MSU-ENGR-82-014
ECOSYSTEM RESPONSES TO ALTERNATIVE PESTICIDES
IN THE TERRESTRIAL ENVIRONMENT:
A SYSTEM APPROACH
Final Technical Report
Environmental Protection Agency
Cooperative Research Agreement R 805624
June 30, 1982
Erik D. Goodman, Ph.D.
Principal Investigator
Department of Electrical Engineering and Systems Science
Jeffrey J. Jenkins, Ph.D.
Robert M. Ron, Ph.D.
Pesticide Research Center
Renate M. Snider, Ph.D.
Department of Zoology
Michigan State University
Project Officer
Jay D. Gile
Toxics and Hazardous Materials Branch
Environmental Research Laboratory, EPA
Corvallls, Oregon
MICHIGAN STATK UNIVERSITY

COLLEGE OF ENGINEERING
MICHIGAN STATE UNIVERSITY
EAST LANSING, MICHIGAN 48824
NSU IS AN AFFIRMATIVE ACTION/EQUAL OPPORTUNITY INSTITUTION

-------
DISCLAIMER
This report does not necessarily reflect the views and policies of the
U.S. Environmental Protection Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
ii

-------
CONTENTS
Page
Disclaimer 		ii
Table of Contents 			iii
Figures 		iv
Tables 		xii
Summary and Recommendations 		xvii
Acknowledgement 		xiv
Introduction 		xx
CHAPTERS
I	Experimental Data for the Azinphosmethyl Fate Model 		1
II	Parameterization of the Azinphosmethyl Fate Model 		32
III	Airborne Loss from the Orchard 		54
IV	Generalization of the Pesticide Fate Model 		84
V	Field Studies of Azinphosmethyl (Guthion ) Effects on
Soil/Litter Invertebrates 		115
VI	Laboratory Toxicity Studies on Selected Soil/Litter
Invertebrates 		122
VII	Descriptions of Spider, Earthworm, and Springtail Models ..	132
VIII	The Ecobiology of Trachelipus Rathkei 		135
IX	The Effects of Azinphosmethyl on a Population of
Trachelipus rathkei (Isopoda) 			156
X	A Study and Model of Azinphosmethyl Effects on a Population
of T. rathkei 				164
References 		194
Appendices
A.	Running the Process and Orcdata Data Analysis Routines 		A-l
B.	The POEM Model User's Guide 		B-l
iii

-------
FIGURES
Number	Page
1-1 Orchard plots 	 4
1-2 Overhead view of horizontal regions of orchard
plots 	 6
I-3(a-d) Azinphosmethyl dislodgeable and soil residues
for the canopy region, 1976 season. Upward
directed arrows show dates of spray application,
and downward directed arrows indicate dates and
amounts of rainfall in mm. Each bar is divided
into two parts; the mean (yg/cm^) is above the
line and its corresponding standard error (S.E.)
below the line 	 22
I-3(e-g) Azinphosmethyl dislodgeable and soil residues
for the alley region, 1976 season. Upward
directed arrows show dates of spray application,
and downward directed arrows indicate dates and
amounts of rainfall in mm. Each bar is divided
into two parts; the mean (yg/cnr) is above the
line and its corresponding standard error (S.E.)
below the 1 ine 			 23
1-4(a-d) Azinphosmethyl dislodgeable and soil residues for
the canopy region, 1977 season. Upward directed
arrows show dates of spray application, and down-
ward directed arrows indicate dates and amounts
of rainfall in mm. Each bar is divided into two
parts; the mean (yg/cm ) is above the line and
its corresponding standard error (S.E.) below
the line 	 26
I-4(e-g) Azinphosmethyl dislodgeable and soil residues for
the alley region, 1977 season. Upward directed
arrows show dates of spray application, and down-
ward directed arrows indicate dates and amounts
of rainfall in mm. Each bar is divided into two
parts; the mean (yg/cm ) is above the line and
its corresponding standard error (S.E.) below
the line 	 27
1-5 Daily temperature range, 1976 season 	 15
1-6 Daily minimum and maximum relative humidity,
1976 season 	 28
iv

