INFORMATION TO USERS T his was p ro d u c ed fro m a c o p y o f a d o c u m e n t s e n t to us fo r m icrofilm ing. W hile th e m o st advanced tech n o lo g ical m eans to p h o to g ra p h and re p ro d u ce this d o c u m e n t have been used, th e q u a lity is heavily d e p e n d e n t u p o n th e q u ality o f th e m aterial su b m itte d . T he fo llo w ing e x p la n a tio n o f tech n iq u e s is provided to help y o u u n d ersta n d m arkings o r n o ta tio n s w hich m ay a p p e ar on th is re p ro d u c tio n . 1 .T h e sign o r " ta r g e t” fo r pages a p p a re n tly lacking fro m th e d o c u m e n t p h o to g rap h e d is “ Missing P age(s)” . If it w as possible to o b tain th e m issing page(s) or se c tio n , th e y are spliced in to th e film along w ith a d jac en t pages. This m ay have n ec essita ted c u ttin g th ro u g h an image and du p licatin g ad jac en t pages to assure y o u o f c o m p le te c o n tin u ity . 2. W hen an im age on th e film is o b lite ra te d w ith a ro u n d black m ark it is an in d icatio n th a t th e film in sp e c to r n o tic e d e ith e r blu rred co p y because o f m o v em en t d u rin g e x p o su re , o r d u p licate c o p y . Unless we m ean t to d elete co p y rig h te d m aterials th a t sh o u ld n o t have been film ed , y o u will fin d a good im age o f th e page in th e a d jac en t fram e. If co p y rig h te d m aterials w ere d eleted y o u will fin d a ta rg e t n o te listing th e pages in th e a d jac en t fram e. 3. W hen a m ap, d raw in g o r c h a rt, e tc ., is p a rt o f th e m aterial being p h o to ­ g raphed th e p h o to g ra p h e r has follow ed a d e fin ite m eth o d in " se c tio n in g ” th e m aterial. It is c u sto m a ry to begin film ing a t th e u p p er le ft h an d co rn e r o f a large sh ee t and to c o n tin u e fro m le ft to right in equal sections w ith small overlaps. If n ecessary, sectio n in g is c o n tin u e d again—beginning below th e firs t row and c o n tin u in g on u n til c o m p lete. 4. F o r any illu stra tio n s t h a t c a n n o t be re p ro d u c e d satisfac to rily by x ero g rap h y , p h o to g rap h ic p rin ts can be pu rch ased a t a d d itio n al co st and tip p e d in to y o u r xero g raphic co p y . R eq u ests can be m ade to o u r D issertations C ustom er Services D e p artm en t. 5. S om e pages in an y d o c u m e n t m ay have in d istin c t p rin t. In all cases w e have film ed th e b est available co p y . University Microfilms International 300 N. ZEEB RD.. ANN ARB OR, Ml 48106 8202456 J e n k i n s , Je f f r e y J a m e s ASSESSMENT OF THE ATTENUATION AND MOVEMENT OF AZINPHOSMETHYL IN A MICHIGAN APPLE ORCHARD ECOSYSTEM Michigan State University University Microfilms International Ph.D. 1981 300 N. Zeeb Road, Ann Arbor, MI 48106 PLEASE NOTE: in all c a s e s th is m aterial h a s b e e n film ed in th e b e st p o ssib le w ay from th e available c o p y . P ro b lem s e n c o u n te re d with th is d o c u m e n t have b ee n identified h e re with a c h e c k m ark V ^ 1. G lossy p h o to g ra p h s o r p a g e s 2. C olored illustrations, p a p e r o r p rin t______ 3. P h o to g ra p h s with d ark b a c k g ro u n d ______ 4. Illustrations a re p o o r c o p y _______ 5. P a g e s with black m arks, n o t original 6. Print s h o w s through a s th e re is tex t on both s id e s of p a g e _______ 7. Indistinct, broken o r sm all print on several p a g e s 8. Print e x c e e d s m argin re q u ire m e n ts ______ 9. . c o p y _____ Tightly b o u n d co p y with print lost in s p in e ______ 10. C o m p u ter printout p a g e s w ith indistinct print_______ 11. P a g e ( s ) _____________ lacking w hen material re ceiv ed , an d not a v a ila b le from s c h o o l or author. 12. P a g e ( s )_____________ s e e m to b e m issing in num bering only a s te x t follows. 13. Two p a g e s n u m b e re d ______________. Text follows. 14. Curling a n d wrinkled p a g e s 15. O ther_____________________________________________________________________ _ i/ University Microfilms International ASSESSMENT OF THE ATTENUATION AND MOVEMENT OF AZINPHOSMETHYL IN A MICHIGAN APPLE ORCHARD ECOSYSTEM By J e ffre y Jam es Jenkins A DISSERTATION Subm itted to M ichigan S ta te U niversity in p a rtia l fulfillm ent of the requirem ents for th e degree of DOCTOR OF PHILOSOPHY D epartm ent of Entomology 1981 ABSTRACT ASSESSMENT OF THE ATTENUATION AND MOVEMENT OF AZINPHOSMETHYL IN A MICHIGAN APPLE ORCHARD ECOSYSTEM By J e ffre y Jam es Jenkins The environm ental behavior of azinphosm ethyl was studied in a Michigan apple orchard w atershed to g a th er d ata on in itia l distribution within th e orchard, v e rtic a l m ovem ent of the pesticide under the influence of rain fall, and loss from the orchard with runoff. The estim ated proportion of a low-volume application initially distributed w ithin th e orchard averaged .624 (standard deviation of .149) over th re e seasons (1976-1978). Exam ination of residues reaching each layer showed the m ajority of the dislodgeable residues were d istrib u ted to th e tre e s and grass-broadleaves. The litter-m o ss and soil contained residue levels roughly ten tim es lower than tre e le a f residues. R unoff studies indicated loss, via this ro u te, of less than 1% of azinphosm ethyl residues p resen t in the orchard. The residue d a ta w ere used to p aram eterize a model for azinphosm ethyl a tten u atio n and m ovem ent in an orchard ecosystem . R ates of a tte n u atio n w ithin, and m ovem ent betw een, specified orchard com partm ents w ere determ ined under various-rainfall regim es. The outp u t of th is model was s tru c tu re d to allow th e estim ation of the tim e course of azinphosm ethyl exposure to ground-dwelling in v e rte b ra tes. Mean squared errors for th e com parison of th e model predictions w ith an independent set of residue d a ta indicated good prediction of azinphosm ethyl fa te w ithin th e tre e and grass-broadleaves layers. Prediction of J e ffre y Jam es Jenkins p esticide dynam ics w ithin th e litte r-m o ss and soil layers was much more d ifficu lt. Model predictions estim ate th a t under dry conditions 25% of th e daily loss of azinphosm ethyl from th e orchard tre e s is due to m ovem ent to o th e r p arts of the orchard. G reater m ovem ent is p red icted under rain fall conditions. E stim ates of daily airborne loss determ ined from deposit residues and d irect sam pling of airborne residues suggest th a t airborne loss is largely responsible for th e early loss of residues, accounting fo r 40% of the daily loss ra te on day 3 of th e first spray period, 1978 season. A m ulti-com ponent k inetic model is p resented for estim ating sim ultaneously the early airborne loss of foliar deposits and the o ften slower dissipation of the rem aining residues. F u rth er model developm ent to include th e e ffe c t o f sele c ted environm ental p aram eters on azinphosm ethyl degradation is also discussed. ACKNOWLEDGMENTS This research com prises a major portion of a p ro ject to study and model the fa te and e ffe c ts of azinphosm ethyl in an apple orchard ecosystem . Funding was provided by th e Environm ental P ro tectio n Agency. The m ultidisciplinary n a tu re of this pro ject required cooperation among many research ers and support personnel. I would especially like to express my appreciation to L ester G eissel for his consultation and invaluable asistance in much of the field operations. I also acknowledge Dr. R obert Kon for his contribution in the adm inistration of sam ple analysis. In addition, I am indebted to Dr. Erik Goodman for his guidance in this research . Finally, I would like to acknowledge Dr. M atthew Zabik for his guidance in this research and my grad u ate studies. My professional developm ent while a graduate student is a d ire c t re su lt of his exam ple. TABLE OF CONTENTS Page LIST OF T A B L E S .............................................................................................................. LIST OF F I G U R E S .............................................................. INTRODUCTION .......................................................................................................... v v ii 1 PART I DEVELOPMENT AND PRESENTATION OF THE EXPERIMENTAL DATA BASE I n t r o d u c t i o n ................................................................................................. 9 A nalytical M e t h o d s .................................................................................... 9 The E xperim ental Site ........................................................................... 12 P reparation of the O rchard fo r R unoff C o lle c ti o n ........................... 13 Spatial S tru ctu re o f th e O r c h a r d ......................................................... 16 Spraying and Sampling Schedule ......................................................... 19 .............................................................. 23 R esults and Discussion PART II PARAMETERIZATION OF THE FIELD-BASED MODEL I n t r o d u c ti o n ................................................................................................. 59 The Model 59 ................................................................................................. Methods for P a ra m e ter E stim ation R esults and Discussion ................................................. 65 ........................................................................... 73 PART in ASSESSMENT OF AIRBORNE LOSS I n t r o d u c ti o n ................................................................................................. iii 93 A nalytical M e t h o d s ......................................................................................... 94 E xperim ental Site and P esticide A pplication Spatial S tructure of the O rchard O rchard Sampling Procedures .........................................95 ...............................................................96 ................................................................... 96 A irborne Residue S a m p l e s R esults and Discussion 97 .............................................................................. 102 PART IV FURTHER MODEL DEVELOPMENT I n t r o d u c t i o n .................................................................................................... 144C hem ical D e g r a d a tio n ...................................................................................147 M icrobial D e g r a d a tio n ...................................................................................155 P h o to d e g ra d a tio n ........................................................................................... 168 P lant U ptake Conclusions ................................................................................................171 ..................................................................... 171 PART V SUMMARY AND CONCLUSIONS Summary Conclusions ........................................................................................................ 173 .................................................................................................... 175 LIST OF R E F E R E N C E S ....................................................................................................177 iv LIST OF TABLES Page PART I 1 Summary o f Significant Events for th e O rchard, 1976 Season . . . . . 21 2 Summary of Significant Events for the O rchard, 1977 S e a s o n ...................22 3 A zinphosm ethyl Initial H orizontal D istribution from the Analysis of T argets, 1976 S e a s o n ................................................................................................. 24 4 A zinphosm ethyl Initial H orizontal D istribution from the Analysis of T argets, 1977 S e a s o n ................................................................................................. 25 5 Initial V ertical D istribution of Azinphosm ethyl Dislodgeable Residues (p g/cm + S E ) .............................................................................................................. 26 6 Estim ation of Initial V ertical Pesticide D istribution, as a Proportion of th e Amount Applied (+ SE), from the Analysis of S a m p l e s ....................... 28 7 Estim ation of Initial V ertical P esticide D istribution, as a Proportion of the Amount Applied, Using T arget D ata to C alculate this Proportion R eaching the O rchard F l o o r .............................................................. 31 8 A zinphosm ethyl in R unoff, 1977 S e a s o n ............................................................... 33 PART II 1 Comparison of the Model Predictions with Azinphosmethyl Residue D ata for th e 1978 S e a s o n ........................................................................................ 90 PART III 1 Summary of Significant Events for the O rchard, 1978 S e a s o n ......................100 2 A zinphosm ethyl Initial H orizontal D istribution from the Analysis of T argets, F irst Spray Period of th e 1978 S e a s o n ...............................................103 3 Initial V ertical D istribution of Azinphosm ethyl Dislodgeable R esidues, F irst Spray Period of th e 1978 Season ...........................................104 4 Estim ation of Initial V ertical P esticide D istribution, as a Proportion of th e Amount Applied (+ SE), from the Analysis of Samples for the F irst Spray Period of the 1978 S e a s o n ................................................................ 105 v 5 6 7 Estim ation of Initial V ertical P esticid e D istribution, as a Proportion of th e Amount Applied, using T arg et D ata to C alcu late the Proportion Reaching the O rchard Floor, fo r th e F irst Spray Period of the 1978 S e a s o n ...................................................................................................... 107 3 Azinphosm ethyl C oncentrations (u g /m ) a t Sampling Heights Between 0.5 and 6.0 M eters above th e O rchard F l o o r ................................ 109 E stim ated A zinphosm ethyl Daily A irborne Loss from th e O rchard, F irst Spray Period of th e 1978 S e a s o n ...............................................................127 PART IV 1 Azinphosm ethyl D egradation when Incubated w ith Various Soils • • i5 8 LIST OF FIGURES Page PA RTI 1 O rchard P l o t s ........................................................................................................ 15 2 Overhead View of H orizontal Regions of O rchard P l o t s .........................18 3 (a-d) A zinphosm ethyl Dislodgeable and Soil R esidues for th e Canopy Region, 1976 Season. Upward d ire c ted arrow s show d a tes o f spray application, and downward d ire c ted arrow s in d icate d ates and am ounts of rain fall in mm. Each bar is divided into tw o p arts; th e mean (u g /c m ) is above the line and its corresponding standard erro r (S.E.) below th e l i n e . ................................ 37 3 (e-g) A zinphosm ethyl Dislodgeable and Soil R esidues for th e Alley Region, 1976 Season. Upward d irected arrow s show d ates o f spray application, and downward d ire c ted arrow s in d icate d ates and am ounts of rain fall in mm. Each bar is divided into two parts; th e m ean (u g /c m ) is above th e line and its corresponding standard erro r (S.E.) below th e l i n e . ................................ 39 4 (a-d) A zinphosm ethyl Dislodgeable and Soil Residues for th e Canopy Region, 1977 Season. Upward d ire c ted arrow s show d ates of spray application, and downward d ire c ted arrow s in d icate d ates and am ounts of ra in fa ll in mm. Each b ar is divided into two p arts; the mean (u g /c m ) is above th e line and its corresponding standard erro r (S.E.) below th e lin e ......................................41 4 (e-g) A zinphosm ethyl Dislodgeable and Soil R esidues for th e Alley Region, 1977 Season. Upward d irected arrow s show d ates of spray application, and downward d irected arrow s indicate d ates and am ounts of rain fall in ram. Each bar is divided into tw o parts; th e mean (u g /c m ) is above the line and its corresponding standard erro r (S.E.) below th e lin e......................................43 5 Daily T em perature Range, 1976 S e a s o n ................................................51 6 Daily Minimum and Maximum R elativ e Hum idity, 1976 Season . . 7 Daily T em perature Range, 1977 S e a s o n ................................................55 8 Daily Minimum And Maximum R elativ e Humidity, 1977 Season . . vii 53 57 PART n 1 C onceptual Model for the D istribution, M ovement, and ....................... 62 A ttenuation of Azinphosm ethyl in an Apple O rchard 2 Flow C hart Showing D ata Processing to Yield Files Suitable for Use in Model P a ram eterizatio n ...............................................................67 3 Flow C hart of th e Process of C alculating th e P aram eters using th e "Processed F i l e s " .................................................................................... 69 ...................................................... 75 4 Azinphosm ethyl A ttenuation M atrix 5 Azinphosm ethyl N on-R ainfall M ovement M a t r i x ................................ 77 6 Azinphosm ethyl Light R ainfall Movement M a t r i x ................................ 79 7 A zinphosm ethyl Heavy R ainfall M ovement M atrix . . . ° . . . 81 8 (a-d) A ctual and P redicted Azinphosmethyl Dislodgeable and Soil Residues for th e Canopy Region, 1978 Season. Upward d irected arrow s show d a tes of spray application, and downward d irected arrow s indicate d ates and am ounts of rain fall in mm. Each bar is divided into two parts; the mean (p g /c m ground area) is above th e line and its corresponding standard erro r (S.E.) below th e line. The solid line running through the bars represents the model's prediction of the change in residues using 1976 and 1977 d a ta only, while th e dashed line shows the p red icted change in residues based on all th re e years d a ta ......................................................... 86 8 (e-g) A ctual and P redicted Azinphosm ethyl Dislodgeable and Soil Residues for th e Alley Region, 1978 Season. Upward d irected arrow s show d ates of spray application, and downward d irected arrow s indicate d ates and am ounts of rain fall in mm. Each bar is divided into tw o parts; the mean (u g /c m ground area) is above th e line and its corresponding standard erro r (S.E.) below th e line. The solid line running through the bars rep resen ts th e m odel's prediction of th e change in residues using 1976 and 1977 d a ta only, while the dashed line shows the predicted change in residues based on all th re e years d a ta ......................................................... 88 PART in 1 O rchard Plots Showing Air Sampling L o c a t i o n s ..................................... 97 2 Diurnal A zinphosm ethyl Airborne Residues a t Sampling H eights Between 0.5 and 6.0 M eters above the O rchard, M easured a t the C en ter Location During the F irst Spray Period of the 1978 S e a s o n ............................................................................................................ 113 3 Diurnal A zinphosm ethyl Airborne Residues a t Sampling Heights Between 0.5 and 6.0 M eters Above th e O rchard, M easured a t the Downwind Locations During the F irst Spray Period of the 1978 S e a s o n ....................................................................................................115 viii 4 (a-d) A zinphosm ethyl D islodgeable and Soil Residues for the Canopy Region, 1978 Season. Upward d irected arrow s show d ates o f spray application, and downward d irected arrow s indicate d ates and am ounts of ra in fa ll in rgm. Each bar is divided into tw o p arts; the m ean (y g /c m ) is above th e line and its corresponding standard e rro r (S.E.) below th e line................................. 117 4 (e-g) A zinphosm ethyl Dislodgeable and Soil R esidues for th e Alley Region, 1978 Season. Upward d irected arrow s show d ates o f spray application, and downward d irected arrow s indicate d ates and am ounts of ra in fa ll in ram. Each bar is divided into tw o p arts; the m ean (y g /c m ) is above the line and its corresponding stan d ard erro r (S.E.) below th e lin e................................. 119 5 D ecline in T otal A zinphosm ethyl Dislodgeable and Soil Residues M easured During th e F irst Spray Period of th e 1978 Season . . 136 D ecline in T otal A zinphosm ethyl Dislodgeable and Soil R esidues M easured During th e Second Spray Period of th e 1978 Season . . 138 6 7 Daily T em perature R ange, 1978 S e a s o n .................................................. 141 8 Daily Minimum and Maximum R elativ e Humidity, 1978 Season . 143 PART IV 1 A zinphosm ethyl Hydrolysis R ate C onstant Versus p H ........................151 2 A zinphosm ethyl Hydrolysis R ate C onstant Versus Air T em perature .................................................................................................. 154 3 A zinphosm ethyl Hydrolysis R a te C onstant Versus Soil T em perature .................................................................................................. 157 4 A zinphosm ethyl M icrobial D egradation Versus Soil P ercen t O rganic M a t t e r ..............................................................................................160 5 A zinphosm ethyl M icrobial D egradation Versus Soil T em perature 6 A m itrole D egradation Versus Soil M oisture for Two Soil Types ix 163 . 166 INTRODUCTION Since the discontinued use of many organochlorine insecticides in th e early seventies, use of the generally m ore toxic and less p ersisten t organophosphate insecticides has increased (Brown, 1978). Because of the g re a te r m am m alian tox icity of these compounds (M atsum ura, 1975), research of th eir fa te in orchards has been prim arily concerned w ith w orker reen try hazard associated w ith foliar dislodgeable residues, soil, and soil dust residues. Most of this work has been done for citru s (G unther e t al., 1977; Nigg e t al., 1977; Thompson and Brooks, 1976; Spear e t al., 1975), but studies have also been conducted for peach (W interlin e t al., 1975; Hansen e t al., 1978) and apple orchards (S taiff e t al., 1975; Hansen e t al., 1978). L ittle work has been done in the area of a tten u atio n and m ovem ent of the organophosphates betw een tre e s, ground cover, and soil, or loss with runoff from deciduous fru it orchards. The lack of cultivation, which is th e p ra c tice in many deciduous fru it growing areas (Haynes, 1980), preserves g re a te r species diversity than is found in many other agroecosystem s (Brown, 1978). This diversity can resu lt in a relativ ely stable deciduous tre e fru it ecosystem if undisturbed by pesticides or other m anagem ent p ra c tice s (Hoyt and Burts, 1974). n atu ral control of many arthropod parasitism is achieved (C roft, 1975). pests Under these conditions, through disease, predation, and However, to produce m arketable fru it, pesticides m ust be used to control a few key pests (Glass and Lienk, 1971). C ontrol of these key pests w ith pesticides may reduce n atu ral enemy populations, prom oting secondary pests to a major pest statu s, o ften requiring 1 2 additional chem ical control (Hoyt and Burts, 1974; C ro ft, 1978; Ware, 1980). In addition, th e increased use of pesticides in orchards, and on all crops, has resu lted in an increased number of re sista n t p est species (Smith, 1976; Brown, 1977, 1978; C roft, 1978). These and o th er problem s asso ciated with th e use of pesticides as a sole means of control have resulted in a g re a te r emphasis being placed on in teg rated pest m anagem ent as a long-term s tra te g y for th e econom ic c ontrol of crop pests (NAS, 1969; Sm ith, 1976). Such program s are designed to use a broad spectrum of control m easures including biological, cu ltu ral and chem ical methods. This approach to pest m anagem ent requires a g re a t deal m ore knowledge not only of the ecobiology of both beneficial and harm ful species, but also of pesticide fa te and e ffe c ts. The success of orchard IPM program s may depend largely on th e judicious use of pesticides allowing maximum b enefit to be gained from biological control m easures (C roft and Brown, 1975; Smith, 1976). This will require a b e tte r understanding of p esticid e fa te throughout the entire orchard ecosystem . A dditional consideration m ust be given to species o th er than those of b e n efit to pest control. Because of th eir re la tiv e perm anence, deciduous orchards are o ften a hab itat for w ildlife indigenous to n a tu ra l ecosystem s (C ro ft, 1978). D etailed studies of pesticide fa te and e ffe c ts, which would be im p ractical or undesirable in a natural settin g , can be done in an orchard, to give som e insight into the ecological hazards associated w ith pesticide exposure to the fauna of these environm ents. Such inform ation may be valuable in th e developm ent of environm ental fa te and ecological e ffe c ts te stin g guidelines as proposed under the Toxic Substance C ontrol A ct and th e recen tly am ended F ederal Insecticide, Fungicide and R odenticide A ct. The Toxic Substances C ontrol A ct authorizes th e Environm ental P ro tectio n Agency (EPA) to obtain from industry d a ta on th e production, use and e ffe c ts on 3 human health and the environm ent of chem ical substances and m ixtures. Testing standards under Section 4 a re d ire c ted a t specific chem ical c h a ra c te ristic s and e ffe c ts, including oncogenicity, terato g en icity , m utagenicity, and other health e ffe c ts, as well as environem ntal fa te , persistence and ecological e ffe c ts. All te sts are substantially sim ilar to m ethods proposed by EPA for its pesticide reg istratio n program and re p re se n t a "base set" of te sts from which to se le c t specific te sts for specific chem icals in te s t rules. To d ate no te s t standards have been proposed for environm ental fa te or ecological e ffe c ts. In addition to the Section 4 standards which are designed to te s t specific chem icals or chem ical groups, th e EPA also plans to issue te stin g guidelines consistent w ith these te stin g standards, but more general in n atu re, encom passing a wider range of chem ical substances and e ffe c ts of concern to EPA. These guidelines will also be used under Section 5 which allows the agency to lim it m anufacturing, processing, distribution, use or disposal of new chem ical substances (or new uses of existing substances), pending developm ent of inform ation. The a d m in istrato r is given this a u th o rity if he has reason to