<%BANNER%>

Modeling shallow groundwater table contribution to soil water retention in the unsaturated zone of a calcareous soil of ...

Permanent Link: http://ufdc.ufl.edu/UFE0025060/00001

Material Information

Title: Modeling shallow groundwater table contribution to soil water retention in the unsaturated zone of a calcareous soil of South Florida
Physical Description: 1 online resource (125 p.)
Language: english
Creator: Barquin, Luis
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: calibration, capillarity, dade, diurnal, equilibrium, florida, groundwater, krome, limestone, lychee
Agricultural and Biological Engineering -- Dissertations, Academic -- UF
Genre: Agricultural and Biological Engineering thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Quantifying the relationship between groundwater table level and root zone soil water content in shallow groundwater conditions is important in South Florida due to the potential changes in groundwater level related to the Comprehensive Everglades Restoration Plan (CERP) and the resulting impacts these changes may have on deep rooted agriculture. Thus, soil water dynamics of the capillary fringe of a Krome very gravelly loam soil profile from a lychee grove (Litchi chinensis) were evaluated using data collected from September 2008 to March 2009. Thirty-two EnviroSCAN capacitance sensors distributed at four depths (10, 20, 40 and 60 cm) and pressure transducers located in four adjacent monitoring wells were installed with the goal of characterizing the shallow groundwater contribution to the soil water content in a soil profile composed of two layers: a scarified layer and an underlying layer of limestone bedrock. The analysis was conducted in three main objectives (1) develop a multi sensor capacitance onsite calibration equation for limestone bedrock by comparing readings of scaled frequency to suction readings of tensiometers, (2) determine the existence, location and statistical significance of diurnal peaks in soil water content and groundwater levels and (3) test the validity of the drain to equilibrium hydrostatic assumptions for predicting the soil water content using groundwater level as a reference. The Krome soil scarified layer?s water holding capacity was primarily attributed to the loam fraction as the limestone fraction had very low water retention properties. Limestone water loss was equivalent to 20% volumetric water content compared to 40% in the Krome scarified soil at 1,000 cm of suction. Four regression models were proposed for calibration purposes in spite of measured spatial and depth variability of the soil. Diurnal peaks of soil water content and groundwater level were statistically identified with circular statistics using Rayleigh test. Mean vectors of soil water content at deeper depths were found to be more related with groundwater fluctuations and soil water content at shallower depths was associated with peaks of solar radiation and soil temperature. A hydrostatic model based on the drained to equilibrium principle and van Genucthen?s equation (1980) was able to capture the general and most representative trends of soil water content changes in response to shallow groundwater fluctuations, accuracy of predictions had an overall Nash Sutcliffe (1970) coefficient of efficiency of 0.72. Results are intended to provide agricultural producers with science-based information on the benefits and challenges of the shallow groundwater table in this area.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Luis Barquin.
Thesis: Thesis (M.S.)--University of Florida, 2009.
Local: Adviser: Migliaccio, Kati W.
Local: Co-adviser: Munoz-Carpena, Rafael.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0025060:00001

Permanent Link: http://ufdc.ufl.edu/UFE0025060/00001

Material Information

Title: Modeling shallow groundwater table contribution to soil water retention in the unsaturated zone of a calcareous soil of South Florida
Physical Description: 1 online resource (125 p.)
Language: english
Creator: Barquin, Luis
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: calibration, capillarity, dade, diurnal, equilibrium, florida, groundwater, krome, limestone, lychee
Agricultural and Biological Engineering -- Dissertations, Academic -- UF
Genre: Agricultural and Biological Engineering thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Quantifying the relationship between groundwater table level and root zone soil water content in shallow groundwater conditions is important in South Florida due to the potential changes in groundwater level related to the Comprehensive Everglades Restoration Plan (CERP) and the resulting impacts these changes may have on deep rooted agriculture. Thus, soil water dynamics of the capillary fringe of a Krome very gravelly loam soil profile from a lychee grove (Litchi chinensis) were evaluated using data collected from September 2008 to March 2009. Thirty-two EnviroSCAN capacitance sensors distributed at four depths (10, 20, 40 and 60 cm) and pressure transducers located in four adjacent monitoring wells were installed with the goal of characterizing the shallow groundwater contribution to the soil water content in a soil profile composed of two layers: a scarified layer and an underlying layer of limestone bedrock. The analysis was conducted in three main objectives (1) develop a multi sensor capacitance onsite calibration equation for limestone bedrock by comparing readings of scaled frequency to suction readings of tensiometers, (2) determine the existence, location and statistical significance of diurnal peaks in soil water content and groundwater levels and (3) test the validity of the drain to equilibrium hydrostatic assumptions for predicting the soil water content using groundwater level as a reference. The Krome soil scarified layer?s water holding capacity was primarily attributed to the loam fraction as the limestone fraction had very low water retention properties. Limestone water loss was equivalent to 20% volumetric water content compared to 40% in the Krome scarified soil at 1,000 cm of suction. Four regression models were proposed for calibration purposes in spite of measured spatial and depth variability of the soil. Diurnal peaks of soil water content and groundwater level were statistically identified with circular statistics using Rayleigh test. Mean vectors of soil water content at deeper depths were found to be more related with groundwater fluctuations and soil water content at shallower depths was associated with peaks of solar radiation and soil temperature. A hydrostatic model based on the drained to equilibrium principle and van Genucthen?s equation (1980) was able to capture the general and most representative trends of soil water content changes in response to shallow groundwater fluctuations, accuracy of predictions had an overall Nash Sutcliffe (1970) coefficient of efficiency of 0.72. Results are intended to provide agricultural producers with science-based information on the benefits and challenges of the shallow groundwater table in this area.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Luis Barquin.
Thesis: Thesis (M.S.)--University of Florida, 2009.
Local: Adviser: Migliaccio, Kati W.
Local: Co-adviser: Munoz-Carpena, Rafael.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0025060:00001


This item has the following downloads:


Full Text

PAGE 1

1 MODELING SHALLOW GROUNDWATER TA BLE CONTRIBUTION TO SOIL WATER RETENTION IN THE UNSATURATED ZONE OF A CALCAREOUS SOIL OF SOUTH FLORIDA By LUIS PABLO BARQUIN VALLE A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Luis Pablo Barquin Valle

PAGE 3

3 To Regina and my family

PAGE 4

4 ACKNOWLEDGMENTS I would lik e to thank the University of Florida, the Agricultural and Biological Engineering Department, the College of Agricultural and Life Sciences and the Tropical Research and Education Center for granting the o pportunity of pursuing my graduate studies and research with such a prestigious and recognized institution. Special thanks to Dr. Kati Migliaccio, for her leadership, trust and patience. But specially for conferring me the opportunity of learning an d developing my interest s in water resources research with her team as a graduate research assistant. Special thanks to Dr. Rafael Muoz Carpena for giving me the first chance to be at UF and develop my inte rests in hydrology during the EARTH undergrad internship as well as all his time patience a nd intellectual s upport in great part of this thesis. Special thanks to Dr. Yuncong Li for giving me the opportunity to return to UF-TREC and work with his team before joinin g graduate school. Special thanks to Dr. Bruce Schaffer and Dr. Jonathan Crane for being not on ly great committee members and professors but also for being such great persons and friends. The author would also like to acknowledge Michael Gutierr ez for all the field support and instrumental training. Special thanks to Sikavas Na-Lampang for sharing his programming skills and special thanks to Tina Dispenza, Nicholas Kiggundu, Isaya Kisseka and Bridgette Castro for all their team support. Last but not least I would like to recognize Regina Incer for all her love, patience and support and my family, specially my parents, fo r giving me all the opport unities and education necessary to pursuit many of my dreams and professional goals.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ............................................................................................................... 4LIST OF TABLES ...........................................................................................................................7LIST OF FIGURES .........................................................................................................................9ABSTRACT ...................................................................................................................... .............13 CHAP TER 1 GENERAL INTRODUCTION .............................................................................................. 15Background .................................................................................................................... .........15Shallow Groundwater Capillarity .................................................................................... 15Drained to Equilibrium Concept...................................................................................... 16Study Site Characteristics: S outheast Florida Hydrology ............................................... 18Soils ......................................................................................................................... ........20Tropical Fruit Industry .................................................................................................... 20Lychee (Litchi chinensis ).................................................................................................21Research Objectives ........................................................................................................... .....222 NON DESTRUCTIVE ONSITE CALI BRATION OF MULTISENSOR CAPACITANCE PROBES IN A LI MESTONE BEDROCK PROFILE .............................. 28Introduction .................................................................................................................. ...........28Materials and Methods ...........................................................................................................30Equipment and Data Collection .......................................................................................30Development of Soil Water Characteristic Curves of Krome Soil Scarified and Limestone Bedrock Layer in Laboratory Conditions ..................................................31Conversion of Suction Values from Tens iometers to Volumetric Water Content Using the Laboratory Soil Water Characteristic Curves of Limestone Bedrock ......... 33Onsite Calibration of Capacitance Sensors in Undisturbed Limestone by Comparing Tensiometers Values to Those of Capacitance Sensors ............................ 33Results and Discussion ........................................................................................................ ...34Development of Soil Water Characteristic Curves of Krome Soil Scarified and Limestone Bedrock Layer in Laboratory Conditions ..................................................34Scarified Krome soil ................................................................................................. 34Limestone bedrock ...................................................................................................35Characterizing the Krome soil profile ...................................................................... 36Conversion of Suction Values from Tens iometers to Volumetric Water Content Using the Laboratory Soil Water Characteristic Curves of Limestone ....................... 37Estimating Onsite Calibration of Capacita nce Sensors in Undisturbed Limestone by Comparing Tensiometers Values to Those of Capacitance Sensors ............................ 37Conclusion .................................................................................................................... ..........39

PAGE 6

6 3 USING CIRCULAR STATISTICS TO ID ENTI FY CAPILLARY RISE DIURNAL FLUCTUATIONS IN CA LCAREOUS SOILS ..................................................................... 48Introduction .................................................................................................................. ...........48Materials and Methods ...........................................................................................................51Experimental Site ............................................................................................................ 51Equipment and Data Collection .......................................................................................52Data Extraction and Compilation of Diurnal Peaks ........................................................ 53Circular Statistical Analysis ............................................................................................54Results and Discussion ........................................................................................................ ...57Data Extraction and Compilation of Diurnal Peaks ........................................................ 57Circular Statistical Analysis ............................................................................................58Further Discussion ...........................................................................................................60Capillary Contribution Relative Significance ................................................................. 63Conclusions .............................................................................................................................634 MONITORING GROUNDWATER LEVELS TO PREDICT SOIL WATER CONTENT IN THE UNSATURATED ZO NE OF A CALCAREOUS SOIL ...................... 79Introduction .................................................................................................................. ...........79Materials and Methods ...........................................................................................................82Experimental Site ............................................................................................................ 82Soil Water and Groundwater Monitoring: Identifying Hydrostatic Conditions ..............84Drained to Equilibrium Soil Water Characteristic Curves .............................................. 85Predicting Soil Water Based on Groundwater Observations........................................... 86Results and Discussion ........................................................................................................ ...87Soil Water and Groundwater Monitoring: Identifying Hydrostatic Conditions ..............87Drained to Equilibrium Soil Water Characteristic Curves .............................................. 88Predicting Soil Water Based on Groundwater Level Observations ................................ 91Simplifying results ................................................................................................... 92Applications of the model ........................................................................................ 93The importance of the drained to equilibrium conditions ........................................ 94Conclusions .............................................................................................................................955 EXECUTIVE SUMMARY ..................................................................................................116Objective 1 ............................................................................................................................116Objective 2 ............................................................................................................................117Objective 3 ............................................................................................................................117LIST OF REFERENCES .............................................................................................................118BIOGRAPHICAL SKETCH .......................................................................................................125

PAGE 7

7 LIST OF TABLES Table page 2-1 Soil physical properties and fitted paramete rs of van Genucth ens model (1980) used to describe laboratory soil water characteristics of a scarified Krome soil partitioned into its gravel a nd loam fraction. ....................................................................................... 412-2 Soil physical properties and fitted paramete rs of van Genucthens model (1980) used to describe laboratory soil water characteristics of the underlying limestone bedrock layer of a typical Krome soil profile using th ree rock samples, sieved gravel and data from all samples. ............................................................................................................. ...412-3 Soil physical properties and fitted paramete rs of van Genucthens model (1980) used to describe laboratory soil water characteristic of the limestone and the scarified layer of a Krome soil profile. Al-Yahyai et al. (2006) Krome scarified layer characterization is shown for comparison. .........................................................................422-4 EnviroSCAN onsite calibration equation coefficients for limestone bedrock from sites 1 and 2 at 40 and 60 cm depths. Re gression parameters are result from the relationship between scaled frequency fr om capacitance sensors and suction from tensiometers converted to volumetric water content. ........................................................422-5 Correlation of capacitance se nsor scaled frequencies ca librated using Al-Yahyai et al. (2006) Krome coefficients and onsite calibration coefficients proposed for each site and depth. Dataset includes only volumetric water content in state of equilibrium. .................................................................................................................. ......423-1 Equipment used to identify diur nal peaks using circular statistics. ................................... 653-2 Circular statistical analys is of the occurrence of maximu m daily soil water content in trenched conditions at 10, 20, 40 and 60 cm depths. ......................................................... 663-3 Circular statistical analys is of the occurrence of maximu m daily soil water content in non trenched conditions at 10, 20, 40 and 60 cm depths. .................................................. 663-4 Circular statistical analysis of the occurrence of maximum daily pooled soil water content at 10, 20, 40 and 60 cm depths. ............................................................................. 673-5 Circular statistical analys is of the occurrence of maximu m daily water table elevation at four wells and all data combined. .................................................................................. 673-6 Circular statistical analysis of the occurrence of maximum daily observations of weather factors. ..................................................................................................................683-7 Circular-circular Pearson product moment correlation tests between soil water sensor data and groundwater level as well as weather variables durin g the study period. ........... 69

PAGE 8

8 3-8 Estimated daily contribution of groundwat er diurnal fluctuations to soil water content at f our different depths and soil conditions of the unsaturated zone of a typical Krome soil profile. Soil water contribut ion is the difference of the daily mean and the maximum daily value at each soil condition. ........................................................ 704-1 Monitoring well locations a nd elevation specifications. .................................................... 974-2 Soil water sensors categorized in 16 different soil conditions. .......................................... 974-3 Summary of all volumetric soil water cont ent daily means at each soil condition from 16 September 2008 to 23 March 2009. .............................................................................. 984-4 Summary of entire dataset of groundwat er daily means and comparison with USGS well 196A historical database. ...........................................................................................994-5 Drained to equilibrium fitted parameters of van Genuchtens model (1980) used to describe field soil water characteristic cu rves at 10 cm depth of Krome soil in a lychee orchard at four different soil conditions. .............................................................. 1004-6 Drained to equilibrium fitted parameters of van Genuchtens model (1980) used to describe field soil water characteristic cu rves at 20 cm depth of Krome soil in a lychee orchard at four different soil conditions. .............................................................. 1014-7 Drained to equilibrium fitted parameters of van Genuchtens model (1980) used to describe field soil water characteristic cu rves at 40 cm depth of Krome soil in a lychee orchard at four different soil conditions. .............................................................. 1024-8 Drained to equilibrium fitted parameters of van Genuchtens model (1980) used to describe field soil water characteristic cu rves at 60 cm depth of Krome soil in a lychee orchard at four different soil conditions. .............................................................. 1034-9 Duncan mean multiple comparison of soil wa ter characteristic curves at each soil condition. .................................................................................................................... .....104

PAGE 9

9 LIST OF FIGURES Figure page 1-1 Divisions of subsurface water (Todd, 1959). ..................................................................... 231-2 Southeastern Florida aquifer allocation from the Water Resources Atlas of Florida (Fernald et al., 1998). ....................................................................................................... ..231-3 General view of the Biscayne Bay Watershed, South Florida. .......................................... 241-4 South Florida Water Manage ment District (1976-2005) m onthly average rainfall of Miami-Dade County and US Geological Su rvey (1950-2009) monthly median water table elevation WTE NGVD1929 of Well 196A at UF-TREC Homestead, Florida. (SFWMD, 2009; USGS, 2009). ......................................................................................... 251-5 Krome soil typical profile consisting of a scarified layer of Krome soil (15 cm) and underlying permeable layer of limestone bedrock. ............................................................ 251-6 Classification of the prin cipal soil types and principal agricultural land use in the study area of South Mi ami-Dade County, FL. ................................................................... 261-7 Dimension of trenches for tropical cr ops in Krome soils of Miami-Dade County. ........... 272-1 Schematic diagram of EnviroSCAN and RS U tensiometers installed at each site. Two EnviroSCAN probes with sensors at 20 and 40 cm and five RSU tensiometers installed at 20 cm (2 units) and 40 cm (3 units). ................................................................ 432-2 Laboratory soil water characteristic curves of Krome scarified layer. Samples with the partitioned gravel and loam fraction ar e compared to a sample without sieve and Al-Yahyai et al. (2006) char acterization. The suction and soil water content were determined with pressure Tempe cells and fi tted with van Genuchten (1980) model. ...... 432-3 Laboratory soil water characteristic curves of limestone bedrock layer. A general curve Limestone (all) was fitted from three limestone samples and compared to the sieved gravel from Krome. The suction a nd soil water content were determined with pressure Tempe cells and fitted with van Genuchten (1980) model. ................................. 442-4 EnviroSCAN capacitance sensor onsite calibration of limes tone bedrock from site 1 at 40 cm depth. Scaled frequency ( SF) was determined by EnviroSCAN and volumetric water content ( ) was derived from suction observations of RSU tensiometers. ................................................................................................................. .....44 2-5 EnviroSCAN capacitance sensor on site calibration of limes tone bedrock from site 2 at 40 cm depth. Scaled frequency ( SF) was determined by EnviroSCAN and volumetric water content ( ) was derived from suction observations of RSU tensiometers. ................................................................................................................. .....45

PAGE 10

10 2-6 EnviroSCAN capacitance sensor on site calibration of limes tone bedrock from site 1 at 60 cm depth. Scaled frequency ( SF) was determined by EnviroSCAN and volumetric water content ( ) was derived from suction observations of RSU tensiometers. ................................................................................................................. .....452-7 EnviroSCAN capacitance sensor onsite calibration of limes tone bedrock from site 2 at 60 cm depth. Scaled frequency ( SF) was determined by EnviroSCAN and volumetric water content ( ) derived from suction observa tions of RSU tensiometers. ... 462-8 Comparison site specific EnviroSCAN cap acitance sensor ons ite calibrations of limestone bedrock from sites 1 and 2 at 40 and 60 cm depths. Scaled frequency (SF) was determined by EnviroSCAN and volumetric water content ( ) was derived from suction observations of RSU tensiometers. .......................................................................462-9 Comparison of site specific EnviroSCAN capacitance sensor onsite calibrations of limestone bedrock from sites 1 and 2 at 40 and 60 cm depths with Al-Yahyai et al. (2006) calibration of Krome soil. Limestone scaled frequency (SF) was determined by EnviroSCAN and limestone volumetric water content ( ) derived from suction observations of RSU tensiometers. ....................................................................................473-1 The circular statistics rectangular components of a mean vector and its determination through the construction of the center of gravity from three observations. ......................................................................................................................713-2 Schematic of selected probes and location of monitoring wells in e xperimental site at University of Florida Tropical Research and Education Center (TREC). A symbolized probes located in a trenched Krome soil condition and B symbolizes probes located in a non trenched limestone soil condition. ............................................... 713-3 Daily means for the entire dataset of groundwater level, barometric pressure and overall soil water content at 10, 20, 40 and 60 cm depths. Study period from16 September 2008 to 23 March 2009. ................................................................................... 723-4 Diurnal fluctuations from November 20 to November 22, 2008; Influence of groundwater level (WTE), solar radiation, re lative humidity and soil temperature in the soil water content of pr obe A11-10 (depth: 10 cm). .................................................... 733-5 Diurnal fluctuations from November 20 to November 22, 2008; Influence of groundwater level (WTE), solar radiation, re lative humidity and soil temperature in the soil water content of pr obe A11-20 (depth: 20 cm). .................................................... 733-6 Diurnal fluctuations from November 20 to November 22, 2008; Influence of groundwater level (WTE), solar radiation, re lative humidity and soil temperature in the soil water content of pr obe A11-60 (depth: 60 cm). .................................................... 743-7 Circular histograms and mean vectors of filtered soil water maximum daily values. ....... 75

PAGE 11

11 3-8 Circular histograms and mean vectors of pooled, filtered soil water m aximum daily values at 10, 20, 40 and 60 cm depths. .............................................................................. 763-9 Circular histograms and mean vectors of filtered groundwater level maximum daily values from all wells (WTE), well 1 (W TE-W1), well 2 (WTE-W2), well 3 (WTEW3) and well 4 (WTE-W4). .............................................................................................. 773-10 Circular histograms and mean vectors of filtered weather maximum daily values. .......... 784-1 Soil water potential in an unsaturated soil column at drained to equilibrium conditions. H=hydraulic potential h=pressure potential, z=gravimetric potential, L=water table elevation....................................................................................................1054-2 Side view schematic of soil conditions studied using monitoring wells and multisensor capacitance probes in the Krome scarified (gray) and limestone bedrock (light) soil profile. ............................................................................................................1054-3 Schematic of experimental site at Univ ersity of Florida Tropical Research and Education Center (TREC), Homestead, FL A symbolized probes located in a trench Krome soil condition and B symbo lizes probes located in a non trenched limestone soil condition. .................................................................................................. 1064-4 Schematic of monitoring well geomet ry and Levelogger location for monitoring groundwater levels. ..........................................................................................................1074-5 Soil water response to groundwater fl uctuations from September 2008 to March 2009 in Lychee trench selected sensors at (a) 10 cm, (b) 20 cm, (c) 40 cm and (d) 60 cm depths. .................................................................................................................... ....1084-6 Soil water characteristic curves for probe s A1 to A4 and B1 to B4 at 10, 20, 40 and 60 cm depths. ................................................................................................................. ..1094-7 Soil water characteristic curves for probe s A5 to A8 and B5 to B8 at 10, 20, 40 and 60 cm depths. ................................................................................................................. ..1104-8 Soil water characteristic curves for probes A9 to A12 and B9 to B12 at 10, 20, 40 and 60 cm depths. ............................................................................................................1114-9 Soil water prediction using groundwater level as reference with the fitted parameters of van Genuchten model (1980) for probes A1 to A4 and B1 to B4 at 10, 20, 40 and 60 cm depths. ................................................................................................................. ..1124-10 Soil water prediction using groundwater level as reference with the fitted parameters of van Genuchten model (1980) for probes A5 to A8 and B5 to B8 at 10, 20, 40 and 60 cm depths. ................................................................................................................. ..113

PAGE 12

12 4-11 Soil water prediction using groundwater level as reference with the fitted param eters of van Genuchten model (1980) for probes A9 to A12 and B9 to B12 at 10, 20, 40 and 60 cm depths. ............................................................................................................1144-12 Estimation of water savings based on drained to equilibrium assessment of irrigation requirements. Drained to e quilibrium model assumes fiel d capacity at 100 H2O cm of suction and compares supplemented irriga tion with average grower (10.8 m3/ha). ... 115

PAGE 13

13 Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master in Science MODELING SHALLOW GROUNDWATER TA BLE CONTRIBUTION TO SOIL WATER RETENTION IN THE UNSATURATED ZONE OF A CALCAREOUS SOIL OF SOUTH FLORIDA By Luis Pablo Barquin Valle August 2009 Chair: Kati White Migliaccio Cochair: Rafael Muoz-Carpena Major: Agricultural and Biological Engineering Quantifying the relationship be tween groundwater table level and root zone soil water content in shallow groundwater conditions is important in South Florida due to the potential changes in groundwater level related to the Comp rehensive Everglades Restoration Plan (CERP) and the resulting impacts these changes may have on deep rooted agriculture. Thus, soil water dynamics of the capillary fringe of a Krome very gravelly loam soil profile from a lychee grove ( Litchi chinensis ) were evaluated using data collect ed from September 2008 to March 2009. Thirty-two EnviroSCAN capacitan ce sensors distributed at four depths (10, 20, 40 and 60 cm) and pressure transducers located in four adjacent monitoring wells were installed with the goal of characterizing the shallow groundwat er contribution to the soil wa ter content in a soil profile composed of two layers: a scarified layer and an underlying layer of limestone bedrock. The analysis was conducted in three main objectives (1) develop a multi sensor capacitance onsite calibration equation for limestone bedrock by compar ing readings of scaled frequency to suction readings of tensiometers, (2) determine the existence, location and statistical significance of diurnal peaks in soil water content and groundwater levels and (3) te st the validity of the drain to equilibrium hydrostatic assumptions for predicti ng the soil water content using groundwater

PAGE 14

14 level as a reference. The Krome soil scarifie d layers water holding capacity was primarily attributed to the loam fraction as the limestone fraction had very low water retention properties. Limestone water loss was equivalent to 20% volu metric water content compared to 40% in the Krome scarified soil at 1,000 cm of suction. Four regression models were proposed for calibration purposes in spite of measured spatial and depth variab ility of the soil. Diurnal peaks of soil water content and groundwater level were statistically iden tified with circular statistics using Rayleigh test. Mean vectors of soil water co ntent at deeper depths were found to be more related with groundwater fluctuati ons and soil water content at shallower depths was associated with peaks of solar radiation and soil temperat ure. A hydrostatic model based on the drained to equilibrium principle and van Genucthens equati on (1980) was able to capture the general and most representative trends of soil water content changes in response to shallow groundwater fluctuations, accuracy of pred ictions had an overall Nash Su tcliffe (1970) coefficient of efficiency of 0.72. Results are intended to provide agri cultural producers with science-based information on the benefits and challenges of the shallow groundwater table in this area.