-------
FIGURES
Number Page
1-7	Daily temperature range, 1977 season 	 29
1-8	Daily minimum and maximum relative humidity,
1977 season 	 30
II-l Conceptual model for the distribution, move-
ment, and attenuation of azinphosmethyl in
an apple orchard 	 33
11-2	Flow chart showing data processing to yield files
suitable for use in model parameterization 	 36
11-3	Flow chart of the process of calculating the
parameters using the "processed files" 	 37
11-4	Azinphosmethyl attenuation matrix 		40
11-5	Azinphosmethyl non-rainfall movement matrix 		41
11-6	Azinphosmethyl light rainfall movement matrix 		42
11-7	Azinphosmethyl heavy rainfall movement matrix 		43
11-8(a-d) . Actual and predicted Azinphosmethyl dislodgeable
and soil residues for the canopy region, 1978
season, using predicted initial distribution
for each spray period. Upward directed arrows
show dates of spray application, and downward
directed arrows indicate dates and amounts of
rainfall in mm. Each bar is divided into two
parts; the mean (yg/cm ground area) is above
the line and its corresponding standard error
(S.E.) below the line. The solid line running
through the bars represents the model's pre-
diction of the residues using 1976 and 1977
data only, while the dashed line shows the pre-
dicted residues based on all three years (1976-
1978) data 	 46
II-8(e-g) Actual and predicted Azinphosmethyl dislodgeable
and soil residues for the alley region, 1978
season. Upward directed arrows show dates of
spray application, and downward directed arrows
indicate dates and amounts of rainfall in mm.
Each bar is divided into two parts; the mean
(pg/cm ground area) is above the line and its
corresponding standard error (S.E.) below the
line. The solid line running through the bars
represents the model's prediction of the
change in residues using 1976 and 1977 data
only, while the dashed line shows the predicted
change in residues based on all three years'	 47
V

-------
FIGURES
Number	Page
II-9(a-d)	Actual and predicted Azinphosmethyl dislodge-
able and soil residues for the canopy region,
1978 season, using the observed initial dis-
tribution for each spray period. Upward directed
arrows show dates of spray application, and down-
ward directed arrows indicate dates and amounts
of rainfall in mm. Each bar is divided into two
parts; the mean (ug/cm ground area) is above the
line and its corresponding standard error (S.E.)
below the line. The line running through the bars
represents the model's prediction of the residues
using 1976 and 1977 data only, while the dashed
line shows the predicted residues based on all
three years (1976-1978) data 	 48
I1-9(e-g) Actual and predicted Azinphosmethyl dislodge-
able and soil residues for the alley region,
1978 season, using the observed initial dis-
tribution for each spray period. Upward directed
arrows show dates of spray application, and down-
ward directed arrows indicate dates and amounts
of rainfall in mm. Ea6h bar is divided into two
parts; the mean (yg/cm ground area) is above
the line and its corresponding standard error
(S.E.) below the line. The solid line running
through the bars represents the model's predic-
tion of the residues using 1976 and 1977 data
only, while the dashed line shows the predicted
residues based on all three years' (1976-1978)
data 		 49
III-1	Orchard plots showing air sampling locations 	
111-2 Diurnal Azinphosmethyl airborne residues at
sampling heights between 0.5 and 6.0 meters
above the orchard, measured at the center
location during the first spray period of
the 1978 season 	 57
II1-3 Diurnal Azinphosmethyl airborne residues at
sampling heights between 0.5 and 6.0 meters
above the orchard, measured at the downwind
locations during the first spray period of
the 1978 season 	 70
56
vi

-------
71
72
83
82
61
62
87
89
90
91
94
96
FIGURES
Azinphosmethyl dislodgeable and soil residues
for the canopy region, 1978 season. Upward
directed arrows show dates of spray applica-
tion, and downward directed arrows indicate
dates and amounts of rainfall in mm. Each«bar
is divided into two parts; the mean (ug/cm )
is above the line and its corresponding stan-
dard error (S.E.) below the line 	
Azinphosmethyl dislodgeable and soil residues
for the alley region, 1978 season. Upward
directed arrows show dates of spray applica-
tion, and downward directed arrows indicate
dates and amounts of rainfall in inm. Eachobar
is divided into two parts; the mean (yg/cm )
is above the line and its corresponding stan-
dard error (S.E.) below the line 	
Decline in total Azinphosmethyl dislodgeable
and soil residues measured during the first
spray period of the 1978 season 	
Decline in total Azinphosmethyl dislodgeable
and soil residues measured during the second
spray period of the 1978 season 	
Daily temperature range, 1978 season 	
Daily minimum and maximum relative humidity,
1978 season 	
Azinphosmethyl hydrolysis rate constant versus pH 	
Azinphosmethyl hydrolysis rate constant versus air
temperature 	
Azinphosmethyl hydrolysis rate constant versus soil
temperature 			
Azinphosmethyl microbial degradation versus soil
percent organic matter 	
Azinphosmethyl microbial degradation versus soil
temperature 	
Amitrole degradation versus soil moisture for two
soil types 		
vii