believe (1) th a t the chem ical substance may present an unreasonable risk or may resu lt in su b stantial human exposure or environm ental release, (2) th e re are insufficient d a ta or experience for determ ining or predicting health or environm ental e ffe c ts, and (3) te stin g is necessary to develop such d a ta. TSCA does not require EPA to provide testin g guidance, except when specifically requested under Section 5 (g). N evertheless, te stin g guidelines are presently being developed in an e ffo rt to stream line th e m onum ental task of TSCA com pliance and to encourage more appropriate and c o st-e ffic ie n t inform ation gathering. So fa r the EPA has proposed general standards for a num ber of human health e ffe c ts. Such standards for environem ntal fa te and ecological e ffe c ts are presently a t the interagency review stag e. To develop generic standards for 4 human health involves only one species, man. inhalation, uniform . ingestion, e tc.), m etabolism , Types of exposure (derm al, and pharm acodynam ics are fairly G eneric te s t standards may need only be modified based on the properties of the chem ical or chem ical categ o ry . G eneric standards for environm ental fa te and ecological e ffe c ts involves all species of flora and fauna. In addition, all bio tic-ab io tic relationships m ust be considered from ocean food chains to soil m icrobial com m unities. This diversity of species (and species interaction) and environm ents makes th e developm ent of generic standards to assess the possible hazards of chem ical substances to the environm ent a form idable task. The first problem would seem to be how many d ifferen t species m ust be te ste d to adequately assess ecosystem e ffe c t; secondly, w hat chem ical properties are im p o rtan t, and, thirdly, w hat and how do environm ental properties influence exposure and possible eco to x icity . Due to th e enormous com plexity, th e re has been much controversy over the ratio n ale for adopting standards in th ese areas where scien tific m ethods are less developed and where no validated techniques are available for c e rta in e ffe c ts. There may be unreasonable risk to an ecosystem when a com ponent or com ponents are exposed to a co ncen tratio n of a chem ical substance which causes harm . Environm ental fa te studies are used to p red ict and e stim ate the presence of potentially harm ful chem ical residues in m an-m ade and n atu ral environm ents. Upon release into the environm ent a chem ical may be m etabolized by living organism s, be transform ed by chem ical or photochem ical reaction, u n altered. or p ersist In some instances degradation or transform ation results in toxic products (Menzie, 1972; Crosby, 1973; Goring e t al., 1975). The goal of environm ent te s t standards should be to identify the dominant pathw ays of chem ical tran sfo rm atio n and tra n sp o rt, and then properties of the chem ical and environm ental re la te this behavior to conditions. This knowledge can 5 then be used to infer what biota may be exposed to the chem ical, th e degree, frequency, and route of exposure. Thus, both th e population a t risk and th e e ffe c t of the chem ical a re determ ined p artly by the d etailed fa te of the chem ical substance. Ju st as in health e ffe c ts te stin g , w here in v itro studies and lab anim als are used to estim ate chem ical e ffe c ts on man, ecosystem e ffe c ts and chem ical fa te are estim ated prim arily from lab o rato ry studies. The foundation for use of laboratory d a ta for environm ental fa te assessm ent is based on the assum ption th a t the accum ulates in the ra te at which a chem ical degrades, dissipates, or environm ent is the sum of the ra te s of known individual chem ical, physical, and biological processes, independently m easurable in th e laboratory. In addition, it is assum ed th a t the lab o rato ry d a ta for individual processes can be in te g ra te d and ex trap o lated to th e ap p ro p riate s e t of "real world" conditions. It is yet to be seen if any s e t of lab o rato ry te sts alone can be used to p redict the environm ental fa te of a chem ical substance to th e ex ten t th a t the EPA may adequately assess the possible risk to an ecosystem . The fa c t is th a t few of th e processes which d eterm in e the environm ental fa te of a chem ical substance have been studied in enough d e tail to p red ict dom inant pathw ays or ra te s of change in th e "real world." Each possible tran sfo rm atio n or tran sp o rt pathw ay can be studied in the lab o rato ry and ra te s may be determ ined under a given s e t of "environm ental conditions" but how a c tu a l environm ental conditions influence ra te s is largely unknown. A few years ago it was thought th a t m icrocosm s, controlled laboratory system s th a t a tte m p t to sim ulate some sele c ted portion of the re a l world, might be used to fill this inform ation gap. In a te rre s tria l m icrocosm environm ental p a ram eters such as te m p e ratu re, light level, w ater c o n ten t, and organism diversity can be controlled and varied by th e in v estig ato r. A common approach to m icrocosm developm ent has been to organize som e elem ents of a selected 6 ecosystem in a container so as to resem ble some aspects of th e ecosystem of in te re st and then sy stem atically to change the com plexity by adding new trophic levels of organism s. P aram eters of the system as they a ffe c t chem ical fa te are m onitored (G illette and W itt, 1979). U nfortunately, developm ent of th e physical models of these developm ent. chem ical, m icrocosm s has not always kept pace More im portantly, physical, biological, com plete and w ith conceptual understanding of the clim atological processes th a t d iscrete control chem ical fa te in a te rre s tria l system is not y et d irectly obtainable from these studies. A nother approach has been to use m ath em atical models. The developm ent and use of m ath em atical models usually involves a system s approach; the form alized analysis of any system or of the general pro p erties of system s. The analysis o f com plex system s as system s, and th e modeling of th ese system s, is co n trary to reductionist trends in science. Laboratory experim ents which isolate and control very sm all com ponents of n atu re have up to th e p resent been th e m ost pow erful investigative tools aiding man in understanding n atu re, but this "science by isolation" has its draw backs as th e larg er environm ent may influence the com ponents of in te re st to such an ex ten t th a t an isolated laboratory experim ent may exclude som e c ritic a l com ponents, or the behavior of a system may not be sim ply the sum of the behaviors of its p arts in isolation (Hall and Day, 1977). M athem atical modeling of environm ental fa te should be based on both reductionist laboratory experim ents and key observations of the fa te of chem icals in th e environm ent to include th e ra te s of tran sp o rt and tran sfo rm atio n as a function of clim atic and o th er environm ental variables. Only a handful of models have been developed which describe chem ical environm ental fa te in general. The present " sta te of the a rt" model is EXAMS, a model of fa te of toxic organic chem icals in aquatic ecosystem s, developed by the 7 EPA's A thens, G eorgia R esearch L aboratory (Lassiter e t al., 1978). This model was used to develop environm ental assessm ents for eleven chem icals in aquatic system s based on lab o rato ry m easurem ents. The model has since been modified to include a library of canonical environm ents and is now in use on a tria l basis in sev eral lab o rato ries and EPA offices. The work presented here is one portion of an e ffo rt to ch aracterize the dynamics and e ffe c ts of an exam ple compound in the te rre s tria l environm ent, utilizing prim arily field m easurem ents and th e methodology of system s modeling and sim ulation. D ata collection, model refin em en t, and revised experim ental design w ere done ite ra tiv e ly , yielding a model which is param eterizable and d ata which are re le v an t to the problem being a tta ck e d . The study of pesticide dynam ics through in situ field studies is d ifficu lt due to the lack of n a tu ra l or planned experim ents (inability to control much of th e variance, i.e., clim atic conditions) and the relativ ely high levels of erro r associated w ith field d a ta . Modeling techniques w ere employed to aid in the understanding of the necessarily large am ount of field d ata needed to co n stru ct a "m eaningful" picture of the pesticide's fa te . The field experim ental program used to investigate the distribution, a tten u atio n and m ovem ent of th e organophosphate insecticide azinphosm ethyl, 0 ,0-dim ethyl-5-(4-oxo-l, 2,3, benzotriazin-3(4H )-ylm ethyl) phosphorodithioate (Guthion ), in a Michigan apple orchard is given in P a rt I. The compound was T> followed from its spray application through the orchard v e g e ta tio n /litte r/so il environm ent and into aq u atic system s. The form of th e model describing azinphosm ethyl m ovem ent and a tte n u atio n , as well as d ata handling procedures and the derived ra te s , are presented in P art II. A companion study was conducted concurrently w ithin the sam e orchard to examine the e ffe c ts of azinphosm ethyl on several ground-dwelling in v erteb rates, including detailed 8 studies of the isopod Tracheolipus rathkei (Snider, 1979; Snider and Shaddy, 1980). Field and laboratory d a ta collected on T. rath k ei w ere used to develop a model describing its ecobiology and tem porally d istrib u ted m o rtality (Goodman e t al., 1981). The output of the fa te model described in P a rt n was used to d eterm ine the tim e-course of azinphosm ethyl exposure. In P a rt III the field experim ental program used to determ ine azinphosm ethyl airborne residues is presented. A m ulti-com ponent k in etic model used in the assessm ent of the contribution of airborne loss to th e overall a tte n u atio n of deposit residues is also described. In P a rt IV degradative losses of azinphosm ethyl are exam ined as a function o f environm ental conditions. PART I ASSESSMENT OF THE ATTENUATION AND MOVEMENT OF AZINPHOSMETHYL IN A MICHIGAN ORCHARD ECOSYSTEM: DEVELOPMENT AND PRESENTATION OF THE EXPERIMENTAL DATA BASE INTRODUCTION R eported here is th e field experim ental program used to in v estig ate th e distribution, a tte n u atio n , and m ovem ent of the organophosphate insecticide azinphosm ethyl, 0,0-dim ethyl-5-(4-oxo-l,2,3, benzotriazin-3(4H )-ylm ethyl) phosphorodithioate (Guthion ), in a Michigan apple orchard. Studies were carried out to g a th er d a ta on both in itia l distribution of azinphosm ethyl w ithin th e orchard and v e rtic a l m ovem ent of the pesticide under th e influence of rain fall, as well as th e a tte n u atio n of the pesticide in various situations. In addition, plots w ere designed such th a t each was a sep a ra te w atershed, allowing runoff collection from them individually. The compound was followed from its spray application through th e orchard v e g e ta tio n /litte r/so il environm ent and into aquatic system s, to determ ine possible exposure of th e orchard ecosystem biota to th e compound. The d a ta presented here w ere used to p a ram eterize a model (presented in P a rt II) for azinphosm ethyl a tte n u atio n and m ovem ent in an orchard ecosystem . ANALYTICAL METHODS O rganic solvents w ere glass-distilled and th e w ater was distilled-deionized. A fte r washing, the glassw are was rinsed w ith aceto n e and hexane followed by overnight heating a t 250°C. 9 10 D islodgeable R esidues. Dislodgeable residues w ere determ ined by th e procedure of G unther e t al. (1973). L eaf discs w ere e x tra c te d with 50 ml portions of w ater containing tw o drops of Triton X-100 in w ater (1:50, v/v). Samples were then a g ita te d on a w rist-action shaker for 15 m inutes. Depending on th e bulkiness of th e grass, litte r , or moss, m ore w ater was used to insure good co n tact w ith the sam ple during e x tra ctio n . One drop of Triton-X 100 solution was added for each additional 25 ml of w ater used. All sam ples w ere given a final rinse by hand w ith a third 50 ml portion of w ater. The combined washes were partitioned th re e tim es against 50 ml portions of hexane which was then co n cen trated on a ro tary evap o rato r and diluted to a proper volume for analysis. Using the above m ethod, recovery and standard deviation from the Triton-X 100 w ater solution fo rtified w ith azinphosm ethyl standard was 95 + 3%. S u rfa ce -P e n e tra te d R esidues. To e stim a te the magnitude of the rem aining residues on or w ithin the cu ticu lar m atrix, sele c ted sam ples co llected on the day of application w ere given an additional e x tractio n in a m anner sim ilar to th a t described by S teffens and Weineke (1976). Three 20 ml volumes of aceto n e w ere successively sw irled w ith th e sam ples ju st a fte r the dislodgeable residue e x tra ctio n had been com pleted. The combined aeeto n e-w ater e x tra c ts w ere c o n ce n tra te d by ro ta ry evaporation to rem ove the acetone. The w ater le ft in th e ro ta ry flask was p artitioned in th e flask against th ree 30 ml volumes of hexane which w ere combined and then dried over anhydrous sodium su lfate and c o n ce n tra te d to about one ml. This procedure was selected as Wieneke and S teffens (1974) w ere able to recover 100.4% of the 14 C azinphosm ethyl applied to bean leaves as an aqueous form ulation, one day following application. This e x tra c t was cleaned up by the silica gel microcolumn described by Kadoum (1967) w ith our elution system . One and a half gram s of u n activ ated Adsorbosil CAB, 100/140 mesh (Applied Sciences, Inc.) was packed as a hexane 11 slurry into a P asteur pipet. The sam ple was added in hexane and th e column eluted w ith 15 ml of 10% benzene in hexane. Ethyl a c e ta te (4%) in benzene then elu ted the azinphosm ethyl in 15 ml. R ecovery and stan d ard deviation of 2 ml of a 5 ppm azinphosm ethyl standard solution added to th e column was 94 + 3%. Soil. E xtraction of soil was based on Schulz e t al. (1970). A pproxim ately 50 g of soil (largest roots rem oved) w ere mixed w ith 10 ml of w ater, if dry, and aceto n e (1.5 m l/g) on a w rist-action shaker for 15 m inutes. The solvent was poured into a ro ta ry flask through a funnel lined w ith W hatman No. 1 filte r paper; as much soil as possible was retain ed in the e x tra ctio n flask. A second ex tractio n followed and the combined e x tra c ts w ere tre a te d e x actly as th e su rfa ce -p e n e tra ted residues of the previous section, including the clean-up column. R ecovery and standard deviation from soil fo rtifie d w ith azinphosm ethyl stan d ard was 84 + 9%. F ilter Paper T arg ets. T argets, consisting of tw o circles of W hatman No. 1 paper, 18.5 cm in d iam eter (attach ed one ato p th e o th e r, by a single stap le to a cardboard backing), w ere e x tra c te d for two hours in a soxhlet e x tra c to r using a m ixture of hexane:acetone, 100 ml:25 ml. R ecovery and stan d ard deviation from ta rg e ts fo rtified with azinphosm ethyl stan d ard was 104 + 5%. P esticide R unoff Studies. A zinphosm ethyl in runoff sam ples co llected during th e 1977 season was exam ined by e x tra c tin g one lite r subsam ples following centrifugation to rem ove sedim ent. The sedim ent was soxhlet e x tra c te d for four hours with a 1:1 hexaneiacetone m ixture, a ir dried and weighed. The su p ern atan t was e x tra c te d with hexane and both sedim ent and su p ern atan t e x tra c ts w ere c o n cen trated by ro ta ry evaporation. R ecovery and stan d ard deviation from runoff w ater fo rtifie d with azinphosm ethyl stan d ard was 93 + 4%. Q uantitation of azinphosm ethyl was accom plished using a Tracor 560 gas chrom atograph having a flam e photo m etric d e te c to r in th e phosphorus mode. A six-foot glass column (2 mm i.d.) packed w ith 3% SE-30 on Gas Chrom Q, 60/80 12 mesh, was op erated a t 195 C; N2, 40 m l/m in; a ir, 90 m l/m in; H2, 60 m l/m in. This system was in te rfa c ed with a dig ital PDP 8 Pam ila PDP 11/40 RSTS com puter for in tegration of the area under the single peak produced. THE EXPERIMENTAL SITE The experim ental apple orchard was lo cated in the vicinity of Grand Rapids, Michigan in K ent C ounty. The tw elv e-y ear old tre e s w ere a m ixture of sem i-dw arf duchess and w ealthy cultivars. u n atten d ed for approxim ately five years. In 1976 th e orchard had been le ft This was a necessary req u irem en t to assure adequate populations of ground-dwelling in v e rte b ra tes, u n a ffe cted by previous seasons' pesticide applications (Snider, 1979). In th e early spring th e tre e s w ere pruned and understory brush was rem oved. J u st prior to the first spray application the a re a was mowed to a height of approxim ately five cm . Due to dry conditions fu rth er mowing was not necessary. In 1977 light pruning was re p e ate d and th e orchard was mowed just prior to th e first, second, and fourth spray applications. A survey of the ground cover by th e lin e -tra n se c t m ethod (Cox, 1974) was carried out in 1976; cover type categ o ries w ere established and th e ir im portance values calcu lated . Above the soil (a m a rle tte sandy clay loam , pH 6.0, 56% sand, 20% silt, 24% clay, and 6.0% o.m . in to p 10 cm) was moss interm ingled w ith litte r; the m ost re c en t litte r deposits covered the moss. The herbaceous grow th covering th e moss and litte r was predom inantly broad-leaved w eeds, w ith grass being of secondary im portance. An exam ination of the ground cover in the areas sam pled for residue analysis yielded sim ilar resu lts. This ground cover com position is not atypical of orchards in the te m p e ra te E astern U nited S tates and C anada, although they vary with th e level of secondary succession (W hittaker, 1975), clim atic, topographic, edaphic conditions, and m anagem ent p ra c tice s (Klingman and Ashton, 1975; Teskey and Shoem aker, 1978; Schubert, 1976). 13 PREPARATION OF ORCHARD FOR RUNOFF COLLECTION To fa c ilita te th e study of azinphosm ethyl loss via runoff, four experim ental plots and one control plot w ere established, each a sep arate w atershed (Figure 1). Slopes in plots 1 and 2 and the control w ere betw een 6 and 12%, slopes in plots 3 and 4 w ere betw een 18 and 25%. All the runoff was collected from a given plot for each event and subsam pled for pesticide analysis. A 600-liter galvanized s te e l livestock tank (61 cm x 61 cm x 183 cm) was placed in an excavation a t the n a tu ra l drainage point for each plot, a t a level to co llect th e runoff. A sm aller tank (40 lite r capacity) was placed inside the larger tank so th a t it would fill and any overflow would be contained in th e larg er tank. Also, an overflow diversion pipe was installed on th e collection tank for plot 1; in the event th e large tank overflow ed, the runoff w ater would be diverted to a point below plot 4 (see Figure 1). R unoff w ater was finally channeled into the tanks by galvanized s te e l collection chutes, 61 cm long with sides 15 cm high, 61 cm wide a t the top and 30.5 cm wide a t the bottom . A 61 cm by 61 cm area of the tank d irectly below the chute was covered with a 6.4 mm mesh screen and th e re s t of th e tank was covered w ith plywood covered w ith plastic, to keep out rain fall, anim als, and debris. The chute and screen w ere also covered with p lastic. Both th e tanks and chutes w ere co ated w ith epoxy paint, as te sts showed th a t a fte r seven days, th e residues rem aining from a five ppb solution of azinphosm ethyl in galvanized s te e l containers w ere 30% less than the residues rem aining in glass, stainless ste e l, or epoxy-painted containers. Azinphosmethyl hydrolysis experim ents (see P a rt IV) indicated little degradation over the seven-day period, a t solution pH values below 7.0 (first order ra te constant of .008 a t 25°C). As th e pH of the runoff w ater collected averaged 5.9 + 0.2, losses by this m echanism prior to sam ple collection and ex tractio n w ere thought to be sm all. Loss via volatilization was also considered to be negligible as azinphosm ethyl has Figure 1 O rchard Plots z -» 15 CONTROL (.158 ha) A PLOT 3 (.052 ha) PLOT 4 (.052 ha) a 8M PLOT 1 (.099 ha) PLOT 2 (.060 ha) RUNOFF COLLECTION POINTS 16 a vapor pressure < 7.5 x 10-6 mm Hg a t 20°C (Mobay Chem ical Co., personal com m unication) and the maximum co n cen tratio n of azinphosm ethyl m easured in the runoff co llected was 22 ppb, well below its w ater solubility of 33 ppm. Plots w ere enclosed and sep arated from one another by a 30.5 cm, 20 gauge aluminum core fence buried approxim ately five cm in the ground. This fence was made continuous with th e collection chute to aid in the final channeling of runoff a t the bottom of the plot slope. Fencing was installed w ith minimal disturbance of the ground cover and in such a m anner th a t p o ten tial runoff would have as little co n tac t as possible w ith the fencing. This collection schem e was designed and im plem ented during th e 1976 season, but due to dry conditions, no runoff was collected in th a t year. SPATIAL STRUCTURE OF THE ORCHARD PLOTS The orchard plots (Figure 1) w ere subdivided both v ertically and horizontally in order to rep resen t adequately the variations which a ffe c t the behavior and im pact of th e pesticide in th e orchard ecosystem . H orizontally, each plot was subdivided into four regions potentially d ifferen t w ith resp ect to in itia l pesticide distribution (see Figure 2). V ertically, each alley region was divided into th re e com partm ents: grass-broadleaves, litte r moss, and soil. The canopy regions contained these com partm ents plus a c om partm ent for the leaves of the tre e . Region 4 (under the canopy) was determ ined to be 32% of the a rea in plot 1, 34% in plot 2, 54% in plot 3, and 45% in plot 4. The rem ainder of th e plot a re a was divided among the th re e alley regions in a ra tio of 2:3:1 for plots 1 and 2, and 1:2:1 for plots 3 and 4. The d ifferen ce in th e ratio s was due to th e closer spacing of the rows along th e N orth-South axis in plots 3 and 4. The num ber of tre e s in each of the plots was as follows: plot 1, 24; plot 2, 15; plot 3, 10; and plot 4, 11. T ree heights averaged 3.0 m eters. 17 Figure 2 Overhead View of H orizontal Regions of O rchard Plots CANOPY REGION (4 ALLEY REGION (2) CANOPY REGION (4 ALLEY REGION (1) ALLEY REGION (3) ALLEY REGION (1) CANOPY REGION (4 ALLEY REGION (2) CANOPY REGION (4 19 SPRAYING AND SAMPLING SCHEDULE The size of the plots and th e ir inaccessibility due to th e aluminum fences made the use of com m ercial spray equipm ent impossible. What was needed was a plot sprayer th a t could: 1) be pulled by a sm all tra c to r, such as a garden tra c to r, and 2) m atch as closely as possible the coverage of a com m ercial spray unit. To m eet these c rite ria , a low pressure, low volume sprayer sim ilar to one developed by H ow itt and Pshea (1965) was constructed. The sprayer was m ounted on a tra ile r and was pulled by a John D eere model 200 garden tra c to r. M ounted on the tra ile r was a double in let, high volume, low velocity blower m anufactured by Dayton, model 3C011. horsepow er Briggs and S tra tto n The blower was pow ered by an eight gasoline engine. Blower output was approxim ately 10,000 cubic fe e t per m inute (at th e highest engine speed settin g ). The air flow was d irected by a fiberglass d eflecto r, co n stru cted for use on Agtech crop sprayers. One side of th e d e fle cto r was widened to accom m odate th re e Beecom ist model 350 mini-spin spray heads. The d eflecto r was fu rth er m odified betw een 1976 and 1977 seasons to d ire c t more of the spray into th e tre e s. Pesticide was delivered to th e spray heads by a M asterflex variable speed tubing pump. The pesticide application ra te was determ ined by the ra te a t which the pesticide m ixture was delivered to th e spray heads, tra c to r speed, and distance covered to spray each plot. The a ctu al application tim e was recorded in th e field for each plot and the a c tu a l am ount of pesticide applied was d eterm ined by tank residual volume a fte r spraying. In addition, during the first spray period of 1978, sam ples of spray m ixture w ere taken, a t th e spray heads, before and a fte r spraying each plot. R esults from this m ethod were com parable to those of the residual volume m ethod. The average application ra te and stan d ard deviation was 1.62 + .33 kg ha-1 50% w.p. lOQJl-1 ha-1 in 1976 and 1.64 + .38 kg ha-1 50% 20 w.p. 1002, * ha 1 in 1977. Plots w ere sprayed in sequence betw een 7 a.m . and 9 a.m . D eposit Residue Samples To determ ine initial horizontal distribution of azinphosm ethyl, a series of 15 to 20 ground-located filte r paper ta rg e ts was deployed among th e four horizontal regions (see Figure 2) in each plot on the d ate of spraying. T argets w ere placed in the cen te r of the alley regions, and a t th e in-row midpoint betw een the tre e trunk and the edge of th e canopy region. As soon as a plot had been sprayed, the filte r papers w ere rem oved from the backing, packaged, and tran sp o rted in picnic coolers cooled by dry ice. Storage was a t -20°C until analysis. The initial distribution of azinphosm ethyl was also determ ined among the th re e ground layers: grass-broadleaves, litte r-m o ss, and soil. In each plot, tw o sam ples of each layer were taken from the tre e and alley regions. A square of sod (15 cm x 15 cm) was cut to a depth of about four cm and th e grass-broadleaf plants, litte r , and moss w ere sep arated into th re e d iscrete sam ples for residue analysis. From the bare sod rem aining, th re e cores (5 cm x 7 cm deep) w ere taken for analysis, providing a sam ple of 40 to 60 g. For analysis of le af deposits a 2.0 cm d iam eter disc was taken from the c en te r of each le a f. discs per tre e . Three tre e s w ere sam pled in each plot, taking 30 or more Samples w ere taken a t shoulder height (distributed within the canopy easily reached) around the e n tire circum ference of the tre e . Sampling d ates for the 1976 season a re shown in Table 1. Infrequent rain fall during this season yielded little inform ation usable to assess m ovem ent due to rainfall. In 1977, interspray sam pling was optim ized for m ovem ent d ata as sam ples w ere collected prim arily a fte r rain fall events (see Table 2). 