PAGE 15

15 CHAPTER 1 GENERAL INTRODUCTION Background Shallow Groundwater Capillarity Contribution to irrigation from shallow wate r tables has been a subject of extensive research around the world. A brief review of su ch studies is provided in Babajimopoulos et al. (2007). Shallow groundwater tables have been s hown to provide a portion of the crop water needs for cotton replacing 60% of the evapotranspiration (ET) (Wallender et al., 1979) and 37% of ET (Ayars and Schoneman, 1986). Others have s uggested that irrigation could be reduced by 80% under shallow groundwater conditions (Prath apar and Qureshi, 1999). Also, assessment tools and models of water movement from a shallow water table to the root zone have been developed such as Upflow (Raes and Deproost, 2003) and DRAINMOD (Skaggs, 1978a). Diurnal fluctuations of groundw ater have been documented by several authors around the world. Most of the groundwater hyd rology literature suggests diurna l fluctuations are principally induced by three factors: evaporation, barometric pressure and tidal actions Diurnal fluctuations of groundwater caused by daily evaporative cy cles are documented by White (1932), Tromble (1977), Bauer et al. (2004), Me rrit (1996) and Chin et al. ( 2008). Temperature effects on groundwater fluctuations were originally studied by Bouyoucos (1915) and developed by Smith (1939), Taylor (1962) and Meyer (1960). Effects of pressure changes on groundwater levels have also been suggested. Peck (1960) attributed the upward movement of groundwater to the effect of atmospheric pressure on entrapped air in the pore space. Turk (1975) in attempt to explain this phenomenon attributed the fluctuations to te mperature induced atmospheric changes acting upon the capillary zone. This process was also supported by Weeks (1979) who described how changes in barometric pressure generally affect diurnal patterns of groundwater levels. The effect

PAGE 16

16 of pressure on water levels was explained by Rasmussen and Crawford (1997) who described how unconfined aquifers generally experience a unit decrease in groundwater level with each unit increase in barometric pressure and vice versa. Previous anecdotal evidence has been reported in Miami-Dade County, Redland agricultural area of Sout h Florida suggesting that groundwater may be contributing to plant water requirements. Al-Yahyai et al. (2005) reported a lack of physiol ogical response between irrigated and non irrigated carambola (Averrhoa carambola ) in a study conducted in the Redlands agricultural area and assumed th is result was due to shallo w groundwater capillary rise. Additional work characterizing the soil water retent ion in this orchard described the difficulty in obtaining suction values greater than 125 cm in the field tr eatments without irrigation (AlYahyai et al., 2006). Similarly, Migliaccio et al. (2008a) reported evidence of capillary influences on soil water status. Th ey investigated soil suction in Ma rl soils of South Florida, and identified diurnal pattern s during the dry season. Drained to Equilibrium Concept In order to model the sh allo w groundwater contribution to the soil water status in the unsaturated zone, it is necessary to understand the process of capillary rise and find alternatives to interpret its fluctuations. A simple approach to interpret the soil water dynamics is through the drained to equilibrium concept explaine d in detail by Wellings and Bell (1982). The unsaturated zone of an unconfined aquifer is generally categorized into three zones: the soil zone, the intermediate zone and the capillary fringe (Todd, 1959). The soil zone contains roots and is characterized by ra pid water content changes due to rainfall and evapotranspiration. The intermediate zone is below the soil zone and varies in thickness accord ing to the water table depth and is the region where water movement is downwards due to percolation. The capillary fringe is the zone extending from the water table into the intermediate zone for a distance,

PAGE 17

17 limited by the pore size distribution of the soil. This zone is partia lly saturated and holds water at negative pressures. The three zones form a dynami c continuum from the water table to the soil surface (Figure 1-1). Thus, capillary rise is a function of porosity and hydraulic conductivity. When water drains, a capillary moisture distributi on curve is formed. After drainage, there is a relationship between water conten t and negative water pressure which relates to the pore size distribution. The concept of potential energy of soil water with respect to an arbitrary reference datum was first suggested by Buckingham (1907). Above the water table, the soil is unsaturated and the matric potential is negative. In a soil with high permeability and a shallow water table, an equilibrium profile is equivalent to the traditional concept of fi eld capacity. When the profile is at equilibrium with the water tabl e, the matric potential equals the height above the water table and its relationship with soil water content is equivalent to the water content characteristic of the soil because the water potential profile will alwa ys move towards equilibrium with the water table. As a result, the water retention curve can be determined either in laboratory or field conditions. The soil water content and suction relationshi p (and also the soil wa ter content and water table relationship) is unique for a given vados e zone. Depending on the path the soil water is moving (drainage or wetting), the soil water content at a given water level can be represented by more than one pressure (Nielsen, 2006). This phenomenon is called hysteresis and is caused by the variation of the por e sizes in the soil. Characterization of the soil water retenti on functions requires parametric models. A commonly used retention model is the Van Genucht en (1980) closed form analytical expression. The closed form equation consists of four independent parameters which have to be estimated

PAGE 18

18 from observed soil water retention data. Many curve fitting and parameter optimization codes such as RETC software (Van Genuchten et al ., 1991) are widely used today. These types of relationships are empirical in na ture with a physical basis. The van Genuchten (1980) equation is expressed as m n rs rh 1 (1-1) where is the volumetric water content; h is the pressure head; s and r represent the saturated and residual water cont ents, respectively; and n and m are empirical shape parameters. The parameters generally adopt values in the following ranges: 0 < < 1, n > 1, 0 < m < 1, h 0. Study Site Characteristics: Southeast Florida Hydrology The Biscayne Aquifer in southeast Florida is an u nconfined coastal aquifer with a shallow layer of highly permeable limestone that covers an area of 10,360 km2 underlying Broward County, Miami-Dade County, Monroe County and Palm Beach County (Klein and Hull, 1978) (Figure 1-2 and 1-3). Groundwater recharge from annual rainfall of 1,448 mm/yr (Ali and Abtew, 1999) is 46% (USGS, 2009 ) (Figure 1-4); approximatel y 70% of the total direct groundwater recharge (Klein and Hull, 1978) occurs through storm events greater than 15 mm during the rainy season (June to Oc tober) (Delin et al., 2000). The aquifers specific yield is 0.26 m of water table rise for every 1 mm of rain (Chin et al., 2005). Thus, the Biscayne Aquifer characteristics of unconfinement, high permeabili ty and shallow depth result in water table fluctuations that are fairly responsive to stor m events and canal system management (Pitt, 1976; Ritter and Muoz-Carpena, 2006). Weather patterns also influence aquifer water lo sses as the principal process of withdrawal from the Biscayne Aquifer is evaporation which has a close correlation with solar radiation and consequently a corresponding seasonal variatio n (Merrit, 1996). This finding was supported by

PAGE 19

19 Abtew (1996) which indicated that solar radiatio n explained 70% of the variations in South Floridas reference evapotranspiration. Likewise Chin et al. (2008) reported that shallow saturated-zone evaporation rate equaled the potentia l evaporation rate to a depth of 1.4 m decreasing to zero at a depth of 2.5 m. One of the most dominant features of flood control and water management in southeast Florida is the hydraulic connect ion of the Biscayne Aquifer with the South Florida Water Management District (SFWMD) canal system. This relationship brings benefits and challenges. The benefits of the system are that it provides an effective and fast response to flood prevention and a protective freshwater head against saltwate r intrusion. The challenges of such a drainage system are related to the reduced retention time or water storage in the watershed during the wet season that results in potential salt water intrusion during drought periods (Klein and Hull, 1978). A great part of the SFWMD s canal management decisions are integrated into the Comprehensive Everglades Restoration Plan (CERP) which represents an ambitious effort to rehabilitate the Everglades National Park (ENP) quantity and di stribution of water (Perry, 2004). The effort of maintaining sustai nable groundwater levels capable of restoring ENP natural flow, protecting agricultural and urba n areas from flooding and prev enting saltwater intrusion along the coast results in a complex and sometime conflicting challenge for canal management. As management goals are adjusted to meet the CERP objectives, there exists a greater potential for elevated groundwater levels and resulting wa ter movement by capillary forces into the underlying unsaturated soil. The potential impact of increased soil water cont ent is of particular concern to deep rooted crops, such as tropical fruits in respect of fl ooding tolerance and water use efficiency (Schaffer, 1998).

PAGE 20

20 Soils Krom e soils are loamy skeletal, carbonatic, hyp erthermic Lithic Udor ethents (Noble et al., 1996) made of rock plowed oolitic limestone. A Krome soils typical profile is composed of a scarified layer (15 cm) of very gravelly loam texture and an underlying layer of limestone bedrock (Li, 2001) (Figure 1-5). Muoz-Carpena et al. (2002) and Al Yahyai et al. (2006) describe the Krome scarified layer physical struct ure in two solid fractions, 51% coarse particles (>2 mm) and 49% loam particles, with a complex bimodal soil water retention patterns where the gravel fraction contributes to very rapid soil wa ter depletion and the loamy fraction provides for soil water retention. Available information on the properties of the limestone bedrock layer is limited since most soil analytical methods are designed for traditional loose soil and not for solid, bedrock material. The information that is available fo cuses on the hydro-geological characterization of the Biscayne Aquifer (Klein and Hull, 1978; Fish and Stewart, 1991; Chin, 2005; Cunningham et al., 2004; 2006) and its geotechni cal properties (Saxena, 1982). Kl ein and Hull (1978) describe the aquifer as a system of limestone, sandst one and sand, highly permeable and extremely porous. The geotechnical specificatio ns for material engineering cl assify the underlying layer as Miami limestone; the bulk density ranges from 1.31 to 1.64. g/cm3 and porosity is between 20 to 50% or greater (Saxena, 1982). Cunningham (2004) characterized the Biscayne Aquifer as a triple porosity: (1) interparti cle vug porosity providing much of the storage, (2) touching vug porosity creating groundwater fl ow passageways and (3) cond uit porosity cavernous vugs. Tropical Fruit Industry According to the 2007 Census of Agriculture Miam i-Dade County has about 27,134 ha of agricultural land distributed am ong 2,498 agricultural producers where subtropical and tropical fruit production is estimated to cover about 4,600 ha with a crop valued at $35.8 million/yr (2002

PAGE 21

21 crop value; 2007 crop value unavailable). Eighty-eigh t percent of these fru it crops are located in the Redlands of Miami-Dade County (Minkow ski and Schaffer, 2001), an area of 18,080 ha (Figure 1-3 and 1-6). Agricultu ral land use encompasses 14,265 ha of this area representing 36.5% of all agriculture in Miami-Dade Count y. Minkowski and Schaffer (2001) reported that 91% of Miami-Dade Countys fruit groves are located in Krome soils (Figure 1-6). Lychee ( Litchi chinensis ) Lychee (Litchi chinensis ) was introduced in Florida in the 1880s and gained importance as a commercial crop in the 1940s. The two main va rieties in South Flor ida are Brewster and Mauritius. It is a polyaxial sp ecies with synchronous growth pa ttern characterized by alternating root and shoot growth and vege tative reproductive growth separated in time (Crane, 2008). The two environmental factors that mainly influence the potential to flower are ambient temperature and available soil water content. However, the most dominant factor to induce consistent flowering is cool non-fr eezing temperatures. Studies in So uth Florida have concluded that around 390 hrs of non-freezing temperatures below 15.5 C correspond to the best crop yield potential (Crane et al., 2004). Vegetative flush usually occurs between June and October. Root flus h is characterized by two cycles, the first cycle occurs between February and May and the second cycle is simultaneous to the vegetative flush (June to October). Flower bud development begins in November and peaks in January and fruit set occurs between December and March peaking in February. The crops season is from March to Ju ne in the case that anthesis is successfully initiated after the quiescence i nduced by non-freezing temperatures. Critical periods to restrict irrigation in South Florida are fr om October to February to li mit vegetative flush. Generally,

PAGE 22

22 drought stress synchronizes the ve getative shoots dormancy and infl uences on flowering increase as long as the plant has been exposed to cold temperatures. As many of the tropical crops in South Fl orida Krome soils, lychees are planted in trenches. They are typically planted at a density of 286 trees/ha. Crane et al. (1994) indicates that growers have found it beneficial to create these soil filled trench es to about 0.6 m deep in the bedrock beneath the rock plowed surface layer to resist greater winds during the hurricane season and to increase the root developm ent (Figure 1-7). Fruit trees grow n in trenched environments in South Florida develop a root system that is characterized by deep roots in a localized area within the trench and a pancake superfic ial root layer that interconnects (over time) with the rest of the trees planted in the orchard (Crane, 2008). The stress physiology cause d by the root growth restriction provides stimulus to produce more carbon and biomass and therefore a better yield (Schaffer, 2008). Research Objectives Em ploying these concepts to interpret and predict the sh allow groundwater contribution due to capillary movement on soil water content in the unsaturated zone will provide some understand and realistic benefits and challe nges posed by potential changes in groundwater levels considering the hydro-geol ogical conditions in the area. Th is research was focused into three objectives: To develop an onsite capacitance sensor cal ibration equation for limestone bedrock by comparing readings of scaled frequency to soil water suction readings of tensiometers. To determine the existence, location and statistical significance of diurnal peaks in soil water content and groundwater level. To test the validity of the drain to equilib rium hydrostatic assumptions for predicting the soil water content using groundwat er level as a reference.

PAGE 23

23 Figure 1-1. Divisions of s ubsurface water (Todd, 1959). Figure 1-2. Southeastern Florida aquifer allocation from the Wate r Resources Atlas of Florida (Fernald et al., 1998).

PAGE 24

24 Figure 1-3. General view of the Bis cayne Bay Watershed, South Florida.

PAGE 25

25 Figure 1-4. South Florida Water Management District (1976-2005) monthly average rainfall of Miami-Dade County and US Geological Su rvey (1950-2009) monthly median water table elevation WTE NGVD1929 of Well 196A at UF-TREC Homestead, Florida. (SFWMD, 2009; USGS, 2009). Figure 1-5. Krome soil typical pr ofile consisting of a scarified layer of Krome soil (15 cm) and underlying permeable layer of limestone bedrock. 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0 50 100 150 200 250 JANFEBMARAPRMAYJUNJULAUGSEPOCTNOVDECWTE (meters) Rainfall (millimeters) Average Rainfall (1976 2005) Median WTE NGVD 1929 (1950 2009) Bedrock Krome soil

PAGE 26

26 Figure 1-6. Classification of the principal soil types and principal agricultural land use in the study area of South Miami-Dade County, FL.

PAGE 27

27 Figure 1-7. Dimension of trenches for tropical crops in Krome soils of Miami-Dade County.

PAGE 28

28 CHAPTER 2 NON DESTRUCTIVE ONSITE CALIBRATION OF MULTISENSOR CAP ACITANCE PROBES IN A LIMESTONE BEDROCK PROFILE Introduction Soil water content m onitoring based on electrom agnetism generally use the soil medium as a circuit component. This type of method is ba sed on the dielectric constant of the soil and relates the apparent dielectric constant of the soil-air-water mixture and the volumetric water content at different electromagne tic field frequencies. This concept is employed in capacitance sensors, where the dielectric constant of the so il is used to measure the soil volumetric water content considering the soil matrix around the sensor as part of the capacitor system (Gardner et al., 1998). EnviroSCAN multi sensor capacitance probes (S entek Ltd. Pty., Stepney, Australia) are commonly used in soil water content monitoring ; the system is detailed in Paltineanu and Starr (1997). Each capacitance sensor consists of two br ass cylindrical rings that function as electrodes (Fares and Alva, 2000). Enviro SCAN operates using a high frequency range (150 Mhz) which minimizes sensitivity to changes in salinity or temperature. Sensors are capable of measuring volumetric water content values ranging from a saturated soil to almost oven dry soil with a resolution of 0.1% (Buss, 1993). However, the multi sensor capacitance system requires soil specific calibration to produce accurate estimates of soil water content due to large variability in soils. The calibration consists of a relationship between volumetri c water content and scaled frequency of a sensor usually determined by regression analysis using a sample dataset. Depending on the nature of the soil, the manufactu rer recommends two calib ration procedures to obtain the absolute values of volumetric soil water content: field calibration and laboratory calibration. Laboratory calibration pr ocedures require the extraction of samples in containers to

PAGE 29

29 track the scaled frequency readings at specifi c volumetric water contents; Paltineanu and Starr (1997) and Mead et al. (1995) desc ribe in detail this method. Field calibration procedures are site specific calibrations described in detail by Morgan et al. (1999) and Polyakov et al. (2005) and can be destructive and laborious. Accurate fi eld calibrations usually require minimizing the uncertainty of soil bulk density and volumetri c water content sampling using sub sampling techniques. The relationship betw een soil water content and scaled frequency can be developed by collecting a soil sample for a range of scal ed frequency values. Th ese samples would be analyzed for bulk density and water content and then results would lead to development of a site specific calibration curve. South Florida (specifically, Miami-Dade County) is characterized by very unique, complex calcareous soils that are derived from limestone A dominant soil type in Miami-Dade Countys agricultural area is Krome soil. Krome soils are loamy skeletal, carbonatic, hyperthermic Lithic Udorethents (Noble et al., 1996) made of rock pl owed oolitic limestone. A Krome soils typical profile is composed by a scarified layer (15 cm) of very gravelly loam texture and an underlying layer of limestone bedrock (Li, 2001). The av ailable information on the properties of the limestone bedrock layer is focused on the hydro geological characteriza tion of the Biscayne aquifer (Klein and Hull, 1978; Fish and St ewart, 1991; Chin, 2005; Cunningham et al., 2004, 2006) and the materials geotechni cal properties (Saxena, 1982) since most of the soil analysis standard methods of physical properties are diffi cult to be applied and the material is not considered for agronomical functions. Al-Yahya i et al. (2006) develope d a capacitance sensor calibration equation for the scarif ied layer of Krome soil using the laboratory method. However the relationship of volumetric water content and scaled frequency of limestone bedrock has not been calibrated for this equipment. Thus, an additional calibration equation for limestone

PAGE 30

30 bedrock is needed to relate EnviroSCAN measur ements to volumetric water content at depths greater than 15 cm in a Krome soil profile. Determination of calibration equations that predict soil water content based on EnviroSCAN measurements requir e non-traditional methods when the media (i.e., bedrock) is not a loose soil or when it is di fficult to maintain structural ch aracteristics when the media is transported to a laboratory for analysis. For th ese types of circumstan ces, the development of procedures to calibrate sensors without removi ng significant amounts of me dia is needed. Jabro et al. (2005) developed on s ite calibration procedures of EnviroSCAN capacitance sensors by comparing readings of scaled frequencies with soil water content of neutr on probes that had been calibrated using the gravimetric method. The appr oach of capacitance sens or calibration using a different soil water monitoring technology can be applied in limestone conditions. The main goal of this study was to devel op a calibration equation for limestone bedrock by comparing readings of scaled frequency to another soil water monitoring technology. The analysis was conducted in four steps: (1) e quipment installation a nd data collection, (2) development of soil water characteristic curves of Krome soil scarified and limestone bedrock layer in laboratory conditions, (3 ) conversion of suction values fr om tensiometers to volumetric water content using the laboratory soil water charac teristic curves of limestone bedrock and (4) onsite calibration of capacitance sensors in limest one by comparing tensiometers values to those of capacitance sensors. Materials and Methods Equipment and Data Collection The experim ent was conducted at the Universi ty of Florida (UF) Tropical Research and Education Center (TREC), Homestea d, Florida. Data were collected from two sites of limestone bedrock. Data from site 1 was collected from 16 September to 4 December 2008 and from site 2

PAGE 31

31 from 22 January to 23 March 2009. Each site was equipped with two EnviroSCAN multi sensor capacitance probes (Sentek Ltd. Pt y., Stepney, Australia) and five automated recording remote sensing tensiometer units (RSU model, Irromete r Co. Riverside, CA, USA). Sensors were localized at 20 and 40 cm depths on each Envi roSCAN capacitance probe. RSU tensiometers were installed so that porous tips were located at similar depths: 20 cm (2 RSU units) and 40 cm depth (3 RSU units) (Figure 2-1). Equipment wa s programmed to collect data every 15 minutes; data collected was categorized by depth and site and then summarized to daily means to avoid lag response between scaled frequencies and sucti on. It was assumed that suction values at 20 cm of depth and scaled frequency values at 20 cm depth were comparable. Likewise, suction values at 40 cm depth and scaled frequency values at 40 cm were paired. Site 1 and site 2 could not be statistically compared because data collection occurred during different time periods. However, relationships among calibration equations permitted the comparison of sim ilarities in sites and depths. Development of Soil Water Characteristic Cu rves of Krome Soil Scarified and Limestone Bedrock Layer in Laboratory Conditions Soil water characteristic curves for the com p lete Krome soil profile were analyzed with Tempe cells using a laboratory soil water retentio n procedure (Klute, 1986). A total of five core samples were collected to characterize the comple te soil profile. Cores were organized into two categories: the Krome scarified layer and the li mestone bedrock layer. Two core samples from Krome scarified layer were analyzed to compare with the current soil water characteristic curve reported in literature for this soil type (Al-Yahyai, 2006). Three core samples from the limestone bedrock were analyzed as part of the ons ite capacitance calibrati on equation procedure. The two samples of Krome scarified layer we re not used for calibration purposes but to characterize the complete soil pr ofile. One of the samples was co llected as an undisturbed core

PAGE 32

32 and the other sample was separated into two subsamples. Using a 2 mm sieve (mesh #10), one core contained the loam fraction (< 2 mm) and th e other contained the grav el fraction of the soil (>2 mm). The gravel and loam fractions were separated considering previous findings that suggest that coarse particles al so contribute to the water holdi ng capacity (Coile, 1953; Berger, 1976; Hanson and Blevins, 1979). Cores of limestone bedrock were extracted near each experimental si te at 40 and 60 cm of depth using an auger, a hammer and a chisel. The cores were shaped according to the Tempe cells dimensions (10 cm diameter and 10 cm le ngth) using an emery wheel. The dimensions of the cores used in this Tempe cells experiment are greater than the standard cores normally used in this type of analysis. These cores are designed for soils with gravel fractions greater than 2 mm, hence core sizes were greater than typical Tempe cell equipment. Sample preparation included saturation with 0.005 M CaSO4 and Thymol solution for 30 days, securing samples in Tempe cell apparatus a nd draining cells at atmo spheric pressure for 2 days, weighing samples at field capacity and gradually applying incr easing pressure while measuring water volumes draining from each Tempe cell. Ten pressure steps were used (i.e., 15, 25, 35, 55, 105, 255, 505, 755, 930, and 980 cm H2O). Pressure steps did not include the first low pressure phase (<15 cm H2O) were soil water is governed by macro pores and drainage which reflects the first step of soil wa ter depletion due to our interest in hydrostatic or drained to equilibrium conditions. Tempe ce ll laboratory soil water retenti on procedure is explained in detail by Klute (1986). Volumes of water depleted were used to ch aracterize water release curves and fitting parameters were calculated using va n Genuchtens (1980) for the soil water suction range previously described: m n rs rh 1 (2-1)

