effect of several environmental factors on carambola have been studied including

low temperatures (Campbell and Malo, 1981; Campbell et al., 1985; Crane, 1994; George

et al., 2002a, 2002b; Watson et al., 1988), drought (Ismail and Awang, 1992; Ismail and

Noor, 1996a; Ismail et al., 1994, 1996; Salakpetch et al., 1990) , flooding (Ismail and

Noor, 1996b; Joyner and Schaffer, 1989), sunlight, (Marler et al., 1994), and wind

(Campbell, 1985; Crane, 1993; Crane et al., 1994a; Marler and Zozor, 1992). In addition

to environmental factors, cultural practices such as cultivar selection (Salakpetch et al.,

1990), mulching (George et al., 2000; 2001), fertilization (Campbell, 1985,1989; Crane,

1994; Knight, 1982), and pruning (Crane et al., 1991; Núñez-Elisea and Crane, 1998,

2000) also influence vegetative and reproductive growth of carambola trees.

In their native, tropical habitat, carambola trees grow, flower, and fruit year-round

(Núñez-Elisea and Crane, 2000). Trees may produce new flushes, flowers, and immature

and mature fruit simultaneously (Galán Saúco et al., 1993). Under subtropical conditions

such as those in south Florida, canopy development is poor and trees defoliate from

February to March presumably due to low temperatures, dry winds, and high light

intensity (Núñez-Elisea and Crane, 1998). Trees begin to refoliate from March to mid-

May and may continue to produce new leaves until mid-October (Núñez-Elisea and

129

Crane, 1998). Shoots of carambola alternate growth with periods of quiescence. Apical

growth results in an extension of the shoot axis, with each growth flush containing

between eight and 12 leaves (Núñez-Elisea and Crane, 1998). Shoot flushing is enhanced

by warm weather and rain. Shoots formed during the spring flush about four to five times

per growth season but those formed in the late summer normally flush once before shoot

extension growth ceases in the fall (Núñez-Elisea and Crane, 1998).

Carambola has several flowering periods and may flower continuously during the

year under favorable environmental conditions (Núñez-Elisea and Crane, 1998, 2000).

The carambola production season in south Florida extends from mid-July to mid-

February (Campbell et al., 1985; Núñez-Elisea and Crane, 1999, 2000). Two reproductive

growth phases occur in south Florida. The first phase is characterized by a flowering peak

in May and with fruit maturing from July to October. The second phase has a peak

bloom in September with fruit maturing from December to February (Campbell et al.,

1985; Núñez-Elisea and Crane, 1998). These flowering phases and the ability of different

wood types to flower indicate that floral initiation is affected by environmental,

physiological, and cultural factors (Núñez-Elisea and Crane, 2000). Manipulation of

growth and production cycles, therefore, requires an understanding of developmental and

phenological properties of the tree.

Soil water content is among the three major growth and productivity determinants

(in addition to temperature and nutrition) that can be regulated throughout the

phenological cycles of tropical fruit crops (Cull, 1986). A few studies have investigated

the effect of soil water depletion (SWD) on the phenology of tropical fruit crops. For

example, water requirements of avocado trees in relation to the growth cycle were studied

130

by Whiley et al. (1988). Núñez-Elisea and Crane (1998) studied the phenological cycles

of carambola trees in South Florida. However, little is known about the response of

carambola phenological cycles to SWD in Krome very gravelly loam soil in a subtropical

climate like Florida. This study was carried out to determine the effect of four levels of

SWD on the phenological cycle of carambola trees grown in the subtropical climate of

South Florida.


Experimental site and plant material. In an orchard, 8-year-old ‘Arkin’

carambola trees grafted on open-pollinated Golden Star rootstock were visually

monitored for phenological growth cycles throughout 2003. The trees were planted at 4.5

m within and 6.0 m between rows and were covered by an artificial (polypropylene

ribbon shade cloth) windbreak on the northern, eastern, and western perimeters and

sapodilla [Manilkara zapota (L.) von Royen] trees on the southern perimeter of the

orchard. All the trees were topped to a 2.7 m height on 23 Apr. 2003.

The orchard was located at the Tropical Research and Education Center in

Homestead, Fla. at 25.5 oN Lat. and 80.5 oW Long. Average monthly temperatures in

2003 were lowest in January (17.9 oC) and the highest (30.4 oC) in July (Fig. 7-1), which

are near the ideal temperature range for growth of carambola (21 to 32 oC) (Ngah et al.,

1989). Historical climatic records of Homestead, Fla. for 41 years showed an average

annual precipitation of 1,571 mm, 79% of which occurred from May through October

(Fig. 7-1). The same pattern occurred during 2003, when 74% of the 1,606 mm total

precipitation occurred during this period. The soil at the experimental site is Krome very

gravelly loam [loamy-skeletal, carbonatic, hyperthermic, Lithic Udorthents (Noble et al.,

131

1996)]. Trees were planted at the intersection of perpendicular trenches that were

approximately 50 to 75 cm wide and 50 cm deep (Colburn and Goldweber, 1961).

A second experiment was also conducted during the same period in 2003 to study

the phenological cycle of 2-year-old ‘Arkin’ carambola trees grafted on Golden Star

rootstock planted in 95 L containers on 4 June 2002. The containers were filled with

Krome very gravelly loam soil and placed in the field adjacent to the orchard used for the

phenological study of mature trees. Fertilizer and pest management were in accordance

with a standard practice for commercial carambola production in south Florida (Crane,

1994).

Soil water depletion treatments. Soil water depletion was determined by

monitoring soil water content with multisensor capacitance probes (EnviroSCAN, Sentek

PTY Ltd., Kent Town, Australia). Three probes per treatment were installed 60 cm north

of the trunk of trees in the orchard and 20 cm from the trunk of trees in containers. Each

probe in the orchard contained four capacitance sensors located at 10, 20, 30, and 50 cm

depths in the soil. Sensors were placed at 10, 20, and 30 cm depths in containers. The

sensors recorded soil water content every 30 min. The data were stored in a datalogger

and later downloaded to a portable computer for analysis. Installation of the system in

Krome soils was described previously by Al-Yahyai et al. (2003) and Núñez-Elisea et al.

(2001), and technical specifications of the multisensor capacitance probes were discussed

by Paltineanu and Starr (1997). Data from the 10, 20 and 30 cm depth sensors were

averaged and plotted using EnviroSCAN software (EnviroSCAN 4.0, Sentek PTY Ltd.,

Kent Town, Australia).

132

Field capacity (FC) values were determined based on the amount of water in the

soil after excess water had drained and the rate of downward movement had decreased.

Trees in the orchard and in containers were irrigated with microsprinklers (Maxijet,

Dundee, Fla., USA) which had a discharge rate 89 L·h-1 and a 360o wetting pattern.

Levels of soil water depletion were predetermined in the orchard and in containers

based on a preliminary study (Al-Yahyai, unpublished data) whereby irrigation water was

withheld and soil water content was monitored with capacitance probes until visual water

stress symptoms (leaf yellowing and abscission) appeared on the trees. Containers were

placed on metal bases above the ground to prevent capillary movement of water from the

soil to the containers.

Irrigation of the orchard trees was initiated when SWD reached one of the

following four levels (where 0% SWD = FC): 0-8% SWD, 9-11% SWD, 12-14% SWD,

or 15-17% SWD. Irrigation of container trees was initiated when SWD reached one of

the following four levels: 0-21% SWD, 22-31% SWD, 32-50% SWD, and 51-60% SWD.

Phenological observations. Orchard and container-grown trees were visually

evaluated weekly for 37 weeks in 2003. Observations of vegetative and reproductive

development included: 1) shoot flushing: shoot growth following bud break or

resumption of growth of quiescent shoots. Following the initial post-winter shoot flush,

observations were made on new axillary bud break (George et al., 2000); 2) shoot

growth: the relative number of continuously growing axillary shoots; 3) flowering:

growth of the panicle from fully developed buds to flower opening (bloom) (Whiley et

al., 1988). Most panicles were formed on branches on the periphery of the canopy, and

quiescent buds on older branches. Flowers were also formed on scaffold branches and on

133

the main trunk (Núñez-Elisea and Crane, 2000); 4) fruit load: the relative number of fruit

per tree. Fruit from six trees per treatment were completely harvested on 18-22 Aug.

(summer harvest) and on 15-19 Dec. (winter harvest), which coincided with the peak

harvesting season in south Florida for 2003.

Experimental design and statistical analysis. Treatments in the orchard and

containers were arranged in a completely randomized design. Each treatment consisted of

three rows. In the orchard, the numbers of trees per treatment were 15 for 0-8% SWD, 11

for 9-11% SWD, 11 for 12-14% SWD, and 14 for the 15-17% SWD treatment. For

containers, each treatment was replicated three times with two trees per replication. The

lack of uniformity in replications among treatments in the orchard was due to some rows

containing unhealthy trees or support posts for the artificial windbreak that resulted in an

uneven number of trees per row. Trees were evaluated based on a score of 0 to 100.