-------
Pa^e
104
105
106
107
108
109
110
112
113
121
FIGURES
Predicted 1978 azinphosmethyl residues in orchard
canopy region layers. Dashed lines represent
predictions based on standard condition attenu-
ation rates (Table IV-2), while solid lines in-
clude predicted effects of daily temperature
and soil moisture changes (the "base line run") 	
Predicted 1978 azinphosmethyl residues in orchard
alley region layers. Dashed lines represent
predictions based on standard condition attenu-
ation rates (Table IV-2), while solid lines in-
clude predicted effects of daily temperature
and soil moisture changes (the "base line run") 	
Predicted effects of lowering daily temperature
by 10 F on azinphosmethyl residues (dashed
lines) versus the base line run (solid lines) 	
Predicted effects (dashed lines) of increasing _7
pesticide volatilization rate by lOx (from 10
to 10 mm Hg). Solid lines are base line run 	
Predicted effects (dashed line) of altering soil
pH from 7 to 9 	
Predicted effects (dashed lines), in canopy layers,
of doubling daily rates of pesticide movement
due to heavy and light rainfall. Solid lines
represent the base line run 	
Predicted effects (dashed lines), in alley layers,
of doubling daily rates of pesticide movement
due to heavy and Tight rainfall. Solid lines
represent the base line run 	
Predicted effects (dashed lines), in canopy
layers, of postulating no inter-layer pesticide
movement under non-rainfall conditions. Solid
lines represent the base line run 	
Predicted effects (dashed lines), in alley layers,
of postulating no inter-layer pesticide movement
under non-rainfall conditions. Solid lines
represent the base line run 	
T. rathkei sampled in 1979 in three orchard plots 	
viii

-------
FIGURES
Number	Page
VI11-1 Growth of Trachelipus females on two diets (mean
weights at 10-day intervals, plus or minus
one S.D. at selected ages) 	 137
VII1-2 First broods of Trachelipus: relationship
between the post-parturient weight of
females and the number o£ young released
(y = 0.859 wt - 6.434; rc = 0.77) 	 140
VI11-3 Second broods of Trachelipus: relationship
between post-parturient female weight and
number oi young released (y = 0.538 wt +
1.154; r = 0.47) 	 140
VI11-4 Relationship between weight gained during gra-
vidity and the number of young produced.
Weight gain = maximum recorded weight -
initial gravid weight (y = 0.264x + 1.154;
rd = 0.47) 	 142
VIII-5(a-c) Mean weights at 10-day intervals of breeding
Trachelipus females reared at three tempera-
tures. Crosses replacing dots indicate that
40 percent or more of all individuals were
carrying broods. At each temperature, the
replicates were grouped according to female
weights after the first brood release: <30 mg,
30-40 mg, and 50-60 mg. Numbers in parentheses:
number of replicates at the beginning of obser-
vation 	 145
VI11-6 Number of days spent by Trachelipus females in
each of four weight classes, at three constant
temperatures 	 I*5
VI11-7 Weight distribution of the Trachelipus field popu-
lation, in percent of all females captured per
date. Total numbers (in parentheses) include
half of the prejuveniles plus all identifiable
females caught. Weight class definitions:
1 = 0.5-5.0 mg; 2 = 5.01-30.0 mg; 3 = 30.01-
50.0 mg; 4 = >50.0 mg 	 150
VI11-8 Distribution of the Trachelipus population over
four weight classes, each divided into the fol-
lowing sub-classes: Class 1: 0.5-0.89; 0.90-2.11;
5.0 mg. Class 2: 5.01-9.79; 9.80-19.17; 19.18-
30.0 mg. Class 3: 30.01-35.57; 35.58-42.17;
42.18-50.0 mg. Class 4: 50.01-59.28; 59.29-70.29;
>70.29 mg. Data are given as percent of all fe-
males caught. Total numbers as in Fig. 7 	 151
ix

-------
FIGURES
Number	Page
VI11-9 Breeding periodicity of Trachelipus in the
field: of all females > 12 mg, percent in
breeding condition. Solid line: females
with empty oostegites included; broken
line: only individuals actually carrying
brood are counted 	
VI11-10 Weight structure of the breeding segment of
the Trachelipus field population, in per-
cent of the total number of females carrying
brood on each collection date 	 153
IX-1 Mean number of T. rathkei collected from boards
placed underneath tree canopies and from
those in alleys between trees 	 1-)8
IX-2 Percent females captured, of the total number
of T. rathkei that could be sexed; exact
collection dates as in Fig. 1 Symbols:
o = Plot 7. X = Plot 8 	 159
IX-3(A,B) Size class distribution, in percent of total
number captured per date, of T. rathkei
populations throughout one year. A: control
plots 7 and 8 combined. B: sprayed plots 2
and 5 combined. In parentheses: number of
animals collected. See text for description
of size classes. Errata in fig. 3B. Plot-
number above the first column should be read
4 instead of 6. At 07.04.78 61 animals were
collected 	
X-l Conceptual model showing processes affecting a
life stage of T. rathkei. Solid lines repre-
sent flow of T. rathkei; dashed lines, flow of
information. Valves control rate of flow,
based on incoming information 	
X-2	Expanded diagram highlighting the breeding pro-
cess and the factors affecting it
170
X-3	Expanded view of the life stages of T. rathkei
as tracked in the model. Each box indicates the
number of bins it contains. Individuals may
also leave TRBRST 1 and TRBRST 2 (post-
parturient "resting" stages) for the top row of
non-breeding classes, although the arrows were
omitted to simplify the diagram 	
x