PAGE 33

33 where is the volumetric water content; h is the applied pressure head; s and r represent the saturated and residual soil water contents, respectively; and n and m are empirical shape parameters. The fitting parameters for all curves were estimated with the RETC program by van Genuchten et al. (1991). For a mo re holistic relationship, the data collected for all limestone cores was used to generate a general limestone release curve representative of the site. Conversion of Suction Values from Tensiomet ers to Volumetric Wa ter Content Using the Laboratory Soil Water Characterist ic Curves of Limestone Bedrock Data from both instruments was summarized to daily means to avoid differences of time response between tensiometers (Towner, 1980) and capacitance sensors (Buss, 1993). Soil water characteristic curves for limestone bedrock were used to relate soil matric potential (suction) values recorded by the RSU tensiometers to vol umetric water content. Using van Genuchtens Equation 2-1 with the respective parameters fo r limestone, daily mean records of volumetric water content from RSU tensiometers were pair ed with simultaneous daily mean records of scaled frequency from the capacitance sensors. Onsite Calibration of Capacitance Sensor s in Undisturbed Limestone by Comparing Tensiometers Values to Those of Capacitance Sensors EnviroSCAN scaled frequency values were nor malized considering the frequencies of air and water. The following equation described by Pa ltineanu and Starr (1997) was used to convert field frequencies into scaled frequency (SF ): WA SAFF FF SF (2-2) where FA is the frequency reading in air, FW is the reading in water and FS are the subsequent readings of the soil water content. The relationship of scaled frequency and volumetric water content was evaluated as: CASFB (2-3)

PAGE 34

34 where is the volumetric water content (independent variable) and A B and C are the regression fitted parameters (Paltineanu and Starr, 1997). Data collected from capacitance sensors and RSU tensiometers was used to determine the fitted parame ters (A, B, C) at each site (site 1 and site 2) and depth (20 and 40 cm) combination. Resulting parameter sets for each site-depth combination were used to predict soil water content based on scaled frequency. BA CSF/1 (2-4) Limestone calibration equations were compared to the calibration equation for Krome soils (AlYahyai et al., 2006) using correlation analysis. Results and Discussion Development of Soil Water Characteristic Cu rves of Krome Soil Scarified and Limestone Bedrock Layer in Laboratory Conditions Scarified Krome soil Soil physical properties and results from the Tempe pressure cell study of gravel and loam subsamples are summarized in Table 2-1. Four soil water characteristic curves shown in Figure 2-2 represent the Krome scarified layer. As depict ed in Figure 2-2, the sieved loam portion of the Krome scarified layer had the greatest water reten tion characteristics. This was expected due to the small particle size of the loam fraction. The gravel fraction of the soil sample in this study constituted 36% of the total vo lume compared to 51% which wa s reported by Al-Yahyai et al. (2006). The effective porosity and bulk density al so differed between Al-Yahyai et al. (2006) and that collected in this study. Soil water characteristic curves in Fi gure 2-2 depict volumetric water content depletion in the first 100 cm of pressure head for the sieved loam, while the gravel remained at relatively constant volumetric water content from 15 to 1,000 cm of pressure. Thus, the drainage of soil water contained in the macro pores of gravel was not captured in the pressure

PAGE 35

35 steps studied. Since the applic ation of study results will be for evaluation of hydrostatic conditions, macro pore gravity drainage charact erization of the limestone is not required. Results for the core soil water characteristic curve of Krome scarified layer suggest its retention properties are between the partitioned lo am and the gravel fractions. The Al-Yahyai et al. (2006) soil water characteristic curve and the Krome scarified soil water characteristic curve from our study were similar. However, some differences were obs erved which could be attributed to the greater loam content of our sa mples (64%) compared to AlYahyai et al. (2006) (51%). Thus, water retention ch aracteristics of our soil allowed for greater volumetric water contents at greater suctions a nd field capacity is primarily de pendent upon the proportion of the soil that is in the loam fraction. Limestone bedrock Physical p roperties and soil water characterist ic curves for the limestone bedrock cores are presented in Figure 2-3. Results indicated that effective porosity of the limestone bedrock was approximately 0.25. The three cores had an average bulk density of 1.4 g/cm3, similar to the range reported by Saxena (1982) of 1.31 and 1.64 g/cm3. The van Genuchten inflection point parameter for limestone was lower than that for the Krome scarified sample ( inf = 0.20 limestone; inf = 0.30 Krome). It is important to consider the excessive volume of water drained between the saturation of thymol and the phase of equilibrium saturated content (zero suction). Macro porosity is not considered in the effectiv e porosity (Table 2-2) a nd it is not included in determining water holding capabilities. As a result, we can conclude that the limest one bedrock has in general very low saturated water content and a characteristic of low volumet ric water content changes to large changes in matric potential (suction) increm ents. This feature was previously described by Muoz-Carpena et al. (2002) for the scarified Krome soil and now is confirmed in the underlying limestone

PAGE 36

36 bedrock as well. The results also indicate that the soil water characteristic of the gravel fraction in Krome scarified layer was similar to the soil water characteristic of the limestone bedrock layer. All the limestone data were pooled to create a general limestone curve (r2 = 0.38) for onsite EnviroSCAN calibration purposes. The low coefficien t of determination suggests that there was heterogeneity in the structure of the limestone profile not only in depth but also in terms of horizontal variability, adding a le vel of complexity and uncertain ty to measuring, predicting and modeling soil water conditi ons in this region. Characterizing the Krome soil profile Physical p roperties and Tempe pressure ce ll results are summarized in Table 2-3. The limestone (all) (r2 = 0.38) and the Krome s carified layer cores (r2 = 0.99) were found to have different soil water retention properties and therefore water ma nagement properties. If we consider the field capacity of a soil as 100 cm of suction, the Krome scarified layers field capacity is approximately 0.27 and at the same suction the volumetric water content of the limestone is 0.22. In addition, at 1,000 cm of pre ssure the Krome soil water content decreases to 60% while a similar pressure change in limestone results in 20% reduction in water content. The Krome soil scarified layer retain s up to 9% and limestone only retains 1.7% of volumetric water content between the field capacity (100 cm of suctio n) and the effective saturated water content. Models for all individual curves (Tables 2-1, 2-2 a nd 2-3) had in general g ood fits with measured data (r2 > 0.9) with the exception of the limestone (all) model (r2 = 0.38). The limestone (all) model is considered an overall fit that incl udes the diverse porositi es found in the study.

PAGE 37

37 Conversion of Suction Values from Tensiomet ers to Volumetric Wa ter Content Using the Laboratory Soil Water Charact eristic Curves of Limestone The soil water characteristic from limestone (all) curve (r2 > 0.38; Table 2-2) was used to convert the measured suction of the RSU tensio meters to volumetric water content. Suction (h) measured by the RSU tensiometers and the fitted parameters (0.00779), n (1.05816), m (0.05496), s (0.2371) and r (0.0) in the Van Genuchtens Equati on (2-1) were used to estimate the volumetric water content regr ession with scaled frequency. Estimating Onsite Calibration of Capacita nce Sensors in Undisturbed Limestone by Comparing Tensiometers Values to Those of Capacitance Sensors Calibration equation s for each site and depth were developed by comparing the capacitance sensors readings (SF) with the converted soil water contents from the RSU tensiometers. Calibration equations are presented in Figures 2-4 to 2-7 and equation coefficients are summarized in Table 2-4. All calibrations were li near regressions (B = 1). Measured scaled frequency from the capacitance sensors and m easured suction from the RSU tensiometers resulted in good fits (r2 > 0.70; Table 2-4). However, when all data was combined a general relationship could not be defined (Figure 2-8). The non-destructive site specific calibration of the scaled frequency values provided a good site and depth specific relationship between the Envi roSCAN observations and the simultaneous RSU suctions (converted to volumetri c water content with th e limestone (all) soil water characteristic curve). Thes e results support the idea of us ing field measurements of other instruments to calibrate sensor readings instead of using more destructive techniques such as field or laboratory calibrations, similar to the methods proposed by Jabro et al. (2005) for calibrating neutron probes. However, the site specific calibration required in limestone conditions makes it unpractical and similar to the field calibration recommended by the manufacturer. Comparison of calib ration results between the new limestone regressions and the

PAGE 38

38 calibration regression used by Al -Yahyai et al. (2006) was eval uated by converting the scaled frequencies from capacitance sensors to volumet ric water content using the new coefficients (Table 2-4) and the typical coefficients fr om Al-Yahyai et al. (2006) (A = 0.011, B = 1, C = 0.5206) at equilibrium conditions (no rainfall or irrigation). Results for the calibration regr ession from site 1 at 40 cm depth from September to December 2008 are plotted in Figure 2-4 (r2 = 0.96). Relationship of the new calibration equation with the Al-Yahyai et al. (2006) calibration e quation at equilibrium conditions was strong (r2 = 0.960). The regression equation for site 2 at 40 cm depth considering data from January to March 2009 (r2 = 0.70) (Figure 2-5) relation with Al-Yahyai et al. (2 006) calibration equation resulted in a lower coefficient of determination of (r2 = 0.51). The deviation between the methods is further observed in storm or ir rigation observations, where the Al-Yahyai et al. (2006) equation showed a greater response to volumetric water content increase. Limestone low response to storm events is due to low wa ter holding capacity expressed in parameters shown in Table 2-3 and Figure 2-3. Regression analysis of scaled frequency and volumetric water content at site 2 (r2 = 0.96) indicated a better fit than site 1 (r2 = 0.74) at the 60 cm depth (Fi gures 2-6 and 2-7). The lower coefficient of determination for Site 1 was due to a parallel scatter attributed to site heterogeneity and instrument uncertainty. Comparisons of the new calibration equations at 60 cm depth with the Al-Yahyai et al. (2006) cal ibration equation (Figures 2-7) are presented in Table 2-5 and indicate there is a relatively similar degree of correlation for site 1 (r2 = 0.82) and site 2 (r2 = 0.83) with Al-Yahyai et al. (2006). If all graphs are combined, similarities and differences of the resulting regression models and respective data for each case are evident (Figure 2-8). Results suggest that horizontal

PAGE 39

39 variability influences volumetric water content mo re than depth. The four site specific limestone regression models are characterized by different slopes. The site specific calibration regressions and the Al-Yahyai et al. (2006) model are presente d in Figure 2-9. It is clear that all of our regression models have a grea ter slope compared to the Al-Yahyai et al. (2006) model. Results indicated that large ch anges in scaled frequencies co rrelate with small changes in volumetric water content. This relationship was derived from the conversion of field suction using the laboratory experiments which observed small changes in volumetric water content with large changes in soil water suction. Finally, de spite the valuable and important relationships found with this technique, in general a sma ll range of volumetric wa ter content (range of < 0.01) was measured (Figures 2-4 to 2-7). A primary difference between the limestone soil water characteristic curve and the curv e presented by Al-Yahyai et al. ( 2006) is their response to water inputs. For purposes of equilibrium conditions, the volumetric water content values seem to be similar. Conclusion The soil water characteristic of a Kro me soil profile was evaluated using pressure Tempe cells. The results of the experiment suggest that most of the wate r holding capacity of the soil is due to the loam fraction. The difference in volumet ric water content of th e Krome scarified layer at effective saturation and volumetric water content at 1,000 cm of suction was 40%. The gravel fraction of Krome scarified soil had a similar soil water characteristic to that of limestone bedrock. The difference in volumetric water co ntent at limestone effective saturation and volumetric water content at 1,000 cm of suction was 20%. EnviroSCAN field calibration usi ng suction from tensiometers as a reference to determine volumetric water content is a site specific tech nique for conditions wher e standard calibration procedures are not possible and soil structure is uniform. The regression equations were different

PAGE 40

40 for all sites and depths but confirmed the dominan ce of spatial variability over depth variability. Thus, four site specific regression models fo r limestone bedrock are proposed. Similarities between the calibration equati on from Al-Yahyai et al. (2006) Krome and the proposed limestone calibration equation were strong dur ing soil water equilibrium conditions while differences were observed during pe riods of rainfall or irrigation.

PAGE 41

41 Table 2-1. Soil physical properties and fitted parameters of van Genucthens model (1980) used to describe laboratory soil water characteristics of a scarified Krome soil partitioned into its gravel a nd loam fraction. Parameter Sieved gravel Sieved loam 0.01473 0.02979 n 1.05522 1.71126 m 0.05233 0.41564 s 0.23264 0.56603 r 0.00000 0.18313 r2 0.970 0.990 Effective Porosity 0.248 0.528 Bulk Density (g/cm3) 2.062 0.907 Table 2-2. Soil physical properties and fitted parameters of van Genucthens model (1980) used to describe laboratory soil water characteristics of the underlying limestone bedrock layer of a typical Krome soil profile using th ree rock samples, sieved gravel and data from all samples. Parameter Limestone(all) Limestone 1 Li mestone 2 Limestone 3 Sieved gravel 0.00779 0.01060 0.00027 0.00744 0.015 n 1.05816 1.09145 1.23822 1.05346 1.055 m 0.05496 0.08379 0.19239 0.05075 0.052 s 0.23712 0.24331 0.28184 0.19149 0.233 r 0.00000 0.00000 0.00000 0.00000 0.000 r2 0.380 0.950 0.970 0.980 0.970 Effective Porosity 0.244 0.253 0.284 0.192 0.248 Bulk Density (g/cm3) 1.524 1.469 1.449 1.117 2.062

PAGE 42

42 Table 2-3. Soil physical properties and fitted parameters of van Genucthens model (1980) used to describe laboratory soil water characteristic of the limestone and the scarified layer of a Krome soil profile. Al-Yahyai et al. (2006) Krome scarified layer characterization is shown for comparison. Parameter Limestone (all)Krome Al-Yahyai (2004) 0.00779 0.024190.09000 n 1.05816 1.765561.46000 m 0.05496 0.433610.31507 s 0.23712 0.365800.47000 r 0.00000 0.194870.10000 r2 0.380 0.990 0.930 Effective Porosity 0.244 0.350 0.470 Bulk Density (g/cm3) 1.524 1.271 1.400 Table 2-4. EnviroSCAN onsite calibration equation co efficients for limestone bedrock from sites 1 and 2 at 40 and 60 cm depths. Regression parameters are result from the relationship between scaled frequency fr om capacitance sensors and suction from tensiometers converted to volumetric water. Site Period Depth A B C r2 1 Sep to Dec 40 3.4857 1 0.00390.96 1 Sep to Dec 60 5.5125 1 -0.40490.74 2 Jan to Mar 40 6.0519 1 -0.63620.70 2 Jan to Mar 60 2.2536 1 0.23430.96 Table 2-5. Correlation of capacitance sensor scaled frequencies calibrated using Al-Yahyai et al. (2006) Krome coefficients and onsite calibra tion coefficients proposed for each site and depth. Dataset includes only volumetric water content in state of equilibrium. Site Period Depth r2 1 Sep to Dec 40 0.96 1 Sep to Dec 60 0.82 2 Jan to Mar 40 0.51 2 Jan to Mar 60 0.83

PAGE 43

43 Figure 2-1. Schematic diagram of EnviroSCAN and RSU tensiometers installed at each site. Two EnviroSCAN probes with sensors at 20 and 40 cm and five RSU tensiometers installed at 20 cm (2 units) and 40 cm (3 units). Pressure head (cm) 0 200 400 600 800 1000 m3/m3) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 Sieved gravel Sieved loam Krome Al-Yahyai (2006) Figure 2-2. Laboratory soil water ch aracteristic curves of Krome scarified layer. Samples with the partitioned gravel and loam fraction ar e compared to a sample without sieve and Al-Yahyai et al. (2006) char acterization. The suction and soil water content were determined with pressure Tempe cells and fitted with van Genuchten (1980) model.

PAGE 44

44 Pressure head (cm) 0 200 400 600 800 1000 m3/m3) 0.00 0.05 0.10 0.15 0.20 0.25 0.30 Limestone 1 Limestone 2 Limestone 3 Sieved gravel Limestone (all) Figure 2-3. Laboratory soil water ch aracteristic curves of limestone bedrock layer. A general curve Limestone (all) was fitted from three limestone samples and compared to the sieved gravel from Krome. The suction a nd soil water content were determined with pressure Tempe cells and fitted with van Genuchten (1980) model. Site 1: September December Depth: 40 cm SF = 0.0039+3.4857 (m3/m3) 0.2020.2040.2060.2080.2100.2120.2140.216 Scaled Frequency (SF) 0.71 0.72 0.73 0.74 0.75 0.76 0.77 Figure 2-4. EnviroSCAN capacitance se nsor onsite calibra tion of limestone bedrock from site 1 at 40 cm depth. Scaled frequency ( SF) was determined by EnviroSCAN and volumetric water content ( ) was derived from suction observations of RSU tensiometers.

PAGE 45

45 Site 2: January March Depth: 40 cm SF = -0.6362+6.0519 (m3/m3) 0.2220.2240.2260.2280.2300.2320.234 Scaled Frequency (SF) 0.71 0.72 0.73 0.74 0.75 0.76 0.77 0.78 0.79 Figure 2-5. EnviroSCAN capacitance se nsor onsite calibra tion of limestone bedrock from site 2 at 40 cm depth. Scaled frequency ( SF) was determined by EnviroSCAN and volumetric water content ( ) was derived from suction observations of RSU tensiometers. Site 1: September December Depth: 60 cm SF = -0.4049+5.5125 (m3/m3) 0.2080.2090.2100.2110.2120.2130.214 Scaled Frequency (SF) 0.745 0.750 0.755 0.760 0.765 0.770 0.775 0.780 Figure 2-6. EnviroSCAN capacitance se nsor onsite calibra tion of limestone bedrock from site 1 at 60 cm depth. Scaled frequency ( SF) was determined by EnviroSCAN and volumetric water content ( ) was derived from suction observations of RSU tensiometers.

PAGE 46

46 Site 2: January March Depth: 60 cm SF = 0.2343+2.2536 (m3/m3) 0.2250.2260.2270.2280.2290.230 Scaled Frequency (SF) 0.742 0.744 0.746 0.748 0.750 0.752 0.754 Figure 2-7. EnviroSCAN capacitance se nsor onsite calibra tion of limestone bedrock from site 2 at 60 cm depth. Scaled frequency ( SF) was determined by EnviroSCAN and volumetric water content ( ) derived from suction observa tions of RSU tensiometers. (m3/m3) 0.2000.2050.2100.2150.2200.2250.2300.235 Scaled Frequency (SF) 0.71 0.72 0.73 0.74 0.75 0.76 0.77 0.78 0.79 Site 1 (40 cm) Site 2 (40 cm) Site 1 (60cm) Site 2 (60 cm) Calibration Regression Figure 2-8. Comparison site speci fic EnviroSCAN capacitance sens or onsite calibrations of limestone bedrock from sites 1 and 2 at 40 and 60 cm depths. Scaled frequency (SF) was determined by EnviroSCAN and volumetric water content ( ) was derived from suction observations of RSU tensiometers.

PAGE 47

47 (m3/m3) 0.20 0.21 0.22 0.23 Scaled Frequency (SF) 0.70 0.72 0.74 0.76 0.78 0.80 Site 1 (40 cm) Site 2 (40 cm) Site 1 (60cm) Site 2 (60 cm) Calibration Regressions Krome Calibration Al-Yahyai (2006) Figure 2-9. Comparison of site specific Enviro SCAN capacitance sensor onsite calibrations of limestone bedrock from sites 1 and 2 at 40 and 60 cm depths with Al-Yahyai et al. (2006) calibration of Krome soil. Limestone scaled frequency (SF) was determined by EnviroSCAN and limestone volumetric water content ( ) derived from suction observations of RSU tensiometers.

PAGE 48

48 CHAPTER 3 USING CIRCULAR STATISTICS TO ID ENTI FY CAPILLARY RISE DIURNAL FLUCTUATIONS IN CALCAREOUS SOILS Introduction Southeast F lorida Biscayne aquifer is an unconf ined coastal aquifer with a shallow layer of highly permeable limestone that covers an area of 10,360 km2 including portions of Broward County, Miami-Dade County, Monroe County and Palm Beach County (Fish and Stewart, 1991). Biscayne aquifer characteristics of unconfinement, high permeability and shallow depth result in water table fluctuations that are fairly responsiv e to storm events and canal system management (Pitt, 1976; Ritter and Muoz-Carpena, 2006). Due to its shallow nature, the groundwater is subject to upwar d movement due to capillary forces into the unsaturated zone of the soil prof ile. The potential contribu tion of groundwater to unsaturated soil water content increases as distan ce to groundwater level decreases (Wellings and Bell, 1982; Raes and Deproost, 2003). As water management goals are adjusted to meet the Comprehensive Everglades Restoration Plan (CER P) objectives, there exists a greater chance of elevated groundwater levels and resulting water movement by capillary forces into overlying unsaturated soil. Given this scenario, char acterizing groundwater fluctuations provide information that could be used to evaluate capillary rise in the unsaturated zone. Anecdotal evidence combined with previous research findings has documented the effect of capillary rise in the agricu ltural area of South Miami-Dade County. Al-Yahyai et al. (2005) reported a lack of physiologi cal response between irrigate d and non irrigated carambola ( Averrhoa carambola ) grown in Miami-Dade County Krom e soil and assumed this result was due to shallow groundwater capi llary rise. Additional work ch aracterizing the soil water retention in this orchard descri bed the difficulty in obtaining suc tion values greater than 125 cm in the field treatments without ir rigation (Al-Yahyai et al., 2006) Similarly, Migliaccio et al.