Ranking were based on the following rating: 0 (none), 1 (very low), 2 (low), 3 (medium),

4 (high), and 5 (very high to maximum). A similar approach was previously used to

evaluate carambola canopy leaf color (George et al., 2000) and flowering (Crane et al.,

1991; George et al., 2001). The ranked data were analyzed using the Kruskal-Wallis non-

parametric test for unequal population sample size (Miller, 1980; Neter et al., 1990). Data

were analyzed using the NPAR1WAY procedure in SAS (SAS Institute, Cary, N.C.).

Multiple mean comparisons were performed as described by Miller (1980) and Neter et

al. (1990).

Results

Shoot flush. In trees in the orchard, shoot flushing began during the last week of

February and reached its peak by mid-March (Fig. 7-2A.). After the initial shoot flush in

the spring, trees continued to flush but at a low rating until the third week of October.

134

Another small increase in shoot flushing occurred following harvest at the end of

September and peaked at the beginning of October. The highest average spring flush

rating, 4.6, occurred in the middle of March for trees in the 9-11% SWD treatment. The

fall flushing was highest for trees in the 12-14% SWD treatment with an average rating

of 1.8 in the first week of October. There were no significant differences in shoot

flushing among treatments during the peak spring and fall shoot flushes. Shoot flushing

during May and June occurred after trees were topped on 23 Apr. Shoot growth ceased

toward the end of October and trees became quiescent during November and December.

There was a significant difference in shoot flushing on 28 Feb. between trees in the 0-8%

SWD that had a high rating (3.8) and those in the 12-14% SWD treatment with a low

rating (2.8). There were no significant differences in shoot flushing ratings among the

other treatments (Table 7-1). There were also no significant differences in shoot flushing

ratings among treatments for the rest of the year.

Shoot growth of container-grown carambola trees began during the third week of

February (Fig. 7-2B). However, unlike orchard trees, shoot flushing of container-grown

trees fluctuated throughout the year and no bud break was observed from mid-May to

mid-July and from mid-November until the end of December. Trees in the 0-21% SWD

and 22-31% SWD treatments had higher shoot flushing ratings than trees in the 32-50%

and 51-60% SWD treatments from mid-February to mid-May. Container-grown trees

started a new shoot flush in August that reached a peak by the first week of September.

Shoot flushing continued at a low level until it ceased by the second week of November.

Significant differences were observed on 25 Apr. between trees in the 0-21% SWD

treatment that had a low shoot flushing rating of 1.7 and trees in the 22-31% SWD

135

treatment that had no shoot flushes (Table 7-2). Although shoot flushing varied among

treatments, no further significant differences among irrigation treatments were observed.

Extension shoot growth. Shoot extension of carambola trees in the orchard was

observed from the first week of March until the third week of December. Extension shoot

growth was very rapid at the beginning of the growing season in March with a gradual

decrease in growth by the end of September followed by a rapid decline from October to

November until growth ceased toward the end of December (Fig. 7-3A). Shoot extension

ratings of the axillary shoots fluctuated from March to September.

During the initial growth of the axillary shoots of trees in the orchard in March and

April, there were no significant differences among treatments (Table 7-1). However, trees

in the 9-11% SWD treatment had a significantly higher shoot growth rating than trees in

the 12-14% SWD treatment on 23 May, 30 May, 7 June, and 13 June. Also significantly

more shoot growth occurred in trees in 0-8% SWD than in the 12-14% SWD treatment.

In addition, trees in the 0-8% and 9-11% SWD treatments had significantly higher shoot

growth rating than trees in the 12-14% SWD treatment on 8 Aug. No significant

differences in shoot growth rating were observed between trees irrigated at 0-8% and 15-

17% SWD (Table 7-1).

Container-grown trees had similar shoot growth patterns to those of orchard trees

throughout the year (Fig. 7-3B). Shoot growth of container-grown trees was high during

the last week of March and the first week of April. Unlike orchard trees, however, shoot

growth ratings of trees in containers did not gradually decrease until September (Fig. 7-

3B). Trees in the 32-50% and 51-60% SWD treatments generally had fewer shoots

growing than trees in the 0-21% or 22-31% SWD treatments.

136

Carambola trees in the 0-21% SWD had significantly fewer axillary shoots

growing than trees in the 51-60% SWD treatment on 25 Apr. (Table 7-2). However, trees

in the 0-21% SWD treatment had more axillary shoots growing than trees in the 32-50%

SWD treatment on 30 May and 7 June, and the 51-60% SWD treatment on 19 Sept.

Axillary shoot growth ratings of these trees were significantly higher than for trees in the

32-50% SWD treatment (Table 7-2) on 30 May and 7 June. Shoot growth ratings of trees

in the 0-21% SWD treatment were significantly higher than that of trees in the 51-60%

SWD treatment.

Flowering. Flowering of trees in the orchard was first observed in the first week of

March (Fig. 7-4A). However, fruit set from this bloom was very low. During the second

half of April, flowering increased and reached its peak during the last 2 weeks of May.

Flowering then continued at a moderate level until the fruit were harvested during the

third week of August. Following harvest, flowering peaked again from mid-September to

mid-October. The amount of flowers decreased following this period and flowering

ceased by the end of December (Fig. 7-4A). Trees in the 9-11% SWD treatment had a

significantly higher flowering rating than trees in the 15-17% SWD treatment on 1 Oct.

(Table 7-1). On 17 Oct., trees in the 9-11% SWD treatment had the highest flowering

rating (Table 7-1).

The first flowering peak in container trees occurred in March and flowering

declined toward the end of April (Fig. 7-4B). Low flowering ratings were observed from

mid-May to the end of July. An increase in flowering began in August which reached its

peak by the end of October, and then declined by December (Fig. 7-4B). Container-

grown trees had lower flowering ratings during the summer and fall compared to that of

137

orchard trees. On 7 and 14 Mar., trees in the 0-21% SWD treatment had a significantly

higher flowering rating than trees in the 51-60% SWD treatment (Table 7-2).

Fruiting. In the orchard, fruit that were not harvested in the winter remained on

the tree until March when they ripened and abscised (Fig. 7-5A). Trees did not set fruit

from newly formed flowers until the second half of April, with a sharp increase in fruit

set during the first two weeks of May. Fruit production continued as new flowers

produced and set fruit for the rest of the year. The major fruit harvest periods were

August and December. However, trees continuously produced fruit during the year.

Orchard trees in the 15-17% SWD treatment had a significantly higher fruit number

rating on 30 May than trees in the 12-14% SWD treatment but fruit number ratings in the

15-17% SWD and 12-14% SWD treatments did not differ from that in the 9-11% SWD

and 0-8% SWD treatments (Table 7-1).

Fruiting was erratic and there was no harvestable yield for carambola trees in

containers until October with a peak yield in mid-November (Fig. 7-5B). During the first

flowering phase there were no harvestable fruit in August, resulting in only one harvest in

December. Fruit number ratings were not significantly different among treatments

throughout the year. This was attributed to the low number of fruit produced and the

variation in fruit yield among container-grown trees in 2003.

Discussion

In the subtropical climate of South Florida, carambola trees did not produce shoot

flushes until mid-February in 2003. Adverse climatic conditions including low

temperatures and dry winds were cited as the reason for this lack of shoot development

during that time of year in south Florida (Campbell et al., 1985; George et al., 2002a,

2002b; Núñez-Elisea and Crane, 1998, 2000) in comparison to tropical climates where

138

carambola continuously produce new shoots (Núñez-Elisea and Crane, 2000; Galán

Saúco et al., 1993). The peak amount of shoot flushing occurred in the spring following a

period of quiescence and a second peak of shoot growth flush occurred in the fall

following the summer harvest, before shoot growth ceased in the winter. Continuous

shoot growth flushing throughout the spring, summer, and winter was due to an alteration

of growth and quiescence as observed by Núñez-Elisea and Crane (1998). Following

topping of orchard trees in April, pruned branches produced 70% of the new growth in

May and June, whereas container-grown trees did not flush during this period. Pruning of

carambola increased vegetative shoot flushing (Núñez-Elisea and Crane, 2000). Shoot

flushing of carambola trees in the orchard and in containers was not affected by irrigation

at four levels of SWD under the subtropical climatic conditions of south Florida in 2003.

The lack of response was presumably due to sufficient soil water from precipitation (Fig.

7-1) in the orchard and in containers. In addition, capillary rise of water from the high

water table (1-2 m below soil surface) to the root zone may have minimized SWD

treatment effects on orchard trees.

Overall shoot growth was not affected by SWD treatments in the orchard. This lack

of response was due to soil water depletion not reaching sufficiently low levels to cause a

reduction in shoot growth. In Malaysia, Bookeri (1996) also found no response of overall

growth of carambola trees in an orchard to irrigation due to sufficient rainfall following a

drought. Similarly, Núñez-Elisea and Crane (1998) reported that, in southern Florida,

vegetative growth of carambola occurred throughout the spring and summer months in

the presence of warm temperatures and rain. For container trees where lateral and

capillary soil water movement into the root zone was restricted, trees irrigated at field

139

capacity had a significantly higher extension shoot growth than trees in the 32-50% and

51-60% SWD treatments. A reduction of shoot growth of fruit crops in response to water

stress is well documented (Hsiao and Acevedo, 1974; Lakso, 1985). A reduction in

vegetative growth in response to water stress was reported previously for container-

grown (Ismail et al., 1994; Ismail and Noor, 1996a) and field-grown carambola trees

(Ismail et al., 1996).

