Incorporating Economic Thresholds and Geospatial Information Technology (GIT) into Pest Management for Twospotted Spider...

MISSING IMAGE

Material Information

Title:
Incorporating Economic Thresholds and Geospatial Information Technology (GIT) into Pest Management for Twospotted Spider Mites in Strawberries
Physical Description:
1 online resource (178 p.)
Language:
english
Creator:
Nyoike, Teresia W
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Entomology and Nematology
Committee Chair:
Liburd, Oscar E
Committee Members:
Lee, Won Suk
Mcsorley, Robert
Grunwald, Sabine

Subjects

Subjects / Keywords:
geostatistics -- gis -- mulch -- nematodes -- neoseiulus -- pcr -- reflectance -- spectroscopy -- strawberries -- tetranychus
Entomology and Nematology -- Dissertations, Academic -- UF
Genre:
Entomology and Nematology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Florida is an important producer of strawberries in the US, supplying 100% of the domestically grown winter crop. The twospotted spider mite (TSSM) mainly feed on the underside of the leaf interfering with photosynthesis, and is the main mite pest affecting strawberry production. A series of experiments were conducted during the 2008 and 2012 strawberry growing season to study the study the ecology and improve the existing pest management practices for TSSM on strawberries. Objective 1 evaluated the effect of TSSM infestation on marketable yields of strawberries, where zero, 5, 10, and 20 mites per strawberry leaf were artificially inoculated onto strawberry plants. Economic losses were detected at the high infestation rate when strawberry plants had accumulated 4,924 mite-days, and a negative correlation was found between TSSM infestation and marketable yield. Objectives 2 & 3 investigated the use of visible and near infrared leaf reflectance spectroscopy to detect TSSM damage on strawberry leaves. Spider mite damage on the leaves was detected using spectral data and the predicting model developed gave an error of 17 mites/leaf. Objective 4 studied spatial and temporal distribution of TSSM and its predatory mite, Neoseiulus californicus McGregor using geographic information system (GIS) and geostatistics to understand movement and dispersal of N. californicus. Results indicated that a sampling spacing of up to 37 m can be used in intensive sampling for site-specific management of TSSM. Neoseiulus californicus ability to disperse was minimal. Objective 5 evaluated the effect of re-using plastic mulch for a second strawberry crop grown within the dead plants from the previous season. Leaving the dead plants increased the incidence of, Colletotrichum acutatum and affected yield in one the second of study. Reusing mulch would save growers up to $3,750 per ha. Finally, in objective 6, molecular and morphological studies were conducted to confirm the identity of a root knot nematode imported with the strawberry transplants during 2010/2011 growing seasons. Polymerase chain reaction (PCR) amplification of the mitochdrial DNA and restriction digestion of the amplified PCR products with Dra1 enzyme produced two fragments at 200 and 250 bp indicating the nematode to be Meloidogyne hapla.
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 Teresia W Nyoike.
Thesis:
Thesis (Ph.D.)--University of Florida, 2012.
Local:
Adviser: Liburd, Oscar E.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-08-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2012
System ID:
UFE0044647:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 INCORPORATING ECONOMIC THRESHOLDS AND GEOSPATIAL INFORMATION TECHNOLOGY (GIT) INTO PEST MANAGEMENT FOR TWOSPOTTED SPIDER MITES IN STRAWBERRIES By TERESIA W. NYOIKE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNI VERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Teresia W. Nyoike

PAGE 3

3 To my loving family

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank my major prof essor, Dr. Oscar Liburd for believing in me and giving me an opportunity to explore into the world of academia from grant writing to conducting research. I thank, Dr. Robert McSorley for his prompt feedback on questions about my research work, Dr. Sabine G runwald for all the talks we had about my research work and life, Dr. Lee Won S uk for introducing me to leaf reflectance spectroscopy. Above all, I thank the four of you for the technical guidance and support through out my research work. In addition to my committee members, I thank Dr. Tesfa Merete for his help in molecular studies of root knot nematodes on strawberri es and Dudley Calfee of Ferris F arm Inc. for his cooperation and support in my field studies on their farm. This experience would not have bee n the same without the help of students and staff in the Small Fruit and Vegetable IPM Laboratory and the staff at the Plant Science Research and Education Unit in Citra for their help in planting, harvesting and frost protection of my strawberry plants at the research station. I thank my family for their love and support, especially Lucy for her persistence, friendship and help with various aspects throughout my study. Despite the distance, your constant phone calls, chats, concerns and good luck wishes k ept me going. I thank all of my friends for their support and encouragement during my studies. Above all, I thank God for this milestone.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ........................... 9 LIST OF FIGURES ................................ ................................ ................................ ....................... 11 ABSTRACT ................................ ................................ ................................ ................................ ... 14 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................. 16 Problem Statement and Justification of the Study ................................ ................................ .. 18 Specific Objectives ................................ ................................ ................................ ................. 22 2 LITERATURE REVIEW ................................ ................................ ................................ ....... 24 Strawberry Production in Florida ................................ ................................ ........................... 24 Twospotted Spider Mite as a Pest of Strawberries ................................ ................................ 25 Biology and Ecology ................................ ................................ ................................ ....... 26 Behavior ................................ ................................ ................................ .......................... 27 Diapause ................................ ................................ ................................ .......................... 27 Movement and Dispersal ................................ ................................ ................................ 28 Management Practices of Twospotted Spider Mites ................................ .............................. 28 Making Management Decisions in Integrated Pest Management ................................ ... 29 The Economic Injury Level Concept ................................ ................................ .............. 29 Geog raphic Information Systems (GIS) in Pest Monitoring ................................ ........... 30 3 EFFECT OF TWOSPOTTED SPIDER MITE, TETRANYCHUS URTICAE KOCH (ACARI: TETRANYCHIDAE) ON MARKETABLE YIELDS OF FIELD GROWN STRAWBERRIES IN N ORTH CENTRAL FLORIDA ................................ ........................ 32 Introduction ................................ ................................ ................................ ............................. 32 Materials and Methods ................................ ................................ ................................ ........... 34 Twospotted Spider Mite Colony ................................ ................................ ..................... 34 Field Experiment Study Site ................................ ................................ ............................ 34 Plot Layout and Experimental Design ................................ ................................ ............. 35 Sampling ................................ ................................ ................................ .......................... 37 Marketable yield ................................ ................................ ................................ ....... 37 Weather factors ................................ ................................ ................................ ........ 38 Data Analysis ................................ ................................ ................................ .......................... 38 Results ................................ ................................ ................................ ................................ ..... 38 Twospotted Spider Mite Population ................................ ................................ ................ 38 Marketable Yield ................................ ................................ ................................ ............. 39

PAGE 6

6 Effe ct of Weather Factors on Twospotted Spider Mite Population Dynamics ............... 40 Insec t Pests and Beneficial Insects ................................ ................................ .................. 41 Discussion ................................ ................................ ................................ ............................... 41 Twospotted Spider Mite Population ................................ ................................ ................ 41 Effe ct of Weather Factors on Twospotted Spider Mite Population Dynamics ............... 42 Marketable Yield ................................ ................................ ................................ ............. 43 Conclusions ................................ ................................ ................................ ............................. 45 4 VISIBLE/NIR REFLECTANCE SPECTROSCOPY FOR TWOSPOTTED SPIDER MITE ( TETRANYCHUS URTICAE KOCH) DETECTION AND PREDICTION ON STRAWBERRY LEAVES ................................ ................................ ................................ ..... 54 Introduction ................................ ................................ ................................ ............................. 54 Materials and Methods ................................ ................................ ................................ ........... 56 Experimental Strawberry Plants ................................ ................................ ...................... 56 Twospotted Spider Mite Colony and Inoculation ................................ ........................... 57 Experiment Set up and Twospotted Spider Mite Sampling ................................ ............ 58 Reflectance Measurements ................................ ................................ .............................. 58 Data Analysis ................................ ................................ ................................ .......................... 59 Correlation Coefficient Spectra ................................ ................................ ....................... 59 Pre processing Treatments ................................ ................................ .............................. 60 PLS Regression Coefficients ................................ ................................ ........................... 61 PLS Models Comparison ................................ ................................ ................................ 61 Results and Discussion ................................ ................................ ................................ ........... 62 Twospotted Spider Mites Population ................................ ................................ .............. 62 Strawberry Leaf Absor bance Spectra ................................ ................................ .............. 63 Correlation Coefficient Spectra ................................ ................................ ....................... 64 PLS Regression Coefficients ................................ ................................ ........................... 64 PLS Models Comparison ................................ ................................ ................................ 65 Conclusions ................................ ................................ ................................ ............................. 68 5 FACTORS AFFECTING REFLECTANCE SPECTROSCOPY AS A MEANS OF DETECTING TWOSPOTTED SPIDER MITE DAMAGE ON STRAWBERRY LEAVES ................................ ................................ ................................ ................................ 79 Introduction ................................ ................................ ................................ ............................. 79 Materials and Methods ................................ ................................ ................................ ........... 81 Strawberry Plants ................................ ................................ ................................ ............. 81 Twospotted Spider Mite Colony and Inoculation ................................ ........................... 82 Strawberry leaf sampling ................................ ................................ ......................... 82 ................................ ................................ ....................... 82 Comparison of Spectral Changes in Mite damage on Strawberry Plants Growing with and without Nitroge n Fertilizer. ................................ ................................ ........... 83 Discriminating TSSM Damage Categories Using Spectral Reading from the Abaxial and the Adaxial Side of Strawberry Leaf ................................ ....................... 84 Pre processing Treatments ................................ ................................ .............................. 84 Data Analysis ................................ ................................ ................................ ................... 84

PAGE 7

7 Results and Discussion ................................ ................................ ................................ ........... 85 Comparison of Spectral Changes in Mite Damaged Strawberry Plants Growing with and without Nitrogen Fertilizer ................................ ................................ ............ 85 Twospotted spider mite population ................................ ................................ .......... 85 Reflectance/absorbance ................................ ................................ ............................ 86 Discriminating TSSM Damage Categories using Spectral Reading from the Abaxial and the Adaxial Side of Strawberry Leaf ................................ ................................ ..... 88 Twospotted spider mite population ................................ ................................ .......... 88 Strawberry reflectance/absorbance ................................ ................................ ........... 88 Conclusions ................................ ................................ ................................ ............................. 91 6 SPATIAL AND TEMPORAL DISTRIBUTION OF TWOSPOTTED SPIDER MITE AND ITS PREDATORY MITE, NEOSEIULUS CALIFORNICUS, ON STRAWBERRIES USING GEOSTATISTICS ................................ ................................ ..... 99 Introduction ................................ ................................ ................................ ............................. 99 Materials and Methods ................................ ................................ ................................ ......... 101 Strawberry Plants and Study Area Description ................................ ............................. 101 Sampling ................................ ................................ ................................ ........................ 102 Geostatistical Analysis ................................ ................................ ................................ .. 102 Data pre processin g ................................ ................................ ................................ 102 Structural analysis of the data ................................ ................................ ................ 103 Results ................................ ................................ ................................ ................................ ... 104 2009/2010 Growing Season ................................ ................................ .......................... 104 2010/2011 Growing Season ................................ ................................ .......................... 105 Discussion ................................ ................................ ................................ ............................. 109 C onclusions ................................ ................................ ................................ ........................... 114 7 CAN TWO YEAR OLD SYNTHETIC MULCH AFFECT MARKETABLE YIELD, ARTHROPOD POPULATIONS, WEEDS AND DISEASES ON FIELD GROWN STRAWBERRIES? ................................ ................................ ................................ .............. 131 Introduction ................................ ................................ ................................ ........................... 131 Materials and Methods ................................ ................................ ................................ ......... 133 Two yr Mulch Preparation ................................ ................................ ............................ 133 Experimental Design and Plots ................................ ................................ ..................... 134 Arthropod Populations Sampling ................................ ................................ .................. 134 Weed Sampling ................................ ................................ ................................ ............. 135 Plant Size and Marketable Yields ................................ ................................ .................. 135 Disease Incidence ................................ ................................ ................................ .......... 135 Data Analysis ................................ ................................ ................................ ................. 136 Results and Discussion ................................ ................................ ................................ ......... 136 Arthropod Populations ................................ ................................ ................................ ... 136 Weeds ................................ ................................ ................................ ............................ 137 Plant Size Growth and Marketable Yields ................................ ................................ .... 138 Disease Incidence ................................ ................................ ................................ .......... 139 Conclusions ................................ ................................ ................................ ........................... 140

PAGE 8

8 8 CONFIRMATION OF MELOIDOGYNE HAPLA ON STRAWBERRY IN FLORIDA USING MOLECULAR AND MORPHOLOGICAL TECHNIQUES ................................ 149 Introduction ................................ ................................ ................................ ........................... 149 Materials and Methods ................................ ................................ ................................ ......... 151 Root knot Sampling and Soil Extraction ................................ ................................ ....... 151 Root knot Nematode Females ................................ ................................ ....................... 152 Morphological Characterization ................................ ................................ .................... 152 Molecular Characterization ................................ ................................ ........................... 152 Polymerase chain reaction ................................ ................................ ...................... 153 Restriction fragment length polymorphism ................................ ............................ 153 DNA sequencing ................................ ................................ ................................ .... 154 Results ................................ ................................ ................................ ................................ ... 155 Root knot Nematode Symptoms and Soil Nematode Abundance ................................ 155 Morphological Characterization ................................ ................................ .................... 155 DNA Extraction and PCR Amplification ................................ ................................ ...... 155 Discussion ................................ ................................ ................................ ............................. 156 Conclusions ................................ ................................ ................................ ........................... 158 9 CONCLUSION ................................ ................................ ................................ ..................... 163 LIST OF REFERENCES ................................ ................................ ................................ ............. 168 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ....... 178

PAGE 9

9 LIST OF TABLES Table page 3 1 Time effect for cumulative marketable yields of strawberries at different T. urticae infestation levels in 2008/2009 field growing season. ................................ ....................... 47 3 2 Time effect for cumulative marketable yields of strawberries at different T. urticae inf estation levels in 2009/2010 field growing season. ................................ ....................... 48 3 3 Mean number of insect pests and beneficial insects recorded on field grown strawberries in 2008/2009 and 2009/2010 growing season in Citrus County, FL. ............ 49 4 1 Descriptive statistics for both calibration and validation data sets for twospotted spider mites on counted per strawberry leaf on two strawberry varieties (Chandl er and Florida Festival). ................................ ................................ ................................ ......... 70 4 2 Results of calibration and prediction of the PLS models with different preprocessing methods for twospotted spider mites on strawberries. ................................ ....................... 71 5 1 Descriptive statistics for the twospotted spider mite counted on strawberry leaves growing with and without nitrogen fertilizer. ................................ ................................ .... 92 5 2 Classification summary of Linear Discriminant Analysis (LDA) with cross validation of principal components (PC) from the PCA of the spectral data collected from mite damaged a nd undamaged strawberry plants. ................................ ................... 93 5 3 Descriptive statistics for the twospotted spider mite counted on the strawberry leaves used for abaxial and adaxial spectra study to classifying twospotted spider mite damage. ................................ ................................ ................................ ...................... 94 5 4 Classification summary of Linear Discriminant Analysis (LDA) with cross validation of principal components (PC) from PCA of the spectral data collected from the adaxial (top) surface of strawberry leaves with three twospotted spider mite infestation levels. ................................ ................................ ................................ .............. 95 5 5 Classification summary of Linear Discriminant Analysis (LDA) with cross validation of principal components (PC) from PCA of the spectral data collected from the a baxial (under)surface of strawberry leaves with three twospotted spider mite infestation levels. ................................ ................................ ................................ ...... 96 5 6 Selecting important wavelengths used in classifying twospotted spider mites from spec tral data collected from the adaxial and abaxial surface of strawberry leaves. .......... 97 6 1 Total counts of arthropod populations from three strawberry leaflets in the 2009/2010 strawberry growing season.. ................................ ................................ .......... 116

PAGE 10

10 6 2 Descriptive statistics of the twospoted spider mites and Neoseiulus californicus counted per strawberry leaflet from 7 Jan to 7 April 2011. ................................ ............. 117 6 3 Summary of semivariogram parameters for ordinary kriging (OK) interpolation method used for twospotted spider mites (TSSM) and Neoseiulus californicus (NC) sampled between January and April on strawberries in 2010/ 2011 growing season. ..... 118 6 4 A summary of mean and root mean square errors (RMSE) comparing prediction statistics for inverse distance weighting (IDW) and ordinary kriging (OK) accuracy in e stimating twospotted spider mites (TSSM) on strawberry leaves. ............................. 119 7 1 Mean arthropod populations per strawberry trifoliate leaf from two treatments (strawberry plants growing with and witho ut old plants from the previous season) on a 2 yr old black plastic mulch. ................................ ................................ ........................ 142

PAGE 11

11 LIST OF FIGURES Figure page 3 1 Mean weekly temperatures recorded in 2008/2009 and 2009/2010 strawberry growing seasons at Citra, Florida. Weeks = time since initiation of sampling. ................. 50 3 3 Accumulated mite days over time for different T. urticae densities o n field grown strawberries for the growing season 2008/2009. ................................ ............................... 51 3 4 Accumulated mite days over time for different T. urticae densities on field grown strawberries for the growing season 2009/ 2010. ................................ ............................... 51 3 5 Mean harvested marketable yields (Kgs) from different T. urticae densities on field grown strawberries for the 2008/2009 growing period. Columns with same letters are not signifi cantly different ( P = 0.05). ................................ ................................ ................ 52 3 6 Mean harvested marketable yields (Kgs) from different T. urticae densities on field grown strawberries for the 2009/2010 growing period. Columns with same letters are not significantly different ( P = 0.05). ................................ ................................ ................ 52 3 7 Harvested marketable yields from different T. urticae densities on field grown strawberries for the growing period between Decembe r 2008 and March 2009. .............. 53 3 8 Harvested marketable yields from different T. urticae densities on field grown strawberries for the growing period between December 2009 and March 2010. .............. 53 4 1 varying twospotted spider mite damage. ................................ ................................ ................................ .................... 72 4 2 Absorbance spectra within visible/NIR spectrum with four twospotted spider mite four infestation levels on strawberry leaflet ....................... 73 4 3 Partial least squares regression coefficient showing important wavelengths for the best PLS models ................................ ................................ ................................ ............... 75 4 4 Twospotted spider mite prediction using multiple linear regressions of the important wavelengths associated with TSSM spectra for strawberry varieties. ............................... 76 4 5 Cross validation results for the best predictive models ................................ .................. 77 4 6 Scatter plot of predicted log mite versus counted log mite for the best predictive models for the two strawberry varieties. ................................ ................................ ............ 78 5 1 Typical absorption spectra from collecting spectral readings from the adaxial (Ad) an d abaxial (Ab) surface of strawberry leaves infested with twospotted spider mites at three infestation levels including a control (C), low (L), and high (H). ........................ 98

PAGE 12

12 5 2 Correlogram spectra of sp ectral data collected from the adaxial (top) and abaxial (underside) surface of strawberry leaves. ................................ ................................ .......... 98 6 1 Spatial distribution for twospotted spider mites on field grown strawberries collect ed on three sampling dates ................................ ................................ .................. 121 6 2 Population densities of the mite populations on the strawberry leaves during the 2010/2011 strawberry growing season. ................................ ................................ ........... 122 6 3 Release points for Neoseiulus californicus population and its establishment on a strawberry field between 19 January and 30 March 2011. ................................ .............. 123 6 4 Ordinar y kriging interpolation of twospotted spider mites on strawberries sampled on 14 March, 30 March, and 7 April in 2010/ 2011 strawberry growing season. ........... 124 6 5 Inverse distance weighting inte rpolation of twospotted spider mites on strawberries sampled on 14 March, 30 March and 7 April in 2010/ 2011 strawberry growing season. ................................ ................................ ................................ .............................. 125 6 6 Interpolated surfaces for the pooled populatio n of twospotted spider mite recorded on strawberry plants in the field for the entire 2010/2011 growing season using two ordinary kriging and inverse distance weighting interpolation methods. ........................ 126 6 7 Inverse distance weighting interpolation of Neoseiulus californicus on strawberries from 14, 30 March and 7 April in 2010/2011 strawberry growing season. ..................... 127 6 8 Predicted Neoseilu lus californicus per strawberry leaf on 7 April using ordinary kriging and inverse distance weighting. ................................ ................................ ........... 128 6 9 Interpolated surfaces for the pooled population of Neoseiulus californicus reco rded per strawberry leaflet for the entire 2010/2011 growing season using ordinary kriging and inverse distance weighting interpolation methods. ................................ ................... 129 6 10 Spatial distribution of twospotted s pider mites and Neoseiulus californicus populations per strawberry leaflet on field grown strawberries in 2010/2011 growing season. Total counts represent pooled populations for the entire season from January to April 2011. ................................ ................................ ................................ ................... 130 7 1 Tetranychus urticae population dynamics on strawberry plants growing on 2 yr old plastic mulch with and without old plants from the previous season in 2010/2011 strawberry growing season. ................................ ................................ ............................. 143 7 2 Total arthropod populations counted on strawberry leaves growing on 2 yr old plastic mulch with and without old plants from the previous season in 2011/2012 strawberry growing season. ................................ ................................ ............................. 144 7 3 Weed species composition recorded in strawberry plots growing with and without old plants from the previous season on 2 yr black plastic mulch. ................................ ... 145

PAGE 13

13 7 4 Average marketable yield harvested for two growing seasons from plots with strawberry plants growing with and without old plants from the previous season on 2 yr old mulch. ................................ ................................ ................................ ................. 146 7 5 Average strawberry marketable yield over time in 2010/2011 growing season. The two treatments evaluated are strawberry plants growing with and without old strawberry plants from the previous season. ................................ ................................ .... 147 7 6 Average strawberry marketable yield over time in 2011/2012 growing season. The two treatments evaluated are strawberry plants growing with and without old strawberry plants from the previous season on a 2 yrs black plastic mulch. ................... 147 7 7 Mean number of missing strawberry plants in plots growing with and without old plants from the previous season in 2011/2012. ................................ ................................ 148 8 1 Strawberry plants showing nematode symptoms (stunted, excessive fibrous root growth and reduced crowns) versus the healthy plants (the last two plants) that were collected from Citra, Marion County, Florida in April 201 1 ................................ .......... 159 8 2 Abundance (per 100cc soil) of Meloidogyne and Trichodorous spp. from soil samples collected in blocks A, B, and C, planted to strawberries. ................................ .. 160 8 3 Female perineal pattern of Meloidogyne hapla isolated from field strawberries collected from Citra, Marion County, Florida in April 2011. ................................ .......... 160 8 4 Amplification products using COII a nd 1RNA primers of the mtDNA from three Meloidogyne spp. females extracted from 2010/2011 and 2011/2012 strawberry plants on 1.8% agarose gel. ................................ ................................ ............................ 161 8 5 Restriction fragment patterns of the PCR amplified products of the region between COII and 1RNA of the mtDNA after digestion with endonuclease Dra1 on 1.8% agarose gel. ................................ ................................ ................................ .................... 162

PAGE 14

14 Abstract of Dissertation Presented to the Graduate School of the University of Florid a in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy INCORPORATING ECONOMIC THRESHOLDS AND GEOSPATIAL INFORMATION TECHNOLOGY (GIT) INTO PEST MANAGEMENT FOR TWOSPOTTED SPIDER MITES IN STRAWBERRIES By Teresia W. Nyoike August 2012 Chair: Oscar E. Liburd Major: Entomology and Nematology Florida is an important producer of strawberries in the US, supplying 100% of the domestically grown winter crop. The twospotted spider mite (TSSM) mainly feed on the undersi de of the leaf interfering with photosynthesis, and is the main mite pest affecting strawberry production. A series of experiments were conducted during the 2008 and 2012 strawberry growing seas on to study the study the ecology and improve the existing pes t management practices for TSSM on strawberries Objective 1 evaluated the effect of TSSM infestation on marketable yields of strawberries, where zero, 5, 10, and 20 mites per strawberry leaf were artificially inoculated onto strawberry plants. Economic lo sses were detected at the high infestation rate when strawberry plants had accumulated 4,924 mite days, and a negative correlation was found between TSSM infestation and marketable yield. Objectives 2 & 3 investigated the use of visible and near infrared l eaf reflectance spectroscopy to detect TSSM damage on strawberry leaves. Spider mite damage on the leaves was detected using spectral data and the predicting model developed gave an error of 17 mites/leaf.

PAGE 15

15 Objective 4 studied spatial and temporal distribu tion of TSSM and its predatory mite, Neoseiulus californicus McGregor using geographic information system (GIS) and geostatistics to understand movement and dispersal of N. californicus R esults indicated that a sampling spacing of up to 37 m can be used i n intensive sampling for site specific management of TSSM. Neoseiulus californicus a bility to disperse was minimal. Objective 5 evaluated the effect of re using plastic mulch for a second strawberry crop grown within the dead plants from the previous seaso n. Leaving the dead plants increased the incidence of, Colletotrichum acutatum and affected yield in one the second of study. Reusing mulch would save growers up to $3,750 per ha. Finally, in objective 6, molecular and morphological studies were conducted to confirm the identity of a root knot nematode imported with the strawberry transplants during 2010/2011 growing seasons. Polymerase chain reaction (PCR) amplification of the mitochdrial DNA and restriction digestion of the amplified PCR products with Dr a1 enzyme produced two fragments at 200 and 250 bp indicating the nematode to be Meloidogyne hapla.

PAGE 16

16 CHAPTER 1 INTRODUCTION Strawberry, Fragaria ananassa Duchesne, is a high value crop commercially produced in several states including California, Florida, M ichigan, North Carolina, New York, Ohio, Oregon, Pennsylvania, Washington, and Wisconsin. Florida is second after California in terms of production, harvested area, and total yield in US strawberry production. During the 2010 2011 growing season, 4006 ha w ere harvested in Florida valued at 366.3 million USD (USDA NASS 2011). Florida strawberry industry supplies 100% of the domestically grown winter crop during the months of November through March (Mossler and Nesheim 2007, Rondon et al. 2005). It is therefo re important to maintain high quality production during these months in order to maintain profitability of this high strawberry industry include pest problems, i.e. insects, mites, d iseases and weeds. There are also problems associated with pesticide resistance, and loss of the fumigant methyl bromide (Mossler and Nesheim 2003, 2007). Twospotted spider mites (TSSM), Tetranychus urticae Koch of the family Tetranychidae are the most da maging and persistent mite pests that affect field grown and greenhouse strawberries in Florida. They are potential pest wherever strawberries are produced (Oatman & McMurtry 1966, Sances et al. 1982, Mossler and Nesheim 2003, 2007, Chandler et al. 2008 ). Spider mites appear in the field very early in the growing season. They can be introduced into the field with strawberry transplants and for this reason growers are advised to routinely inspect transplants received from nurseries to ensure they are pest fr ee. Twospotted spider mites damage strawberry plants by piercing and removing cell contents of the leav chlorophyll and reduced photosynthetic rates, leaving the leaf with white or yellow spots or

PAGE 17

17 ves, low densities of T. urticae mainly cause damage to the spongy mesophyll tissues, while high densities intensify the damage, causing severe injury to the palisade parenchyma cells (Sances et M on host plants include injection of toxic saliva and changes in the chemical composition of host plants after mite potassium and magnesium in chrysanthemum after TSSM damage. In addition to cytological and chemical changes in mite damaged plants, physiological processes are also affected. Sances at el. (1979, 1982) observed that strawberry leaf damage by TSSM resulted in a decrease in the intensity of transpiration. H igh infestations of TSSM can significantly affect plant growth and substantially reduce plant yields, especially if the plants are attacked early in the growing season (Wyman et al. 1979, Sance et al. 1982, Rhodes and Liburd 2006, Satos et al. 2007). The u se of pesticides is the most common method of control to manage TSSM populations. Integrated pest management (IPM) relies on field monitoring of pest population levels and ntrol action carried out only when needed (Flint and Gouveia 2001). In this regard, the economic threshold (ET), also known as action threshold, is used as a guide t o apply a management action in order to avoid economic losses (Pedigo 1996). Thresholds are determined based on the relationship between pest population levels and the damage they cause to a crop. Once the ET level is known, management tactics are applied at this level to prevent the pest from reaching economically damaging levels. The economic threshold levels for TSSM in strawberry vary from region to

PAGE 18

18 region, cultivar, stage of crop growth, time of season, control options and market prices (Walsh and Zal om 1996, Pedigo 1996). Problem Statement and Justification of the Study S trawberry growers routinely scout for the various life stages of TSSM during the growing season. Currently, the economic threshold of TSSM in strawberries is based on the presence of any life stage on 5% of the leaflets (Mossler and Nesheim 2003 & 2007). However, this approach is not sufficient because the different life stages of spider mites offer varying risks to the strawberry plant. For instance, a threshold based on egg populati on is not the same as one based on the adult stage, especially where the population of females is high. In order to stages except eggs) that are already causing damage to the plant. Therefore, there is a need to determine the mite numbers that would cause economic losses on field grown strawberries in Florida. Sampling and monitoring are the two most important steps in determining pest population levels in order to guide for sound IPM programs. Early pest detection is therefore preferred so that management actions can be carried out before pest populations begin to cause economic losses. Due to their small size, TSSM may remain unnoticed until they cause visible d amage to the leaves (Opit et al. 2009). As a result, strawberry growers spray often on a calendar basis to avoid unforeseen losses. Frequent spraying kills non target organisms and can also contaminate the environment. Moreover, spraying on a calendar basi s increases the selection pressure and cause pest resistance to pesticides. Currently, field sampling and detection for spider mites on strawberry leaves involve the use of magnifying lens to look for mites on the undersides of the leaves. This procedure is labor intensive and some growers try to avoid it due to the strain on the eye. Development of a rapid

PAGE 19

19 and reliable method for early detection of spider mites and other strawberry pests would facilitate timely implementation of control measures where the y are most needed. Activities that are likely to improve TSSM monitoring and assessment would greatly enhance pest management practices in strawberries. Remote sensing provides a non destructive, rapid, and cost effective means of identifying and quantify ing plant stress from changes in spectral characteristics of the leaves (Mirik et al. 2006). S tudies have demonstrated that insect and mite damage can be detected using leaf reflectance as shown for the greenbug Schizaphis graminum (Rondani) on wheat (Miri k et al. 2006), European corn borer Ostrinia nubilalis (Hbner) infestation on corn (Carroll et al. 20 0 8), cotton aphid, Aphis gossypii Glover and spider mite damage on cotton (Fitzgerald et al. 2004, Reisig and Godfrey 2006), and mites on strawberries (Fr aulo et al. 2009). Generally, green leaves have a typical reflectance signature between 0.4 2.4 m with a high reflectance in the short wave near infrared (NIR) wavelengths (0.7 1.0 m) due to internal scattering by leaf cells and five absorption regio ns; one at 0.4 0.7 m due to chlorophyll, and four others at 0.97, 1.20, 1.40, and 1.94 m due to the bending and stretching of the O H bond in water (Gates et al. 1965, Curran 1989). Changes in the leaf, such as those caused by insect damage, can alter d iffuse reflectance of the leaf, causing an increase in reflectance within the visible region and a reduction in reflectance in the NIR region (Nilsson 1995). Removal of chloroplasts (hence chlorophyll) through TSSM feeding results in a decrease in the use of radiant energy and hence reduced vegetative growth and yield (Sances et al. 1979, Sances et al.1982, Kielkiewicz 1985, Reddall et al. 2004, Nyoike and Liburd unpublished data). In addition, TSSM feeding damage also affects the interaction of the leaf wi th light energy in the NIR wavebands.