PAGE 49

49 (2008) reported evidence of capillary influences on soil water status. They investigated soil suction in Marl soils of South Florida and id entified diurnal patterns during the dry season. Diurnal fluctuations of groundw ater have been documented by several authors around the world. Most of the groundwater hyd rology literature suggests diurna l fluctuations are principally induced by three factors: evaporation, ba rometric pressure and tidal actions. Evaporation of the water table is a negative recharge and thus an important process to consider during extended periods of no rainfall Diurnal fluctuations of groundwater caused by daily evaporative cycles were documented by White (1932), Tromble (1977) and Bauer et al. (2004). Merrit (1996) recognized the principal process of water flux from the Biscayne aquifer as evaporation which has a close co rrelation with solar radiation and consequently a corresponding seasonal variation. This findi ng was supported by Abtew (1996) confirming solar radiation explains 70% of the variations in South Floridas reference evapotranspiration. Groundwater fluctuations may also be a due to temperature gradie nts where soil water moves from warmer to cooler profiles in res ponse to vapor pressure differences. This was originally suggested by Bouyoucos (1915) and developed by Smith (1939) and Taylor (1962). Meyer (1960) also found diurnal fluctuations of shallow gro undwater accompanied by capillary rise of soil water induced by temperature ch anges. South Floridas distribution of mean temperature is strongly related to solar radiati on distribution and, in c onsequence, temperature fluctuations are associated to the evaporative pro cesses (Abtew, 1996). Pressure change impacts on groundwater leve ls have also been suggested by several authors. For example Peck (1960) attributed the upward movement of groundwater to the effect of atmospheric pressure on entrapped air in the pore space. Turk (1975), in attempt to explain this phenomenon, attributed the fluctuations to temperature induced atmo spheric changes acting

PAGE 50

50 upon the capillary zone. This process was s upported by Weeks (1979) who described how changes in barometric pressure generally aff ect diurnal patterns of groundwater levels. The influence pressure has on water levels was al so explained by Rasmusse n and Crawford (1997) which described how unconfined aquifers generall y experience a unit decrease in water level with unit increase in barometric pressure and vice versa. Groundwater level can also be influenced by tidal changes in co astal areas (Todd, 1959). Todd (1959) also explained how mo nitoring wells installed near ti dally influenced water bodies could experience periodic fluctuations in water levels. This type of signal is not only observed in sites close to the co ast but also in locations like Southeast Florida, where groundwater levels are responsive to canal level change s (Ritter and Munoz-Carpena, 2006). As a result, several factors are involved in water table daily recharge and decline. The fluxes of water from the groundwater into the uns aturated zone of the local soil profiles are currently uncharacterized and the impact of weather parameters on this process is not understood specifically in Southeast Florida Krome soils. Characterization of groundwater level fluctuations can be used to assist with interpreting fluctuations in soil water content, particularly in regions with unconfined shallow aquifers (Bouyoucos, 1915; Migliaccio et al., 2008) during hydrostatic conditions (i.e., absence of rainfall or irrigation). The relationship between groundwater level and soil water content during hydrostatic conditions is governed predominantly by capillary forces. However, other factors (previously discussed) may also pl ay a role in this relationship. Determination of peak occurrence can be used to explain soil water content changes due to daily groundwater fluctuations or to identify other factor s that are influenci ng soil water content. Thus, a useful approach is needed to evaluate diurnal cycles, such as circular statistics. Circular

PAGE 51

51 statistics are methods used to in terpret the frequency of occurrence for each daily peak identified in a time series. This method relates the time of the day to a location on the circumference of a circle and data points distributed on a circle are analyzed as dir ectional data. The mean direction or mean vector of a circle ( ) is a measure based on trigonometry using a center of gravity, C (Figure 3-1). In this example, three equal ma sses are represented as vertices of a triangle (representing peak frequencies in a dataset) and the direction of this center of gravity to the origin O is the mean direction of the dataset or un it vector. Circular statistical analyses are described in detail by Batschel et (1981). To the authors knowledge, this method has not been previously published in l iterature to evaluate capillary rise. Howe ver, circular statistics have been used to evaluate seasonality and timing variab ility of flooding (Magilligan and Graber, 1996; Black and Werritty, 1997), meteorol ogy (Hassan et al., 2009) and pr incipally biology and animal behavior (Mennill and Ratcli ffe, 2004; Oliveira et al., 1998; Paton et al., 2003; Peach, 2003). The goal of this study was to determine the ex istence, location and statistical significance of diurnal peaks in soil water content of Krome soil and groundwater levels Circular statistics were used to answer these questions and to fi nd a relationship among the di rections found in soil water content and groundwater level with weather variables. The specific objectives were to: (1) extract and compile diurnal peaks of soil water contents, groundwater levels and weather variables, (2) perform circular statistical analysis and (3) interpreta tion of mean vectors. Materials and Methods Experimental Site Diurnal fluctuations of soil water contents, gr oundwater levels and weather variables were collected in Hom estead South Mi ami-Dade County, Florida, at th e University of Florida (UF) Tropical Research and Edu cation Center (TREC) (80.5 W, 25.5 N). The site is located 16 km west of Biscayne Bay and 250 m south of th e South Florida Water Management District

PAGE 52

52 (SFWMD) Mowry Canal C-103. The mean eleva tion is 3.12 m Nationa l Geodetic Vertical Datum (NGVD) 1929 and is identified by a subt ropical, marine climate. The study was conducted in a lychee ( Litchi chinensis ) orchard. The dominant soil type in the area is Krome soil; classified as loamy skeletal, carbonatic, hyperthermic Lithic Udorethents (Noble et al., 1996) made of rock plowed oolitic limestone. The complete soil profile is composed of a scarified layer (15cm) of very gravelly loam texture and an underlying layer of limestone bedrock. The lychee orchard (as most of the tropical fruit crops in the area) is planted in rock plowed trenches 0.40 m wide and 0.60 m deep filled with scarified Krome soil. The experimental site (as seen in Figure 3-2) cons ists of six rows: two rows with lychee trees planted in trenches (hence after referred as lychee trench), one ro w without lychee trees but with trenched soil (hence after referred as trench), two rows with out lychee trees but with surficial roots and no trenched soil (hence after referr ed as lychee no trench) and one row without lychee trees roots and without trenches (hence after referred as no trench). Each ly chee row consists of 12 trees spaced at 4.6 m within row and 7.6 m between rows with a planting density of 286 trees/ha. Data collected from this site represent diurnal fluctuat ions of a typical tropical fruit orchard in this region. Equipment and Data Collection To track soil water content, three data loggers recorded m easurements from 24 EnviroSCAN multi sensor capacitance probes (Sente k Ltd. Pty., Stepney, Australia) distributed in the six rows previously described. Each row had four probes installed within 13.8 m and each probe had four sensors positioned at 10, 20, 40 and 60 cm depths (Figure 3-2). The installed system of 96 sensors measured soil water conten t as a function of the apparent bulk dielectric constant of the soil, imposed frequency and elect rode configuration, as described in detail by Paltineanu and Starr (1997). Sensors were progr ammed to record volumetric water content

PAGE 53

53 values every 15 min at a resolution of 0.1% (Buss, 1993; Fares and Alva, 2000). A sensor normalization procedure was used to determine th e scaled frequency (SF). A regression equation for the Krome soil, previously developed by Al-Ya hyai et al. (2006), was used to calibrate scaled frequency ( SF ) to volumetric water content ( ). Considering that the 24 probes were installed in the experimental plot for a systematic sampling, six probes were randomly selected to represent 24 of the 96 sensors (Figure 3-2). Thus, six selected probes were used to capture the daily peaks of soil water content per depth. Probe sites were named so that locations in trenches were assigned as A and locations ou tside trenches were assigned as B. Sensor identification included the probe id and insta llation depth (e.g., A11-20 corresponds to a sensor at probe A11 located at 20 cm depth). Groundwater level was measured using Le velogger LT 3001 (Solinst Ltd., Ontario, Canada) pressure transducers. The Leveloggers were suspended and submerged at 3.6 m depth (accuracy = 0.1%, -10 C to 40 C) in four monitoring wells lo cated along the perimeter of the study site (Figure 3-2). Groundwater levels were compensated using data collected by Barologger air barometric pressu re. Groundwater levels were also adjusted and validated based on weekly manual water table depth readings from each monitoring well. Barometric pressure records from the Barologger were included in the da ta analysis as part of the weather variables. Additionally, weather records (i.e., solar radiation, relativ e humidity, temperature) from the Florida Automated Weather Network (FAWN) stat ion located at UF TREC within 100 m of the site, were used. Data Extraction and Compilation of Diurnal Peaks Data were collected every 15 m in for soil wa ter content at 10, 20, 40 and 60 cm depths, water table elevation, barometric pressure, rela tive humidity, solar radiation and soil temperature (Table 3-1). The dataset for each instrument contained 18,144 observations corresponding to189

PAGE 54

54 days from 16 September 2008 to 23 March 2009. Th e dataset collected wa s filtered to include only data points that were representative of equilibrium conditions so that potential capillary influences could be evaluated. Thus, data poi nts associated with storm events and freeze protection were removed. The filtering criter ia reduced the representative data to 12,480 observations per variable corresponding to 130 da ys. Each day was summarized according to the time its maximum observed daily value (or peak) occurred. For each daily peak identified in the time series, the corresponding time of day was transl ated into a location on the circumference of a circle (e.g., 360 corresponds to 0:00 hr). The procedure was repeated for each variable (soil water content, groundwater level, barometric pre ssure, relative humidity, solar radiation and soil temperature). Data from the 24 soil water sensors were also combined to obtain a mean vector according to the installed depth (i.e., 10, 20, 40 and 60 cm depths). Water table elevation observations from the four wells were also me rged to obtain a mean vector of water table elevation. All circular data analyses were performed using the ORIANA software (Kovach, 2009). Circular Statistical Analysis Frequencies of occurrence of m aximum daily va lues during the day were interpreted with circular statistics. Temporal m easurements were converted to angular measurements to calculate the mean vector length ( r ). 2 1 2 1sin cos 1 n i i n i in r (3-1) where r is the mean vector length, n is the sample size and are the angular measurements. When events were distributed uniformly during th e entire period (i.e., all day) the results for r were close in value to 0. When a significant co ncentration of events was located at a specific time during the day, values of r were closer to 1. The mean vector length ( r ) is a useful measure

PAGE 55

55 of data concentration but an additional test was needed to co nfirm if the mean vector length significantly exceeded the length of a uniform distribution. The Rayleigh test of significance of the mean vectors was used to test the null hypothesis of uniform distribution (no mean direction) for e ach of the six variables. This method tested the existence of a significant mean angle or mean direction and consequently the location of a consistent peak during the day. For each rejected null hypothesis, a diurnal peak was statistically confirmed and identified. The Rayleigh test of uniform distribution is calculated through the z value of the population: z=nr2 (3-2) Thus, the greater the mean vector, the larger the value of z and the greater the concentration of the observations will be around the m ean. If the critical level of P from the Rayleigh test of randomness table is less than the assigned al pha level, the null hypothesis is rejected. Significance occurs when z z ( ). The statistical significance of r was first studied by Lord Rayleigh in 1880 and its statistica l applications were developed in detail by Beran (1969) and Greenwood and Durand (1955). One assumption of the ci rcular parametric test (for purposes of comparison) is that the data is random with a von Mises distribution (unimodal). To test the validity of such an assumption w ith our dataset, the Watsons U test was used to test the null hypothesis of a theoretical unimodal distribution for each dataset. Frequencies of daily peak da ta were summarized in circ ular histograms showing the number of cumulative maximum values observed at each hour of the day during the study period. The mean vector and vector lengt h were determined according to the frequency distributions and concentration of observations. Thus, length of the mean vector at each histogram was illustrated with an arrow symbol and concentrations of fre quencies were represented by bars projected from

PAGE 56

56 the circles center. For each vari able, the following circular statistic parameters were calculated: mean vector, length of mean vector, median, ci rcular variance, circular standard deviation, standard error mean, 95% confidence interval, Watson U test for theoretical distribution and Rayleigh test for uniform distribution. Significance of a presumed bivariate depe ndence (e.g. between groundw ater level and soil water content) was tested with correlation analysis. Usually this procedure is completed when the angles of both variables are uniformly distribut ed on the circle and the difference between the observed angles of the two variables (Equation 3-3) can be used to fit Equation 3-4 and determine the mean vector length r A high value of r indicates a strong positive correlation and a low value indicates that the angles from the tw o variables are weakly correlated. The values of r range from 0 to 1 and can be used as a correlation coefficient. iii (3-3) 2 1 2 1sin cos 1 n i i n i in r (3-4) where is the difference of the angles and from two different variables. The is then used in Equation 3-2 instead of in order to get the vector length of the two variables. For datasets that violated the assumption of uniform distri butions, the circular-cir cular correlation method (explained in detail by Batschelet [1981]) was used to calculate the stre ngth of the relationship between two circular variables. Circular-circular correlation us es the Fisher and Lee (1983) method, analogous to the Pearson product moment correlation normally used in linear data analysis. This coefficient ranges from -1 to +1, according to a perfect negative or positive correlation and tests the null hypot hesis that correlation between two variables is 0. Circularcircular correlation analysis te sts the hypothesis that there is no correlation between two circular variables. The main objective of completing this statistical analysis was to determine the possible

PAGE 57

57 statistical relationships between weather and grou ndwater level and soil wa ter content variables. This statistic evaluation is intended to help identify the main factors influencing the capillary effect or diurnal peaks identified. Results and Discussion Data Extraction and Compilation of Diurnal Peaks A tim e series plot of soil water content, groundwater level and barometric pressure is depicted in Figure 3-3. Some genera l features of the datasets can be observed in this figure. One pattern that was identified is the decline in groundwater table level, observed from October 2008 to March 2009. This decline has been previ ously documented and is characteristic of groundwater levels in southeast Florida dur ing the dry season (K lein and Hull, 1978). Groundwater level data from wells 2 and 4 also ca ptured the influence of barometric pressure changes. This relationship was more evident in da ta from wells 2 and 4 than that from wells 1 and 3. This can be contributed to instrumentatio n, as equipment used to measure groundwater levels from wells 2 and 4 were more sensitive to temperature and pressu re changes due to the greater range of barometric compensation ( 5000 cm) compared to wells 1 and 3 (500 cm). Therefore compensated groundwater levels were smoother at smaller ranges of barometric compensation (Figure 3-3). Plots considering a greatly reduc ed time scale were created (F igures 3-4, 3-5 and 3-6) to identify patterns that might be less evident at the greater time scale. The period plotted for the reduced time scale represents a randomly selected pair of days from the study period (21 November to 22 November 2008). Each figure incl udes six datasets: soil water content, water table elevation (WTE), solar ra diation, relative humidity, soil temperature and barometric pressure. Soil water contents plotted in these fi gures were from probe A11 (trenched no lychee). Peaks of soil water content at 10 and 20 cm dept hs coincided with peaks of soil temperature and

PAGE 58

58 solar radiation and minimum values of WTE a nd barometric pressure (Figure 3-4 and 3-5). Alternatively, the minimum values in soil water c ontent at 60 cm depth coincided with peaks of solar radiation and soil temperature and minimu m values of WTE and barometric pressure (Figure 3-6). The relationship between barometric pressure and groundwater level was anticipated as it has previously been observed (Peck, 1960). Similarly, it was expected that solar radiation and soil temperature would follow a similar pattern. While Figure 3.3 includes the enti re period of record, values corresponding to storm events or freeze protection were discarded using a filter ing process. The remaining, filtered dataset was further evaluated to identify prevalent patterns. One pattern identified was that ranges of soil water fluctuations were less at greater depths For example, ranges of soil water content for probe B8 (lychee no trench) at 10, 20, 40 and 60 cm depths were 0.002, 0.003, 0.0006, and 0.0004, respectively. Similarly, for probe A6 (lychee trench) shallower depth (10 and 20 cm) ranges of soil water content (0.012 and 0.02) were greater than soil wa ter content (0.006 and 0.0025) at deeper points ( 40 and 60 cm depths). Circular Statistical Analysis Results f rom circular statistical analysis are summarized in Tables 3-2 to 3-6 and simplified using circular histograms in Figures 3-7 to 3-10. All mean vectors had statistical significance based on the Rayleigh test ( = 0.05). The significant differe nces of mean directions found in all the variables using Rayleigh test is attributed in great part to the number of observations. For datasets that were characterized by multimodal distributions, vector lengths decreased and the von Mises di stribution hypothesis was reject ed in all cases (P<0.05). Circular histograms of water table elevation (Figure 3-9, Table 3-5) suggest a different mean vector for maximum daily values found at we lls 1 and 3 compared to wells 2 and 4. In this case the differences in water leve l ranges for each pair of sensors were influenced by the level of

PAGE 59

59 sensitivity to temperature and atmospheric pres sure fluctuations (as previously discussed). Results considering a pooled data set of water table elevation (WTE) data indicated a significant mean vector and median value of 5:00 am (r = 0.61). The significant mean vectors for wells 1 and 3 were consistently observed at 4:11 am (r = 0.76) and 4:09 am (r = 0.79), respectively. Alternatively, the mean vectors for well 2 (7:03 am) and well 4 (7:31 am) had lower mean vector lengths (r = 0.36 and 0.34, respectively) due to greater variances and standard deviations in the dataset (see Figure 3-3). Thus, gr oundwater levels whose range of pressure compensation is too high tend to have greater standa rd deviations due to a poor co mpensation of barometric and temperature fluctuations. The weather variables were more consistent in the frequency concentrations compared to WTE and soil water content result s. Figure 3-10 and Table 3-6 mean vector lengths were greater than 0.81 except for relative humidity (r = 0.57) where the dataset was characterized by a bimodal distribution. The greatest frequency for relative humidity occurred at 7:00 am. This frequency location is likely associated with th e cooler temperatures observed during morning hours and is a typical weather response. The second greatest frequenc y in relative humidity occurred at 11:00 am. One possibl e explanation for this is that ponded water (residual from storm events or irrigation where days did not meet removal criteria during data set filtering process), evaporation, and increasing temperature combine to form optimum conditions for a mid-morning peak in relative humidity. However, this was not confirmed and further inve stigation is needed to identify true causes of the observed bimodal di stribution. The mean vectors for barometric pressure, solar radiation and soil temperature were found at 11:14 am (r = 0.81), 12:42 pm (r = 0.97) and 3:58 pm (r = 0.98), respectively.

PAGE 60

60 Circular histograms of soil wa ter content (Figure 3-7 and Figure 3-8) illustrate the variability in peak values that were observed. This variability translat ed into reduced mean vector lengths (r values) for mean vectors desc ribing soil water content peaks (Table 3-2, Table 3-3 and Table 3-4). For example, mean vectors of soil water contents at 10 and 20 cm depths (Figure 3-8 and Table 3-4) were located at 4:10 pm (r = 0.46) and 11:16 pm (r = 0.46), respectively. Mean vectors for soil water content datasets at 40 and 60 cm depths were at 3:23 am (r = 0.39) and 7:23 am (r = 0.31), respectively. The 40 and 60 cm soil water content depths were found to be similar to mean vectors found in groundwater levels of well 2 (7:03 am) and well 4 (7:31 am). This corresponds with general observations previously discussed for Figures 34 to 3-6. Given the rejected null hypotheses of Rayleigh test in all variables (Tables 3-2 to 3-6) circular-circular correlation analysis was used to determine the strength of the relationship between soil water content and groundwater level, barometric pressure, soil temperature, relative humidity and solar radiation. Fo r all variable combin ations except soil wa ter probe A11-10 and barometric pressure (r = -0.000105), the null hypot hesis of zero correl ation was rejected. Although the probability of zero relationship was rejected, most of the r coefficients were low (r < 0.5) and therefore the relations hip between soil water content, groundwater level and the rest of the weather variables were not strong. Thus correlation analysis confirmed relationships among all variables but th e results did not provided a good level of interpretation due to the weak relationships. Further Discussion The circular m ethods applied captured significant patterns in all variables of diurnal fluctuations expressed as mean vectors. Visual ex amination of the variables mean vector results provided some interesting concepts to consider.

PAGE 61

61 Soil water content in the shallow layer of the soil profile (10 cm) seemed to be more influenced by weather than soil water content at the 40 and 60 cm depths. This concept is based on the fact that the soil water content mean vect or for 10 cm depth was at 4:10 pm. This peak value corresponded to mean vectors for solar ra diation (12:42 pm) and soil temperature (3:58 pm). (It is expected that sola r radiation would peak prior to soil temperature peak as soil temperature is a function of heat transfer primar ily driven by solar radiation.) Additionally, these peaks coincided with minimum values for groundw ater level. These combined events could be explained by the following phenomenon: groundwate r evaporation, transport of water vapor up through the soil profile, and subsequent c ondensation (at shallower soil depths) when temperatures start to decrease. Groundwater evap oration of the shallow aquifer is supported by results in Chin et al. (2008) that indicated the shallow saturated zone evap oration rate equals the potential evaporation rate to a depth of 1.4 m decreasing to zero at a depth of 2.5 m. While the 10 cm depth peak mean vector may be explained (in some part) by this phenomenon, the 20 cm depth peak vector values were not directly correlated with othe r measured variables. The 20 cm depth peak mean vector for soil water content, wh ich occurred at 11:16 pm, could potentially be explained by other physical phenomenon. For exam ple, soil water movement has been observed due to temperature gradients where soil water mo ves from warmer to cooler and wet to dry profiles (Bouyoucos, 1915). Thus, the later peak at 20 cm could possibly be due to water movement from 10 cm to 20 cm depths due to these gradients and ther efore the 20 cm depth peak vector for soil water conten t would occur after the peak v ector for the 10 cm depth, which was observed. While soil water content mean vectors for sha llower depths could poten tially be linked to weather factors, results for deeper depths suggest ed a different relationship. Our results indicated

PAGE 62

62 that soil water mean vectors at depths of 40 (3:23 am) and 60 cm (7:23 am) were similar to groundwater mean vectors (WTE-W1 = 4: 11 am; WTE-W2 = 7:03 am). Although the multimodal response in most of the soil water content and groundwater level data analyzed suggests that more than one variable is im portant in describing soil water content and groundwater level fluctuations, pr obes located in lychee trenches at 40 and 60 cm depth (A1-40, A1-60, A6-40 and A6-60) captured mean vectors with unimodal distributions with consistent peaks of soil water content similar to groundwater peak mean vectors (Table 3-2 and Table 3-5). The results suggest a tree effect which isolates the groundwater eff ect from other variables. Also, mean vectors for trenches with lychee trees had greater vector lengths (r > 0.5) than mean vectors found at trenches w ithout lychee trees (r < 0.5). To understand diurnal cycles in soil water content, it is esse ntial to characterize diurnal fluctuations in groundwater level. Groundwater levels in our study were influenced by different factors as evident by the multimodal distributions (Figure 3-9). Primary recharge for the aquifer is from precipitation during the ra iny season (June to October), as approximately 70% of the total direct groundwater recharge (Klein and Hull, 1978) is from storms greater than 15 mm (Delin et al., 2000). The time series data collected during this study (dry season) captured a declining trend in groundwater levels. Mean vectors and frequencies observed in early morning hours could be associated with rates of depletion greater than diurnal peaks (i.e., Figure 3-7, probe B8-40, B3-40 and B9-40). Another possibility is that groundwater peaks vect ors occurring during this time might be related to an inverse effect of the m ean vectors found in barometric pressure (11:14 am). As barometric pressure reaches minimum values, a reduction of entrapped air in the unsaturated zone as described by Turk (1975) and Peck (1960) may result. Other peaks observed in groundwater circular histogram s might be attributed to cana l stage management. Studies by

PAGE 63

63 Merrit (1996), Ritter and Muoz -Carpena (2006) and Klein and Hull (1978) have noted the interconnection of canal stages and groundwater levels. As suggested, data collected could not provide a definite explanation for observed pattern s. Further investigation is needed to accurate access driving factors influence groundwater level fluctuations. Capillary Contribution Relative Significance Although Rayleigh test determ ined the existence and location of the diur nal peaks in all the variables studied, the relative importance of this increase of soil wate r content during the day needs to be studied in detail to determine its contribution to plan t water uptake purposes. According to the results shown in Table 3-8, the soil water cont ent daily difference between the maximum daily values and the daily mean soil water content was less th an 1% of volumetric water content in all cases. This daily peak <1% of soil water content could be attributed to the relatively low groundwater daily mean diurnal range (< 4 cm). Thus, despite statistical significance of the mean vector s found, suggested hypothesis of possible daily contribution to plant water uptake through diurna l capillarity requires further re search. Considering an actual ET of 3.3 mm/day (Migliaccio et al ., 2008b) groundwater diurnal fluctuations would be able to contribute an average of 0.58 0.46 mm of daily soil water content equivalent to 18% of mean actual daily ET. Conclusions Location an d statistical significan ce of diurnal fluctuations we re confirmed with circular statistical analysis. The technique represents a practical approach to locate and confirm the existence of mean vectors of soil water content and groundwater level usi ng the Rayleigh test of uniform distribution. The mean vectors had statis tical significance in all the tests and circular histograms provided a method for better evaluating the large dataset.

PAGE 64

64 Many of the mean vectors found were characte rized by multimodal distributions attributed to multivariate effects. The mean vector of gro undwater was closer to the mean vectors of soil water content at 40 and 60 cm depths which also corresponded with the mean vectors of relative humidity at dew point hours. Re sults suggest that soil water at 10 and 20 cm depths were more similar to solar radiation (evapot ranspiration) and soil water temper ature mean vectors. Most of the variables evaluated did not met unimodal Von Mises distribution for purposes of parametric comparisons among variables. Circ ular-circular correlati on analysis confirmed in general a weak relationship between soil water content and the weather and groundwater p eak mean vectors.