Flowering of carambola trees in an orchard and in containers in South Florida in

2003 occurred in March. Flowering in March may have resulted from above average

temperatures in February and March. The average monthly temperature in February was

22.3 oC (annual average is 19.9 oC) and maximum temperature was 30.8 oC. In March,

the average monthly temperature was 25.9 oC (annual average is 22.1 oC) and maximum

temperature was 32.3 oC. In the subtropical climate of south Florida, George et al. (2001)

found that flowering of carambola trees in an orchard increased in response to increased

soil temperature. In contrast, flowering of young container-grown carambola was less

affected by temperature than by water stress in a greenhouse experiment in Australia

(Salakpetch et al., 1990). For orchard trees in our study, two flowering periods were

observed, the first was from April to June with a peak in May, and the second flowering

period was from August to October with a peak in September. These two flowering peak

periods in South Florida were also observed by Campbell et al. (1985) and Núñez-Elisea

and Crane (1998). Under optimum environmental conditions, however, carambola has

several flowering periods throughout the year (Samson, 1980; Tidbury, 1976; Núñez-

Elisea and Crane, 1998; 2000).

140

Orchard trees were able to sustain more flowers during the first flowering period

(April to June) than container trees. It can be speculated that due to the limited

carbohydrate reserves in young container trees following initial flowering in March,

flowering ratings were very low in the first flowering period from April to June. At

critical stages of the growth cycle, competition for carbohydrate reserves between

reproductive and vegetative growth leads to depletion of these reserves (Whiley et al.,

1988). In container trees, vegetative growth may have taken precedence over

reproductive growth to replenish these carbohydrate reserves (Anuar et al., 1992), which

explains the absence of the first flowering phase in container-grown carambola trees.

Overall carambola flowering in orchard and container-grown trees was not affected by

SWD levels except for two observation dates when trees receiving the least amount of

water had the fewest flowers. The inconsistent response of flowering of carambola trees

to SWD in this study was similar to observations for avocado, lychee and mangoes. In

avocado and lychee, Chaikiattiyos et al. (1994), reported that water stress reduced

vegetative growth but did not induce flowering. Flowering of mango in response to water

deficit was also found to be inconsistent, where trees in the orchard flowered but those in

a greenhouse did not flower following rewatering (Whiley, 1993).

Fruit production of carambola trees was not influenced by soil water depletion

levels. This lack of response is contrary to what was found in lychee (Menzel et al.,

1995), where fruit set and growth was reduced by water stress. Adequate soil water for

fruit growth and development of carambola trees in an orchard in south Florida resulted

in no differences in fruit number among SWD treatments. Fruit production of container-

141

grown trees was negligible and varied greatly, thus no significant differences were

observed among treatments.

Conclusions

Phenological cycles of carambola trees including shoot flushing, extension shoot

growth, flowering, and fruiting showed little response to irrigation at four SWD levels.

This lack of response was likely caused by sufficient soil water due to precipitation and

capillary rise from the high water table located about 1-2 m below the soil surface.

Environmental factors such as temperature and cultural practices such as pruning

appeared to have altered the annual phenological cycle by influencing vegetative and

reproductive growth. The phenological cycles of young trees grown in containers were

not as well-defined as those of mature trees in an orchard. The response of container-

grown trees to irrigation varied among trees rather than among treatments. Further studies

over longer periods are needed to better understand the phenology of these trees.

Nonetheless, information on phenology of carambola in this study may provide useful

guidelines for orchard management in South Florida.

142

Table 7-1. The effect of soil water depletion (SWD) on shoot flushing, axillary shoot growth, flowering, and fruit load of orchard-grown carambola trees in 2003.

Ratings

Shootflushing

Axillary shoot growth Flowering Fruit load

Treatments 28 Feb.z 23-May 30-May 7-Jun 13-Jun 3-Jul 8 Aug. 1 Oct. 17 Oct. 30-May

0-8% SWD 3.8 ay 4.6 ab 4.5 ab 4.4 ab 4.3 ab 4.6 a 3.9 a 4.1 ab 3.2 b 2.9 ab

9-11% SWD 3.7 ab 4.7 a 4.8 a 4.8 a 4.7 a 4.4 ab 3.8 a 4.6 a 3.8 a 3.3 ab

12-14% SWD 2.8 b 3.9 b 4.1 b 3.9 b 3.6 b 3.7 b 3.2 b 3.8 ab 3.0 b 2.7 b

15-17% SWD 3.4 ab 4.5 ab 4.5 ab 4.1 ab 3.7 b 3.9 ab 3.3 ab 3.5 b 3.0 b 3.6 a

142

zDates listed when significant differences among treatments were detected (total number of observation dates was 37). yDifferent letters within columns indicate significant differences among means according to the Kruskal-Wallis non-parametric test, α ≤ 0.1.

143143

Table 7-2. The effect of soil water depletion (SWD) on shoot flushing, axillary shoot growth, and flowering of container-grown carambola trees in 2003.

Ratings

Shoot flushing

Axillary shoot growth Flowering

Treatments 25 Apr.z 25 Apr. 30-May 7-Jun 19 Sept 8 Aug. 7 Mar. 14 Mar.

0-21% SWD 1.7 ay 3.3 b 4.8 a 4.8 a 4.0 a 2.8 a 2.8 a 3.2 a

22-31% SWD 0.0 b 4.8 ab 4.4 ab 4.4 ab 3.8 ab 2.0 ab 2.0 ab 2.2 ab

32-50% SWD 1.4 ab 4.6 ab 3.4 b 3.2 b 3.6 ab 1.4 ab 1.4 ab 1.8 ab

51-60% SWD 0.3 ab 5.0 a 4.2 ab 4.0 ab 3.0 b 1.2 b 1.2 b 1.0 b zDates listed if significant differences among treatments were detected (total number of observation dates was 37). yDifferent letters within columns indicate significant differences among means according to the Kruskal-Wallis non-parametric test, α ≤ 0.1.

144

Month

Jan Feb Mar April May June July Aug Sept Oct Nov Dec

Pre

cipitat

ion (

mm

)

0

50

100

150

200

250

300

350

400

450

500

Tem

pera

ture

(o C)

0

5

10

15

20

25

30

35

2003 average monthly rainfall41 yr average monthly rainfall2003 average monthly temp.

Fig. 7-1. Average 41-year and 2003 monthly precipitation and average 2003 monthly temperature in Homestead, Fla. Source: Florida Automatic Weather Network (FAWN) (2003) and The National Oceanic and Atmospheric Administration (NOAA) (2003).

145

Month

Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Shoot

flush

rat

ing

0

1

2

3

4

5

0-8% SWD 9-11% SWD12-14% SWD15-17% SWD

Month


Shoot

flush

rat

ing

0

1

2

3

4

5

0-21% SWD22-31% SWD32-50% SWD51-60% SWD

A

B

Fig. 7-2. Shoot flush ratings of carambola trees in Krome very gravelly loam soil of South Florida in 2003 irrigated at four levels of soil water depletion (SWD) in an orchard (A, n = 51) and in containers (B, n = 22).

146

Month


Shoot

gro

wth

rat

ing

0

1

2

3

4

5 0-8% SWD 9-11% SWD12-14% SWD15-17% SWD

Month


Shoot

gro

wth

rat

ing

0

1

2

3

4

5 0 -21% SWD22-31% SWD32-50% SWD51-60% SWD

A

B

Fig. 7-3. Shoot growth ratings of carambola trees in Krome very gravelly loam soil of South Florida in 2003 irrigated at four levels of soil water depletion (SWD) in an orchard (A, n = 51) and in containers (B, n = 22).

147

Month


Flow

erin

g r

atin

g

0

1

2

3

4

5 0-8% SWD 9-11% SWD12-14% SWD15-17% SWD

Month


Flow

erin

g r

atin

g

0

1

2

3

4

5 0-21% SWD22-31% SWD32-50% SWD51-60% SWD

A

B

Fig. 7-4. Flowering ratings of carambola trees in Krome very gravelly loam soil of South Florida in 2003 irrigated at four levels of soil water depletion (SWD) in an orchard (A, n = 51) and in containers (B, n = 22).

148

Month


Fruit loa

d r

atin

g

0

1

2

3

4

5

0-8% SWD 9-11% SWD12-14% SWD15-17% SWD

Month


Fruit load

rat

ing

0

1

2

3

4

5

0-21% SWD22-31% SWD32-50% SWD51-60% SWD

A

B

Fig. 7-5. Fruit load rating of carambola trees in Krome very gravelly loam soil of South Florida in 2003 irrigated at four levels of soil water depletion (SWD) in an orchard (A, n = 51) and in containers (B, n = 22).