PAGE 20

20 Cellular damage by TSSM in the mesophyll layer interferes with the ability of the leaf to reflect NIR energy (Jensen 2005). There are factors that can cause spectral changes in the visible/near infrared region that cou ld confound TSSM injury detection and classification on strawberry leaves using spectral characteristics. One example is the interaction between leaf nitrogen (N) content (as supplied from nitrogenous fertilizers) and chlorophyll removal as induced by TSSM feeding on strawberry leaves. It is known that the majority of leaf N is contained in chlorophyll molecules and therefore leaf chlorophyll and leaf N contents are closely related (Yoder and Pettigrew Crosby 1995). The chlorophyll content in corn leaves do ubled as N fertilizer was increased (Daughtry et al. 2000). Alternatively, it is known that insect feeding damage can also cause differences in adaxial (top) and abaxial (under side of the leaf) leaf reflectance between 400 and 2500 nm (Carroll et al. 2008 ). Furthermore, spider mites mostly inhabit the underside of the leaf surface on strawberries where they feed and only come to the top surface if the population is very high. Sampling programs for TSSM on strawberries involve turning the leaf to scout for spider mites. Most previous studies e.g. Mirik et al. 2006, Reisig and Godfrey 2006 and Fraulo et al. 2009 collected leaf reflectance from the adaxial (top surface) surface of the leaf. The question is could leaf reflectance from the abaxial side of the le af offer better precision in determining TSSM infestation level? Lastly, could leaf spectral characteristics vary between strawberry varieties? To manage TSSM on strawberries, Neoseiulus californicus McGregor (Acari: Phytoseiidae) has been recommended as a potential biological control agent in F l orida (Rhodes and Liburd 2006, Fraulo and Liburd 2007, Liburd et al. 2007, Fraulo et al. 2008). Collectively, these studies demonstrated that N. californicus was able to persist in the field and give season

PAGE 21

21 long co ntrol of TSSM. However, the spatial distribution of N. californicus in relation to its prey TSSM and the ability to move and disperse on strawberry plantings after localized inoculative releases has not been established on Florida strawberries. Spatial var iability studies of N. californicus and its prey would elucidate their population dynamics and determine how predatory mites disperse from a point of release after inoculative studies. In order to understand the movement and dispersal of N. californicus af ter inoculative releases to control TSSM, we studied their spatial and temporal distribution using geographic information system (GIS) and geostatistics. Other than pest related problems, strawberry is one of the most expensive crops to produce, with an av erage variable cost per ha estimated at $24,478 ($9791/acre) [Santos et al., 2011]. These costs include land preparation, labor, strawberry transplants, fertilizer, weeding, insect and mite management, and disease control. To remain profitable, strawberry production practices must be easy to accomplish, sustainable, and economically feasible. Double cropping (using the plastic mulch for a second crop) is one strategy that could be used to reduce cost of production (Waterer et al. 2007). Double cropping with a second strawberry crop demands leaving the plastic mulch intact through the summer months for use in the fall season. Using the plastic mulch for a second crop is beneficial in that the cost of purchasing new plastic mulch, drip tubing, laying and dispo sing the mulch is spread across two growing seasons. In addition, the plastic is not hauled to a landfill or burned at the end of the season, reducing contamination to the environment. Although a few growers have double cropped strawberries with other vege table crops (such as cucumber, watermelon, and squash), growing a second strawberry crop on the same plastic is a relatively new practice in Florida. In 2009, approximately a third of Florida's

PAGE 22

22 strawberry land was used for double cropping (Noling and Whide n 2009, Noling 2011). Currently, growers use a desiccant to kill the strawberry plants at the end of the season and the dead old plants are removed in the fall before transplanting new plants. The question that is being addressed is what would be the effec t of planting new strawberry plants within the dead plants from the previous season? Leaving the dead plants would potentially save growers extra labor to pull out the plants before transplanting the new ones. There is a need to determine if leaving the ol d plants in the field will affect arthropod populations on the second crop, act as a reservoir for weed seeds, and/ or a source of disease inocula that can ultimately affect marketable yields. Could the thatch that is left from the dead plants (in old hole s) be sites for diapausing insects and mites or a source for disease inoculum in the field? Twospotted spider mites have a diapausing stage that is able to survive through hot summer months. Therefore, the ultimate goal of my study was to improve the exis ting pest management practices for TSSM. Studies involved determining infestation levels, per strawberry leaf, that could cause economic losses, evaluating the effectiveness of leaf reflectance spectroscopy in detecting TSSM damage on leaves, studying the spatio temporal distribution of TSSM and its predatory mite, N. californicus and assessing double cropping as a cost effective strawberry production system were conducted over a 4 year period. Specific Objectives 1. To determine the relationship between TSSM infestation and damage on field grown strawberries in order to develop an economic threshold level for management of TSSM 2. To develop a spectral model to predict mite counts using visible/near infrared spectroscopy 3. To examine factors affecting TSSM damage d etection on strawberries using leaf reflectance spectroscopy 4. To map spatial and temporal distribution of TSSM and its predatory mite (Neoseiulus californicus McGregor) using geographic information system (GIS) and geostatistics

PAGE 23

23 5. To evaluate the effect of re using plastic mulch for a second strawberry crop grown with the dead old plants from the previous season. 6. To confirm the identity of a root knot nematode on strawberries using molecular and morphological studies

PAGE 24

2 4 CHAPTER 2 LITERATURE REVIEW Strawbe rry Production in Florida In Florida strawberries are grown in open fields in double rows on raised beds with drip lines covered with black plastic mulch (Chandler et al. 2008). The soil is fumigated before planting with methyl bromide/chloropicrin to kil l soil pathogens and weeds. During the season weed management is achieved with regular herbicide application between the beds or hand weeding. Weed problems not only impact the quality and quantity of harvested yield but also harbor spider mites (Mossler a nd Nesheim 2003, 2007). irrigation is used for first three weeks for plant est ablishment, and later on in the season for freeze protection. Drip irrigation is used after plant establishment as well as for fertigation purposes. Usually, the harvest period runs from late November through early April. During harvesting berries are hand picked every other day or every two days. Strawberry plants are susceptible to a wide array of pathogens, insects, birds and mites. The main diseases that pose a problem for growers are Botrytis (gray mold) and Colletotrichum (anthracnose, fruit rot, and crown rot). The gray mold affects all stages of strawberry fruit both during development in the field and in transportation. To manage gray molds, growers mostly use Captan and Thiram (fungicides) in a prophylactic control program. Other sporadic and emerg ing fungal diseases include powdery mildew and Phytophthora (Mossler and Nesheim 2007). The only commonly encountered bacterial disease is Angular leaf spot caused by Xanthomonas fragariae Kennedy and King and can be problematic in cold and wet conditions

PAGE 25

25 especially at transplant (due to overhead irrigation for plant establishment) and freezing periods (White and Liburd 2005). Strawberries are attacked by a host of insect pests. Early season pests include Spodoptera frugiperda J. E. Smith, S. eridania Cram er and Helicoverpa zea Boddie, causing damage to the foliage and flowers (Mossler and Nesheim 2003). Growers especially those using predatory mites, apply Bacillus thuringiensis (B.t) and spinosad for the control of these lepidopteran larvae. Later in the season, aphids and thrips may be found damaging the developing flowers and fruits. These two pests can be controlled with broad spectrum pesticides such as Malathion although if left untreated natural predators and parasites can regulate their populations. Bird predation has been viewed as a sporadic problem. According to Mossler and Nesheim 2007 and O. E Liburd (personal experience), bird predation losses have increased significantly over the last few years. The main bird species include American robin, c edar waxwing, and crows. Birds have been estimated to cause over $2 million (USD) in losses (Mossler and Nesheim 2007). The main control method currently in use include propane cannon but crows can quickly become accustomed to them and subsequently become less effective. Additionally, deterrents such as reflective ribbon, and scare crow such as Terror Eye. Twospotted Spider Mite as a Pest of Strawberries Twospotted spider mite (TSSM) belongs to the family Tetranychidae and is a worldwide pest of many agr icultural crops including strawberries [ Fragaria spp.] ( Wyman et al. 1979 Sances et al. 1982). Given the optimum conditions; high temperature, low humidity, host plant and leaf age, TSSM can reach damaging population levels resulting in adverse effects on the host plant.

PAGE 26

26 Biology and Ecology Life cycle of TSSM comprises of an egg, larva, protonymph, deutonymph, and an adult (male and female). The three immature stages have a feeding (active) and quiescent (resting) stage (Huffaker et al. 1969). Of the imma ture stages, only the larva has 3 pairs of legs while the other two; proto and deutonymph have 4 pairs like their adult stages. Developmental period is highly dependent on the temperature (White and Liburd 2005, Osborne et al. 1999). Given optimum conditi ons (approximately 26.7C), spider mites can complete their development in five to twenty days (Fasulo and Denmark 2000). The lower temperature limit for development is about 12 o C while the upper threshold is 40 o C (Jeppson et al. 1975). There are up to thr ee times more females than males in any given colony of TSSM (Osborne et al. 1999). Males are elliptical but smaller than the females with their bodies pointed posteriorly. The female is also oval in shape and about 0.4 mm in length with a rounded posterio r end. Females occur in different colors ranging from yellow or green to dark green, straw color, brown, black, translucent to various shades of orange mainly in the summer months (Osborne et al. 1999). Adult female life span has a preoviposional and ovipo sitional period where the later begins after she has laid her first egg. Ovipositional period can last from 10 days at 35 o C to 40 days at 15 o C. A female can lay up to 100 eggs in her lifetime (Huffaker et al. 1969). The eggs are spherical (approximately 0. 2 mm in diameter) and clear to tan in color. They are laid on the underside of leaves by the females. Usually, the eggs are attached to the leaves with fine silk webbing. Egg hatch in about three days (Fasulo and Denmark 2000). Fertilized eggs hatch into f emales while males develop from unfertilized eggs (Osborne et al. 1999). Generally, a mated female can produce both male and female progeny but unmated females only develop into males.

PAGE 27

27 Behavior Quiescent female deutonymph are known to release a sex pherom one that attracts and maintain males close to them (Penman and Cone 1972). Males will stay near the quiescent the females, which may result in a fight in cases whe re more than one male is present (Cone 1985). The male attraction/response to quiescent deutonymphs is increased by the presence of female deutonymph web (Penman and Cone 1972). Another interesting behavior in TSSM is the ability to develop genetic resista nce to insecticides and miticides. This is greatly enhanced by the type of reproduction manifested by the TSSM: arrhenotoky. Arrhenotoky is a kind of parthenogenesis where unfertilized eggs develop into haploid males.Twospotted spider mite males have only one set of chromosomes (haploid) and therefore genetic mutations will be expressed immediately. These new genetic variations can produce resistant genes that are likely to be added quickly to the population through natural selection (Helle and Overmeer 197 3). Other factors contributing to the increased probability to develop pesticide resistance in TSSM population include their high reproductive rates, short generational time, sib mating, and increased selection pressure due high usage of pesticides with si milar modes of action (Osborne et al. 1999). Diapause Twospotted spider mite will overwinter as adults (mated females before they lay eggs) on host plants or enter into a state of diapause (Veerman 1985). Diapause is a genetically determined state of supp ressed development that may be controlled by environmental factors. Decreased photoperiod, low temperatures and deteriorating host plant quality induce diapause in mite populations. Once in diapauses, the mites will neither feed nor reproduce and they usua lly move from the host to hibernation sites (Veerman 1985) such as cracks and crevices (Osborne et

PAGE 28

28 al. 1999). In southeastern states where winter temperatures are mild, reproducing populations can remain active on alternative hosts throughout the winter mo nths. The state of diapause is terminated once favorable conditions (increase in both photoperiod and temperature) are achieved in the spring. Movement and Dispersal Twospotted spider mites will disperse to new hosts once the current host plant quality dec lines due to high mite populations, plant aging, or crop removal. Dispersal over short distances such as within plant and inter plant movement is achieved through crawling. Additionally, TSSM populations are able to move from weeds or other hosts to invade crops in the fields through crawling (Kennedy and Smitley 1985). Long distance dispersal is achieved through wind assisted aerial dispersal. Twospotted spider mites move up the plant and occupy the edges of their host where they are more exposed and prone to aerial dispersal. In the presence and Smitley 1985). Males are not known to disperse. Mites can also drop off the infested plants or move over soil surfaces to new hosts (Osborne et al. 1999). Management Practices of Twospotted Spider Mites The use of chemical pesticides is the most popular method for managing TSSM in s trawberries. In an effort to reduce the effects of pesticides in the environment, human beings and reduce resistance problems, the use of predatory mites in strawberries such as Phytoseiulus persimilis Athias Henriot and Neoseiulus californicus McGregor ha ve been studied extensively in Florida (Rhodes and Liburd 2006, Fraulo and Liburd 2007, Liburd et al. 2007,). Prior to introduction of bifenazate, hexythiazox and etazole, up to 40% of strawberry growers were estimated to have adopted the use of predatory mites. These numbers have dropped to about 25%

PAGE 29

29 after registration of these more effective and newer compounds (Mossler and Nesheim 2007). A survey conducted in 2007 reported that growers see the cost of using predatory mites as one of the limiting factor t o their adoption (Fraulo 2007) Making Management Decisions in Integrated Pest Management The idea of economic damage, economic injury level and economic threshold were borne out of the need for more judicious use of pesticides for pest management. The goal was to conserve natural enemies, protect human and the environment. It is already established that pests do not cause significant losses at all times when present. Therefore, establishing when management action should be applied is very important in order to avoid losses due to pest. Economic injury level (EIL) is the lowest number of insects/mites that will cause economic damage (Pedigo 1996). In order to avoid losses, management actions should be applied before pest population reaches the EIL. Population will need to be controlled at the economic threshold (ET). This is also known as action threshold (AT) and it acts as a guide as to when control measures are necessary. Because EIL forms the basis for determining economic thresholds (ETs), it is important to determine how EIL is calculated. The Economic Injury Level Concept Economic injury level is characterized using the pest numbers even though the term suggests that EIL is an injury level. Pedigo and Rice (2009) argued that as the name suggests, EIL i s a level but because injury is hard to measure across a field pest numbers can therefore be used as an index. In order to determine the EIL, it is important to define the terms and procedures used. First, economic damage has been defined as the amount of injury that would As Pedigo and Rice (2009) reported economic

PAGE 30

30 damage begins to occur when the cost of management is equal to the monetary loss that can be realized from the pest population. Economic injury level can therefore be expressed as: EIL = C / V x b where V = ma rket value per unit of produce D = damage per unit injury (e.g. bushels lost/acre) C = cost of management per area b = is a coefficient of a linear regression representing yields and number of mites per unit area Since ET is calculated from EIL, it follows that ET = EIL x C x where C = factor of increase per unit time and x = time period in weeks (Pedigo and Rice 2009). Geographic Information Systems (GIS) in Pest Monitoring Geospatial Information Technology (GIT) is a set of tools associated with site s pecific precision pest management. The components of GIT include Global Positioning System (GPS), imagery systems, and Geographic Information Systems (GIS). Geospatial tools are usually integrated together. Global Positioning System is used for precise and accurate location of features/sampling points; remote sensing imagery can be use d to map conditions on the field, spatial analysis to understand the underlying patterns of the data, and all these data can be stored in or analyzed in GIS (Kelly and Guo 200 7). These tools have now become more available to pest management practitioners and can be used to analyze spatial distribution of insects and integrate spatial information into management decisions (Fleischer et al. 1999, Park et al. 2007). Spatial and te mporal dynamics of pest populations can be investigated by generating interpolated distribution maps of species at specific time intervals and comparing these maps to detect changes in the spatial patterns over time (Garcia 2006). Interpolation is the math ematical estimation of values at unsampled locations based on the neighboring sampled locations

PAGE 31

31 (Fleischer et al. 1999, Ess and Morgan 2003). Interpolation methods (such as kriging, local average and inverse distance weighting, and nearest neighbor) create maps that could be used to help us understand population dynamics of a pest. Remote sensing and digital image analysis are methods of acquiring data from an object without being in physical contact with the object (Ess and Morgan 2003). Remote sensin g provides a harmless, rapid, cost effective means of identifying and quantifying plant stress from changes in spectral characteristics of the canopy (Mirik et al. 2006). Plant stress, due to pest damage, causes changes in absorption in the visible spectra l region due to a decrease in chlorophyll content as induced by insect/mite damage. Similarly, change occurs at the near infrared (NIR) region of the electromagnetic spectrum due to changes to the internal cell structure of the leaves (Kelly and Guo 2007). These changes in reflectance and absorption can be used to differentiate pest infested plants from non infested plants in the field (Mirik et al. 2006). Various remote sensing platforms including aircraft, satellite and ground and lab based methods can be used to acquire the spectral characteristics of plants.

PAGE 32

32 CHAPTER 3 EFFECT OF TWOSPOTTED SPIDER MITE, TETRANYCHUS URTICAE KOCH (ACARI: TETRANYCHIDAE) ON MA RKETABLE YIELDS OF F IELD GROWN STRAWBERRIES IN NORTH CENTRAL FLORIDA Introduction The market value o f strawberries ( Fragaria ananassa Duchesne) depends on the quality of fruits produced, which is influenced by a number of factors including physical injuries from insect pests and pathogens. Management strategies that will reduce strawberry damage from pes ts and pathogens are therefore critical in enhancing the quality of strawberries produced. Such efforts require adequate knowledge of the strawberry pest complex and beneficial insects and their biology. The twospotted spider mite [TSSM], Tetranychus urtic ae Koch (Acari: Tetranychidae), is one of the most damaging and persistent mite pests that affect field grown strawberry in Florida (Poe 1971, Mossler and Nesheim 2003 & 2007, Chandler et al. 2008) and other areas where it is produced (Oatman and McMurtry 1966, Sances et al. 1982 Walsh et al. 1998, Greco et al. 2005, Satos et al. 2007 ). The piercing and removal of cell contents from strawberry leaves by T. urticae (Tomczyk and Kro 1979, DeAngelis et al. 1982, Park and Lee 2005). At high infestation rates, T. urt icae can suppress flower and leaf development, and ultimately affect the quality and quantity of berries produced (Sances et al. 1982, Fraulo et al. 2008). The economic loss incurred by T. urticae and associated damage is enormous, making it a major target for integrated pest management (IPM) in strawberry systems ( Sances et al. 1982 Liburd et al. 2007). The principles of IPM demand that pest management actions be delayed until it is economically sound to implement them. The term economic threshold(ET) is used to define the

PAGE 33

33 pest population density level at which management actions should be applied to avoid economic losses (Pedigo 1996).Thresholds are determined based on the relationship between pest population levels and the damage they cause to a crop. O nce the ET level is known, management tactics are applied at this level to prevent the pest from reaching economically damaging levels. Economic threshold levels of TSSM in strawberry vary from region to region, cultivar, stage of crop growth, time of seas on, control options and market prices (Walsh and Zalom 1996, Pedigo 1996). An ET of 50 mites per leaflet was reported for a short day strawberry variety in California (Wyman et al. 1979) compared with 1 mite per leaf for the day neutral cultivar Selva (Wal sh et al. 1998). Similarly, an ET of 5 mites per mid tier leaflet was reported during the first five months of growth in winter compared with 10 mites per mid tier leaflets in summer and 20 mites per leaflets after harvest was initiated (Walsh and Zalom 19 96). Finally, T. urticae densities higher than 5 mites per leaflet during the early growth stages can negatively affect the number of berries produced and hence the overall yield (Sances et al. 1982, Gimenez Ferrer et al. 1994). Accurate data on ET for mi tes are lacking or unavailable for north central Florida because mite sampling and management is based on presence/absence or percent infestation per leaf (Mossler and Nesheim 2003) as opposed to actual mite numbers. More precise information on mite number s could improve management efficiency and subsequently avoid economic damage. To provide more accurate information on TSSM, T. urticae densities were manipulated in strawberries under field conditions and their effect on marketable yields of strawberries w ere quantified. Our ultimate goals were to determine the mite numbers that would cause economic losses on field grown strawberries in north central Florida, and to quantify how various

PAGE 34

34 environmental factors influenced ET values. We hypothesized that ET val ues would be affected by the prevailing environmental factors. Materials and Methods Twospotted Spider Mite Colony Twospotted spider mites (TSSM) were obtained from the Small Fruit and Vegetable IPM Laboratory colony at the University of Florida, Gainesvil le, FL. The colony was maintained on bean ( Phaseolus vulgaris L.) plants and strawberry transplants contained in 3.78 liters polyethylene pots. The plants were kept under two 60 watt incandescent bulbs with 14:10 h (light: dark) photoperiod and 60% relativ e humidity. Old mite damaged plants were replaced with new bean plants every other week. Plants were watered manually three times per week or as needed. Field Experiment Study Site The study was conducted at the University of Florida, Plant Science Researc h and Education Center located in Citra, Marion County, Florida, during the 2008/2009 and 2009/2010 strawberry growing seasons. The growing season in north central Florida runs from mid September to March or April the following year depending on the enviro nmental conditions. Two weeks before planting, a soil fumigant consisting of a mixture of methyl bromide and chloropicrin (50:50) (Hendrix and Dail, Palmetto, FL) was applied to the beds at the rate of 36.5 liters/hectare. During land preparation and bed f ormation, fertilizer (10 10 10; N P 2 O 5 K 2 O) [Southern States CO OP, Cordele, GA] was applied at the rate of 227.3 kg per acre on the center cover. A fertigation pro gram using a 16.8 kg per ha of nitrogen and phosphorous (11 37 0) fertilizer (Dyna Flo, Chemical Dynamics, Inc. Plant City, FL) was adopted after transplanting. These rates were increased to 22 kg per ha at flowering to meet the increasing demands of the

PAGE 35

35 p lants. Other growing practices including weeding and removing runners were done according to the standard production practices in North Central Florida (Peres et al. 2006). Bare root, green d from Strawberry Tyme Farms Inc.,Ontario, Canada. After transplanting strawberries, over head irrigation was used for the first three weeks for plant establishment. Drip irrigation was then used for the remaining part of the season. In both years, an inse cticide Dipel (Bacillus thuringiensis) (Valent BioSciences Corporation, Libertyville, IL), applied approximately 3 4 weeks after transplanting strawberries to control cutworms. Several fungicides, including Abound (azoxystrobin) (Syngenta Crop Protection Greensboro, NC), Aliette (aluminum tris) (Bayer Crop Science, Research Triangle Park, NC), Serenade ( Bacillus subtilis ) Agraquest, Davis, CA), and Cabrio EG (pyraclostrobin) (BASF CorporationResearch Triangle Park, NC), were rotated weekly to control common fungal diseases, including Botrytis fruit rot ( Botrytis cinerea ), anthracnose crown rot ( Colletotrichum acutatum ), powdery mildew ( Podosphaera aphanis ) and Colletotrichum crown rot ( Colletotrichum gloeosporides ). Plot Layout and Experimental Design Experimental plots were 7.3 X 6.1 m planted with 6 double rows of strawberries spaced at 0.35 m between plants (along the row) and between the two rows on the bed. Plots were spaced at 11 m apart to provide a buffer zone between individual plots. Experimen tal design was a completely randomized block with four treatments ( T. urticae infestation levels), and five and four replicates per treatment in 2008/2009 and 2009/2010 growing seasons, respectively. The reduction in number of replicates used in 2009/2010 was due to increased labor costs and subsequent land charges. Treatments consisted of three T. urticae infestation levels including a high infestation rate (20 spider mites per leaf), medium infestation rate (10 spider mites per leaf), low infestation rate (5 spider mites per leaf) and a control (zero

PAGE 36

36 mites). The control treatment plots were maintained at close to zero T. urticae / leaf with two applications of the reduced risk acaricide, bifenazate (Acramite, Chemtura, Middlebury, CT) at the rate of 0.89 kg/ha. Bifenazate is one of the recommended miticides for twospotted spider mites on strawberries in Florida and well timed applications of Acramite successfully control populations of T.urticae on strawberries (Rhodes and Liburd 2006, Liburd et al. 2007 ). In order to achieve the three T. urticae infestation levels, spider mites were introduced into the plots from the laboratory colony. The inoculum source for strawberry plants was infested T. urticae bean leaf. The average SE mite counts per bean leaf were 96.6 8.1 per trifoliate. The number of T urticae introduced per plot was determined based on the number of fully grown leaves per strawberry plant per plot in each season. Tetranychus urticae infested bean leaves were clipped onto strawberry leaves with plastic paper clips (Vinyl paper clips, Wal Mart stores, Inc. Bentonville, AR) to allow the mites to crawl/walk onto the new hosts. These numbers were 5760, 2880, and 1440 mites per plot in 2008/2009 growing season and 4320, 2160, and 1080 mites per plot in 2009/2010 growing season for the high, medium and low treatments, respectively. Plants in 2009/2010 growing season were smaller and had fewer leaves than those in 2008/2009 growing season. All treatments including the control were applied four week s after transplanting strawberries. Prior to the treatment applications, 10 mature trifoliate leaves were collected from each plot to determine if there were pre existing (natural infestations) spider mite populations in the field. During the 2008/2009 gro wing season, two foliar applications of bifenazate were required to keep mite populations near zero in the control but in 2009/2010 only one application of bifenazate was needed to keep mite population at or near zero.

PAGE 37

37 Sampling The sampling of T. urticae w as conducted between 12 December 2008 and 2 March 2009 for the 2008/2009 field season and between 17 December 2009 and 8 April 2010, for the 2009/2010 field season. These sampling periods began two weeks after initial infestation and continued to the end o f the growing seasons. Each week, 10 mature trifoliate leaves from 10 randomly selected plants in each replicate were selected for sampling. Leaves were placed in Ziploc bags (Racine, WI) and transported to the laboratory under cool conditions and examine d for TSSM using a dissecting microscope (Leica MZ 12 5, Leica Microsystems, Houston, TX). From these leaves, T. urticae motiles (all stages except eggs), insect pests and beneficial insects, and other mite species were recorded. Cumulative mite days, wher e 1 mite day is equivalent to one mite feeding for 1 day (Wyman et al. 1979) were calculated using weekly mite counts. Cumulative mite days were calculated by averaging mite counts per plant, per week multiplied by 7 days (number of days between two consec utive samplings). i + x i+ 1 (3 1) W here x i is the number of mites at week i of sampling, x i+ 1 is the mite population on next over the total number of wee ks. Marketable yield Strawberries were harvested twice per week from the two inner alternate rows that were not used for leaf sampling. Marketable fruits consisted of berries that weighed more than 10 grams and without physical evidence of damage. Berries less than 10 gram, or berries showing spider mite injury or other damage including bird damage; weather related damage (freezing,

PAGE 38

38 frost), or fungal diseases were considered unmarketable. Harvesting was conducted from 30 December 2008 to 2 March 2009 and fr om 29 December 2009 to 12 April 2010. Weather factors Weather data for Citra, FL was downloaded from Florida Automated Weather Network (FAWN 2011), a University of Florida IFAS Extension website. Daily minimum and maximum temperatures and total rainfall re ceived were obtained for the two growing seasons. Weekly temperature averages and total rainfall received were computed (Figs. 3 1 & 3 2). Data Analysis PROC MIXED (SAS Institute 2002) was used to run a two way analysis of variance (ANOVA) with treatments and time as the main effects and their interactions were also tested. Mean comparisons for cumulative mite days and yield among treatments were evaluated using a PROC LSMEANS statement adjusted with a Bonferonni correction using SAS (SAS Institute, 2002). Treatments were considered significant when P REG (SAS Institute, 2002) was used to establish the relationship between mite infestation levels and harvested marketable yield. A regression was used to evaluate the effect of weather factors on spider mite densities A regression analysis using PROC REG (SAS Institute, 2002) was also used to determine the effect of rainfall and temperature on T. urticae population. Results Twospotted Spider Mite Population In 2008/2009, T. urticae population increased with time and t he different infestation levels resulted in significant differences in mite densities among treatments ( F = 568.48; df=3, 288; P < 0.0001) and over time ( F = 44.56; df =17, 288; P < 0.0001 Fig. 3 3) .High infestation accumulated up to 9042 mite days, which w as significantly different from all other treatments

PAGE 39

39 throughout the season. Strawberry plants in the medium infestation resulted in 6460 mite days while the low density accumulated 4056 mite days at the end of the season in March. In the control plots, mit e populations remained low throughout the season, barely accumulating 200 mite days on the strawberry plants. In 2009/2010 growing season, fewer mite days were accumulated on strawberry plants as compared with the 2008/2009 growing season (Fig. 3 4). Never theless, significant ( F = 41.43; df= 3, 192; P < 0.0001) differences in the number of accumulated mite days among the treatments were recorded, as well as differences over time ( F = 27.60; df=15, 192; P < 0.0001 Fig. 3 4 ) However, there was no significant ( F = 0.21; df=45, 192; P = 1.000) interaction between treatment and time. The high mite infestation level treatment accumulated more mite days (2645) than all other treatments, which was significantly different from the control treatment throughout the sea son. At the end of the growing season, low had accumulated 1058 mite days, medium had 1140 mite days, and throughout the season both treatments were not significantly different from each but recorded more mite days than the control. Marketable Yield There were significant ( F = 22.83; df=3, 288; P < 0.0001) differences in marketable yields harvested from strawberry plants within different mite infestation levels during 2008/2009 growing season ( Fig. 3 5 ) Overall, the low infestation level had the highest m ean marketable yield, which was significantly different from all other treatments. The high infestation level had numerically the lowest yield and was lower than the control, but it was not significantly different from the medium mite infestation level. In 2009/2010, there was no significant ( P = 0.05) interaction between treatment and time, ( F = 0.41, df = 45,189; P = 0.9996); therefore, the main effects are reported. Significant differences in marketable yields harvested occurred at different T. urticae i nfestation levels (F =

PAGE 40

40 20.12; df = 3, 192; P< 0.0001 Fig. 3 6 ) and over time ( F = 156.77; df =15, 189; P < 0.0001). Overall, control treatment had significantly higher yields than all other treatments during the 2009/2010 growing season. Strawberry plants growing within medium and low infestation mite levels had similar marketable yields that were significantly higher than the high infestation level. A negative correlation was obtained between the marketable yields of strawberries and accumulated mite day for both years. This correlation was only significant (r = 0.49; P = 0.03) during the 2008/2009 growing season but not in 2009/2010 (r = 0.44; P = 0.09). During the 2008 2009 growing season, the effect of mite populations on marketable yields was observe d from 16 February 2009 until the end of the season (2 March) [Table 3 1]. A significant reduction in yields was detected when strawberry plants had accumulated 4924 mite days on the high infestation level treatment and when the low and control infestation s had 1890 and 133 mite days respectively. In 2009/2010, the effect of T. urticae on the marketable yields was only recorded on the last two sampling dates in April (5 and 12) when control had significantly more marketable yields than all the other treatme nts (Table 3 2). On both sampling dates, strawberry plants within the high mite infestation level had accumulated 1939 mite days as compared to those within the control with 210 mite days. In terms of marketable yields, the control and the medium mite infe station treatment levels had significantly higher marketable yields than all other treatments on both dates. Effe ct of Weather Factors on Twospotted Spider Mite Population Dynamics Both temperature and rainfall significantly affected T. urticae population growth in 2009/2010 growing season ( F = 30.82; R 2 = 0.1959; P < 0.0001) but not in 2008/2009 ( F =1.66; R 2 = 0.0092; P = 0.1921).