PAGE 65

65 Table 3-1. Equipment used to identify di urnal peaks using circular statistics. Equipment Variable Number of Instruments EnviroSCAN Soil water content 24 Levelogger LT 3001 Water table elevation 4 Barologger Barometric pressure 1 FAWN1 weather station Relative humidity 1 FAWN weather station Solar radiation 1 FAWN weather station Temperature 1 1 Florida Automated Weather Network (FAWN) weather station at Homestead, Miami-Dade County, Florida.

PAGE 66

66 Table 3-2. Circular statistical analysis of the occurrence of maximum daily soil wate r content in trenched conditions at 10, 20 40 and 60 cm depths. Variable A6-10 A6-20 A6-40 A6-60 A1-10 A1-20 A1-40 A1-60 A11-10 A11-20 A11-40 A11-60 Mean Vector () 11:42 PM 6:55 AM 8:33 AM 9:09 AM 3:14 PM 11:03 PM 7:43 AM 9:24 AM 4:13 PM 9:06 PM 2:50 AM 6:48 AM Length of Mean Vector (r) 0.347 0.419 0.894 0.896 0.669 0.878 0.577 0.766 0.654 0.637 0.448 0.282 Median 12:00 AM 9:00 AM 8:45 AM 9:15 AM 3:15 PM 12:00 AM 8:45 AM 9:45 AM 4:00 PM 8:15 PM 2:00 AM 8:00 AM Concentration 0.74 0.923 5.022 5.091 1.83 4.386 1.419 2.503 1.753 1.668 1.001 0.588 Circular Variance 0.653 0.581 0.106 0.104 0.331 0.122 0.423 0.234 0.346 0.363 0.552 0.718 Circular Standard Deviation 83.361 75.545 27.088 26.88 51.354 29.28 60.069 41.826 52.781 54.45 72.615 91.165 Standard Error of Mean 9.917 8.078 2.371 2.353 4.541 2.561 5.552 3.629 4.692 4.876 7.505 12.342 95% Confidence Interval (-/+) 10:25 PM 5:52 AM 8:15 AM 8:50 AM 2:38 PM 10:43 PM 6:59 AM 8:55 AM 3:36 PM 8:28 PM 1:51 AM 5:11 AM 1:00 AM 7:59 AM 8:52 AM 9:27 AM 3:49 PM 11:23 PM 8:26 AM 9:52 AM 4:49 PM 9:44 PM 3:49 AM 8:25 AM Watson's U Test (von Mises, U) 1.708 1.385 2.592 2.89 0.389 1.114 1.701 1.525 1.683 0.903 0.914 0.711 Watson's U Test (p) < 0.005 < 0.005 < 0.005 < 0.005 < 0.005 < 0.005 < 0.005 < 0.005 < 0.005 < 0.005 < 0.005 < 0.005 Rayleigh Test (Z) 15.654 22.853 103.961 104.317 58.218 100.121 43.31 76.297 55.642 52.69 26.084 10.338 Rayleigh Test (p) 1.59E-07 1.19E-10 < 1E-12 < 1E-12 < 1E-12 < 1E-12 < 1E-12 < 1E-12 < 1E-12 < 1E-12 4.70E-12 3.24E-05 Table 3-3. Circular statistical analysis of the occurrence of maximum daily soil water content in non trench ed conditions at 10 20, 40 and 60 cm depths. Variable B8-10 B8-20 B8-40 B8-60 B3-10 B3-20 B3-40 B3-60 B9-10 B9-20 B9-40 B9-60 Mean Vector () 6:15 PM 4:16 AM 12:47 AM 1:56 AM 3:11 PM 10:44 PM 1:00 AM 1:54 PM 3:22 PM 9:45 PM 12:17 AM 2:10 AM Length of Mean Vector (r) 0.312 0.403 0.492 0.305 0.795 0.831 0.633 0.2 0.593 0.683 0.902 0.646 Median 5:00 PM 2:52 AM 12:15 AM 11:45 PM 2:45 PM 11:45 PM 12:15 AM 1:45 PM 3:00 PM 11:45 PM 11:45 PM 1:30 AM Concentration 0.656 0.879 1.127 0.64 2.799 3.292 1.653 0.408 1.482 1.903 5.41 1.714 Circular Variance 0.688 0.597 0.508 0.695 0.205 0.169 0.367 0.8 0.407 0.317 0.098 0.354 Circular Standard Deviation 87.483 77.281 68.241 88.343 38.847 34.914 54.766 102.793 58.554 50.078 25.975 53.528 Standard Error of Mean 11.111 8.444 6.748 11.384 3.37 3.039 4.912 17.584 5.36 4.409 2.274 4.774 95% Confidence Interval (-/+) 4:48 PM 3:10 AM 11:54 PM 12:27 AM 2:44 PM 10:20 PM 12:21 AM 11:36 AM 2:40 PM 9:10 PM 11:59 PM 1:32 AM 7:42 PM 5:22 AM 1:40 AM 3:26 AM 3:37 PM 11:07 PM 1:39 AM 4:12 PM 4:04 PM 10:19 PM 12:35 AM 2:47 AM Watson's U Test (von Mises, U) 0.682 0.754 0.529 0.674 2.392 1.111 0.938 0.945 1.653 1.925 1.839 0.998 Watson's U Test (p) < 0.005 < 0.005 < 0.005 < 0.005 < 0.005 < 0.005 < 0.005 < 0.005 < 0.005 < 0.005 < 0.005 < 0.005 Rayleigh Test (Z) 12.632 21.079 31.468 12.063 82.092 89.676 52.138 5.201 45.747 60.559 105.849 54.311 Rayleigh Test (p) 3.27E-06 7.01E-10 < 1E-12 5.77E-06 < 1E-12 < 1E-12 < 1E-12 0.006 < 1E-12 < 1E-12 < 1E-12 < 1E-12

PAGE 67

67 Table 3-4. Circular statistical analysis of the occurrence of maximum daily pooled soil water content at 10, 20, 40 and 60 cm d epths. Variable Soil Water 10 cmSoil Water 20 cmSoil Water 40 cm Soil Water 60 cm Mean Vector () 4:10 PM11:16 PM3:23 AM 7:23 AM Length of Mean Vector (r) 0.464 0.466 0.399 0.315 Median 3:30 PM12:00 AM2:00 AM 8:45 AM Concentration 1.045 1.05 0.871 0.665 Circular Variance 0.536 0.534 0.601 0.685 Circular Standard Deviation 71.0470.85277.623 87.055 Standard Error of Mean 2.948 2.934 3.478 4.482 95% Confidence Interval (-/+) 3:47 PM10:53 PM2:55 AM 6:48 AM 4:33 PM11:39 PM3:50 AM 7:59 AM Watson's U Test (von Mises, U) 3.171 2.17 2.053 1.574 Watson's U Test (p) < 0.005< 0.005< 0.005 < 0.005 Rayleigh Test (Z) 167.666169.037124.449 77.536 Rayleigh Test (p) < 1E-12< 1E-12< 1E-12 < 1E-12 Table 3-5. Circular statistical analysis of the occurrence of maximum daily water tabl e elevation at four wells and all data co mbined. Variable WTE-W1WTE-W2WTE-W3WTE-4 WTE Mean Vector () 4:11 AM7:03 AM4:09 AM7:31 AM4:41 AM Length of Mean Vector (r) 0.7680.3650.7920.3410.614 Median 4:45 AM8:45 AM4:45 AM9:00 AM5:00 AM Concentration 2.5210.7832.770.7271.569 Circular Variance 0.2320.6350.2080.6590.386 Circular Standard Deviation 41. 62681.39239.10883.99156.551 Standard Error of Mean 3.6119.4053.39213.0232.955 95% Confidence Interval (-/+) 3:43 AM5:49 AM3:43 AM5:49 AM4:18 AM 4:39 AM8:16 AM4:36 AM9:13 AM5:04 AM Watson's U Test (von Mises, U) 0.2791.5270.2720.380.991 Watson's U Test (p) < 0.005< 0.005< 0.005< 0.005< 0.005 Rayleigh Test (Z) 76.68617.2881.5859.096147.227 Rayleigh Test (p) < 1E-123.13E-08< 1E-121.12E-04< 1E-12

PAGE 68

68 Table 3-6. Circular statistical analysis of the occurre nce of maximum daily observa tions of weather factors. Variable Barometric PressureSoil TemperatureRelative HumiditySolar Radiation Mean Vector () 11:14 AM3:58 PM4:58 AM12:42 PM Length of Mean Vector (r) 0.814 0.983 0.57 0.97 Median 11:00 AM4:00 PM7:00 AM12:45 PM Concentration 3.05 30.523 1.39217.045 Circular Variance 0.186 0.017 0.43 0.03 Circular Standard Deviation 36.70710.458 60.75214.091 Standard Error of Mean 3.188 0.917 5.6411.236 95% Confidence Interval (-/+) 10:49 AM3:51 PM4:13 AM12:32 PM 11:39 AM4:05 PM5:42 AM12:52 PM Watson's U Test (von Mises, U) 2.586 0.317 1.6450.155 Watson's U Test (p) < 0.005< 0.005 < 0.005< 0.025 Rayleigh Test (Z) 86.237 125.74 42.235122.37 Rayleigh Test (p) < 1E-12< 1E-12 < 1E-12< 1E-12

PAGE 69

69 Table 3-7. Circular-circular Pearson product moment correlation tests between soil water sensor data and groundwater level as w ell as weather variables during the study period. Variable WTE-W1 WTE-W2 WTE-W3 Barometric Pressure So il Temperature Relative HumiditySolar Radiation r P r P r P r P r P r P r P A6-10 -0.075 < 0.05 0.043 < 0.05-0.02< 0.050.019< 0.05 -0.073< 0.05-0.117< 0.05-0.026< 0.05 A6-20 0.065 < 0.05 0.064 < 0.050.072< 0.05-0.089< 0.05 -0.225< 0.050.077< 0.050.002< 0.05 A6-40 0.324 < 0.05 -0.046 < 0.050.253< 0.05-0.042< 0.05 0.093< 0.050.427< 0.050.011< 0.05 A6-60 0.34 < 0.05 2.98E-04 < 0.050.238< 0.05-0.046< 0.05 -0.007< 0.050.37< 0.050.067< 0.05 B8-10 -0.102 < 0.05 0.031 < 0.05-0.114< 0.050.072< 0.05 -0.022< 0.05-0.095< 0.05-0.043< 0.05 B8-20 0.045 < 0.05 -0.028 < 0.050.138< 0.05-0.048< 0.05 0.134< 0.050.073< 0.050.037< 0.05 B8-40 -0.06 < 0.05 0.004 < 0.05-0.042< 0.050.044< 0.05 -0.118< 0.05-0.051< 0.05-0.112< 0.05 B8-60 0.025 < 0.05 0.018 < 0.050.019< 0.05-0.019< 0.05 6.24E-04< 0.05-0.03< 0.050.07< 0.05 A1-10 -0.079 < 0.05 -0.151 < 0.05-0.015< 0.05-0.053< 0.05 0.192< 0.05-0.015< 0.05-0.011< 0.05 A1-20 -0.186 < 0.05 -0.125 < 0.05-0.163< 0.050.038< 0.05 -0.214< 0.05-0.083< 0.05-0.121< 0.05 A1-40 0.059 < 0.05 -0.034 < 0.050.078< 0.050.025< 0.05 -0.039< 0.050.208< 0.050.026< 0.05 A1-60 0.095 < 0.05 -0.069 < 0.050.177< 0.05-0.021< 0.05 0.126< 0.050.165< 0.050.087< 0.05 B3-10 -0.006 < 0.05 -0.011 < 0.05-0.123< 0.050.002< 0.05 0.069< 0.05-0.112< 0.05-0.078< 0.05 B3-20 -0.134 < 0.05 -0.082 < 0.05-0.131< 0.050.122< 0.05 -0.12< 0.05-0.087< 0.05-0.157< 0.05 B3-40 0.01 < 0.05 0.006 < 0.050.003< 0.05-0.049< 0.05 -0.009< 0.05-0.011< 0.050.008< 0.05 B3-60 0.016 < 0.05 -0.023 < 0.05-0.002< 0.050.023< 0.05 -0.015< 0.05-0.037< 0.05-0.002< 0.05 A11-10 -0.002 < 0.05 -0.085 < 0.05-0.063< 0.05-1.05E-04ns 0.257< 0.050.019< 0.05-0.065< 0.05 A11-20 0.001 < 0.05 -0.119 < 0.05-0.059< 0.050.166< 0.05 0.031< 0.05-0.044< 0.05-0.148< 0.05 A11-40 0.069 < 0.05 0.046 < 0.050.059< 0.050.037< 0.05 -0.062< 0.050.054< 0.050.007< 0.05 A11-60 0.058 < 0.05 0.066 < 0.050.064< 0.05-0.059< 0.05 -0.025< 0.050.063< 0.050.069< 0.05 B9-10 -0.023 < 0.05 0.039 < 0.050.027< 0.050.065< 0.05 0.043< 0.050.031< 0.05-0.003< 0.05 B9-20 -0.127 < 0.05 -0.008 < 0.05-0.216< 0.050.062< 0.05 0.007< 0.05-0.147< 0.05-0.095< 0.05 B9-40 0.013 < 0.05 -0.017 < 0.05-0.017< 0.050.011< 0.05 0.086< 0.050.044< 0.050.106< 0.05 B9-60 0.075 < 0.05 -0.085 < 0.050.133< 0.05-0.094< 0.05 -0.055< 0.050.104< 0.050.171< 0.05

PAGE 70

70 Table 3-8. Estimated daily contribution of groundwater diurnal fluctuations to soil water content at four different depths and soil conditions of a typical Krome so il profile. Soil water contribution is the difference of the daily mean and the maximum daily value at each soil condition. Depth (cm) Overall Trench No Trench %mean %stdev %mean %stdev %mean %stdev 10 0.490 0.331 0.592 0.454 0.388 0.191 20 0.239 0.111 0.217 0.100 0.261 0.138 40 0.150 0.101 0.152 0.140 0.148 0.075 60 0.087 0.032 0.082 0.031 0.092 0.040

PAGE 71

71 Figure 3-1. The circular statis tics rectangular components of a mean vector and its determination through the construction of the center of gravity from three observations. Figure 3-2. Schematic of selected probes and locati on of monitoring wells in experimental site at University of Florida Tropical Research and Education Center (TREC). A symbolized probes located in a trenched Krome soil condition and B symbolizes probes located in a non trenched limestone soil condition.

PAGE 72

72 14 16 18 20 22 24 26 28 10 cm 20 cm 40 cm 60 cm 10/1/200811/1/200812/1/20081/1/20092/1/20093/1/2009 WTE NGVD 1929 (cm) 60 80 100 120 140 160 180 Barometric Pressure (cm) 0 20 40 60 80 100 120 Well 1 Well 2 Well 3 Well 4 Atmospheric pressure Figure 3-3. Daily means for the entire dataset of groundwater level, barometric pressure and overall soil water content at 10, 20, 40 and 60 cm depths. Study period from16 September 2008 to 23 March 2009.

PAGE 73

73 A11-10 11/20 00:0011/20 12:0011/21 00:0011/21 12: 0011/22 00:00 % 26.2 26.3 26.4 26.5 26.6 26.7 WTE NGVD 1929 (cm) 111 112 113 114 115 116 117 118 119 Solar Radiation (w/m^2) 0 20 40 60 80 100 Relative Humidity % 40 50 60 70 80 90 100 Soil Temperature (C) 12 14 16 18 20 22 Barometric Pressure 88 90 92 94 96 98 100 102 Soil Water WTE Solar Radiation Relative Humidity Soil Temperature Barometric Pressure Figure 3-4. Diurnal fluctuati ons from November 20 to November 22, 2008; Influence of groundwater level (WTE), solar radiation, re lative humidity and soil temperature in the soil water content of pr obe A11-10 (depth: 10 cm). A11-20 11/20 00:0011/20 12:0011/21 00:0011/21 12:0011/22 00:00 % 24.56 24.58 24.60 24.62 24.64 24.66 24.68 24.70 24.72 WTE NGVD 1929 (cm) 111 112 113 114 115 116 117 118 119 Solar Radiation (w/m^2) 0 20 40 60 80 100 Relative Humidity % 40 50 60 70 80 90 100 Soil Temperature (C) 12 14 16 18 20 22 Barometric Pressure 88 90 92 94 96 98 100 102 Soil Water WTE Solar Radiation Relative Humidity Soil Temperature Barometric Pressure Figure 3-5. Diurnal fluctuati ons from November 20 to November 22, 2008; Influence of groundwater level (WTE), solar radiation, re lative humidity and soil temperature in the soil water content of pr obe A11-20 (depth: 20 cm).

PAGE 74

74 A11-60 11/20 00:0011/20 12:0011/21 00:0011/21 12: 0011/22 00:00 % 19.65 19.70 19.75 19.80 19.85 19.90 19.95 20.00 WTE NGVD 1929 (cm) 111 112 113 114 115 116 117 118 119 Solar Radiation (w/m^2) 0 20 40 60 80 100 Relative Humidity % 40 50 60 70 80 90 100 Soil Temperature (C) 12 14 16 18 20 22 Barometric Pressure 88 90 92 94 96 98 100 102 Soil Water WTE Solar Radiation Relative Humidity Soil Temperature Barometric Pressure Figure 3-6. Diurnal fluctuati ons from November 20 to November 22, 2008; Influence of groundwater level (WTE), solar radiation, re lative humidity and soil temperature in the soil water content of pr obe A11-60 (depth: 60 cm).

PAGE 75

75 Figure 3-7. Circular histograms and mean vector s of filtered soil water maximum daily values.

PAGE 76

76 Figure 3-8. Circular histograms and mean vector s of pooled, filtered soil water maximum daily values at 10, 20, 40 and 60 cm depths.

PAGE 77

77 Figure 3-9. Circular histograms and mean vect ors of filtered groundwater level maximum daily values from all wells (WTE), well 1 (WTE-W1), well 2 (WTE-W2), well 3 (WTEW3) and well 4 (WTE-W4).

PAGE 78

78 Figure 3-10. Circular histograms and mean vectors of filtered weather maximum daily values.

PAGE 79

79 CHAPTER 4 MONITORING GROUNDWATER LEVELS TO PREDICT SOIL WATER CONTENT IN THE UNSATURATED ZONE OF A CAL CAREOUS SOIL Introduction The hydrology of South Florida was significan tly m odified during th e 1950s by a complex land drainage project to protec t the agricultural and urban deve loping areas from flooding. This project consisted of building a drainage canal ne twork that influenced we tlands contiguous to the Everglades National Park (ENP). An effort is currently in place through the Comprehensive Everglades Restoration Plan (CER P) to rehabilitate the natural flow and water deliveries to the ENP without affecting the developed adjacent area s. To achieve these goals, the South Florida Water Management District (SFWMD) may need to raise groundwater levels (Poole, 1996; Perry, 2004). The Biscayne aquifer is the main source of fres hwater in Southeast Flor ida. It is described as an unconfined aquifer consisting principally of porous limestone that increases in thickness from about 14 m in the western part of MiamiDade County to about 30 m in the eastern part (Fish and Stewart, 1991). Typical porosity is about 0.26 and hydr aulic conductivity is generally greater than 305 m/day (Chin et al., 2005). Its characterist ics of unconfinement, high permeability and shallow depth result in water tabl e fluctuations that are fairly responsive to storm events and canal system management (Pitt, 1976; Ritter and Muoz-Carpena, 2006). According to Chin et al. (2008), the aquifers sp ecific yield averages 0.23 as response to direct recharge of groundwater from perc olation of rainwater, and sha llow saturated zone evaporation rate equals the potential evapor ation rate to a depth of 1.4 m decreasing to zero at a depth of 2.5 m. The shallow nature of the Biscayne Aquifer re sults in the potential for it to influence soil water conditions in the unsaturated zone. Despite its high permeability, a triple porosity characteristic described by C unningham et al. (2006) in conjunction with the local shallow

PAGE 80

80 groundwater conditions suggest a capill ary rise effect. The capillary rise is related to the forces in the soil that wick water upward ag ainst gravity to form a zone a bove the water table that remains saturated called capillary fringe. The length of the capillary fringe is defined by the pore size distribution of the soil (Todd, 1959) and its potential contributi on to plant water uptake is dependent on water table depth and site elevation (Wellings and Bell, 1982). Shallow water table contribution to plant water needs has been a subject of extensive research world-wide. A brief review of literature describing shallow groundwater contributes to crop water needs is provided in Babajimopoulos et al. (2007); shallow groundwater tables were shown to provide a portion of the crop wa ter needs for cotton re placing 60% of the evapotranspiration (ET) (Wallender et al., 1979) and 37% of ET (Ayars and Schoneman, 1986) in San Joaquin Valley, California. A study in Pakist an suggested that irrigation could be reduced by 80% under shallow groundwater conditions (Prathapar and Qureshi, 1999). The interest in understanding this phenomenon is evident by the de velopment of models that simulate shallow groundwater dynamics, such as Upflow (Raes and Deproost, 2003) and DRAINMOD (Skaggs, 1978a). The relationship between soil water content and shallow groundwater level is a dynamic process of alternate cycles of wetting and drying. In a field scale situation, this relationship can be more simply evaluated by assuming hydrostatic conditions. Hydrostatic conditions refer to soil water in equilibrium (without movement). Th e concept of using hydrostatic conditions or drain to equilibrium conditions was introduced and explained in detail by Wellings and Bell (1982). If equilibrium conditions occur, water pre ssure potential is determined by the distance of a specific point in the soil to the water table, with positive values when the point is below the water table (saturation and hydrau lic head) and negative when the point is above the water table

PAGE 81

81 due to its matric potential in terms of soil water suction (B uckingham, 1907; Klute, 1986; Skaggs, 1978b) (Figure 4-1). Thus, a soil water char acteristic curve in a homogenous soil profile drained to equilibrium can be obtained by collecting a series of simultaneous observations of soil water contents and groundwater levels. Soil water point of saturation would be reached when water table is zero, and its volum etric water is equal to the effective porosity (Figure 4-1) (Wellings and Bell, 1982). Capillarity effects have been previously studied using the drained to equilibrium concept. Sumner (2007) used the idea to evaluate specif ic yield response to micro topography in the Everglades National Park wetlands, while Nachabe et al. (2004) tested this concept to model soil water storage capacity in Myakka fine sands of Tampa Bay, Florida. A lthough no studies (to the authors knowledge) have focused on capillary cont ribution to the soil ro ot zone in Southeast Florida agricultural soils, previous anecdotal evidence has been reported in the region suggesting that groundwater may be contributing to plant wate r requirements. For example, Al-Yahyai et al. (2005) reported a lack of physiological response between irrigated and non irrigated carambola ( Averrhoa carambola ) and assumed this result was due to shallow groundwater capillary rise. Additional work characterizing the soil water retent ion in this orchard described the difficulty in obtaining suction values greater than 125 cm in the field tr eatments without irrigation (AlYahyai et al., 2006) suggesting that groundwater was contribu ting to soil water content. Similarly, Migliaccio et al. (2008) reported evidence of capillary influences on soil water status. They investigated soil suction in Marl soils of South Florida and identified diurnal patterns during the dry season (November to May). The potential impact of increased soil water content by shallo w groundwater is of particular concern to deep rooted crops, such as tropical fruits due to flooding tolerance and