CHAPTER 8 SUMMARY AND CONCLUSIONS

The objective of this research project was to determine the effects of irrigation at

different soil water depletion (SWD) levels, on physiology, growth and yield of 2-year-

old, container-grown, and 8-year-old, field-grown 'Arkin' carambola trees and to relate

SWD to physiology, growth, and yield of carambola. Soil water depletion treatments

were determined by continuous measurement of soil water content with multisensor

capacitance probes (EnviroSCAN). Measurements of soil water content using multisensor

capacitance probes were compared to a neutron probe measurements and tensiometer

readings. Calibration equations for multisensor capacitance probes (EnviroScan) and a

neutron probe in Krome very gravelly loam soil were also determined. The calibration

equations obtained for the capacitance probe was y = 0.011x + 0.5206; r2=0.98, and that

for the neutron probe was y = 29.05x-6.4; r2=0.95. In Krome very gravelly loam soil,

water retention curves of tension determined with tensiometers vs. volumetric water

content using capacitance probes or a neutron probe were fitted with van Genuchten’s

model (1980). The r2 values indicated that the tensiometer tension readings were more

highly correlated with neutron probe measurements of volumetric soil water content than

with multisensor capacitance probe measurements. This suggests that capacitance probe

measurements of soil water content were more variable in response to changes in soil

texture throughout the soil profile than neutron probe measurements. The use of

tensiometers was limited to a soil tension of 20 kPa due to air entry into the water column

of the tensiometer and water column discharge. The use of a neutron probe by growers is

149

150

not practical because its radioactive source requires health and safety monitoring and a

radiation license. Soil water content determined continuously with multisensor

capacitance probes and computer software designed for irrigation management may be a

practical method for irrigation scheduling in tropical fruit tree orchards.

An increase in SWD resulted in reductions in ΨS, A, E, and gs of carambola trees

grown in containers in Krome very gravelly loam soil. In an orchard, SWD never reached

a point where there was a negative impact on ΨS, A, E, or gs. For trees in containers in

Krome very gravelly loam soil, SWD levels above 50% caused a reduction in ΨS reduced

gs. This resulted in a linear reduction in E and a sharp decline in A when gs fell below 50

mmol·m-2·s-1. Leaf gas exchange was more closely correlated with ΨS than with SWD

level. Therefore, ΨS may be better basis than soil water content for irrigation scheduling

for carambola in calcareous soils.

There was no effect of SWD treatments on shoot growth of carambola trees

grown in an orchard or in containers. In the orchard, sufficient soil water content from

precipitation and the shallow water table possibly resulted in sufficient soil water content

to obtain adequate vegetative growth and yields. For trees grown in containers, where

lateral water movement and capillary rise were prevented, trunk diameter was decreased

in trees irrigated at high SWD levels. Soil water depletion of 51-60% reduced total dry

weight of trees in containers. Tree root:shoot ratios were not significantly different

among SWD treatments because there were no significant differences in leaf and root dry

weights among treatments. Fruit length, fruit number, and fruit weight did not differ

significantly among SWD treatments in an orchard. Total soluble solids content and fresh

151

and dry weights of fruit were not related to SWD treatments in an orchard or in

containers.

Phenological cycles including shoot flush, extension shoot growth, flowering, and

fruiting of carambola trees showed little response to irrigation at four SWD levels. This

lack of response was likely caused by sufficient soil water due to precipitation and

capillary rise from the high water table located at about 1-2 m below the soil surface of

the orchard. Further studies over longer time periods are needed to better understand the

phenology of these trees. Nonetheless, information gathered in this study on phenology of

carambola may provide useful guidelines for orchard management in South Florida.

APPENDIX SOIL WATER DEPLETION, PHYSIOLOGY, GROWTH, AND YIELD, OF

CARAMBOLA TREES IN KROME SOIL

Table A-1. Effect of soil water depletion (SWD) on total number and weight of

carambola trees grown in containers.

Treatments Totalz fruit no. Totalz fruit wt. (kg)0-21% SWD 29.17y a 2.26 a 22-31% SWD 20.33 a 1.66 a 32-50% SWD 24.33 a 1.54 a 51-60% SWD 34.17 a 2.18 a

zTotal of two harvests per tree (2002 and 2003), excluding immature fruit. yMeans within columns followed by the same letter are not significantly different by Duncan’s multiple range test, P ≤ 0.05.

152

153

Fig. A-1. Soil water depletion treatments (SWD) as determined from measurements of

soil water content using multisensor capacitance probes (EnviroSCAN) in an 8-year-old carambola orchard in Krome soil in South Florida. Field capacity was determined based on the pattern of soil water depletion over a period of 7 d. Visual stress symptoms (VSS) occurred when the leaves became wilted and chlorotic (photo inset in top right corner of graph).

154

Table A-2. Average soil water depletion (SWD) prior to irrigation of carambola trees in containers in the field and in an orchard in 2002 and 2003.

Container average SWD (%) Orchard average SWD (%)

Treatment 2002 2003 Average Treatment 2002 2003 Average

0-21 SWD 17 20 19 0-8 SWD 9 8 8

22-31 SWD 30 31 30 9-11 SWD 10 11 10

32-50 SWD 47 50 49 12-14 SWD 14 14 14

51-60 SWD 59 59 59 15-17 SWD 17 16 17

155

Greenhouse Orchard0.00.20.40.60.81.01.21.41.61.82.02.22.4

Location

VP (k

Pa)

Fig.. A-2. Vapor pressure (VP) difference between saturated and ambient air vapor pressure in the greenhouse and the orchard.

156

ΨS = - 0.57 MPa A = 7.48 µmol m-2 s-1

Onset of water stress

Field capacity

Soi

l wat

er c

onte

nt (%

)

Calendar date Fig. A-3. Soil water depletion (SWD) as measured with capacitance probes

(EnviroSCAN) in an 8-year-old carambola orchard in Krome soils. Points of ‘field capacity’ and ‘onset of water stress’ were determined based on the pattern of soil water depletion. Measurements of stem water potential (ΨS) and net CO2 assimilation (A) indicated by the vertical double-line on the x-axis were chosen randomly at a point beyond the theoretical ‘onset of water stress’ point.

157

Shoo

t flu

sh ra

ting

0

1

2

3

4

5

Shoo

t gro

wth

ratin

g

0

1

2

3

4

Flow

erin

g ra

ting

0

1

2

3

4

Months


Frui

t loa

d ra

ting

0

1

2

3

4

Fig. A-4. Ratings of the phenological cycles of carambola trees in Krome very gravelly loam soil in an orchard in South Florida in 2003 Trees were irrigated at four levels of soil water depletion (SWD) in an orchard (n = 51).

LIST OF REFERENCES

Al-Yahyai, R., B. Schaffer, and F.S. Davies. 2003. Monitoring soil water content for irrigation scheduling in a carambola orchard in a gravelly limestone soil. Proc. Fla. State Hort. Soc. 116: 37-41.

Anuar, A., O. Yaacob, and M.R. Noor. 1992. The nutritional status and the tree phenology of selected fruit trees in three different agro-ecological zones of Malaysia. Acta Hort. 321: 545-552.

Barbera, G. and B. Cascio. 1985. Effects of water stress on lemon summer bloom: The 'Forzatura' technique in the Sicilian citrus industry. Acta Hort. 171:391-370.

Batten, D.J., C.A. McConchie, and J. Lloyd. 1994. Effect of water deficit on gas exchange characteristics and water relations of orchard lychee (Litchi chinensis Sonn) trees. Tree Physiol. 14:1177-1189.

Baumhardt, R.L., R.J. Lascano, and S.R. Evett. 2000. Soil material, temperature, and salinity effects on calibration of multisensor capacitance probes. Soil Sci. Soc. Amer. J. 64:1940-1946.

Begg, J.E. and N.C. Turner. 1970. Water potential gradients in field tobacco. Plant Physiol. 46:343-346.

Begg, J.E. and N.C. Turner. 1976. Crop water deficits. Adv. Agron. 28:161-217.

Boman, B.J. 1996. Effect of micro irrigation frequency on Florida grapefruit. Proc. Intl. Soc. Citricult. 2:678-682.

Bookeri, M. 1996. Effect of irrigation on carambola (Averrhoa carambola) production in drought prone areas. Proc. Intl. Conf. Trop. Fruit. Kuala Lampur, Malaysia. p. 317-320.

Braun, P., D. Aspinall, and F. Lenz. 1989. The use of physiological indicators of water status in irrigation scheduling in citrus. Acta Hort. 240:267-270.

Bray, E.A. 1997. Plant responses to water deficit. Trend. Plant Sci. 2:48-54.

Bünemann, E. and J.H. Crane. 2000. Diverse production of tropical fruit in South Florida. Acta Hort. 531:53-56.

158

159

Burman, R.D., R.H. Cuenca, and A. Weiss. 1983. Techniques for estimating irrigation water requirements. Adv. Irr. 2:335-394.

Buss, P. 1989. Irrigation scheduling for horticulture: An integrated approach. Acta Hort. 240:261-270.

Campbell, C.A. and K.E. Koch. 1989. Sugar/acid composition and development of sweet and tart carambola fruit. J. Amer. Soc. Hort. Sci. 114:455-457.

Campbell, C.W. 1985. Tropical fruit in Florida. HortScience 20:16.

Campbell, C.W. 1989. Carambola production in the United States. Proc. Interamer. Soc. Trop. Hort. 33:47-54.

Campbell, C.W. and S.E. Malo. 1981. The carambola. Fact Sheet FC-12. Fla. Coop. Ext. Serv. Fruit Crops.