PAGE 41

41 Insect Pests and Beneficial Insects In both years, the main pests recorded were melon aphids, Aphis gosspyii Glover and whiteflie s Bemisia tabaci (Gennadius) (immature) on the leaves. There were no significant differences among the treatments for aphids and whiteflies in 2008/2009 growing season ( F = 2.38; df=3, 195; P = 0.0710 and F = 1.30; df=3, 195; P = 0.2766) and in 2009/2010 f or aphids ( F = 0.74; df=3, 268; P = 0.5291) and for whitefly immature ( F = 0.38; df=3, 268; P = 0.7642), respectively [Table 3 3]. The two main beneficial insects recorded were six spotted thrips, Scolothrips sexmaculatus (Perg. ) and big eyed bugs (BEB) ( G eocoris spp.). In 2008/2009, significantly more six spotted thrips were recorded within the high and medium mite infestation levels than in the control ( F = 3.12; df =3, 195; P = 0.0272) while the low infestation level was not significantly different from any other treatment (Table 3 3). On the contrary, in 2009/2010 no significant differences were observed on the six spotted thrips recorded across the treatments ( F = 2.37; df =3, 268; P = 0.0713) [Table3]. No significant differences were observed among in BEB eggs between treatments during the 2008/2009 growing season ( F = 1.28; df=3, 195; P = 0.2842). However, during the 2009/2010 field season, the number of BEB recorded in the high mite infestation treatment was significantly higher than the control ( F = 3.66; df=3, 268; P = 0.0130) [Table 3 3]. Discussion Twospotted Spider Mite Population Tetranychus urticae population increased with time during both growing seasons and more growth was observed when the temperature began to increase above 20 o C after mid February in 2008/2009 and in March in 2009/2010. Strawberry plants accumulated more mite days in the first growing season than in the second season (2009/2010). The reduction in mite days

PAGE 42

42 experienced during the 2009/2010 growing season was attributed to lo w / freezing temperatures experienced during the second growing season. In particular, January and February were the coldest months with 13 and 9 days below freezing temperatures, respectively in 2009/2010 growing season compared to 9 and 4 days in 2008/20 09 (Fig. 3 1). As a result, T. urticae development in 2009/2010 was heavily impacted in January and no treatment differences were observed in four out of the five sampling dates during that month. In fact, mite populations during 2009/2010 growing season b arely reached 500 mite days by end of January, as compared with the previous season where populations reached 2375, 1307, and 823 mite days in high, medium, and low infestation levels, respectively. Effe ct of Weather Factors on Twospotted Spider Mite Popul ation Dynamics Temperature is one of the main factors affecting T. urticae life cycle (Shih et al. 1976, White and Liburd 2005). Optimum temperatures for development are between 23 29 o C and developmental period from egg to mature adult can range from 7 to 12 days (Shih et al. 1976).Consequently, the low temperatures experienced during the 2009/2010 growing season significantly affected the mite population establishment and development. More rainfall was recorded during the second growing season than in the first season. During the 2008/2009 growing season, the maximum rainfall received 34.5 mm per week, with a 10.5 mm per week average for the entire season, while in 2009/2010 up to 83.6 mm of rainfall k. Rainfall patterns affect soil moisture levels, which are known to influence T. urticae population levels in field grown strawberries (White and Liburd 2005). For instance, White and Liburd (2005) found 3 times as many eggs and motiles on strawberry plan ts exposed to low or moderate soil moisture levels compared with high soil moisture. These weather related factors contributed significantly to the differences observed in T. urticae population during both growing seasons.

PAGE 43

43 Marketable Yield Strawberry yields were higher in the 2009/2010 growing season than in 2008/2009. Differences in yields could partially be attributed to differences in mite populations between the two seasons. As previously indicated, mite populations in the second year remained low during early growing period and only increased toward the end of the season. Sances et al. (1981) showed that the time of mite infestation on strawberry leaves can affect the amount of damage from mite injury. High mite numbers would be required to cause more da mage when infestation occurs late in the season. In 2009/2010, mite infestation levels increased at the end of the season and hence less damage were realized from mites over the course of that season. Yield reduction was detected when strawberry plants had accumulated 4923 mite days in 2008/2009; at this point mites per trifoliate leaf were averaging 220 motiles per leaf (equivalent to 73 mites per leaflet). In 2009/2010, yield reduction was observed when strawberry plants had accumulated 1938 mite days, av eraging 83 motiles per trifoliate leaf within the high infestation mite density level. Yield at this point was affected by mite days accumulated at the time of flower initiation (approximately 2.5 wk before). On these dates, strawberry plants had accumulat ed 2375 mite days in 2008/2009 and 1055 mite days in 2009/2010. In order to avoid losses, taking into consideration weather differences between the seasons, treatment applications would need to be done much earlier in 2008/2009 growing season than in 2009/ 2010. This would have been on 31 January 2009 when mites were averaging 45 motiles per trifoliate leaf and on 19 March 19 2010 at 50 motiles per trifoliate leaf. Our recommended treatment threshold is a conservative one compared to what has been reported b efore but it is worth mentioning the differences in growing periods and location. For instance, an ET level of 50 active mites per leaflet was reported as a level that would provide an effective treatment on a short day strawberry variety in California (Wy man et al. 1979).

PAGE 44

44 initiates flower buds when there is 14 h of daylight (or less), or does well in cooler temperatures and or if strawberry buds are produced in temperatures cooler than 15.5 o C (60 o F) (Ano nymous 2011). The strawberry variety evaluated in this study is Festival, which is also a short day plant, and in 2008/2009 flowering season ended abruptly after daily averages of 21.1 o C (70 o F) were reached in March. Development of economic threshold for T urticae on strawberries should also take into consideration factors such as weather, market price, and strawberry variety (Raworth 1986). In Florida, strawberry prices remain high until strawberry fruits from California become available on the market. Ma to spray against mite populations. In some cases, the grower might decide to leave the plants without the spray if the prices are not too favorable. Two predatory insects, six spotted thrips and big eyed bugs (BEB) were recorded during the two growing seasons. These predators tended to be high in abundance in plots where T. urticae populations were high in both season. These predatory insects are generalists in their feeding beh avior, a factor that has been cited as their limitation for use in the biological control systems (Oatman and Mcmurtry1966). Therefore, we do not think that these predatory insects contributed to any significant reduction in numbers of T. urticae on those plots. Furthermore, laboratory studies evaluating the predatory ability of big eyed bugs ( Geocoris punctipes Say) among other predatory insects on twospotted spider mites concluded that BEB preferred to feed on other plant insects (Rondon et al. 2004). Non etheless, six spotted thrips and BEB have been reported previously in field strawberry plantings (Raworth 1990, Fraulo et al. 2008). Therefore, their presence in un sprayed high TSSM plots maybe typical in this area.

PAGE 45

45 Aphids and whiteflies were present in t he early growing seasons, but as the temperatures decreased these two pests were observed to decrease. These two pests were not present in significant numbers in either growing season and therefore, it is likely they had only a negligible effect on the mar ketable yields. Furthermore, beneficial insects belonging to Chalcidoidea and Syrphidae that are associated with aphids and whiteflies were observed in the field and could have potentially contributed to their reduction in numbers. Conclusions This study c onfirms that control of T. urticae on field grown strawberries is important in preventing yield losses and economic damage. In addition, early infestation coupled with favorable temperatures for T. urticae will result in economic losses during mid season. Therefore, early treatment would be the best time to manage spider mites in order to prevent strawberry yield losses later in the season. Prevailing weather conditions are also an important factor to consider when making management decision for T. urticae A lower treatment threshold of 45 motiles per trifoliate leaf (15 mites per leaflet) is recommended during warm seasons as recorded in 2008/2009 as compared with 50motiles per trifoliate leaf (~17 mites per leaflet) for cooler seasons (2009/2010) when hig h mite numbers per leaf can be tolerated. Management decision may be affected by other factors such as strawberry market prices, crop condition (vigor), and targeted market (fresh or processing berries). Weather parameters such as temperature and rainfall, as well as host plant conditions can affect T. urticae populations physiologically and mechanically (direct effects) or indirectly through host quality changes (Knapp et al. 2006). In this study we report a lower threshold than that previously indicated fo r California; at 50 mites per leaflet by (Wyman et al. 1979). We recommend a moderate treatment threshold of 45 and 50 motiles per trifoliate leaf (~15 17 mites per leaflet) during warm weather and cooler

PAGE 46

46 months, respectively, which is much better than cal endar sprays that some growers are using (Nyoike et al. unpublished data). Further studies would be necessary to relate our ET numbers to the critical threshold (Cp) (threshold based on presence of any life stage on five percent of the leaflets collected) [ Mossler and Nesheim 2007] that are used by some growers in Florida.

PAGE 47

47 Table 3 1. Time effect for cumulative marketable yields of strawberries at different T. urticae infestation levels in 2008/2009 field growing season Effect Date F Value Pr> F Infestati on level*date 12/30/08 0.87 0.459 Infestation level*date 1/4/09 0.09 0.9671 Infestation level*date 1/8/09 0.22 0.8799 Infestation level*date 1/11/09 0.35 0.7872 Infestation level*date 1/14/09 0.39 0.7613 Infestation level*date 1/20/09 0.44 0.7244 Inf estation level*date 1/23/09 0.58 0.6303 Infestation level*date 1/27/09 0.87 0.459 Infestation level*date 1/31/09 1.47 0.2232 Infestation level*date 2/2/09 1.65 0.1788 Infestation level*date 2/6/09 1.53 0.2069 Infestation level*date 2/10/09 2.16 0.0923 Infestation level*date 2/16/09 2.70 0.0458 Infestation level*date 2/19/09 2.98 0.0317 Infestation level*date 2/23/09 3.21 0.0233 Infestation level*date 2/26/09 3.32 0.0204 Infestation level*date 3/3/09 3.71 0.012 Infestation level*date 3/5/09 3.80 0 .0107 Numerator degrees of freedom = 3 Denominator degrees of freedom = 288

PAGE 48

48 Table 3 2. Time effect for cumulative marketable yields of strawberries at different T. urticae infestation levels in 2009/2010 field growing season Effect Date F Value Pr> F Infestation level*date 12/29/2009 0.08 0.9686 Infestation level*date 1/7/2010 0.20 0.8992 Infestation level*date 1/14/2010 0.31 0.8149 Infestation level*date 1/19/2010 0.73 0.5328 Infestation level*date 1/25/2010 1.16 0.3282 Infestation level*date 2/1 /2010 1.20 0.3115 Infestation level*date 2/10/2010 1.20 0.3115 Infestation level*date 2/18/2010 1.13 0.3383 Infestation level*date 2/24/2010 1.33 0.2672 Infestation level*date 3/1/2010 1.42 0.2378 Infestation level*date 3/8/2010 1.18 0.3204 Infestati on level*date 3/16/2010 1.12 0.3403 Infestation level*date 3/26/2010 1.19 0.3132 Infestation level*date 3/29/2010 1.88 0.1349 Infestation level*date 4/5/2010 4.60 0.0039 Infestation level*date 4/12/2010 7.46 <0.0001 Numerator degrees of freedom = 3 De nominator degrees of freedom = 288

PAGE 49

49 Table 3 3. Mean number of insect pests and beneficial insects recorded on field grown strawberries in 2008/2009 and 2009/2010 growing season in Citrus County, FL Mean SEM # 2008/2009 Insect pests Beneficial insec ts Treatment Aphid a Whitefly nymph b Six spotted thrips c BEB egg d High 6.8 0.9 1.2 0.3 2.3 0.5a 1.6 0.3 Medium 9.3 0.9 0.7 0.3 1.7 0.5a 1.3 0.3 Low 7.6 0.9 1.3 0.3 1.5 0.5ab 1.2 0.3 Control 5.7 0.9 0.6 0.3 0.3 0.5b 0.8 0.3 2009/2010 High 1.6 0.6 3.6 0.9 0.4 0.1 1.9 0.3a Medium 1.3 0.6 2.9 0.9 0.3 0.1 1.5 0.3ab Low 2.4 0.6 4.3 0.9 0.4 0.1 1.9 0.3a Control 2.1 0.6 3.8 0.9 0.02 0.1 0.8 0.3b Big eyed bug. Means with the sa me letter are not significantly different ( P = 0.05) # Standard error of mean 2008/2009 a = F = 2.38; df = 3, 195; P = 0.0710 b = F = 1.30; df = 3, 195; P = 0.2766 c = F = 3.12; df = 3, 195; P = 0.0272 d = F = 1.28; df = 3, 195; P = 0.2842 2009/2010 a = F = 0.74; df = 3, 268; P = 0.5291 b = F = 0.38; df = 3, 268; P = 0.7642 c = F = 2.37; df = 3, 268; P = 0.0713 d = F = 3.66; df = 3, 268; P = 0.0130

PAGE 50

50 Figure 3 1. Mean weekly temperatures recorded in 2008/2009 and 2009/2010 strawberry growing seasons at Citra, Florida. Weeks = time since initiation of sampling. Figure 3 2. Weekly total rainfall recorded in 2008/2009 and 2009/2010 strawberry growing seasons at Citra, Florida -10 -5 0 5 10 15 20 0 5 10 15 20 25 mean weekly min temperature (oC) Week 2008/2009 2009/2010 0 10 20 30 40 50 60 70 80 90 50 51 52 53 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Weekly total rainfall (mm) Calendar Week 2008/2009 2009/2010

PAGE 51

51 Figure 3 3. Accumulated mite days over time for different T. urticae densities on field grown strawberries for the growing season 2008/2009 Figure 3 4. Accumulated mite days over time for different T. urticae densities on field grown strawberries for the growing season 2009/2010 0 2000 4000 6000 8000 10000 12000 14000 30-Dec 13-Jan 27-Jan 10-Feb 24-Feb 10-Mar Accumulated mite days Sampling date High Medium Low Control 0 1000 2000 3000 4000 Accumulated mite days Sampling date High Low Medium Control

PAGE 52

52 Figure 3 5. Mean harvested marketa ble yields (Kgs) from different T. urticae densities on field grown strawberries for the 2008/2009 growing period. Columns with same letters are not significantly different ( P = 0.05). Figure 3 6. Mean harvested marketable yields (Kgs) from different T urticae densities on field grown strawberries for the 2009/2010 growing period. Columns with same letters are not significantly different ( P = 0.05). 0 1 2 3 4 5 6 7 8 9 High Medium Low Control Mean marketable yields per plot Treatment bc c a b 0 1 2 3 4 5 6 7 8 9 10 High Medium Low Control Mean marketable yield per plot (Kg) Treatment c b b a

PAGE 53

53 Figure 3 7. Harvested marketable yields from different T. urticae densities on field grown strawber ries for the growing period between December 2008 and March 2009 Figure 3 8. Harvested marketable yields from different T. urticae densities on field grown strawberries for the growing period between December 2009 and March 2010 0 2 4 6 8 10 12 14 16 12/30 1/13 1/27 2/10 2/24 3/10 Marketable yields(Kgs) Sampling date High Medium Low Control 0 5 10 15 20 25 12/29 1/12 1/26 2/9 2/23 3/9 3/23 4/6 4/20 Marketable yield (Kg) High Medium Low Control

PAGE 54

54 CHAPTER 4 VISIBLE/NI R REFLECTANCE SPECTR OSCOPY FOR TWOSPOTTE D SPIDER MITE ( TETRANYCHUS URTICAE KOCH) DETECTION AND PREDICTION ON STRAWB ERRY LEAVES Introduction Sampling and monitoring are two important steps in determining pest population levels to guide for sound management programs. Early pest detection is therefore preferred so that management actions can be carried out before pest populations begin to cause economic losses. Twospotted spider mite [TSSM], Tetranychus urticae Koch (Acari: Tetranychidae) is the main mite pest affecting field grown strawberries ( Fragaria ananassa Duchesne) in Florida (Poe 1971, Mossler and Nesheim 2007). Tetranychus urticae usually inhabit the lower side of the leaf from where they feed from causing damaging the mesophyll cells as they remove c hlorophyll (Kielkiewicz 1985, ) On strawberry leaves, low densities of T. urticae mainly cause damage to the spongy mesophyll tissues while high densities intensifies the damage causing severe injury to the palisade parenchyma cells (Sances et al. 1979, T Removal of chloroplasts (hence chlorophyll) through TSSM feeding results in a decrease in the use of radiant energy with consequent reduction in vegetative growth and yield (Sances et al. 1979, Sances et al.1982, Kielkiewicz 1 985, Reddall et al. 2004, Nyoike and Liburd, unpublished data). The removal of chlorophyll also causes the leaf to have white or yellow spots also known as a T. urticae on host plants may include injection of toxic saliva (Kielkiewicz 1985), chemical changes in the leaf decreased transpiration on strawberry (Sances at el.1979, 1982) and on cotton (Redall et al. 2004).

PAGE 55

55 Tetranychus urticae are relatively small mites and may remain unnoticed until they cause vi sible damage to the leaves (Opit et al. 2009). As a result, strawberry growers tend to spray more often than would be necessary to avoid unforeseen losses. Currently, field sampling and detection for spider mites on strawberry leaves involve the use of mag nifying lens to look at mites on the undersides of the leaves. This procedure is labor intensive and some growers try to avoid it due to the strain on the eye. Development of a rapid and reliable method for early detection of spider mites and other strawbe rry pests would facilitate timely implementation of control measures where they are most needed. Although the use of remote sensing is still in its exploratory stages in pest damage detection, various studies have shown it can be used to detect insect, mi te and disease related damages on plants (Riedell and Blackmer 1999, Fitzgerald et al. 2004, Mirik et al. 2006, Reisig and Godfrey 2006, Reisig and Godfrey 2007, Fraulo et al. 2009, Jones et al. 2010). In these studies, spectral changes as a result of art hropod feeding or disease infection and their associated injury on the plant was to differentiate damaged from undamaged leaves. Basically, a spectral reflectance (the way leaves interact with light) signature for healthy leaves is known between 0.4 2.4 m. Principal features in this region include a high reflectance in the shortwave near infrared (NIR) wavelengths (0.7 1.0m) due to internal scattering by leaf cells and five absorption regions; one at 0.4 0.7 m due to chlorophyll, and four others at 0.97, 1.20, 1.40, and 1.94 m due to the bending and stretching of the O H bond in water (Gates et al. 1965, Curran 1989). Leaf damage as induced by pests affects how the leaf interacts with light energy and therefore spectral changes in response to damag e by arthropods could be used as an indicator of crop vigor and health. Various spectral vegetation indices (SVI) derived either from airborne

PAGE 56

56 hyperspectral imagery or ground based platforms have used to derive changes on leaves/plant canopy as induced by arthropod damage (Reisig and Godfrey 2006, Reisig and Godfrey 2007, Mirik et al. 2007, Carroll et al. 2008). Mirik et al. (2007) showed that a strong to a weak (0.91 0.01) positive relationship correlation coefficient between Russian wheat aphid (Diuraphis noxia Mordivilko) densities on wheat and various spectral indices. Reisig and Godfrey (2007) investigated the important wavelengths for detecting cotton aphid ( Aphis gossypii Glover) and T. urticae damaged cotton leaves using a portable hyperspectral spec trometer. They found that spider mite and aphid damaged leaves had a higher reflectance in the near infrared particularly at wavelength 0.85 m as compared to non damaged leaves. Fraulo et al. (2009) were able to detect T. urticae damage on strawberry leav es with a high coefficient of determination of 0.85 and root mean square error of 12.2 mites per leaf using partial least squares regression. Collectively, these studies demonstrate that visible/NIR reflectance spectroscopy could offer a rapid cost effecti ve method for monitoring strawberry damage by TSSM. Two experiments were conducted to study spectral changes as induced by TSSM damage on two strawberry varieties with the ultimate goal of developing prediction models for TSSM numbers on the leaves. The o bjectives of the study were to use visible/near infrared region (VIS/NIR) diffuse reflectance spectroscopy to predict TSSM numbers causing injury on strawberry leaves using partial least squares (PLS), to evaluate different pre pretreatment transformations for the spectral data, and to validate the prediction models developed using an independent data set. Materials and Methods Experimental Strawberry Plants The greenhouse experiments were conducted in the spring of 2009 and 2012 using two different represe

PAGE 57

57 and Nematology Department, University of Florida, Gainesville, Florida. Plants were grown in 1 liter plastic pots (Nurseries Supplies Inc. Chambersburg, PA) using a 1:1 mixtur e of Metromix (SunGro Horticultural Distributors Inc., Bellevue, WA) and Jungle Growth potting mix with fertilizers (0.21% N, 0.07% P 2 O 5 0.14% K 2 O and 0.10% Fe) [Jungle Growth, Statham, GA]. In addition, plants were top dressed with a nitrogen based liqu id fertilizer at the rate of 1 tablespoon (15 mL) per 3.75 L of water every other week during the growing period. The plants were watered every day up to one week post transplanting and every third day thereafter until the end of the experiment. Liquid cop per fungicides (Southern AG, Palmetto, FL) in rotation with pyraclostrobin (Cabrio EG, BASF Ag Products, Research Triangle Park, NC) were sprayed twice during the growing season as a prophylactic treatment for fungal diseases on strawberry plants. If need the growing season to control aphids and whiteflies. Twospotted Spider Mite Colony and Inoculation To achieve the desired TSSM infestation levels (treatments) on the strawberry p lants, spider mites were artificially introduced from an existing laboratory colony. A colony of twospotted spider mite was maintained in the greenhouse and / or laboratory reared on beans or strawberries. T o achieve low, medium, and high mite infestation s as would naturally occur in the field, with 5, 15, and 25 TSSM per leaf, respectively were applied on three fully grown strawberry leaves growing in the greenhouse. Initial mite infestations were achieved by placing a mite infested leaf disc with a known number of mites on the test plants and allow the mite to crawl onto the new host. In addition, a control treatment with zero mites was established. Each treatment was placed in mite free screen cage to avoid cross contamination among the treatments.

PAGE 58

58 Exper iment Set up and Twospotted Spider Mite Sampling The first experiment was conducted during the spring 2009 between 30 Jan. and 24 April the experiment and sampling (for TSSM and leaf reflectance) was conducted at weekly intervals for five weeks. Similarly, during the 2011/2012 season, sampling was conducted at weekly int erval and 32 leaves were scanned weekly for four weeks from 27 Jan. to Feb. 24. For both experiments, leaf sampling was initiated two weeks after introducing the mites onto the test plants. Leaflets were taken each week (45 in 2009 and 32 in 2012) from the plants and placed into individual labeled bags (Ziploc, S. C. Johnson & Son Inc., Racine, WI). The number of mites on each leaflet were counted under a dissecting microscope (Leica MZ 12 5, Leica Microsystems, Houston, TX) and then leaflets were placed b ack into their respective bags for spectral reading. Reflectance Measurements To reduce dehydration spectral data were collected within 2 hours of leaflet collection at the Agricultural and Biological Engineering Laboratory at the University of Florida, G ainesville. A spectrophotometer (Cary 500 Scan UV VIS NIR spectrophotometer, Varian Inc., Palo Alto, Ca.) with an integrating sphere (DRA CA 5500, LabSphere, Inc., North Sutton, NH ) was used to collect spectral reflectance from 200 to 2500 nm, in 1 nm incr ement. The diameter of the sample port was 5 cm, and the coating of the material inside the integrating sphere was polytetrafluoroethylene (PTFE) (Spectralon, Labsphere, Inc., North Sutton, NH). Spectral values from each leaflet were collected from 19.6 cm 2 per leaflet. Before every scan, each leaflet was positioned on the sample port such that spectral values were collected from the areas that best represented that mite infestation level. A 50 mm diameter PTFE disk was used to obtain the

PAGE 59

59 optical reference standard each time before spectral measurements were taken. Ultraviolet (UV) and mercury lamps were used as light sources and were warmed up for 30 min before starting the scanning session. A total of 2301 variables were generated for each strawberry leafl et by taking their reflectance measurement from the top (adaxial) surface of the leaf. Data Analysis Each observation had 2300 absorbance values (from 200 to 2500 nm); altogether there were 120 samples in 2009 and 128 in samples in 2012 experiments, respe ctively. The reflectance values were converted into absorption spectra using Eqn. 1 (Williams and Norris, 2001). A = log (1/R) (4 1) where A is absorbance and R is reflectance. Correlation Coefficient Spectra Correlation coefficient spectra were computed between absorbance and actual mite numbers on the strawberry leaves at each wavelength using PROC CORR (SAS Institute, 2002). High correlation coefficients indicate that at those specific wavelengths, TSSM numbers can be calculated by measuring the absorb ance (Yoder and Pettigrew Crosby 1995, Bogrekci and Lee 2005). To identify the important wavelengths in TSSM damage detection, wavelengths with high correlation coefficients (both negative and positive valu es) were selected. Therefore wavelengths correlation coefficient spectra. Both spectra were very different in terms of the r value s obtained and therefore to select the important wavelengths (with high |r | values), |r | was lowered to 0.2 for the Chandler variety, whereas 0.6 was used for the Florida Festival in 2012.