21 Table 1. Spray D ate Summary of Significant Events for the O rchard, 1976 Season Sample D ate R ainfall D ate R ainfall Amount mm in 36.3 1.40 Days Since Spray O rchard Mowed 28 June 30 June 2 July 2 6 9 14 July July July July 0 4 7 12 15 July 15 July 20 July 0 5 20 July o 22.9 .90 06.4 0.25 22 July 7 26 July 12 27 July 28 July 28 Julyb 11.4 0.45 31 July 04.6 .18 29 July 5 7 2 August 4 August 5 August 10.2 .40 14 August 27.9 1.10 13 10 August 17 August 16 August 19 17 August 20 August 24 August 0 3 7 28 August 30 August O" SB 0 1 R ainfall s ta rte d im m ediately a fte r spraying. R ainfall occurred a fte r sampling. 06.4 0.25 13 22 Table 2. Spray D ate 26 May Summary of Significant Events for the O rchard, 1977 Season Sample D ate R ainfall D ate R ainfall Amount mm m Orchard Mowed 23 May 5 .1 25 May 0.20 30 May 20.3 0.80 4 June 25.4 1.00 12.7 0.50 26 May 1 June 6 June 11 June and 12 June 13 June Days Since Spray 0 4 6 10 11 17 18 O rchard Mowed 15 June 16 June 0 16 June 17 June and 18 June 6 .4 0.25 27 June 5.1 0.20 30 June 25.4 1.00 4 July 7 July 12.7 5.1 0.50 0.20 18 July 22.9 0.90 20 June 28 June 1 July 7 July 7 July 19 July 2 4 11 12 14 15 20 22 0 11 12 O rchard Mowed 20 July 22 July 22 July 24 July 2.5 0.10 29 July 2 .5 -.1 0 3 August 4 August 6 August 2 .8 13.7 3 .8 0.11 0.54 0.15 8 August 10 August 2.5 22.9 0.10 .90 13 August 12.7 0.50 25 July 1 August 7 August 12 August 12 August 15 August 18 August a R ainfall occurred ju st prior to spraying. 0 2 3 6 10 11 13 15 16 17 0 1 3 6 23 R esidues in R unoff A zinphosm ethyl m ovem ent out of the orchard w ith runoff was also determ ined. R unoff w ater was rem oved from th e orchard in fo u r-lite r, dark glass b o ttle s and stored a t 5°C until analysis. Excess runoff was m easured volum etrically in the field and discarded. R ainfall was m easured with s ta tic rain gauges and this inform ation was c o rre la te d w ith d a ta from recording rain gauges m aintained 3/4 mile from th e o rch ard by the Michigan S ta te U niversity In stitu te of W ater R esearch. T em p eratu re and hum idity d a ta shown in Figures 5, 6, 7 and 8 w ere obtained from the Michigan W eather Service, Grand Rapids, M ichigan. RESULTS AND DISCUSSION Initial H orizontal D istribution Tables 3 and 4 show the results of the analysis of the filte r paper ta rg e ts for th e first spray day of th e 1976 and 1977 seasons. The average am ount and 2 stan d ard erro r (SE) of pesticide are expressed in u g /c m ground a re a and proportions shown are proportions of dose applied. The grand average across plots 1 through 4 for both seasons shows th a t alley region 2 and the canopy region receiv e th e m ajority of th e residues reaching the ground; alley regions 1 and 3 received less. P esticide distribution was m ore uniform in the 1977 season, and residue levels w ere proportionally lower across all regions. believes The author this was due to m odification of the sprayer betw een seasons, a fte r which m ore of th e spray was d irected into th e atm osphere and aw ay from th e ground. Initial V ertical D istribution A zinphosm ethyl in itial v e rtic a l distribution for the 1976 and 1977 seasons is given in Table 5. Each plot is rep resen ted by only two regions; alley and canopy. In 1976 the alley was random ly sam pled. In 1977 the average am ounts for the alley rep resen t w eighted averages of alley regions 1, 2, and 3, w eighted Table 3. Azinphosmethyl Initial H orizontal D istribution from th e Analysis of T argets, 1976 Season Plot 1 2 3 4 Alley (1) Alley (2) Alley (3) Canopy (4) 2.48 + 0.22 3.35 + 0.45 2.76 + 0.34 2.89 + 0.51 Sample size 10 14 7 13 Proportiona .071 .131 .026 .093 1.36 +0.31 2.46 + 0 .4 4 1.65 + 0.39 3.20 + 0.37 Sample Size 9 6 9 11 Proportion8 .050 .079 .026 .136 .75 + 0.20 3.81 + 1.00 1.36 + 0.50 5.10 + 0.77 Sample Size 5 5 5 7 Proportion8 .009 .076 .023 .226 2.51 +0.95 3.06 + 1.28 2.14 + 0.65 2.90 + 1.10 Sample Size 9 4 3 10 Proportion8 .042 .111 .065 .175 1.78 + 0.44 3.17 +0.28 1.98 +0.31 3.52 + 0.53 Average Amount (p g/cm Average Amount (p g/cm Average Amount (p g/cm Average Amount (p g/cm 2 2 2 2 + SE) + SE) + SE) + SE) cr & Grand Average Amount (p g/cm'* + SE) Proportion of am ount applied Standard error b Table 4. Azinphosmethyl Initial H orizontal D istribution from the Analysis of T argets, 1977 Season Alley (1) Alley (2) Alley (3) Canopy (4) 2.75 + 0.63 3.12 + 0.77 2.63 + 0.50 3.55 + 1.14 Sample size 7 7 13 6 Proportiona .070 .114 .023 .095 1.34 + 0.24 1.96 + 0.12 1.24 + 0.26 1.83 + 0.44 9 3 6 4 .047 .109 .026 .103 1.03 + 0.22 2.51 + 0.13 1.34 + 0.28 1.39 + 0.45 7 5 8 6 .105 .064 .020 .088 .91 + 0.43 2.27 + 0.43 .978 + 0.28 1.43 + 0.09 Sample Size 3 5 8 3 Proportion® .017 .079 .024 .093 1.51 + 0.42 2.47 + 0.25 1.55 + 0.37 2.05 + 0.51 Plot 1 2 A verage Amount (p g /c m 2 + SE)b Average Amount (p g/em o + SE) Sample Size Proportion 3 Average Amount (p g/cm o + SE) Sample Size Proportion® 4 Average Amount (p g/cm 2 + SE) Grand Average Amount (p g /c m 2 + SE) Proportion of am ount applied Standard error 26 Table 5. Initial V ertical D istribution of A zinphosm ethyl Dislodgeable Residues (y g /c m + SE) T ree Y ear 1976 G rass-B roadleaves L itter-M oss Soil 1.49 + .40 .18 + .04 .20 + .05 9 9 8 1.72 + .33 .15 + .08 .19 + .0 2 8 5 8 2.88 + .42 .55 + .04 .29 + .04 .29 12 6 8 1 1.15 + .23 .27 + .05 8 8 Canopy Average Amount 2.92 +.41 Sample Size 11 Alley A verage Amount Sample Size 1977 Canopy A verage Amount Sample Size Alley Average Amount Sample Size o Soil residues w ere determ ined on a whole basis, value rep resen ts y g/cm top 10 cm. 2 in th e 27 by p ercen t of plot area. These d ata indicate th a t for both years the m ajority of th e dislodgeable pesticide residues deposited in th e orchard was in itially d istributed vertically to the tre e and the grass-broadleaf layer. A zinphosm ethyl residues distributed to th e litte r-m o ss layer and soil w ere roughly te n tim es low er than le a f residues. Using the v ertic al distribution of the dislodgeable residue d ata along w ith th e pesticide application ra te s and the plot c h arac te ristic s, th e proportion of th e applied pesticide reaching each of the v e rtic al regions was calcu lated on a plot basis and then averaged across plots (see Table 6). To d eterm ine the am ount o f p esticide distributed to the tre e leaves, the tre e leaf surface area was e stim ated by rem oving all th e leaves from two tre e s, one in plot 2 and one in plot 4, following the 1978 sam pling season (an exhaustive search for other methods of determ ining tre e leaf surface was unsuccessful). The leaves from each tre e w ere contained in six 50 1 bags. Each bag was weighed and sub sam ples of ten leaves w ere random ly selected for each 500 g of leaves. d eterm ined for each sub sam ple. Weights and areas w ere T otal le af area was determ ined from th ese m easurem ents and the to ta l le a f w eight. The average for these tw o tre e s was approxim ately 400,000 cm (one side of le af only). This is a crude e stim ate due to sam pling lim itations and is specific for th e tre e size and vigor found in this orchard. As the pesticide was not evenly distributed v ertically in th e tre e , e stim a te s based on le af residue d ata taken from the lower half of the tre e in th e 1976 and 1977 seasons would tend to o v erestim ate the proportion distributed to th e e n tire tre e . In 1978, th e upper half of th e tre e was sam pled sep a ra te ly . From these d ata, a ra tio was determ ined by dividing the average am ount of dislodgeable residue found in the e n tire tre e (upper and lower) by th e average am ount of dislodgeable residue found in the low er half of the tre e . This ra tio was determ ined using an aeross-plot average for each of the sam ple d ates for th e 28 Table 6. Estim ation of In itial V ertical P esticid e D istribution as a ] of the Amount Applied (+ SE), from th e Analysis of Samples 1976 Alley Dislodgeable S urfaceP e n e tra te d Canopy SurfaceDislodgeable P e n e tra ted Tree*3 .376 + .082 .027 Grass .112 + .015 .021 .060 + .014 .011 L itter-M oss .014 + .005 .006 .013 + .004 .003 Soil0 .015 + .0 0 2 T otal .661 + .093 .010 + .006 1977 Tree .337 + .020 .022 Grass .070 + .016 .013 .033 + .009 .005 Litter-M oss .018 + .002 .007 .014 + .004 .005 Soil0 .015d .015 OS jQ V T3 D ata averaged across plots for th e firs t application of th e season. The proportion distributed to the tre e includes both leaves and bark. Soils residues w ere determ ined on a whole sam ple basis. 1976 value used as th ere are in su fficien t d a ta fo r 1977. .556 +.011 29 first spray period of the 1978 season. A least squares linear regression of these ratio s w ith tim e gave an in te rc e p t of .7027 and a slope of -.0063T, T being th e tim e since spray in days (r = .836). The value for th e in te rc ep t was then m ultiplied by the average lower le a f values for th e 1976 and 1977 seasons to arrive a t an estim ate of the average le a f residue level for the e n tire tre e (data in Table 6 and Figures 3 and 4 have been adjusted accordingly). Due to pre-1977 sprayer design the 1976 d a ta are probably s till slightly o v erestim ated . This 2 e stim a te was then m ultiplied by 400,000 cm /tr e e and by th e num ber of tre e s per plot. The resu lt was divided by th e dose applied to d eterm ine th e proportion of pesticide distributed to th e leaves. The proportion of p esticid e distributed to th e tre e included residues deposited on both leaves and bark. The proportion distributed to the bark was estim ate d a t 15% of th a t going to leaves, based on th e work of Steiner (1969). The proportion of the pesticide applied th a t is distributed as su rfacep e n e tra te d residues was estim ate d from m atched sam ples from th e first spray of th e 1978 season. The su rfa c e -p e n e tra te d to dislodgeable residue ratio s and standard errors w ere as follows; leaves, .081 + .015; grass, .190 + .002; litte r, .421 + .018; moss, .272 (one sam ple). These ratio s w ere m ultiplied by the appropriate dislodgeable residue proportions to e stim a te th e su rfa ce -p e n e tra ted residue proportion. All soil residues w ere determ ined on a whole sample basis. To e stim a te the am ount of pesticide deposited in th e orchard during spray application, all the dislodgeable, su rfa c e -p e n e tra te d and soil residue proportions w ere sum m ed across both th e alley and canopy regions. In 1976, 66.8% (SE .093) o f the pesticide applied was e stim ated to be initially deposited to the various orchard layers. Confidence in this conservative e stim a te is subject to the following sources of error; accuracy of am ount applied, hom ogeneity of application, rep resen tativ e sam pling of residue distribution, and the tre e surface 30 a re a e stim ate . The 33.2% not accounted for is assumed to be due to airborne loss, prim arily as d rift a t application, but also as v olatilization and wind erosion in the four hours betw een application and sampling. A nother possible pathw ay of residues not accounted for is foliar pen etratio n and subsequent unavailability to su rfa ce strip e x tractio n . Numerous studies have com pared dislodgeable to w hole-leaf tissue residues (W interlin e t al., 1975; Elliot e t al., 1977; G unther e t al., 1977). Their applicability to the present study is u n certain , as th e degree of pesticide le af uptake may vary with p esticide, form ulation, mode of application, am ount applied, le af type, and environm ental conditions (Hull, 1970). The efficiency of the solvent strip procedure used h ere, as com pared to w hole-leaf ex tra ctio n procedures, in rem oving freshly deposited azinphosm ethyl residues is not known. Use of the solvent strip procedure was based on the resu lts of a detailed study conducted by Weineke and S teffens (1974). found th a t 100.4% of th e 14 These research ers C azinphosm ethyl applied to bean leaves in an aqueous form ulation could be recovered w ith a w ater followed by a benzene strip , for sam ples taken one day following application. In 1977 only 55.4% (SE .001) was estim ate d to be deposited in the orchard w ith 45.6% as airborne loss. A sim ilar pesticide mass balance shown in Table 7 uses th e ta rg e t d a ta to e stim a te the proportion of the pesticide applied th a t reaches the orchard floor. Tree proportions w ere e stim ated in the sam e m anner as fo r Table 6. This tre a tm e n t of the d a ta shows th a t 74.1% (SE .099) of pesticide applied in 1976 was deposited in the orchard, on the average, w ith 25.9% as airborne loss. In 1977 61.0% (SE .034) was e stim ated to be deposited in the orchard w ith 39.0% as airborne loss. Analysis of th e plot to tals (Table 7) using ta rg e ts to e stim a te the proportion of the applied dose distributed to the orchard floor suggests th a t the am ount rem aining in the orchard following application is g re a te r in 1976 than in 1977 (paired t- te s t of plot to ta l proportions gave a P < .15). A sim ilar analysis 31 Table 7. Estim ation of Initial V ertical P esticide D istribution, as a Proportion of th e Amount Applied, using T arget D ata to C alculate Proportion R eaching the O rchard Floor 1976 O rchard Floor Dislodgeable Tree_____ _ _ S urfaceP e n e tra ted Plot Alley Canopy T otal 1 .228 .093 .428 .030 .780 2 .155 .138 .374 .026 .694 3 .156 .181 .155 .011 .506 4 .218 .178 .547 .039 .985 .741 +.099 _____________________ 1977_______________________ 1 .207 .096 .291 .021 .615 2 .182 .105 .387 .028 .702 3 .088 .101 .337 .024 .550 4 .120 .094 .334 .024 .572 .610 + .034 32 indicated th a t this d ifferen ce was not due to am ounts distributed to the tre e s but ra th e r th a t proportion distributed to th e orchard floor. The author believes th a t this difference may be due to th e betw een-season sprayer m odification th a t d irected more of th e spray upward into th e atm osphere and away from the ground. Wind speed and atm ospheric stab iity during application may also have contributed to this d ifferen ce. There is also som e evidence th a t ta rg e t residues as a proportion of dose applied w ere g re a te r than the combined residues distributed to the various orchard floor layers (paired t- te s t of plot averages across both years gave a P < .15). As these d a ta w ere norm alized for ex tractio n efficiency, this resu lt would indicate a su b stan tial loss in residues during the approxim ately four hours betw een spray application and sampling. The most likely sources of this loss are thought to be v olatilization and/or wind erosion (G unther and Blinn, 1955; Taylor e t al., 1977). Losses in Runoff A zinphosm ethyl levels found in runoff co llected during the 1977 season are given in Table 8. No runoff was found in the orchard plots as a resu lt of th e rainfall events th a t occurred on June 27, July 5, 7, 24, 28, and August 2, 4 (Table 2). These rainfall events averaged 5.8 + 3.5 mm (intensity, 2.6 + 1.2 m m /hr) which is considerably low er than the average runoff-producing rainfall event of 21.2 + 4.7 mm (intensity, 6.5 + 1.8 m m /hr) shown in Table 8. R ainfall events th a t occurred on June 11-12 and 17-18 produced sm all am ounts of runoff in some of the plots, but sam ples w ere not analyzed. R ainfall in ten sity for th ese two events was sim ilar to those th a t produced no runoff, but th e duration of rainfall was slightly longer. The 25.4 mm (intensity, 8.4 m m /hr) rain fall th a t occurred on August 8-10 produced runoff in all but plot 4, but no analysis was undertaken as sam ples w ere not collected u n til a fte r the spray application on August 12. No runoff-producing events occurred during the 1976 season. The concentration of Table 8. Azinphosmethyl in Runoff, 1977 Season D ate Days Since Spray 30 May mm R ainfall m m /H r. Plot Runoff 20.3 5.1 mm x 10 L iters PPbc Control _o _2 0.80 6 .0 10.50 5.00 20.0 0.70 0.60 3.5 21.70 1.00 4 .0 18.80 2.00 14.0 0.74 2.50 9 .8 0.08 6.70 40 .0 0.26 10.00 40.0 0.25 1.00 8 .0 0.40 5.60 22.5 0.03 1.70 10.0 0.24 0.10 0.5 0.73 1.30 10.7 3.90 5.50 22.0 0.15 2.20 13.0 4.10 1.00 4 .0 2.25 0.50 3 .0 3.40 3.10 12.5 b a a a 4,5 June 10 25.4 9.8 mm x 10 L iters PPb 30 June 14 25.4 5.4 18 July 11 22.9 7.1 3,4,6 August 13 20.3 5.1 mm x 10-2 L iters PPb _2 mm x 10 L iters PPb _2 mm x 10 L iters PPb 0.06 0 .5 b 2.60 10.5 1.30 mm x 10-2 L iters PPb 0.50 4 .0 14.35 0.80 3 .0 cr P= 13 August 12.7 No runoff collected. c R unoff not analyzed. PPb of Azinphosmethyl in runoff w ater. 6 .4 b a a 1.00 20.4 0.03 a 0.70 10.0 0.20 34 azinphosm ethyl in the sedim ent (sedim ent am ounts w ere less than 0.5 gram s per lite r of runoff) was very sm all com pared to levels in the w a ter, and these d a ta are not rep o rted here. The lack of sedim ent is to be ex p ected as the dense ground cover on th e orchard floor allows little soil erosion (Asmussen e t al., 1977; H arrold e t al., 1970). The am ount of runoff is also expected to be less than might be found from tilled fields of sim ilar slope c h a ra c te ristic s. This may be a ttrib u te d to e ffe c t of ground cover, which increases in filtra tio n and slows overland flow, thus decreasing to ta l runoff and loss of pesticide. Many research ers have rep o rted th a t concentrations of pesticides in runoff are highest for ra in fa ll events occurring soon a fte r application (Hall e t al., 1972; Baur e t al., 1972; Glass e t al., 1974; R itte r e t al., 1974; Caro e t al., 1974). This is indeed w hat was observed for th e pesticide levels in runoff for the o rchard studied. The levels of azinphosm ethyl in runoff co llected on June 6, 10 days a fte r application, w ere 10 to 100 tim es less than the levels found in runoff co llected on June 1, just 4 days a fte r application. Tim e since spray has an e ffe c t on th e am ount of pesticide in runoff only as it is re la te d to the am ounts of pesticide residue available to runoff. The am ount of rainfall preceding th e rain fall event th a t produces runoff also influences th e am ount of pesticide th a t is lost to runoff. A com parison of azinphosm ethyl concen tratio n s in runoff co llected from plots 1, 3, and 4 on June 6 (10 days following application) w ith azinphosm ethyl concentrations in runoff from th e sam e plots co llected July 19 (11 days following application) shows much higher levels in th e July 19 runoff. This may be explained by the fa c t th a t although both events occurred approxim ately the sam e num ber of days following aplieation, no rain fall occurred betw een application and the event producing runoff on July 19, w hereas 20.3 mm of rain fall occurred ju st 4 days prior to th e June 6 event. The intervening rain fall event not only produced runoff, rem oving pesticide residues from the orchard, 35 but also red istrib u ted those residues rem aining such th a t they would be less likely to be rem oved w ith successive runoff. The fa c t th a t tre e le af residues for June 6 averaged .399, 1.267, and 1.225 p g /c m 2 for plots 1, 3, and 4, resp ectiv ely , and le af residues found on July 19 averaged 1.942, 2,286, and 2,374 U g/cm 2 for plots 1, 3, and 4, resp ectiv ely , supports this hypothesis. In general, increased runoff caused increased pesticide loss so th a t th e concentration of pesticide in varying am ounts of runoff on a given day rem ained relativ ely co n stan t. plot 2. The d a ta, for the m ost p a rt, re fle c te d this relationship, except for Plot 2 was c h a ra c te riz e d by consistently high runoff com pared to the o th e r plots, but upon analysis, azinphosm ethyl levels d e te c te d w ere considerably low er than those in the o th er plots. explanation for this. I have not y et arrived a t a satisfa c to ry One m ight hypothesize th a t co llected runoff was coming from outside the plot or possibly was interflow from upslope th a t surfaced just prior to th e collection point. O verall it appears th a t the contribution of runoff to the azinphosm ethyl loss from th e orchard was quite sm all. For exam ple, the 20.3 mm of rain on May 30 produced only .01 mm of runoff from plot 4, which contained 75 mg of azinphosm ethyl, w hereas 44 g w ere applied ju st four days prior to the event. This loss was well under 1% o f the residues present a t the tim e of the rainfall. R esidue D ata Figures 3 and 4 re p re se n t azinphosm ethyl dislodgeable residues for the various above ground layers and soil residues, for alley and canopy regions. L eaf residues are shown as e n tire tre e e stim ates using the m ethod described for Table 6. Prior to averaging across plots, s ta tis tic a l outliers w ere determ ined and elim inated using the te s t proposed by Grubbs (1969). A to ta l of 954 sam ples for 1976 and 637 sam ples for 1977 are rep resen ted , with the average num ber of sam ples per sam pling d ate as follows: leaves, 11, grass-broadleaves, 14; litte r - 36 Figure 3 (a-d) A zinphosm ethyl D islodgeable and Soil Residues for th e Canopy Region, 1976 Season. Upward d ire c ted arrow s show d ates of spray application, and downward d ire c ted arrow s in d icate dates and am ounts of ra in fa ll in mm. Each b ar is divided into tw o p arts; the m ean (u g/cm ) is above th e line and its corresponding stan d ard erro r (S.E.) below th e line. 37 X 96 £ CANOPY LEAVES 1976 219 n11 •*1i0 .*■ i 1. 8 uiest w 5 0«9 o •x z a 2 3 ( 1 ,0 a 0 .9 i.e A» ■< t »>• 1 1 L-w—i---- ----- --- i___i___i___ i __i___i_ -i T t JULIAN OATE T CANOPY GRASS-BROADLEAVES r a in cnni • X36 1 I l- 1976 TP] « J 29 11 10 i.e _i = 1.2 W CSI «o 5s. 0>6 | = o.o a J—I 0*6 1.2 J 1f 1 r 1 t JULIAN DATE CANOPY L I T T E R - M 0 S S 1976 28 UJ x 0.6 176 CANOPY S O I L 1976 TXT X 28 2 .0 1 .6 1.0 I CM : S o.5 : di r =»0.0 ! -U- 0 .5 1.0 x X 266 JULIAN DATE 1 f 1 't 1 1 1JULIAN OflTE r r 38 Figure 3 (e-g) Azinphosmethyl Dislodgeable and Soil R esidues for the Alley Region, 1976 Season. Upward d irected arrow s show d ates of spray application, and downward d irected arrow s in d icate d a tes and am ounts of rain fall in ram. Each bar is divided into tw o p arts; the mean (y g /c m ) is above th e line and its corresponding standard e rro r (S.E.) below th e line. 39 ALLEY RRIN (MM) GRASS-BROADLEAVES T TT I 6 5 T 36 23 2.4 U 1976 I 26 10 1.8 1.2 U J CM = *: w u n0.6c o N X z < £ UJ X o £ => o.o r*j 0.6 1.2 » I I 175 I i t * L T 1_____ 1 T f l LLEY 1 JULIRN OflTE 1 --------- 1--------- 1----------1----------1----------1----------1--------- ) T LITTER-MQSS RRIN I MM) T FT 6 5 V/ 36 1 .2 I 23 11 265 1976 ~ T ~ 28 10 0 .9 0 .6 U J CM x r: n X £ Q 3 _ z ◦c UJ X to o 0 .3 o\ 0.0 0.3 UJ to 0.6 i I ,» I J I I I 175 f l LLEY 2.0 I I 1 JULIRN ORTE 9. RRIN (MM) I SOIL I I I I I 1 I 265 1976 T 36 T 28 23 11 10 1.5 1 .0 U J CM X X _ _ to o 0 . 5 O -s, X Ci> % =>0.0 £ 0.5 1 .0 175 T— t JULIRN ORTE 265 40 Figure 4 (a-d) A zinphosm ethyl Dislodgeable and Soil R esidues for the Canopy Region, 1977 Season. Upward d ire c ted arrow s show d ates of spray application, and downward d irected arrow s indicate d ates and am ounts of rain fall in mm. Each bar is divided into two p a rts; th e mean (y g/cm ) is above the line and its corresponding stan d ard erro r (S.E.) below th e line. 41 CANOPY LEAVES i 101i |13 i 3 .6 2 .7 _j 1977 [T]J13 1 i |3 ijil .1 8 26 5 23 . £ 1. 6 t*JN 5j 5 z o 0 .9 £ 3 0 .0 O.o 0.4 UJ 1 «-- I , « 0.8 140 f l LLEY RRIN (UN) LITTER-M0SS 1977 T 25 T 25 20 0.8 240 JULIRN ORTE 13 13 T 13 8 23 8 25 0.6 0.4 U J CM co 5-N 0.2 O X o 1 t =>0.0 1 0 .2 i 0 .4 i i i i 140 ALLEY TTT 5 . 