PAGE 82

82 water use efficiency (Schaffer, 1998). Understa nding the hydrological interactions between groundwater level and soil water cont ent is needed so that soil wa ter content scenarios at given water table levels can be evaluated in terms of their benefits and challenges from water management and agronomical perspectives. The objective of this study was to test the va lidity of the drain to equilibrium hydrostatic assumptions for predicting the soil water content based on shallow groundwater elevation in a typical tropical fruit grove in southeast Florida. The analysis was conducted in three steps: (i) monitoring of soil water content and groundwater level in a synchronized manner and identify when hydrostatic conditions developed; (ii) determ ination of soil water char acteristic curves of soil profiles that were considered drained to equilibrium; and (iii) construction of simple hydrostatic models to predict soil water content based on groundwater level observations. Materials and Methods Experimental Site The study was conducted in Hom estead south Miami-Dade County, Florida, in a lychee ( Litchi chinensis ) orchard at the University of Florida (UF) Tropical Research and Education Center (TREC) (80.5 W, 25.5 N). The site has a mean elevation of 3.12 m National Geodetic Vertical Datum (NGVD) 1929 and a subtropical marine climate. Annual rainfall is 1,448 mm/yr (Ali and Abtew, 1999) with 80% occurring during the wet season (June to October). The dry season (November to May) is characterized by a decreasing trend in groundwater levels (Chin, 2008). The groundwater level near the site (100 m south) is monitored by USGS and depth ranges are -0.38 to 2.99 m NGVD (1929) from 1950 to 2008 (USGS, 2009). The soils at the study site are classified as Krome calcareous very gravelly loam soils. Krome soils are loamy skeletal, carbonatic, hyperthermic Lithic Udorethents (Noble et al., 1996) made of rock plowed oolitic limestone; a techni que introduced during the early 1950s to increase

PAGE 83

83 the soil depth and improve agricultural productio n in the zone (Colburn and Goldweber, 1961). Muoz-Carpena et al. (2002) a nd Al Yahyai et al. (2006) de scribe Krome soils physical structure in two solid fractions, 51% coarse part icles (>2 mm) and 49% loam particles. The soil is characterized by a complex bimodal water retention pattern, where the gravel fraction contributes to very rapid soil wa ter depletion and the loamy frac tion retains water. Krome soils typical profile is constituted by a scarified layer (15 cm ) of very gravelly loam texture and an underlying layer of limestone bedrock. The lychee orchard (as most tropical fruit crops in the area) is planted in rock plow ed trenches 0.40 to 0.45 m wide and 0.45 to 0.60 m deep (Figure 42); this practice is needed to increase the soil de pth for rooting growth and tree stability (Crane et al., 1994). The trenches are usually dug in parallel lines for the tree rows with another set of parallel trenches perpen dicular to the first for trees spacing. Trees are planted at the intersection of the trenches (Li, 2001). The experimental site depicted in Figure 4-2 consists of si x rows: two rows with lychee trees planted in trenches (hence after referred as lychee trench ), one row without lychee trees but with trenched soil (hence afte r referred as trench), two rows without lychee trees but with surficial roots and non trenched soil (hence afte r referred as lychee no trench) and one row without lychee trees or r oots and without trenches (hence afte r referred as no trench). Each lychee row consisted of 12 trees planted at a spacing of 4.6 m within rows and 7.6 m between rows; plant density was 286 trees /ha. Considering that the ma in purpose of the study was to explain soil water dynamics, the se lection of this crop represente d in general the planting and spacing of tropical fruits in Southeast Florida. Assuming hydrostatic conditions (in equilibriu m and without water movement) the soils profile hydraulic potential ( H ) is equivalent to the differen ce between its pressure potential ( h )

PAGE 84

84 and gravimetric potential (z ) (Figure 4-1). The pr essure potential (h) at a given point is defined as the distance of the point to the water table and the gravimetric potential (z ) is the distance of the given point above the reference leve l. If the given point is located in the unsaturated zone above the water table, its pressure potential will be negative (suction) and the hydraulic potential is expressed as the distance of the re ference level to the water table ( L ) (Muoz-Carpena and Ritter, 2005). Soil Water and Groundwater Monitoring : Identify ing Hydrostatic Conditions Soil water content was measured using EnviroSCAN multi sensor capacitance probes (Sentek Ltd. Pty., Stepney, Australia). Three data loggers monitored 24 probes distributed in the six rows included in the study (Figure 4-3). Each row had four probes spaced at 13.8 m and each probe had four sensors positioned at 10, 20, 40 and 60 cm depths (Figure 4-2). The system of 96 sensors measured soil water content as a function of the apparent bulk dielectric constant of the soil described in detail by Paltineanu and Starr (1997) at a resolution of 0.1% (Buss, 1993; Fares and Alva, 2000). A sensor normalization procedure was used to determine the scaled frequency ( SF ): WA SAFF FF SF (4-1) where FA is the frequency reading inside th e PVC probe while suspended in air, FW is the reading inside the PVC probe in a water bath and FS are the subsequent readi ngs inside the PVC probe installed in the soil. Then, a re gression equation for the specific soil type (i.e., Krome soil) was used to relate the scaled frequency ( SF ) to the volumetric water content ( ). A calibration (Equation 4-2) previously devel oped by Al-Yahyai et al. (2006) through the gravimetric water content method was used:

PAGE 85

85 BA CSF/1 (4-2) where is the volumetric water content and A B and C are coefficients fitted through nonlinear regression coefficient values are 0.011, 1 and 0.5206 for A, B and C, respectively. Groundwater level was measured using Le velogger LT 3001 (Solinst Ltd., Ontario, Canada) pressure transducers suspended and submerged at 3.6 m depth (accuracy = 0.1%, -10 C to 40 C). The leveloggers were positioned in four monitoring wells located along the perimeter of the study site (Figure 4-2). Location coordinates and surveyed elevations of the monitoring well risers and their surface refe rences are provided in Table 4-1 and Figure 4-4. Water levels were compensated using air barometric pressure measurements from a Barologger. Groundwater levels were also adjusted based on week ly manual water table depth readings. Both soil water content and groundwater leve l monitoring devices were synchronized to collect readings every 15 min from 16 September 2008 to 23 March 2009. Irrigation was suspended during the study to promote drained to equilibrium conditions although irrigation for freeze protection did occur. Daily rainfall and re ference evapotranspiration (ETo) data were collected from the Florida Automated Weather Network (FAWN) station located at UF TREC, less than 100 m from the study si te. Data collected when events occurred that would likely violate the drain to equilibrium assumption (e.g., storms, freeze protection events and remaining moisture greater than daily evapotranspira tion) were discarded from the analysis. Drained to Equilibrium Soil Water Characteristic Curves The filtered groundwater level a nd so il water content datasets were summarized to daily means. The soil water content (after filtering) were categorized according to the installed depth and soil conditions. Thus, the soil water content dataset was subdivided in to 16 different soil conditions (Table 4-2, Figure 42). Groundwater level data from the four piezometers were

PAGE 86

86 averaged and summarized to daily means. Water ta ble depth (WTD) values were adjusted to the distance between the soil water sensor and the water table depth (DSM). This adjustment was made to relate the pressure head or soil water content at each sensor depth to the groundwater level. Thus, the filtered soil water content data from each sensor with its correspondent DSM (pressure head) defines a field soil water rela tionship. This relationship was defined using the van Genuchten-Mualem model (van Genuchten, 1980). The Van Genuc hten et al. (1991) RETC program was used to fit collected data by applyi ng least squares parameter optimization and the van Genuchten (1980) equation: m n rs rh 1 (4-3) where is the volumetric water content; h is the pressure head; s and r represent the saturated and residual water cont ents, respectively; and n and m are empirical shape parameters. The parameters generally adopt values in the following ranges: 0 < < 1, n > 1, 0 < m < 1, h 0. Predicting Soil Water Based on Groundw ater Observations The fitted parameters ( s, r, n and m ) estimated from the soil water retention data at each sensor were used to predict soil wate r according to the DSM daily observations. The difference between observed values and model predictions were evaluated using the Nash and Sutcliffe (1970) coefficient of efficiency ( Ceff) described in Equation (4-4). n i i n i iiOO PO Ceff1 2 1 2)( 0.1 (4-4) where Oi are the observed values, Pi is predicted values and is the mean of the measured data; values for Ceff range between 1.0 (perfect fit) and negative infinity. Ceff values less than zero indicate that the mean value of the measured time series would be a better predictor of the model (Nash and Sutcliffe, 1970).

PAGE 87

87 Results and Discussion Soil Water and Groundwater Monitoring : Identify ing Hydrostatic Conditions During the study period more than 1.7 million readings of soil volumetric water content ( ) were collected from 96 capacitance sensors a nd about 72,500 groundwater le vel readings were collected from four pressure transducers. Ranges and variances found among sensors from the same soil conditions varied significantly. Thus, each sensor data set was used to develop a sensor specific soil water characteristic curve. Variability observed in the same soil conditions could be attributed to contact interferen ce of the probe with the soil due to air gaps, instrumentation deficiencies and/or site speci fic conditions of soil profile. From the soil water summary in Table 4-3 of all collected data, the greatest means of volumetric soil water content were at the shallowe st depth (i.e., 10 cm) for lychee trench, lychee no trench and trench soil conditions The standard deviations were also greater for data collected at this depth which was likely due to greater exposure of shallo w soil depths to weather factors including rainfall and freeze protection irrigation. The spikes of soil water content observed in January 2009 were related to freeze protection a nd sensors responded differently according to their location and distan ce from the sprinklers. The selected representative se nsors in Figure 4-5 show the soil water dynamic response to rainfall events and mean groundwater table elevation (WTE) fluctuati ons. The plotted data indicates the influence of storms from the end of September 2008 to the beginning of October 2008. Data collected from the beginning of dry season (November) unt il March 2009 provided a period of fewer rain events. This dry period offered the conditions to collect most of the data representing a soil water state of equilibrium. This filtered data set is summarized in Table 4-3. The groundwater level data range from field measurements wa s verified by comparing the range of the collected data to that of groundwater monitoring we ll (well 196A) ma intained by the

PAGE 88

88 United States Geological Survey (USGS, 2009), lo cated less than 100 m south of the study site. Monthly groundwater records from USGS (2009) had very similar monthly means with the 4 monitoring wells (Table 4-4). The historical monthly median s (1950-2008) reported by USGS were below the monthly medians measured in the groundwater monitoring wells every month except September. All the collected values were w ithin the USGS historical range and it could be concluded that the data collected accurately represented the groundw ater level. Considering that the drained to equilibrium models dependent variable (x-axi s) is the DSM or adjusted groundwater table depth. The leve ls obtained during the study we re higher than historical medians. However, distances from the deepest se nsors (60 cm) to the groundwater level less than 120 cm were not obtained during the studied period. Drained to Equilibrium Soil Water Characteristic Curves Results of fitted curves and parameters are shown in Figures 4-6 to 4-8 and Tables 4-5 to 4-8. Filtering the collected dataset for points that met the drain to equilibrium assumption resulted in 75 data points (representing mean values) per sensor. Data from each sensor were used with the van Genuchten model (1980) and fitte d parameters were determined (Tables 4-5 to 4-8). For each soil condition, a model was develope d considering all daily mean values of the same category (e.g., one category was all data fr om lychee trench 10 cm depth conditions). The fitted parameters at 10 cm in Table 4-5 indicate a greater variance in the n fitting parameter compared to the remaining soil condit ions at the same dept h (Table 4-6 to 4-8). Parameter n represents the slope of the curves an d generally indicates the differences in rate of soil water depletion. The va riability observed in the lychee tren ch treatment could be due to the variability in root systems and soils expected in lychee trenched conditions. Alternatively, most of the soil conditions had similar values for th e saturated water content (average 0.26) and air entry ( = 0.005). From the fitted van Genuchten e quation, the inflection point derived from

PAGE 89

89 Dexter and Bird (2001) for all mean soil conditio ns in Table 4-5 were very similar for all the sites ( infl = 0.14). Soil water results from all sensors showed certain patterns of water replenishment and depletion related to depth. Genera lly it is expected to find more variation in the vertical soil profile as compared to horizont al soil profile due to variabi lity associated with structure, compaction and layering. Skaggs et al. (1978b) remarked about the vertical heterogeneity feature and explained that superficial prof iles tend to have higher porosity than the deeper layers. Bruce and Luxmoore (2003) explained situ ations where the curves may vary more with depth than with area in response to compaction. For this study, many of the soil water ch aracteristic curves illustrated this phenomenon (Figure 4-6, 4-7 and 4-8). Vertical variability was observed for most sites as illustrated by the decreasing pore size distribution parameter (n ) with increasing depth (T able 4-5). For example, n values calculated for all data in lychee trenched conditions were 21.123, 16.093, 10.808, and 2.115 for 10, 20, 40, and 60 cm depths, respectively. However, soil water c ontent results also varied in small horizontal distances although it was the same soil type and soil condition. Soil horizontal variability as discussed by Becket and Webster (1971) and Warrick et al. (1977) requires interpretation according to common factors. Variation was found to be greater in sample points close to the soil surface. The horizontal variations identified in the same soil condition suggest the presence of structural heterogeneity. The lychee trench soil water characteristic cu rves from probes A1 to A8 (Figures 4-6 and 4-7) generally suggest a break poi nt of depletion at 180 cm of pr essure head (DSM) not seen in the trench probes or the non trench ed soil conditions. This breakpoi nt suggests a threshold were

PAGE 90

90 bubbling pressure breaks and depletion begins. Thus, pressure head close to 180 cm could indicate the limit of capillary contribution. Data from the lychee no trenched soil conditi on had similar patterns among sensor depths but different saturated water c ontents. In spite of the limited groundwater table range obtained during the study, the continuity f ound among certain curves of differe nt depths in the same probe suggested similar soil conditions at certain depth ranges and therefore the possibility to combine datasets to increase the range of pressure head observations av ailable to characterize the soil water content characteristic curve. Specifically, probes B1, B2, B4, B5, B6 and B7 at 40 and 60 cm depth (Figures 4-6 and 4-7) represented si milar strata of the limestone and could be combined to capitalize on this benefit. These cu rves had a more steady depletion rate where large changes in suction responded to small changes in soil water content. In the graphs from the mentioned probes, a different curve was identified for the 10 cm depth sensor which represented the Krome scarified layer. Water content at field capacity by definition refers to an average suction of 100 kPa (100 cm of water pressure head). If it is assumed that the saturated soil water content is close to the effective porosity and we compare the results from all the fitted curves of the study (which had a similar value, generally close to 0.26), it can be concluded that most of the curves had a very limited water holding capacity where the field capacity was very close to the saturated soil water content. These results are supported by previous in situ soil water characteristic curves by Nuez-Elisea et al. (2001) were tensiome ters at 0 suction presented the similar s parameters found in this study at 100 cm of pressure hea d. This data could also help to discard the possibility of a bimodal porosity at lower suction rates. However, more information is required to determine if the soil water characteristic curv es would be better described using a bimodal

PAGE 91

91 porosity where more than one soil water inflect ion point is identifie d. For this condition, multimodal retention functions as described in deta il by Durner (1994) could be used. If this is the case, groundwater levels at closer depths wi ll be required to capture the gravel macro pore fraction of the soil water depletion. Finding soil water dynamics at closer DSM observations (< 0.1 m) would be required to capture this phenom enon, for the case of a dry season were drained to equilibrium conditions prevail these lengths are unusual. Considering that fruit tree root systems that gr ow in trenches of calcareous soils are located in the top 30 cm of the soil prof ile (Zekri et al., 1999) and that mo st of the roots concentrate in the top 10 cm layer (Nez-Elisea et al., 2000), it can be concluded that from all the fitted parameters in this study, lychee trenched curves at depths at 10 and 20 cm can contribute to assess irrigation practices based on capillary rise contribution. Predicting Soil Water Based on Groundw ater Level Observations Differences among soils in terms of retention and water movement depend to a large extent on pore size and shape distributions The interpretation of the drain to equilibrium data as a field soil water characteristic curve using van Ge nuchtens model (1980) adds meaning to the evaluation because the van Genuchten model is c onsidered an analytical expression with fitted hydraulic parameters that have a physical basi s. The models proposed for each sensor were determined by fitting the parameters shown in Tables 4-5, 4-6, 4-7 and 4-8 to Equation 4-3. The models capacity to explain the relationship between groundwater le vel and soil volumetric water content was estimated using the Ceff accuracy of predictions whic h was in general greater than the mean value of the observed data. Time series comparisons of the 96 models (Figures 4-9, 410 and 4-11) indicate that groundwater level has a greater influence on the soil water content of the deepest layers of the trench (40 and 60 cm) as compared to the shallower layers of the trench (10 and 20 cm). This is evident by observing the beginning of the time series between the months

PAGE 92

92 of November and December 2008. From this point, model predictions tend to underestimate soil water content compared to the predictions occurr ing near the end of January. During this period, soil water observations were not completely rela ted to groundwater fluc tuations as remaining moisture from precipitation and fr eeze protection events also influe nced soil water content. This situation is more evident in the 10 and 20 cm depths where greater exposure to evapotranspiration rates occurs. Simplifying results The results from this study have the potential to be converted into a practical approach to assess capillary rise of a shallow groundwater table. The vertical variability in the observed soil conditions required a different pr ediction model for different depths. To simplify the results found in this study, a fitted parameter from all data is presented at the end of each soil condition from Table 4-5 to Table 4-8. Pair-wise multiple comparisons among means of these fitted curves was completed with Duncan test ( = 0.05; Table 4-9). The resu lts grouped the categories generally by depth. Similarities were found between the soil c onditions of lychee trench and lychee no trench at 10 cm depth, likely due to the Krome scarifie d layer both categories have in common. Duncan groupings indicated that three of the four 20 cm depth conditions were not significantly different. The same result was found with the 60 cm depth. However, the 40 cm depth conditions for soil water content were show n to be significantly different. The use of Duncan means multiple comparisons suggest that different soil condi tions may use a common soil water characteristic curve for only the 20 and 60 cm depth. Difference observed at 10 and 40 cm depths may be due to differences in soil st ructure; however, additiona l work is needed to verify this.

PAGE 93

93 Applications of the model The drained to equilibrium interpretation of cap illary rise captured the m ain trends of soil water status according to water table depth. Howeve r, the model is not appropriate for estimating sub-daily soil water conditions. The methodology used in this study is applicab le to different soil types, elevations and locations which makes it a practical tool. To best parameterize a model, selection of a representative site for soil water content monitoring is critical and should consider surface elevation and soil structure variations. Thes e factors represent some of the most sensitive parameters influencing capillary co ntribution to plant water uptake. Contributions from groundwater to soil water content could be assessed using the model developed in this study. This contribution could be integrated into irrigation management through the following steps: 1. Calculate soil volume according to area and depth to be irrigated. 2. Multiply the soil volume by porosity to obtain water volume to be filled. 3. Use drained to equilibrium fitted paramete rs in van Genuchtens equation (1980) to estimate soil water status (pick one set of all data parameters according to the soil conditions). 4. Estimate the field capacity of the soil with drained to equilibrium model (assuming field capacity refers to average suction of 100 cm of water pressure head). 5. Measure water table depth and use it as refere nce to estimate soil water status with the drained to equilibrium model. 6. Subtract the soil water status according to the water table depth fr om the desired field capacity. 7. Multiply the difference of soil water to be supplemented with the soil volume to be irrigated. 8. Estimate the water to be applied per hectare. Implementing this model to assess irrigation requirements could help reduce standard irrigation. Considering that the average grow er applies about 37.9 L/ day, the drained to

PAGE 94

94 equilibrium procedure using the lychee trench parameters at 20 cm depth could potentially reduce irrigation rates according to the water table depth (Figure 4-12). From the soil water holding capacity perspective, irriga tion requirements would not be ne cessary if the water table is close to the root zone up to a depth of 200 cm. However, water requirements according to the phenological stage of the plant and daily ET is not considered for these purposes. If this is the case, given the low changes of soil water content in large changes of pr essure, low volume high frequency irrigation is recommended if water ta ble is between 100 and 200 cm depth were the bubbling pressure appears to break. The importance of the drained to equilibrium conditions In order to extend the m odels capacity of prediction beyond the dr ained to equilibrium conditions, selected parameters from the lychee tren ch soil conditions were tested to predict soil water content during the entire study period. The Ceff results were not as satisfactory as those reported in Tables 4-5 to 4-8. The accuracy of the predictions in all the depths was reduced underestimating observations because the sensor s were reporting remaining soil water from water inputs. This point highlights the importance of properly identifying drain to equilibrium for an accurate prediction. In many circumstances, the soil water holding characteristics make it difficult to capture the drain to equilibrium condition; the effect of hysteresis is di fficult to distinguish between the drying of wetting processes with fi eld data. Nonetheless, the aquifers high permeability and the soils low water holding capacity are driving factors that allow the drained to equilibrium conditions to be obtained. As a result, the re lationship found between soil water content and groundwater level using this simple and practic al one-dimensional model is maintained when hydrostatic assumptions are valid. In this study, this corresponds to periods when water table is declining and low rainfall is reported, coincidi ng with the dry season in South Florida. This

PAGE 95

95 season corresponds to the period wh en capillary rise has the poten tial to supply a part of the irrigation requirements, or if in excess damage the crop roots (unlikely to happen during the dry season at current levels at the study site). The simple model presented allows the determination of possible soil water content with groundwater level alteration. It is important to understand that the values of the parameters obtained from rete ntion models are based on field observations of soil water content and pressure he ad. Model validity can only be assumed within the limits of the observed results (DSM range). Extrapolation of model values outside the range of input (DSM) are not reliable and should only be used as an indication of the trends. To extend the model, future research should focus on gathering field da ta for conditions when the groundwater level is closer to the installed sensors or to perform studies in a laboratory setting where conditions can be controlled. Finally, our results cl early indicate that groundwater le vel is an important factor to consider for explaining the soil water content variati on in this area and that th is effect needs to be considered in CERP management alternatives. Conclusions Soil water content is strongl y related to the shallow groundwater level. Hydrostatic assum ptions can be used to investigate capillar y rise in shallow groundwater table conditions. This was successfully done by measuring soil water content and groundwater level and fitting data using the van Genuchten (1980) equation. Hori zontal and vertical va riations in the data required that each sensor be described using a different model. Field capacity was found to be very close to the saturated water content parameter confirming the low water holding capacity of the soils. Using groundwater level as a reference to pred ict soil water content is possible and can be considered as a simple and useful one dimensional model technique. Accuracy of predictions had an overall Ceff 0.72 and responded better at greater dept hs. The proposed model was able to

PAGE 96

96 capture the general and most representative tren ds of soil water content changes in response to the shallow groundwater fluctuati ons. Applications of this approach to assess the soil water status using water table depth as a reference seemed to satisfy the soil water requirements when the water table depth range s are less than 200 cm of the reference point.