PAGE 60

60 Pre processing Treatments Pre processing transformations carried out included Norris gap derivatives (Norris and Williams, 1984), first and second Savitzky Golay derivatives and Savitzky Golay smoothing (Savitzky Golay, 1964). These transformations were used on the raw absorbance spectra to remove random noise. Norris g ap (N G) derivative was carried out using a 9 nm search window band whereas a search window of 5 observations to the left and the right was used for first and second Savitzky Golay (S G) derivatives using polynomial order 1 and 2 pre treatment methods. The se methods were selected based on the most commonly used transformations (Vasques et al. 2008a & b) but the search window was preferred to avoid losing too much information through the smoothing processes. After transformation, the first and the last wavel engths were found to be noisy and therefore removed. Wavelengths between 355 to 2400 nm and 355 to 2200 nm were used for the N G and S G transformations, respectively and were selected for further analysis. Pre processing treatments were carried out using UnScrambler software v9.5 (Camo, AS, Norway). Partial least squares (PLS) regression (Martens and Ns 1989) was used to relate changes in mite population on the strawberry leaves with changes in spectral reading corresponding to wavelengths between 200 2500 nm using PLS in Unscrambler (Unscrambler Program, Camo, Norway). The data was randomly divided into a calibration and a validation set of 89 and 31 samples for Chandler, and 90 and 38 samples for Florida Festival varieties in 2009 and 2012, respective ly. The PLS model was generated using PLS1 (PLS, Unscrambler), an algorithm for sample data with one Y variable (mite counts) with a full cross validation on the calibration data set. During the model calibration, the data were inspected for any outliers that could potentially affect the model. One outlier was identified from the 2009 data set and two observations in 2012 data set that were removed and the models recalculated. The numbers of latent variables retained

PAGE 61

61 were chosen based on the minimum root m ean square error of cross validation (RMSECV), [PLS Unscrambler]. The validation data set was used to evaluate robustness of the models by predicting the unknown mite numbers. PLS Regression Coefficients In PLS regression coefficients are used to summari ze the relationship between the all the predictors (spectral reading at the wavelengths) with the response variable (mite numbers per leaf). The cumulative importance of each wavelength can be calculated from the B matrix of the X loadings and X weights us ing equation 2 (Min and Lee 2005). The size of coefficient is used to determine how important that X variable is in predicting Y and hence the bigger the coefficient the more important that variable is (PLS, Unscrambler). (4 2) Where B is the cumulative importance of each wavelength, w is the X weight, p is the X loading, and q is the Y weight. Important variables (wavelengths) in both strawberry varieties TSSM spe ctra were selected 2002) was conducted to derive the relationship betw een the important variables and TSSM counts on the strawberry leaves to develop a prediction equation for each variety. The mite count data was log transformed to meet normality requirements for regression before analysis The accuracy of the equations was tested in predicting mite numbers for independent data sets for each variety. PLS Models Comparison Six models were generated for each strawberry variety, Chandler in 2009 and Florida Festival in 2012. Comparison between calibration models were carried out using c orrelation

PAGE 62

62 coefficient of calibration and of cross validation ( r c and r cv ), and root mean squared error of calibration (RMSEC) [Eq. 3] (Jamshidi et al. 2012) The ability to accurately predict unknown samples was evaluated based on the calculat ed coefficient of determination (R 2 ), root mean squared error for cross validation (RMSECV or RMSEP) [Eq. 4] (Jamshidi et al. 2012) and the RPD is calculated from standard deviation of validation set (2010) describes RPD values between 2 RPD is a better method for comparing models because the range (dispersion) of the da ta is considered during calculation unlike using R 2 (Dunn et al. 2002). RMSEC = (4 3) RMSECV or RMSEP = (4 4) RMSECV or RMSEP = (4 5) where / is the number of observation in the calibration and validation set; and are the predicted and observed values of the observation, respectively; is the standard deviation of the validation set, and = number of observations Results and Discussion Twospotted Spider M ites Population Figure 1 shows strawberry leaves with varying TSSM damage from which spectral characteristics were collected from, while Table 4 1 shows the de scriptive statistics associated with the damage categories. The mean number ( standard error, SE) of TSSM on the leaves of (t = 0.87, P = 0.387) from 108 20.9 for the validation data set. Similarly, the mean number of

PAGE 63

63 differ significantly (t = 0.43, P = 0.666) from 76.3 14.9 for the validation data set. Th us the two means were acceptable for use in PLS analysis. In both years, mite distribution was positively skewed with median lower than mean value, and so a log transformation was applied before the regression analysis. Strawberry Leaf Absorbance Spectra Typical absorbance spectra for raw absorbance and the preprocessed spectra for wavelengths between 355 and 2200 nm are provided are shown in Fig. 4 1. Overall, the averages of the four TSSM infestation levels (control, low, medium and high) were very simi lar and could not be differentiated on the graphs (Fig. 4 1) unlike previously observed by Fraulo et al. (2009).In their study they reported that spider mite damage caused a characteristic increase in between 0.80 1.3 m and less changes in the green visible region (0.52 0.58 m) on strawberry leaves. The equipment used to collect spectral data in the two studies was different and also the mite numbers recorded on the leaves could have been different Cotton leaves that were damaged by the cotton aphid ( Aphis gossypii Glover) and the two spotted spider mite ( T. urticae ) had either increased (Reisig and Godfrey 2007) or reduced (Reisig and Godfrey 2006, Fitzgerald et al. 2004) leaf reflectance in the N IR region. On wheat, studies showed a decrease in reflectance in the NIR region and an increase in reflectance in the visible region due to greenbug [ Schizaphis graminum (Rondani)] and Russian wheat aphid damage as compared to non damaged canopies (Riedell and Blackmer 1999, Mirik et al. 2006, Mirik et al. 2007). The pre processed TSSM strawberry damaged and undamaged leaf spectra are shown in Fig. 41b 1d. Savitzky Golay smoothing (SG S) [Fig 1b] was very similar to the raw absorbance (Fig. 4 1a) showing a similar trend in absorption with key regions such as nitrogen absorption (~

PAGE 64

64 550 nm), reduced absorption in the short wave NIR region (700 1000 nm) and two water absorption bands at around 1430 and 1914 nm. On the other hand, Savitzky Golay first deriva tive (SG 1) [Fig. 1c], and Norris gap (N G) derivative [Fig. 1d] showed similar absorbance peaks at 516, 560, and 699 nm with slight variations at 1350 and 1875 (N G) and at 1389 and 1881 nm on the SG 1 curve. In addition, an absorbance peak was observed at around 873 nm on the SG 1 spectra curve but that was missing on the N G curve. Some of the peaks observed in this study have been reported before as important wavelengths associated with arthropod damage on the leaves. Frau lo et al. (2009) reported that wave NIR region between 800 1300 nm were the most diagnostic characteristic of spider damage on strawberry leaves. The main difference between the studies was the area of the leaf scanned to collect spectral data. In their st udy spectral data was collected from an area of 3.2 cm 2 versus 19.63 cm 2 that is almost the actual size of a strawberry leaflet. This could explain the lack of mite infestation level separation in the NIR region. The low absorption in the NIR region (735 1341 nm) al. 1965, Curran et al. 1989). Correlation Coefficient Spectra Selection of the wavelengths were highly correlated with TSSM damage a total of 56 wavel engths selected from the correlation coefficient spectra collected from Florida Festival (2012 study) and 27 wavelengths from Chandler (2009 study) strawberry varieties. PLS Regression Coefficients Regression coefficients for the two strawberry varieties are shown in Fig. 4 4. The two peaks at wavelengths 419, 443, 456, 483, 548, 670, 776, 789, 889, and 933 nm and 710, 879, 1904, 1979, and 1996 nm an indication

PAGE 65

65 that those wavelengths were important in detecting TSSM damage on the strawberry leaves. Wavelengths within the range 419 680 nm are related to chlorophyll absorption while wavelengths greater than 1000 n m are related to water absorption due to leaf water content. Multiple regression analyses results indicated that the important variables had either a positive or a negative correlation with the mite numbers on the strawberry leaves. The regression model f : log mite = 3.12 + 50.75* wv419 + 22.70*wv443 69.21*wv456 43.65*wv483 8.13*wv548 +50.12*wv670 +102.75*wv776 8.55*wv789 41.55*wv889 44.49*wv933. Figure 4 4A shows prediction results obtained for the Chand ler variety using regression with a weak coefficient of determination (R 2 ) of 0.42. On 2 = 0.543) [Fig. 4 4B]. These wavelengths for the Florida Festival var iety were selected from the Savitzky Golay first derivative pre processed spectra while the raw absorbance was used to the Chandler variety. In both cases, these were the best models from the PLS results. The regression iety with the important wavelengths was as follows: log mite = 18.11 96.24* wv695 +54.75*wv692 + 26.35 *wv710 + 14.96*wv879 8.32*wv1904 21.01*wv1979 + 33.06*wv1996 PLS Models Comparison The results of the calibration and validation PLS models of the twospotted spider mite (TSSM) on strawberries are presented in Table 4 2. There were six models generated for each selected wavelengths (27 for Chandler an d 56 for Florida Festival), and four models from a combination of the two pre processing methods (derivative and smoothing). The lowest RMSEC for the Chandler variety was obtained from the Savitzky Golay first derivative using polynomial order 1(SG 1) [0.2 4] followed by the untransformed raw absorbance [0.27]. Good calibration

PAGE 66

66 models are characterized by large r and smaller RMSEP (Natsuga and Kawamura 2006). In this respect, the raw absorbance model was one of the best models for the Chandler strawberry var iety (2009) with the highest correlation coefficient (r = 0.71) and a relatively low RMSEP (0.58) (Table 4 2 & Fig. 4 5a).When evaluated on the ability to predict TSSM numbers, this model had a low coefficients of determination (R 2 = 0.20) [Fig. 4 6a] and residual predictive deviation (RPD) value of 1.29 (Table 4 2). A similar RPD (1.29) was obtained from the PLS model with SG 1 transformation that had had an R 2 of 0.46 (Table 4 2). Even though PLS model with Norris gap (NG) pre processing had the highest R 2 (0.54), it had RPD value (1.25) was not the highest. As suggested by Dunn et al. (2002), R 2 is not the best measure when comparing predictive ability of different models. Therefore, based on the RPD values for the PLS developed for Chandler strawberry va based on their RPDs values between 1.5 and 2.0 [Table 4 2] The PLS model with SG 1 was the best prediction model for Florida Festival with RMSEC = 0.34, r c = 0.94, RMSEP = 0.54, r cv = 0.82, R 2 = 0.61, and RPD = 1.82 (Table 4 2, Figs. 4 5b & 4 6b). Comparable RPD values were obtained from PLS models with SG S and NG preprocessing treatments or the raw absorbance spectra. The three models had RPD > 1.6. Critical RPD in VIS/NIR analysis for foliar pests has not been set but based on the RMSE (17 mites per leaf) obtained from the best model there is potential in usin g this model. Currently, most of the RPD range available have been set for NIR analysis for soil characteristics as used by Chang et al. 2001, Dunn et al. 2002, Vasques et al. 1.6 2.0 is

PAGE 67

67 ability of models with less than 1.4 may not be reliable and for those between1.4 and 2.0, the models can be improved during calibration. In com paring model performance, starting from the best model to the worst in each year, ranking was as follows raw absorbance = SG 1 > N G > SG S > SG 2 > selected wavelengths for Chandler variety in 2009, while for Florida Festival in 2012 as SG 1> SG S > N G > raw absorbance > SG 2 > selected wavelengths. Other than the differences in the varieties used between the two years, all other procedures were the same and therefore this shows that spectral differences exist between strawberry varieties. Important wavel engths as indicated by PLS coefficient show similar wavelength regions are affected by TSSM damage in the two strawberry varieties spectra (Figs. 4 4). The different predictive model obtained for each variety indicated that in a mixed variety strawberry pl anting, predictive models should be generated independently according to the number of varieties planted. Most strawberry growers use different varieties on their farm for various reasons including early maturing varieties, market preferences (fresh vs pro cessing fruits) or targeted market (local market for U pick vs. for export market). The PLS model with SG 1 and N G pre processing treatment performed consistently well in both years of this study. Therefore, use of derivatives proved to be one of the best pre processing for TSSM spectra on strawberries. Similar results have been reported before, where f irst derivative gave much better prediction than raw reflectance or log 1/R while using leaf spectra to predict nitrogen and chlorophyll content on fresh bi gleaf maple leaves by Yoder and Pettigrew Crosby (1995). Similarly, Vasques et al. (2008) found that S G and N G derivatives were the best pre processing transformations in a study where 30 methods in combination with different multivariate methods were co mpared. In the analysis, only one search window was used during transformation and possibly calibration models could be improved by evaluating pre

PAGE 68

68 processing transformation combinations (size of the search window), as suggested by Chang et al. 2001. Howeve r, during pre processing, there is loss of information in smoothing the spectra curves but it is not known how much of the lost information is useful in calibration (Unscrambler v9.5) and this process should be carried out cautiously. Relatively good resul ts were obtained with the raw spectra (RPD = 1.68 in 2012) and this means also that raw spectra data could also be used to give satisfactory results without transformation. Conclusion s Twospotted spider mite detection on strawberry leaves from the spectral data was achieved from a model that gave an error 17 mite/leaf. It seems that differences in spectral changes in different strawberry varieties and therefore different predictive model should be developed for each variety especially in mixed strawberry pl antings. The accuracy of the model could be mm) and an error of 17 mites per leaf is could be within accuracy range especially when the mite population is very high or in its initial early stages with many protonymphs (first instar of spider mites life stage). In addition, strawberry leaves can support a very high number of mites (1000 mites per leaf) and error of 17 mites per leaf is acceptable (personal observation ). However, mite infestation on strawberry plants during the early growing season will affect the yields (Sances et al. 1982) and some growers want zero mites on the leaves. Twospotted spider damage detection and therefore prediction model developed will v ary among strawberry varieties and observed between Florida Festival and Chandler varieties. The best prediction model developed for the Chandler variety from multiple linear regression analysis with the important wavelengths/variables that have the highe st coefficient of determination. For the Florida festival, the entire pre processed spectra gave the best model with an R 2 of 0.606.

PAGE 69

69 The use of derivatives in particular, SG 1and N G proved to be the best pre processing method for removing noise from the s trawberry leaf spectra to predict TSSM followed by SG S method. These methods gave a consistent performance on both years. It may not be advisable to pre select wavelengths for TSSM detection. The PLS models from the selected wavelengths performed the wors t in both years (with the lowest RPD) indicating that some important information (wavelengths) could have been removed during the wavelength selection. The RPD

PAGE 70

70 Table 4 1. De scriptive statistics for both calibration and validation data sets for twospotted spider mites on counted per strawberry leaf on two strawberry varieties (Chandler and Florida Festival) 2009 (Chandler) 2012 (Florida Festival) Calibration Validation Ca libration Validation Control 3.6 1.5 1.0 0.6 0.0 0.0 0.0 0.0 Low 74.0 12.8 24.7 5.4 25 2.6 24.7 5.4 Medium 126.8 16.3 95.6 40.9 77.6 3.0 71.3 5.5 High 161.9 37.6 195.2 43.0 187.5 13.8 211 19.3

PAGE 71

71 Table 4 2. Resu lts of calibration and prediction of the PLS models with different preprocessing methods for twospotted spider mites on strawberries LVs Calibration set Cross Validation Validation RPD Year Wavelengths Preprocessing RMSEC r c RMSEP r cv R 2 200 9 200 2500 None 8 0.27 0.94 0.58 0.71 0.20 1.29 Selected (27) None 4 0.56 0.73 0.63 0.63 0.34 1.18 355 2200 Savitzky Golay 1st derivative (polynomial order 1) (SG 1) 4 0.24 0.96 0.58 0.71 0.46 1.29 355 2200 Savitzky Golay 2nd derivative (polynomia l order 2) (SG 2) 2 0.30 0.93 0.62 0.65 0.26 1.2 355 2200 Savitzky Golay smoothing (SG S) 5 0.55 0.75 0.62 0.66 0.36 1.21 355 2400 Norris gap derivative (N G) 4 0.28 0.94 0.60 0.69 0.54 1.26 2012 200 2500 None 5 0.41 0.90 0.58 0.79 0.54 1.68 Sel ected (56) None 1 0.71 0.66 0.73 0.64 0.51 1.34 355 2200 Savitzky Golay 1st derivative (polynomial order 1) (SG 1) 4 0.34 0.94 0.54 0.82 0.61 1.82 355 2200 Savitzky Golay 2nd derivative (polynomial order 2) (SG2 2) 4 0.12 0.99 0.71 0.67 0.46 1.38 355 2200 Savitzky Golay smoothing (SG S) 4 0.50 0.85 0.55 0.82 0.58 1.79 355 2400 Norris gap derivative (NG) 3 0.42 0.89 0.56 0.81 0.61 1.76 r c or r cv c orrelation coefficient of calibration or validation ; RMSEC or RMSEP root mean squared error (RMSE for cross validation or RMSE for prediction;RPD residual predictive deviation

PAGE 72

72 Figure 4 1. twospotted spider mite damage. A) control with no damage, B) low infestat ion showing some damage on the top left side of the leaf, C) medium infestation most of the damage along the mid vein, D) high infestation with TSSM damage visible on almost the entire leaflet.

PAGE 73

73 Figure 4 2. A bsorbance spectra within visible /NIR spectrum with four t wospotted spider mite four infestation levels on strawberry leaflets ( Florida Festival A) raw absorbance, B) its S avitzky Golay smoothing (SG S), C) Savitzky Golay first derivative (SG 1) and D) N orris gap (N G) deriva tive. 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 355 855 1355 1855 2355 Raw Absorbance Wavelength (nm) Control High Low Medium 0 0.2 0.4 0.6 0.8 1 1.2 1.4 355 855 1355 1855 2355 Transformed absorbance (SG S) Wavelength (nm) Control High Low Medium A B

PAGE 74

74 Figure 4 2. Continued -0.02 -0.01 0 0.01 0.02 0.03 355 855 1355 1855 2355 Transformed Absorbance (SG 1) Wavelength (nm) Control High Low Medium -0.04 -0.02 0 0.02 0.04 355 855 1355 1855 2355 Transformed absorbance (NG) Wavelength (nm) Control High Low Medium D C

PAGE 75

75 Figure 4 3. Partial least squares regression coefficient showing important wavelengths for the best PLS models A ) raw absorbance spectra in 2009 and B ) for the Savitzky Golay first derivative pr e processed spectra in 2012 for Chandler and Florida Festival strawberry varieties, respectively. -20 -10 0 10 20 30 wv355 wv655 wv955 wv1255 wv1555 wv1855 wv2155 Regression coefficient (B) Wavelength (nm) -20 -10 0 10 20 wv355 wv655 wv955 wv1255 wv1555 wv1855 wv2155 Regression coeficient (B) Wavelength (nm) A B

PAGE 76

76 Figure 4 4. Twospotted spider mite prediction using multiple linear regressions of the important wavelengths associated with TSSM spectra for strawbe rry variet ies, A) Chandler and B) Florida Festival y = 0.544x + 0.955 R = 0.42 0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3 Predicted TSSM (Log mite) Counted TSSM (Log mite) y = 0.53x + 0.603 R = 0.543 -0.5 0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3 Predicted TSSM (Log mite) Counted TSSM (Log mite) A B

PAGE 77

77 Figure 4 5. Cross validation results for the best predictive models where A) is the entire raw absorbance spectra for Chandler and B) for the Savitzky Golay first derivative pre processed s pectra for Florida Festival stra wberry variety. A B

PAGE 78

78 Figure 4 6. Scatter plot of predicted log mite versus counted log mite fo r the best predictive models for the two strawberry varieties where A) shows the entir e raw absorbance spectra for Chandle r and B) the Savitzky Golay first derivative pre processed spectra for the F lorida Festival y = 0.452x + 1.0597 R = 0.201 -1 -0.5 0 0.5 1 1.5 2 2.5 3 3.5 0 0.5 1 1.5 2 2.5 3 Predicted TSSM (Logmite) Counted TSSM (Logmite) y = 0.783x + 0.335 R = 0.606 -1 -0.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 0.5 1 1.5 2 2.5 3 Predicted TSSM (Log mite) Counted TSSM (Log mite) A B

PAGE 79

79 CHAPTER 5 FACTORS AFFECTING RE FLECTANCE SPECTROSCO PY AS A MEANS OF DETECTING TWOSPOTTED SPIDER MITE DAMAGE O N STRAWBERRY LEAVES Introduction The twospotted spide r mite (TSSM) [ Tetranychus urticae Koch] is a pest of many agricultural crops including strawberries (Oatman and McMurtry 1966, Gimnez Ferrer et al. 1994). Spider mites feed on the underside of leaves damaging the mesophyll cells as they remove chloroplas ts that contain chlorophyll (Kielkiewicz 1985). Due to their small size, T. urticae may remain unnoticed until they cause visible damage to the leaves (Opit et al. 2009). Incorporating leaf reflectance into the existing monitoring and sampling programs fo r TSSM on strawberries could make damage detection faster. Remote sensing provides a non destructive, rapid, and cost effective means of identifying and quantifying plant stress from changes in spectral characteristics of the leaves (Mirik et al. 2006). S t udies have demonstrated that insect and mite damage can be detected using leaf reflectance as shown for greenbug [ Schizaphis graminum (Rondani)] on wheat (Mirik et al. 2006), European corn borer [ Ostrinia nubilalis (Hbner) ] infestation on corn (Carroll et al. 2008), cotton aphid ( Aphis gossypii Glover) and spider mite damage on cotton (Fitzgerald et al. 2004, Reisig and Godfrey 2006), and spider mite on strawberries (Fraulo et al. 2009). Generally, green leaves have a typical reflectance signature between 0.4 2.4 m with a high reflectance in the short wave near infrared (NIR) wavelengths (0.7 1.0 m) due to internal scattering by leaf cells and five absorption regions; one at 0.4 0.7 m due to chlorophyll, and four others at 0.97, 1.20, 1.40, and 1.94 m due to the bending and stretching of the O H bond in water (Gates et al. 1965, Curran 1989). Changes in the leaf, such as those caused by insect damage, can alter diffuse reflectance of the leaf, causing an increase in reflectance within the visible reg ion and a reduction in reflectance in the NIR region (Nilsson 1995). Removal of

PAGE 80

80 chloroplasts (hence chlorophyll) through TSSM feeding results in a decrease in the use of radiant energy and hence reduced vegetative growth and yields (Sances et al. 1979, San ces et al.1982, Kielkiewicz 1985, Reddall et al 2004, Nyoike and Liburd, unpublished data). In addition, TSSM feeding damage also affects the interaction of the leaf with light energy in the NIR wavebands. Cellular damage by TSSM in the mesophyll layer int erferes with the ability of the leaf to reflect NIR energy (Jensen 2005). Several studies have shown an inverse relationship between TSSM numbers (hence damage) and plant productivity (Sance et al. 1979, Sances et al. 1981, Nyoike and Liburd unpublished). Leaf spectral changes in the green wavelength and NIR band could be used to detect TSSM damage on strawberry leaves (Fraulo et al. 2009). However, the interaction between N fertilizer and chlorophyll removal as induced by TSSM feeding has not been determin ed in strawberries. It is known that majority of leaf N is contained in chlorophyll molecules and therefore between leaf chlorophyll content and leaf N content are closely related (Yoder and Pettigrew Crosby 1995). Chlorophyll content in corn leaves double d as N fertilizer was increased (Daughtry et al. 2000). Most previous studies e.g. Mirik et al. 2006, Reisig and Godfrey 2006 and Fraulo et al. 2009) involved collecting leaf reflectance from the adaxial (top surface) surface of the leaf. However, insect feeding damage can also cause differences in adaxial and abaxial (under side of the leaf) leaf reflectance between 400 and 2500 nm (Carroll et al. 2008). Furthermore, spider mites mostly inhabit the abaxial side of the leaf surface on strawberries from whe re they feed from and only come to the adaxial surface if the population is very high. Sampling programs for TSSM on strawberries involve turning the leaf to scout for spider mites. We hypothesize that leaf reflectance from the abaxial side of the leaf cou ld offer better precision in determining TSSM

PAGE 81

81 infestation level. Currently growersuse presence/absence sampling program that involves using a hand lens to observe for spider mites on the under of the leaves (Mossler and Nesheim 2007). Use of damage infesta tion categories (no mite, low, medium, high) is desirable to strawberry grower for quick mite detection on the plants (Fraulo et al. 2009). In this study, we evaluated spectral changes on strawberry leaves as affected by TSSM infestation levels growing wit h and without nitrogen fertilizers. We also studied the differences in spectral reflectance collected from the adaxial or abaxial side of the leaf surface. The main goal was to establish if N fertilizer could be a confounding factor in detecting and classi fying TSSM damage on strawberry leaves using spectral characteristics. Secondly, we wanted to determine if collecting leaf reflectance readings from the abaxial side of the leaf had a better correlation with TSSM damage on the leaves. The objectives of the study were: 1) to compare spectral changes in mite damaged leaves from strawberry plants growing with and without nitrogen fertilizer, 2) to discriminate TSSM damage categories using spectral reading from the abaxial and the adaxial sides of the leaf, and 3) to identify the important wavelengths that could be used to classify different infestation levels. Materials and M ethods Strawberry Plants Greenhouse experiments were conducted in the spring of 2009, 2011, and 2012 using University of Florida, Gainesville. Unless otherwise indicated, plants were grown in 1 liter plastic pots (Nurseries Supplies Inc. Chambersburg, PA) using a 1:1 mixture of Metromix (SunGro Horticul tural Distributors Inc., Bellevue, WA) and Jungle Growth potting mix with fertilizers (0.21% N, 0.07% P 2 O 5 0.14% K 2 O and 0.10% Fe) [Jungle Growth, Statham, GA]. In addition, plants were top dressed with a nitrogen based liquid fertilizer at the rate of 1

PAGE 82

82 tablespoon per 3.75 L of water every other week during the growing period. Liquid copper fungicide (Southern AG Palmetto, FL) and pyraclostrobin (Cabrio EG, Research Triangle Park, NC) were sprayed in rotation twice during the growing season as a prophy lactic treatment for fungal diseases on strawberry plants. If needed, acetamiprid (Assail 70 WP, ) was sprayed early in the growing season to control aphids and whiteflies. Twospotted Spider Mite Colony and Inoculation In order to ach ieve the desired TSSM infestation levels (treatments) on the strawberry plants, spider mites were artificially introduced from an existing laboratory colony. A colony of twospotted spider mite was maintained in the greenhouse and or laboratory reared on b eans or strawberries. T o achieve low, medium, and high mite infestations as would naturally appear in the field, 5, 15, and 25 TSSM per leaf, respectively were applied to the strawberry plants in the greenhouse. Initial mite infestations were done by placi ng a mite infested leaf disc with a known number of mites on the test plants to allow the mite crawl onto the new host. In addition, a control treatment with zero mites was established. Each treatment was placed in mite free screen cage to avoid cross cont amination between the treatments. Strawberry leaf s ampling Sampling was initiated two weeks after introducing the mites onto the test plants. Ten strawberry leaves were picked from each treatment and placed in individual labeled bag (Ziploc, S. C. Johnson & Son Inc., Racine, WI). The number of mites on each leaf were counted under a dissecting microscope (Leica MZ 12 5, Leica Microsystems, Houston, TX) before taking the spectral reading. Spectral data were collected within 2 hours after collection to reduce dehydration at the Agricultural and Biological Engineering Laboratory at the University of Florida, Gainesville. A

PAGE 83

83 spectrophotometer (Cary 500 Scan UV Vis NIR spectrophotometer, Varian Inc., Palo Alto, CA) with an integrating sph ere (DRA CA 5500, LabSphere, Inc.) was used to collect spectral reflectance from 200 to 2500 nm, in 1 nm increment. The diameter of the sample port was 5 cm, and the coating of the material inside the integrating sphere was polytetrafluoroethylene (PTFE). A 50 mm diameter PTFE disk was used to obtain the optical reference standard each time before spectral measurements were taken. Ultraviolet (UV) and mercury lamps were used as light sources and were warmed up for 30 minutes before collecting spectral data. A total of 2301 variables were generated for each leaf taking the reflectance measurement from the top (adaxial) side of the leaves unless otherwise noted. Comparison of Spectral Changes in Mite damage on Strawberry Plants Growing with and without Nitrog en Fertilizer. The effects of nitrogen fertilizer on spectral values of strawberry leaves were evaluated between 8 Dec 2010 and 25 Feb 2011, using four treatments: mite infested with and without nitrogen fertilizer and a control (zero mites) with and witho ut nitrogen fertilizer, replicated four times. Thirty six pots were set up with a potting mix (SunGro Horticultural Distributors Inc., Bellevue, WA), and half of them (18 pots) had a a teaspoon (~1 gm) of 10 10 10 (N P 2 O 5 K 2 O) [Southern States CO OP, Cor dele, GA] incorporated into the soil before transplanting strawberry plants After plant establishment (~ 3.5 wks), a liquid based nitrogen fertilizer was applied to these plants every other week at the rate of 1 tablespoon /3.75 L until the end of the exp eriment. The rest of the pots were supplied with teaspoon of the 10 10 10fertilizer incorporated at planting, and no other fertilizers were applied for the rest of the experiment. Thirty two leaves were sampled weekly collecting mite data and the leaf re flectance from Jan 20 to Feb 24 2011 at weekly intervals.