25 SOIL T 20 3.8 i 1977 r p 25 13 i 240 JULIAN DATE 9. RAIN (MM) _i .5 T 13 8 13 23 8 25 2.7 1 .8 llj CM 5° i5 0.9 x e> % =>0.0 x -I U -U 1- -L-L- 0 .9 1.8 -I 140 I L _ J I I I _1 I JULIAN ORTE I _J I I I _J L. 240 44 moss, 14; and soil, 4. Each bar is divided into two p arts. The mean is above the line and its corresponding standard e rro r below th e line. Time is rep resen ted by the Julian date with upward d irected arrow s indicating spray application dates. Downward directed arrow s above th e bars in d icate d ates and am ounts of rain fall in m illim eters. (These d a ta are also given in Tables 1 and 2.) The co n cen tratio n scale is designed so th a t the highest bar for a given layer, over both canopy and alley regions, is nearly full scale. The concentration of dislodgeable p esticid e residue found in a given layer a t any point in tim e throughout th e season is a function of tw o processes; m ovem ent and atten u atio n . A ttenuation is thought to include all degradative processes (chem ical, photochem ical, and m icrobial), airborne loss, p en etratio n into plant subsurfaces, and irreversib le soil binding (Ebling, 1963; Hull, 1970; K atan e t al., 1976). orchard. M ovement mainly red istrib u te s the pesticide w ithin th e This redistribution, prim arily by rain fall, is confounded w ith the a tten u atio n processes, making determ ination of a tte n u atio n ra te s in the field d ifficu lt. A tre a tm e n t of this problem using m ath em atical modeling techniques is described in P a rt II. An "eyeball" exam ination of these d a ta indicates contrasting ra te s of both m ovem ent and a tte n u atio n among the various v e rtic al s tra ta and horizontal regions of th e orchard. Comparison of le a f pesticide residues for the first spray period in 1976 (Figure 3a) w ith le af p esticid e residues for th e first spray period in 1977 (Figure 4a) shows th a t th e residues rem aining a fte r eleven days in 1977 a re roughly tw o thirds the value of those pesticide residues rem aining a fte r 12 days in th e 1976. This d ifference is possibly due to the 20 mm and 25 mm rainfall events occurring in the in terv al betw een pesticide application and the sampling eleven days la te r in 1977, while no rain fell during the first spray period of the 1976 season. T hat rain fall is responsible for the reduction of foliar residue deposits has been suggested by a num ber of 45 re sea rc h ers, many of the earlier studies are sum m arized by Ebling (1963). m ore re c e n t studies, McMechan e t al. In (1972) rep o rted th a t azinphosm ethyl applied as a w ettab le powder form ulation to apples was lost a t a much m ore rapid r a te from foliage during w et w eath er as opposed to dry. Sim ilar resu lts w ere rep o rted by Williams (1961) for azinphosm ethyl and carbaryl applied to apples, and by Thompson and Brooks (1976) for dislodgeable residues of azinphosm ethyl as and four other organophosphate insecticides applied em ulsifiable c o n cen trates to oranges in Florida. G unther e t al. (1977) also noted the influence of rain fall on the decline of parathion dislodgeable residues applied as a w ettab le powder form ulation to oranges in C alifornia. Nigg e t al. (1977) used m ultiple linear regression to exam ine th e relationship betw een residue decline of ethion dislodgeable residues (applied as E.C.) and the variables: degree-days, cum ulative le a f w etness, and ordinary tim e. They found, for the experim ent w here rain fall occurred, th a t residue decay was m ost highly o c o rre la te d w ith cum ulative rain fall (r = .963). In the present study, the le a f residue d a ta also show th a t rain fall had a significant e ffe c t on pesticide residues in the tre e . This is indicated by th e results of a 23 mm rain occurring six days a fte r th e second spray application, which reduced th e residue levels to less than h alf th e ir form er value. Sim ilar resu lts w ere rep o rted by McMechan e t al. (1972) for ra in fa ll events occurring much closer to application. In th e ir study, a 17.5 mm rain fall (1.8 m m /hr) th a t s ta rte d six hours a fte r a 50 % w.p. azinphosm ethyl application (.23 kg a .i. ha-1 ) to sem i-dw arf apple tre e s, rem oved 41% of the in itial deposit. A much lig h ter rain fall of 3.0 mm (0.9 mm hr-1 ) th a t s ta rte d five hours a fte r a sim ilar application, rem oved 12% of th e in itial deposit. Van Dyk (1976) exam ined th e e ffe c t of a rtific ia l rain fall on parathion residues applied as both w e ttab le powder and em ulsifiable c o n ce n tra te to orange, lem on, and g ra p e fru it leaves and fru it. R ainfall was applied a t 33 mm hr-1 . F acto rial 46 analysis of variance showed significant in teractio n (P < 0.01) betw een parathion residues and the sim ulated rain fall. The type of form ulation was not significant. This is not surprising considering th e high ra te of rain fall applied. By c o n tra st, a 28 mm rain occurring 17 days following th e th ird spray of the 1976 season (present study) had little e ffe c t on the apparent residue decline, suggesting th a t ra in fa ll events occurring close to application have a g re a te r influence on le a f pesticide dislodgeable residue levels. This phenomenon has also been suggested by Ebling (1963), McMechan e t al. (1972) for azinphosm ethyl applied to apple foliage, and by S teffens and Weineke (1975) for ^ C azinphosm ethyl applied to bean foliage. One theory th a t m ight explain this o ccurrence was first proposed by G unther and Blinn (1955), also by Ebling (1963), and most recen tly by E llio tt e t al. (1977). These research ers suggest th a t pesticide deposits on foliar su rfaces are lost a t d ifferen t ra te s due to th e degree a t which they adhere or p e n e tra te th e le af su rface. Initial rapid loss is a resu lt of erosion of loosely bound deposits, possibly adsorbed to the form ulations or dust on the plant. More tightly bound residues are lost prim arily through volatilizatio n , decom position, and p en etratio n into subsurface tissues. A ccording to this theory, the p ercentage of the deposit th a t is loosely bound decreases rapidly following application. It is this fractio n of the deposit th a t is m ost susceptible to erosion processes, including rain fall. As th e loosely bound residues a re lost, those m ore tightly bound residues th a t rem ain are less susceptible to loss w ith rainfall. Com parison of 1976 canopy le af residue d a ta (Figure 3a) with canopy g rassbroadleaves residue d a ta (Figure 3b) shows little d ifference in residue decline, under th e no-rainfall conditions of th e first spray period. The e ffe c t of the 23 mm rain in the second spray period of the 1976 season on residue decline in the grass-broadleaves was much less than was seen in the canopy leaves (Figure 3a). This might indicate th a t in addition to pesticide moving downward out of the 47 grass-broadleaves w ith rain fall, p esticid e is moving in from the tre e above. Comparison of canopy litte r-m o ss p esticid e residues during th e first spray period of 1976 (Figure 3c) w ith those residues found in the grass-broadleaves (Figure 3b) shows a m arked difference in residue decline, as pesticide levels rem ain a t a ra th e r constant level throughout the spray period. This would suggest th a t under no-rainfall conditions, pesticide is moving into this layer a t th e sam e ra te th a t pesticide is moving out and/or being a tte n u a te d . An a lte rn a tiv e hypothesis would be no m ovem ent and no or very slow atten u atio n , which seem s highly unlikely. Residue levels in the canopy litte r-m o ss layer increased following the 23 mm rainfall event occurring during th e second spray period of th e 1976 season, again suggesting m ovem ent of p esticid e into the litte r-m o ss from layers above, in am ounts g re a te r than m ovem ent out and/or a tte n u atio n losses. In general, canopy litte r-m o ss residue values rem ained fairly co n stan t throughout both the 1976 and 1977 seasons. Under n o -rain fall conditions (first spray period, 1976 season, Figure 3d), canopy soil residues did not increase w ith tim e following application as did the litte r-m o ss residues. As azinphosm ethyl has been shown to degrade fa ste r in non-sterile soils as com pared to sterile soils (Yaron e t al., 1974), m icrobial degradation may play an im p o rtan t role in losses from this layer. In addition, root uptake and tran slo catio n (Al-Adil e t al., 1973) may have been partially responsible for th e observed soil residue p a tte rn . Pesticide movem ent into the canopy soil w ith ra in fa ll is again pronounced, as indicated by the 23 mm rainfall event during th e second spray period of th e 1976 season (Figure 3d). The first spray period o f the 1977 season shows canopy soil residue levels increasing w ith each ra in fa ll ev en t (Figure 4d). Alley grass-broadleaves pesticide residues show a sy ste m a tic decline over th e first spray period of th e 1976 season (Figure 3e). The decline is slightly ste e p e r than the decline shown for canopy grass residues during the sam e no­ 48 ra in fa ll period (Figure 3b). m ovem ent from th e This differen ce is probably due to less pesticide tre e s into the alley grass-broadleaves. Increased a tte n u a tio n losses in the alley regions due in p a rt to g re a te r exposure to wind and solar radiation must also be considered. fa c to rs influencing dieldrin volatilization radiation as th e m ost significant. Taylor e t al. (1977) in discussing from orchard grass, cited solar A ssociated su rface tem p eratu res may also influence ra te s of chem ical and m icrobial degradation and foliar uptake (Ebling, 1963; Hull, 1970). A zinphosm ethyl applied to plant surfaces has been shown in one study (Liang and L ichtenstein, 1976) to be susceptible to photodegradation by sunlight. This degradative pathw ay is d irectly re la te d to solar radiation. R ainfall e ffe c ts are clearly shown when com paring alley grass residue to decline for th e first spray period in 1976 (Figure 3a) w ith th e first spray period of the 1977 season (Figure 4e). The differen ce in rain fall e ffe c ts betw een canopy grass-broadleaves and alley grass-broadleaves is shown in the second spray period o f the 1976 season (Figures m ovem ent, following the 3b and 3e) which indicates th a t th e pesticide 23 mm rain fall ev en t, out of th e alley grass- broadleaves is probably g re a te r due to d ire c t exposure to rain fall, and also th a t th e re is little or no pesticide m ovem ent in from th e tre e s. a tte n u a tio n due to shading cannot be disregarded. Again, d ifferen tial Alley litte r-m o ss pesticide residues show a definite decline during th e first and last spray periods of the 1976 season (Figure 3c). While canopy litte r-m o ss residue values increased with ra in fa ll during th e first spray period of th e 1977 season (Figure 4c), th e alley litte r-m o ss residue values show a decline, indicating th a t residue m ovem ent out and/or aten u atio n is g re a te r than residue m ovem ent into this alley layer. Alley soil residues for the first spray period of the 1976 season (Figure 3g) are sim ilar to canopy soil residues for th e sam e period (Figure 3d, note the differen ce in scaling). There appears to be a decrease in alley soil residue levels following the 49 23 mm rain during the second spray period of the 1976 season (Figure 3c) w hereas th e canopy soil residues (Figure 3d) increase following this rain fall ev en t. The high variability of these d ata discourages speculation as to the processes taking place. The general p a tte rn of th ese residue levels over th e season, c h a ra c te riz e d by a buildup of residues tow ard m id-season, is surprisingly sim ilar to th e azinphosm ethyl orchard soil residue p a tte rn rep o rted by Kuhr e t al. (1974). A general observation of the d a ta presented in Figure 3 indicates th a t, as th e season progressed, the interrelationship betw een pesticide m ovem ent and a tten u atio n becam e more difficu lt to follow. This com plexity is possibly a function of changes in w eather p a tte rn s in addition to rain fall (tem p eratu re, hum idity and wind). Gunther e t al. (1977) a tte m p te d to re la te te m p e ratu re d a ta to the decline of azinphosm ethyl residues applied to citru s in southern C alifornia. They found it difficult to make any m eaningful in te rp re ta tio n of the d ata, but s ta te d th a t azinphosm ethyl dissipation was slightly more rapid during w arm er w eather. No correlations betw een residue d ata and te m p e ratu re or hum idity w ere a tte m p te d in the present study, but w eather d a ta was provided fo r possible future use (Figures 5, 6, 7, and 8). Changes in plant physiology and soil m icrobial a c tiv ity or a buildup of tightly bound or p e n etrate d residues, m ay also have an e ffe c t on pesticide residue dynamics throughout the season. However, from th e lim ited analysis of th e d a ta, th e re is some indication th a t azinphosm ethyl is redistributed throughout the orchard w ith tim e following application. R edistribution is indicated during periods w ithout rain fall, but is m ore pronounced following rainfall events. L ittle more inform ation can be gained w ithout th e aid of a more sophisticated technique of analysis. analysis of the d a ta is presented in P a rt H. Such an Using m ath em atical m odeling techniques, the change in observed residue levels in each layer and region is e tim a ted as a function of both m ovem ent and a tten u atio n . 50 Figure 5 Daily T em perature R ange, 1976 Season June A u g u st 52 Figure 6 Daily Minimum and Maximum R elativ e H um idity, 1976 Season 100 90 80 70 60 50 40 30 20 June July August 54 Figure 7 Daily T em perature Range, 1977 Season 100 35 90 30 80 25 70 20 IO in 60 15 50 10 40 30 Ju ne July August 56 Figure 8 Daily Minimum And Maximum R elativ e Hum idity, 1977 Season 100 90 80 70 R.H. 60 50 40 30 June July A u gust 58 Inform ation on pesticide co n cen tratio n s as a function of m ovem ent and atten u atio n is essen tial to th e estim atio n of possible exposure to th e biota of this agroecosystem . If p esticides a re to be used in an e ffe c tiv e m anner in conjunction with biological contro l techniques, possible exposure to ben eficial populations, as well as th e ta rg e t species, m ust be b e tte r understood under a variety of clim atic conditions. PART II ASSESSMENT OF THE ATTENUATION AND MOVEMENT OF AZINPHOSMETHYL IN A MICHIGAN ORCHARD ECOSYSTEM: PARAMETERIZATION OF A FIELD-BASED MODEL INTRODUCTION In P a rt I a description of th e field d a ta co llected on th e distribution, m ovem ent, and atten u atio n of azinphosm ethyl in an experim ental apple orchard was given. These d ata w ere gath ered sp ecifically in order to allow developm ent, refin em en t, and p a ram eterizatio n of a model describing th e sp atial and tem poral distribution of azinphosm ethyl in th e orchard in response to rain fall. While th e tim e series of concentrations observed w ere rep o rted in th e earlier p a rt, th e m odel and th e p a ram eterizatio n process, to g e th e r w ith th e p a ra m ete r values g en erated , are described in this p a rt. Validation of this model using a third season's d a ta is presented. The form of this model was chosen to allow: (1) use of model output to provide pesticide exposures for models of organism s dwelling in the orchard floor (Goodman, 1980) and (2) fu tu re developm ent to rep resen t th e dynamics of other pesticides and th e e ffe c ts of additional environm ental facto rs. THE MODEL In order to stru c tu re th e experim ental program and th e d ata analysis, a conceptual model for the distribution, m ovem ent, and fa te of the pesticide in the orchard was form ulated. Inform ation g ath ered has resu lted in continual 59 60 refinem ents of the model stru c tu re . The co n cep tu al model utilizes a sp atial subdivision of the orchard into a canopy region and th re e alley regions (Figure 2 of P a rt I). V ertically, four s tr a ta (not necessarily all p resen t a t a given sam pling location) are identified. These are called tre e , grass-broadleaves, litte r-m o ss, and soil. The conceptual m odel (see Figure 1) describes the dynamics of the pesticide from its spray application to its u ltim a te disappearance from th e orchard via d rift, atten u atio n (including airborne loss, photolysis, chem ical degradation, m icrobial degradation, and p e n etratio n of su rface residues) and runoff. Each day, a v ecto r C of co n cen tratio n s of pesticide in each of seven regions (two horizontal by four v e rtic al, minus one fo r non-existent alley trees) is calculated, based on m anagem ent actions (spraying, mowing) and rain fall. Figure 1 describes the processes a ffe c tin g th e p esticid e con cen tratio n in only one of the seven regions. For analysis of the field d a ta, the processes in Figure 1 are lum ped into th re e categories: initial distribution, m ovem ent, and atten u atio n . M ovement is fu rth er subdivided according to rain fall in ten sity (none, light, and heavy). The m ath em atical form s of the various com ponents of this model are described below: (1) Initial D istribution The spray ra te (in Kg/ha) is supplied as an input. D rift, including losses from th e orchard during spraying and up to th e tim e of post-spray sam pling, was e stim ate d using mass conservation, based upon the known application r a te , e stim a te d pre-spray residue levels, and measured residues following application (see P a rt I). The proportion d rift averaged .376 + .149 over th e th re e seasons. Because 61 Figure 1 C onceptual Model for the D istribution, M ovem ent, and A ttenuation o f A zinphosm ethyl in an Apple O rchard r^PMtlcJSS^V LX j Spray Rate Degradation Products "" Microbial Degradation Agricultural Alterations Photolysis Non-rain Vertical Movement CM (O nvironment and Compound Characteristics Chemical Degradation Penetration Rainfall Vertical Movement e---------^Orchard Physical^ Param etara * Lateral Movement e— ! TT _ l_ Runoff J 63 the standard deviation observed among the d ata was higher than an tic ip ate d , an a tte m p t was made to re la te th e d rift loss to wind speed a t the tim e of application from the available d ata. Mean wind speeds during application w ere estim ate d from d a ta of th e M ichigan W eather Service, Grand Rapids, M ichigan. The following relationship was determ ined: D = .021 WS + .091 (r = .475) w here D is proportion d rift and WS is wind speed in km /hr. The residue deposited in the orchard is apportioned into the seven regions according to a spray distribution v ecto r, in which each en try specifies the proportion of the spray th a t is cap tu red by th e corresponding region, and added to any rem aining residue from e a rlie r sprays in the pesticide co n cen tratio n v ecto r C. All units are expressed as y g pesticide/cm (2) 2 ground a rea. A ttenuation A ttenuation is tre a te d in th e model as a s e t of daily proportion losses—a single proportion for each layer. Thus it is conveniently representable as a diagonal m atrix A which pre-m ultiplies the pesticide distribution v ecto r C, a seven-elem ent column v ecto r o containing th e concentration of pesticide (y g/cm ) in each region a t a p a rtic u la r tim e. Each day, a tte n u atio n is e x tra c te d via equation ( 1): C (after) = (I-A)C(before) (1) w here I is th e 7 x 7 iden tity m atrix. and Crj, the pesticide concentrations in canopy and alley soils, resp ectiv ely , represen t to ta l soil residues, while C^, C C g , Cg, and Cg rep resen t only dislodgeable residues. Thus th e 64 a tte n u a tio n for non-soil layers includes su rface pen etratio n of residues. (3) M ovement Daily redistrib u tio n of th e p esticid e within the orchard is m odeled using th re e m atrices distribution colum n v ecto r C. to pre-m ultiply the pesticide Daily m ovem ent not a ttrib u te d to ra in fa ll is m odeled by equation (2): C (after) = PC (before) (2) w here P is a 7 x 7 colum n-stochastic low er triangular m atrix known as the non-rainfall pure m ovem ent m atrix . The m atrix P is re s tric te d to containing a t m ost 19 non-zero e n trie s, as th e tre e layer is th e only canopy lay er from which m ovem ent to alley layers is modeled, so Py = 0, for i = 5, 6, 7 and j = 2, 3, 4. Several p a ra m ete rs are n ecessary to adequately describe a rain fall event; in ten sity , e tc . for exam ple, duration, average intensity, peak U nfortunately the sm all num ber of rain fall events during a spray season precluded using so fine a description. R ainfall events w ere classified on a daily basis into only tw o categories (heavy and light) in order to obtain enough instances of each category to p a ra m ete riz e the model. Heavy rain was m ore than 10 mm rain fall or m ore than o th e r m easurable rain classified as light. defined as any event of 5 m m /hour in a day, with Equations (3) and (4) show th e pesticide redistrib u tio n s caused by heavy and light rain fall, respectively: C (after) = H C(before) (3) C (after) = LC(before) (4) w here H and L a re 7 x 7 colum n-stochastic low er triangular m atrices. 65 As in equation (1), H and L do not model m ovem ent betw een alley and canopy ground layers; thus pesticide m ovem ent with overland runoff is not included in th ese equations. A ttenuation and non-rainfall m ovem ent w ere m odeled (and param eterized) as daily phenom ena. Thus, to rep resen t the n atu ral changes in pesticide distribution from one day to th e nex t, (C(k) to C(K+1)), e x actly one of the following relationships is used: C(k+1) = P(I-A)C(k) (no rain) (5) C(k+1) = L P(I-A)C(k) (light rain) (6) C(k+1) = H P(I-A)C(k) (heavy rain) (7) METHODS FOR PARAMETER ESTIMATION This section describes the techniques used to param eters from the d a ta base of field d eterm inations. e stim ate the model The routines described are designed to be used repeatedly , i.e., as e n tries are added to the d ata base, new p aram eters including th e ir e ffe c ts can be quickly g en erated . This capability for dynamic re p a ram ete riz a tio n of the model gives it the flexibility to begin w ith relativ ely few crude d a ta, producing prelim inary outputs, and to produce m ore a cc u ra te results as the d a ta base grows. D ata Base The d ata base consists of a large num ber of sequential disk files, which are updated and utilized by various program s. R ainfall d a ta, orchard plot physical c h ra c te ristic s, and records o f sam ples analyzed for pesticide are stored in these files. Figure 2 shows how sam ple and ta rg e t d a ta , en tered as "raw" outputs from gas chrom atograph m easurem ents, are tran sfo rm ed into new files of "processed" sam ple and ta rg e t d a ta . In this stag e, ap p ro p riate corrections for an aly tical technique, sam ple w eight, averages across sam ples, e tc . are made, yielding files suitable for use in calculating th e model p a ram eters. Figure 3 is a flow chart of 66 Figure 2 Flow C hart Showing D ata Processing to Yield Files Suitable for Use in Model P a ra m e teriza tio n 67 INPUT FILES OUTPUT FILES / INPUT 1 New I D a ta c a r d DATA SAM PLES / I ( O ld F ile s < \ c DATA A TARGETS J OATA A SAM PLES J C NONRAIN TARGETS HILO AVG f CORRECTED \ (s a m p le s I ( C N O N R A IN ^ \ TA RGETS J /C O R R E C T E D ^ (s a m p le s J DAY AVGS OUTLIER DAY AVGS AVERAGE PR O C E SSE D ] SA M PLES J ( f ----( da ta t a r g e t s1 AVERAGE NONRAIN ^ TARGETS AVERAGE ( h > U p d a te d DATA ^ TA RGETS DATA SAM PLES CORRECT > f PRO CESSED ^ (T A R G E T S J J ' F lie s 68 Figure 3 Flow C hart of the Process of C alculating th e P a ra m e ters using th e "Processed Files" 69 ( ME ASURED \ . LEAF AREAS J LEAF T = ----------- c PLOT AREAS ^ SPR AY RATE ^ ----------- ( ( P R O C E S S E D ^ ----------TARGETS Q / T R E l \ ---------------- » ( I NTERCEPTION ) V INDEX 1£1 ) --------------- j AREA J P R OC E S S E D SAMPLES V -•P \ Ml MANUAL SELECTION RAINFALL INTERVALS J r SELECTED \ RAINFALL ) INTERVALS J V ADAPT 1 -~—* f X5 ■< PEN RATE S P R DIST vy J \ IMMEDIATE j n )] P R E / P O S T RAIN S A MP L E VAL. — !■ -!— ■ —» —> NON-RAIN MOVEMENT - ( a t ten u a tio n ! s^r m ADAPT 2 9 ) * / " S ELECTED \ --------------- * ( N O N - R A I N F A L L ) _ A . INTERVALS ' *( ADJUST ) NON-RAINFALL^ INTERVALS i £ : FIND EVENT RAIN DATES \ areas y TARGET ~ A.PISTRIBUTION J ! | TARGETS j PS E UDO I J RAINFALL MOVEMENT ) INITIAL PEN A ATTENUATION J < _y s p ra y V d istribution \ J ----------I___ — P SIT REP < SITUATION REPORT 'N J _____ the process of calcu latin g th e p aram eters using th e "processed" files. Files used by a given routine are shown with a dashed arrow into th e routine, and files produced are shown w ith a dashed arrow to the files. Flow of the program is via th e solid arrow s. Subroutines LEAF and AREA perform needed conversions of some residue m easurem ents (for exam ple, o p esticide/cm ground area). from yg p esticid e/cm leaf a re a to yg TARGETS uses "processed" ta rg e t values to determ ine a proportion o f pesticide reaching each plot and region. The process of p aram eterizin g this model continues w ith FINDEV, a subroutine which searches th e d ata files for se ts of sam ples useful for c alculating red istribution and losses of the p esticide. For estim atin g loss and m ovem ent in the absence of rain, it prepares a file of sam ple sets each of which consists of a vecto r of concentratio n s (in each region and layer) a t the beginning of a non-rainfall in te rv a l and a sim ilar v ecto r a t th e end of the interval. Of course, e ith e r rain or a spray intervening will exclude a pair of sequential sam ple vectors from being used for this purpose. FINDEV also prepares sim ilar files of vectors rep resen tin g sam ples before and a fte r rain fall events of various inten sities to allow estim atio n of the e ffe c ts of th ese rains on the pesticide distribution. As shown in th e diagram , m anual selection is utilized to determ ine rain events to be classified as sim ilar, and to exclude sam ple d ates which are unacceptable for the task to be perform ed (when num erous d a ta are missing for a given sam ple date). The m atrices A and P (equations (5), (6) and (7)) are d eterm ined from a m atrix Q = P(I-A). A fte r Q is found (see equation (9) below), the e n trie s A ^ are c alcu lated as: 71 From equation (8), Q should re la te the pesticide v ecto r C(k) to C(k+1) for each day kj on which no rain fall occurred. C ^ X - .C ^ ) ] Thus, given m atrices B = [C(k1) and A = [C(k^ + 1) C(k^ + l)...C (k n + 1 )], kj 2 [days w ithout rainfall] , th e following relationship should hold: A = QB. (9) So long as a t le ast seven sets of b e fo re /a fte r concentrations are available, Q should be uniquely determ inable using Gaussian elim ination. However, owing to both random and sy stem atic variations in th e d ata sets, such a procedure did produce a feasible solution. It is necessary to introduce additional co n strain ts th a t a "best fit" m ust satisfy . Entries in Q m ust be re s tric te d to th e range [0, 1 ], so th a t pesticide rem oved from a region is lim ited to w hat is available for m ovem ent. Also, because too few soil sam ples w ere analyzed to allow for a reliable solution for the pesticide a tten u atio n ra te in soils, a ra te determ ined from th e incubation of the orchard soil in th e lab was introduced into th e solution m atrix before the rem ainder of the 19 w ere determ ined. Solution for the en tries in Q to optim ize th e fit to all available d ata for non-rainfall intervals was carried out in program ADAPT1, using the adaptive optim ization technique o f Holland (1975), also described in DeJong (1980). The optim ization was done on a broadened version of equation (9) allowing varying in tervals betw een sam pling days, since: C(k.+m) = QmC(k.), m > 1, so long as no rain falls on days k j,..., k j. Once Q was determ ined, it was used, to g eth er w ith Equation (6) and d a ta surrounding light rain fall events, to calcu late L, th e light rain m ovem ent m atrix. Because sam ples w ere typically collected a t intervals of th re e to five days, the e ffe c ts of degradation and non-rainfall m ovem ent were rem oved from th e d a ta surrounding each rain fall before th e rain e ffe c t could be determ ined. equations (5) and (6), we obtained: Using 72 C(k.+p+q) = Qq L Qp C(kj), w here C was known a t days kj and kj+p+q and light rain o ccurred on day k^+p. Thus (Q C(k.