Campbell, C.W., R.J. Knight, and R. Olszack. 1985. Carambola production in Florida. Proc. Fla. State Hort. Soc. 98:145-149.

Campbell, G.S. and M.D. Campbell. 1982. Irrigation scheduling using soil moisture measurements: Theory and practice. Adv. Irr. 1:25-42.

Castel, J. and A. Buj. 1990. Response of Salustiana oranges to high frequency deficit irrigation. Irr. Sci. 11:121-127.

Chaikiattiyos, S., C.M. Menzel, and T.S. Rasmussen. 1994. Floral induction in tropical fruit trees: Effect of temperature and water supply. J. Hort. Sci. 69:397-415.

Chalmers, D.J., P. Mitchell, and P.H. Jerie. 1984. The relation between irrigation, growth, and productivity of peach trees. Acta Hort. 173:283-288.

Chanasyk, D.S. and M.A. Naeth. 1988. Measurements of near-surface soil moisture content with a hydrogenously shielded neutron probe. Can. J. Soil Sci. 68:171-176.

Chanasyk, D.S. and M.A. Naeth. 1996. Field measurement of soil moisture using neutron probes. Can. J. Soil Sci. 76:317-323.

Chartzoulakis, K., B. Noitsakis, and I. Therios. 1993. Photosynthesis, plant growth and dry matter distribution in kiwifruit as influenced by water deficits. Irr. Sci. 14:1-5.

Cimato, A., C. Cantini, and G. Sani. 1990. Climate-phenology relationships on olive cv. Frantoio. Acta Hort. 286:171-174.

Cohen, M., T. Ameglio, P. Cruiziat, P. Archer, C. Valancogne, and S. Dayau. 1997. Yield and physiological responses of walnut trees in semi-arid conditions: application to irrigation scheduling. Acta Hort. 449:273-280.

160

Colburn, B. and S. Goldweber. 1961. Preparation of oolitic limestone soil for agricultural use. Proc. Fla. State Hort. Soc. 74:343-345.

Cole, P.J. 1989. Irrigation management of temperate fruit crops in Australia. Acta Hort. 240:241-247

Cooper, A., S. Poirier, M. Murphy, M. J. Oswald, and C. Procise. 1998. South Florida Tropicals: Carambola. Fact Sheet FCS 8520. Inst. Food Agr. Sci., Univ. of Fla., Gainesville.

Comstock, J.P. 2002. Hydraulic and chemical signaling in the control of stomatal conductance and transpiration. J. Expt. Bot. 53:195-200.

Crane, J.H. 1989. Acreage and plant densities of commercial carambola, mamey sapote, lychee, longan, sugar apple, atemoya, and passion fruit plantings in South Florida. Proc. Fla. State Hort. Soc. 102:239-242.

Crane, J.H. 1993. Commercialization of carambola, atemoya, and other tropical fruit in South Florida. p. 448-460. In: J. Janick and J. E. Simon (eds.), New crops. Wiley, New York, NY.

Crane, J.H. 1994. The carambola (starfruit). Fact Sheet HS-12. Inst. Food Agr. Sci., Univ. of Fla., Gainesville.

Crane, J.H., L. Willis, S.P. Lara, and M. Ellenby. 1991. Short-term effects of manual topping of five-year old 'Arkin' carambola trees. Proc. Fla. State Hort. Soc. 104:57-60.

Crane, J.H., Campbell, R.J., and C.F. Balerdi. 1993. Effect of Hurricane Andrew on tropical fruit trees. Proc. Fla. State Hort. Soc. 106:139-144.

Crane, J.H., A.J. Dorey, R.C. Ploetz, and C.W. Weekly, Jr. 1994a. Post-Hurricane Andrew effects on young carambola trees. Proc. Fla. State Hort. Soc. 107:338-339.

Crane, J.H., C. Balerdi, R. Campbell, C. Campbell, and S. Goldweber. 1994b. Managing fruit orchards to minimize hurricane damage. HortTechnology 4:21-27.

Crane, J.H., R. J. Knight, Jr., O. Rodriguez, and L.C. Crane. 1998. Cultivar tree growth and fruit quality evaluation of young carambola ( Averrhoa carambola L.) trees. Proc. Fla. State Hort. Soc. 111:299-302.

Cuenca, R.H. 1989. Irrigation system design: An engineering approach. Irr. Sys. 115-188.

Cull, B.W. 1986. A phenological cycling approach to tree crop productivity research. Acta Hort. 175:151-156.

161

Davies, F.S. and A.N. Lakso. 1978. Water relations in apple seedlings: Changes in water potential components, abscisic acid levels and stomatal conductances under irrigated and non-irrigated conditions. J. Amer. Soc. Hort. Sci. 103:310-313.

Davies, F.S. and A.N. Lakso. 1979. Water stress responses of apple trees. 1: Effects of light and soil preconditioning treatments on tree physiology. J. Amer. Soc. Hort. Sci. 104:392-395.

Davies, F.S. and J. Bower. 1994. Water stress, gas exchange, and fruit set of 'Olinda' valencia orange trees in eastern Transvaal area of South Africa. Acta Hort. 365:121-125.

Degner, R.L., S.D. Moss, and D.W. Mulkey. 1997. Economic impact of agriculture and agribusiness in Dade County, Florida. Fla. Agr. Mkt. Res. Ctr. Ind. Rpt. 97-I, Food and Resource Econ. Dept., Inst. Food Agr. Sci., Univ. of Fla., Gainesville.

Degner, R.L., T.J. Stevens III, and K. Morgan. 2002. Miami-Dade County agricultural land retention study. Fla. Agr. Mkt. Res. Ctr., Inst. Food Agr. Sci., Univ. of Fla., Gainesville.

Domingo, R. and M.C. Ruiz-Sanchez. 1996. Water relations, growth, and yield of Fino lemon trees under regulated deficit irrigations. Irr. Sci. 16:115-123.

Evett, S.R. 2000. Some aspects of time domain reflectrometry (TDR), neutron scattering, and capacitance methods of soil water content measurement. p. 5-49. In: IAEA, Comparison of soil water measurement using the neutron scattering, time domain reflectometry and capacitance methods. International Atomic Energy Agency, Vienna, Austria.

Evett, S.R. and J.L. Steiner. 1995. Precision of neutron scattering and capacitance type soil water content gauges from field calibration. Soil Sci. Soc. Amer. J. 59:961-968.

Evett, S., J.P. Laurent, P. Cepuder, and C. Hignett. 2002. Neutron scattering, capacitance, and TDR soil water content measurements compared on four continents. 17th World Congr. Soil Sci, 14-21 Aug. 2002, Thailand.

Failla, O., G. Zocchi, C. Treccani, and S. Cocucci. 1992. Growth, development and mineral content of apple fruit in different water status conditions. J. Hort. Sci. 67:265-271.

Fares, A. and A.K. Alva. 2000. Evaluation of capacitance probes for optimal irrigation of citrus through soil moisture monitoring in an entisol profile. Irr. Sci. 19:57-64.

Fereres, E. 1997. Irrigation scheduling of horticultural crops. Acta Hort. 449:253-258.

162

Flore, J.A. and A.N. Lakso. 1989. Environmental and physiological regulations of photosynthesis in fruit crops. p. 111-157. In: J. Janick (ed.). Hort. Rev. Timber Press, Portland, Ore.

Flore, J.A., A.N. Lakso, and J. W. Moon. 1985. The effect of water stress and vapor pressure gradient on stomatal conductance, water use efficiency, and photosynthesis of fruit crops. Acta Hort. 171:207-218.

Galán Saúco, V., U. Menini, and H. Tindall. 1993. Carambola Cultivation. FAO Plant production and protection paper 108. FAO, Rome, Italy.

Gardner, C.M., J.P. Bell, J.D. Cooper, T.J. Dean, M.G. Hodnett, and N. Gardner. 1991. Soil water content. p. 1-73. In: K.A. Smith, and C.E. Mullins (eds.). Soil analysis: Physical methods. Marcel Dekker, New York, N.Y.

Gardner, W. and D. Kirkham. 1952. Determination of soil moisture by neutron scattering. Soil Sci. 73:391-401.

Garnier, E. and A. Berger. 1985. Testing water potential in peach trees as an indicator of water stress. J. Hort. Sci. 60:47-56.

George, A.P. and. R.J. Nissen. 1994. Carambola. p. 206-211. In: B. Schaffer and P. C. Andersen (eds.). Handbook of environmental physiology of fruit crops. Vol. 2: Subtropical and tropical crops. CRC Press, Boca Raton, Fla.

George, H.L., J.H. Crane, B. Schaffer, Y. Li, and F.S. Davies. 2000. Effect of polyethylene and organic mulch on growth and yields of 'Arkin' carambola (Averrhoa carambola L.) in south Florida. Proc. Fla. State Hort. Soc. 113:5-11.

George, H.L., J.H. Crane, B. Schaffer, and Y. Li. 2001. Effect of polyethylene and organic mulch on flowering, pollen viability, percent fruit set, and fruit production of 'Arkin' carambola (Averrhoa carambola L.) in South Florida. Proc. Interamer. Soc. Trop. Hort. 44:72-77.