PAGE 84

84 Discriminating TSSM Damage Categories Using Spectral Reading from the Abaxial and the Adaxial Side of Strawberry Leaf This experiment was conducted between 15 Dec 2011 and 24 Feb 2012. Eighteen plan ts were used for this experiment as varying mite population were introduced as explained above to achieve a no mite, low, medium and high mite infestations levels. Plant maintenance and sampling is as explained above. Twospotted spider mite data and their respective spectral readings were collected weekly from 27 Jan to 24 Feb 2012 where 8 leaflets were sampled each week for a period of four. Sixty four leaflets were sampled resulting in 128 observations of spectral readings collected from both the abaxial and adaxial side of the strawberry leaf. Pre processing Treatments The reflectance values were converted into absorption spectra using Eqn. 1(Williams and Norris 2001). A = log (1/R) (5 1) where A is absorbance and R is reflectance. Data Analysis Correlat ion coefficient ( r ) spectra were computed between absorbance and actual mite numbers on the strawberry leaves at each wavelength using PROC CORR (SAS Institute, 2002). High r indicates that, at those specific wavelengths, TSSM numbers can be calculated by measuring the absorbance (Yoder and Pettigrew Crosby 1995, Bogrekci and Lee 2005). In order to identify the important wavelengths in TSSM damage detection, wavelengths with r greater than 0.6 or less than 0.6 (or 0.7 for the abaxial surface) were selected from the correlation coefficient spectra. The value or |r| was adjusted so that at least 30 wavelengths were selected and not more than 60. The selected wavelengths were then subjected to a stepwise regression (PROC REG STEPWISE, SAS) and discriminant ana lysis (PROC DISCRIM, SAS) to identify

PAGE 85

85 the significant wavelengths with the SAS options for entry and exit ( SLSTAY and SLENTRY) set at to specify the significance level for entry into and removal out of the model Principal component analysis (PCA) was carr ied out on the spectral data to reduce the data dimensionality using PROC PRINCOMP (SAS 2004). Linear Discriminant Analysis (LDA) was applied on the retained PCs in order to classify mite damage levels from the spectral data. Linear discriminant analysis i s a statistical technique to classify objects into groups based on the differences between the groups such that variance within the group is minimized while variance between the groups is maximized. Damage categories from Fraulo (2007) were modified to inc lude control (zero mites), low/moderate (10 100 mites/ leaf), and h categories. To select the PCs greatest difference between treatments, PROC STEPDISC (SAS, 2004) was used on the retained PCs and PROC DISCRIM (SAS, 2004) with CROSSVALIDATION option was applied on the identified PCs to classify mite d amage levels on the strawberry leaves. Results and Discussion Comparison of Spectral Changes in Mite Damaged Strawberry Plants Growing with and without Nitrogen Fertilizer Twospotted spider m ite population Descriptive statistics for TSSM counted on the st rawberry leaves are shown in Table 5 1. No mites were recorded on the control treatment growing with and without fertilizer and TSSM numbers recorded on the mite inoculated strawberry plants were comparable (85.9 and 89.0 per leaflet) between the two treat ments (TSSM infested plants growing with and without N fertilizer) (Table 5 1). In addition, the ranges of TSSM numbers mite in the two treatments were high due to delayed mite population establishment as affected by low temperatures recorded in January 20 11.Because mite populations were very low (less than 5 mites per leaf) during the first week of

PAGE 86

86 the experiment, only the data collected in the last three weeks (96 observations) of the experiment were used for analysis. Reflectance/absorbance Principal com ponent analysis reduced the spectral data into 6 PCs that captured 92.8% of the data variation, with the first PC explaining 66.3% of the total variation. The first six PCs were used to discriminate the mite infested leaves into four categories: mite damag ed growing with (Mite Fert) or without fertilizer (Mite No Fert) or non damaged leaves with (Con Fert) or without fertilizer (Con No Fert). Discriminant analysis correctly classified 54 out of 94 egory had the highest number of 2). Nine leaves were correctly classified as control (no mites) with fertiliz without fertilizer (5) or mite infested with (1) or without (5) fertilizer (Table 5 2). Almost half of hich was the highest misclassification in the data set. Cross validation results indicated an error rate of 44.6% when classifying new samples. The error rate in classifying TSSM damaged leaves from strawberry plants growing with and without fertilizer was relatively high and no clear pattern was observed. There seems to be no major difference between plants growing with and without fertilizer whether infested or un infested with TSSM. We speculate that this may be different on field grown strawberries wher e plants are grown for an extended period, longer than the 3 months used in this experiment. The experiments were carried out for only half of the growing period and nitrogen requirements could be different when the plants are left to bear fruits for the e ntire season. A regular strawberry growing season in north central Florida runs from mid September to March or April

PAGE 87

87 the following year depending on the environmental conditions. It is worth mentioning that the plants in this experiment were applied with s ome base fertilizer for plant establishment and did not produce a lot of fruits. The important wavelengths that could be used to distinguish between mite damaged or undamaged strawberry plants growing with and without fertilizer were selected based on the wavelengths that showed high correlation with mite numbers. Using the criterion r 0.5, a total of 57 wa velengths were selected, and further screening using a stepwise regression identified 6 wavelengths including 611, 620,627, 640,674, and 691 nm as the most important wavelengths. Step wise discriminant analysis with 57 se lected wavelengths indicated that 8 (645, 643, 689, 642, 647, 614, 658, and 650 nm) of those wavelengths were the most important in classifying mite infestation levels. Both methods revealed wavelengths within the same range (611 to 691 nm) as the most imp ortant wavelengths to distinguish mite infested plants growing with and without N fertilizer. A relationship exists between chlorophyll content and N content in the leaf (Daughtry et al. 2000, Yoder and Pettigrew Crosby 1995). Chlorophyll in the leaves is responsible for light absorption in the blue (450 nm) and red (670 nm) wavelength regions and therefore could be used to detect low N levels in the leaves. Our study shows that wavelengths in the visible range, particularly red, were the best indicators in differentiating damaged leaves from undamaged strawberry plants growing with and without fertilizer. Wavelengths within the visible waveband were identified as the best predictors for chlorophyll and short near infrared (900 1300 nm) as the best predicto r for N on fresh bigleaf maple (Yoder and Pettigrew Crosby 1995).Daughtry et al. (2000) found an inverse relationship between reflectance and chlorophyll content in corn at wavelengths 550 and 715 nm but minimal changes at wavelengths 450 and 670 nm. Our r esults

PAGE 88

88 indicate that wavelengths in the 600s nm were depicted as the important wavelengths in differentiating the different TSSM damage categories, which we believe was as a result of mite damage. This may be an indicator that chlorophyll removal by TSSM d amage can be differentiated from N deficiencies in the strawberry leaves. However, further investigation needs to be conducted to confirm this result. Discriminating TSSM Damage Categories using Spectral Reading from the Abaxial and the Adaxial Side of Str awberry Leaf Twospotted spider mite population The descriptive statistic of the TSSM counted from the strawberry leaves are presented in Table 5 3. Twospotted spider mite numbers varied between 10 to 97 and 118 to 406 mites per leaf with means of 51.4 and 231.8 for the low and high infestation damage levels, respectively (Table 5 3). The control plant had no mites for the entire sampling period. Strawberry reflectance/absorbance A typical absorbance spectra observed on the abaxial and adaxial spectral are shown in Figure 5 1. Generally, diffuse reflectance is higher on the adaxial side of the leaf with notable differences at UV VIS (200 648 nm) and the two water absorptions bands (~ 1423 and 1910 nm) where leaves scanned on the adaxial side and therefore sh owed a higher absorption at those wavelengths. Overall, the graphs shows similar characteristics defined by a nitrogen absorption band near 550 nm, low absorption region (NIR around 688 1380 nm) and at the water absorption bands (in 1423 and 1940nm). Diffe rences in TSSM infestation levels are more obvious on the abaxial spectra where the control had the highest absorbance and high infestation with the lowest absorbance (Figure 5 1). Leaf absorbance was negatively correlated with the level of mite damage.

PAGE 89

89 Pr incipal component analysis on the adaxial spectral data reduced the data dimension into 7 principal components (PCs) that captured 92.3% of the variation in the data. The 7 retained PCs were used to create a model that correctly classified 80, 70, and 50% of the control, high and low strawberry damaged leaves into their infestation levels categories (Table 5 4). The low infestation level had a 50% misclassification where 10 of the samples were classified as either high or control. Whereas none of the contro l samples were classified as high infestation, two of the high damaged leaves were classified as control (Table 5 4). Cross validation results indicate an error rate of 33.3% when using the model to classify new samples with up to a 50% misclassification o f the low infestation levels. Alternatively, PCA on the abaxial spectral data reduced the data into 5 PCs that explained 95.7% of the variation in the data. Four of the retained PC were used in LDA and resulted in 85, 70, and 35% correct classification of the control, high, and low damage levels (Table 5 5). Similarly, the low infestation level had the highest misclassification with 7 and 6 observations placed in the control and high infestation categories, respectively. There was a clear separation of high and control infestation levels and none of the highly damaged leaves were classified as control and vice versa. Overall, 36.7% of the strawberry leaves would be misclassified and the lowest accuracy (65%) was recorded with the low mite infestation (Table 5 5). Collecting spectra readings from abaxial side of the leaf resulted in a higher accuracy for delineating between high and non infested strawberry leaves (control) but a lower accuracy in classifying low infested. Both abaxial and adaxial spectra would have a problem in classifying low infested leaves from non infested leaves (control) and also from the high infested leaves. This is a drawback in using spectral readings to classify mite damaged strawberry leaves. Growers would want to detect the damage at very early stages so that the appropriate measures

PAGE 90

90 would be taken to control TSSM populations. Furthermore, time of TSSM infestation on strawberries is very critical and early infestation causes more injury to the plants than late infestation. Studies c onducted in north central Florida recommended a treatment threshold for TSSM at 45 and 50 motiles per trifoliate leaf (~15 17 mites per leaflet) during warm weather and cooler months (Chapter 3). It is therefore important that the proposed sampling tool (u sing spectral characteristics) would be able to detect TSSM damage resulting from less than 20 mites per leaflet. The day neutral cultivar Selva has a very low economic threshold (ET) level of 1 mite per leaf (Walsh et al. 1998) compared to a moderate leve l of 50 mites per leaflet (150 mites/leaf) for a short day strawberry variety in California (Wyman et al. 1979). Such low ET values (1 mite/leaf) would limit TSSM detection using spectral reflectance, whereas it would be applicable on the short day strawbe rry variety. It is worth mentioning that ET vary from region to region, cultivar, stage of crop growth, time of season, control options, and market prices (Pedigo 1996, Walsh and Zalom 1996) and therefore the use of a portable spectrometer that can be moun ted on tractor may still be an option in detecting TSSM damage on strawberries in some situations. From the correlogram (Figure 5 2) the maximum r values obtained were 0.73 at 581 nm and 0.60 at 530 nm for the abaxial and adaxial sides of the leaf, respect ively. There were notable differences between the two spectra especially on the visible wavebands where a negative correlation was observed between mite numbers and absorbance on the adaxial spectra and a positive correlation around the same wavelengths on the abaxial spectra. Selecting wavelengths that had |r| = 0.6 resulted in 38 wavelengths in the adaxial (top) spectra (between 763 498 nm) and 458 wavelengths in the abaxial side of the strawberry leaf. When |r| was increased to 0.7 for the abaxial spectr a, a total of 49 wavelengths between 737 nm and 1980 nm were selected.

PAGE 91

91 Important wavelengths were selected using stepwise regression on the retained wavelengths 693 and 763 nm from the adaxial spectra and 748, 752, 759, 766, and 1980 nm from the abaxial sp ectra (Table 5 6). On the other hand, if stepwise discriminant was used, wavelengths 637, 642,672, and 763 nm were selected from the adaxial spectra and 738, 745, 763, 771, and 1964 nm from the abaxial spectra. Only wavelength 763 nm was selected by both m ethods as an important wavelength that could be used to classify mite damage on the leaves (Table 6). O ur results show that within the NIR, wavelength around 763 nm can be used to detect TSSM damage on strawberry leaves. In addition, we find that the water absorption waveband at 1964 nm (due to leaf water content) could be also used to detect mite damage on the leaves. Leaf water content affects light reflectance at wavelengths > 1000 nm. Conclusions In summary, although the abaxial had a better accuracy in classifying high infested leaflet and non infested leaflets, the results for detecting low TSSM infested leaflet were lower than those obtained from the adaxial spectra. Therefore collecting leaf reflectance from the abaxial surface may not be desirable b ecause growers would want to detect TSSM at low infestation levels. Wavelengths in the NIR region and water absorption band were important in TSSM damage detection. Conversely, TSSM infestation level classification on strawberry plants growing with and wit hout fertilizer had a high error rate and therefore further studies are needed to confirm the results. Wavelengths near the red band were determined as the most important wavelengths in classifying TSSM infestation categories.

PAGE 92

92 Table 5 1. Descriptive stati stics for the twospotted spider mite counted on strawberry leaves growing with and without nitrogen fertilizer. With Fertilizer Without Fertilizer Control Mite infested Control Mite infested Mean 0.0 85.9 0.0 89.0 Std Error 0.0 23.1 0.0 27.0 Minimum 0.0 0 0.0 1 Maximum 0.0 477 0.0 519 Median 0.0 34 0.0 22 Skewness 1.9 1.8 Kurtosis 3.4 2.3 Std Dev 0.0 120.1 0.0 140.1 Range 0.0 477 0.0 518

PAGE 93

93 Table 5 2. Classification summary of Linear Discriminant Analysis (LDA) with cross validation of principal components (PC) from the PCA of the spectral data collected from mite damaged and undamaged strawberry plants. Results shows that 45, 70, 48 and 59% of cont Number of Observations and Percent Classified into Mite Damage Levels From tr eatment Con Fer t Con No Fert Mite Fert Mite No Fert Total Con Fert 9 5 1 5 20 45 25 5 25 100 Con No Fert 3 14 2 1 20 15 70 10 5 100 Mite Fert 2 3 13 9 27 7 11 48 33 100 Mite No Fert 4 1 6 17 27 15 4 .7 22. 59 100 Total 18 23 22 31 94 19 24 23 3 3 100 Priors 0.21 0.21 0.29 0.29 Error rate (C) 0.55 0.25 0.48 0.41 0.41 Error rate (CV) 0.55 0.3 0.52 0.41 0.45

PAGE 94

94 Table 5 3. Descriptive statistics for the twospotted spider mite counted on the strawberry leaves used for abaxial and adaxial spectra study to classi fying twospotted spider mite damage. TSSM Infestation levels Statistic control Low High N 20 20 20 Mean 0.0 51.4 231.8 Std Error 0.0 6.6 18.1 Minimum 0.0 10 118 Maximum 0.0 97 406 Median 0.0 51.5 227.5 Skewness 0.02 0.49 Kurtosis 1.6 0.25 Std Dev 0.0 29.7 80.9 Range 0.0 87 288

PAGE 95

95 Table 5 4. Classification summary of Linear Discriminant Analysis (LDA) with cross validation of principal components (PC) from PCA of the spectral data collected from the adaxial (top) surface of strawberry lea ves with three twospotted spider mite infestation levels. Results show that 80, 70, and 50% of the control, high and low damage categories, respectively were correctly classified into their infestation levels categories. Number of Observations and Percent Classified into Mite Damage Levels From treatment Control High Low Total Control 16 0 4 20 80 0 20 100 Low 5 5 10 20 25 25 50 100 High 2 14 4 20 10 70 20 100 Total 23 19 18 60 38.33 31.67 30 100 Priors 0.33 0.33 0.33 Error rate (C) 0.15 0.2 5 0.45 0.28 Error rate(CV) 0.2 0.3 0.5 0.33

PAGE 96

96 Table 5 5. Classification summary of Linear Discriminant Analysis (LDA) with cross validation of principal components (PC) from PCA of the spectral data collected from the abaxial (under)surface of strawberry leaves with three twospotted spider mite infestation levels. Results show 85, 70, and 35% of the control, high and low damage categories, respectively were correctly classified into their infestation levels categories Number of Observations and Percent Classified into Mite Damage Levels From treatment Control High Low Total Control 17 0 3 20 85 0 15 100 Low 7 6 7 20 35 30 35 100 High 0 14 6 20 0 70 30 100 Total 24 20 16 60 40 33 27 100 Priors 0.33 0.33 0.33 Error rate (C) 0.1 0.25 0.65 0. 33 Error rate(CV) 0.15 0.3 0.65 0.37

PAGE 97

97 Table 5 6 S electing important wavelengths used in classifying twospotted spider mites from spectral data collected from the adaxial and abaxial surface of strawberry leaves. Side of the leaf scanned Adaxial (top) Abaxial (Under) |r| 0.6 0.7 Wavelengths selected 38 49 Step discriminant analysis (nm) 627, 763, 642, 637 745, 738,1964, 771, 763 Stepwise regression (nm) 693, 763 748, 1980, 766, 759, 752

PAGE 98

98 Figure 5 1. Typical absorption spectr a from collecting spectral readings from the adaxial (Ad) and abaxial (Ab) surface of strawberry leaves infested with twospotted spider mites at three infestation levels including a control (C), low (L), and high (H). Figure 5 2. Correlogram spectra of spectral data collected from the adaxial (top) and abaxial (underside) surface of strawberry leaves 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 200 700 1200 1700 2200 Absorbance Wavelength (nm) H_Ad H_Ab L_Ad L_Ab C_Ad C_Ab -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 200 700 1200 1700 2200 Correlation Coefficient Wavelength (nm) Adaxial (top) Abaxial (under)

PAGE 99

99 CHAPTER 6 SPATIAL AND TEMPORAL DISTRIBUTION OF TWOS POTTED SPIDER MITE A ND ITS PREDATORY MITE, NEOSEIULUS CALIFORNI CUS, ON STRAWBERRIES USIN G GEOSTATISTI CS Introduction Neoseiulus californicus McGregor (Acari: Phytoseiidae) is an important predatory mite for the twospotted spider mite (TSSM), Tetranychus urticae Koch [Acari: Tetranychidae], which is the main mite pest damaging strawberries ( Fragaria ananas sa Duchesne) in Florida. Biological control as a means of managing pest is desirable because it is safe to consumers, workers, and poses no hazardous effects to non target organisms in the environment. Therefore, the use of N. californicus to control T. ur ticae on strawberries has been used widely in different parts of the world (Mari and Zamora 1999, Greco et al. 1999, Sato et al. 2007, Cakmak et al. 2009). In Florida, N. californicus has been recommended as a potential biological control agent for T. urti cae on strawberries (Rhodes and Liburd 2006, Fraulo and Liburd 2007, Liburd et al. 2007). Collectively, the studies in Florida demonstrated that N. californicus was able to persist in the field and give season long control of TSSM. Furthermore, Fraulo and Liburd (2007) showed that early release of N. californicus during the strawberry growing season was the best time to release these mites in the field as compared to mid or late season releases. However, the spatial distribution of N. californicus in relati on to its prey T. urticae and the ability to move and disperse on strawberry plantings after localized inoculative releases has not been established on Florida strawberries. The ability of natural enemies to disperse within a cropping system is an importan t aspect in utilizing the organism in biological control (Buitenhuis et al. 2010). The patchy distribution of T. urticae (Greco et al. 1999) in the field requires that a predatory mite is able to move and disperse in order to locate its host. Spatial varia bility studies of N. californicus and its prey T.

PAGE 100

100 urticae would elucidate their population dynamics and determine how predatory mites disperse can be used to explain spatial variation of insect pests and their distribution using the relationship between mean and variance but not their location (Leibhold et al. 1993). Geographic information systems (GIS) and geostatistics have been suggested in ecological studi es to examine spatial uses sample values and their locations to characterize their spatial patterns and predict sample values at un sampled locations (Lie also recognizes the spatial relationships between the locations and that samples close to each other are more similar than those farther away (Tobler 1970, Isaaks and Srivastava 1989). The use o f geostatistics to study spatial distribution of insect is not a new tool in pest management. Geostatistics has been applied in spatial distribution studies for Gypsy moth [ Lymantria dispar (L.)] (Liebhold et al. 1991), Lygus hesperus Knight on lentils ( S chotzko and enerate prescription maps for management zones for corn rootworm ( Diabrotica virgifera virgifera LeConte), [ Park and To llefson 2005], spatial sampling programs for corn rootworm ( Park and To llefson 2006) and Colorado potato beet le ( Leptinotarsa decemlineata L) (Weisz et al. 1995), and to determine trap spacing for insect monitoring programs (Rhodes et al. 2011). One of the commonly used methods in describing spatial relationships in variability studies is through the use of a sem ivariogram ( 1989, Hubbard et al. 2001). A semivariogram (variogram) is a graphical representation of semivariance (calculated as half the variance of the differences between sample points) and plotted against the distance that separate them ( Isaaks and Srivastava, 1989). A model is fitted to the semivariogram and is used to estimate unknown sample values. Several methods can be used

PAGE 101

101 to estimates values at un sampled locations (also known as interpolation) [Bolstad 2008]. The principle b ehind interpolation is that things that are close together are more similar than those that are farther apart (Tobler 1970). Interpolation methods are based on either the distance between the unknown value and its neighboring known values or both the dist ance and statistical relationships between the points (Fleischer et al. 1999). Alternatively, in Inverse Distance Weighting [IDW], points that are far from the unknown value are weighed less as compared to the ones that are nearby (Ess and Morgan 2003). Kr iging is a geostatistical method that utilizes values of the known samples and the statistical relationships between the observations to predict unknown sample value (Schotzko s, kriging also uses weights determined by the degree of relatedness of pairs of point as a function of distance between them (Fleischer et al. 1999). The goal of the study is to incorporate geostatistical techniques into sampling and monitoring for T. urt icae and its predatory mite in order to understand their temporal and spatial distribution patterns in the field. This is important to understand how N. californicus disperse and move in the field in response to T. urticae population. The objectives of the study were: 1) to determine the spatial distribution of N. californicus and T. urticae in a strawberry field, 2) to detect temporal changes in distribution patterns for the mites over the growing season, and 3) to map their distribution in order to deter mine dispersal and movement of N. californicus Materials and Methods Strawberry Plants and Study Area Description growing seasons in 2009 2011 using a 1,265 m 2 (0 .13 ha) sampling grid. The sampling grid had 8 strawberry beds. Strawberry beds were 0.7 m wide at the top, 0.25 m high and 115 m long,

PAGE 102

102 each planted with a double row of strawberries. The two strawberry varieties grown were transplanted on 24 Sept. 2009 and 29 Sept. 2010, respectively. Management practices including weed management, pest and disease control and harvesting Sampling A sampling grid with 104 sampling points at spacing of 5 m within the row and additional 25 points were randomly established within the sampling grid in each year. Simply, there were 26 sampling points on each strawberry bed established on every other row of strawberry of the eight rows in the exper imental plot. The study area borders and sampling point locations were geo referenced using a Trimble GeoXT GPS receiver, Trimble, Sunnyvale, CA. Sampling for both TSSM and N. californicus was conducted every two weeks but the sampling interval increased t o 3 weeks during cold months (Jan/Feb when TSSM and predator activity was low) or reduced to one week when the mite / predator population increased. At every sampling point, 3 strawberry leaflets were collected and placed into individual Ziploc bag and the n taken to the laboratory to assess Leica MZ 12 5, Leica Microsystems, Houston, TX) from 5 Nov. 2009 to 23 March 2010 and from 30 Nov. to 7 April 2011, during the 2009/2010 and 2010/2011 straw berry growing season, respectively. Geostatistical Analysis Data pre processing The sampling point data from the GPS receiver was imported into ArcMap 9.3 (ESRI 2003). The GPS data collected in the Geographic Coordinate System [GCS] World Geographic

PAGE 103

103 System WGS 1984 were projected to Universal Transverse Mercator (UTM) 17 N, coordinate used to join the sampling data to the georeferenced sampling points. Descrip tive statistics including mean, standard error, skewness, and kurtosis were generated for the strawberry mite data. Structural analysis of the data Spatial structure of the mite data in 2010/2011 was determined using a semivariogram, a graphical represen tation that shows spatial correlation of sample points in relation to their neighboring points. The semivariogram function can be defined as: (6 1) where is the semivariance at distance h, Z( x i ) and Z( x i +h ) are measu red values at x i and x i + h and N( h ) is the total number of sample pairs within the distance interval h The three important features of a semivariogram are the range (marks the distance where data are spatially correlated), sill (the semivariance when ran ge is attained), and nugget effect (represents measurement errors or data variation at very fine scale) [Isaaks and Srivastava 1989]. Semivariograms were only generated for mite populations sampled in the 2010/2011 strawberry growing season but not in 2009 /2010 season due to low mite populations. Temporal spatial variability of mites was identified using semivariograms for each sampling date separately. A semivariogram for the pooled mite populations for the entire season was also generated to examine the r elationship between the two mite species (predator and prey). Prediction of number of mites at unsampled locations was achieved by using oridinary kriging (OK) and IDW. Ten neighbors with a minimum of two neighboring points was used in a

PAGE 104

104 one sector to pred ict for unsampled locations for TSSM and at least two points and a maximum of points in a one sector to predict for N. californicus To avoid overcrowding on the semivariogram due to too many points, semivariances are grouped into bins also known as neighbors was used as done in OK and weights assigned to known points when predicting unknown values set to as p = 2. The mathematical formula used in IDW is: where is the unknown value to predict at location N is the number of measured (known samples) to be used in prediction, are the weights assigned to the sample that are going to be use d in prediction and is the known value at location Cross validation errors (mean prediction and root mean square prediction errors) were used to compare the two interpolation methods as well as to evaluate their prediction performan ce for each sampling date. Results 2009/2010 Growing Season Twospotted spider mite population was very low during the 2009/2010 growing season and therefore N californicus was not released in the field. Table 6 1 shows total counts of TSSM motiles and egg s from strawberry leaves during the season. Only two sampling points had spider mites on the first day of sampling, and by 23 March 2010, the number of sites had only increased to 11 points. Total counts of TSSM (motiles and eggs) increased drastically and on 23 March 2010, TSSM were aggregated on a few leaves recorded at two sampling points with total mite counts of 231 and 123 per 3 leaflets. Similarly, the highest number of TSSM eggs (748) was counted from one of these sampling points.

PAGE 105

105 The only two insec t pests recorded were melon aphids ( Aphis gossypii Glover) and whitefly immatures ( Bemisia spp.) that were also in low numbers per three leaflets (Table 6 1). There were no natural enemies recorded during the growing season. 2010/2011 Growing Season Althou gh sampling was initiated on 30 November 2010, TSSM infestation was not observed before 7 January 2011, when four sampling points showed positive for TSSM (Table 2 & Fig 6 1). At this point, mite numbers were averaging 0.33 motiles per leaf. As the season progressed, TSSM population was observed to spread to other neighboring points from the hot spot observed on January 7 (Fig 6 1) to occupy the entire field by the end of the season. On 9 Feb, spider mites were observed to spread along the row and occupy approximately 25 m. Mite dispersal also occurred to neighboring rows (Fig 6 1A). By mid March, TSSM occupied almost half of the north part of the field (Fig 6 1B). Spider mite population spread to cover most of the field by early April with only a few sa mpling points with no mites in the middle of the field. Twospotted spider mites eggs were only counted during the first three sampling dates after which the population became too high to count, and percent leaf coverage was adopted (data not shown). In ter ms of TSSM population density per leaf, the highest mean (82.2 17.5) was observed on 14 March 2011, and then population decreased slightly on 30 March and at the end of the season mite population was observed to increase again (Table 6 2 & Fig 6 2A). O n the other hand, N. californicus population was low and the highest number per strawberry leaflet for the season was recorded on 30 March (Fig 6 2B).Figure 6 3 shows inoculative sites for N. californicus on the strawberry field and population distributio n on two sampling dates in the 2010/2011 growing season (1 & 30 March 2011). Neoseiulus californicus were released at every 20 m along the strawberry row but their population growth and establishment was much slower compared to TSSM. The predatory mites we re introduced in the

PAGE 106

106 field on January 19 th but their motiles or eggs were not recovered on the strawberry leaves until 1 March (Table 6 2). On this sampling date, N. californicus motiles were only counted at 2 sampling points (Fig 6 3). By mid season (14 March), predatory mites were only recorded at 4 sampling points in the field and they did not reach an average of 1 motile per leaf (Table 6 2). On 30 March, N. californicus were recorded on 19 sampling points mostly in the north central region of the fie ld (Fig 6 2). At the end of the growing season, the mean number N. californicus recorded per strawberry leaflet had decreased to 0.3 from 0.52 and the number of sampling points positive for this mite also decreased (Table 6 2). Similar to 2009/2010 growi ng season, the only two insect pests that were recorded were melon aphids and whitefly immatures also in very low numbers. Mean counts per leaf for the two insect pests were < 1.0 per leaf, and a total of 86 aphids and 6 whitefly immatures were recorded fo r the entire season. Spherical models were used in semivariograms generated to study the spatial structure of TSSM population counted between 7 Jan to 7 April 2011 and for the pooled population combining all the sampling dates together. Both mite species (TSSM & N. californicus ) populations were positively skewed (Table 6 2) and therefore log transformation was applied before variography. Semivariograms modeled on the first three sampling dates (7 Jan, 27 Jan and 9 Feb) were almost pre nugget with high nug get to sill ratio (72, 69 and 78% Table 6 3 ) The spatial range on these dates was above 50 meters. Although the spatial range was the same for 7 and 27 Jan both the nugget and partial sill increased between the dates. On 14 March, the spatial range decre ased to 31.85 m and the semivariogram modeled for the TSSM counted from strawberry leaves on this date, had a low nugget to sill ratio which is an indication that most of the variation in the data was captured within the range. On 30 March, the spatial ran ge decreased

PAGE 107

107 to 17.85 m and the nugget to sill ratio increased to 53% as compared to 14 March. Spider mite numbers recorded on the strawberry leaves on this date were lower than those recorded on the previous sampling date (Table 6 2). On 7 April, when TSS M had dispersed to occupy most of the field, the spatial range decreased to 13.86 m and at least 60% of the variance in the data was attributed to spatial dependence. Similarly, semivariograms modeled for TSSM population early in the growing season showed a random distribution in the field and distribution became aggregated as the mite population increased. The semivariogram for the TSSM pooled population had a spatial range of 20.25 m and nugget to sill ratio of 35%. Similarly, spherical models were used in semivariograms to describe spatial structure of N. californicus for the last two sampling dates (30 March and 7 April 2011) and for the pooled population the 2010/2011 strawberry growing season. Semivariograms modeled were almost pure nugget with a very high nugget to sill ratios (< 70%) (Table 6 3) and therefore no spatial patterns was detected in this mite species population. Spatial range varied between 23 to 44 m and the semivariograms showed that N. californicus population a random distribution in t he strawberry field. Interpolated TSSM density maps for 14 March, 30 March and 7 April 2011 are shown on for ordinary kriging (OK) [Fig 6 4] and inverse distance weighting (IDW) [Fig 6 5], field and a second one in the north end of the field. Mite numbers were more than 300 per strawberry leaflet in those hot spots. In addition to the two hot spots, IDW maps showed two other areas in the south region of the field where mite population had started to establish. These two areas were not observed on the OK maps. By March 30, spider mite population had also

PAGE 108

108 edge of the f ield (Figs 6 4 & 6 5). Both the IDW and OK maps showed that in the center of the strawberry field, mite numbers had reduced to less than 20 mites per strawberry leaflet on 30 March. Inverse distance weighting interpolated maps highlighted more hot spots t han OK on 7 April (Figs 6 4 & 6 5). A similar trend was observed on the two interpolated maps (IDW and OK) for the pooled TSSM population data for the entire season (Figure 6 6). Table 6 4 shows prediction error statistics associated with IDW and OK inte rpolation methods for TSSM numbers. Of all the sampling dates, the root mean square error (RMSE) for TSSM was highest on 14 March, which was the day with the highest number of mites per strawberry leaflet. Both IDW and OK overestimated TSSM numbers per str awberry leaf on 14 March and underestimated mite numbers on the first three sampling dates when mite population was low. On 7 Jan, 27 Jan, 9 Feb and 30 March 2011, both methods had similar RMSEs within a difference of less one mite per leaf (Table 6 4). Or dinary kriging had the highest mean prediction error when used to predict spider mites on strawberry leaves for the pooled TSSM population data (Table 6 4). On average, both interpolation methods (IDW and OK) had similar means indicating similar accuracy i n predicting mite numbers per leaf. Prediction maps for N. californicus population for the last three sampling dates (14, 30 March and 7 April 2011) are shown in Fig 6 7. Neoseiulus californicus population was observed end of the season (7 April) N. californicus population started to decrease in the center of the field that had the highest numbers on the previous sampling dates. Interpolated surfaces for N. ca lifornicus using IDW highlighted more spots present with the predatory mites than those where OK method was used (Figs 6 8 & 6 9). However, prediction statistics errors for the two interpolation methods were very similar indicating similar accuracy. On 30 March, remnant

PAGE 109

109 mites from prediction and RMSE for IDW was 0.06 and 2.62 while for OK was 0.14 and 2.34, respectively. On 7 April, the error statistics were 0.02 and 0.89 for IDW and 0.04 and 0.92 for OK method. Fig 6 10 shows pooled populations for bot h the two mite species on strawberry leaves. Neoseiulus californicus was found to be in close association with their prey (TSSM) where areas with high spider mites numbers also had high predatory mites numbers depending on when the mites established. Disc ussion The two strawberry growing seasons were somewhat different in that predatory mites were only released in 2010/2011 growing season and not in 2009/2010 due to low TSSM population. In 2010/2011 growing season, TSSM distribution in the field was random early in the season in the field. In predicting TSSM at unsampled locations, two interpolation methods (IDW and OK) were used. Results indicate IDW interpolation method was better than OK in highlighting from the two methods were comparable giving equal accuracy in predicting mite numbers. Neoseiulus californicus was int roduced at localized release points/areas in the field with the hope the predatory mites will be able to disperse in search of TSSM to occupy the entire field. Neoseiulus californicus only established and persisted at the sampling points that had TSSM popu lation at the time of predatory mite release. Neoseiulus californicus population was low through the season and showed a random distribution in the field towards the end of the season when their population started to increase. Where N. californicus was int roduced at a site with TSSM infestations, they eventually reduced their numbers by the end of the season. The N.