+p+q) = L Qp C(kj), and m atrices o f b efo re- and a fte r-ra in fa ll e stim a te d co ncentration vectors were assem bled as -1 q 1 -1 q A l = [(Q ) AC(k1+p1+q1) ... (Q ) nC(kn+Pn +qn)l Bl = |(QP l C(k1) ... (QPnC(kn)] and the m atrix L in ADAPT1. =L was optim ized by ADAPT2, a program sim ilar to It also solves, in an analogous fashion, for the heavy rain fall m ovem ent m atrix H of equation (8). Once the m ovem ent m atrices are p aram eterized by the adaptive routines, th e in itial pesticide distribution fo r spray applications subsequent to the first application of the season a re e stim ate d . The a tten u atio n and m ovem ent m atrices are used, along w ith th e residue distribution on th e last sam pling day of the preceding spray period, to e stim a te the dislodgeable residues present just prior to application. Pen r a te (Figure 3) calcu lates th e e stim ate of th e am ount of su rfa c e -p e n e tra te d residues presen t a t th e tim e of application based upon its estim ate of the dislodgeable residue and a ra tio of su rfa ce -p e n e tra ted to dislodgeable residues. This ra tio was d eterm ined from m atched sam ples taken th e la st sam pling day for se le c te d spray periods over th e 1976, 1977 and 1978 seasons. The su rfa c e -p e n e tra te d to dislodgeable residues and standard errors w ere as follows: leaves, 0.110 + .003; grass, 1.075 + .195; litte r, 2.367 + .263; moss, 1.701 + .184. These ra tio s w ere m ultiplied by th e appropriate dislodgeable residues to e stim a te th e su rface p e n e tra te d residues present just prior to application. To e stim ate the in itial pesticide distributon, the estim ate s of dislodgeable and su rfa c e -p e n e tra te d residues present a t th e tim e of application w ere su b tra cte d from the to ta l residues m easured the afternoon following application. Assuming these e stim ate s are conservative (see P a rt I), d rift is determ ined as the difference betw een the e stim a te of the am ount deposited in th e orchard a t application and the am ount applied. Using the m ethod described 73 above, the average proportion and standard deviation of the dose initially distributed to th e orchard, for all spray applications over th e th re e seasons (except two applications for which th e re was rain fall following th e la st sampling day of the preceding spray period) was .624 + .149, giving an average d rift e stim ate as reported earlier. The final routine, situ atio n re p o rt, produces a daily record of both m easured and predicted residue values. The d a ta is output in both tabular and graphical form , as shown in Figures 8a-g. RESULTS AND DISCUSSION P aram eterizatio n of th e A ttenuation and M ovem ent M atrices The model was param eterized with azinphosm ethyl residue d a ta, co llected over two seasons (1976, 1977) as previously described in P a rt I, plus a third season (1978). The m atrices g en erated (for equations (1) to (4)) are shown in Figures 4, 5, 6, and 7. The m atrices describe the daily a tten u atio n and m ovem ent of pesticide under th re e specified ra in fa ll conditions; none, light, and heavy. The no-rainfall m ovem ent m atrix (P) shows some pesticide m ovem ent from th e tre e , possibly due to wind erosion, dew or g u ttatio n , w ith th e m ajority of the residue being trapped in the litte r-m o ss layer. Also of in te re st is the fa c t th a t th e model estim ates th a t equal am ounts of residues (approxim ately 0.9%) move from the tre e to the canopy litte r-m o ss and alley litter-m o ss layers. G unther e t al. (1977) estim ate d th e m ovem ent of parathion from orange tre e s to the ground over a five-day dry period to be less than 1% of the applied dose. In order to directly d e te c t this m ovem ent in th e orchard used in this study, filte r paper ta rg e ts were placed on the orchard floor (as described in P a rt I) for periods w ithout rainfall of up to four days. R ecovery of approxim ately 0.5% of in itial tre e residues confirm ed th e existence of som e pesticide m ovem ent. The 74 Figure 4 A zinphosm ethyl A tten u atio n M atrix ATTENUATION CANOPY LEAVES CANOPY GRASSBROAOLEAVES CANOPY CANOPY LITTER-MOSS SOIL ALLEY ALLEY GRASSLITTER-MOSS BROADLEAVES ALLEY SOIL .0 4 9 .041 .1 6 7 .0 7 9 .0 6 7 .2 2 3 .0 7 9 Figure 5 A zinphosm ethyl N on-R ainfall M ovement M atrix FROM N O N -R A IN FA L L M O V EM EN T MATRIX CANOPY LEAVES CANOPY CANOPYCANOPY GRASS* LITTER-MOSSSOIL BROADLEAVES ALLEY ALLEY GRASSLITTER-MOSS BROADLEAVE8 CANOPY LEAVES .983 CANOPY GRASS* BROAOLEAVES 0 .0 .935 CANOPY LITTER-MOSS .008 .065 .907 CANOPY SOIL 0 .0 0 .0 .093 ALLEY GRASS* BROAOLEAVES 0 .0 ALLEY LITTER-MOSS .0 0 9 .0 2 7 .9 9 6 ALLEY SOIL 0 .0 0 .0 .0 0 4 ALLEY SOIL h* N rf\ I s 1 .0 .973 1 .0 78 Figure 6 A zinphosm ethyl Light R ainfall M ovem ent M atrix FR.QM LIGHT RAINFALL CANOPY LEAVES „ TO CANOPY LEAVES CANOPY CANOPY GRASSLITTER-MOSS BROAOLEAVES CANOPY SOIL ALLEY ALLEY GRASSUTTER-MOSS BROADLEAVES ALLEV SOB. .9 3 2 CANOPY GRASS* BROADLEAVES .0 3 0 .7 4 7 CANOPY LITTER-MOSS .0 0 2 .071 1 .0 CANOPY SOIL .0 3 6 .1 8 2 0 .0 ALLEY GRABSBROAOLEAVES 0 .0 .669 ALLEY LITTER-MOSS 0 .0 0 .0 1 .0 ALLEY BOB. 0 .0 .131 0 .0 1 .0 1 .0 A zinphosm ethyl Heavy R ainfall M ovement M atrix FROM HEAVY RAINFALL Tn LU CANOPY LEAVES CANOPY CANOPY GRASSLITTER-MOSS BROADLEAVES CANOPY LEAVES .822 CANOPY GRASS* BROADLEAVES .020 .613 CANOPY LITTER-MOSS .096 0 .0 1 .0 CANOPY SOIL .008 .3 8 7 0 .0 ALLEY GRASSBROAOLEAVES ALLEY LITTER-MOSS ALLEY SOIL CANOPY SOIL ALLEY ALLEY GRASSLITTER-MOSS BROAOLEAVES ALLEY SOIL 1 .0 .0 2 7 .2 8 0 0 .0 .447 1 .0 .027 .2 7 3 0 .0 1 .0 82 unknown ra te of loss from the ta rg e ts precludes a d ire c t assessm ent of the m ovem ent ra te based only on these d ata. Figure 4 shows an a tte n u atio n ra te residues in the tre e of 4.9% day *. for azinphosm ethyl dislodgeable Hall e t al. (1975) exam ined th e loss of azinphosm ethyl dislodgeable residues applied to apples using tw o types of air b last sprayers. Trees w ere sprayed on one side only and each tre e was divided into nine sites for sampling. The average residues rem aining for sites one to five (which m ost closely correspond to th e tre e a re a sam pled in the p resen t study) over a 14-day period w ere used to d eterm in e th e loss ra te . w ere 7.6% day R ates determ ined -1 2 -1 2 (r = .914) for the high flo w -rate application and 5.5% day (r = .887) for the low air flo w -rate application. These loss ra te s are higher than the a tte n u atio n ra te alone as determ in ed in th e p resen t study, but are in ex cellen t ag reem en t with the overall loss ra te of azinphosm ethyl dislodgeable residues from the tre e of 6.7% day 1 (obtained by sum ming atten u atio n and m ovem ent from the tre e under dry conditions). These resu lts suggest th a t approxim ately 25% of the daily loss of azinphosm ethyl dislodgeable residues from th e tre e is red istrib u ted within the orchard, under dry conditions. Figure 5 shows th a t less m ovem ent o ccu rred from th e grass-broadleaf to litte r-m o ss layer in the alley region than in th e canopy, suggesting th a t conditions under the tre e canopy may be m ore favorable to m ovem ent, in th e absence of rainfall. The atten u atio n ra te for th e canopy grass-broadleaves is sim ilar to th a t observed in th e tre e . However th e a tten u atio n ra te for the alley grass-broadleaf layer is considerably higher. Increased exposure to solar radiation resulting in higher su rface te m p e ratu res, which in turn influence losses by volatilization, degradation, and plant uptake (Ebling, 1963; Hull, 1970; Bukovac, 1970) may be responsible for th e observed d ifferen ce. M ovement out of the canopy litte r-m o ss layer in the absence of rainfall was sim ilar to th a t 83 from the grass-broadleaves, w ith virtually no m ovem ent from the alley litte r moss. A ttenuation ra te s observed in th e litte r-m o ss layer a re higher than m ight be expected, but little is known about p esticid e atten u atio n on th ese su rfaces. Again th e ra te observed in th e alley was g re a te r than th a t in th e canopy. The reason for this difference is thought to be th e sam e as for th e grass-broadleaf layer. The laboratory-determ ined ra te for azinphosm ethyl a tte n u atio n in soil (7.9% day *) is only an approxim ation of the field ra te . However this ra te is in rough agreem ent with th e ra te of 6.3% day- '1 determ ined from th e field d a ta o f Ruhr e t al. (1974), in which sam ples w ere tak en from beneath apple tre e s in an u p state New York orchard (ignoring any m ovem ent of azinphosm ethyl into th e soil). A loss ra te of 5.8% day -1 was determ ined from the field d a ta of Schulz e t al. (1970), in which azinphosm ethyl was applied as an em ulsifiable c o n c e n tra te to th e soil surface. The experim ent was run in th e early spring in Wisconsin (soil te m p e ratu re 5-15°C). Only a general com parison can be m ade with th ese field studies, as soil type, pH, available m oisture, ground cover, and te m p e ra tu re , as well as am ount and mode of application, may influence azinphosm ethyl loss from soil (Ham aker, 1972; Schulz e t al., 1970). R ainfall-induced losses of azinphosm ethyl and oth er deposit residues in orchards have been indicated by a num ber of research ers (McMechan e t al., 1972; Williams, 1961; Thompson and Brooks, 1976; G unther e t al., 1977; Nigg e t al., 1977). The influence of rain fall on th e rem oval of azinphosm ethyl from apple foliage is discussed in P a rt I. The present tre a tm e n t of th e d ata calcu lates azinphosm ethyl loss as a function of heavy or light rain fall (as defined above). No a tte m p t was made to re la te susceptibility to rain fall rem oval w ith th e age of th e residue deposit. Figures 6 and 7 show th a t approxim ately 7% and 12% of the residues are moved out of the tre e by light and heavy rain fall, resp ectiv ely . The light rain moves all these residues into th e canopy region, while th e heavy 84 ra in fa ll moves tw o-thirds to the canopy and one-third to the alley region. Values for proportions moved to th e individual layers are probably not as reliable, but generally light rain moves the residues to the canopy grass-broadleaf and soil layers while the heavy rain moves th e larg est proportion o t the canopy litte r moss layer. am ounts. Canopy and alley grass-broadleaves and alley soil receive lesser Eighteen and 38% of th e residues deposited on th e canopy grass- b roadleaf layer are moved to the soil following light and heavy rainfalls, resp ectiv ely . T hirteen p e rc en t is moved from th e alley grass-broadleaves to the soil as a resu lt of light rain fall, w hereas 27% is moved as a resu lt of heavy rain fall. The larg est proportion of alley grass-broadleaf residue (45%) is moved to th e litter-m o ss following a heavy rain fall. No m ovem ent is indicated from the litte r-m o ss layer as a resu lt of e ith e r light or heavy rain fall. C ertain types of m ovem ent are difficult to distinguish based on the available d a ta. For exam ple, a large am ount of m ovem ent from the canopy grass-broadleaves to the soil, accom panied by a sim ilar am ount of m ovement from th e tre e to the canopy grass-broadleaves, can produce th e sam e resu lt as a d ire c t m ovem ent from the tre e to the soil. Only a widely varying s e t of in itial pesticide distributions would enable th e p aram eterizatio n routes to distinguish th e se two processes definitively. However, in the operation of the model, the n e t resulting distribution will be sim ilar in e ith er case, so long as the in itial distribution is sim ilar to the one used to p aram eterize the model. More cred ib ility should be a tta c h e d to th e m odel-generated distributions than to th e individual m atrix en tries. Com parison of the Model O utputs w ith the Field D ata Figures 8a-8g show a te s t of th e model's predictions as com pared to th e field d a ta obtained during the 1978 season. The continuous lines (solid or dashed) re p re se n t th e model's daily prediction of azinphosm ethyl dislodgeable foliar and 85 Figure 8 (a-d) A ctual and P redicted A zinphosm ethyl Dislodgeable and Soil Residues for th e Canopy Region, 1978 Season. Upward d irected arrow s show dates of spray application, and downward d irected arrow s indicate dates and am ounts of rain fall in mm. Each bar is divided into two p arts; the mean (p g /c m ground area) is above th e line and its corresponding standard e rro r (S.E.) below th e line. The solid line running through th e bars rep resen ts the model's prediction of the change in residues using 1976 and 1977 d a ta only, while the dashed line shows the p red icted change in residues based on all th re e years d ata. 86 (nm CANOPY LEAVES 1978 CANOPY G RASS-BROADLEAVES CANOPY LITTER -M O SS 13.2 9.9 | 6.6 n 5 3.3 c o ^ J = 0 .0 a 3.3 (nn) 2. 0 l.S | 1.0 2 g 5 0.E % ON. x a % 3 0.0 . | I II 3 III f vV 2 17 11 GRASS-BROADLERVES T n r T t r 41 2 51 1978 T 14 U J Csl rw 0 .5 XL - 1.0 . 1 . 1___ » T 150 " - ALLEY rrin 1.2 2 17 11 I 51 6 250 JULIRN ORTE LITTER-M0SS emu ’ H ' t i- V 5 1978 TITTT" 3f w3 w 8I 4 1 2 8 14 0.9 0 .6 U JM CANOPY SOIL 1978 2 .8 JULIRN ORTE 1978 1978 118 Figure 4 (e-g) A zinphosm ethyl Dislodgeable and Soil R esidues fo r the Alley R egion, 1978 Season. Upward d irected arrow s show d ates of spray application, and downward d irected arrow s in d icate d ates and am ounts of rain fall in ram. Each bar is divided into two p arts; the mean (y g/cm ) is above th e line and its corresponding stan d ard e rro r (S.E.) below th e line. 119 ALLEY RAIN (Mfl) -j £ 2 0 .9 _ 0 .6 , T 77 1V 1.2 17 GRASS-BR0ADLEAVES 11 T u n 3r 3 3 SI 4 1 1978 T 2 14 U J CM w O X 5-S . 0 . 3 © £ =»o.o M cr J-L . 0.3 . 0.6 . i i r 150 ALLEY RAIN (MM) • 2 17 1---------1— __ I_____ I_____ I_____ L f JULIAN DATE 11 J SI .I - I 1 1 I T LITTER-M 0SS T H 0.8 —I 1 1r 245 1978 T 4 1 2 8 14 0.6 * 0.4 ui eg £ 5 0.2 O v. X CD X 3 S -1—L 4 ■U ° * ° - 4- 0. 2 0.4 ff f ‘---------1--------i ------- I---------1---------- 1 ISO A------1 rain (nm SOIL Tin ttr 3 .? V 2 1. 6 1---------- L _ — I-------1--------- 1 f L -- i I 17 11 1978 I . mr 14 4 1 51 I 245 JULIAN OATE ALLEY 9. j 1. 2 o.e U I CM 5 o 0.4 O •s. X CD 6 3 0.0 -H-J- I I I I I t.. -liu 0.4 0.8 ISO t ' ' ' t ' ' JULIAN OATE 245 120 solar radiation input. The authors qualify this sta te m e n t with the hypothesis th a t solar radiation is one fa c to r th a t may be singled out as a source of vapor flux variation, as solar radiation input a ffe c ts all o th er p aram eters o f th e com plex relationship betw een the crop m icroclim ate and the ad jacen t atm osphere. One p a ra m ete r which solar radiation m ost definitely a ffe c ts is th e te m p e ratu re a t the le a f su rface, which has a d irect e ffe c t on th e ra te of volatilization. A sim ilar conclusion was reachd by Phillips (1974) from th e resu lts of laboratory experim ents in which the volatilization ra te of dieldrin from glass and co tto n le af surfaces was m easured a t various tem p eratu res and wind speeds. Phillips sta te d th a t "clearly the fa c to r having the most d ram atic e ffe c t on loss was te m p e ratu re." These observations would suggest th a t a possible p a rtia l explanation for differences in airborne concentrations m easured in th e afternoon on successive spray d ates may be the differences in solar radiation input. The average solar radiation (cal cm -2 min -I ) m easured during the afternoon sam pling periods on each of th e four application days was as follows: 1.11 on June 9, 1.06 on June 29, 0.57 on July 21, and 1.00 on August 10. The value of 1.11 cal cm 2 min -1 on June 9 was 100% of th e available solar radiation (cloudless day). June 29 was also clear, July 21 was cloudy, and August 16 was partly cloudy. Comparison of solar radiation with th e average airborne pesticide concentration m easured a t the cen te r location (Location 6, Figure 1) for these application days (Table 1) shows a nonlinear trend with th e highest airborne residues on June 9, which had the highest solar radiation input, and the low est airborne residues on July 21, which had th e low est solar radiation input. In addition to solar radiation, the influence of wind speed on the airborne pesticide concentrations m easured was also exam ined. In th e com parison of th e mean wind speed a t 6.0 m eters (see Table 1) m easured during th e afternoon sam pling period (1300-1500 hr) with th e average airborne pesticide co n cen tratio n 121 m easured a t the c e n te r location for th e four application d ates (Table 1), a nearly linear relationship was observed. Again the June 9 application d a te, on which th e highest airborne pesticide concentrations w ere m easured, also had th e highest mean wind speed. The July 21 application d ate, on which the low est airborne pesticide concentrations w ere m easured, had the low est mean wind speed (none d e tec te d , < 0.5 m s *). Phillips (1971) d em onstrated using wind tunnel experim ents th a t th e re was a m arked increase in th e ra te of loss of dieldrin deposits on glass su rfaces a t wind speeds of 3.2 km hr-1 (.89 m s_1) as com pared to still air. Dieldrin was deposited as a thin film in a m anner designed to prevent m echanical loss from wind. I fe e l th a t increased wind speed is responsible for increased loss in azinphosm ethyl deposits in th e orchard. Both v olatilization and wind erosion are m echanism s thought to be a ffe c te d by wind speed. What is not clear is w hether increased airborne loss of deposits due to increased wind speed should be re fle c te d in the increased airborne pesticide concentrations m easured, as increased wind speed, although rem oving m ore p esticide, should have a diluting e ffe c t on concentrations m easured in the air. It is possible th a t wind erosion of freshly deposited residues is responsible for the m arked difference betw een airborne residues m easured the afternoon of June 9 and those m easured th e afternoon of July 21. The combined e ffe c t of wind speed and solar radiation input m ust also be considered. F urther d ata collection under a v ariety of environm ental conditions (actu al and/or sim ulated) is needed to b e tte r describe the processes contributing to pesticide airborne loss. On sam pling dates following th e application d a te, more o ften than not, airborne pesticide residue levels m easured a t 6.0 m eters exceeded those m easured a t 0.5 or 1.0 m eter. Six m eters is tw ice the average tre e height. One wonders how far above the orchard airborne residue levels are m easurable. The d a ta recorded did not allow estim ation of horizontal airborne residue flux by 122 d ire c t m ethods. Estim ation of e ith er pesticide v e rtic a l airborne flux or mass horizontal flow of pesticide vapor by aerodynam ic m ethods requires knowledge of wind speed profiles. R eliable in te rp re ta tio n of these profiles requires m inim al flu ctu atio n in atm ospheric turbulence. While this criterio n may be satisfied by uniform stands of short crops (Parm ele e t al., 1972; Taylor e t al., 1977), the orchard profile does not satisfy it. An in d irect e stim ate utilizing both deposit p esticide residues and airborne residues is presented as an altern ativ e approach. H artley (1969) proposed th a t the v o latilization ra te (F^) of one pesticide, from a non-adsorbing su rfa ce , could be p red icted from th a t (F ) of another a pesticide, given the vapor pressures (P) and m olecular weights (m) of th e two compounds, by th e following equation: V mb>1/2 F, = —--- ——T“7K • F a (1) a This relationship assum es th a t vapor flux is solely a function of m olecular diffusion through the stag n an t a ir closely surounding the su rface. Taylor (1978) points out th a t when exam ining vapor flux from plant surfaces, this equation is valid only during th e tim e when th ere is com plete coverage of th e plant su rface, i.e., when th e pesticide vapor pressure is not reduced due to adsorption and is in e ffe c t volatilizing from itse lf. As the deposits becom e increasingly sm aller the vapor flux is no longer s tric tly a function of m olecular diffusion, but also a function of the degree of adsorption to the plant su rface. In addition residue may p e n e tra te le a f tissues or accu m u late in cracks and fissures in the epicu ticu lar wax su p erstru ctu re or le af specialized stru c tu re s. With these lim itations in mind, equation (1) was used to p red ict the airborne loss of azinphosm ethyl from the orchard using dieldrin vapor flux over orchard grass determ ined by Taylor e t al. (1977), along with vapor pressure d a ta found in the 123 lite ra tu re . Using the aerodynam ic m odel described by Parm ele e t al. (1972), Taylor e t al. (1977) e stim ated the dieldrin vapor flux to be 80.4 g /h a /h r during th e 1300 to 1500 hours sam pling period 3.5 hours following application. The vapor pressure rep o rted dieldrin is 2.6 x 10-6 mm Hg a t 20°C (Spencer and C laith, 1969). Only an upper lim it for th e vapor pressure of azinphosm ethyl has been reported; < 7.5 x 10 ”6 mm Hg (no te m p e ra tu re given, Schrader, 1963). _7 The vapor pressure of the ethoxy analog, azinphos-ethyl is given as 2.2 x 10 mm Hg a t 20°C by Spencer (1968). To b e tte r e stim a te the vapor pressure of azinphosm ethyl from th e d a ta available, th e relationship betw een the possible c ry stal stru c tu re s of these tw o analogs was exam ined. Rohrbaugh e t al. (1976) exam ined, using x-ray diffractio n crystallography, th e c ry stal and m olecular s tru c tu re of azinphosm ethyl. The unit cell stereograph depicted indicates th a t c ry stal stru c tu re is d ic ta ted , in p a rt, by overlapping of th e nearly planar ring system s and by interm olecular repulsion e ffe c ts of the m ethoxy groups; the repulsion working against the packing forces of th e cry stal. The increased size of th e ethoxy group of azinphos-ethyl should resu lt in w eaker packing forces by increasing the allowed distance betw een m olecules. This e ffe c t is in turn re fle c te d in the h eat of sublim ation, which varies d irectly w ith th e m agnitude of th e packing forces, and the observed vapor pressure, which varies inversely with the m agnitude of the h eat of sublim ation (Barrow, 1966). In addition, the m elting point of azinphosm ethyl (73-74°C) is approxim ately 18° higher than azinphosm ethyl (56°C), indicating stro n g er packing forces and th e re fo re a lower vapor pressure. Based on th ese observations, it was estim ate d th a t th e vapor pressure of azinphosm ethyl was slightly less than 2.2 x 10 -7 o mm Hg a t 20 C, the vapor pressure of its ethoxy analog. This value was used in equation (1) as P^, along w ith the dieldrin vapor pressure of 2.6 x 10 ® mm Hg a t 20°C as P , and 80.4 g ha-1 hr-1 , th e field diurnal vapor flux (1300-1500 hr) as Ffl, to give an 124 upper lim it e stim ate for azinphosm ethyl diurnal airborne flux (1300-1500 hr) of 6.2 g ha-1 hr- 1 . d a ta , th e To exam ine the sen sitiv ity of equation (1) to vapor pressure value of 1.0 x 10 -7 mm Hg, re p o rte d as th e maximum for azinphosm ethyl vapor pressure by Schrader (1963), was also used, resulting in an e stim ate d azinphosm ethyl diurnal airborne flux of 2.82 g ha-1 hr-1 . These p redicted values for azinphosm ethyl diurnal flux (1300-1500 hr) from th e orchard a re by no means precise e stim ate s, but ra th e r crude approxim ations, due in p art to the general natu re of the azinphosm ethyl vapor pressure d a ta. O ther sources of variation are the d ifferences in application ra te (5.6 kg ha 1 a.i. for dieldrin vs. .65 kg ha~* a.i. for azinphosm ethyl), and the increased influence of wind th a t tre e deposits may receive as com pared to grass deposits. In theory, if th e re was com plete coverage in both experim ents the application ra te should have little e ffe c t, as th e vapor flux is independent of th e am ount of pesticide applied. As both pesticides were applied as w ater base sprays, which are deposited on plant surfaces as droplets, com plete coverage is doubtful (Taylor, 1978). Under these "non-ideal" conditions, vapor flux is no longer independent of the am ount of pesticide present. As s ta te d e a rlie r, one explanation for th e dependency of vapor flux on pesticide deposit residues is the hypothesis th a t incom plete coverage a ffe c ts the ap p aren t vapor pressure of th e pesticide due to adsorptive e ffe c ts of the foliar surface. Depending on th e degree of coverage, p a rt of the pesticide deposit will be unencum bered by e x tern al adsorptive e ffe c ts and will ev ap o rate a t the maximum allow able ra te (governed only by m olecular diffusion through th e stag n an t layer); th e rem aining portion of th e pesticide deposit will volatilize a t some slow er r a te , depending on th e degree of adsorption. To approxim ate the e ffe c t of foliar adsorption (i.e., incom plete coverage) on th e vapor flux p redicted in equation (1), it was assum ed th a t the higher the application ra te the g re a te r the p ercen tag e of th e pesticide deposit th a t is 125 a ffe c te d by foliar adsorption. This is a necessarily sim ple approxim ation of a com plex phenomenon, as additional facto rs including form ulation, spray droplet size, clim atic conditions during application, and c h a ra c te ristic s of th e fo liar su rface may influence the s ta te of the pesticide deposit and the volatilization ra te (Ebling, 1963; H artley, 1969; Hull, 1970). With th ese lim itatio n s in mind, equation (1) was m odified in the following manner _ V " b >1/2 " " aw w here a 171 Db • ^ is the application ra te of pesticide a and Db is th e application ra te of pesticide Db* As 75% of the pesticide deposited in the orchard was applied to th e tre e s, tre e surface a re a must be considered when com paring airborne loss from the orchard w ith th a t of a grass field. T herefore th e application ra te of azinphosm ethyl was based not only on th e .263 ha of ground a re a but also th e .276 ha of tre e surface a re a. The average application ra te for th e first spray period of 1978 was 0.65 kg ha * a.i. Three fourths was applied to .276 ha tre e su rface area and one fourth to .263 ground su rface a re a. If the tre e and ground surface areas are w eighted by the proportion of th e p esticid e th ey receive and averaged, the adjusted application ra te is 0.63 kg ha- * a.i. This value was divided by 5.6 hg ha * a.i. of dieldrin applied to th e grass pastu re to give a Db/D a ra tio in equation (2) of .112. In using this ra tio as a crude approxim ation of foliar adsorptive e ffe c ts on vapor flux, it must also be assum ed th a t both application ra te s represent pesticide deposits th a t are less than or equal to com plete coverage. The azinphosm ethyl flux e stim ate s d eterm ined from equation (2) w ere 0.697 g ha 1 hr, using 2.2 x 10~^ mm Hg and 0.317 g ha-1 hr- * _7 using 1.0 x 10 mm Hg for the vapor pressure of azinphosm ethyl. D iurnal airborne loss during this sam ple period, for the rem aining d a tes sam ples, was estim ate d by assuming this loss is directly proportional to th e m easured 126 horizontal flux (g ■ m -2 ■s -1 ) a t 3.0 m e te rs, d eterm ined from airborne pesticide concentrations (Table 6) and wind speed m easurem ents a t 3.0 m eters taken during sam pling (Table 1). Daily airborne loss was estim ated on th e work of Taylor e t al. (1977), who rep o rted a m arked diurnal change in volatilization ra te s of dieldrin and heptachlor, when applied to orchard grass. Sim ilar diurnal variations w ere rep o rted by Taylor e t al. (1976) for soil incorporated dieldrin. These research ers m easured a peak flux early in th e afternoon with virtually no v o latilization m easured before 0500 or a fte r 2300 hr. Based on these findings, daily diurnal airborne loss was e stim ated assum ing peak flux occurred during the tw o-hour sam pling period betw een 1300 and 1500 hr, th a t no airborne loss occurred before 0500 or a fte r 2300 hr, and th a t th e loss ra te varied linearly betw een the end points and the peak. Integration of the area under the triangle form ed gives a daily airbone loss of 9.0 tim es th e peak hourly flux m easured betw een 1300 to 1500 hours. To e stim ate