-------
FIGURES
Number	Page
X-4 Time response of C- (k). Vertical axis shows
the proportion of'the distance between (0)
and m. which C. (k) has moved by day k, for
various choice$'Sf A, the daily change rate 	 180
X-5	Mean numbers of isopods captured/board, in
control vs. sprayed plots, broken into size
groups of £ 12 mg, >12 mg 	
2
X-6	Simulated populations (number/m ) of T. rathkei
for three years in a control (unsprayed) plot:
(a) eggs (daily hatch), (b) 0.5-5 mg, (c)
5-12 mg, (d) 12-30 mg, (e) 30-50 mg, (f) 50-
80 mg, (g) >80 mg, (h) totals, excluding eggs 	
X-7	Simulated populations (number/m ) of T. rathkei
for three years in a plot sprayed with azinphos-
methyl (see Pesticide Fate Model section): (a)
eggs (daily hatch), (b) 0.5-5 mg, (c) 5-12 mg,
(d) 12-30 mg, (e) 30-50 mg, (f) 50-80 mg, (g)
>80 mg, (h) totals, excluding eggs 	 iyu
A-l	Flowchart of the processing of raw data to
create processed files 	
A-2	Data flow in parameterization routines (0RCDATA) 	 A-1°
XI

-------
TABLES
Number	Page
1-1	Summary of Significant Events for the Orchard,
1976	Season 	 8
1-2	Summary of Significant Events for the Orchard,
1977	Season 	 9
1-3	Azinphosmethyl Initial Horizontal Distribution
from the Analysis of Targets, 1976 Season 	 10
1-4	Azinphosmethyl Initial Horizontal Distribution
from the Analysis of Targets, 1977 Season 	 H
1-5	Initial Vertical Distribution of Azinphosmethyl
Dislodgeable Residues (ug/cm + SE) 		 12
1-6 Estimation of Initial Vertical Pesticide Distri-
bution as a Proportion of the Amount Applied
(+ SE), from the Analysis of Samples 	 M
1-7 Initial Vertical Pesticide Distribution, as a
Proportion of the Amount Applied, Using
Target Data to Calculate Proportion Reaching
the Orchard Floor 	 17
1-8	Azinphosmethyl in Runoff, 1977 Season 	 18
11-1 Comparison of the Model Predictions (Inclu-
ding Drift) with Azinphosmethyl Residue
Data for the 1978 Season 	 51
II-2 Comparison of the Model Predictions (Exclu-
ding Spray Dates) with Azinphosmethyl
Residue Data for the 1978 Season 	 53
III-1	Summary of Significant Events for the
Orchard, 1978 Season 	 59
II1-2 Azinphosmethyl Initial Horizontal Distribu-
tion from the Analysis of Targets, First
Spray Period of the 1978 Season 	 63
111-3 Initial Vertical Distribution of Azinphos-
methyl Dislodgeable Residues, First Spray
Period of the 1978 Season 	 65
111-4	Estimation of Initial Vertical Pesticide Dis-
tribution as a Proportion of the Amount
Applied (+ S.E.) from the Analysis of
Samples for the First Spray Period of the
1978	Season 	 66
xii

-------
TABLES
Number	^a9e
111-5 Estimation of Initial Vertical Pesticide Dis-
tribution as a Proportion of the Amount
Applied, Using Target Data to Calculate
the Proportion Reaching the Orchard Floor,
for the First Spray Period of the 1978
Season 	
3
II1-6	Azinphosmethyl Concentrations (yg/cm ) at
Sampling Heights Between 0.5 and Six Meters
Over Orchard 	 68
II1-7 Estimated Azinphosmethyl Daily Airborne Loss
from the Orchard, First Spray Period of
the 1978 Season 	
IV-1	Azinphosmethyl Degradation when Incubated with	fi7
Various Soils 	
IV-2	Daily Azinphosmethyl Attenuation Proportions,
Under Standard Conditions, Subject to User
Alteration 	 101
IV-3	Daily Azinphosmethyl Attenuation Proportions,
Separated by Process, Under Standard Condi-
tions, Subject to User Alteration 	 102
V-l	Plots Sampled by Pit-Trapping in 1977-79 	 H6
V-2	Results of Analysis of Variance of Selected
Groups of Invertebrates Found in Pit-Trap
Samples (NS = Not Significant; Dashes =
Factors Not Included in the Statistical
Model in 1979) 	 H7
V-3	Mean Number of A. caliginosa Per Sample in
1976 and 1977. Three Samples Were Taken
Per Plot. The Figures Include All Worms
Found in the Upper 30 cm of Soil (Means
+ S.E.) 			
V-4 Average Percent of Worms Found in 3 Depth In-
crements (Mean + S.E., Number of Samples in
Parentheses) in the 1976 and 1977 Sampling
Seasons 	 119
VI-1	Percent Mortality of Folsomia Candida After 6
Days Exposure to Field-Collected, Guthion-
Contaminated litter at 21 C. Replication:
10 Individuals per each of 10 replicate con-
tainers 	 I23
xiii