PAGE 110

110 californicus to control TSSM on strawberries may n ot be very successful. Maps generated by mid season that showed that TSSM population had decreased in the center of the field, which had high population at the beginning of the season. Twospotted spider mite population build up was very slow during the 200 9/2010 growing season probably due to low temperatures and long freezing days experienced during January and February 2010. For this reason, N californicus were not released in the field during that season. In addition, due to reasons beyond our control s ampling was only conducted until March 23 rd 2010 instead of to the end of the season. The grower chose to use a miticide, fenbutatin oxide (Vendex 50WP, DuPont, ld by mid season and therefore sampling program was terminated. During the 2010/2011 strawberry growing season, TSSM population was low early in the season and gradually increased throughout the season. Semivariograms were therefore computed to study spat ial dependence in the 2010/2011 TSSM population data. Spatial dependence was not detected in the TSSM data collected in the first three sampling dates (7 Jan, 27 Jan, and 9 Feb l. 1991, Liebhold et al. 1993) with very high nugget to sill ratios (> 69%). Lack of spatial dependence on these dates is probably due to high aggregation of TSSM numbers on strawberry leaves in the Variograms computed with data that has very large or very small values will have noise that may mask structure and produce pure nugget effect (Liebhold et al. 1993). In such case where spatial structure is lacking, sample variance may be sufficient to de scribe the data (Liebhold et al. 1993). Nugget represents measurement errors or data variation at very fine scale beyond the minimum lag distance (Isaaks and Srivastava 1989, Liebhold et al. 1993).

PAGE 111

111 Spatial structure was detected when the mite population i ncreased (from 1 March to the end of the season) and dispersed to occupy most of the field. Spatial structures of biological occurrences such as insect populations are affected by time of sampling and therefore proper timing is very important (Schotzko and population was recorded, the semivarigram modeled had the lowest nugget to sill ratio (14%) and relatively high spatial range of dependence (31.8 m). Spatial range (i.e., distance at which data show spa tial dependence), indicates that sampling points shorter than the range (31.8 m) had similar number of TSSM per leaf. On March 30 when TSSM population decreased from the previous sampling date as a result of chemical control, the spatial structure changed with a range of only 17 m and a high nugget to sill ratio of 53%. On 7 April, TSSM population had dispersed to occupy the entire field (Fig s. 6 1B, 6 4, 6 5) and the semivariogram modeled was linear an indication that spatial autocorrelation range could no t be identified. When the TSSM population for the season was pooled, the modeled semivariogram had a spatial range of 20.25 m and a moderate high nugget to sill ratio of 35%. Overall, the range varied between 13 to 37 m when spatial structure was detected over the whole sample period. The spatial autocorrelation range can be used to determine sampling intervals for TSSM in intensive sampling programs for TSSM site specific pest management. Four common types of semivariograms were described by Schotzko and relation to insect distribution. In this context, TSSM population showed a random distribution on the first three sampling dates, as characterized by linear semivariograms with little or no slope and nugget (localized discontinuity) equa l to sill. On 1, 14 and 30 March 2011, TSSM population showed clumped distributions as indicated by semivariograms that had a gradual reduction in

PAGE 112

112 The spatial dis tribution of N. californicus was examined using semivariograms only for 30 March and 7 April 2011. On both dates, N. californicus population had a random distribution and o. Neoseiulus californicus population was very low through out the season and unlike TSSM (prey), the predators did not disperse to occupy the entire field. Inverse distance weighting and OK interpolation methods performed similarly and showed the same are methods had similar numbers giving same level of accuracy. However, it was noted that IDW was better than OK in highlighting small TSSM hotspots forming especially when the mit e population was low. For example, OK map for TSSM predicted for 14 March 2011, (Fig 6 4) does not show two small spots developing on the half south of the field or on Fig s. 6 8 or 6 9 comparing N. californicus prediction that only show areas with high mi te numbers only. In these the maps. On 14 March, both methods had lower RMSE (135 147 mites) in relation to the variance of the interpolation variance (196). It s eems that both methods had lower accuracy in predicting mite numbers counted on this date. This could be attributed to presence of a few very high mite numbers counted on a few strawberry leaves on that date. In addition, the highest number of TSSM for the season was recorded on this date. However, OK seemed to be slightly better than IDW in predicting TSSM on this date. With exception of this date, both methods were comparable in predicting TSSM numbers. Rhodes et al. (2011) found that IDW and OK interpola tions were similar in modeling flower thrips distribution in blueberry fields. In strawberries, TSSM IPM programs require frequent sampling for pest control decision and

PAGE 113

113 therefore a simpler method of interpolation such as IDW would be easier to use. Furthe rmore, its capability in highlighting early developing hot spots on the maps generated would be useful in decision making for early management actions before the population become too high. Prediction maps for TSSM can be used to define management zones fo r TSSM in the field. Treatment threshold for TSSM on strawberry is determined by several factors such as time of infestation, strawberry variety grown, market price and control method to be used. Economic threshold for TSSM is lower early in the season for a susceptible strawberry variety, and also lowered if biological control agents are to be used for management purposes as opposed to chemical control. Treatment threshold level is usually lower than 20 mites per leaf early in the season and higher than 20 late in the season. On the other hand, prediction maps for N. californicus can be used to define areas that should not be treated or where pesticides harmful to predatory mites. Neoseiulus californicus population increased on the point of release that had high prey (TSSM) population in the center of the field. Recovery from other release point was very low and even from the point that was not used as release points as shown on Figure 6 3. This means that N. californicus had very minimal dispersal and was o nly able to reduce TSSM population at the point of release if the prey was present. Some Phytoseiids such as Phytoseiulus permisilis, have been shown to disperse and distribute themselves in relation to their prey (TSSM) [Nachman 1981]. In the 2010/2011 gr owing season, N. californicus were introduced in the field when only 5 out of the 129 sampling points were recorded positive for TSSM population. We found that at one sampling point, TSSM populations reduced from 274 motiles to zero mites per leaf (between 9 Feb and 30 March) and N. californicus increased to 50 motiles per leaf by 14 March. Despite

PAGE 114

114 high numbers of N. californicus counted per leaf (50 and 19 motiles) on 30 March and 7 April, respectively, their spatial distribution in the field was very poor These mites were found aggregated on a few leaves only in a few sampling points. It can also be observed that aggregation of N. californicus they diminished their prey at inoculation site when present. In small plot studies, N. californicus dispersed at the rate of ~ 5 cm per day to cover a distance of 2.2 m within a strawberry row and they hardly moved despite presence of high TSSM numbers at neighboring points (plants) (Nyoike unpublishe d data). In this study, we report that localized release of predatory mites with expectation that they will disperse to control spider mites may not be relied on by commercial strawberry growers. However, site specific releases can be done targeting mite p predatory mite such as P. persimilis will work better than N. californicus in situations where TSSM mites are randomly dispersed in the field. Twospotted spider mites has a high reprodu ctive capacity (up to 100 eggs in her lifetime) coupled with reproduction cycle that is highly dependent on temperature (optimum temperature is 80 F) and be as short as 5 days (Osborne et al. 1999, Fasulo and Denmark 2000, Huffaker et al. 1969). It is clea r that N. californicus with the inoculation at the localized release points could not manage /keep up with the increasing TSSM populations. Conclusions This study indicate that the minimum distance predatory mites can be released with anticipation to spre ad throughout the field should be lower than the 20 meters (within the row) pred atory mites on the entire field. Res ults also shows that a sampling distance h igher than 5 m used in the experiment can be used in intensive sampling for TSSM especially in site specific

PAGE 115

115 pest management based on the minimum range (13 m) obtained. Site specific pest management requires intensive sampling that can be cost prohibitive. However, cost involved (management and sampling) can be justified if site specific approach results in reduced pesticide input and effective control of the pest (Park and Tollefson 2005). Furthermore, the aggregated behavior of TSSM makes it an excellent candidate for site specific pest management (Park et al. 2007). On the other hand, it is obvious that further studies should be conducted at shorter distances than the 20 meters (within the row) used in this study to understand spatial distribution of N. c alifornicus in relation to its prey.

PAGE 116

116 Table 6 1. Total counts of arthropod populations from three strawberry leaflets in the 2009/2010 strawberry growing season. The table shows sampling points that were positive for twospotted spider mites, whiteflies or aphids and sampled between November 2009 and March 2010. Twospotted spider mites Insect pests Sampling points Sampling date Motiles Eggs Whitefly nymphs Aphids Positive sites 5 Nov 5 23 37 0 2 19 Nov 10 9 59 0 3 5 Dec 0 9 33 4 1 17 Dec 0 14 83 8 2 18 Jan 46 67 14 7 5 17 Feb 29 93 0 10 4 23 March 370 1497 0 4 11

PAGE 117

117 Table 6 2. Descriptive statistics of the twospoted spider mites and Neoseiulus californicus counted per strawberry leaflet from 7 Jan to 7 April 2011 Mite Date Mean SEM 3 Std De v Max Skewness Kurtosis TSSM 1 7 Jan 0.33 0.27 3.05 34 10.94 121.49 27 Jan 1.97 1.81 20.52 232 11.27 127.38 9 Feb 3.94 2.51 28.56 274 8.39 72.71 1 Mar 14.42 3.93 44.09 300 4.32 21.06 14 Mar 82.20 17.47 196.90 1092 3.56 13.48 30 Mar 57.05 8.98 10 1.59 581 3.45 13.49 7 Apr 79.94 9.68 109.57 582 2.15 4.73 NC 2 motiles 27 Jan 0.00 0.00 0.00 0 9 Feb 0.00 0.00 0.00 0 1 Mar 0.10 0.09 0.99 11 10.74 117.96 14 Mar 0.41 0.39 4.44 50 11.26 126.79 30 Mar 0.52 0.21 2.41 19 6.56 45.09 7 Apr 0 .30 0.08 0.91 5 3.41 11.55 NC eggs 27 Jan 0.00 0.00 0.00 0 9 Feb 0.00 0.00 0.00 0 1 Mar 0.05 0.04 0.40 4 8.98 84.17 14 Mar 0.27 0.27 3.02 34 11.27 127.00 30 Mar 0.28 0.16 1.86 20 9.74 102.23 7 Apr 0.34 0.12 1.41 12 6.16 43.62 1 TSSM = tw ospotted spider mites motiles 2 NC = Neoseiulus californicus 3 SEM = standard error of mean

PAGE 118

118 Table 6 3. Summary of semivariogram parameters for ordinary kriging (OK) interpolation method used for twospotted spider mites (TSSM) and Neoseiulus californicus (N C) sampled between January and April on strawberries in 2010/2011 growing season Sampling Date Range Partial sill Nugget Nugget: Sill Lag size # of lags TSSM 1 7 Jan 56 0.05 0.13 72 1 56 27 Jan 56 0.15 0.33 69 1 56 9 Feb 54.5 0.19 0.59 78 1 56 1 Ma rch 37.24 1.76 1.18 40 2 28 14 March 31.85 4.64 0.77 14 3 18 30 March 17.83 1.28 1.47 53 2 28 7 April 13.86 1.69 1.15 40 2 28 All dates 20.25 1.31 0.71 35 2 28 NC 2 30 March 44 0.05 0.24 83 1 56 7 April 23.11 0.05 0.14 74 1 56 All date s 42.72 0.18 0.4 98 1 28 1 twospotted spider mites 2 Neoseiulus californicus

PAGE 119

119 Table 6 4. A summary of mean and root mean square errors (RMSE) comparing prediction statistics for inverse distance weighting (IDW) and ordinary kriging (OK) accuracy in estimati ng twospotted spider mites (TSSM) on strawberry leaves Date Statistic (error) Mean (mites) Root Mean Square Error (Mites) 7 Jan IDW 0.03 3.09 OK 0.2 2.98 27 Jan IDW 0.16 20.9 OK 1.55 20.34 7 Feb IDW 0.34 28.04 OK 2.93 28.23 1 Mar IDW 2.4 45.78 OK 6.22 40.42 14 Mar IDW 9.4 147.6 OK 11.43 135.11 30 Mar IDW 6.15 95.30 OK 6.43 95.15 7 Apr IDW 2.31 99.04 OK 14.4 113 All dates IDW 3.03 229.80 OK 48.54 235.59

PAGE 120

120 A

PAGE 121

121 Figure 6 1. Spatial distribution for twospotted spider mites on field grown strawberries collected on three sampling dates A) in January and February 2011, and B) in March and April 2011. B

PAGE 122

122 Figure 6 2. P opulation densities of the mite populations on the strawberry leaves d uring the 201 0/2011 strawberry growing season, A) for the t wospotted spider mites (TSSM) and, B) for the predatory mites ( Neoseiulus californicus (NC) 0 20 40 60 80 100 120 1/7 1/21 2/4 2/18 3/4 3/18 4/1 Mean # of TSSM per leaf Sampling date 0 0.1 0.2 0.3 0.4 0.5 0.6 1/27 2/10 2/24 3/10 3/24 4/7 Mean # of NC motiles / eggs per leaf Sampling date Motiles Eggs B A

PAGE 123

123 Figure 6 3. Release points for Neoseiulus californicus population and its establishment on a strawberry field between 19 January an d 30 March 2011. Neoseiulus californicus release points indicate the points between where predatory mites were introduced.

PAGE 124

124 Figure 6 4. Ordinary kriging interpolation of twospotted spider mites on strawberries sampled on 14 March, 30 March, and 7 April i n 2010/ 2011 strawberry growing season.

PAGE 125

125 Figure 6 5. Inverse distance weighting interpolation of twospotted spider mites on strawberries sampled on 14 March, 30 March and 7 April in 2010/ 2011 strawberry growing season.

PAGE 126

126 Figure 6 6. Interpolated surface s for the pooled population of twospotted spider mite recorded on strawberry plants in the field for the entire 2010/2011 growing season using two ordinary kriging and inverse distance weighting interpolation methods.

PAGE 127

127 Figure 6 7. Inverse distance weighti ng interpolation of Neoseiulus californicus on strawberries from 14 30 March and 7 April in 2010/2011 strawberry growing season.

PAGE 128

128 Figure 6 8. Predicted Neoseilulus californicus per strawberry leaf on 7 April using ordinary kriging and inverse distance we ighting

PAGE 129

129 Figure 6 9. Interpolated surfaces for the pooled population of Neoseiulus californicus recorded per strawberry leaflet for the entire 2010/2011 growing season using ordinary kriging and inverse distance weighting interpolation methods.

PAGE 130

130 Figu re 6 10. Spatial distribution of twospotted spider mites and Neoseiulus californicus populations per strawberry leaflet on field grown strawberries in 2010/2011 growing season. Total counts represent pooled populations for the entire season from January to April 2011.

PAGE 131

131 CHAPTER 7 CAN TWO YEAR OLD SYNTHETIC M ULCH AFFECT MARKETAB LE YIELD, ARTHROPOD POPULATION S, WEEDS AND DISEASE S ON FIELD GROWN STRAWBERRIES? Introduction Florida is a very important producer of strawberries in the US. The state is ranked second after California and produces 100% of the domestically grown winter crop (Mossler and Nesheim 2007). During the 2010 2011 growing season, 15% of the total strawberry crop in the US was harvested in Florida from 4006.4 ha valued at $366.3 million USD. Overa ll acreage under strawberries in Florida has increased from 7300 to 9900 between 2005 and 2011 (USDA NASS 2011). In order to sustain profitability of this high value crop, the state must maintain high quality production during the critical winter months of the year when Florida is the only local supplier/producer of fresh market strawberries (Mossler and Nesheim 2007). Most of production in Florida is carried out in open fields. Unlike temperate regions such as Europe where strawberry planting and harvestin g can be spread over two growing seasons (winter and summer) (Freeman and Gnayem 2005), in Florida this takes place in one growing season. Usually, the crop is planted between late September and early October, and harvested from December through March or A pril when the season ends. The standard growing procedures for strawberries in Florida is to use black plastic mulch on raised beds that are removed at the end of the growing season (Chandler et al. 1993). Plastic mulches are used on raised bed to conserve water, prevent weed germination and reduce herbicide, and keep fruit clean ( Kasperbauer 2000, Freeman and Gnayem 2005, Waterer et al. 2007). Strawberry is one of the most expensive crops to produce, with an average variable cost per ha estimated at $24,47 8 ($9791/acre) [Santos et al. 2011]. These costs include land

PAGE 132

132 preparation, labor, strawberry transplants, fertilizer, weeding, insect and mite management, and disease control. To remain profitable, strawberry production practices must be easy to accomplish sustainable, and economically feasible. Double cropping (using the plastic mulch for a second crop) is one strategy that could be used to reducing costs of production (Waterer et al. 2007). Although a few growers have double cropped strawberries with oth er vegetable crops such as cucumber, watermelon, and squash, growing a second strawberry crop on the same plastic is a relatively new practice in Florida. In 2009 approximately a third of Florida's strawberry land was used for double cropping (Noling and W hiden 2009, Noling 2011). Double cropping with a second strawberry crop demands leaving the plastic mulch intact through the summer months for use in the fall season. Using the plastic mulch for a second crop is beneficial in that the cost of purchasing ne w plastic mulch, drip tubing, laying and disposing the mulch is spread across two growing seasons. In addition, the plastic is not hauled to a landfill or burned at the end of the season, reducing contamination to the environment. Generally in double cropp ing the strawberry plants are desiccated at the end of the season using glyphosate (Roundup) and/ or a soil fumigant without pulling out the crowns with the roots. The beds are kept weed free throughout the summer months. At transplant, manual labor is us ed to pull out old strawberry plants (dead crowns and roots) from the previous season before making new holes for the second crop. Using the plastic mulch for a second strawberry crop with the old strawberry plants (thatch) would be advantageous because th is would save the growers the additional labor costs (~ $70/acre) to remove the old plants. Enhanced methods for strawberry production that will reduce labor costs and increase production and profits with adequate resource management are highly encouraged for strawberry production in Florida (Florida Strawberry Research and Education Foundation [FSREF] 2011 & 2012).

PAGE 133

133 Studies were therefore conducted to determine if leaving the old plants in the field will affect arthropod populations on the second crop, act as a reservoir for weed seeds, and/ or as a source of disease inocula that can ultimately affect marketable yields. Could the thatch (in old holes) be sites for diapausing insects and mites or a source for disease inoculum in the field? For example, twospo tted spider mites, the main mite damaging strawberries, have a diapausing stage that is able to survive through hot summer months. In addition to arthropod populations, the study also investigated how the presence of old roots in the soil affected the grow th and development of new transplants. Materials and Methods Two field studies were conducted during 2010/2011 and 2011/2012 strawberry growing seasons in a commercial strawberry field located in Citrus County, Florida using 2 yr old black plastic mulch ( I ntergro, Clearwater, FL ). Planting beds were 28 inches (0.71 m) wide on top and 10 inches (0.25 m) high were prepared with a standard bedder. Each bed was fitted with one drip line for irrigation before covering them with black plastic mulch (1.25 mil). Th e beds were fumigated with Telone Inline ( dichloropropene and chloropicrin) (Dow AgroSciences, Indianapolis, IN) at the rate of 35 gpa before transplanting strawberries. Two strawberry varieties 2010 and 12 Oct. 2011, respectively. Each bed was planted with a double row of strawberries. Growing procedures and management with respect to weed, disease management, and harvesting yields were carried out according to the standard strawberry productio n practices in Florida (Peres et al. 2010). Two yr Mulch Preparation To prepare the plastic mulch for a second crop, strawberry plants were terminated/desiccated using a fumigant, Telone EC (12 gpa) through the drip line. To maintain the beds with the dr ip tubing through the summer months, drip tapes were flushed with sulfuric

PAGE 134

134 acid once a month, irrigated once a week and kept weed free using herbicides whenever needed to control weeds. Prior to planting the second strawberry crop, torn plastic mulch was r epaired and beds fumigated with Telone Inline. New planting holes were made on the mulch surfaces at random, meaning that some holes were exactly where the previous plant was or on a new location. Experimental Design and Plots The experimental design was a randomized complete block with two treatments and four replications during the two growing seasons. Treatments included strawberry plants growing with and without old plants from the previous season. Experimental area in 2010/2011 consisted of 8 strawbe rry beds (rows) each 125 m long. Four of the rows had the old strawberry plants from the previous season pulled out before transplanting the new crop, and the other four rows had the old plants remaining (not pulled out). The same two treatments were evalu ated on a different block with 2 yr mulch old plastic mulch in 2012/2012 growing season. The experimental area consisted of 16 strawberry beds that were 100 m long. Experimental plots were 9.1 m (30 ft) long with four rows of strawberry. Arthropod Populati ons Sampling Leica MZ 12 5, Leica Microsystems, Houston, TX) at 10X magnification

PAGE 135

135 Weed Sampling Plant Size and Marketable Yields During each g rowing season, 10 strawberry plant size dimensions, including height and width, were taken at early, mid, and late growing seasons. In 2010/2011 growing season, harvesting was conducted from 22 Nov 2010 to 26 March 2011 and in 2011/2012 from 21 Nov to 8 Ma rch 2012. A total of 50 plants from each plot was harvested twice per week (or as required), graded according to marketing standards and weighed separately as per treatment (Peres et al. 2010). Disease Incidence Soon after strawberry plant establishment, the number of missing plants per plot was recorded. A prognosis was conducted to differentiate plant death due to disease related causes or as a function of the planter due to human error (since strawberry transplants were planted by hand). Plant death due to fungal diseases was monitored from transplanting until the end of the season. The number of missing plants was taken at different stages of crop development, from fruit initiation to harvest. Plants showing fungal symptoms were collected and samples se nt to the Division of Plant Industry (DPI), Florida Department of Agriculture and Consumer Services, Gainesville, for the disease causing pathogen identification and /or confirmation.

PAGE 136

136 Data Analysis A series of t tests were conducted using PROC TTEST progra m in SAS (SAS Institute 2002) was used to test for significant ( P < 0.05) differences between the treatments in arthropod populations, plant size, and marketable yields. Results and Discussion Arthropod Populations Twospotted spider mites were not recorded on strawberry plants growing with the old plants or in plots where the old plants were removed until 26 January 2011. Although the mite population increased over time in the 2010/2011 growing season (Fig. 7 1), there were no significant ( differences between the two treatments (Table 7 1). Neoseiulus californicus a predatory mite for TSSM; was recorded on both treatments in very low numbers per leaf. Two other predatory (beneficial) insects for TSSM were recorded within the treatments including six Melon aphids ( Aphis gossypii Glover) were the only other insect pest recorded during the season but like all the other arthropods recorded on the straw berry leaves, there were no significant differences between the two treatments (Table 7 1). Melon aphids are commonly encountered on strawberry leaves but they are not considered a major pest of field grown strawberries (Fraulo et al. 2008; Nyoike and Libu rd unpublished data).

PAGE 137

137 Acramite, Chemtura, Middlebury, CT Weeds On average, significantly more weeds per quadrant (2.95 0.4 versus 1.5 0.2) were recorded on plots where the old plants were removed as compared with plots where they were left in the field before transplanting the new plants ( On plots where the old plants were removed, weeds were observed to germinate or grow from the planting hole where the previou s plants grew and/ or at the base of the strawberry plant. Generally, more weeds were observed in double cropping (using a two yr old mulch) compared with a new plastic mulch (Waterer et al. 2007). The increase of weeds has been associated with the loss of [98:2] methyl bromide: chloropicrin, a formulation of methyl bromide that was very effective on weeds. Although a 50: 50 methyl bromide formulation is available, its use has been limited due to the increase in price and reduced concentration affecting its effectiveness (Noling and Whiden

PAGE 138

138 D and chloropicrin marketed as Telone Inline is an effective alternative to methyl bromide but chloropicrin alone is effective against diseases but not nemat odes or weeds (Noling 2011). The three dominant weed species recorded were Carolina geranium ( Geranium carolinianum L.) black medic ( Medicago spp.) and wild carrot ( Daucus spp.) [Figure 7 3]. Carolina geranium and black medic are some of the most importan t weeds in strawberry production (Rosskopf et al. 2005, US EPA 2005). These weeds were included in the list of weeds to justify use of methyl bromide as a critical pest management tool in strawberry production (US EPA 2005). Other minor weed species record ed were sow thistle ( Sonchus spp.), Mexican tea ( Chenopodium ambrosoides L.) and wondering cudweed ( Gnaphalium pensylvanicum Willde.) (Fig.7 3). In both growing seasons, hand weeding was the only option of weed control during the growing season but it may not be very cost effective at ~ $70 per acre. Plant Size Growth and Marketable Yields Leaving or removing the old plants from the previous season did not affect strawberry plant size (width and height) on the two yr old black plastic mulch. During both sea sons no significant differences were recorded in terms of plant height or width between the two treatments evaluated. Marketable yield harvested in strawberry plots where the old plants were left in the field was lower than that recorded where the old plan ts were removed before transplant (Figure 7 4). However, the differences between the treatments were only significant in 2011/2012 growing season (t = 2.63; P = 0.0390) but not in 2010/2011 (t = 0.37; P = 0.7237). Over time, both treatments showed a simi lar trend in terms of monthly averages and harvesting strawberries (Figs. 7 5 & 7 6). As indicated by the high standard errors on the strawberry plots with old plants, at least one of the plots varied from the other plots in terms of yield recorded. For

PAGE 139

139 in stance, during the 2010/2011 growing season, one plot yielded 7.3 kg (16 lbs) in March as compared with 22.7 kg (50 lbs) from other plots of the same treatment. This same plot with low yield had the highest number of plant mortality due to fungal diseases. Disease Incidence On average, significantly more plants were missing in plots where old plants from the previous season were left to grow with the new transplants (t = 2.23; P = 0.0388) during 2010/2011 growing season. There were 3.2 1.2 missing plant s on plots with old plants as compared with 0.4 0.2 where the old plants were removed. Similarly, in 2011/2012 significantly more strawberry plants were lost due to fungal mortality in plots where old plants were left in the field than where they were re moved (t = 4.15; P < 0.0001). We observed that plant mortality increased overtime throughout the season especially on the plots where old plants were left in the field (Fig. 7 7). During both growing seasons, diseased plants were submitted for identificati on at the Department of Plant Industry (DPI), Gainesville, Florida. The main fungal pathogens identified were Colletotrichum acatatum, Rhizoctonia solani, and Verticillium spp., causative agents for crown rot, root and crown rot, and Verticillium wilt, res pectively. Soil borne disease was problematic on the second crop and especially where the old plants were left in the field. In double cropping, planting beds were left intact throughout the summer months as would have been done in solarization. It is well documented that soil solarization is one method of disinfesting soil from fungal inoculum and weeds and nematodes due to the elevated temperatures in the beds (Katan 1981, McSorley and Gill 2010). Noling and Whiden (2009) reported that temperatures under the mulch varied between 85 and 115 o F (29.4 o C and 46.1 o C) under the black plastic mulch during summer months in Florida. Such high temperatures in absence of food were sufficient to kill sting nematodes ( Belonolaimus longicaudatus Rau); the

PAGE 140

140 most damaging nematode on strawberries in Florida (Noling 1999). However, these temperatures are generally insufficient for control of many nematodes, weeds, and plant pathogens and greater temperatures would typically be expected from solarization under clear plastic mulch (McSorley and Gill 2010). Furthermore, Colletotricum acutatum has been found to survive on dry strawberry berries through the summer hot months in Florida (Mertely and Peres 2010). Elsewhere, solarization was effective in Israel in killing Colletotri chum acutatum infected mummified berries buried at 20 cm below ground soil (Freeman and Gnayem 2005); reduced soil inocula of Verticillium microsclerotia and Phytophthora cactorum in California (Hartz et al. 1993, Himelrick et al. 1995). Conclusion s Leavin g the old plants on the plastic mulch in double cropping with strawberries did not increase arthropod populations or affect plant growth of the second strawberry crop. However, it appears that these old plants could be a source of fungal inoculum and hence an increase in plant mortality (due to fungal diseases). Alternatively, pulling out the old plant created more holes on the plastic mulch that permitted more weeds to germinate from the soil. It would be advisable to utilize the same planting holes and av oid making new holes on the plastic mulch. However, at this point we do not know if this will have adverse effects on the new transplants. The savings on double cropping with a second strawberry crop include savings on poly cover at the end of the season, costs associated with summer cover crop (millet) such as planting the cover, fertilizing, irrigation, turning in the cover crop after the season, and land preparation procedures at the time of planting in the fall. However, there are costs associated with double cropping such as increased hand weeding, repairing the plastic mulch, flushing the drip lines once a month, injecting the line cleaner with sulfuric acid once a month, and watering the beds once a week. The beds also must be maintained weed free thr oughout the summer months,

PAGE 141

141 usually requiring herbicide such as glyphosate. The plastic mulch (1.25 mil) used for double cropping is more expensive than the standard/regular (1mil) commonly used. Overall, it is estimated that double cropping with a second s trawberry crop would save growers (~ $1500/ acre [$3750/ha]. The strawberry crop is one of the most expensive crop to produce with an average variable cost per ha estimated at $24,477.5 ($9791.02/acre) [Santos et al., 2011]. Some growers may agree that dou ble cropping with a second strawberry crop may be worthwhile pursuing even if they have to spend an extra $70 to pull the old plants before transplanting.