PAGE 97

97 Table 4-1. Monitoring well locations and elevation specifications. Station name ID Latitude N Longitude W Well diameter (m) Well depth to surface (m) Riser elevation NGVD1 1929 (m) Surface elevation NGVD 1929 (m) Reference above benchmark (m) (decimal degrees) Well 1 W1 25.5105 80.4990 0.051 8.530 3.106 3.149 0.043 Well 2 W2 25.5101 80.4990 0.051 8.530 2.941 2.954 0.013 Well 3 W3 25.5105 80.4993 0.051 8.530 3.164 3.228 0.064 Well 4 W4 25.5101 80.4993 0.051 8.530 3.057 3.057 0.000 1 NGVD: National Geodetic Vertical Datum. Table 4-2. Soil water sensors categorized in 16 different soil conditions. Soil conditions Depth (cm) Number of sensorsProbes Lychee Trench 10 8 A1 to A8 Lychee Trench 20 8 A1 to A8 Lychee Trench 40 8 A1 to A8 Lychee Trench 60 8 A1 to A8 Lychee No Trench 10 8 B1 to B8 Lychee No Trench 20 8 B1 to B8 Lychee No Trench 40 8 B1 to B8 Lychee No Trench 60 8 B1 to B8 Trench 10 4 A9 to A12 Trench 20 4 A9 to A12 Trench 40 4 A9 to A12 Trench 60 4 A9 to A12 No Trench 10 4 B9 to B12 No Trench 20 4 B9 to B12 No Trench 40 4 B9 to B12 No Trench 60 4 B9 to B12

PAGE 98

98 Table 4-3. Summary of all and filtered volumetric soil water c ontent daily means at each soil condition from 16 September 2008 to 23 March 2009. Soil conditions Depth (cm) (m3/m3) All (m3/m3) Filtered Mean Std Dev MinimumMaximumMean Std Dev MinimumMaximum Lychee Trench 10 0.2340.0800.0510.3660.2240.0880.0510.364 Lychee Trench 20 0.2040.0450.0690.2780.2000.0460.0790.276Lychee Trench 40 0.2140.0310.0660.2610.2150.0250.1140.258Lychee Trench 60 0.2190.0150.1960.2600.2170.0140.1960.257Lychee No Trench 10 0.2560.0330.2060.3400.2540.0330.2090.337 Lychee No Trench 20 0.2060.0270.1290.2720.2040.0260.1460.269Lychee No Trench 40 0.2120.0190.1670.2620.2110.0190.1670.255Lychee No Trench 60 0.2180.0300.1440.2660.2180.0290.1440.260Trench 10 0.2400.0320.1310.2950.2380.0300.1660.290 Trench 20 0.2360.0210.1820.2800.2350.0200.1980.278Trench 40 0.2400.0210.1950.2830.2410.0200.2030.281Trench 60 0.2300.0140.1980.2640.2300.0130.2010.259No Trench 10 0.1970.0440.0210.3090.1890.0390.0220.300 No Trench 20 0.1720.0390.0090.2480.1650.0310.0320.242No Trench 40 0.1890.0380.1080.2640.1850.0380.1330.245No Trench 60 0.2270.0200.1950.2740.2250.0210.1950.261

PAGE 99

99 Table 4-4. Summary of entire dataset of groundw ater daily means and comparison with USGS well 196A historical database. Month Year Well ID Surface Elevation NGVD 1929 (m) Water Table Elevation NGVD 1929 (m) Mean Std Dev Minimum Maximum Median September 2008 W1 3.15 1.14 0.06 1.01 1.39 1.12 September 2008 W2 2.95 1.14 0.06 1.02 1.38 1.13 September 2008 W3 3.23 1.20 0.06 1.08 1.43 1.18 September 2008 W4 3.06 . September 2008 USGS 196A 3.15 1.14 1.06 1.36 September 1950-2008 USGS 196A 3.15 0.72 2.15 1.28 October 2008 W1 3.15 1.31 0.09 1.16 1.57 1.29 October 2008 W2 2.95 1.30 0.09 1.15 1.61 1.29 October 2008 W3 3.23 1.37 0.09 1.23 1.64 1.36 October 2008 W4 3.06 1.32 0.08 1.21 1.60 1.31 October 2008 USGS 196A 3.15 1.33 1.22 1.60 October 1950-2008 USGS 196A 3.15 0.88 1.83 1.24 November 2008 W1 3.15 1.15 0.03 1.06 1.21 1.15 November 2008 W2 2.95 1.09 0.05 0.91 1.21 1.09 November 2008 W3 3.23 1.19 0.03 1.13 1.25 1.19 November 2008 W4 3.06 1.14 0.06 0.96 1.29 1.14 November 2008 USGS 196A 3.15 1.15 1.08 1.21 November 1950-2008 USGS 196A 3.15 0.70 1.89 1.05 December 2008 W1 3.15 1.07 0.02 1.02 1.15 1.06 December 2008 W2 2.95 1.04 0.03 0.95 1.12 1.04 December 2008 W3 3.23 1.13 0.02 1.08 1.18 1.12 December 2008 W4 3.06 1.08 0.03 0.99 1.16 1.09 December 2008 USGS 196A 3.15 1.06 1.04 1.09 December 1950-2008 USGS 196A 3.15 0.43 1.32 0.96 January 2009 W1 3.15 1.02 0.03 0.91 1.06 1.03 January 2009 W2 2.95 0.96 0.04 0.85 1.04 0.96 January 2009 W3 3.23 1.08 0.03 1.00 1.12 1.09 January 2009 W4 3.06 1.02 0.04 0.88 1.12 1.02 January 2008 USGS 196A 3.15 1.01 0.94 1.05 January 1950-2008 USGS 196A 3.15 0.30 1.39 0.89 February 2009 W1 3.15 0.87 0.02 0.81 0.94 0.87 February 2009 W2 2.95 0.83 0.06 0.72 0.97 0.80 February 2009 W3 3.23 0.95 0.04 0.85 1.04 0.94 February 2009 W4 3.06 0.86 0.07 0.73 1.06 0.84 February 2009 USGS 196A 3.15 0.87 0.83 0.93 February 1950-2008 USGS 196A 3.15 0.23 1.19 0.83 March 2009 W1 3.15 0.84 0.08 0.74 1.05 0.81 March 2009 W2 2.95 0.80 0.08 0.69 1.02 0.76 March 2009 W3 3.23 0.89 0.08 0.79 1.11 0.85 March 2009 W4 3.06 0.83 0.09 0.70 1.07 0.80 March 2009 USGS 196A 3.15 0.87 0.74 1.02 March 1950-2008 USGS 196A 3.15 0.07 1.15 0.82

PAGE 100

100 Table 4-5. Drained to equilibrium fitted paramete rs of van Genuchtens model (1980) used to describe field soil water characteristic cu rves at 10 cm depth of Krome soil in a lychee orchard at four different soil conditions. Soil conditions Probe ID s a r b c nd me r2 f Ceff g Lychee Trench A1-10 0.257 0.000 0.005 14.987 0.933 0.786 0.781 Lychee Trench A2-10 0.229 0.000 0.004 5.950 0.832 0.834 0.834 Lychee Trench A3-10 0.405 0.000 0.004 3.570 0.720 0.645 0.646 Lychee Trench A4-10 0.285 0.000 0.005 14.973 0.933 0.819 0.819 Lychee Trench A5-10 0.266 0.000 0.005 43.027 0.977 0.887 0.887 Lychee Trench A6-10 0.274 0.000 0.005 29.074 0.966 0.734 0.734 Lychee Trench A7-10 0.373 0.000 0.004 9.791 0.898 0.546 0.546 Lychee Trench A8-10 0.251 0.000 0.005 56.777 0.982 0.733 0.733 Lychee Trench All data* 0.286 0.000 0.005 21.122 0.953 0.811 0.809 Trench A9-10 0.235 0.000 0.005 16.767 0.940 0.823 0.823 Trench A10-10 0.299 0.000 0.004 9.620 0.896 0.759 0.759 Trench A11-10 0.351 0.000 0.004 2.320 0.569 0.759 0.753 Trench A12-10 0.260 0.000 0.005 15.511 0.936 0.780 0.779 Trench All data* 0.270 0.000 0.005 12.582 0.921 0.797 0.497 Lychee No Trench B1-10 0.261 0.000 0.005 11.926 0.916 0.677 0.677 Lychee No Trench B2-10 0.272 0.000 0.004 5.761 0.826 0.637 0.638 Lychee No Trench B3-10 0.269 0.000 0.005 13.090 0.924 0.795 0.794 Lychee No Trench B4-10 0.293 0.000 0.004 10.455 0.904 0.799 0.793 Lychee No Trench B5-10 0.406 0.000 0.006 2.012 0.503 0.690 0.689 Lychee No Trench B6-10 0.393 0.039 0.009 1.251 0.201 0.168 0.169 Lychee No Trench B7-10 0.298 0.000 0.004 4.469 0.776 0.743 0.744 Lychee No Trench B8-10 0.271 0.000 0.004 7.501 0.867 0.704 0.704 Lychee No Trench All data* 0.359 0.000 0.005 2.457 0.593 0.631 0.631 No Trench B9-10 0.304 0.000 0.005 4.582 0.782 0.891 0.891 No Trench B10-10 0.383 0.000 0.005 1.831 0.454 0.650 0.649 No Trench B11-10 0.233 0.000 0.005 7.251 0.862 0.710 0.710 No Trench B12-10 2.537 0.000 0.115 1.804 0.446 0.788 0.788 No Trench All data* 0.352 0.000 0.006 4.521 0.779 0.848 0.848 a Saturated water content (s). b Residual water content (r). c Inverse of the air entry value ( ). d Dimensionless measure of por e size distribution (slope) (n). e Curve shape parameter ( m=1-1/n ) (m ). f Coefficient of determination of observed values and fitted curve ( r2). g Nash-Sutcliffe (1970) coeffi cient of efficiency of observed and predicted values ( Ceff). Fitted parameters considering all daily mean values of the soil condition.

PAGE 101

101 Table 4-6. Drained to equilibrium fitted paramete rs of van Genuchtens model (1980) used to describe field soil water characteristic cu rves at 20 cm depth of Krome soil in a lychee orchard at four different soil conditions. Soil conditions Probe ID s a r b c nd me r2 f Ceff g Lychee Trench A1-20 0.267 0.000 0.005 10.077 0.901 0.796 0.796 Lychee Trench A2-20 0.227 0.000 0.005 7.087 0.859 0.844 0.844 Lychee Trench A3-20 0.281 0.000 0.005 11.013 0.909 0.854 0.854 Lychee Trench A4-20 0.260 0.000 0.005 12.604 0.921 0.836 0.836 Lychee Trench A5-20 0.278 0.000 0.005 31.136 0.968 0.864 0.863 Lychee Trench A6-20 0.253 0.000 0.006 25.541 0.961 0.806 0.806 Lychee Trench A7-20 0.256 0.000 0.005 18.278 0.945 0.837 0.838 Lychee Trench A8-20 0.234 0.000 0.005 19.841 0.950 0.607 0.607 Lychee Trench All data* 0.256 0.000 0.005 16.093 0.938 0.836 0.836 Trench A9-20 0.256 0.000 0.005 12.603 0.921 0.852 0.852 Trench A10-20 0.261 0.000 0.005 7.950 0.874 0.901 0.901 Trench A11-20 0.341 0.000 0.005 2.021 0.505 0.856 0.857 Trench A12-20 0.625 0.000 0.017 1.750 0.429 0.697 0.697 Trench All data* 0.266 0.000 0.005 10.339 0.903 0.872 0.872 Lychee No Trench B1-20 0.219 0.000 0.005 12.133 0.918 0.747 0.746 Lychee No Trench B2-20 0.220 0.000 0.005 7.582 0.868 0.775 0.775 Lychee No Trench B3-20 0.256 0.000 0.005 7.928 0.874 0.762 0.762 Lychee No Trench B4-20 0.276 0.000 0.005 9.195 0.891 0.836 0.835 Lychee No Trench B5-20 0.260 0.000 0.005 8.006 0.875 0.763 0.762 Lychee No Trench B6-20 0.258 0.000 0.005 6.272 0.841 0.741 0.741 Lychee No Trench B7-20 0.247 0.000 0.004 7.254 0.862 0.763 0.763 Lychee No Trench B8-20 0.250 0.000 0.005 6.651 0.850 0.772 0.771 Lychee No Trench All data* 0.254 0.000 0.005 6.281 0.841 0.741 0.740 No Trench B9-20 0.301 0.000 0.006 4.971 0.799 0.899 0.898 No Trench B10-20 0.328 0.000 0.007 1.900 0.474 0.524 0.522 No Trench B11-20 0.313 0.000 0.006 3.428 0.708 0.825 0.825 No Trench B12-20 0.401 0.000 0.009 2.536 0.606 0.832 0.832 No Trench All data* 0.290 0.000 0.006 4.883 0.795 0.828 0.825 a Saturated water content (s). b Residual water content (r). c Inverse of the air entry value ( ). d Dimensionless measure of por e size distribution (slope) (n). e Curve shape parameter ( m=1-1/n ) (m ). f Coefficient of determination of observed values and fitted curve ( r2). g Nash-Sutcliffe (1970) coeffi cient of efficiency of observed and predicted values ( Ceff). Fitted parameters considering all daily mean values of the soil condition.

PAGE 102

102 Table 4-7. Drained to equilibrium fitted paramete rs of van Genuchtens model (1980) used to describe field soil water characteristic cu rves at 40 cm depth of Krome soil in a lychee orchard at four different soil conditions. Soil conditions Probe ID s a r b c nd me r2 f Ceff g Lychee Trench A1-40 0.268 0.000 0.006 10.078 0.901 0.775 0.775 Lychee Trench A2-40 0.270 0.000 0.005 4.587 0.782 0.822 0.822 Lychee Trench A3-40 0.260 0.000 0.005 11.043 0.909 0.866 0.866 Lychee Trench A4-40 0.335 0.000 0.006 2.185 0.542 0.798 0.799 Lychee Trench A5-40 0.253 0.000 0.005 10.057 0.901 0.825 0.825 Lychee Trench A6-40 0.238 0.000 0.006 25.638 0.961 0.872 0.872 Lychee Trench A7-40 0.244 0.000 0.005 10.726 0.907 0.879 0.879 Lychee Trench A8-40 0.314 0.000 0.005 2.567 0.610 0.877 0.877 Lychee Trench All data* 0.256 0.000 0.005 10.808 0.907 0.869 0.869 Trench A9-40 0.245 0.000 0.005 15.336 0.935 0.820 0.819 Trench A10-40 0.246 0.000 0.005 10.617 0.906 0.908 0.907 Trench A11-40 0.342 0.000 0.006 1.584 0.369 0.880 0.881 Trench A12-40 0.261 0.000 0.005 14.539 0.931 0.872 0.874 Trench All data* 0.272 0.000 0.005 5.365 0.814 0.843 0.843 Lychee No Trench B1-40 0.280 0.000 0.007 1.679 0.404 0.798 0.799 Lychee No Trench B2-40 0.253 0.000 0.005 5.863 0.829 0.890 0.890 Lychee No Trench B3-40 0.242 0.000 0.005 6.062 0.835 0.880 0.880 Lychee No Trench B4-40 0.331 0.000 0.006 2.040 0.510 0.881 0.880 Lychee No Trench B5-40 0.307 0.000 0.006 2.207 0.547 0.835 0.835 Lychee No Trench B6-40 0.307 0.000 0.007 2.224 0.550 0.869 0.869 Lychee No Trench B7-40 0.324 0.000 0.007 1.755 0.430 0.871 0.871 Lychee No Trench B8-40 0.218 0.166 0.008 8.318 0.880 0.658 0.657 Lychee No Trench All data* 0.306 0.000 0.007 2.045 0.511 0.862 0.862 No Trench B9-40 0.207 0.000 0.006 2.827 0.646 0.728 0.728 No Trench B10-40 0.287 0.000 0.009 1.539 0.350 0.553 0.553 No Trench B11-40 0.465 0.000 0.387 1.169 0.144 0.169 0.436 No Trench B12-40 0.275 0.160 0.011 3.907 0.744 0.786 0.882 No Trench All data* 0.375 0.163 0.027 2.048 0.512 0.798 0.796 a Saturated water content (s). b Residual water content (r). c Inverse of the air entry value ( ). d Dimensionless measure of por e size distribution (slope) (n). e Curve shape parameter ( m=1-1/n ) (m ). f Coefficient of determination of observed values and fitted curve ( r2). g Nash-Sutcliffe (1970) coeffi cient of efficiency of observed and predicted values ( Ceff). Fitted parameters considering all daily mean values of the soil condition.

PAGE 103

103 Table 4-8. Drained to equilibrium fitted paramete rs of van Genuchtens model (1980) used to describe field soil water characteristic cu rves at 60 cm depth of Krome soil in a lychee orchard at four different soil conditions. Soils conditions Probe ID s a r b c nd me r2 f Ceff g Lychee Trench A1-60 0.263 0.000 0.006 5.845 0.829 0.830 0.830 Lychee Trench A2-60 0.228 0.209 0.008 41.366 0.976 0.753 0.754 Lychee Trench A3-60 0.331 0.000 0.008 1.735 0.424 0.753 0.754 Lychee Trench A4-60 0.318 0.000 0.007 2.009 0.502 0.839 0.839 Lychee Trench A5-60 0.301 0.000 0.007 2.390 0.582 0.832 0.832 Lychee Trench A6-60 0.466 0.099 0.020 2.198 0.545 0.895 0.895 Lychee Trench A7-60 0.249 0.000 0.006 5.856 0.829 0.903 0.903 Lychee Trench A8-60 0.326 0.000 0.008 1.879 0.468 0.847 0.848 Lychee Trench All data* 0.313 0.000 0.007 2.115 0.527 0.870 0.870 Trench A9-60 0.295 0.000 0.007 2.236 0.553 0.807 0.807 Trench A10-60 0.280 0.000 0.005 1.776 0.437 0.471 0.777 Trench A11-60 0.318 0.000 0.009 1.551 0.355 0.918 0.918 Trench A12-60 0.337 0.000 0.008 1.679 0.404 0.842 0.842 Trench All data* 0.325 0.000 0.009 1.630 0.386 0.856 0.856 Lychee No Trench B1-60 0.513 0.216 0.024 5.889 0.830 0.024 -2.391 Lychee No Trench B2-60 0.285 0.000 0.008 1.709 0.415 0.936 0.936 Lychee No Trench B3-60 0.252 0.000 0.012 1.708 0.415 0.812 0.812 Lychee No Trench B4-60 0.308 0.000 0.007 1.534 0.348 0.948 0.949 Lychee No Trench B5-60 0.349 0.024 0.037 1.266 0.210 0.923 0.922 Lychee No Trench B6-60 0.288 0.108 0.109 1.173 0.147 0.147 0.397 Lychee No Trench B7-60 0.271 0.000 0.007 1.336 0.252 0.560 0.766 Lychee No Trench B8-60 0.428 0.218 0.026 2.523 0.604 0.417 0.416 Lychee No Trench All data* 0.268 0.000 0.006 1.514 0.339 0.757 0.757 No Trench B9-60 0.288 0.000 0.012 1.350 0.259 0.529 0.528 No Trench B10-60 0.239 0.080 0.073 1.092 0.084 0.052 0.052 No Trench B11-60 0.237 0.361 0.004 1.971 0.493 0.049 0.049 No Trench B12-60 0.193 0.226 1.244 1.122 0.109 0.001 -0.028 No Trench All data* 0.270 0.086 0.027 1.130 0.115 0.281 0.281 a Saturated water content (s). b Residual water content (r). c Inverse of the air entry value ( ). d Dimensionless measure of por e size distribution (slope) (n). e Curve shape parameter ( m=1-1/n ) (m ). f Coefficient of determination of observed values and fitted curve ( r2). g Nash-Sutcliffe (1970) coeffi cient of efficiency of observed and predicted values ( Ceff). Fitted parameters considering all daily mean values of the soil condition.

PAGE 104

104 Table 4-9. Duncan mean multiple comparison of soil water characteristic curves at each soil condition. Duncan Grouping mean (m3/m3) Soil condition Depth (cm) A 0.276 Lychee No Trench10 B A 0.269 Lychee Trench 10 B C 0.262 Trench 10 B C 0.257 Trench 20 C 0.253 No Trench 10 D 0.229 Trench 40 D 0.225 Lychee Trench 20 E D 0.223 Trench 60 E D 0.222 Lychee No Trench20 E D 0.222 No Trench 60 E D 0.217 No Trench 20 E D 0.215 Lychee No Trench40 E D 0.215 Lychee No Trench60 E 0.210 Lychee Trench 40 F 0.194 No Trench 40 F 0.189 Lychee Trench 60

PAGE 105

105 Figure 4-1. Soil water potential in an unsaturat ed soil column at drained to equilibrium conditions. H=hydraulic potential h=pressure potential, z=gravimetric potential, L=water table elevation Figure 4-2. Side view schematic of soil conditions studied using monitoring wells and multisensor capacitance probes in the Krome scarified (gray) and limestone bedrock (light) soil profile.

PAGE 106

106 Figure 4-3. Schematic of experimental site at University of Florida Tropical Research and Education Center (TREC), Homestead, FL A symbolized probes located in a trench Krome soil condition and B symbo lizes probes located in a non trenched limestone soil condition.

PAGE 107

107 Figure 4-4. Schematic of monitoring well geom etry and Levelogger location for monitoring groundwater levels.

PAGE 108

108 Sep Oct Nov Dec Jan Feb Mar Apr Rainfall (mm) 0 2 4 6 8 10 12 14 16 18 Soil moisture (m 3 /m 3 ) 0 5 10 15 20 25 30 WTE NGVD 1929 (m) 0.6 0.8 1.0 1.2 1.4 1.6 1.8 Rainfall Sensor 5 (10 cm) Sensor 6 (10 cm) Sensor 8 (10 cm) Mean WTE (a) Sep Oct Nov Dec Jan Feb Mar Apr Rainfall (mm) 0 2 4 6 8 10 12 14 16 18 Soil moisture (m3/m3) 0 5 10 15 20 25 30 WTE NGVD 1929 (m) 0.6 0.8 1.0 1.2 1.4 1.6 1.8 Rainfall Sensor 1 (20 cm) Sensor 7 (20 cm) Sensor 8 (20 cm) Mean WTE (b) Sep Oct Nov Dec Jan Feb Mar Apr Rainfall (mm) 0 2 4 6 8 10 12 14 16 18 Soil moisture (m3/m3) 12 14 16 18 20 22 24 26 28 WTE NGVD 1929 (m) 0.8 1.0 1.2 1.4 1.6 1.8 Rainfall Sensor 1 (60 cm) Sensor 4 (60 cm) Sensor 5 (60 cm) Sensor 7 (60 cm) Sensor 8 (60 cm) Mean WTE (c) Sep Oct Nov Dec Jan Feb Mar Apr Rainfall (mm) 0 2 4 6 8 10 12 14 16 18 Soil moisture (m3/m3) 12 14 16 18 20 22 24 26 28 WTE NGVD 1929 (m) 0.6 0.8 1.0 1.2 1.4 1.6 1.8 Rainfall Sensor 1 (40 cm) Sensor 2 (40 cm) Sensor 3 (40 cm) Sensor 5 (40 cm) Sensor 7 (40 cm) Mean WTE (d) Figure 4-5. Soil water response to groundwater fluctuations from September 2008 to March 2009 in Lychee trench selected sensors at (a) 10 cm, (b) 20 cm, (c) 40 cm and (d) 60 cm depths.

PAGE 109

109 Lychee Trench Probe A1 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Lychee No Trench Probe B1 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Lychee Trench Probe A2 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Lychee No Trench Probe B2 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Lychee Trench Probe A3 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Lychee No Trench Probe B3 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Lychee Trench Probe A4 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Lychee No Trench Probe B4 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Figure 4-6. Soil water characterist ic curves for probes A1 to A4 and B1 to B4 at 10, 20, 40 and 60 cm depths.

PAGE 110

110 Lychee Trench Probe A5 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Lychee No Trench Probe B5 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Lychee Trench Probe A6 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Lychee No Trench Probe B6 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Lychee Trench Probe A7 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Lychee No Trench Probe B7 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Lychee Trench Probe A8 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Lychee No Trench Probe B8 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Figure 4-7. Soil water characterist ic curves for probes A5 to A8 and B5 to B8 at 10, 20, 40 and 60 cm depths.

PAGE 111

111 Trench Probe A9 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm No Trench Probe B9 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Trench Probe A10 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm No Trench Probe B10 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Trench Probe A11 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm No Trench Probe B11 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Trench Probe A12 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm No Trench Probe B12 Pressure head (cm) 100 150 200 250 300 (m3/m3) 0.0 0.1 0.2 0.3 0.4 10 cm 20 cm 40 cm 60 cm Curve 10 cm Curve 20 cm Curve 40 cm Curve 60 cm Figure 4-8. Soil water characteristic curves for pr obes A9 to A12 and B9 to B12 at 10, 20, 40 and 60 cm depths.

PAGE 112

112 Lychee Trench Probe A1 10/1/08 11/1/08 12/1/08 1/1/09 (m3/m3) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 ( m ) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Lychee No Trench Probe B1 10/1/08 11/1/08 12/1/08 1/1/09 Rainfall (mm) 0 20 40 60 80 100 (m3/m3) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Lychee Trench Probe A2 10/1/08 11/1/08 12/1/08 1/1/09 (m3/m3) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 ( m ) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Lychee No Trench Probe B2 10/1/08 11/1/08 12/1/08 1/1/09 Rainfall (mm) 0 20 40 60 80 100 (m3/m3) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Lychee Trench Probe A3 10/1/08 11/1/08 12/1/08 1/1/09 (m3/m3) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 ( m ) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Lychee No Trench Probe B3 10/1/08 11/1/08 12/1/08 1/1/09 Rainfall (mm) 0 20 40 60 80 100 (m3/m3) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Lychee Trench Probe A4 10/1/08 11/1/08 12/1/08 1/1/09 (m3/m3) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 ( m ) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Lychee No Trench Probe B4 10/1/08 11/1/08 12/1/08 1/1/09 Rainfall (mm) 0 20 40 60 80 100 (m3/m3) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Figure 4-9. Soil water prediction us ing groundwater level as referen ce with the fitted parameters of van Genuchten model (1980) for probes A1 to A4 and B1 to B4 at 10, 20, 40 and 60 cm depths.