-------
TABLES
Number	Page
VI-2 Percent Mortality of Collembola After 7 Days
Exposure to Guthion in Yeast and in Soil
at Two Moisture Regimes, at 15.5 C. Repli-
cation: 40 Individuals of Each Species in
Each of 5 Replicate Containers per Dosage 	 123
VI-3 Percent Mortality of Collembola After 24 Hours
Exposure to Guthion in Soil (25$ moisture)
and in Yeast, at 15.5 C. Replication: 200
Animals of Each Species per Dosage 			 124
VI-4	Percent Mortality of Tidabius tivius (Chilopoda)
After 7 Days on Guthiop-Contaminated Soil
(25% moisture) at 26.7 C. Replication: 20
Individuals per Stage and Concentration 	 125
VI-5 Toxicity of Guthion to Adult Isopods at 15.5°C
(Soil: 25% Moisture), as Percent Mortality
After 7 Days Exposure 	 126
VI-6	Replication in Diazinon Toxicity Experiments
With T. rathkei: Numbers of Animals Observed
per Weight Bin and ppm Diazinon, at 26.7 C 	 127
VI-7	Results of GLC Analyses of Diazinon-Contaminated
Soil (Adjusted for GLC Extraction Efficiency)
Prior to Using the Soils for Toxicity Exper-
iments with T. rathkei 	128
VI-8	Percent Mortality of T. rathkei Exposed to
Diazinon-Contaminated Soil at Three Tempera-
tures. Values Not in Parentheses are Cumula-
tive 7-Day Mortalities; Values in Parentheses
are 10-Day Mortalities (Same Animals, 3 Days
Later). Weight Bins as Specified in Table
VI-6 Were Combined Into Larger Weight Classes,
to Obtain a Minimum of 5 Replicate Animals
per Class 	 129
VI-9	7-Day Mortality Percentages (Corrected for Con-
trol) for T. rathkei Reared on Guthion- and
Diazinon-Contaminated Soil. Data Included
Only if a Minimum of 10 Replicate Test Animals
Had Been Observed Per Weight Class and
Dosage 	 130
VI-10	Relative Toxicity of Guthion and Diazinon to
Trachelipus rathkei at Three Temperatures.
For Details See Table VI-8 and Chapter X 	 131
xiv

-------
Page
131
139
139
139
141
141
143
146
146
147
154
160
TABLES
Toxicity of Diazinon-Contaminated Soils tg
II. Instar Trachelipus rathkei at 15.6 C,
Contrasted to Guthion Toxicity at the same
Temperature. All Data Corrected for Control
by Abbott's Formula. Replication: Minimum
of 40 Animals Per Run, (Two Per Replicate
Container) 	
Duration of Gravidity, in Days, of Trachelipus
Females Kept at Three Temperatures 		
Duration, in Days, of the Rest Interval Between
Brood Release and the Next Brood Molt of
Trachelipus 	
Number of Consecutive Broods Produced by
Trachelipus Females in the First Breeding
Period (Percent in italics) 	
Second Breeding Period of Trachelipus Females
Reared at Constant Temperatures 	
Percent Unsuccessful Broods in the First Period
of Trachelipus, per Weight Class and Temper-
ature*	
Number of Days Needed by Trachelipus Females to
Reach the Final.Mean Weights (+ S.D.) of Five
Weights Classes 	
Cumulative Total Mortality of Non-Breeding
Trachelipus, in Percent of all Individuals
Observed Through Each of the Four Weight
Classes (Only the Lowest Class Includes Males
and Females, in Unknown Proportions) 	
Percent Mortality Occurring Either During Gravi-
dity or During the Post-Parturient Molt of
Breeding Trachelipus Females 	
Summary of Board Collections of Trachelipus in 1978 .
Breeding Periodicity of Trachelipus Females in the
Field, Categorized by Estimated Weights 	
Mean Number of Trachelipus rathkei Per Board
on Each Collection Date 	
XV