PAGE 142

142 Table 7 1. Mean arthropod populations per strawberry trifoliate leaf from two treatments (strawberr y plants growing with and without old plants from the previous season) on a 2 in Citrus County, Florida in 2010/2011 strawberry growing season. Mean arthropod populations per str awberry leaf (Mean SEM) With dead plants Without dead plants t value P value TSSM motiles 18.4 4.2 24.2 6.6 0.74 0.4608 N. californicus 0.07 0.04 0.01 0.01 1.30 0.1945 Aphid 0.4 0.1 0.4 0.3 0.04 0.9655 Six spotted thrips 0.02 0 .01 0.0 0.0 1.43 0.1549 Big eyed bug eggs 0.03 0.02 0.02 0.01 0.39 0.6984 *TSSM = Twospotted spider mites

PAGE 143

143 Figure 7 1. Tetranychus urticae population dynamics on strawberry plants growing on 2 yr old plastic mulch with and without old plants from the previous season in 2010/2011 strawberry growing season 0 20 40 60 80 100 120 140 12/30 1/20 2/10 3/3 3/24 4/14 Mean TSSM motiles per leaf 2010/2011 growing season With dead plants Without dead plants

PAGE 144

144 Figure 7 2. Total arthropod populations counted on strawberry leaves growing on 2 yr old plastic mulch with and without old plants from the previous season in 2011/2012 strawberry grow ing season. 0 5 10 15 20 25 30 TSSM Motiles Whitefly nymphs Lacewing eggs Thrips BEB* Total #. of arthropod counted on strawberry leaves With dead plants Without dead plants

PAGE 145

145 Figure 7 3. Weed species composition recorded in strawberry plots growing with and without old plants from the previous season on 2 yr black plastic mulch 51% 17% 15% 7% 6% 2% 2% Weed Species Composition Calorina geranium Black medic Wild carrot Wandering cudweed Sow thistle Field pennycress Mexican tea

PAGE 146

146 Figure 7 4. Average marketable yield harvested for two growing seasons from p lots with strawberry plants growing with and without old plants from the previous season on 2 yr old mulch 0 10 20 30 40 50 60 2010/2011 2011/2012 Marketable yield (Kg) With dead plants Without dead plants

PAGE 147

147 Figure 7 5. Average strawberry marketable yield over time in 2010/2011 growing season. The two treatments evaluated are strawberry plants grow ing with and without old strawberry plants from the previous season Figure 7 6. Average strawberry marketable yield over time in 2011/2012 growing season. The two treatments evaluated are strawberry plants growing with and without old strawberry plan ts from the previous season on a 2 yrs black plastic mulch 0 5 10 15 20 25 30 35 Nov Dec Jan Feb March Avg. Strawberry Marketable yield per month (Kg) 2010/2011 growing season With dead plants Without dead plants 0 5 10 15 20 25 Nov Dec Jan Feb March Avg. Strawberry Marketable yield/month (Kg) 2011/2012 growing season With dead plants Without dead plants

PAGE 148

148 Figure 7 7. Mean number of missing strawberry plants in plots growing with and without old plants from the previous season in 2011/2012 0 2 4 6 8 10 12 14 26-Oct-11 26-Nov-11 26-Dec-11 26-Jan-12 26-Feb-12 26-Mar-12 Mean number of missing plants Sampling Date With dead plants Without dead plants

PAGE 149

149 CHAPTER 8 CONFIRMATION OF MELOIDOGYNE HAPLA ON STRAW BERRY IN FLORIDA USI NG MOLECULAR AND MORPHO LOGICAL TECHNIQUES Introduction Root knot nematodes, Meloidogyne spp., are known to infect many agricultural crops including strawberry ( Fragaria x ananassa Duch.). Meloidogyne incognita, M. javanica, M. arenaria, and M. hapla are among the most economically important species of root knot nematodes ( Qui et al. 2006 ) In particular, M. hapla the northern root knot nematode, is a serious pest of strawberries in the northeastern United States (LaMondia 2002) and a minor pest in California (Westerdahl 2009). Other Meloidoigyne spp. that are potential pests of strawberries include M. javanica and M. incognita in California (Westerdahl 2009). However, M. incognita, despite being found on all soils in North Carolina, does not cause damage on strawberries (Averre et al. 2011). In Florida, the most economically important nematode damaging strawberries is the sting nematode, Belonolaimus longicaudatus as reported since the 1950s (Noling 1999) Typical above ground sympt oms for sting nematodes include stunted plants, dormant plants with no new growth, and leaves dying off starting from the older leaves and progressing to the younger leaves (Noling 1999). One of the ways in which nematodes can be introduced into a clean field is by using infected planting material. This is a very important source of nematodes especially when dealing with crops that are vegetatively propagated such as strawberries and bananas. The common practice while planting strawberries is to fumigate the soil with fumigants such as methyl bromide [50:50; methyl bromide: chloropicrin] or Telone products (Noling and Whiden 2010) for protection against soil borne diseases as well as nematodes and weeds. This practice is expected to provide season long c ontrol for plants by maintaining nematode populations below damaging levels. However, if endoparasitic nematodes enter the production system via infested

PAGE 150

150 transplants, they may not be killed by the fumigant and therefore are a potential risk to the growers. Moreover, endoparasitic nematodes are within the roots and are hard to target with any fumigant without killing the plant. In spring 2011, localized stunting was observed on beds planted with strawberries in Citra, Marion County, Florida. The strawberry p lant roots at those locations were also observed to be galled. Around the same time, nematode symptoms associated with root knot nematode were reported on strawberries in south Florida, and presumed to be caused by M. hapla (Noling and Whiden 2010). Genera lly, B. longicaudatus is considered the most damaging nematode to strawberries in the state. Typically, M. hapla is found in cooler climates, while M. incognita M. javanica and M. arenaria are the predominant root knot nematodes in regions between 30 o N an d 35 o S latitude (Taylor and Sasser 1978). Since several species of root knot nematodes occur in Florida, accurate identification of the root knot nematode on strawberries is critical. To implement an effective management program including crop rotation, c hoice of cultivar, monitoring pest status and its spread in the field; correct identification of the nematode species is important (Adam et al. 2007, Orui 1998). There are several methods available to identify nematodes including polymerase chain reaction (PCR), PCR RFLP (Harris et al. 1990; Powers and Harris 1993), isozyme electrophoresis (Esbenshade and Triantaphyllou 1990), RAPD PCR (Orui 1998), specific sequence characterized amplified region (SCAR) PCR (Fourie et al. 2001), and use of perineal patter ns (Hartman and Sasser 1985). The use of perineal patterns to distinguish between Meloidogyne spp. needs to be supplemented with other methods of identification due to high interspecies similarities and high intraspecies variation (Hu 2011). Reliable molec ular techniques for nematode diagnostics are well established and can be used to complement morphological

PAGE 151

151 studies. DNA can be obtained from a single juvenile or female nematode and amplified through PCR (Harris et al. 1990). The amplified DNA is then diges ted using restriction enzymes to obtain different fragment sizes using restriction fragment length polymorphism (Harris et al., 1990; Powers et al. 2005). The objective of this study was to conduct both m orphological studies and PCR RFLP to identify the Me loidogyne spp. associated with the damage on strawberries grown under Florida conditions. Materials and Methods A field study was conducted at the University of Florida, Plant Science Research and Education Center located in Citra, Marion County, Florida, during the 2010/2011 strawberry growing season (October April). As a standard planting procedure for strawberries, a soil fumigant methyl bromide: chloropicrin (50:50) was injected into tarped planting beds at the rate of 36.5 liters per hectare before tra nsplanting strawberries. Two weeks after soil fumigation, strawberry variety Festival imported from Ontario, Canada as bare root transplants, was planted on 20 October 2010. Strawberries were planted into three blocks, each 21 by 6.3 m with 6 double rows o f strawberries at spacing of 0.35 m by 0.35 m. Growing practices including fertilization, weeding, harvesting, and removal of runners were done according to the standard production practices for north central Florida (Peres et al. 2010). Root knot Sampling and Soil Extraction At the end of the harvest season, on 25 April 2011, six soil samples were collected from each block using a cone shaped auger (Cole Parmer, Vernon Hills, IL) and combined into a composite sample per block. Soil samples were stored at 1 0C for four days, and nematodes were extracted from a subsample (100 cc) using a centrifugal flotation method (Jenkins, 1964).

PAGE 152

152 Root knot Nematode Females At the end of the harvest season, 24 strawberry plants were gently dug out of the soil leaving the r oots system as intact as possible. The plants were cleaned by removing any soil adhering to the roots using tap water, before wrapping them with moist paper towel. Root knot nematode females were hand picked from the strawberry roots at X20 under a dissect ing microscope (Leica MZ 12 5, Leica Microsystems, Houston, TX) and processed for morphological examination or placed in 1% saline water in a sterile micro tube and frozen at 6 o C until needed. Root knot nematode females were collected at the end of 2010/ 2011 strawberry growing season in April; and during the 2011/2012 season, females were recovered from roots upon receiving transplants from strawberry nurseries in October 2011. Morphological Characterization Root knot nematode females were prepared and p erineal patterns studies conducted as described by Hartman and Sasser (1985). Molecular Characterization DNA was extracted from single females per vial using the DNeasy blood and tissue extraction kit (Qiagen, Santa Clarita, CA) according to the manufactu slight modifications. Briefly, individual female nematodes were hand picked in 90 l ATL buffer and incubated overnight at 56 o C after adding 20 l of proteinase K. The females were then vortexed for 15 sec before adding 100 l Buff er AL and 100 l ethanol (96%). The mixture was placed in a DNeasy mini spin column and centrifuged at 8000 rpm for 1 min and flow through sample discarded. With a new receiving tube, the centrifugation procedure was repeated adding 200 l Buffer AW1 and 2 00 l Buffer AW2 to dry the DNeasy membrane at 14000 rpm for 3 min, discarding the flow through each time. The DNeasy Mini spin column was placed

PAGE 153

153 onto a 1.5 ml microcentrifuge tube and 50 l Buffer AE was directly pipetted on to the DNeasy membrane to coll ect the DNA. Polymerase chain reaction Amplification of the mitochondrial DNA (mtDNA) between cytochrome oxidase subunit II ( COII) TACCTTTGACCAATCACGCT GGTCAATGTTCAGAAATTTGT GG (Powers and Harris, 1993). PCR was carried out according to the manufacturers recommendation (DNeasy Blood & Tissue Handbook, Qiagen, Santa Clarita, CA); 25 l volumes containing 12.5 l of PCR multiplex (Qiagen, Santa Clarita, CA ), 1.5 l of 10 each forward and reverse primers, 4.5 l deionized water, and 5 l template DNA. All PCR reactions were run in a thermal Gradient cycler (PTC 200 Peltier Thermal Cycler, Watertown, Massachusetts). The PCR conditions were: denaturation at 95 C for 15 min; 40 cycles of 94 C for 30 sec, 62 C for 90 sec, and 72 C for 90 sec; and a final extension at 72 C for 10 min (Joyce et al.1994). A portion (8 l) of the amplification product was electrophoresed on a 1.8% agarose gel and stained with ethidium bromide. The sizes of amplified products were determined by comparison with a 1 Kb molecular weight ladder (Invitrogen, Carlsbad, CA). Restriction fragment length polymorphism Restriction digestion of the PCR products was conducted using Dra1 enzyme (Promega, Madi son WI). A total volume of 20 l including, 10l of PCR product, 2 l of Dra1 enzyme, 2 l of buffer (Promega, Madison, WI) and 6 l water was used for the restriction reaction and incubated at 37 o C for four hr. The digestion products were separated using 1.8% agarose gel in electrophoresis.

PAGE 154

154 DNA sequencing Additional PCR reactions were carried out to amplify the partial ITS spacer 1, complete sequences of 5.8S and ITS2, and partial 28S RNA gene region. Forward primer TW81 (GTTTCCGTAGGTGAACCTGC) and reverse primer AB28 (ATATGCTTAAGTTCAGCGGGT) were used as described by Joyce et al. (1994). A 10 l portion of the DNA suspension from the previous extraction was added to the PCR reaction mixture containing: 12.5 l PCR supermix (Qiagen, Santa Clarita, CA ), 1.5 l of 10 each forward and reverse primers, 4.5 l deionized water, and 5 l template DNA. All PCR reactions were run in a thermal Gradient cycler (PTC 200 Peltier Thermal Cycler). The PCR conditions were: denaturation at 95 C for 15 min; 35 cycles of 94 C for 1 min, 58 C for 1.5min, and 72 C for 1.5 min; and a final extension at 72 C for 10 min (Joyce et al., 1994). A portion (8 l) of the amplification product was electrophoresed on a 1.8% agarose gel and stained with ethidium bromide. The sizes of amplified products were determined by comparison with a 1 Kb molecular weight ladder (Invitrogen, Carlsbad, CA). For direct sequencing, PCR products were purified with the QIAquick PCR purification kit (QIAGEN, Santa Clarita, CA) and sequenced at the Inter disciplinary Center for Biotechnology Research (ICBR) at the University of Florida. The primers used for sequencing were the same as those used for amplification. Sequences were edited with BioEdit (BioEdit version 7.0.9, Carlsbad, CA). The sequences were compared with those in Genbank ( http://blast.ncbi.nlm.nih.gov/Blast ) by means of a Basic Local Alignment Search Tool (BLAST) search, and sequences were aligned with ClustalW2 (CLUSTAL 2.0, Dublin, Ireland ).

PAGE 155

155 Results Root knot Nematode Symptoms and Soil Nematode Abundance Above ground symptoms associated with nematode infestation were observed on the strawberry plants growing in the field in 2010/2011 growing season. The plants were stunted, with reduced cro wns, and produced small berries (Fig. 8 1A). Below ground symptoms included root galling and excessive growth of fibrous roots compared to healthy roots (Figure 8 1A). In the 2011/2012 growing season, transplants showing root knot galling (Figure 8 1B) wer e recovered from bare root green top strawberry transplants from the nursery. The dominant plant parasitic nematodes recovered from the soil samples were root knot nematodes ( Meloidogyne spp.) and stubby root nematodes ( Trichodorus spp.) [Fig. 8 2]. The hi ghest mean number of root knot nematode juveniles (232 J2/100 cc soil) was recorded in block C, followed by 21 juveniles in block B and only one J2 in block A. Trichodorous spp. numbers ranged from 2 to 5 nematodes per 100 cc soil across the blocks. Morpho logical Characterization Morphological examination of females extracted from strawberry roots revealed a round perineal pattern with fine, undulating longitudinal lines resembling wrinkles. An invaginated line around the vulva as described by Yik and Birch field (1978) was also observed (Fig. 8 3). These are typical characteristics of Meloidogyne hapla that distinguish it from the other root knot nematode species. DNA Extraction and PCR Amplification DNA was extracted from a single female in the 2010/2011 st rawberry growing season and two females (each extracted separately) in the 2011/2012 strawberry transplants shipment. The PCR amplified the mtDNA region, yielding a single PCR product with the length of ca 540 bp for all three samples tested (Fig. 8 4). Di gestion of the PCR product using DraI restriction

PAGE 156

156 enzyme generated RFLP fragments at 200 and 250 bp for all the samples (Fig. 8 5), confirming the nematode identity as M. hapla PCR with primers TW81 and AB28 yielded a single fragment of ca 990 for the tes ted isolates. A BLAST search of Genbank revealed that our M. hapla sequences were identical to those of M. hapla from Netherlands (NCBI accession # AY281854), California (NCBI accession # AY268108), Australia (NCBI accession # AF516722), and others, with m aximum identity up to 98%. Discussion During the 2010/2011growing season, several strawberry plants were observed to be stunted and with low yield despite similar management practices across the blocks. The unevenly distributed nematode infestation across the blocks suggests the possibility that nematode infested transplants were only in a few of the boxes received from nurseries. The block with the highest number of juveniles in the soil recorded significantly lower yield than the block with the lowest nu mber of nematodes (Nyoike, unpublished data). PCR RFLP analysis identified the nematode as M. hapla PCR amplification resulted with a band formation at 540 bp and two bands after endonuclease digestion at 200 and 250 bp. Orui (1998) reported similar resu lts using Dra1 for discrimination among10 different species of Meloidogyne When suspecting more than one Meloidogyne species to be present or due to intraspecific variations within one species, other endonucleases should be used to confirm the results of the species identity (Orui 1998). Our results were further confirmed with DNA sequencing that yielded up to 98% identity when our samples were compared with known gene sequences in the Genbank. Both samples collected at the end of 2010/2011 growing season and at the beginning of 2011/2012 (soon after importing the transplants) tested positive for M. hapla This study

PAGE 157

157 confirms that M. hapla was imported with the transplants from Ontario, Canada. There have been previous reports of M. hapla imported into Flo rida from the northern areas but no confirmation studies of the species had been performed (Howard et al 1985; Noling and Whiden 2010). At the end of the 2010 strawberry growing season, nematode symptoms associated with root knot nematode ( M. hapla) infest ation were reported on strawberries but no species confirmation was done (Noling and Whiden, 2010). Strawberry plants were observed to be stunted, low yielding, with shortened growth life and galled roots. Similarly, in North Carolina, M. hapla has also be en introduced to the state through transplants (Averre et al. 2011). The current study confirms the presence of M hapla on transplants in Florida by both molecular and morphological methods. Morphological and molecular studies have also been used to chara cterize a population of M. hapla found damaging on coffee in Hawaii (Handoo et al. 2005). Their study demonstrates that various morphological characters of second stage juveniles, males, and females can be combined with molecular studies to compare differe nt M. hapla populations. The occurrence of M hapla on transplants is an indication that stricter import sanitation rules should be put in place, but worth it is mentioning that M. hapla is not considered a orida, growers are expected to buy strawberry transplants only from certified nurseries. The strawberry transplants used in this study were obtained from a certified nursery in Canada. Meloidogyne hapla is more prevalent in colder latitudes and in high ele vations of the tropics (Powers and Harris 1993). This makes the survival of M. hapla through hot summer months in Florida on cultivated and non cultivated host crops questionable. However, during this past summer (2011), M. hapla was able to survive in the field on peanut for a period of 4 months (D. W. Dickson, personal comm.). At this point, we can speculate that if the nematode

PAGE 158

158 establishes in Florida, it may infect the same hosts as other M. hapla populations. Moreover, M. hapla is also known to have a w ide host range affecting more than 550 crop and weed species (Jepson 1987). Currently, Meloidogyne spp. are not the most damaging nematodes on strawberries in Florida, but this could potentially change if more nematode infested transplants are imported int o the state. It is therefore important to put stricter regulations on clean strawberry transplants from the nurseries to ensure that growers are safe from such sources of nematode inoculum. Introduction of endoparasitic nematodes such as root knot nematode with the planting materials can be quite problematic because soil fumigation is only carried out before planting. Furthermore, endoparasitic nematodes are hard to target with any contact nematicide. Strawberry season in Florida runs from October to April, and after final harvest growers wait about 4 5 months before planting the next crop. Currently, quite a number of growers in the state are exploring the use of 2 yr old plastic mulch. In this approach, strawberry planting beds are irrigated intermittently throughout summer. In such a case, survival of nematodes under these conditions in the soil is unknown. Until further studies are conducted on the survivability over two or more years, it remains unclear whether this nematode can survive under Florida con ditions. Alternatively, growers may plant a cover crop and use a susceptible host such as a leguminous crop, which could only lead to an increase in pest numbers if the nematode is able to survive during the summer season. Conclusions This study ascerta ins that M. hapla recorded on strawberry plants in the field was imported with the transplants from the nurseries in Ontario, Canada. M eloidogyne hapla is commonly found in cooler climates and high altitudes, and has been reported as a common pest for str awberries in northeastern United States.

PAGE 159

159 Figure 8 1. Strawberry pl ants showing nematode symptoms ( stunted, excessive fibrous root growth and reduced crowns) versus the healthy plants (the last two plants) that were collected from Citra, Marion County, Florida in April 2011 A) and B) galled strawberry roots recovered from the bare root green top strawberry transplants shipment from Ontario, Canada nurseries in October 2011 A B

PAGE 160

160 Figure 8 2. Abundance (per 100cc soil) of Meloidogyne and Trichodor ous spp. from soil samples collected in blocks A, B, and C, planted to strawberries. Figure 8 3. Female perineal pattern of Meloidogyne hapla isolated from field strawberries collected from Citra, Marion County, Florida in April 2011 0 50 100 150 200 250 A B C Abundance / 100cc soil Strawberry Block Meloidogyne Trichodorus

PAGE 161

161 Figure 8 4. Amplification products using COII and 1RNA primers of the mtDNA from three Meloidogyne spp. females extracted from 2010/2011 and 2011/2012 strawberry plants on 1.8% agarose gel. All samples formed a band at 540 bp. The females are: 1, 201 0/2011 from strawberry plants after the growing season; 2 and 3, 2011/2012 from strawberry transplants upon arrival from a nursery in Ontario, Canada, and MK on the first and last loading lines is 1 kb DNA ladder (Invitrogen, Carlsbad, CA).

PAGE 162

162 Figure 8 5 Restriction fragment patterns of the PCR amplified products of the region between COII and 1RNA of the mtDNA after digestion with endonuclease Dra1 on 1.8% agarose gel. The samples formed two bands at 200 and 250 bp. The samples are: 1, 2010/2011 from st rawberry plants after the growing season; 2 and 3, 2011/2012 from strawberry transplants from a nursery in Ontario, Canada, and MK on the first and the last loading lines is 1 kb DNA ladder (Invitrogen, Carlsbad, CA).

PAGE 163

163 CHAPTER 9 CONCLUSION This study confi rms that control of twopsotted spider mites (TSSM) [ Tetranychus urticae ] on field grown strawberries is important in preventing yield losses and economic damage. Yield reduction was detected when strawberry plants had accumulated 4,923 mite days in 2008/20 09; at this point mites per trifoliate leaf were averaging 220 motiles per leaf (equivalent to 73 mites per leaflet). In 2009/2010, yield reduction was observed when strawberry plants had accumulated 1,938 mite days, averaging 83 motiles per trifoliate lea f within the high infestation mite density level. Early infestation coupled with favorable temperatures for TSSM will result in economic losses during mid season. Therefore, spider mites should be treated early in the season to prevent strawberry yield los ses later in the season. Prevailing weather conditions are also an important factor to consider when making management decision for TSSM. Weather parameters including temperature and rainfall, as well as host plant conditions could affect TSSM establishmen t and development. Management decision is also related to other factors such as strawberry market prices, crop condition (vigor), and targeted market (fresh or processing berries). Twospotted spider mites feeding on strawberry leaves did cause changes in s pectral characteristics within the visible and near infrared regions. Twospotted spider mites damage detection on strawberry leaves was possible from the spectral data, and a prediction model was developed. This model could predict TSSM numbers on a strawb erry leaf with relatively good accuracy (root mean square error of 17 mites/leaf). This model was found to be of medium accuracy. Since TSSM is very tiny (< mm), an error of 17 mites per leaf is still within an acceptable range especially when one consider s the first instar of spider mites (larva). In addition, strawberry leaves can support a very high number of mites (1000 mites per leaf [personal observation]). It is worth mentioning that TSSM damage detection and therefore

PAGE 164

164 prediction model developed will vary among strawberry varieties (Florida Festival and Chandler). Among the various pre processing methods for spectral data before analysis, derivatives, in particular Savitzky Golay first derivative using polynomial order 1 and Norris gap, proved to be the best methods for removing noise from the strawberry leaf spectra to predict TSSM, followed by Savitzky Golay smoothing method. It may not be advisable to pre select wavelengths for TSSM damage detection and in absence of pre processing methods, raw abs orbance data can be used. In a related study, factors that could be confounding in detecting and classifying TSSM damage on strawberry leaves using spectral characteristics were investigated. The accuracy of discriminating TSSM damage categories from spect ral data using TSSM damaged strawberry plants grown with and without nitrogen fertilizer was evaluated. The error rate in classifying TSSM damaged leaves from strawberry plants growing with and without fertilizer was relatively high and no clear pattern wa s observed. There seems to be no major difference between plants growing with and without fertilizer whether infested or not infested with TSSM. However, a field experiment could be used to verify the results where strawberry plants are grown for an extend ed period, longer than the 3 months used in this experiment. Another factor investigated was the difference between collecting spectral data from the adaxial (top) or abaxial (under) side of the leaf, and if this would improve TSSM injury detection on the strawberry leaves. Collecting spectra readings from abaxial side of the leaf resulted in a higher accuracy for delineating between high and non infested strawberry leaves (control) but a lower accuracy in classifying leaves that had a low TSSM infestation level. Both abaxial and adaxial spectra would have a problem in classifying low infested leaves from non infested leaves

PAGE 165

165 (control) and also from the high infested leaves. This is the drawback in using spectral readings to classify mite damaged strawberry l eaves. Growers would want to detect the damage at very early stages so that the appropriate measures could be taken to control TSSM populations. Furthermore, the time of TSSM infestation on strawberries is very critical and early infestation causes more in jury to the plants than late infestation. In order to understand the movement and dispersal of Neoseiulus californicus after inoculative releases to control TSSM, we studied their spatial and temporal distribution using geographic information system (GIS) and geostatistics. The goal was to understand the movement and dispersal of N. californicus after inoculative releases to control TSSM. A geo referenced sampling grid of 104 sampling points at intervals of 5 meters was established with an additional 25 sam pling points. Our result shows that a sampling distance of 13 37 m can be used in intensive sampling for TSSM, especially in site specific pest management. The results indicate that the ability of N. californicus to disperse is minimal and therefore rele ase distances shorter than the 20 meters used in this study should be investigated in order to understand their TSSM at those sites to zeros, an indication that N californicus also suited for site specific of management of spider mites. Two field studies were conducted to evaluate the effect of double cropping strawberry fields with a second strawberry crop, a production system that would save growers on their co st of production. It is estimated that growers can save up to $3,750/ha by re using the plastic mulch. Results indicate that leaving the old dead plants on the plastic mulch in double cropping with strawberries did not increase arthropod populations or aff ect plant growth of the second strawberry crop. However, it appears that these old dead plants could be a source of fungal

PAGE 166

166 inoculum and hence an increase in plant mortality (due to fungal diseases). Alternatively, pulling out the old plants created more ho les on the plastic mulch that permitted more weeds to germinate from the soil. It would be recommended that the same planting holes be utilized for the new crop to avoid making new holes in the plastic mulch. However, at this point we do not know if this w ill have adverse effects on the new transplants. Although there are costs involved in re using the plastic mulch such as weed control during summer, a thicker plastic mulch (1.25 mil) which is more expensive than regular mulch, maintaining the beds and dri p tapes, and repairs on the plastic mulch for wear and tear. Alternatively, the savings on double cropping with a second strawberry crop include cost of new plastic mulch for the second season, costs associated with summer cover crop (millet) such as plant ing the cover, fertilizing, irrigation, turning in the cover crop after the season, and land preparation procedures at the time of planting in the fall. Some growers may agree that double cropping with a second strawberry crop may be worth pursuing even if they have to spend an extra $70 to pull the old plants before transplanting. Overall, it is estimated that double cropping with a second strawberry crop would save growers (~ $1500/ acre [$3750/ha]. Lastly, molecular and morphological studies were conduct ed to confirm the identity of a root knot nematode imported with the strawberry transplants from Ontario, Canada in 2010 and 2011growing seasons. During the 2010/2011growing season, several strawberry plants were observed to be stunted and with low yields despite similar management practices across the blocks. The unevenly distributed nematode infestation across the blocks suggested the possibility that nematode infested transplants were only in a few of the boxes that were received from nurseries. Polymera se chain reaction amplification (PCR) of the mitochdrial DNA identified the root knot nematode, as Meloidogyne hapla and DNA sequencing of the internal transcribed

PAGE 167

167 spacer (ITS) region confirmed the identity with 98% similarity with other known M. hapla in the Genbank. This study confirmed that M. hapla was imported with the transplants from Ontario, Canada. Even though M hapla is known to thrive well in colder latitudes and in high elevations of the tropics, they significantly affected strawberry yields i n two of the plots that had a high M. hapla juveniles in the soil. This occurrence on transplants is an indication that stricter import sanitation rules must be put in place, particularly with respect to M. hapla which is not buy strawberry transplants only from certified nurseries as were the plants used in our experiments.