the loss occurring during th e afternoon of th e first day (from 1300 to 2300) one-half th e daily e stim ate , or 4.5 tim es th e peak flux, was used. Table 7 shows the e stim ated daily airborne loss of azinphosm ethyl from th e orchard for th e first spray period of th e 1978 season. Loss is shown both as an absolute loss in g/ha and also as a p ercentage of the to ta l dislodgeable residues and soil residues p resent on a given sam ple d a te. No daily loss e stim ate on a percentage basis was made for the application d ate as absolute loss shown is the estim ate d loss from 1300 to 2300 hr, ra th e r than th e e n tire day. These e stim ates suggest a m arked decline in the daily airborne loss over th e 18-day sam ple period. As a p a rt of th e p aram eterizatio n of th e model presented in P a rt II, first-o rd e r ra te constants w ere determ ined for the disappearance of azinphosm ethyl foliar dislodgeable residues and soil residues in each canopy and alley layer (accounting fo r m ovem ent). If these individual ra te constants are w eighted by th e proportion of th e to ta l dislodgeable residues and Table 7. Days A fter Application E stim ated Azinphosmethyl Daily Airborne Loss from the O rchard, F irst Spray Period of th e 1978 Season Mean H orizontal Flux a t 3.0 m During Sampling (1300-1500) (y g m s" ) E stim ated Lossa During Sampling g ha hr T b .jEstinjated Daily Loss g ha' % day Downwind Edge 0 3 6 13 18 10.6 2 .4 0 .4 0 .4 1 .697 .158 .026 .026 2 .317 .072 .012 .012 1 .313C 1.42 0.24 0.24 2 1 .4 3 C 0.65 0.11 0.11 0 .4 0.24 0.43 0.34 0.11 0.20 .317 .157 0.33 .014 .011 3.1 3 e 3.11 0.65 0.27 0.22 1.4 3 c 1.41 0.30 0.12 0.10 1.51 0.34 0.27 0.39 0.68 0.16 0.12 0.18 1 2 C enter 0 3 6 13 18 o 11.5 5 .7 1.2 0.5 0 .4 .697 .345 .073 .030 .024 _ f7 Q Based on th e estim ated airborne loss a t peak flux (1300-1500 hr) on day 0, calculated using 2.2 x 10 mm Hg a t 20 C (column 1 and 1.0 x 10 mm Hg (column 2) for the V.P. of azinphosm ethyl in equation (2). Diurnal loss for the rem aining sam ple dates was estim ated by assuming this loss is directly proportional to the m easured horizontal flux a t b3m. Daily diurnal loss was estim ated assuming peak flux during the 2 hr sampling period (1300-1500 hr), th a t no airborne loss occurred before 05^0 and a fte r 2300, and th a t th e loss ra te varied linearly betw een the end points and th e peak. The daily loss in % da" was determ ined by dividing the daily loss in gram s by th e to ta l g ha dislodgeable residues (including soil residues) rem aining. O ne-half day e stim ate. 128 soil residues which they a tte n u a te , the average a tten u atio n ra te is 5.1% /day. The estim ated air loss ra te on day 3, based on m easurem ents made a t the downwind edge (Table 7, Column 1), is 1.86% day 1. This loss ra te rep resen ts approxim ately 36% of the to ta l a tte n u a tio n . By day 6 a ir loss is 15%, day 13 it is 5%, and on day 18 air loss rep resen ts 8% of th e to ta l atte n u atio n . These observations indicate th a t initially airborne loss plays a m ajor role in the overall atten u atio n dislodgeable residues, how ever, this contribution drops off rapidly betw een 3 and 6 days following application. The change in the ra te of airborne loss is believed to be a function of th e s ta te of residue deposits w ith th e loosely bound residues being rapidly lo st to the atm osphere soon a fte r application. The rem aining m ore tig htly adsorbed and/or p ro te c te d dislodgeable residue deposits volatilize a t a slow er ra te . The e ffe c ts of rain fall on th e s ta te of the pesticide deposits within th e orchard m ust also be considered (see P a rt I). C ertainly rain fall events th a t occurred during th e first spray period (see Figure 5) had an e ffe c t on the pesticide airborne co n cen tratio n s and deposit residues m easured. Again, loosely bound residues should be m ost susceptible to rain fall e ffe c ts redistributing these residues to o th er p arts of th e orchard or as runoff leaving th e orchard. These red istrib u ted residues may no longer be loosely bound, as they may be m ore evently distribued over a larg er foliar su rface than a t application. By c o n tra st, residues th a t w ere initially d istrib u ted to p ro tected plant areas could be re d istrib u te d w ith ra in fa ll as to increase th e ir susceptibility to airborne loss. R ainfall may also enhance volatilization if the soil m oisture level is low prior to a ra in fa ll ev en t, th e increased soil w ater may displace adsorbed pesticide, increasing its e ffe c tiv e vapor pressure a t the soil surface (Spencer e t al., 1973). Heavy rains on u n satu rated soils may also move initially displaced pesticide to lower soil depths (Spencer and C laith, 1977; Helling e t al., 1971). 129 In the model described in P a rt II, m ovem ent is sep arated from overall a tte n u atio n only, and this a tten u atio n process is assum ed to be first o rder. Modeling pesticide disappearance using first-o rd e r kinetics assum ed th a t the individual atten u atio n processes are first order and stric tly additive, giving an exponential decay of pesticide residues w ith tim e. As early as 1955, G unther and Blinn proposed th a t the first-o rd e r loss curve was an approxim ation of a bilinear or trilin ea r loss curve. This hypothesis was again proposed by G unther e t al. (1969, 1977), by Hill (1971), and Van Dyk (1974, 1976). Taylor e t al. (1977) indicated th a t the disappearance of heptachlor and dieldrin from orchard grass and soil could be a ttrib u te d solely to airborne loss. They rep resen ted pesticide loss as a bilinear process w ith two regression equations; one for days 1 to 5 and another for days 5 to 107. Stam per e t al. (1979) proposed an a lte rn a tiv e to firs torder kinetics for foliar applied insecticides, showing for a num ber of foliar applied organophosphate insecticides, th a t In co n cen tratio n versus In tim e gave a linear relationship w ith a b e tte r co rrelatio n co efficien t than the linear relationship established by plotting In co n cen tratio n versus tim e (which indicates first-o rd e r kinetics). The author suggests the reason the In - In plots give a b e tte r linear relationship is th a t one form of th e equation describing th e fitte d line is in agreem ent with equations describing m olecular diffusion from sm all volumes. Their solution is in terestin g , but th e use of In tim e would not be com patible with the algorithm s used in th e model presented in P a rt II. A third approach to modeling airborne loss was proposed by Phillips (1971). He suggested th a t a double exponential equation T = A -k t + B k,t e e (3) may apply when tw o volatilization processes a re occurring sim ultaneously a t d ifferen t ra te s. This model was la te r proposed by Popendorf and Leffingw ell 130 (1978) to explain the observed bilinear loss of parathion dislodgeable residues on citru s foliage. The approach used here is an extension of th e double exponential equation. A t the present, residues deposited in th e orchard are classified as e ith e r dislodgeable, su rfa ce -p e n e tra ted , or soil. This classification is based on an aly tical procedures for th e recovery of th ese residues as outlined in th e experim ental section. Dislodgeable residues a re p ictu red as residing on the foliar su rface, w hereas su rfa c e -p e n e tra te d residues a re em bedded on or in th e cu ticular m atrix. Further p enetratio n , resu ltin g in tissue bound residues, is also thought to occur (Wieneke and Steffens, 1974). To m odel airborne loss, it was assumed th a t a portion of the dislodgeable residues are loosely bound. The loss of th ese residues to the atm osphere th e re fo re being only slightly a ffe c te d by the adsorptive e ffe c ts of the le a f surfaces. The rem aining dislodgeable residues are more tightly bound and volatilize m ore slowly. A sim ilar th eo ry for th e d ecrease in the daily airborne loss ra te of foliar applied pesticides is p resen ted by Taylor (1978). V olatilization of su rfa c e -p e n e tra te d and tissue-bound residues is thought to be negligible and the kinetics of the disappearance of th ese residues is not tre a te d here. Making these assum ptions, p esticid e airborne loss from fo liar su rfaces is represented in the d ifferen tia l form as: = - a 1D(t) - a 2L(t) - a gT(t) (4) w here t is tim e in days, D is dislodgeable residues, L is loosely bound residues, T is to ta l residues (dislodgeable plus loosely bound residues, and excluding p e n etrate d residues), a^ is th e air loss ra te of th e tig h tly bound dislodgeable residues, a 2 is the air loss ra te of loosely bound residues, and a^ is the atten u atio n ra te of the to ta l residues (other than air loss and assum ed to be independent of binding s ta te ). The co n stan t a g is thought to be prim arily a 131 function of degradation a t th e foliar su rface and p en etratio n to subsurface tissues. As T = D + L, equation (4) can be reduced to: -(a x + a 3)D(t) - (a2 + a 3)L(t) (5) and th erefo re: (6) TfF = - (al + a 3)D(t) and (7) As loosely bound residues are w hat th e ir nam e im plies, these residues are m ore easily lost to wind erosion and volatilize a t a ra te governed only by diffusion across the stag n an t air layer. C onsequently it is assum ed th a t airborne loss of th e se residues is responsible for the observed early rapid decay of to ta l residues; this loss is rep resen ted by G unther e t al. (1958) and more recen tly by Popendorf and Leffingw ell (1978) as the first phase of a bilinear loss p a tte rn . With the disappearance of these residues from th e fo liar su rface th e ra te of decline of rem aining dislodgeable residues is prim arily responsible for th e observed ra te . The loss ra te of these m ore tig h tly bound residues is determ ined by th e com bination of airborne loss (a^), degradation and penetration (a3) of the rem aining dislodgeable residues. This rep resen ts th e second phase of an observed bilinear loss p a tte rn . distinguish betw een As cu rren t sam pling or analytical techniques cannot loosely bound and m ore tig h tly adhered dislodgeable residues, D, L and the ra te constants a^, a 2 and a 3 w ere determ ined indirectly. In determ ining th e tw o r a te constants in th e double exponential equation, Popendorf and Leffingw ell (1978) used dislodgeable residue d a ta and a constrained optim ization procedure th a t com pared predicted and observed values. As the hypothesized mechanism for the observed bilinear loss of foliar 132 applied pesticides presented here is based on the assum ption th a t early pesticide disappearance is prim arily due to airborne loss, th e change in th e estim ated daily airborne loss ra te was used to e stim ate when th e loosely bound residue deposits approached zero. If the daily a ir loss ra te s (average of c e n te r and downwind edge e stim ate s in Table 7, Column 1) are exam ined, it is noted th a t th e re is a sharp d ecrease in the ra te s betw een day 3 (1.69%) and day 6 (0.54%). The average ra te and standard deviation for days 6, 13, and 18 is 0.40% + 0.18%. It was th erefo re assumed th a t the loosely bound residues disappeared betw een 3 and 6 days following application and th a t the average daily air loss ra te of 0.40% rep resen ted the air loss ra te (a^) of tightly bound dislodgeable residues. The to ta l loss ra te (a1 + a^) of th e tightly bound dislodgeable residues was determ ined from a least squares linear regression of log g ha * to ta l residues versus tim e since applications, for days 3, 6, and 13 of the first spray period. T otal residues included foliar dislodgeable residues and soil residues. Soil was included to give a conservative e stim ate w ith re sp ec t to m ovem ent of residues present in the orchard following rain fall ev en ts. Day 18 was o m itted from this analysis as it was fe lt th a t th e 51.3 mm (2.1") of rain fall the orchard received on day 16 red istrib u ted the residues in such a m anner as to not allow a conservative e stim ate of residues p resent. From this analysis th e daily loss ra te (a^ + a^) was determ ined to be 7.5% (r = .958). This ra te is considerably fa s te r than th e daily o atten u atio n ra te of 4.2% (r = .985), d eterm ined from to ta l residues for days four to 20 of the second spray period of the 1978 season. Only one rain fall event occurred during th e second spray period (5.1 mm on day 10), w hereas 17.3 mm of rain fall fell just prior to day 3, and 16.8 mm fell betw een days 3 and 13 of the first spray period (see Table 1 and Figure 4). This would suggest th a t th e disappearance of these residues was influenced by rainfall (although oth er environm ental param eters such as wind speed, solar radiation, hum idity, or som e 133 com bination of th e above also may have contributed). The results may have been th a t the pesticide residues m easured during th e first spray period w ere not conservative (i.e., m ovem ent of pesticid e to unsam pled soil depths, pesticide carry -in with m oisture uptake by litte r and foliar surfaces, increasing the p e n e tra te d residue pool). Also, as s ta te d e arlier, rain fall may have red istrib u ted foliar residues as to allow increased v o latilizatio n . An influx of m oisture a t the soil su rface may have also caused an increase in v o latilization. That this may be th e case is suggested by the in crease in the e stim ated average daily a ir loss ra te betw een day 13 (0.26%) and day 18 (0.41%) (Table 7, Column 1). U ntil th e e ffe c ts of rainfall and o th er clim atic conditions on pesticide disappearance can be adequately d iffe re n tia te d , uniqueness of th e daily atten u atio n ra te (a^ + a^) to this s e t of environm ental conditions m ust be assum ed. Given a^ to 0.40% and a.^ + a the am ount and ra te of airborne loss of the loosely bound residues, w ere estim ated using an optim ization routine sim ilar to th a t employed by Popendorf and Leffingw ell (1978). Equation (3) was rearran g ed to solve for T (to ta l residues) a t som e sam ple date L: T(k.) = D(ki_1) • (1 - (ax + a 3» L(k._i) • ( l - ( a 2 + a 3)) (8) As T is known a t sam ple d a tes k. = 0, 3, 6, 13, and a^ and a 3 have previously been e stim ate d , values for L and a 2 w ere trie d in equation 8 to give a minimum value fo r th e te s t sta tistic : T. - (L. + Dj)2 T. (9) 134 Using this s ta tis tic , it was estim ate d th a t 37.5% of azinphosm ethyl dislodgeable residues, in itially deposited in the orchard, w ere loosely bound. The daily a ir loss ra te (a2) of th e loosely bound frac tio n was e stim a te d to be 90% day This indicates a rapid loss of approxim ately one th ird of the residues to the atm osphere, resulting in the disappearance of this frac tio n in approxim ately 3 days. It is also suggested th a t loss of the rem aining dislodgeable residues is governed by a much slower a ir loss ra te and a tte n u a tio n a t th e foliar su rface. Figure 5 shows the solution to equation (5) as a line through the observed residue values (to tal dislodgeable residues and soil residues) for the first spray period of th e 1978 season. The proportion of azinphosm ethyl deposit residues estim ated to be loosely bound and the ra te of loss of this frac tio n is thought to be a function of facto rs associated with c h arac te ristic s. application, environm ental conditions, and orchard This is suggested by th e d ifferen ce in th e decline of to ta l dislodgeable and soil residues betw een th e first and second spray periods as shown in Figures 5 and 6. P a ra m e teriza tio n o f equation (4) using th e residue d a ta from the second spray resu lted in an e stim a te of 7.7% fo r the loosely bound frac tio n (L) with an estim ate d daily a ir loss ra te (a2) of 30% day * for this frac atio n . The ra te constant a^ + a^ was d eterm ined from linear regression of th e to ta l dislodgeable and soil residues for days 4 to 20 to be 4.2% day -1 (r 2 = .985). The air loss ra te (a2) o f the dislodgeable fractio n (D) was assumed to be th e sam e as determ ined for th e first spray period, 0.4% day The lack of an in itial period of rapid loss during th e second spray period may have been due solely to th e lesser am ount of rain fall receiv ed , as com pared to th e first spray period. That rainfall may a ffe c t airborne loss has been discussed earlier. In addition, rew ettin g of the residue deposit may enhance p en etratio n and plant uptake (Hull, 1970; Bukovac, 1976), resulting in an increase in the observed loss 135 Figure 5 Decline in T otal A zinphosm ethyl Dislodgeable and Soil Residues M easured During th e F irst Spray Period of the 1978 Season 136 RAINFALL (mm) 500 + 17.3 * J ♦ ♦ 51.3 2.6 ♦ 3.0 11.2 400 300 200 g/ha 100 3 6 12 9 DAYS 15 18 21 137 Figure 6 D ecline in T otal A zinphosm ethyl Dislodgeable and Soil R esidues M easured During th e Second Spray Period of the 1978 Season 138 RAINFALL (mm) 500 5 .6 5.1 400 300 200 g/ha 100 DAYS 139 ra te of dislodgeable residues. The e ffe c ts of o th er environm ental p aram eters (i.e., solar radiation, am bient te m p e ra tu re , re la tiv e hum idity, wind speed and atm ospheric stability) should also be considered. To b e tte r understand th e influence of those fa c to rs associated w ith the form ulated pesticide, its applicatio n , and environm ental conditions on th e a tten u atio n of residue deposits in orchards, a more thorough sam pling program of both airborne and deposit residue is required. The p esticid e should be applied and its disappearance m onitored under a v ariety of properly c h aracterized "natural conditions." The m ulti-com ponent k in etic model presented here provides a tool for investigating those p aram eters th a t influence pesticide airborne loss and the overall a tte n u a tio n of deposit residues. 140 Figure 7 Daily T em perature Range, 1978 Season 100 90 80 70 60 50 40 - 30 June July A u g u st 142 Figure 8 Daily Minimum and Maximum R elativ e Hum idity, 1978 Season 142 Figure 8 Daily Minimum and Maximum R elativ e Hum idity, 1978 Season 143 June July August PART IV ASSESSMENT OF THE ATTENUATION AND MOVEMENT OF AZINPHOSMETHYL IN A MICHIGAN APPLE ORCHARD ECOSYSTEM: FURTHER MODEL DEVELOPMENT INTRODUCTION It would be unrealistic to believe th a t a model of pesticide fa te in any te rre s tria l ecosystem , driven by tim e and ra in fa ll conditions only, would be capable of explaining all the variation observed in the field d a ta. To b e tte r explain the variation observed, both m ovem ent and a tten u atio n should be decom posed to represent th e ir com ponent physical, chem ical, and biological processes, each a function of the re le v an t environm ental p aram eters (i.e., solar radiation, tem p eratu re, hum idity, wind speed, atm ospheric stab ility , etc.). The conceptual model shown in Figure 1 o f P a rt II was developed to rep resen t those processes which d eterm ine the distribution and fa te of a pesticide applied to an orchard ecosystem . The in itia l m ath em atical form of the model describes azinphosm ethyl fa te in term s of a tten u atio n and m ovem ent, based only on field d ata from th re e seasons. The atten u atio n ra te constants determ ined represent the com bined e ffe c ts of a num ber of individual processes as indicated in the conceptual model. These processes a re thought to be individually influenced by d ifferen t com binations of environm ental p aram eters. In addition, because these processes occur sim ultaneously, th e use of field experim ents to study th e ir individual contribution to the overall atten u atio n ra te is difficult. 144 145 Laboratory experim ents, designed to isolate and study individual processes under controlled environm ental conditions, are a t p resen t th e p referred m ethod for studying environm ental fa te . Such experim ents, while giving th e research er much g re a te r control of the variables involved, o ften do not give results th a t can be extrap o lated to the re a l world. However, much progress has been made re c en tly in determ ining ra te constan ts for a num ber of atten u atio n processes under various sets of environm ental conditions (Zepp e t al., 1975; Sm ith e t al., 1977; Freed e t al., 1979). One approach to the developm ent of a m ore precise model of pesticide fa te would employ both in situ field m easurem ents and d a ta from laboratory experim ents. L aboratory experim ents should be designed to isolate and exam ine th e influence of sele c ted environm ental p a ram eters on the individual processes th a t in com bination resu lt in the observed field atte n u atio n ra te . In addition, in situ field ra te d ata should be co llected under a v ariety of "n atu ral conditions" to allow the p a rtia l isolation of individual environm ental e ffe c ts. F u rth er model developm ent and ex perim ental design should address th e following questions: (1) How are various loss processes influenced by individual environm ental p aram eters, and (2) What is the contribution of each process to the overall a tten u atio n ra te under a given s e t of environm ental conditions. In P a rt n, a m atrix A of daily a tten u atio n ra te s of azinphosm ethyl in various orchard layers was developed. The orchard is tre a te d as four v e rtic a l layers (tre e , grass-broadleaves, litte r-m o ss, and soil) by two horizontal areas (canopy, alley). Seven regions (no alley tre e region exists) appear in each of the m atrices and vectors below, beginning w ith canopy tre e s and proceeding to alley soil. Each diagonal elem ent A~ rep resen ts th e fractio n of azinphosm ethyl lost from region i in one day, excluding any m ovem ent to or from another region 146 (which is accounted for in m atrices P, L, and H, th e nonrain m ovem ent, light rain m ovem ent, and heavy rain m ovem ent m atrices, respectively). To investigate th e influence of rain fall and o th e r environm ental p a ram eters of pesticide loss, the atten u atio n m atrix A may be decom posed into th e diagonal m atrices: A=A +A +A +A +A p c v m u w here the summand m atrices A^, Ae , Ay, Am , and Ay rep resen t photolysis, chem ical degradation, v o latilzatio n , m icrobial degradation, and plant uptake, respectively, and are called the atten u atio n com ponent m atrices. Each of these m atrices is a 7 x 7 diagonal m atrix with one non-zero e n try fo r each region. Each atten u atio n com ponent m atrix is decom posable into a diagonal m atrix of co n stan ts (the ra te s under "standard conditions") and a s e t of functions which modify those ra te s based on environm ental conditions. The "standard" ra te s for each process should sum to the field-determ ined values. To exam ine th e feasib ility of this approach to fu rth er model developm ent, th e relationship betw een azinphosm ethyl degradation kinetics and a num ber of environm ental p a ram eters was determ ined from laboratory d a ta and the resu lts of oth er laboratory experim ents rep o rted in the lite ra tu re . The v o latilizatio n kinetics of azinphosm ethyl were discussed in P a rt III. P esticide degradative mechanism s have been review ed by a num ber of research ers (Ebling e t al., 1963; Crosby, 1973; Leonard e t al., 1976). P esticide degradation has been traditionally divided into th re e major areas: chem ical, photochem ical, and biological. Pesticides may undergo a num ber of chem ical transform ations in the environm ent to include: hydrolysis and o th e r nucleophilic reactio n s, oxidation, isom erization, reduction and fre e rad ical reactio n s (Goring e t al., 1975). For the organophosphates, including azinphosm ethyl, hydrolysis and oxidation are thought to be th e m ost commonly occurring. Biological 147 degradation includes plant and anim al uptake and m etabolism . M icrobial degradation is a major pathw ay for the disappearance of many pesticides in the soil (K earny and Helling, 1969). The diverse m icrobial populations of m ost soils a re capable of degrading pesticides w ith little difficulty, e ith e r by adaption, or m ore commonly, by co-m etabolism (M atsum ura, 1975). M icrobial degradation on plant su rfaces must not be excluded in assessing possible causes of pesticide disappearance (Wieneke and S teffens, 1975). Photochem ical degradation of pesticid es has been d em onstrated in w ater and a ir and on soil and foliar surfaces (Crosby, 1969; Nilles and Zabik, 1975; Liang and L ichtenstein, 1976; Zepp and C line, 1977). C hem ical D egradation As w ith the overall a tte n u a tio n ra te , the ra te s for the individual processes a re assum ed to be first-o rd e r or pseudo first-o rd e r. C hem ical degradation of azinphosm ethyl is assum ed to occur prim arily by hydrolysis, but oxidation is also possible (Eto, 1974). Oxidation may occur in all regions. The oxidation ra te is rep resen ted by DP Dt (oxidation) = , der R ox =K ox C The m ost likely oxidation pathw ay is through th e reactio n w ith fre e radicals, assum ing an excess of free radicals available for in teractio n (r generation > Kqx) then Kqx is ra te lim iting and th e re a ctio n is pseudo first-o rd e r (Smith e t al., 1977). Photooxidation, as a resu lt of reactio n with photochem ically form ed fre e radicals, the oxygen trip le t diradical, or th e more re a ctiv e singlet oxygen, may be m ore responsible for many pesticide non-biological oxidations (Crosby, 1973; Khan, 1976). Soil fre e radicals may also be im portant in the oxidation of pesticid es in this medium (Plim m er e t al., 1967; A rm strong and Konrad, 1974). Spear e t al. (1978) indicated th a t paraoxon production may be re la te d to both ozone and dust levels on citru s in c e n tra l C alifornia. Spencer e t al. (1975) also 148 noted paraoxon form ation on dust and dry soil beneath citru s tre e s in southern C alifornia. Oxidation of azinphosm ethyl dislodgeable residues on southern C alifornia citru s foliage has also been in d icated (Gunther e t al., 1977). However, azinphosm ethyl-oxon levels never exceeded 1.0% of the azinphosm ethyl present. The oxon form ed was m ore stab le, but dissipated rapidly following rain fall. The likelihood th a t oxidation would co n trib u te significantly to th e degradation of azinphosm ethyl during th e rela tiv e ly w et and humid sum m er months norm ally experienced in the te m p e ra te e aste rn U nited S tates is doubtful. Hydrolysis has been shown to be an im portant mechanism of organophosphate degradation in both soil and aqueous environm ents (Freed e t al., 1979). Hydrolysis resulting from re a ctio n w ith m oisture on foliar su rfaces must also be considered. A discussion of hydrolytic mechanism s for the organophosphates in w