-------
TABLES
Number	Page
X-l Number of Days Needed for T. rathkei to Grow
Through Each Weight Class at Three Tempera-
tures. In Parentheses: Final Mean Weights
+ S.E	 165
X-2 Percent Mortality, Within Four Weight Groups,
of Non-Breeding T. rathkei Females at Three
Constant Temperatures 	 167
X-3 Percent Mortality of T. rathkei Females Asso-
ciated With Each Brood Within One Breeding
Period (Three Consecutive Broods Occurred
at 26.7 C Only). Assignment to Weight Classes
is Based on Weights Just Prior to Each Brood 	 167
X-4	Degree-Day-Based Weight Class Durations for
Breeding and Non-Breeding T. rathkei Females 	 169
X-5	Field-Collected Litter: Results of Analyses
for Total Azinphosmethyl Residues, and Rela-
tionship of Collection Date to the Last Pre-
vious Spray Application Date 		 I7*
X-6	Cumulative Mortality Percentages for T. rathkei
Reared on Contaminated Soil for 7 Days. Data
Corrected for Control by Abbott's Formula.
*) Indicates that the Values are Means Obtained
From at Least 2 Replicate Runs (Containing 20
Animals Each) 		 175
X-7 Daily Mortality Records (Cumulated Percent) for
II. Instar T. rathkei Reared on Azinphosmethyl-
ContaminatecT Soil at Three Temperatures (20
Animals Per Pesticide Level and Temp) 	 176
X-8	Cumulative Percent Mortality, Over 6 Days, of
Trachelipus of Various Weights Exposed to
Field-Collected Litter Contaminated With
Guthion (21 C Constant Temperature), (n =
Total Number of Animals Per Run, Distributed
as 2 Per Container) 	 178
X-9	Comparison of Field Data With Model Predictions.
Columns 3 and 4 are the Ratios of Numbers of Pre-
Reporductive (<12 mg) T. rathkei Collected Per
Sampling Board From Control Plots to Number Per 5
Board From Sprayed Plots, for Each Sampling Day.
Last Two Columns are the Model's Prediction,
Given the Environmental and Pesticide Applica-
tion Inputs Observed in the Field, of the Ratios
of <12 mg and ^ 12 mg Populations From Unsprayed
and Sprayed Plots on the Same Day 	 191
xv i

-------
Summary and Recommendations
A conceptual model was developed to describe aspects of the fate and
effects of a pesticide In an orchard ecosystem. In order to refine, para-
meterize, and test a mathematical model based upon this conceptual model, a
program of field and laboratory experiments was undertaken.
The environmental behavior of azinphosmethyl was studied in a Michigan
apple orchard watershed to gather data for the model on initital distribu-
tion within the orchard, vertical movement of the pesticide under the influ-
ence of rainfall, loss from the orchard with runoff, and effects of the
pesticide on several orchard Invertebrate populations. The estimated pro-
portion of a low-volume application initially distributed within the orchard
averaged .624 (standard deviation of .149) over three seasons (1976-1978).
Examination of residues reaching each layer showed the majority of the dis-
lodgeable residues were distributed to the trees and grass-broadleaves. The
litter-moss and soil contained residue levels roughly ten times lower than
tree leaf residues. Runoff studies Indicated loss, via this route, of less
than 1% of azinphosmethyl residues present in the orchard. The residue data
were used to parameterize a model for azinphosmethyl attenuation and move-
ment in an orchard ecosystem. Rates of pesticide attenuation within, and
movement between, specified orchard compartments were determined under various
rainfall regimes. The output of this model was structured to allow the esti-
mation of the time course of azinphosmethyl exposure to ground-dwelling
invertebrates. Mean squared errors for the comparison of the model predic-
tions with an independent set of residue data indicated good prediction of
azinphosmethyl fate within the tree and grass-broadleaves layers. Prediction
of pesticide dynamics within the litter-moss and soil layers was much more
difficult. The model estimates that under dry conditions 25% of the daily
loss of azinphosmethyl from the orchard trees is due to movement to other
parts of the orchard. Greater movement is predicted under rainfall conditions.
Estimates of daily airborne loss determined from deposit residues and direct
sampling of airborne residues suggest that airborne loss is largely responsi-
ble for the early loss of residues, accounting for 40% of the daily loss
rate on day 3 of the first spray period, 1978 season. A multi-component
kinetic model is presented for estimating simultaneously the early airborne
loss of foliar deposits and the often slower dissipation of the remaining
residues. Further model development to include the effect of selected en-
vironmental parameters on azinphosmethyl degradation was also undertaken.
The generalized model developed, entitled the Pesticide Orchard Ecosystem
Model (POEM), includes as a special case the model for the azinphosmethyl
applications under the conditions of this field study. POEM also includes
facilities for altering parameters to describe effects of other formulations,
other pesticides, other application procedures and/or other field conditions.
The field sampling program for invertebrates included pit-trappping,
block sampling for earthworms, and cryptozoan board sampling for isopods.
Relative population densities in control plots versus plots sprayed with
azinphosmethyl, followed in some cases by diazinon, were determined.
xvii