George, H.L., F.S. Davies, J.H. Crane and B. Schaffer. 2002a. Root temperature effects on 'Arkin' carambola (Averrhoa carambola L.) trees. 1. Leaf gas exchange and water relations. Scientia Horticulturae 96:53-65.

George, H.L., F.S. Davies, J.H. Crane and B. Schaffer. 2002b. Root temperature effects on 'Arkin' carambola (Averrhoa carambola L.) trees. II. Growth and mineral nutrition. Scientia Horticulturae 96:67-79.

Germanà, C. and V. Sardo. 1996. Relationship between net photosynthesis and transpiration rate in citrus trees. Proc. Intl. Soc. Citricult. 2:1117-1121.

Girona, J., M. Mata, D.A. Goldhamer, R.S. Johnson, and T.M. DeJong. 1993. Patterns of soil and tree water status and leaf functioning during regulated deficit irrigation scheduling in peach. J. Amer. Soc. Hort. Sci. 118:580-586.

163

Green, J.G. 1987. Carambola production in Malaysia and Taiwan. Proc. Fla. State Hort. Soc. 100:275-278.

Halevy, A.H. 1972. Water stress and the timing of irrigation. HortScience 7:113-114.

Halle, F., R.A. Oldeman, and P.B. Tomlinson. 1978. Tropical trees and forests: An architectural analysis. Springer-Verlag, New York, N.Y.

Hanan, J. 1972. Repercussions from water stress. HortScience 7:108-112.

Hanson, B.R. and G.L. Dickey. 1993. Field practices affect neutron moisture meter accuracy. Calif. Agr. 47:29-31.

Hanson, B.R. and D. Peters. 2000. Soil type affects accuracy of dielectric moisture sensors. Calif. Agr. 54:43-47.

Hawaii Agricultural Statistics Services (H.A.S.S.). 2001. Hawaii tropical specialty fruit. 1 August 2001. Hawaii Dept. Agr., U.S. Dept. Agr., Honolulu, Hawaii.

Heng, L.K., G. Cayci, C. Kutuk, J.L. Arrillaga and P. Moutonnet. 2002. Comparison of soil moisture sensors between neutron probe, Diviner 2000, and TDR under tomato crops. 17th World Congr. Soil Sci., 14-21 Aug. 2002, Thailand.

Higgs, K.H. and H.G. Jones. 1984. A microcomputer-based system for continuous measurement and recording fruit diameter in relation to environmental factors. J. Expt. Bot. 35:1646–1655.

Hilgeman, R. 1962. Trunk growth of the Valencia orange in relation to soil moisture and climate. Proc. Amer. Soc. Hort. Sci. 82:193-198.

Hillel, D. 1998. Environmental soil physics. Acad. Press, New York, N.Y.

Hsiao, T.C. 1973. Plant responses to water stress. Annu. Rev. Plant Physiol. 24:519-570.

Hsiao, T. 1990. Measurements of plant water status. p. 243-279. In: B. Steward and D. Nielsen (eds.). Irrigation of agricultural crops. Agron. Monogr. No. 30, ASA, Madison, Wis.

Hsiao, T.C. 1993. Growth and productivity of crops in relation to water status. Acta Hort. 335:137-148.

Hsiao, T.C. and E. Acevedo. 1974. Plant responses to water deficits, water-use efficiency, and drought resistance. Agr. Meteorol. 14:59-84.

Hsiao, T.C. and K.J. Bradford. 1983. Physiological consequences of cellular water deficits. p. 227-265. In: Howard M. Taylor, Wayne R. Jordan, and Thomas R. Sinclair (eds.). Limitations to efficient water use in crop production. ASA, Madison, Wis.

164

Huang, X., H. Huang, and F. Gao. 2000. The growth potential generated in citrus fruit under water stress and its relevant mechanisms. Sci. Hort. 83:227-240.

Huguet, J.G., S. H. Li, J. Y. Lorendeau, and G. Pelloux. 1992. Specific micromorphometric reactions of fruit trees to water stress and irrigation scheduling automation. J. Hort. Sci. 67:631-640.

Ismail, M.R. and M. Awang. 1992. Growth and physiological changes of Averrhoa carambola as influenced by water availability. Pertanika 15:1-7.

Ismail, M.R. and K.M. Noor. 1996a. Growth, water relations, and physiological processes of starfruit (Averrhoa carambola L ) plants under root growth restriction. Sci. Hort. 66:51-58.

Ismail, M.R. and K.M. Noor. 1996b. Growth and physiological processes of young starfruit (Averrhoa carambola L.) plants under soil flooding. Sci. Hort. 65:229-238.

Ismail, M.R., S.W. Burrage, H. Tarmizi and M.A. Aziz. 1994. Growth, plant water relations, photosynthesis rate, and accumulation of proline in young carambola plants in relation to water stress. Sci. Hort. 60:101-114.

Ismail, M.R., M.K. Yusaf, and A. Masturi. 1996. Growth and flowering of water stressed starfruit plants and response to ameliorated water stress. Proc. Intl. Conf. Trop. Fruit. Kuala Lampur, Malaysia. 2:97-106.

Jackson, M. 1997. Hormones from roots as signals for the shoots of stressed plants. Trends Plant Sci. 2:22-28.

Jones, H.G. 1990. Physiological aspects of the control of water status in horticultural crops. HortScience 25:19-26.

Jones, H.G., M.T. Luton, K.H. Higgs and P.J. Hamer. 1983. Experimental control of water status in an apple orchard. J. Hort. Sci. 58:301-316.

Jones, H.G., A.N. Lakso, and J.P. Syvertsen. 1985. Physiological control of water status in temperate and subtropical fruit trees. p. 301-344. In: J. Janick (ed.). Hort. Rev. AVI Publ., Westport, Conn.

Joyner, M.E. and B. Schaffer. 1989. Flooding tolerance of 'Golden Star' carambola trees. Proc. Fla. State Hort. Soc. 102:236-239.

Katerji, N., O. Bethenod and P. Quetin. 1994. Behavior of maize stem diameter during drying cycles: Comparison of two methods for detecting water stress. Crop Sci. 34:165-169.

Knight, Jr., R.J. 1965. Heterostyly and pollination in carambola. Proc. Fla. State Hort. Soc. 78:375-378.

165

Knight, Jr., R.J. 1982. Response of carambola seedling populations to Dade County's oolitic limestone soil. Proc. Fla. State Hort. Soc. 95:121-122.

Knight, Jr., R.J. and J.H. Crane. 2002. The 'Arkin' carambola in Florida. Proc. Fla. State Hort. Soc. 115:93-93.

Knight, Jr., R.J., M. Lamberts, and J.S. Bunch. .1984. World and local importance of some tropical fruit crops grown in Florida. Proc. Fla. State Hort. Soc. 97:351-354.

Kozlowski, T.T. 1968. Diurnal changes in diameter of fruit and tree stems of Montmorency cherry. J. Hort. Sci. 43:1-15.

Kramer, D. R. 1980. Drought stress and the origin of adaptations. p. 7-20. In: N.C. Turner and P.J. Kramer (eds.). Adaptation of plants to water and high temperature stress. Wiley, New York, N.Y.

Kristensen, K.J. 1973. Depth intervals and topsoil moisture measurement with the neutron depth probe. Nordic Hydrol. 4:77-85.

Krome, W.H., R.C. J. Koo, and T.W. Young. 1970. Irrigation experiments with 'Tahiti' lime in Dade County. Proc. Fla. State Hort. Soc. 380-381.

Lahav, E. and D. Kalmar. 1977. Water requirements of avocado in Israel. 2: Influence on yield, fruit growth, and oil content. Aust. J. Agr. Res. 28: 869-877.

Lahav, E. and A. Whiley. 2002. Irrigation and mineral nutrition. p. 259-266. In: A.W. Whiley, B. Schaffer, and B.N. Wolstenholme (eds.). The avocado: Botany, production, and uses, CABI, Oxon, UK.

Lakso, A.N. 1985. The effect of water stress on physiological processes in fruit crops. Acta Hort. 171:275-290.

Lamberts, M. and J.H. Crane. 1990. Tropical fruit. p. 337-355. In: J. Janick and J. E. Simon (eds.). Advances in new crops. Timber Press, Portland, Ore.

Lankes, C. 1985. Effect of water stress on transpiration and CO2 gas exchange of the apple leaf and fruit. Acta Hort. 171:305-314.

Larson, K., B. Schaffer, and F.S. Davies. 1989. Effect of irrigation on leaf water potential growth and yield of mango trees. Proc. Fla. State Hort. Soc. 102:226-228.

Levy, Y. 1980. Effect of evaporative demand on water relations of Citrus limon. Ann. Bot. 46:695-700.

Li, Y. 2000. Irrigation scheduling for tropical fruit groves in South Florida. Fact Sheet TR001. Inst. Food Agr. Sci., Univ. of Fla., Gainesville.

166

Li, Y., R. Rao, H. Bryan, and T. Olczyk. 1998. Optimized irrigation schedule to conserve water and reduce nutrient leaching for tomatoes grown on a calcareous gravelly soil. Proc. Fla. State Hort. Soc. 111:58-61.