PAGE 168

168 LIST OF REFERENCES Adam, M. A. M., M. S. Phillips, and V. C. Blok. 2007. Molecular diagnos tic key for identification of juveniles of seven and economically important species of root knot nematode ( Meloidogyne spp). Plant Pathol. 56:190 197. Anonymous 2011. Short day strawberry plants.http://strawberryplants.org/2010/10/short day strawberry pla nts/ Accessed 23 September 2011. Averre, C. W., W. O. Cline, R. K. Jones, and R. D. Milholland. 2011. Diagnosis of strawberry diseases. North Carolina Cooperative Extension Service, North Carolina State University, Raleigh, NC http://www.ces.ncsu.edu/depts/hort/consumer/agpubs/ag 386.pdf Accessed December 2011. Binns, M. R., and P. Nyrop. 1992. Sampling insect populations for the purpose of IPM decision making. Annu. Rev. Entomol 37: 427 453. Bogrekci, I., and W. S. Lee. 2005. Spectral measurement of common soil phosphates. Trans. ASAE 48: 2371 2378. Bolstad, P., 2008. GIS fundamentals a first text on geographic information systems, 3rd Ed. College of Food, Agricultural and Natur al Resources, University of Minnesota, St. Paul, Minnesota. Buitenhuis, R. L., Shipp and C. S Dupree. 2010. Dispersal of Amblyseuis swirskii Athias Henriot (Acari: Phytoseiidae) on potted greenhouse chrysanthemum. Biol. Control 52: 110 114. Cakmak, I., A. Janssen, M. W. Sabelis, H. Baspinar. 2009. Biological control of an acarine pest by single and multiple natural enemies. Biol. control 50: 60 65. Carroll, M. W., J. A. Glaser, R. L. Hellmich, T. E. Hunt, T. W. Sappington, D. Calvin, K. Copenhaver, and J. F ridgen. 2008. Use of spectral vegetation indices derived from airborne hyperspectral imagery for detection of European corn borer infestation in Iowa plots. J. Econ. Entomol. 101: 1614 1623. Chandler, C. K., E. E. Albregts, and T. E. Crocker. 1993. Is it t ime to develop a new system for growing strawberries in Florida? Proc. Fla. State Hort. Soc. 106: 144 146. Chandler, C.K., T.E. Crocker, J. F. Price, and E.E. Albregts. 2008. Growing strawberries in the home garden. Cooperative Extension, Services, IFAS, U niversity of Florida, Gainesville, FL. http://strawberry.ifas.ufl.edu/productionguideintro.htm Accessed June 2009 Chang, W. C, D. A. Laird, M. J. Mausbach, an d C. R. Hurburgh, Jr. 2001. Near infrared reflectance spectroscopy principal components regression analyses of soil properties. Soil Sci. Soc. Am. J. 65:480 490. Cloyd, R. A. 2008. All Predatory Mites Are Not Created Equal. Kansas State University, Kansas. http://www.greenhousegrower.com/article/17974 Accessed October 2011. Curran, P. J. 1989. Remote sensing of foliar chemistry. Remote Sens. Environ. 30: 271 278. Curran, P. J, .J. L. Dungan, B. A. Macler, and S. E. Plummet. 1991. The effect of a red leaf pigment on the relationship between red edge and chlorophyll concentration. Remote Sens. Environ. 35:69 76.

PAGE 169

169 Cone, W.W. 1985. Mating and chemical communication, pp. 243 247. In : W. Helle & M. W. S abelis (eds.). Spider Mites: Their Biology, Natural Enemies and Control. Elsevier, Amsterdam. Mediterranean islands with similar climatic environments by means of mid inf rared diffuse reflectance spectroscopy. Eur. J. Soil Sci. 61:865 876. Daughtry C. S. T., C. L. Walthall, M. S. Kim, E. Brown de Colstoun and J. E. McMurtrey III. 2000. Estimating corn leaf chlorophyll concentration from leaf and canopy reflectance Remote S ens. Environ. 74:229 239. DeAngelis, D. J., K. C. Larkson, R. E. Berry, and G. W. Krantz. 1982. Effects of spider mite injury on transpiration and leaf water status in peppermint. Environ. Entomol. 11: 975 978. Dunn, B. W., H. G. Beecher, G. D. Batten, an d S. Ciavarella. 2002. The of near infrared reflectance spectroscopy for soil analysis a case study from Riverine plain of south eastern Australia. Aust. J. Exp. Agric. 42: 607 614. Esbenshade, P R., and A. C. Triantaphyllou. 1990. Isozyme phenotypes for the identification of Meloidogyne species. J. Nematol. 22:10 15. Redlands, California. http://dusk.ge o.orst.edu/gis/geostat_analyst.pdf Ess, D., and M. Morgan. 2003. The precision farming guide for agriculturists, 138pp. John Deere, Moline, IL. Fasulo, T. R and H. A. Denmark. 2000. Twospotted Spider Mite, Tetranychus urticae Koch (Arachnida: Acari: Tetra nychidae). EENY 150. University of Florida, IFAS Extension, Gainesville, FL. Fitzgerald, G. J., S. J. Maas, and W. R. Detar. 2004. Spider mite detection and canopy component mapping in cotton using hyperspectral imagery and spectral mixture analysis. Preci s. Agric. 5: 275 289. Fleischer, S. J., P. E. Blom, and R. Weisz. 1999. Sampling in precision IPM: when the objective is a map. Phytopathalogy 88: 1113 1118. Flint, M. L., and P. Gouveia. 2001. IPM in practice principles and methods of integrated pest mana gement. University of California statewide integrated pest management project agriculture and natural resources. Publication 3418, University of California, Davis, CA. Florida Automated Weather Network (FAWN). 2011.http://fawn.ifas.ufl.edu/Accessed Septemb er 23rd 2011 Fourie, H., C. Zijlstra, and A. H. McDonald. 2001. Identification of root knot nematode species occurring in South Africa using the SCAR PCR technique. Nematology 37:675 680. Fraisse, C., and N. Peres. 2011. Cold La Nia surprises meteorologi sts and growers. Berry/Vegetable Times, February 2011, University of Florida, UF IFAS Extension. http://strawberry.ifas.ufl.edu/BerryTimes/BVT0810.pdf

PAGE 170

170 Fraulo, A. B. 2007. Timed releases with Neoseiulus californicus as a biological control agent for Tetran ychus urticae Koch and its ecological impact on North Florida strawberry fields, M.S. Thesis University of Florida, Gainesville, FL. Fraulo, A. B., and O. E. Liburd. 2007. Biological control of twospotted spider mite, Tetranychus urticae with predatory mi te, Neoseiulus californicus in strawberries. Exp. Appl. Acarol. 43: 109 119. Fraulo, A. B., M. Cohen, and O. E. Liburd. 2009. Visible/Near Infrared reflectance (VNIR) spectroscopy for detecting twospotted spider mite (Acari: Tetranychidae) damage in straw berries. Environ. Entomol. 38:137 142. Fraulo, A. B., R. McSorley, and O. E. Liburd. 2008. Effect of the biological control agent Neoseiulus californicus (Acari: Phytoseiidae) on arthropod community structure in North Florida strawberry fields. Fla. Entomo l. 91: 436 445. Freeman, S. and N. Gnayem. 2005. Use of plasticulture for strawberry plant production. Small Fruits Rev. 4: 21 32. Garcia, F. J. M. 2006. Analysis of the spatio temporal distribution of Helicoverpa armigera Hb. In a tomato field using stoch astic approach. Bio Syst. Eng. 93: 253 259. Gates, D. M., H. J. Keegan, J. C. Schleter, and V. R. Weidner. 1965. Spectral properties of plants. Appl. Optica 4:11 20. Gimnez Ferrer, R. M, W.A. Erb, B.L. Bishop, and J.C. Scheerens. 1994. Host pest relations hips between the two spotted spider mite (Acari: Tetranychidae) and strawberry cultivars with different levels of resistance. J. Econ. Entomol. 87:168 175. Greco, N. M., G. G. Liljesthrm, and N. E. Snchez. 1999. Spatial distribution and coincidence of Ne oseiulus californicus and Tetranychus urticae (Acari: Phytoseiidae, Tetranychidae) on strawberry. Exp. Appl. Acarol. 23: 567 580. Greco, N. M., N.E. Sanchez and G. G. Liljesthrom. 2005. Neoseiulus californicus (Acari: Phytoseiidae) as potential control age nt of Tetranychus urticae (Acari: Tetranychidae): effect of pest/predator ratio on pest abundance on strawberry. Exp. Appl. Acarol. 37:57 66. Grunwald, S. (ed.), 2006. Environmental soil landscape modeling geographic information technologies and pedometr ics. p. 488. CRC Press, New York. Handoo, Z, A., A. M. Skantar, L. K. Carta, and D. P. Schmitt. 2005. Morphological and molecular evaluation of a Meloidogyne hapla population damaging coffee ( Coffee arabica ) in Maui, Hawaii. J. Nematol. 37:136 145. Harris, T. S., L. J. Sandall, and T. O. Powers. 1990. Identification of single Meloidogyne juveniles by polymerase chain reaction amplification of mitochondrial DNA. J. Nematol. 22:518 524. Hartman, K. M., and J. N. Sasser. 1985. Identification of Meloidogyne spe cies on the basis of differential host test and perineal pattern morphology. Pp. 69 77 in Barker K. R., C. C. Carter, and J. N. Sasser, eds. An advanced treatise on Meloidogyne Volume II: Methodology, eds. NC State University Graphics, Raleigh, NC.

PAGE 171

171 Hartz, T.K., J. E. De Vay, and C. L. Elmore. 1993. Solarization is an effective soil disinfestations technique for strawberry production. Hortscience 28: 104 106. Helle, W., and W. P. J. Overmeer. 1973. Variability in tetranychid mites. Ann. Rev. Entomol. 18:97 120. Herbert, H. J. 1981. Biology, life tables and innate capacity for increase of the two spotted spider mite Tetranychus urticae (Acarina: Tetranychidae). Can. Entomol. 113:371 378. Himelrick, D. G., F. M. Woods, W. A. Dozier, Jr., and J. D. Williams. 1995. Soil fumigation and soil solarization in the annual hill plasticulture system. Adv. Strawberry Res. 14: 69 72. Howard, C. M., A. J. Overman, J. F. Price, and E. A. Albregts. 1985. Diseases, nematodes, mites, and insects affecting strawberries in Flor ida. Bulletin 857. Agricultural Experiment Stations, Institute of Food and Agricultural Sciences (IFAS), University of Florida, Gainesville, FL. Hu, M. X., K. Zhuo, and J. L. Liao. 2011. Multiplex PCR for the simultaneous identification and detection of M eloidogyne incognita M enterolobii and M javanica using DNA extracted directly from individual galls. Phytopathology 101:1270 1277. Hubbard, J. B., K. L. Flanders, R. W. Mankin, E. A. Guertal, and R. L. Crocker. 2001. Mapping of soil insect infestatio ns sampled by excavation and acoustic methods. J. Econ. Entomol. 94:1452 1458. Huffaker, C.B., M. van de Vrie and J.A. McMurtry. 1969. The ecology of tetranychid mites and their natural control. Ann. Rev. Entomol. 14:125 74. Isaaks E. H. and M. R. Sriva stava.1989. An introduction to applied geostatistics. Oxford University Press, New York. Jamshidi, B., S. Minaei, E. Mohajerani, and H. Ghassemian. 2012. Reflectance Vis/NIR spectroscopy for nondestructive taste characterization of Valencia oranges. Comp. Electron. Agric. 85: 64 69. Jenkins, W. R. 1964. A rapid centrifugal floatation technique for separating nematodes from soil. Plant Dis. Rep. 48:692. Jensen, J. R. 2005. Introductory digital image processing: a remote sensing perspective. Pearson Prentice Hall, Upper Saddle River, NJ. Jeppson, L. R., H. H. Keifer and E. W. Baker. 1975. Mites injurious to economic plants. University of California Press, Los Angeles, California. Jepson, S. B. 1987. Identification of root knot nematodes ( Meloidogyne species). CAB International, Wallingford, UK. Joyce, S.A., A. P. Reid, F. Driver, and J. Curran. 1994. Application of polymerase chain reaction (PCR) methods to the identification of entomopathogenic nematodes. Pp.178 187 I n : A.M. Burnell, R.U. Ehlers, and J.P. Mass on, eds. Genetics of entomopathogenic nematode bacterium complexes College, Maynooth, Co, Kildare, Ireland. Luxembourg, DGXII, European commission. Kasperbauer, M. J. 2000. Strawberry yield over red ver sus black plastic mulch. Crop Sci. 40:171 174.

PAGE 172

172 Katan, J. 1981. Solar heating (solarization) of soil for control of soil borne pests. Ann. Rev. Phytopathol. 19: 211 236. Kelly, M., and Q. Guo. 2007. Integrated agricultural pest management through remote se nsing and spatial analysis, pp. 191 207. In : A. Ciancio & K. G. Mukerji [eds.], General concepts in integrated pest and disease management. Springer, Netherlands. Kennedy, G. G., and D. R. Smitley. 1985. Dispersal, pp. 233 242. In : W. Helle and M.W. Sabeli s [eds.], Spider mites, their biology, natural enemies and control, Vol. 1A. Elsevier, Amsterdam Kielkiewicz, M. 1985. Ultrastructural changes in strawberry leaves infested by two spotted spider mites. Entomol. Exp. Appl. 37: 49 54 Knapp, M., I. Sarr, G Gilioli. 2006. Population models for threshold based control of Tetranychus urticae in small scale Kenyan tomato fields and for evaluating weather and host plant species effects. Exp. Appl. Acarol. 39: 195 212. LaMondia, J. A. 2002. Seasonal populations of Pratylenchus penetrans and Meloidogyne hapla in strawberry roots. J. Nematol. 34:409 413. Liburd, O. E., J. C. White, E. M. Rhodes, and A. A. Browdy. 2007. The residual and direct effects of reduced risk and conventional miticides on twospotted spider m ites, Tetranychus urticae (Acari: Tetranychidae) and predatory mites (Acari: Phytoseiidae). Fla. Entomol 90:249 257. Liebhold, A. M., R. E Rossi, and W. P. Kemp 1993. Geostatistics and geographic information systems in applied insect ecology. Ann. Rev. En tomol. 38: 303 327. Liebhold, A. M., X. Zhang M. E. Hohn, J. S. Elkinton M. Ticehurst M G. L. Benzon, and W. Campbell 1 991. Geostatistical analysis of Gypsy moth (Lepidoptera: Lymantriidae) egg mass populations. Environ. Entomol. 20:1407 1417. Mari, G F and G. J. E. Zamora. 1999. Biological control of Tetranychus urticae (Acari: Tetranychidae) with naturally occurring predators in strawberry plantings in Valencia, Spain. Exp. Appl. Acarol. 23:487 495 Martens, H. and T. Ns. 1 989. Multivariate calibrati on, 419 pp. John Wiley & Sons, Chichester, United Kingdom. McSorley R., and H. K. Gill. 2010. Introduction to soil solarization. ENY 062. University of Florida, IFAS extension, Gainesville, FL. Mertely, J., and N. Peres. 2010. Plant disease considerations when replanting strawberries on old plastic. Berry/Vegetable Times, August/September 2010, University of Florida UF IFAS Extension. http://strawberry.ifas.ufl.edu/BerryTimes/BVT0810.pdf Mirik, M., G.J. Michelis (Jr), S. K Mirik, and N. C. Elliot. 2007. Reflectance characteristics of Russian wheat aphid (Hemiptera:Aphididae) stress and abundance in winter wheat. Comp. Electron. Agric. 51: 81 86 Mirik, M., G.J. Michelis (Jr), S. K Mirik, N. C. Elliot and R. Bowling. 2006. Hyperspectral spectrometry as a m eans to differentiate uninfested and infested winter wheat by greenbug (Hemiptera:Aphididae). J. Econ. Entomol. 99: 1682 1690.

PAGE 173

173 Mossler, M. A. and O. N. Nesheim. 2003. Strawberry pest management strategic plan (PMSP). Cooperative Extension, Services, IFAS, University of Florida, Gainesville, FL. Mossler, M. A., and O. N. Nesheim. 2007. Strawberry pest management strategic plan (PMSP), CIR1443. Cooperative Extension, Services, IFAS, University of Florida, Gainesville, FL. Nachman, G. J. 1981. Temporal and spatial dynamics of an acarine predator prey system. J. Anim. Ecol. 50: 435 451 National Agricultural Statistical Service (NASS) USDA. 2008. Non citrus fruits and nuts, pp. 28 31 2008 preliminary summary. http://www.nass.usda.gov/. Accessed September 2009 National Agricultural Statistical Service (NASS) USDA. 2011. Non citrus fruits and nuts, pp. 28 31 2011 preliminary summary (March 2012). http://www.nass.usda.gov/. Accessed April 2012. Natsuga, M. and S. Kawamura. 2006. Visible and near infrared refl ectance spectroscopy for determining physicochemical properties of rice. Transactions of the ASABE 49: 1069 1076. Nicola, B. M., K. Beullens, E. Bobelyn, A. Peirs, W. Saeys, K. I. Theron, and J. Lammertyn 2007. Nondestructive measurement of fruit and ve getable quality by means of NIR spectroscopy: a review. Postharvest Biol. Tec. 46, 99 118. Nilsson, H. E. 1995. Remote sensing and image analysis in plant pathology. Annu. Rev. Phytopathol. 15:489 527. Noling, J. W. 1999. Nematode management in strawberri es ENY 031. IFAS extension, University of Florida, Gainesville, FL http://edis.ifas.ufl.edu/pdffiles/NG/NG03100.pdf Accessed on December 14 2011. Noling, J. W., and A. Whiden. 2009. Methyl bromide transition strategy: A fall 2009 focus on drip fumigation under old and new plastic. Berry/Vegetable Times, September 2009, University of Florida UF IFAS Extension. http://strawberry.ifas.ufl.edu/BVT0909.pdf. Noling, J.W., and A. Whidden. 2010. I nternational issues and growing concerns for new nematode problems in the Florida strawberry industry. Berry/Vegetable Times, June 2010, UF IFAS Extension, University of Florida, Gainesville, FL. http://strawberry.ifas.ufl.edu/BerryTimes/BVT0610.pdf Noling, J. W. 2011. How do you replace methyl bromide? Vegetable Growers News Volume 45 No. 6. Norris, K. H., and P. C. Williams 1984. Optimization of mathematical treatments of raw near inf rared signal in the measurement of protein in hard red spring wheat,: influence of particle size. Cereal Chem. 62, 158 165. Oatman, E. R.,and J. A. Mcmurtry. 1966. Biological control of the twospotted spider mite on strawberry in southern California. J. Ec on. Entomol. 59: 433 439. Opit, G. P., J. Perret, K. Holt, J. R. Nechols, D. C. Margolies, and K. A. Williams. 2009. Comparing chemical and biological control strategies for twospotted spider mites (Acari: Tetranychidae in commercial greenhouse production of bedding plants. J. Econ. Entomol. 102: 336 346.

PAGE 174

174 Orui, Y. 1998. Identification of Japanese species of genus Meloidogyne (Nematoda: Meloidogynidae) by PCR RFLP analysis. Appl. Entomol. Zool. 33:43 51. Osborne, L.S., L. E. Ehler, and J. R. Nechols. 1999. B iological control of the twospotted spider mite in greenhouses. Bulletin 853 (technical) 12/85 (updated 7/12/99). University of Florida Central Florida Research and Education Center, Apopka, FL. Park, Y. L., and J. H. Lee. 2005. Impact of twospotted spider mites (Acari: Tetranychidae) on growth and productivity of glasshouse cucumbers. J. Econ. Entomol. 98: 457 463. Park, Y. L., and J. J. Tollefson. 2005. Spatial prediction of corn rootworm (Coleoptera: Chrysomelidae) adult emergence in Iowa cornfields. J. Econ. Entomol. 98: 121 128. Park, Y. L., and J. J. Tollefson. 2006. Development and economic evaluation of spatial sampling plans for corn rootworm Diabrotica spp. (Col., Chrysomelidae) adults. J. Appl. Entomol. 130: 337 342. Park, Y. L., R. K. Krell, and M. Carroll. 2007. Theory, technology, and practice of site specific insect pest management. J. Asia Pacific Entomol. 10: 89 101. Pedigo, L. P. 1996. Entomology and pest management, 2nd Edition, 679 pp. Prentice Hall, Upper Saddle River, New Jersey. Pedigo, L. P., and L. G. Higley. 1996. Theory, pp.3 8 In : Higley, L. G., and L. P. Pedigo [Eds]. Economic thresholds for integrated pest management, 327 pp. University of Nebraska, Lincoln, Nebraska. Pedigo, L. P., and M. E. Rice. 2009. Entomology and pest manage ment, 6th Edition, 784 pp. Pearson Prentice hall, Upper Saddle River, New Jersey. Penman, D. R., and W. W. Cone. 1972. Behavior of the male twospotted spider mite in response to quiescent female deutonymphs and to web. Ann. Entomol. Soc. Am. 65: 1289 1293 Peres, N. A., J. F. Price, W. M. Stall, C. K. Chandler, S. M. Olson, T. G. Taylor, S. A. Smith, and E. H. Simonne. 2006. Strawberry production in Florida. Pp. 375 405 In : S. M.Olson and E. H. Simonne, eds. Vegetable production handbook for Florida 2006 2 007. University of Florida, IFAS publication, Gainesville, FL. Peres, N. A., J. F. Price, W. M. Stall, C. K. Chandler, B. M. Santos, S. M. Olson, S. A. Smith, and E. H. Simonne. 2010. Strawberry production in Florida. Pp. 263 272 In : S. M.Olson and B. Sant os, eds. Vegetable production handbook for Florida 2010 2011. University of Florida, IFAS publication, Gainesville, FL. Poe, S. L. 1971. Influence of host plant physiology on populations of Tetranychus urticae (Acarina: Tetranychidae) infesting strawberry plants in Peninsular Florida. Fla. Entomol. 54: 183 186. Powers, T. O., and T. S. Harris. 1993. A polymerase chain reaction method for identification of five major Meloidogyne species. J. Nematol. 25:1 6. Powers, T. O., P. G. Mullin, T. S. Harris, L. A. S utton, and R. S. Higgins. 2005. Incorporating molecular identification of Meloidogyne spp. into a large scale regional nematode survey. J. Nematol. 37:226 235.

PAGE 175

175 Qui, J. J., B. B. Westerdahl, C. Anderson, and V. M. Williamson. 2006. Sensitive PCR detection o f Melodoigyne arenaria M. incognita and M. javanica extracted from soil. J. Nematol. 38:434 441. Raworth, D. A. 1986. An economic threshold function for the twospotted spider mite, Tetranychus urticae (Acari: Tetranychidae), on strawberries. Can.Ent. 118 : 9 16. Raworth, D. A. 1990. Predators associated with the twospotted spider mites, Tetranychus urticae on strawberry at Abbotsford, BC, and development of non chemical mite control. J. Entomol. Soc. British Columbia 87: 59 67. Reddall, A., V. O. Sadras, L. J. Wilson, and P. C. Gregg. 2004. Physiological responses of cotton to twospotted spider mite damage. Crop Sci. 44: 835 846 Reidell, W. E., and T. M. Blackmer. 1999. Leaf reflectance spectra of cereal aphid damaged wheat. Crop Sci. 39: 1835 1840. Reis ig, D., and L. Godfrey 2006. Remote sensing for detection of cotton aphid (Homoptera: Aphidae) and spider mite (Acari: Tetranychidae) infested cotton in the San Joaquin Valley. Environ. Entomol. 35:1635 1646. Reisig, D., and L. Godfrey. 2007. Spectral r esponse of cotton aphid (Homoptera: Aphidae) and spider mite (Acari: Tetranychidae) infested cotton: controlled studies. Environ. Entomol. 36:1466 1474. Rhodes, E. M., and O. E. Liburd. 2006. Evaluation of predatory mites and Acramite for control of twos potted spider mites in strawberries in north central Florida. J. Econ. Entomol. 99: 1291 1298. Rhodes, E. M., O. E. Liburd, and S. Grunwald. 2011. Examining the Spatial Distribution of Flower Thrips in Southern Highbush Blueberries by Utilizing Geostatisti cal Methods. Environ. Entomol. 40:893 903. Rondon, S. I., D. J. Cantliffe, and J. F.Price. 2004. The feeding preferences of the bigeyed bug, minute pirate bug, and pink spotted lady beetle relative to main strawberry pests. Environ. Entomol. 33: 1014 1019. Rondon, S. I., D. J. Cantliffe, and J. F. Price. 2005. Population dynamics of the cotton aphid, Aphis gossypi (Homoptera: Aphididae), on strawberries grown under protected structure. Fla. Entomol. 88: 152 158. Rosskopf, E. N., D. O Chellemi, N. Kokalis Burelle, and G. T. Church. 2005. Alternatives to methyl bromide: A Florida perspective. APS net feature story June 2005. http://www.apsnet.org/publications/apsnetfeatures/Documents/2005/MethylBromideAlter natives.pdf Sances, F. V., J. A. Wyman, and I. P. Ting. 1979. Morphological responses of strawberry leaves to infestations of twospotted spider mite. J. Econ. Entomol. 72: 710 713. Sances, F. V., J. A. Wyman, I. P. Ting, R. A. Van Steenwyk and E. R. Oatman. 1981. Spider mite intera ctions with photosynthesis, transpiration and productivity of strawberry. Environ. Entomol. 10: 442 448.

PAGE 176

176 Sances, F.V, N. C Toscano, L. F LaPre, E. R Oatman, and M. W. Johnson. 1982. Spider mites can reduce strawberry yields. Calif. Agric. 36:15 16. Santos B.M., N. A. Peres, J. F. Price, C. K. Chandler, V. M. Whitaker, W. M. Stall, S. M. Olson, S. A. Smith and E. H. Simonne. 2011. Strawberry production in Florida. Horticultural Sciences Department Document HS736, University of Florida, Institute for Food and Agricultural Sciences and Florida Cooperative Extension Service. http://edis.ifas.ufl.edu/pdffiles/CV/CV13400.pdf SAS Institute. 2002. SAS (Statistical Analysis System) for Windows, Version 9.2. SAS Institute, Inc., Cary, NC, USA. Sato, M. E., M. Z. D Silva, F. D.S. Filho, A. L. Matioli, and A. Raga. 2007. Management of Tetranychus urticae (Acari: Tetranychidae) in strawberry fields with Neoseiulus californicus (Acari: Phytoseiidae) and acaricides. Exp. Appl. Acarol. 42:107 120. Savitzky, A., and M. J E. Golay 1964. Smoothing and differentiation of data by simplified least squares procedures. Anal. Chem. 36:1627 1639. Schotzko, D. J., and L. E. O'Keeffe. 1989. Geostatistical description of the spatial distribution of Lygus Hesperus (Heteroptera: Miri dae) in lentils. J. Econ. Entomol. 82: 1277 1288. Shih, C. T., S. L. Poe, and H. L. Cromroy. 1976. Biology, life table, and intrinsic rate of increase of Tetranychus urticae Ann. Entomol. Soc. Am. 69: 362 364. Taylor, A. L., and J. N. Sasser. 1978. Biolog y, identification and control of root knot nematodes ( Meloidogyne species). North Carolina State University Graphics, Raleigh, NC. Tobler, W. 1970. A computer movie simulating urban growth in the Detroit region. Econ. Geogr. 46:234 40. Tomczyk, A., and D. 329. In : W. Helle and M.W. Sabelis [eds]. Spider mites, their biology, natural enemies and control, Vol. 1A. Elsevier, Amsterdam. US. EPA. 2005. U.S. Nomination for methyl bromide critical use exemption s from the 2007 phase out of methyl bromide. U. S. Environmental Protection Agency. http://www.epa.gov/ozone/mbr/2007_nomination.html Accessed April 2012. Vasques G. M., S. Grunwald and J. O Sickman. 2008a. Comparison of multivariate methods for inferential modeling of soil carbon using visible/near infrared spectra. Geoderma 146: 14 25. Vasques G. M., S. Grunwald and J. O. Sickman. 2008b.Modeling of soil organic carbon fractions using visib le near infrared spectroscopy. Soil Sci. Soc. Am. J. 73: 173 184. Veerman, A. 1985. Diapause, pp. 279 316. In W. Helle and M.W. Sabelis (eds). Spider mites, their biology, natural enemies and control, Vol. 1A. Elsevier, Amsterdam. Walsh, D., and F. Zalo m. 1996. Chemical control of the two spotted spider mite. University of California, Davis. http://www.calstrawberry.com/research_library/96 11.pdf. Accessed October 2009.

PAGE 177

177 Walsh D.B, F.G. Zalom, and D.V. Shaw. 1998. Interaction of the two spotted spider mit e (Acari: Tetranychidae) with yield of day neutral strawberries in California. J. Econ. Entomol. 91:678 685. Waterer, D., W. Hrycan, and T. Simms. 2007. Potential to double crop plastic mulch. Can. J. Plant Sci. 88: 187 193. Weisz, R., S. Fleischer, and Z. Smilowitz. 1995. Map generation in high value horticultural integrated pest management: appropriate interpolation methods for site specific pest management of Colorado potato beetle (Coleoptera: Chrysomelidae). J. Econ. Entomol. 88: 1650 1657. Westerdahl, B. B. 2009. Nematodes. UC IPM Pest management guidelines: Strawberry. University of California, Davis, CA. http://www.ipm.ucdavis.edu/PMG/r734200111.html Accessed Ja nuary 2012 White, J. C., and O. E. Liburd. 2005. Effect of soil moisture and temperature on reproduction and development of twospotted spider mite (Acari:Tetranychidae) in strawberries. J. Econ. Entomol. 98: 154 158. Williams, P., and K. Norris. 2001. Ne ar infrared technology in the agricultural and food industries. 2nd ed. St. Paul, Minn.: American Association of Cereal Chemists. Wyman, J. A., E. R. Oatman, and V. Voth. 1979. Effects of varying twospotted spider mites infestation levels on strawberry yie ld. J. Econ. Entomol. 72: 747 753. Yik, C. P and W. Birchfield. 1978. Scanning electron microscopy of perineal patterns of three species of Meloidogyne. J. Nematol. 10: 118 122. Yoder, B. J., and R. E. Pettigrew Crosby. 1995. Predicting nitrogen and chlorophyll content and concentra tions from reflectance spectra (400 2500 nm) at leaf and canopy scales. Remote Sens. Environ. 53: 199 211.

PAGE 178

178 BIOGRAPHICAL SKETCH Teresia Nyoike was born in the Thika District of Central Kenya. She attended University of Nairobi, Kenya where she attained her Bachelors in Agriculture (Crop Protection) on August 2001. Upon finishing her BSc, she immediately joined Dudutech Kenya Limited, an Integrated Pest Management (IPM) Company that produced bio control agents for local and export markets. During her four yea rs with the company, she held specific responsibilities of developing bio control agents for root knot nematodes and also advising farm managers on integrated pest University of Florida in 2007. Her MS research evaluated the use of living and synthetic mulches to control key pests of zucchini squash. She enrolled for her doctoral studies in the same institution in August 2008 after working as a research associate at the Small Fruit and Vegetable IPM Laboratory, University of Florida for 8 months. Her doctoral research was focused on improving sampling techniques and management practices for twospotted spider mites on strawberries. She received her PhD in ento mology in August 2012. She hopes to continue her interest in IPM and contribute to sustainable pest management of key pests.