-------
Laboratory cultures of the isopod Trachellpus rathkel and the collem-
bolan Folsomia Candida were established. Laboratory data were obtained to
aid in parameterizing both life cycle models and azinphosmethyl toxicity
models for populations of these organisms. Limited lab studies of azinphos-
methyl toxicity to earthworms (Allolobophora caliginosa) and spiders (Lycosa
lugubris) were also performed. The four models developed for invertebrates
are components of the POEM model, and use daily predicted pesticide levels
in various environmental compartments (e.g., grass/broadleaves, litter/moss,
soil) to predict pesticide-induced mortality of the populations over long
periods of time. The model, laboratory data, and field data for T. rathkei
are presented in detail in this report. For the other organisms, the field
and laboratory data are presented, but the models are described only in more
general terms.
Two appendices describe (A) the procedure used to process the field
data on pesticide residues to provide parameters for the fate portion of the
POEM model, and (B) the Users' Manual, including sample output, for the POEM
model.
Recommendations
(1)	Further work to refine, parameterize, and test the components of the
POEM model, or similar models, for other pesticides and other conditions
should be undertaken. In many cases, the current froms are derived based
on very sparse data in the literature. While predictions based on these
forms may be very informative and useful in some contexts, they are not
likely to be very accurate for predicting actual fate and Impacts of pesti-
cides until they have been carefully refined based on currently non-existent
data. Nevertheless, the present model may be helpful because it allows the
user to determine the Implications of various sets of assumptions about
pesticide dynamics and effects.
(2)	Work on models for the long-term effects of pesticide exposure on
populations of Invertebrates should be continued. While this study Includes
a reasonable model for effects of azinphosmethyl on isopods and less refined
models for collembola, earthworms, and spiders, the methodology should be
extended and refined through application to other pesticides and organisms.
(3)	The model presented here does not provide an overall Indicator of the
ecosystem-level impact of a pesticide in a particular situation. While
impacts on individual populations are likely to be key components of any
sound measure of overall impact, the importance and role of each population
In the ecosystem must also be defined and incorporated in the measure.
Research aimed at identifying key populations and modeling their functions
should be undertaken. The search for integrating measures or indicators of
ecosystem stress or damage for terrestrial systems should be broadened and
intensified.
xviii

-------
ACKNOWLEDGEMENTS
The authors wish to acknowledge the assistance and guidance of their
project officers at the Environmental Research Laboratory in Corvallis, Dr.
James Gillett and later Dr. Jay Gile. The contributions of many at M.S.U.
are deeply appreciated. Jay Hestbeck, Mehrdad Tabatabai and Paul Weston,
graduate students in Systems Science, contributed heavily to the modeling
work. Programming was done by Tom Julien, Steve Walsh, and Mike Sayen.
Chemical analyses, while directed by Drs. Kon, Jenkins and Zabik, involved
many student laboratory assistants, whose contributions are appreciated.
Roseann Miller and Kay Grimnes assisted in laboratory rearing of invertebrates.
The contributions of Drs. A. W. A. Brown, W. E. Cooper and J. W.
Butcher, particularly to the early phases of the project, are gratefully
acknowledged.
The authors are grateful the excellent contribution of Gwen Counseller
and Julia Pond to the preparation of this report.
xix

-------
INTRODUCTION
Since the discontinued use of many organochlorine insecticides in the
early seventies, use of the generally more toxic and less persistent
organophosphate insecticides has increased (Brown, 1978). Because of the
greater mammalian toxicity of these compounds (Matsumura, 1975), research
of their fate in orchards has been primarily concerned with worker reentry
hazard associated with foliar dislodgeable residues, soil, and soil dust
residues. Most of this work has been done for citrus (Gunther et al.,
1977; Nigg et al., 1977; Thompson and Brooks, 1976; Spear et al., 1978),
but studies have also been conducted for peach (Winterlin et al., 1975;
Hansen et al., 1978) and apple orchards (Staiff et al., 1975; Hansen
et al., 1978). Little work has been done in the area of attenuation and
movement of the organophosphates between trees, ground cover, and soil, or
loss with runoff from deciduous fruit orchards.
The lack of cultivation, which is the practice in many deciduous fruit
growing areas (Haynes, 1980), preserves greater species diversity than is
found in many other agroecosystems (Brown, 1978). This diversity can
result in a relatively stable deciduous tree fruit ecosystem if undisturbed
by pesticides or other management practices (Hoyt and Burts, 1974). Under
these conditions, natural control of many arthropod pests through disease,
predation, and parasitism is achieved (Croft, 1975). However, to produce
marketable fruit, pesticides must be used to control a few key pests
(Glass and Llenk, 1971). Control of these key pests with pesticides may
reduce natural enemy populations, promoting secondary pests to a major
pest status, often requiring additional chemical control (Hoyt and Burts,
1974; Croft, 1978). In addition, the increased use of pesticides in
orchards, and on all crops, has resulted in an Increased number of resistant
pest species (Smith, 1976; Brown, 1977, 1978; Croft, 1978). These and
other problems associated with the use of pesticides as a sole means of
control have resulted in a greater emphasis being placed on integrated
pest management as a long-term strategy for the economic control of crop
pests (NAS, 1969; Smith, 1976). Such programs are designed to use a broad
spectrum of control measures including biological, cultural and chemical
methods. This approach to pest management requires a great deal more
knowledge not only of the ecobiology of both beneficial and harmful species,
but also of pesticide fate and effects. The success of orchard IPM pro-
grams may depend largely on the judicious use of pesticides allowing maximum
benefit to be gained from biological control measures (Croft and Brown,
1975; Smith, 1976). This will require a better understanding of pesticide
fate throughout the entire orchard ecosystem.
Additional consideration must be given to species other than those of
benefit to pest control. Because of their relative permanence, deciduous
orchards are often a habitat for wildlife indigenous to natural ecosystems
(Croft, 1978). Detailed studies of pesticide fate and effects, which
would be impractical or undesirable in a natural setting, can be done in
xx