ater can be found in Faust and Gomaa (1972) and Smith e t al. (1977). The reactio n is a function of pH, and can be e ith e r n eu tral, acid, or base-cataly zed . Again, th e re a ctin g species (HgO, H+, OH- ) w ere assum ed to be in excess of the pesticide, and the reactio n first order, a t a given pH. Azinphosm ethyl aqueous hydrolysis as a function of pH was determ ined by the procedure of Freed e t al. (1979). Buffers used w ere as follows: pH 1.0, 0.01 m KC1 and 0.01 m HC1; pH 3.0 and 5.0, 0.01 m potassium hydrogen p h th alate and 0.01 m NaOH; pH 7.0, 7.5, 8.0, 8.5, 9.0, 0.01 m TRIS and 0.01 m HC1. A zinphosm ethyl co n cen tratio n was d eterm ined by analysis of residual p aren t compound by GLC (see P a rt I). F irst-o rd er ra te constants a t 25°C were determ ined from linear regression of log concentrations versus tim e, over a 20day incubation period. Hydrolysis ra te constants a t pH 1, 3, 5, and 7 averaged 0.8% day * (standard deviation, 0.1%). The observed stab ility under n e u tra l and acid conditions is in agreem ent w ith the observations of Liang and L ichtenstein (1972). Faust and Gomaa (1972) re p o rt th a t many organophosphates are stable 149 under acid conditions. (A zinphosm ethyl hydrolysis as a function of pH is shown in Figure 1). The relationship betw een pH and the b ase-cataly zed reactio n ra te (pH 7.5-9.0) is rep resen ted by th e linear regression equation: Kh = 0.095 pH - .713 (r2 = .996) w here (1) is the aqueous hydrolysis ra te co n stan t (tim e- *). The ra te of hydrolysis as a function of tem p eratu re was represented using the Arrhenius equation: Kh = Ae-E a/R T w here (2) is th e aqueous hydrolysis ra te co n stan t (tim e-1 ) a t te m p eratu re T, A is a constant depending on the chem ical and oth er non-therm al facto rs, R is the gas constant, and Ea is th e energy of a ctiv atio n . If is determ ined over some te m p e ratu re range, Ea can be calcu lated from the slope of the line, for a plot of log versus 1/T. To e stim a te te m p e ratu re e ffe c ts on azinphosm ethyl hydrolysis the d a ta of Liang and L ichtenstein (1972) was analyzed using equation (2) as shown in Figure 2. The value determ ined for Ea, over the te m p eratu re range 5-50°C, was 12.5 K cal/m ole. This is an approxim ate value and will vary w ith pH and th e te m p e ratu re range used. Buffer com position and stren g th may also influence hydrolysis ra te s and associated values for Ea (Smith e t al., 1977). Hydrolysis of azinphosm ethyl is assum ed to be im p o rtan t in all regions. Although little is known about reactio n s occurring on foliar surfaces, it is assumed th a t, due to tran sp iratio n , th e re is an environm ent with su fficien t m oisture to allow hydrolysis to occur (Wieneke and S teffens, 1974). A n eu tral pH is assumed in all but th e soil regions. Hydrolysis is tre a te d as a function of air tem p eratu re only. This again is an approxim ation as tem p eratu res a t the foliar surface will o ften exceed th e am bient air te m p e ratu re. In th e soil regions hydrolysis is tre a te d as a function of both soil tem p eratu re and pH. Properties 150 Figure 1 Azinphosm ethyl Hydrolysis R ate C onstant Versus pH .1 6 .14 .12 .10 RATE CONSTANT DAY-1 .08 .06 .04 .02 152 of th e soil which influence hydrolysis are discussed by Freed e t al. (1979). Adsorption to clay and organic m a tte r is thought to play an im portant role in degradation, as adsorbed organophosphates may be p ro te c ted from hydrolysis or, in som e cases, may resu lt in increased reactio n ra te s due to surface catalysis (Crosby, 1970). The use of th e Arrhenius equation (2) to rep resen t the e ffe c ts of te m p e ra tu re on hydrolysis and oth er non-biological degradative m echanisms in th e soil environm ent is discussed in H am aker (1972). This author suggests th a t th e heterogenous n atu re of soils as a reactio n medium may not allow the use of equation (2) as it has traditio n ally been applied to reactio n s in homogenous solutions. This is indicated by the fa c t th a t the distribution co efficien t, Kd (pesticide adsorbed/pesticide in solution), may change as the pesticide is tran sfo rm ed , and also by th e exotherm ic n a tu re of the adsorption process resulting in an equlibrium sh ift tow ards sorption to organic m a tte r with increasing te m p e ra tu re (Felsot and Daum, 1979). Because both these phenomena will influence the concenration of pesticide in th e soil solution and, th e o re tic ally , th e ra te of hydrolysis, this suggests th a t the "A" term in equation (2) is not a c o n stan t, but a function of both pesticide co ncentration and soil te m p e ra tu re . U ntil the influence of soil pro p erties on hydrolysis is b e tte r understood, and for th e purposes of this study, "A" is assum ed to be a constant. The d a ta of Yaron e t al. (1974) w ere used to determ ine the relationship betw een te m p e ra tu re and th e azinphosm ethyl hydrolysis ra te constant in soil. D egradation ra te s for w et soil (50% of satu ratio n ) incubated a t 6, 25, and 40 °C , w ere used to p a ra m ete riz e equation (2) as shown in Figure 3. The Ea determ ined from this d a ta was 13.5 K cal/m ole. The soil used in this study was a silty loam , pH 8.4, and < 1% organic m a tte r. Both pH and soil type should be considered when com paring ra te s determ ined from the d ata in this p articu lar study with o th e r research . Soil in th e orchard used in th e p resent study was a m a rle tte 153 Figure 2 A zinphosm ethyl T em perature Hydrolysis R a te C onstant Versus Air 154 0 .2 r - DEG RATE d a y -1 0 60 T E MP 155 sandy clay loam , pH 6.0, 56% sand, 20% silt, 24% clay, and 6.0% organic m a tte r in the top 10 cm . The results of degradation studies in the laboratory, using soil from th e orchard (m oisture c o n ten t 30% = 48% of satu ratio n ) ste riliz e d w ith sodium azide, fo rtified w ith 10 ppm azinphosm ethyl, and incubated a t 25 + 1°C for 20 days, showed no degradation. This is not surprising considering th e aeid pH and the clay and organic m a tte r co n ten t of the soil. M icrobial D egradation Although m icrobial degradation cannot be ruled out as contributing to the overall a tten u atio n of residues deposited on foliar su rfaces, m icrobial degradation is thought to be a major degradative pathw ay in the litter-m o ss and soil layers. In some instances, pesticides can be used as the sole food source of m icroorganism s, but more o ften they are co-m etabolized w ith other organics. If th e pesticide is used as a prim ary n u trien t by th e microorganism s of th e soil, then a lag period may be observed following application while th e soil m icroorganism s population adapts to the new food source. becom e shorter with successive This period may applications (H am aker, 1972). With co­ m etabolism , as long as the pesticide rep resen ts a sm all fractio n of the to ta l food source, no lag period should occur. This is the situation assum ed to be present in th e orchard soil and litte r , with regard to azinphosm ethyl biodegradation. The r a te of biological degradation will th erefo re vary with the to ta l available food source (i.e., organic m ater co n ten t of the soil), te m p e ratu re, and m oisture (H am aker, 1972). To determ ine the m icrobial degradation ra te , a 20-day incubation a t 25 + 1°C, of the non-sterilized orchard soil (m oisture co n ten t 30% = 48% o f saturation) fo rtifie d with 10 ppm azinphosm ethyl, was perform ed. The first-o rd e r ra te co n stan t, determ ined from linear regression of co n cen tratio n of azinphosm ethyl rem aining versus tim e, was 7.9% day -I 2 (r = .969). 156 Figure 3 A zinphosm ethyl T em perature Hydrolysis R a te C onstant Versus Soil 157 0 .2 DEG RATE d a y -1 0 50 TEMP 158 The influence of organic m a tte r co n ten t of the soil on m icrobial degradation of azinphosm ethyl was based on th e d a ta o f Iw ata e t al. (1975). Soils in this study were passed through a 100 mesh sieve and m oisture was added to 40% satu ratio n . incubated a t The soils w ere fo rtifie d a t 450 ppm azinphosm ethyl and 30°C. C h aracteristics of th ese soils and th e corresponding azinphosm ethyl degradation ra te s over a 20-day period are as follows. Table 1. Azinphosm ethyl D egradation when Incubated w ith Various Soilsa % Organic M atter Soil 1 2 3 4 5 0 .8 1.8 2.1 2 .3 6 .0 a M echanical Analysis, % Sand Silt Clay pH K b m r2 53.6 56.0 12.5 22.4 56.0 6 .9 7 .6 7.3 7 .3 6 .0 -.0 1 0 -.0 3 3 -.0 5 0 -.0 6 1 -.0 7 9 .992 .996 .991 .969 .969 31.0 33.0 50.7 34.5 20.0 15.4 11.0 36 .8 43.1 24 .0 a D ata for soils 1-4 rep o rted in Iw ata e t al. (1975). Soil 5 was tak en from th e .o rc h a rd used in the present study. Proportional daily loss determ ined from the d a ta of Iw ata e t al. (1975) (Soils 1-4) and the orchard soil used in th e present study (soil 5). Figure 4 shows a plot of percen t organic m a tte r versus th e azinphosm ethyl degradation ra te . D egradation is assum ed to be prim arily m icrobial, due to th e n e u tra l or acidic pH values for these soils. H ow ever, hydrolysis and other form s of chem ical degradation may also have co n trib u ted to th e observed ra te s derived from the d a ta of Iw ata e t al. (1975) A m odified from of the V erhulst-Pearl logistic equation (Pielou, 1969) was fit to th e d a ta to give th e following relationship: K m = .079 [1 + e-1,85(% ° ' m* “ l*87) i -1 1 The correlation of azinphosm ethyl degradation w ith a single soil fa c to r (organic m a tte r) must be in terp reted w ith som e caution, as the organic m a tte r 159 Figure 4 A zinphosm ethyl M icrobial D egradation Versus Soil P ercen t O rganic M atter 0.1 091 DEG RATE d a y "1 0 10 % O.M. 161 co n ten t may be c o rrelated to o th e r soil conditions effectin g degradation. High organic m a tte r is usually accom plished by a low soil pH, and adsorption has been shown to be positively re la te d to organic m a tte r co n ten t (Saltzm an e t al., 1972). Even though the d ata presented here shows a positive correlation of degradation ra te to organic m a tte r c o n te n t, a num ber of research ers have found th a t a t very high organic m a tte r contents of p e a t and muck soils, degradation is decreased, presum ably due to adsorption (Beynon e t al., 1966; H am aker, 1972; Kaufm an, 1964). In addition, the relationship betw een the d a ta of Iw ata e t al. (1975) and th a t of the present study m ust be viewed in light o f the d ifferen ces in in itial concentrations used. H am aker (1972) c ite s a num ber of studies which indicate th a t the degradation ra te , on a p ercen tag e basis, increases w ith decreasing concentration. The azinphosm ethyl m icrobial degradation ra te of 7.9% day- * determ ined using an in itia l co n cen tratio n 10 ppm, may have been low er if th e in itial concentration of 450 ppm, em ployed by Iw ata e t al., (1975) was used. T em perature differences betw een th e tw o studies are not thought to be significant, as indicated by th e te m p e ratu re relationship shown in Figure 5. The relationship betw een te m p e ratu re and soil m icrobial degradation was determ ined from the d a ta of Yaron e t al. (1974). The difference in degradation ra te s betw een sterile and n o n -sterile w et soil (50% of satu ratio n , incubated a t 6°, 25°, and 40°C, was used to determ ine this relationship. The d a ta was fit to the following polynomial equation: Km(t) = .00298T - .00005T2 - .013 w here Km is the m icrobial degradation ra te a t a given tem p eratu re, T. This fit shows the optimum te m p e ratu re to be approxim ately 30°C. This is in ag reem ent w ith the d ata of Day e t al. (1961) for degradation of am itrole in soil. This te m p e ratu re relationship is thought to re fe lc t a response to te m p e ratu re by th e microorganism population. Changes in population size, distribution, or adaption 162 Figure 5 A zinphosm ethyl M icrobial D egradation Versus Soil T em perature 163 0.5 i— DEG RATE d a y -1 0 TE MP 164 to degradation of th e chem ical m ay all be responsible for the observed te m p e ratu re e ffe c t. m echanism . The response may also be due in p a rt to an Arrhenius type A t te m p e ratu res above th e optim um (30°C) th e ra te no longer shows a positive relationship to te m p e ra tu re , possibly due to the h e a t lability of the m icrooganism s. Soil m oisture has also been shown to be an im portant fa c to r in soil degradation of pesticides. The d a ta assem bled by H am aker (1972) show much slow er ra te s in dry as com pared to m oist soils and th a t ra te s tend to level o ff a t higher m oisture levels 30% of satu ratio n ). A t satu ratio n levels ra te s o ften drop due to the lack o f oxygen necessary fo r aerobic degradation. pesticides are capable of being degraded mechanism m ust not be n eglected . satu ratio n levels only b riefly. anaerobic m icroorganism anaerobically, so this Many possible Soil m oisture in the orchard may rem ain a t Under these conditions th e developm ent of an population able to significantly a lte r pesticide c o ncentration is not likely. The d a ta of Yaron e t al. (1974) shows little or no degradation of azinphosm ethyl for both s te rile and n o n -sterile dry soils. Increased adsorption, due to the dry conditions, may be p artially responsible for the observed resu lts. A t 50% of satu ratio n , th e degradation ra te for the n o n -sterile soil a t 25°C was tw ice th a t of the s te rile soil. relationship betw een No o th er d a ta was found in the lite ra tu re on the azinphosm ethyl degradation and soil m oisture. As m icrobial degradation has been shown to be the prim ary pathw ay of am itro le loss in soils (Kearny and Helling, 1969), th e tre a tm e n t by H am aker (1972) of th e d a ta of Day e t al. (1961) for am itro le is presented as a crude e stim ate of the influence o f soil m oisture on azinphosm ethyl m icrobial degradation (Figure 6). The e ffe c t of both soil m oisture c o n ten t and te m p e ratu re on the degradation of a num ber of herbicides was exam ined by Walker (1974, 1976a, 165 Figure 6 A m itrole D egradation Versus Soil M oisture for Two Soil Types 100 O 80 % MAX. RATE 60 991 40 •V ista Sandy Loam O Chino Silt Loam 20 J L 10 20 30 40 50 60 70 % FIELD CAPACITY 80 _L 90 J 100 167 1976b, 1976c), Smith and Walker (1977), and Walker and Smith (1979). In these studies no distinction was made betw een chem ical and biological d egradative pathw ays. In all cases a positive co rrelatio n was observed betw een th e degradation ra te and both soil m oisture co n ten t and soil te m p e ratu re. These d a ta w ere used to develop and te s t a model for herbicide persistence. Assuming first-o rd e r kinetics, herbicide h alf-life as a function of soil m oisture was rep resen ted by th e following em pirically derived equation (Walker, 1974): H = am _ b t j w here H is th e herbicide h a lf-life, m is th e soil m oisture co n ten t, and a and b are co n stan ts. The A rrhenius equation was used to rep resen t the e ffe c t of te m p e ratu re on the degradation ra te . H erbicide persistence in th e field was e stim ated from these two equations using both lab o rato ry determ ined values and sim ulated seasonal and diurnal soil te m p e ratu re and m oisture flux (Walker, 1974; Walker and Barnes, 1981). The model was te ste d against herbicide loss when applied to bare soil, and th e model worked best when th e herbicide was incorporated. A pplicability to cropped fields and p articu larly perennial crops such as orchards lies in the ability of the model to rep resen t th e e ffe c ts of th e crop and/or ground cover on soil te m p e ratu re and m oisture flux. The model in its present s ta te does not have this capability. F u rth er developm ent to rep resen t the e ffe c t of soil m oisture and te m p e ratu re on chem ical and biological d egradative pathways individually would also be desirable. pH and soil te x tu re may also be im p o rtan t facto rs influencing m icrobial degradation. Soils with extrem es in pH will m ost likely have developed m icrobial populations adapted to these conditions. The influence of pH m ight be more im p o rtan t, how ever, if a soil am endm ent which d rastically a lte rs th e soil pH is used, thereby requiring adaption by the m icrobial com m unity. Soil tex tu re describes the soil aggregate size and th e pore space of th e soil, which may 168 influence the availability of m oisture and air to the m icrobial population. The in terrelatio n sh ip betw een all soil properties and the microorganism population m ust be considered in assessing pesticide degradation. Photodegradation Solar radiation is known to be im portant to the atten u atio n of pesticides in th e environm ent, as it supplies th erm al energy which influences pesticide v o latilization and the ra te of many d egradative reactions. The energy of th e photons may also be adsorbed by th e bond (electronic) energy of the m olecule, which m ay resu lt in transform ation. Much research has been done in the laboratory on the photolysis of pesticides and th e photoproduets form ed (Zabik e t al., 1976). Only recen tly have a tte m p ts been made to determ ine ra te s of photodegradation under n a tu ra l conditions. com puter modeling techniques to conditions (Zepp and Cline, 1977). The m ost promising approach uses ex trap o late laboratory findings to field These authors have a tte m p te d to p red ict photolysis of a num ber of compounds in th e aquatic environm ent. The study of photolysis in solution is p referred as th e hom ogeneity of this medium allows th e re sea rc h er to accu rately m onitor and vary the chem ical environm ent. Experim ents to determ ine th e photochem istry of compounds in the solid or sorbed s ta te , such as pesticides applied to th e soil and plant surfaces, are fa r less m anageable (Zabik and Ruzo, 1981). R elatively few studies have a tte m p te d to determ ine the ra te of photolysis on plant and soil surfaces under n atu ral conditions. One of th e m ajor d ifficu lties associated with this kind of study is d ifferen tia tio n betw een photodegradation, v o latilization, and m etabolism . In addition, the presence of sen sitizers and quenchers may g reatly a lte r th e ra te of photolysis determ ined in the absence of these substances. For exam ple, Liang and L ichtenstein (1976) rep o rted th a t following an eight-hour exposure to sunlight, 2.8% of the 14 C azinphosm ethyl applied to bean leaves was determ ined 169 to be the oxygen analog, w hereas no oxygen analog was found for applications to corn leaves or glass p lates. No oxygen analog was p resent in th e dark controls. The authors suggest th a t a com ponent of th e bean le a f may have enhanced th e form ation of azinphosm ethyl oxon in the presence of sunlight. N in ety -th ree, 88, and 94% of the radiocarbon was recovered from th e glass, corn, and bean le af dark controls. In all cases, the radiocarbon recovered was determ ined to be u naltered azinphosm ethyl. Sixty-six, 86, and 72% was recovered from th e glass, corn, and bean leaves following th e eight-hour sunlight exposure with 5.7, 8.9, and 4.3% of th a t recovered determ ined to be photoproducts, d ifferen tia l ex tractio n and thin layer chrom atography. based on Substantial loss to volatilization of azinphosm ethyl on glass and fo liar su rfaces as a resu lt of th e sunlight exposure is indicated. The g re a te r v o latilizatio n of the azinphosm ethyl on exposed surfaces, as com pared to th e co n tro l (covered w ith black cloth), may have been due in p a rt to higher te m p e ratu re and g re a te r air exchange a t the foliar surface. V olatilization o f photoproducts m ust also be considered. If it is assum ed th a t only a sm all am ount of the photoproducts are lost to volatilization, then a crude e stim ate of th e in itial ra te o f photolysis can be made from th e percent of photoproduets form ed. As the eight-hour exposure period (0900 to 1700 hr) used by Liang and L ichtenstein (1970) rep resen ts only a portion of th e daily solar radiation exposure during the sum m er months in N orth A m erica, the relationship betw een the photolysis ra te o f carbaryl and tim e of day (July) rep o rted by Zepp and Cline (1977) was used to e stim ate th e initial daily loss ra te . Based on th e relationship rep o rted by these authors i t was assum ed th a t th e peak ra te of photolysis occurred a t approxim ately 1300 hr, th a t no photolysis occurred before 0600 hr or a fte r 2000, and th a t th e loss ra te varied linearly betw een th e end points and the peak. Integrating the a rea under the trian g le form ed yields an e stim ate th a t 170 exposure to sunlight betw een 0900 and 1700 hr is responsible for approxim ately 82% of the daily loss due to photolysis, during th e sum m er months. The am ount of photoproduets form ed (estim ates of azinphosm ethyl loss due to photolysis during the eight-hour exposure), as a p ercen t of dose applied to each su rface, was adjusted accordingly to arriv e a t an e stim ate of azinphosm ethyl in itia l daily loss due to photolysis. The in itia l daily loss ra te s calcu lated w ere 6.9, 10.8, and 5.2% day- * for azinphosm ethyl on glass, corn, and bean leaves, resp ectiv ely . The azinphosm ethyl degradation ra te for dislodgeable residues on apple foliage and the orchard grass determ ined in P a rt n was approxim ately 4.5% day- *. The initial daily losses due to photolysis, as determ ined above, would in d icate th a t photolysis may contrib u te significantly to th e disappearance of azinphosm ethyl foliar deposits. F acto rs th a t must be considered when estim atin g d ire c t photolysis under n atu ral conditions are discussed by Zepp and Cline (1977) and Sm ith e t al. (1977). These include: the incident light intensity, as a function of season, la titu d e , tim e of day, cloud cover, p ercen t of light adsorbed by th e pesticide (as com pared to its surroundings) and quantum yield (fraction of photons adsorbed th a t results in transform ation). As many pesticides show maximum adsorbance in th e u ltra violet region, outside the range of w avelengths reaching the su rface of the e a rth , sensitized reactions are o fte n im p o rtan t to pesticide photodegradation. A sen sitizer adsorbs, light a t a w avelength p resen t, followed by an energy tra n sfe r to the pesticide (Khan, 1974). The com ponents of a plant su rface which may a c t as sensitizers are largely unknown. The d a ta of Liang and L ichtenstein (1976) also indicated th a t azinphosm ethyl on th e soil su rface undergoes photodegradation. E stim ated daily initial ra te s of photodegradation for the th re e soil types used (sand, loam , muck) w ere slightly fa ste r than those estim ated for th e glass and foliar su rfaces. Due 171 to lim ited d ata available, no a tte m p t was made to e stim a te the ra te of azinphosm ethyl photodegradation on plant or soil su rfaces as a function of environm ental conditions. Much research needs to be done in this a re a , before such e stim ates will be possible. P lan t U ptake Greenhouse studies on plant uptake and m etabolism of 14 C azinphosm ethyl applied to bean leaves (Steffens and Wieneke, 1976) suggest th a t this loss m echanism may be sin im portant pathw ay in the a tten u atio n of foliar deposit residues in an apple orchard. The influence of environm ental fa c to rs such as solar radiation, te m p e ratu re, hum idity and oth er conditions influencing le af w etness on plant uptake of pesticides is discussed by Hull (1970), S teffens and Wieneke (1975) and Bukovae (1976). However, no q u an tita tiv e relationships are p resented. A com puter model for foliar uptake of pesticides was rep o rted by Bridges and Farrington (1974). This model rep resen ts plant uptake by both diffusion and mass flow through th e observable s tru c tu re of a w heat leaf. The model stru c tu re does not allow for the influence of environm ental conditions on th e ra te of foliar uptake. A general lack of d ata in th e lite ra tu re points to th e need for further research to investigate the contribution of plant uptake and m etabolism to th e overall atten u atio n of foliar residues as a function of environm ental conditions. CONCLUSIONS The lite ra tu re review ed and the results of the prelim inary experim ents rep o rted provide a basis for additional work. F u rth er m odel developm ent, as a m ethod for b e tte r understanding of pesticide environm ental dynam ics, will require th e form ulation of laboratory and field experim ents th a t a re specifically designed to provide d a ta to be used in the num erical analysis of pesticide fa te as a function of environm ental conditions. The conceptual m odel presented 172 identifies th e key processes in th e assessm ent o f pesticide fa te in an orchard ecosystem and investigation. the working model provides a fram ew ork for fu rth er PART V SUMMARY AND CONCLUSIONS SUMMARY A model for azinphosm ethyl atten u atio n and m ovem ent in a Michigan apple orchard ecosystem was developed to assess th e fa te and exposure of this compound to the orchard flora and fauna. The model was param eterized prim arily through th e use of field d ata g ath ered over a num ber of seasons. R ates of a tten u atio n w ithin, and m ovem ent b etw een, specified orchard com partm ents w ere determ ined under various rain fall regim es. The output of this model was stru c tu re d to allow the estim atio n of the tim e course of azinphosm ethyl exposure to ground-dwelling in v e rte b ra tes. Mean squared erro rs for th e com parison of the model predictions w ith an independent se t of residue d ata indicated good prediction of azinphosm ethyl fa te w ithin th e tre e and grassbroadleaves layers. Prediction of pesticide dynamics within the litter-m o ss and soil layers was much m ore d ifficu lt. This is thought to be due in p a rt to the lesser am ounts of residues distributed to th ese layers, variability in the com position of th e litter-m o ss layer, and th e strong dependence of pesticide dynam ics in these layers on environm ental facto rs oth er than rain fall. The model in its p resen t form provides a new approach to the analysis of field residue d ata. To the author’s knowledge, this is the first study of its kind to sim ultaneously determ ine pesticide a tte n u atio n and m ovem ent agroecosystem through the num erical analysis of field d ata. 173 within an In addition, the 174 output of th e fa te model was specifically designed to be used as input in a model describing the ecobiology and tem porally d istrib u ted m o rtality of the grounddwelling isopod Trachelipus rath k ei (Goodman e t al., 1981). This model was p artially param eterized using d a ta from field studies conducted concurrently within the sam e orchard as th e fa te studies (Snider, 1979; Snider and Shaddy, 1980). F a te model predictions e stim ate th a t under dry conditions 25% o f the daily loss of azinphosm ethyl from th e orchard tre e s is due to m ovem ent to o th er parts of the orchard. G re ater m ovem ent is p red icted under rain fall conditions. This m ovem ent resulted in increased pesticide exposure and additional m o rtality , especially among the im m ature age classes rep resen ted in the T. rath k ei model (Goodman, 1980). The results of the in itial distribution studies indicated th a t airborne loss played a major role in the early a tte n u atio n of azinphosm ethyl residue deposit residues. E stim ates of daily airborne loss d eterm ined from deposit residues and d irect sampling of airborne residues suggested th a t airborne loss was in fa c t largely responsible for the early loss of residues (40% of th e daily loss ra te on day 3 of the first spray period, 1978 season), under th e sp ecific s e t of environm ental conditions th a t prevailed during and following application. A m ulti-com ponent kinetic model was presen ted for estim atin g sim ultaneously th e early airborne loss of foliar deposits and the o ften slow er dissipation of the rem aining residues. Variation in th e s e t of equation p aram eters d eterm ined sep arately from d a ta gathered during th e first two spray periods of the 1978 season suggests th a t th e ra te of airborne loss and o th e r atten u atio n m echanism s (i.e., degradation, plant uptake and m etabolism ) a re dependent, in p a rt, on rain fall and other environm ental p aram eters y e t to be c h arac te riz ed . 