Li, Y., J. Crane, B. Boman, and C. Balerdi. 2000. Irrigation management survey for tropical fruit crops in South Florida. Proc. Fla. State Hort. Soc. 113:40-42.

Lim, T. 1996. Carambola. 2: Growing and Marketing. Agnote 663-D32. Northern Territory, Australia.

Ludlow, M.M. 1989. Strategies of response to water stress. p. 269-281. In: K. Kreeb, H. Richter, T. and M. Hinckley (eds.). Structural and functional responses to environmental stresses: Water shortage. SPB Acad. Publ., The Netherlands.

Mansfield, T.A. and C.J. Atkinson. 1990. Stomatal behaviour in water stressed plants. p. 241-264.In: R.G. Alscher and J.R. Cumming (eds.). Stress responses in plants: Adaptation and acclimation mechanisms. Wiley-Liss, New York. N.Y.

Martin, F.W., C.W. Campbell, and R.M. Ruberte. 1987. Perennial edible fruit of the tropics: An inventory. U.S. Dept. Agr., Agr. Hdbk 642:45-46

Marler, T.E., M.V. Mickelbart. 1992. Repeated mechanical stress from leaf cuvette influences leaf gas exchange. HortScience 27:432-434.

Marler, T.E., and Y. Zozor. 1992. Carambola growth and leaf gas exchange responses to seimic or wind stress. HortScience 27:913-915.

Marler, T.E., and P.D. Lawton. 1995. Movement influences carambola leaflet chlorophyll fluorescence and temperature under sunny conditions. J. Amer. Soc. Hort. Sci. 360-361.

Marler, T.E., B. Schaffer, and J. H. Crane. 1994. Developmental light level affects growth, morphology, and leaf physiology of young carambola trees. J. Amer. Soc. Hort. Sci. 119:711-718

Matthews, R.F., C.A. Sims, and P.F. West. 1988. Utilization of carambola in a tropical beverage. Proc. Interamer. Soc. Trop. Hort. 32:25-33.

McCutchan, H. and K. Shackel. 1992. Stem-water potential as a sensitive indicator of water stress in prune trees (Prunus domestica L. cv. French). J. Amer. Soc. Hort. Sci. 117:607-611.

Mead, R.M., I.C. Paltineanu, J.E. Ayars, and J. Liu. 1994. Capacitance probe use in soil moisture measurements. ASAE Mtg., Kansas City, Miss., 20-22, June 1994.

Menzel, C.M., J.H. Oosthuizen, D.J. Roe, and V.J. Doogan. 1995. Water deficits at anthesis reduce CO2 assimilation and yield of lychee (Litchi chinensis Sonn ) trees. Tree Physiol. 15:611-617.

167

Merriam, R.A. and K. Knoerr. 1961. Counting time required with neutron soil moisture probes. Soil Sci. 92:394-395.

Miller, R.G. 1980. Simultaneous statistical inference. Springer-Verlag, New York, N.Y.

Mills, T.M., B.E. Clothier, and M.H. Behboudian. 1997. The water relations of 'Braeburn' apple fruit growth under deficit irrigation. Acta Hort. 449:385-392.

Morgan, K.T., L.R. Parsons, T.A. Wheaton, D.J. Pitts, and T.A. Obreza. 1999. Field calibration of a capacitance water content probe in fine sand soils. Soil Sci. Soc. Amer. J. 63:987-989.

Morton, J.1987. Carambola (Averrhoa carambola L.). p. 25-128. In: Julia F. Morton (ed.). Fruit of warm climates. Southern Book Service, Miami, Fla.

Moseley, A.E., R.L. Degner, S.D. Moss, D.L. Hull, R.L. Clouser, and W.D. Mulkey. 1990. Economic impact of agriculture and agribusiness in Dade County, Florida. Food and Resource Econ. Dept., Inst. Food Agr. Sci., Univ. of Fla., Gainesville.

Mostert, P.G. and J.E. Hoffman. 1996. Comparison of four scheduling methods for irrigation of citrus. Proc. Intl. Soc. Citricult. 2:688-691.

Muñoz-Carpena, R., Y. Li, and T. Olczyk. 2002. Alternatives of low cost soil moisture monitoring devices for vegetable production in south Miami-Dade County. Bul. ABE 333. Inst. Food Agr. Sci., Univ. of Fla., Gainesville.

Muñoz-Carpena, R., J. Crane, G. Israel, and C. Yurgalevitch. 2003. Water conservation survey of Miami-Dade County agricultural and golf course commercial water users. Proc. Fla. State. Hort. Sci. 116: 15-21.

Nagarajah, S. 1989. Physiological responses of grapevines to water stress. Acta Hort. 240:249-256.

Nakasone, H.Y. and R.E. Paull. 1998. Tropical fruit. CABI., Oxon, UK.

Nand, D. 1970. Flowering and bearing behaviour of carambola (Averrhoa carambola Linn.). Ind. J. Hort. 27:147-152.

Naor, A. 2000. Midday stem water potential as a plant water stress indicator for irrigation scheduling in fruit trees. Acta Hort 537:447-454.

Naor, A., I. Klein, and I. Doron. 1995. Stem water potential and apple fruit size. J. Amer. Soc. Hort. Sci. 120:577-582.

Naor, A., I. Klein, H. Hupert, Y. Grinblat, M. Peres, and A. Kaufman. 1999. Water stress and crop level interactions in relation to nectarine yield, fruit size distribution, and water relations. J. Amer. Soc. Hort. Sci. 124:189-193.

168

Neter, J., W. Wasserman, and M.H. Kutner. 1990. Applied linear statistical models: Regression, analysis of variance, and experimental designs. Irwin, Boston, Mass.

Neuhaus, A. 2003. Managing irrigation for yield and fruit quality in avocado. Univ. of Western Australia., Australia, PhD Diss.

Ngah, W., I. Ahmad, and A. Hassan. 1989. Carambola production, processing and marketing in Malaysia. Proc. InterAmer. Soc. Trop. Hort. 33:30-43.

Noble, C.V., R.W. Drew, and V. Slabaugh. 1996. Soil survey of Dade County area, Florida. U.S. Dept. Agr., Natural Resources Conservation Serv., Washington, DC.

Núñez-Elisea, R.B. and J.H. Crane. 1998. Phenology, shoot development, and floral initiation of carambola (Averrhoa carambola L. cv. Arkin) in a subtropical environment. Proc. Fla. State Hort. Soc. 111:310-312.

Núñez-Elisea, R. and J.H. Crane. 2000. Selective pruning and crop removal increase early-season fruit production of carambola (Averrhoa carambola L.). Sci. Hort. 86:115-126.

Núñez-Elisea, R., B. Schaffer, M. Zekri, S. K. O’Hair, and J.H. Crane. 2000. Monitoring soil water content in tropical fruit orchards in south Florida with multisensor capacitance probes and tensiometers. HortScience 35:487.

Núñez-Elisea, R., B. Schaffer, M. Zekri, S.K. O’Hair, and J.H. Crane. 2001. In situ soil-water characteristic curves for tropical fruit orchards in trenched calcareous soil. HortTechnology 11:65-69.

Ould Mohamed, S., P. Bertuzzi, A. Bruand, L. Raison, and L. Bruckler. 1997. Field evaluation and error analysis of soil water content measurement using capacitance probe method. Soil Sci. Soc. Amer. J. 61:399-408.

Paltineanu, I.C. and J.L. Starr. 1997. Real-time soil water dynamics using multisensor capacitance probes: Laboratory calibration. Soil Sci. Soc. Amer. J. 61:1576-1585.

Palma, L., V, Novello, and F. Monastra. 1997. First observation on physiological effects of irrigation in pistachio. Acta Hort. 449:545-550.

Peng, Y. H. and E. Rabe. 1998. Effect of differing irrigation regimes on fruit quality, yield, fruit size and net CO2 assimilation of 'Mihowase' satsuma. J. Hort. Sci. Biotech. 73:229-234.

Phene, C.J., R.J. Reginato, B. Itier, and B.R. Tanner. 1990. Sensing irrigation needs. p. 208-261. In: G. J. Hoffman, T. A. Howell, and K. H. Solomon (eds.). Management of farm irrigation systems. ASAE., Joseph, Mich.

Plaut, Z. and A. Meiri. 1994. Crop Irrigation. p. 47-63. In: K.K. Tanji and B.Yaron (eds.). Management of water use in agriculture. Springer-Verlag, New York, N.Y.

169

Popenoe, W. 1948. Manual of tropical and subtropical fruit. p.428-432. Macmillan, New York. N.Y.

Razi, M.I., M. Awang, and S. Razlan. 1992. Effect of water stress on growth and physiological processes of Averrhoa carambola. Acta Hort. 321:505-509.

Reich, P.B. and R. Borchert. 1984. Water stress and tree phenology in a tropical dry forest in the lowlands of Costa Rica. J. Ecol. 72:61-74.

Richards, L.A. 1942. Soil moisture tensiometer materials and construction. Soil Sci. 53:241-248.

Robinson, M. and T.J. Dean. 1993. Measurement of near surface soil water content using a capacitance probe. Hydrol. Processes. 7:77-86.