PAGE 113

113 Lychee Trench Probe A5 10/1/08 11/1/08 12/1/08 1/1/09 (m 3 /m 3 ) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Lychee No Trench Probe B5 10/1/08 11/1/08 12/1/08 1/1/09 Rainfall (mm) 0 20 40 60 80 100 (m3/m3) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Lychee Trench Probe A6 10/1/08 11/1/08 12/1/08 1/1/09 (m 3 /m 3 ) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Lychee No Trench Probe B6 10/1/08 11/1/08 12/1/08 1/1/09 Rainfall (mm) 0 20 40 60 80 100 (m3/m3) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Lychee Trench Probe A7 10/1/08 11/1/08 12/1/08 1/1/09 (m 3 /m 3 ) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Lychee No Trench Probe B7 10/1/08 11/1/08 12/1/08 1/1/09 Rainfall (mm) 0 20 40 60 80 100 (m3/m3) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Lychee Trench Probe A8 10/1/08 11/1/08 12/1/08 1/1/09 (m 3 /m 3 ) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Lychee No Trench Probe B8 10/1/08 11/1/08 12/1/08 1/1/09 Rainfall (mm) 0 20 40 60 80 100 (m3/m3) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Figure 4-10. Soil water predicti on using groundwater level as reference with the fitted parameters of van Genuchten model (1980) fo r probes A5 to A8 and B5 to B8 at 10, 20, 40 and 60 cm depths.

PAGE 114

114 Trench Probe A9 10/1/08 11/1/08 12/1/08 1/1/09 (m 3 /m 3 ) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE No Trench Probe B9 10/1/08 11/1/08 12/1/08 1/1/09 Rainfall (mm) 0 20 40 60 80 100 (m3/m3) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Trench Probe A10 10/1/08 11/1/08 12/1/08 1/1/09 (m 3 /m 3 ) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE No Trench Probe B10 10/1/08 11/1/08 12/1/08 1/1/09 Rainfall (mm) 0 20 40 60 80 100 (m3/m3) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Trench Probe A11 10/1/08 11/1/08 12/1/08 1/1/09 (m 3 /m 3 ) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE No Trench Probe B11 10/1/08 11/1/08 12/1/08 1/1/09 Rainfall (mm) 0 20 40 60 80 100 (m3/m3) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Trench Probe A12 10/1/08 11/1/08 12/1/08 1/1/09 (m 3 /m 3 ) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE No Trench Probe B12 10/1/08 11/1/08 12/1/08 1/1/09 Rainfall (mm) 0 20 40 60 80 100 (m3/m3) 0.0 0.1 0.2 0.3 0.4 WTE NGVD 1929 (m) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Rainfall Obs (10 cm) Obs (20 cm) Obs (40 cm) Obs (60 cm) Pred (10 cm) Pred (20 cm) Pred (40 cm) Pred (60 cm) WTE Figure 4-11. Soil water predicti on using groundwater level as reference with the fitted parameters of van Genuchten model (1980) for probes A9 to A12 and B9 to B12 at 10, 20, 40 and 60 cm depths.

PAGE 115

115 Figure 4-12. Estimation of water savings base d on drained to equilibrium assessment of irrigation requirements. Drained to equilibri um model assumes field capacity at 100 H2O cm of suction and compares supplemen ted irrigation with average grower (10.8 m3/ha). 0% 20% 40% 60% 80% 100% 120% 0100200300400500Water Savings (%)Water Table Depth (cm)

PAGE 116

116 CHAPTER 5 EXECUTIVE SUMMARY Quantifying the relationship be tween groundwater table level and root zone soil water content in shallow groundwater cond itions is important in South Florida due to the potential changes in groundwater level related to the Comp rehensive Everglades Restoration Plan (CERP) and the resulting impacts these changes may have on deep rooted agriculture. The topic requires long and comprehensive research in multiple disciplines. This study represents a small contribution through three differe nt objectives: soil water instru ment calibration, methods of capillary rise interpretation and modeling the uns aturated zone of the soil profile. Results are intended to provide growers with science-based information on the benefits and challenges of the shallow groundwater table in this area. Objective 1 Laborato ry soil water characteristic curves helped to identify similarities in soil water retention patterns and low wate r holding capacity prope rties of gravel and limestone bedrock, attributing Krome soils water holdi ng capacity to the loam fracti on of the soil. Although gravel and limestone retention patterns were low, the soil water depletion was characterized to have very low volumetric water content changes in large increments of suction. Onsite calibration of multi sensor capacitance probes using synchronized suction values from tensiometers as a reference to determine volumetric water content is a promising technique for conditions were standard calibration procedures are not possible. However limestone heterogeneity limited the results to an unpractical site specific calibrati on. Regression equations for instrument calibration in limes tone bedrock conditions were ch aracterized by their spatial and depth variability. As a result four regression m odels were proposed for calibration purposes. Relationships between Al-Yahyai et al. (2006) calibration regre ssion model of Krome soils and

PAGE 117

117 the proposed models were similar at equilibriu m conditions, confirming the possibility to use a single calibration equation in Krome scarified and limestone bedrock Objective 2 Circular statistics repres ent a practical approach to interpre t diurnal fluctuations of soil water content and groundwater level. Locations of statistically significant mean vectors were confirmed in all instruments a nd variables with Rayleigh test although many of the vectors were influenced by multimodal distributions attributed to multivariate effects. Mean vectors of soil water content at deeper depths were more related to mean vectors of groundwater; mean vectors of soil water content at shallower depths were more related to sola r radiation and soil temperature. Circular-circular correlation analysis confirm in general a weak relationship between diurnal peaks of soil water content a nd the peaks found in groundwater and weather variables. Objective 3 A hydrostatic m odel based on the drained to equilibrium principle was proposed to use groundwater level as a reference to predict soil water content. Pred ictions indicated this approach can be considered as a simple and useful re ference of one-dimensiona l model technique. The proposed model was able to capture the general and most representative trends of soil water content changes in response to the shallow groundwater fluc tuations although accuracy of predictions responded better at gr eater depths. Results have the potential to improve irrigation practices by considering groundwater contribution to the soil wate r status in the unsaturated zone.

PAGE 118

118 LIST OF REFERENCES Abtew, W 1996. Evapotranspiration measuremen ts and modeling for three wetland systems in South Florida. Water resources bulletin 32: 465. Ali, A., and Abtew. W. 1999. Regional rainfall fre quency analysis for central and South Florida. hydrologic reporting unit, South Florida, We st Palm Beach, FL.: Water management district water resources evaluation. Al-Yahyai, R., B. Schaffer, F.S. Davies, and J.H. Crane. 2005. Four levels of soil water depletion minimally affect carambola phenological cycles Horttechnology 15: 623-630. Al-Yahyai, R., B. Schaffer, F.S. Davies, a nd R. Munoz-Carpena. 2006. Characterization of soilwater retention of a very gravelly loam soil varied with determination method Soil Science 171: 85-93. Ayars, J.E., and R.A. Schoneman. 1986. Use of saline water from a shallow water table by cotton Transactions of the ASAE 29: 1674-1678. Babajimopoulos, C., A. Panoras, H. Georgoussis, G. Arampatzis, E. Hatzigiannakis, and D. Papamichail. 2007. Contribution to irrigati on from shallow water table under field conditions. Agricultural Water Management 92: 205-210. Batschelet, E. 1981. Circular Statistics in Biology. New York: Academic Press. Bauer, P., G. Thabeng, F. Stauffer, a nd W. Kinzelbach. 2004. Estimation of the evapotranspiration rate from di urnal groundwater level fluctuat ions in the Okavango Delta, Botswana Journal of Hydrology 288: 344. Beckett and Webster. 1971. Soil variability: a review Soils and Fertilizers. 34: 1. Beran, R.J. 1969. Asymptotic theory of a class of tests for uniformity of a circular distribution. Annals of Mathematical Statistics 40: 1196-1206. Berger, E. 1976. Partitioning the pa rameters of stony soils, important in moisture determinations into their constituents Plant and Soil 44: 201-207. Black, A. and A. Werritty. 1997. Seasonality of flooding: a case study in North Britain. Journal of Hydrology 195: 1-25. Bouyoucos, G. T. 1915. Effect of temperature on the movement of water vapor and capillary moisture in soil. Journal of Agricultural Research 5: 141-172. Bruce, R.R., and R.J. Luxmoore. 2003. Water Retention: Field Methods. In Methods of Soil Analysis, Soil Science Societ y of America Book Series 5: 663-686. Buckingham, E. 1907. Studies on the movement of soil moisture Govt. print. off., Washington.

PAGE 119

119 Buss, P. 1993. The use of capacitance based meas urements of real time soil water profile dynamics for irrigation scheduling. In National Conference of the Ir rigation, Australia and the National Committee on Irrigation and Drainage, Launceston, Tasmania. Chin, D.A., and R.D. Patterson. 2005. Quantificat ion of hydrologic processes and assessment of rainfall-runoff models in Miam i-Dade County, Florida. Information Services, Reston, VA, Denver, CO.: U.S. Geological Survey. Chin, D.A. 2008. Phenomenological models of hydrologic processes in South Florida Journal of Hydrology 349: 230-243. Coile, T.S. 1953. Moisture Content of Small Stone in Soil. Soil Science 75(3):203-208. Colburn, B., and S. Goldweber. 1961. Preparatio n of oolitic limestone soil for agricultural Proceedings of the Florida State Horticultural Society 74: 343-345. Crane, J. 2008. Tropical fruit crop production and research; HOS 5555 course notes, pp. 267. University of Florida. Crane, J. H., and B. Schaffer. 2004. Increased ex posure to cool ambient temperatures increased yield of 'Mauritius' lychee ( Litchi chinensis ) in Homestead Florida. In Proceedings of the Florida State Horticultural Society 117: 206-208. Crane, J., C. Balerdi, R. Campbell, C. Ca mpbell, and S. Goldweber. 1994. Managing Fruit Orchards to Minimize Hurricane Damage HortTechnology 4: 21-27. Cunningham, K. J., R. A. Renken, M. A. Wacker, M. R. Zygnerski, E. Robinson, A. M. Shapiro, and G. L. Wingard. 2006. Application of carbonate cyclostratigraphy and borehole geophysics to delineate porosity and preferen tial flow in the karst limestone of the Biscayne aquifer, South Florida. Geological Society of America 404: 191-208. Cunningham, K.J., J.I. Carlson, and N.F. Hu rley. 2004. New method for quantification of vuggy porosity from digital optical borehole images as applied to the karstic Pleistocene limestone of the Biscayne aquifer, southeastern Florida Journal of Applied Geophysics 55: 77-90. Delin, G.N., R.W. Healy, M.K. Landon, and J.K. Bohlke. 2000. Effects of topography and soil properties on recharge at two sites in an agricultural field. American Water Resources Association 36: 1401-1416. Dexter, A.R., and N.R.A. Bird. 2001. Methods for predicting the optimum and the range of soil water contents for tillage based on the water retention curve. Soil and Tillage Research 57: 203-212. Durner, W. 1994. Hydraulic conductivity estimation for soils with heterogeneous pore structure. Water Resources Research 30: 211-223.

PAGE 120

120 Fares, A., and A.K. Alva. 2000. Soil water co mponents based on capacitance probes in a sandy soil. Soil Science Society of America Journal 64: 311-318. Fernald, E.A., and E.D. Purdum. 1998. Water Resources Atlas of Florida. Tallahassee, Florida State University, Institute of Scienc e and Public Affairs. Available at: www.evergladesvillage.net/atlas_of_fla /atlas.html. Accessed 19 February 2009. Fish, J. E., and M. Stewart. 1991. Hydrogeology of the surficial aquifer system, Dade County, Florida. Report 90-4108, 50 p.: U. S. Geologica l Survey Water Resources Investigations. Fisher, N.I., and A.J. Lee. 1983. A corre lation coefficient for circular data Biometrika 70: 327332. Gardner, C.M.K., T.J. Dean, and J.D. Cooper. 1998. Soil water content measurement with a high-frequency capacitance sensor Journal of Agricultural Engineering Research 71: 395403. Greenwood, J. A., and D. Durand. 1955. The distri bution of length and components of the sum of n random unit vectors. Annals of Mathematical Statistics 26: 233-246. Hanson, C.T., and R.L. Blevins. 1979. Soil water in coarse fragments Soil Science Society of America Journal 43: 819-820. Hassan, SF, Hussin, AG, Zubairi, YZ, 2009. Analysis of Malaysian wind direction data using ORIANA. Modern Applied Science 3(3):115-119. Jabro, J.D., Lieb, B.G., and Jabro, A.D. 2005. Estim ating soil water content using site-specific calibration of capacitance measurements from sentek enviroscan systems. Applied Engineering in Agriculture 21(3): 393-399. Klein, H., and J.E. Hull. 1978. Biscayne aquifer, southeast Florida. Tallahassee, Florida.: U.S. Geological Survey, Water Resources Division. Klute, A. 1986. Water Retent ion: Laboratory Methods. In Methods of Soil Analysis, Soil Science Society of America Book Series 1: 635-662. Kovach, W. L. 2009. Oriana for Windows, version 3.01. Pentraeth, Wales, UK.: Kovach Computing Services. Li, Y. 2001. Calcareous soils in Miami-Dade County. University of Florida. Cooperative extension service, Institute of Food and Agriculture Scien ces, EDIS SL183, Gainesville, Fla. Magilligan, F. and B. Graber. 1996. Hydroclimatological and geomorphic controls on the timing and spatial variability of floods in New England, USA. Journal of Hydrology, 178: 159180.

PAGE 121

121 Mead, R. M., J. E. Ayars, and J. Liu. 1995. Eval uating the influenceog soil texture, bulk density and soil water salinity on a capacitance probe calibration. ASAE Pap. 95-3264, St. Joseph, MI.: ASAE. Mennill, D. J. and Ratcliffe, L. M. 2004. Nest cavity orientation in black-capped chickadees Poecile atricapillus: do the acous tic properties of cavities infl uence sound reception in the nest and extra-pair matings? Journal of Avian Biology 35: 477-482. Merritt, M.L. 1996. Simulation of th e water-table altitude in the Biscayne aquifer, southern Dade County, Florida, water years 1945-89.: U.S. Geological Survey Wa ter-Supply Paper 2458. Meyer, A. F. 1960. Effect of temperature on groundwater levels. Journal of Geophysical Research 65 (6): 1747-1752. Migliaccio, K., B. Schaffer, J. Crane, Y. Li, and R. Muoz-Carpena. 2008a. Assessing capillary rise in a field nursery consid ering irrigation management. In Proceedings of ASABE. ASABE Annual International Meeting, June 29 July 2, 2008, Providence, Rhode Island. Migliaccio, K.W., B. Schaffer, Y.C. Li, E. Evans, J.H. Crane, and R. Munoz-Carpena. 2008b. Assessing benefits of irrigation and nutrient management practices on a Southeast Florida royal palm (Roystonea elata) field nursery. Irrigation Science 27: 57-66. Minkowski, K., and B. Bruce Schaffer. 2001. Sect ion 1: A geographic information system for Miami Dade County agriculture In Miami-Dade County agri cultural land retention study, Vol. 2 Florida agricultural mark et research center, Gainesville, FL. p. 110: Institute of Food and Agricultural Sciences, University of Florida. Morgan, K. T., L. R. Parsons, T. A. Wheatson, D. J. Pitts, and T. A. Obreza. 1999.Field calibration of a capacitance water content probe in fine sand soils Soil Science Society of America Journal 63 (4): 987-989. Muoz-Carpena, R., and A. Ritter. 2005. Hidrologa Agroforestal Canarias, Spain: MundiPrensa. Muoz-Carpena, R., Y. Li, and T. Olczyk. 2002. Alternatives of low cost soil moisture monitoring devices for vegetable production in South Miami-Dade County. University of Florida Cooperative Extension Service, Institu te of Food and Agriculture Sciences, EDIS ABE333, Gainesville, FL. Nash, J.E., and J.V. Sutcliffe. 1970. River flow forecasting through conceptual models part I A discussion of principles. Journal of Hydrology 10 : 282-290. Nielsen, D. 2006. Environmental site characterization and ground-water monitoring 208-244. 2nd editon. Boca Raton, FL.: Taylor & Francis.

PAGE 122

122 Noble, C.V., R.W. Drew, and J.D. Slabaugh. 1996. So il survey of Dade County area, Florida. In: IFAS USDA/NRCS in cooperation with the University of Florida, Agricultural Experiment Stations, Soil and Water Science Department t.F.D.o.A.a.C. Services, t.S.D.S.a.W.C. District and a.t.F.D.o. Transportation (Eds). Nuez-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: 65-68. Nez-Elisea, R., B. Schaffer, M. Zekri, S.K. OHair, and J.H. Crane. 2000. Monitoring soil water content in tropical fruit orchards in South Florida with multisensor capacitance probes and tensiometers Hort-Science 35: 487. Oliveira, E.G., Srygley, R.B. and R. Dudley. 1998. Do neotropical migran t butterflies navigate using a solar compass? Journal of Experimental Biology 201: 3317-3331. Paltineanu, I.C., and J.L. Starr. 1997. Real -time Soil water dynamics using multisensor capacitance probes: laboratory calibration Soil Science Society of America Journal 61: 1576. Paton, Peter W. C., Timm, Brad and Tupper, Todd, 2003. Monitoring pond breeding amphibians. Report from the USGS Patuxent Wildlife Research Center Wellfleet, MA. Peach, M. B., 2003. Rheotaxis by epaulett e sharks, Hemiscyllium ocellatum ( Chondrichthyes hemiscylliidae ), on a coral reef flat. Australian Journal of Zoology 50(4) 407-414. Peck, A.J. 1960. The water table as affected by atmospheric pressure. Journal of Geophysical Research 65: 2383. Perry, W. 2004. Elements of South Florida's Comprehensive Everglad es Restoration Plan. Ecotoxicology 13: 185-193. Pitt, W.. 1976. Response of ground-water levels to flood control operations in three basins, Southeastern Florida, Geological Survey, Ta llahassee: U.S. Dept. of the Interior. Polyakov, V., A. Fares, and M. Ryder. 2005. Cali bration of a capacitance system for measuring water content of tropical soil Vadose Zone Journal 4: 1004-1010. Poole, CE. 1996. Environmental resources for South Florida ecosystem restoration State University System of Florida Monographic holdings. Prathapar ,and Qureshi. 1999. Model ling the effects of deficit irri gation on soil salinity, depth to water table and transpiration in semi-arid zones with monsoonal rains International Journal of Water Re sources Development 15: 141-159. Raes, D., and P. Deproost. 2003. Model to assess water movement from a shallow water table to the root zone. Agricultural Water Management 62: 79-91.

PAGE 123

123 Rasmussen, T.C., and L.A. Crawford. 1997. Identifying and removing barometric pressure effects in confined and unconfined aquifers Ground Water 35: 502-511. Ritter, A., and R. Muoz-Carpena. 2006. Dynami c factor modeling of ground and surface water levels in an agricultural area adjacent to Everglades National Park Journal of Hydrology 317: 340-354. Saxena, S.K. 1982. Geotechnical properties of ca lcareous rocks in Southern Florida. In Geotechnical properties, behavior, a nd performance of calcareous soils, ASTM STP 777, 340-358. American Society for Testing and Materials. Schaffer, B. 2008. Ecophysiology of subtropical a nd tropical fruit crops; HOS 5555 course notes. University of Florida. Schaffer, B. 1998. Flooding responses and water-use efficiency of subtropi cal and tropical fruit trees in an environmentally-sensitive wetland. Annals of Botany 81: 475-481. Skaggs, R.W. 1978a. A water management model for shallow water table soils. Water Resources Research Institute of the University of North Carolina, Raleigh. Skaggs, R.W., L.G. Wells, and S.R. Ghate. 1978b. Predicted and measured drainable porosities for field soils. Transactions of the ASAE 21: 522-528. Smith, W. O. 1939. Thermal transf er of moisture in soils. Transactions American Geophysical Union 24(2): 511-524. Sumner, D.M. 2007. Effects of capillarity and microtopogr aphy on wetland specified yield Wetlands 27: 693-701. Taylor, S. A. 1962. Influence of temperature upon the transfer of water in soil systems. Mededelingers, Landbouwhogeschool, Gent. Belgium. 27: 535-551. Todd, D.K. 1959. Groundwater hydrology New York: Wiley. Towner, G. 1980. Theory of time response of tensiometers European Journal of Soil Science 31(4): 607 621. Tromble, J.M. 1977. Water requirements for mesquite (Prosopis juliflora). Journal of Hydrology 34: 171. Turk, L. J. 1975. Diurnal fluctuations of water tables induced by atmos pheric pressure changes. Journal of Hydrology 26: 1-16. USGS. 2009. Ground-water watch:site number:253029080295601-S-196A. US Geological Survey Ground-Water Watch. Available at: http://groundwaterwatch.usgs.gov. Accessed 28 March 2009.

PAGE 124

124 Van Genuchten, M.T. 1980. A closed-form equation for predicting the hydr aulic conductivity of unsaturated soils Soil Science Society of America Journal 44: 892-898. Van Genuchten, M.T., F.J. Leij, S.R. Yates, and S.K.E.R.L. Robert. 1991. The RETC code for quantifying the hydraulic functi ons of unsaturated soils. Ro bert S. Kerr Environmental Research Laboratory, Office of Research and Development, U.S. E.P.A., Ada, Okla. Wallender, W.W., D.W. Grimes, D.W. Henderson, and L.K. Stromberg. 1979. Estimating the Contribution of a perched water table to the seasonal evapotranspiration of cotton Agronomy Journal 71: 1056-1060. Warrick, A. W., G. J. Mullen, and D. R. Nielse n. 1977. Scaling field-measured soil hydraulic properties using a similar media concept Water Resources Research 13(2): 355. Weeks, E.P. 1979. Barometric fluctuations in wells tapping deep unconfined aquifers Water Resources Research 15(5): 1167. Wellings, S. R., and J. P. Bell. 1982. Physical co ntrols of water movement in the unsaturated zone. Quarterly Journal of Engineering Geology 15: 235 241. White, W.N.1932 A method of estimating groundwater supp lies based on discharge by plants and evaporation from soil. Utah.U.S. Geological Survey Water-Supply Paper. Results of Investigations in Escalante Valley, 659-A 165 p. Zekri, M., R. Nez-Elisea, B. Schaffer, S.K. OHair, and J.H. Crane. 1999. Multi-sensor capacitance probes for monitoring soil water dyn amics in the oolitic limestone soil of South Florida. In Proceeding of the Florida State Horticultural Society 112: 178.

PAGE 125

125 BIOGRAPHICAL SKETCH Luis Pablo Barquin Valle was born in Guatem ala de la Asuncion, Guatemala. Luis went to EARTH University in Costa Rica for his underg raduate education in Agronomy and Natural Resources in 2001. During his undergraduate intern ship in 2003, Luis worked for Dr Rafael Muoz Carpena and Dr Bruce Schaffer with the hydrology and environmental physiology laboratories. This experience helped him to defi ne his interests for wate r resources engineering which motivated him to do his agronomy degree thesis related to th e response of banana production with improved drainage conditions. After graduating from EARTH University in December 2004, Luis worked as a farm supervisor for two years with AVCMI Group, the biggest poultry corporation in Central America. In 2007, Luis came to the U.S. and worked as a research assistant for Dr Yuncong Li with UF-TREC Soil and Water Science Laboratory. In August 2007, Luis was accepted to the University of Florida graduate school at the Agricultural and Biological Engineering Depart ment, joining Dr. Kati Migliaccio s research team to work on his Master of Science thesis focused on modeling capillary rise of groundwater in South Florida calcareous soils.