175 CONCLUSIONS U nderstanding pesticide fa te throughout th e en tire orchard ecosystem is now gaining im portance as a resu lt of th e in te re st m anagem ent as a long-term stra te g y fo r pest control. in in teg rated p est In orchards and many o th er crops, where chem icals are still heavily relied upon, the success of biological control as a m ajor com ponent of IPM may depend largely on the judicious use of pesticides. The tren d in p est m anagem ent research has been to use modeling techniques to b e tte r understand th e ecobiology of pest, host, and beneficial species so as to m axim ize th e effectiv en ess of biological control m easures. A com parable e ffo rt in modeling pesticide fa te and e ffe c ts on both harm ful and beneficial species is essen tial to the developm ent of e ffe ctiv e and e ffic ie n t orchard in teg rated pest m anagem ent program s. The use of modeling techniques to b e tte r understand pesticide fa te in orchards may also provide a basis for th e developm ent of models describing the fa te of organic toxicants in te rre s tria l ecosystem s in general. Ecosystem s use energy from the sun to constantly recy cle th e ir w ater, a ir, m ineral, plant and anim al resources. N atural m echanism s thought to be essential to these system s include: biogeochem ial cycling, h o st-p a ra site relations, co m p etitito n for food and h a b ita t, predator-prey relationships, food chains, symbioses, com m unitydiversity and succession, and n atu ral selection (Odom, 1971; Pim entel and Goodman, 1974; Southwick, 1976; NSF/RA, 1976). The e ffe c t of th e introduction of chem ical substances into th e environm ent on the above-m entioned n atu ral m echanism s of ecosystem s is largely unknown. No te s ts are currently available which can address ecosystem e ffe c ts a t this level. The EPA adm its th a t proposed te s ts under FIFRA and TSCA are only screening te sts and th a t th e re is an urgent need for developm ent of higher level ecosystem te sts. The question now arises as to w hether th ese screening te sts can adequately se le c t those 176 compounds which will m ost likely pose no hazard to ecosystem s. Does a negative re su lt in an acu te to x icity te s t preclude behavioral e ffe c ts on a population influencing its ability to com pete for food and/or h a b ita t? There is no evidence th a t even chronic to x icity or life cycle studies can give th e answ er to this question. It is disturbing th a t negativ e resu lts fo r all of th e proposed "ecosystem e ffe c ts" te sts may allow continuous low level exposure of a chem ical substance when nothing is known about th e e ffe c ts of this exposure on any of th e aforem entioned "essential n a tu ra l mechanism s" of ecosystem s. A large e ffo rt by th e scien tific com m unity is needed to b e tte r understand ecosystem e ffe c ts above th e species level. This e ffo rt seem s param ount in light of the apparent inability of present screening te sts to preclude higher level e ffe c ts. Rapid developm ent of ecosystem e ffe c ts te s t standards should include th e use of m athem atical modeling techniques to deal w ith th e com plexity of th e in teractio n s involved. These ecosystem e ffe c ts models can then be coupled w ith chem ical exposure models producing a very pow erful tool in ecosystem hazard assessm ent. The developm ent of ecosystem models appears to be much fa rth e r off than exposure models. In th e near fu tu re, much em phasis m ust be placed on rapid developm ent o f exposure models, in screening chem icals for ecosystem e ffe c ts. A ccurate predictions of exposure may o ffse t much of th e u n c ertain ty in ecosystem e ffe c ts screening te sts. LIST OF REFERENCES LIST OF REFERENCES Al-Adil, K. M., White, E. R., W interlin, W. L. and K ilgore, W. W. (1973) U ptake and translocation of G uthion by beans and barley. J. Agric. Food Chem. 21, 376-379. Asmussen, L. E., White, A. W., J r., Hauser, E. U. and Sheridan, J . M. (1977) R eduction of 2,4-D load in su rface runoff down a grassed w aterw ay. J . Environ. Qual. 6, 159. A rm strong, D. E. and Konrad, J . G. (1974) Nonbiological degradation of pesticides. In P esticides in Soil and W ater. (Ed. by Guenzi, W. D.), p. 123— 131. SSSA, Madison, Wisconsin. Barrow, G. M. (1966) Physical C hem istry. M cGraw-Hill, Inc., New York, 843 p. Bauer, J. R ., Bovey, R. W. and M erkle, M. G. (1972) C oncentration of Picloram in runoff w ater. Weed Sci. 20, 309-314. Beynon, K. I., Davies, L., and Elgor, K. (1966) Analysis of crops and soils for residues of diethyl 1— (2,4 dichlorophenyl)-2-chlorovinyl phosphate. J . Sci. Food Agr. 17, 167-174. Bridges, R. C. and Farrington J . A. (1974) A co m p artm en tal com puter model of foliar uptake of pesticides. P estic. Sci. 5, 365-381. Brown, A. W. A. (1977) How have entom ologists d ealt with resistan ce. 68th Ann. M eet. Am. Phytopathol. Soc. 3, 67-74. Proc. Brown, A. W. A. (1978) Ecology of P esticid es. John Wiley and Sons, New York, 525 p. Bukovac, M. J . (1976) H erbicide en try into plants. In Herbicides: Physiology, B iochem istry, Ecology. (Ed. by Audus, L. J.), p. 335-364. Academ ic Press, London. C aro, J. H., Taylor, A. W. and Lemon, E. R. (1971) M easurem ent of pesticide concentrations in air overlying a tre a te d field. In tern atio n al Symposium on Id entification and M easurem ent of Environm ental P ollutants, O ttaw a, C anada. C aro, J . W., Freem an, H. P., and Turner, B. C. (1974) P ersistence in soil and losses in runoff of soil-incorporated carbaryl in a sm all w atershed. J . A gric. Food Chem. 22, 860. 177 178 C laith, M. M., Spencer, W. F., F arm er, W. J., Shoup, T. D., and G rover, R. (1980) V olatilization of S-ethyl n,n-dipropylthiocarbam ate from w ater and w et soil during and a fte r food irrigation of an a lfa lfa field. J . Agric. Food Chem. 28, 610-613. Cox, G. W. (1974) L aboratory Manual of G eneral Ecology. Company, Dubuque, Iowa, 196 p. Wm. C. Brown C ro ft, B. A. (1975) Tree fru it pest m anagem ent. In Introduction to P est M anagem ent. (Eds. M etcalf, R. L. and Luckmann, W. H.). Wiley Interscience, New York. C ro ft, B. A. (1978) P o ten tials for research and im plem entation of in teg rated p est m anagem ent on deciduous fru it-tre e s. In P est C ontrol S trateg ies (Eds. Smith, E. H. and P im entel, David), p. 101-115. A cadem ic Press, New York. C ro ft, B. A. and Brown, A. W. A. (1975) Responces of arthropod n atu ral enem ies to insecticides. Ann. Rev. Entomol. 20, 285-335. Crosby, D. G. (1969) E xperim ental approaches to pesticide photodecom position. Residue Reviews 25, 1-12. Crosby, D. G. (1970) The non-biological degradation of pesticides in soils. In Pesticides in th e Soil: Ecology, D egradation, and M ovement, p. 86-94, Mich. S ta te Univ., East Lansing, MI. Crosby, D. G. (1973) The fa te of pesticides in the environm ent. P lant Physiol. 24, 467. Day, B. E., Jordon, L. J. and Hendrixson, R. T. (1961) am itrole in C alifornia soils. Weeds 9, 443-456. Ann. Rev. The decom position of DeJong, K. (1980) Adaptive system design: a genetic approach. Syst. Man C ybern., SMC-10, p. 566-574. IEEE Trans. Ebling, W. 1963. Analysis of the basic processes involved in the deposition, degradation, p ersistence and effectiv en ess of pesticides. Residue Reviews 3, 35. E lliot, D. L., Iw ata, Y., C arm an, G. E., and G unther, F. A. 1977. Post­ application slough-off o f pesticide deposits on orange tre e s. Bull. Environ. C ontam . Toxicol. 18, 711-718. Eto, M. (1974) Organophosphate pesticides: organic and biological chem istry. CRC Press, C leveland, Ohio. 368 p. F aust, S. D. and Gom aa, H. M. 1972. C hem ical hydrolysis of some organic phosphorus and carb am ate pesticides in aq u atic environm ents. Environ. L e tt. 3, 171-201. Felsot, A. and Dahm, P. A. (1979) Sorption of organophosphorus and carb am ate insecticides by soil. J . A gric. Food Chem . 27, 557-563. 179 Freed, V. H., Chiou, C. T. and Schmedding, D. W. 1979. D egradation of selected organophosphate pesticides in w ater and soil. J. Agric. Food Chem. 27, 706-708. G ille tte , J. W. and W itt, J . M., Eds. (1979) T e rrestria l M icrocosms Proc. Symp. Workshop on T e rrestria l M icrocosm s, NSF/RA 790034. Glass, B. L.. and Edwards, W. M. (1974) Picloram in ly sim eter runoff and percolation w ater. Bull. Environ. C ontam . Toxicol. 11, 109-112. Glass, E. H. and Lienk, S. E. (1971) Apple in sect and m ite populations developing a fte r discontinuance of in secticid es: 10 year record. J . Econ. Entomol. 63/23-26. Goodman, E. D. (1980) Modeling p esticid e fa te and e ffe c ts on so il/litte r in v erteb rates in an orchard ecosystem . A b stract. Environm ental Risk Assessment - An In teg rated A pproach. Society of Environm ental Toxicology and C hem istry F irst Annual M eeting, A rlington, Virginia. Goodman, E. D., J. J . Jenkins, R. M. Snider, R. T. Kon, P. A. Weston, M. J. Zabik, T. Julian and M. T abatabai. (1981) A study and model of azinphosm ethyl fa te and e ffe c ts in a Michigan apple orchard ecosystem . Final Technical R eport, EPA C ooperative A greem ent R805624. (In preparation). Goring, C. A. I., Laskowski, D. A., H am aker, J . W. and M eikle, R. W. (1975) Principles of pesticide degradation in soil. In Environm ental Dynamics of Pesticides (Eds. Haque, R. and F reed, V. H.), p. 135-172. Plenum Press, New York. Grubbs, F. E. (1969) Procedures for d e tec tin g outlying observations in sam ples. Technom etrics 11, 1. G unther, F. A. 1969. Insecticide residues in C alifornia citru s fru it and products. Residue Reviews 28. G unther, F. A. and Blinn, R. C. (1955) Analysis of Insecticides and A carieides. Interscience, New York. G unther, F. A., Jeppson, L. R ., Barkley, J . H., E llio tt, L. M., Blinn, R. C., and Dunn, C. L. (1958) P ersisten ce of residues of 2,3-p-dioxane-dithol S,Sbis(0,0-diethyl phosphorodithioate) as an acaricid e on and in m ature lemons and oranges. J. Agr. Food Chem. 6, 210. G unther, F. A., W estlake, W. E., Barkley, J . H., W interlin, W. and Langbehn, L. 1973. Establishing dislodgeable p esticid e residues on le af su rfaces. Bull. Environ. C ontam . Toxicol. 9, 243-249. G unther, F. A., Iw ata, Y., C arm an, G. E., and Smith, C. A. 1977. The citrus reen try problem: research on its causes and e ffe c ts, and approaches to its m inim ization. Residue Reviews 67, 1-139. Hall, A. S. and Day, J. W., J r., Eds. (1977) Ecosystem Modeling - In Theory and P rac tic e. W iley-Interscience, New York. 684 p. 180 H all, F. R., R eichard, D. L., and K rueger, H. R. (1975) Dislodgeable azinphosm ethyl residues from air b last spraying of apple foliage in Ohio. A rch. Environ. C ontam . Toxicol. 3, 352-363. Hall, J. K., Paulus, M. and Higgins, E. R. (1972) Losses of a tra zin e in runoff w ater and soil sedim ent. J . Environ. Q uality 1, 172-176. H am aker, J. W. (1972) D ecom position: q u a n tita tiv e asp ects. In Organic C hem icals in the Soil Environm ent. (Eds. Goring, C. A. I. and H am aker, J. W), p. 253-340. M arcel D ekker, New York. Hansen, J. D., Schneider, B. A., Olive, B. M., and B ates, J . J . (1978) Personnel sa fe ty and foliage residue in an orchard spray program using azinphosm ethyl and cap tan . Arch. Environ. C ontam . Toxicol. 7, 63-71. H arrold, L. L., T riplet, G. B., J r., and Edwards, W. M. (1970) N o-tillage cornc h a ra c te ristic s of the system . Agric. Eng. 51, 128. H artley, G. S. (1969) Evaporation o f pesticides. In Pesticide Form ulations R esearch. (Ed. Gould, R. F), p. 115-134. Am. Chem . Soc. 'Advances in C hem istry Series,' 86. Haynes, R. J . (1980) Influence of soil m anagem ent p ra c tice on th e soil agro­ ecosystem . A gro-Ecosystem s 6, 3-32. Helling, C. S., K earney, P. C., A lexander, M. soils. Adv. Agron. 23, 147-241. (1971) Behavior of pesticides in Hill, K. R. (1971) R eport of h alf-life working p arty . C hem ists 54, 1316-1317. J. Ass. O ffic. Anal. Holland, J . H. (1975) A daptation in n a tu ra l and a rtific ia l system s. U niversity of Michigan Press, Ann Arbor. 183 p. The H ow itt, A. J . and Pshea, A. (1965) The developm ent and use of u ltra low volume ground sprayers for pests a tta c k in g fru it. Mich. Q uart. Bull. 48, 144. H oyt, S. C. and Burts, E. C. (1974) In teg rated control of fru it pests. Ann. Rev. Entom ol. 19, 231-252. Hull, H. G. (1970) Leaf stru c tu re as re la te d to absorption of pesticide and oth er compounds. Residue Reviews 31, 1-150. Iw ata, Y., Ousch, M. E., W estlake, W. E., and G unther, F. A. 1975. Behavior of five organophosphorus pesticides in dust derived from several soil types. Bull. Environ. C ontam . Toxicol. 14, 49. Iw ata, Y., W estlake, W. E., Barkley, J . H., C arm an, G. E., and G unther, F. A. (1977) Behavior of phenthoate (cidial) deposits and residues on and in grap efru its, lemons and lem on leaves, oranges and orange leaves, and in th e soil beneath orange tre e s. J . A gric. Food Chem . 25, 362-368. 181 Kadoum, A. M. (1967) A rapid m icrom ethod of sam ple cleanup for gas chrom atographic analysis of in secticid al residues in plant, anim al, soil, and su rface and ground w a ter e x tra c ts. Bull. Environ. C ontam . Toxicol. 2, 264. K atan, J ., Fuhrm ann, T. W. and L ichtenstein, E. P. (1976) Binding of parathion in soil: a reassessm ent of pesticide persistence. Science 193, 891-894. K aufm an, D. D. (1964) M icrobial degradation of 2,2-dichloro-propionic acid in five soils. Can. J . M icrobiol. 10, 843-852. K earney, P. C. and H elling, C. S. Residue Reviews 25, 25-44. (1969) R eactions of pesticides in soils. Khan, M. A. Q., Khan, H. M., and Southerland, D. J . (1976) Ecological and health e ffe c ts of th e photolysis of insecticides. In Survival in Toxic Environm ents (Eds. Khan, M. A. Q. and Bederka, J. P., Jr.), p. 333-355. A cadem ic Press, New York. Klingman, G. C. and Ashton, F. M. (1975) Weed Science: P rac tic es. Wiley Interscience, New York, p. 32-33. Principles and Kuhr, R. J., Davis, A. C. and Bourke, J. B. (1974) Dissipation of Guthion, Sevin, Polyram , Phygon and Systox from apple orchard soil. Bull. Environ. C ontam . Toxicol. 11, 224-230. L assiter, R. R., Baughman, G. L. and Burns, L. A. (1978) F a te of toxic organic substances in the aq u atic environm ent. In S ta te -o f-th e -A rt in Ecological Modeling (Ed. Jorgensen, S. E.), 219-245. Inter. Soc. Ecol. Mod., Copenhagen. Leonard, R. A., Bailey, G. W., Swank, R. R., J r. (1976) T ransport, detoxification, fa te and e ffe c ts of pesticides in soil and w ater environm ents. In Land A pplication of Waste M aterials, 49-78. Soil C onservation Society of A m erica, Ankeny, Iowa. Liang, T. T. and L ichtenstein, E. P. (1972) E ffect of light, te m p e ratu re, and pH on the degradation of azinphosm ethyl. J. Econ. Entomol. 65, 315-321. Liang, T. T. and L ichtenstein, E. P. .(1976) E ffects of soils and le af surfaces on th e photodecom position of [ C] azinphosm ethyl. J. A gric. Food Chem. 24, 1205-1210. M atsum ura, Fumio. Toxicology of Insecticides. Plenum Press, New York, 503 p. Maybank, J ., Yoshida, K. and G rover, R. (1977) The M agnitude and Im portance of Spray D rift from A gricultural A pplcations. 70th Ann. M eeting Air Poll. C ontrol Assoc., T oronto, C anada. M cM echan, A. D., M organ, C. V. G. and Wardle, G. A. (1972) Erosion of azinphosm ethyl from apple leaves by rain and o v e rtre e irrigation. J. Entomol. Soc. B rit. Colum bia, 69, 5-8. 182 M enzie, C. M. (1972) F ate of pesticides in th e environm ent. Ann. Rev. Ent. 17, 199-221. N ational Academy of Sciences. 1969. Insect P est M anagem ent and C ontrol. Publ. 1695. Washington, D. C. 508 pp. Nigg, H. N., Allen, J. C., King, R. W., Thompson, N. P., Edwards, G. J . and Brooks, R. F. (1977) Dislodgeable residues o f ethion in Florida citru s and relationships to w eather variables. Arch. Environ. C ontam . Toxicol. 6, 257-267. N illes, G. P. and Zabik, M. J . (1975) Photochem istry of bioactive compounds. M ultiphase photodegradation and mass sp e c tra l analysis o f basagran. J. A gric. Food Chem. 23, 410-414. NSF/RA. (1976) Ecosystem processes and organic co ntam inants, research needs and an interdisciplinary p ersp ectiv e. NSF/RA 760008. Odom, E. P. (1971) Fundam entals of Ecology. 3rd Edition, W. B. Saunders Co., Philadelphia. 547 p. Parm ele, L. H., Lemon, E. R. and Taylor, A. W. (1972) M icrom eteorological m easurem ent of pesticide vapor flux from b are soil and corn under field conditions. W ater Air Soil Poll. 1, 433-451. Phillips, F. T. (1971) P ersistence of organochlorine in secticides on d ifferen t su b strates under d ifferen t environm ental conditions. I. The ra te s of loss of dieldrin and aldrin by v o latilizatio n from glass su rfaces. P estic. Sci. 2, 255-266. Phillips, F. T. (1974) Some asp ects of v o latilizatio n of organochlorine insecticides. Chem. Ind. 5, 197-201. Pielou, F. C. (1969) An Introduction to M ath em atical Ecology. Interscience, New York, p. 19-22. Wiley Pim entel, D. and Goodman, N. (1974) Environm ental im pact of pesticides. In Survival in Toxic Environm ents. (Eds. Khan, M. A. Q. and Bederka, J . B., Jr.), 25-52. Plim m er, J. R., K earney, P. C ., K aufm an, D. D. and G uardia, F. S. (1967) A m itrole decom position by fre e rad ical g en eratin g system s and by soils. J. Agric. Food Chem. 15, 996-999. Popendorf, W. J. and Leffingw ell, J . T. (1978) N atural variations in the decay and oxidation of parathion foliar residues. J . A gric. Food Chem. 26, 437441. R itte r, W. F., Johnson, H. P., Lovely, W. G. and Molnau, M. (1974) A trazine, propachlor, and diazinon residues on sm all a g ric u ltu ra l w atersheds, runoff losses, persistence and m ovem ent. Environ. Sci. Technol. 8, 38-42. Rohrbaugh, W. J ., M eyers, E. K. and Jacobson, R. A. m olecular stru c tu re of organophosphorus insecticides. 1976. C ry stal and 183 Saltzm an, S., L. Kliger and Yaron, B. (1972) A dsorption-desorption of parathion as a ffe c te d by soil organic m a tte r. J . Agric. Food Chem. 20, 1224-1226. Schrader, G. (1963) Die entw ichlung neuer insecktizider phosphazsaure-ester. W einhein/B ergstr.: V erlag Chem ie GmbH. Schubert, O. E. (1972) P lant cover changes following herbicide applications in orchards. Weed Sci. 20, 124-127. Schulz, K. R., L ichtenstein, E. P., Liang, T. T. and Fuhrm ann, T. W. (1970) P ersistence and degradation of azinphosm ethyl in soils, as a ffe c te d by form ulation and mode of application. J . Econ. Entom ol. 63, 432-438. Sm ith, A. E. and Walker, A. (1977) A q u a n tita tiv e study of asulam p ersisten ce in soil. P estic. Sci. 8, 449-456. Sm ith, J . H., M abey, W. R., Bohonos, N., H olt, B. R., Lee, S. S., Chou, T. W., Bom berger, D. C. and Mill, T. (1977) Environm ental Pathw ays of Selected C hem icals in F reshw ater System s. EPA R eport 600/7-77-113. Sm ith, R. F. (1976) Insecticides and in te g ra te d pest m anagem ent. In The Future for Insecticides, Needs and P rospects. Wiley In terscien ce, New York, p. 489-506. R Snider, R. M. (1979) The e ffe c ts of azinphosm ethyl (Guthion ) on a population of Trachelipus rathkei (Isopoda) in a Michigan orchard. Pedobiol. 19, 99105. Snider, R. M. and Shaddy, J. H. (1980) The ecobiology of Trachelipus rathkei (Isopoda). Pedobiol. 20, 394-410. Southwick, C. H. (1976) Ecology and the Q uality of Our Environm ent. 2nd Ed., D. Van N ostrand, London. 426 p. Spear, R. C., Popendorf, W. J., Leffingw ell, J . T., and Jenkins, D. (1975) Parathion residues on citru s foliage. D ecay and com position as re la te d to worker hazard. J. A gric. Food Chem . 23, 808-810. Spencer, W. F. and C laith, M. M. (1975) Vaporization of chem icals. In Environm ental Dynamics of P esticides. (Eds. Haque, R. and Freed, V. H.), p. 61-78. Plenum Press, New York. Spencer, W. F. and C laith, M. M. (1977) The solid-air in te rfa c e : tra n sfe r of organic pollutants betw een the solid-air in te rfa c e . In F a te of P ollutants in th e Air and W ater Environm ents, P a rt I. (Ed. S uffet, I. H), p. 107-126. John Wiley and Sons, Inc., New York. Spencer, W. F., F arm er, W. J . and C laith, M. M. (1973) Pesticide V olatilization. R esidue Reviews 49, 1-47. S ta iff, D. C., C om er, S. W. and F o ster, R. J . (1975) Residues of parathion and conversion products on apple and peach foliage resu ltin g from rep eated spray applications. Bull. Environ. C ontam . Toxicol. 14, 135-139. 184 Stam per, J . H., Nigg, H. N. and Allen, J . C. (1979) Organophosphate insecticide disappearance from le af surfaces: an alte rn ativ e to first-o rd e r kin etics. Environ. Sci. Tech. 13, 1402-1405. S teffens, W. and W ieneke^ J . (1975) Influence of hum idity and rain on uptake and m etabolism of C-azinphosm ethyl in bean plants. A rch. Environ. C ontam . Toxicol. 3, 364-370. S teffens, W. and J . W ieneke. (Ift76) Studies on the uptake, m etabolism and degradation of C -labelled Gusat.hion in kidney beans. P flanzenschultznachrichten 29, 1-17. Steiner, P. W. (1969) The distribution of spray m aterial betw een ta rg e t and non­ ta rg e t areas of a m ature apple orchard by airblast equipm ent. M.S. Thesis, Cornell Univ. Taylor, A. W. (1978) Post-application volatilization of pesticides under field conditions. J. Air Poll. C ontrol Assoc. 28, 922-927. Taylor, A. W., G lotfelty, D. E., Glass, B. L., Freem an, H. P. and Edwards, W. M. (1976) V olatilization of dieldrin and heptachlor from a m aize field. J. Agric. Food Chem. 24, 625-631. Taylor, A. W., G lo tfelty, D. E., Turner, B. C., Silver, R. E., Freem an, H. P. and Weiss, A. (1977) V olatilization of dieldrin and heptachlor residues from field vegetation. J. Agric. Food Chem . 25, 542-548. Teskey, B. J. E. and Shoem aker, J. S. (1978) Tree F ru it Production. Publishing Co., Inc., W estport, Conn., 19-25. AVI Thompson, N. P. and Brooks, R. F. (1976) D isappearance of dislodgeable residues of five organophosphate pesticides on citru s leaves and fru it during dry and w et w eather in Florida. Arch. Environ. C ontam . Toxicol. 5, 55-61. Turner, B. C. and G lo tfelty, D. E. (1977) Field air sampling of pesticide vapors with polyurethane foam . Anal. Chem. 49, 7-10. Van Dyk, L. P. (1974) The persistence of som e pesticides on South A frican citru s. Environ. Qual. S afety, suppl. Ill, p. 330-335. Van Dyk, L. P. (1976) Parathion persistence on South A frican citru s. Environ. C ontam . Toxicol. 4, 289-311. Arch. Walker, A. (1974) A sim ulation model for th e prediction of herbicide persistence. J. Environ. Q uality 3, 396-401. Walker, A. (1976a) Sim ulation of herbicide persistence in soil. I. Sim azine and Prom etryne. P estic. Sci. 7, 41-49. Walker, A. (1976b) Sim ulation of herbicide persistence in soil. n. Sim azine and Linuron Long-Term Experim ents. Pestic. Sci. 7, 50-58. 185 W alker, A. (1976c) Sim ulation of herbicide persistence in soil. HI. Propyzam ide in D ifferent Soil Types. P estic. Sci. 7, 59-64. Ware, G. W. (1980) E ffects of pesticides on n o n targ et organism s. Reviews 76, 173-201. Residue White, A. W., J r., H arper, L. A., Leonard, R. A. and Turnbull, J . W. (1977) Trifluralin v o latilization losses from a soybean field. J . Environ. Qual. 6, 105-110. W hittaker, R. H. (1975) Co., New York. C om m unities and Ecosystem s. MacMillan Publishing 14 Wineke, J. and S teffens, W. (1974) D egradation of C-azinphos on glass plates as com pared to th e m etabolism on bean leaves. Environ. Qual. Safety, Suppl. ffl., 544-548. Williams, K. (1961) Note on the e ffe c t of rain, and sprinkler irrig atio n , on the persistence of spray residues of Guthion and Sevin on apple leaves. Can. J. Plant Sci. 41, 449-451. W interlin, W., Bailey, J. B., Langbehn, L. and M ourer, C. (1975) D egradation of parathion applied to peach leaves. P est. M ont. J. 8, 263-269. Woodrow, J . E., Seiber, J . N., Crosby, D. G., M oilanen, K. W., Soderquist, C. J. and M ourer, C. (1977) Airborne and surface residues of parathion and its conversion products in a tre a te d plum orchard environm ent. Arch. Environ. C ontam . Toxicol. 6, 175-191. Yaron, B., H ever, B. and Birk, Y. (1974) K inetics of azinphosm ethyl losses in th e soil environm ent. J . A gric. Food Chem. 22, 439-441. Y ates, W. E. and N. B. Akesson. (1973) Reducing pesticide chem ical d rift. In Pesticide Form ulations. (Ed. Van Valkenburg, W.), p. 275-339. M arcel D ekker, Inc. Zabik, M. J. and L. O. Ruzo. (1981) F acto rs A ffectin g Pesticide Photodecom position Studies. In T est Protocols for Environm ental F ate and M ovement of Toxicants, p. 20-27. Proc. Symp. 94th Ann. M eeting Ass. of O ffic. Anal. C hem ists, W ashington, D.C. Zabik, M. J., L eav itt, R. A. and Su, G. C. (1976) Photochem istry of bioactive compounds: a review o f pesticide photochem istry. Ann. Rev. Entomol. 21, 61-79. Zepp, R. G., Wolfe, N. L., Gordon, J . A., and Baughman, G. L. (1975) Dynamics of 2,4-D e ste rs in su rface w aters hydrolysis, photolysis, and vaporization. Environ. Sci. Tech. 9, 1144-1150. Zepp, R. G. and Cline, David M. (1977) R ates of d ire c t photolysis in aquatic environm ent. Environ. Sci. Tech. 11, 359-366.