Salakpetch, S. 1987. Flower ontogeny and pollen germination in carambola (Averrhoa carambola cv. Fwang Tung). MSc. (Agric.) Prelim. Rpt. Univ. of Western Australia.

Salakpetch, S., D.W. Turner, and B. Dell. 1990. The flowering of carambola (Averrhoa carambola L.) is more strongly influenced by cultivar and water stress than by diurnal temperature variation and photoperiod. Sci. Hort. 43: 83-94.

Salisbury, F.B. and C.W. Ross. 1985. Plant physiology (3rd Ed.). Wadsworth, Belmont, Calif.

Salomon, E. 1984. Phenology of flowering in citrus and avocado and its significance. Acta Hort. 149:53.

Samson, J.A. 1980. Tropical fruit. Longmans, New York, N.Y.

Schaffer, B. 1998. Flooding response and water-use efficiency of subtropical and tropical trees in an environmentally-sensitive wetland. Ann. Bot. 81:475-481.

Schaffer, B. and P.C. Andersen. 1994. Handbook of environmental physiology of fruit crops. Vol. 2: Sub-tropical and tropical crops. CRC Press, Boca Raton, Fla.

Schaffer B, A. W. Whiley, J.H. Crane. 1994. Mango. p. 165-197. In: B. Schaffer and P. C. Andersen (eds.). Handbook of environmental physiology of fruit crops. Vol. 2: Sub-tropical and tropical crops. CRC Press, Boca Raton, Fla.

Shackel, K.A., H. Ahmadi, W. Biasi, R. Buchner, D. Goldhamer, S. Gurusinghe, J. Hasey, D. Kester, B. Krueger, B. Lampinen, G. McGourty, W. Micke, E. Mitcham, B. Olson, K. Pelletrau, H. Philips, D. Ramos, L. Schwankl, S. Sibbett, R. Snyder, S. Southwick, M Stevenson, M Thorpe, S Weinbaum, and J. Yeager. 1997. Plant water status as an index of irrigation need in deciduous fruit trees. HortTechnology 7:23-29.

170

Shackel, K.A., B. Lampinen, S. Southwick, W. Olson, S. Sibbett, W. Krueger, J. Yeager, and D. Goldhamer. 2000. Deficit irrigation in prunes: Maintaining productivity with less water. HortScience 35:1063-1066.

Scholander, P.F., H.T. Hammel, E.D. Bradstreet, and E.A. Hemmingsen. 1965. Sap pressure in vascular plants. Science. 148:339-346.

Schulze, E.D. 1986. Whole-plant responses to drought. Austral. J. Plant Physiol. 13:127-141.

Seeley, S. 1990. Hormonal transduction of environmental stresses. HortScience 25:1369-1376.

Simonneau, T., R. Habibi, J.P. Goutouly, J.G. Huguet, and A. Lecomte. 1993. Diurnal changes in stem diameters and plant water content in peach trees. Acta Hort. 335:191-196.

Smajstrla, A.G. and D.J. Pitts. 1997. Tensiometer service, testing, and calibration. Fla. Coop. Ext. Serv. Bulletin 319. Inst. Food Agr. Sci., Univ. of Fla., Gainesville.

Smajstrla, A.G. and D.S. Harrison. 1998. Tensiometers for soil moisture measurement and irrigation scheduling. Circular 487. Inst. Food Agr. Sci., Univ. of Fla., Gainesville.

Southwick, S.M. and T.L. Davenport. 1986. Characterization of water stress and low temperature effects on flower induction in citrus. Plant Physiol. 81:26-29.

Southwick, S.M. and T.L. Davenport. 1987. Modification of the water stress-induced floral response in ‘Tahiti' lime. J. Amer. Soc. Hort. Sci. 112:231-236.

Sritharan, R. and F. Lenz .1989. The influence of long-term water stress and fruiting on photosynthesis and transpiration in apple. Gartenbauwissenschaft. 54:150-154.

Starr, J.L. and I.C. Paltineanu .1998a. Real-time soil water dynamics over large areas using multisensor capacitance probes and monitoring system. Soil & Tillage Res. 47:43-49.

Starr, J.L. and I.C. Paltineanu. 1998b. Soil water dynamics using multisensor capacitance probes in nontraffic interrows of corn. Soil Sci. Soc. Amer. J. 62:114-122.

Stern, R.A., M. Meron, A. Naor, R. Wallach, B. Bravdo, and S. Gazit. 1998. Effect of fall irrigation level in ‘Mauritius’ and ‘Floridian’ lychee on soil and plant water status, flowering intensity, and yield. J. Amer. Soc. Hort. Sci. 123:150-155.

Tan, S.C. and R.E. Layne .1991. Soil water content and stomatal conductance in a mature peach orchard as influenced by various irrigation regimes. Can. J. Soil Sci. 71:253-258.

171

Tidbury, G.E. 1976. Averrhoa Spp. p. 291-303. In: R.J. Garner, R.J. and S.A. Chandhri (eds.). The propagation of tropical fruit trees. Cmwlth. Bur. Hort. Plantat. Crops, East Malling, UK.

Tomer, M.D. and J.L. Anderson. 1995. Field evaluation of a soil water-capacitance probe in a fine sand. Soil Sci. 159:90-97.

Topp, G.C. and J.L. Davis. 1985. Time-domain reflectometry (TDR) and its application to irrigation scheduling. pp. 107-127. In: D. Hillel (ed.). Advances in irrigation. Vol. 3. Acad. Press, Inc., New York, N.Y.

Topp, G.C., J.L. Davis, and A.P. Annan. 1980. Electromagnetic determination of soil water content: Measurement in coaxial transmission lines. Water Resources Res. 16:574-582.

Torrecillas, A., R. Domingo, R. Galego and M.C. Ruiz-Sanchez. 2000. Apricot tree response to withholding irrigation at different phenological periods. Sci. Hort. 85:201-215.

Turner, N.C. 1979. Drought resistance and adaptation to water deficits in crop plants. p. 343-372. In: H. Mussell and R.C. Staples (eds.). Stress physiology in crop plants. Wiley, New York, N.Y.

Turner, N.C. 1981. Techniques and experimental approaches for the measurement of plant water status. Plant & Soil. 58:339-366.

Turner, N.C. 1986. Adaptation to water deficits: A changing perspective. Austral. J. Plant Physiol. 13:175-190.

Turner, N.C. and M.J. Long. 1980. Errors arising from rapid water loss in the measurement of leaf water potential by the pressure chamber technique. Austral. J. Plant Physiol. 7:527-537.

van Genuchten, M. Th. 1980. A closed-form equation for predicting the hydraulic conductivity of unsaturated soil. Soil Sci. Soc. Amer. J. 44:892-898.

Wagner, C.J., W.L. Bryan, and R.E. Berry. 1975. Carambola selection for commercial production. Proc. Fla. State Hort. Soc. 88:466-469.

Wang, Z. and G.W. Stutte. 1992. The role of carbohydrates in active osmotic adjustment in apple under water stress. J. Amer. Soc. Hort. Sci. 117:816-823.

Watson, B.J., A.P. George, R.J. Nissen, and B.I. Brown. 1988. Carambola: A star on the horizon. Queensland Agr. J. 114:45-51.

Werkhoven, C. 1993. Time domain reflectometry and tensiometers for detecting soil moisture content. Acta Hort. 335:491-496.

172

Whiley, A. 1993. Environmental effects on phenology and physiology of mango: A review. Acta Hort. 341:168-176.

Whiley, A., J. Saranah, B. Cull, and K. Pegg. 1988. Manage avocado tree growth cycles for productivity gains. Queensland Agr. J. 114:29-36.

Wu, K. 1998. Measurement of soil moisture change in spatially heterogeneous weathered soils using a capacitance probe. Hydrol. Processes. 12:135-146.

Zekri, M., and L.R. Parsons. 1999. Determination of field capacity in a Florida sandy soil and drainage time at different depths. HortTechnology. 9:258-261.

Zekri, M., R. Núñez-Elisea, B. Schaffer, S.K. O’Hair, and J.H. Crane. 1999. Multisensor capacitance probes for monitoring soil water dynamics in the oolitic limestone soil of South Florida. Proc. Fla. State Hort. Soc. 112:178-181.

Zimmermann, U. 1978. Physics of turgor- and osmo-regulation. Annu. Rev. Plant Physiol. 29:121-148.

BIOGRAPHICAL SKETCH

Rashid Al-Yahyai was born in Al-Hamra, Oman, on February 8, 1971. He obtained

his high school diploma in 1989 from Majid bin Khamis Secondary School at Al-Hamra,

Oman. After graduation, he was enrolled in the College of Agricultural and Marine

Sciences at Sultan Qaboos University, Oman. He earned his bachelor’s degree in May

1993 with a major in lant cience. Rashid earned a Master of Science degree in

omology from the Department of Horticulture at Cornell University, Ithaca, New York,

in May 1998. In January 2000, he was enrolled in the Department of Horticultural

Sciences at the University of Florida. He conducted his research at the Tropical Research

and Education Center, Homestead, Florida. He will be holding a faculty position at the

Department of Crop Sciences, College of Agricultural and Marine Sciences at Sultan

Qaboos University in Oman.

173

soil water depletion, growth, physiology, and yield of carambola trees in krome soil

Documents