-------
an orchard, to give some insight into the ecological hazards associated
with pesticide exposure to the fauna of these environments. Such informa-
tion may be valuable in the development of environmental fate and ecological
effects testing guidelines as proposed under the Toxic Substance Control
Act and the recently amended Federal Insecticide, Fungicide and Rodenticide
Act.
The Toxic Substances Control Act authorizes the Environmental Pro-
tection Agency (EPA) to obtain from industry data on the production, use
and effects on human health and the environment of chemical substances
and mixtures. Testing standards under Section 4 are directed at specific
chemical characteristics and effects, including oncogenicity, teratogenicity,
mutagenicity, and other health effects, as well as environmental fate,
persistence and ecological effects. All tests are substantially similar
to methods proposed by EPA for its pesticide registration program and rep-
resent a "base set" of tests from which to select specific tests for
specific chemicals in test rules. To date no test standards have been
proposed for environmental fate or ecological effects.
In addition to the Section 4 standards which are designed to test
specific chemicals or chemical groups, the EPA also plans to issue testing
guidelines consistent with these testing standards, but more general in
nature, encompassing a wider range of chemical substances and effects of
concern to EPA. These guidelines will also be used under Section 5 which
allows the agency to limit manufacturing, processing, distribution, use
or disposal of new chemical substances (or new uses of existing substances),
pending development of information. The administrator is given this
authority if he has reason to believe (1) that the chemical substance may
present an unreasonable risk or may result in substantial human exposure
or environmental release, (2) there are insufficient data or experience
for determining or predicting health or environmental effects, and (3) testing
is necessary to develop such data. TSCA does not require EPA to provide
testing guidance, except when specifically requested under Section 5 (g).
Nevertheless, testing guidelines are presently being developed in an effort
to streamline the monumental task of TSCA compliance and to encourage more
appropriate and cost-efficient information gathering.
So far the EPA has proposed general standards for a number of human
health effects. Such standards for environmental fate and ecological
effects are presently at the interagency review stage. To develop generic
standards for human health involves only one species, man. Types of
exposure (dermal, inhalation, ingestion, etc.), metabolism, and pharmaco-
dynamics are fairly uniform. Generic test standards may need only be
modified based on the properties of the chemical or chemical category.
Generic standards for environmental fate and ecological effects involves
all species of flora and fauna. In addition, all biotic-abiotic relation-
ships must be considered from ocean food chains to soil microbial communities.
This diversity of species (and species interaction) and environments makes
the development of generic standards to assess the possible hazards of
chemical substances to the environment a formidable task. The first problem
would seem to be how many different species must be tested to adequately
assess ecosystem effect; secondly, what chemical properties are important,
and, thirdly, what and how do environmental properties influence exposure
and possible ecotoxicity. Due to the enormous complexity, there has been
much controversy over the rationale for adopting standards in these areas
where scientific methods are less developed and where no validated techniques
are available for certain effects.
xxi

-------
There may be unreasonable risk to an ecosystem when a component or
components are exposed to a concentration of a chemical substance which
causes harm. Environmental fate studies are used to predict and estimate
the presence of potentially harmful chemical residues in man-made and natural
environments. A chemical released into the environment may be transformed
by chemical or photochemical reaction, be metabolized by living organisms or
persist unaltered. Some degradation and transformation products can be
hazardous in their own right (Menzie, 1972; Crosby, 1973; Goring et al.,
1975). The goal of environmental test standards should be to identify the
dominant pathways by which a chemical degrades, dissipates, and accumulates,
and then relate this behavior to various chemical, physical, and biological
conditions. This knowledge can then be used to infer what biota may be
exposed to the chemical, by what route, how often, and in what concentrations.
Thus, both the population at risk and the effect of the chemical are deter-
mined partly by the detailed fate of the chemical substance. Just as in
health effects testing, where jbn vitro studies and lab animals are used to
estimate chemical effects on man, ecosystem effects and chemical fate are
estimated primarily from laboratory studies. The foundation for use of
laboratory data for environmental fate assessment is based on the following
assumptions:
(1)	The rate of transformation or transport of a chemical in or from
an environmental system is the sum of the rates of known individual
chemical, physical, and biological processes and the rate of
equilibrium constants for these processes can be predicted or
measured independently in the laboratory.
(2)	The laboratory data for individual processes can be integrated and
extrapolated to the appropriate set of environmental conditions.
It is yet to be seen if any set of laboratory tests can be used to predict
the environmental fate of a chemical substance to the extent that the EPA
may adequately assess the possible risk to an ecosystem. The fact is that
few of the processes which determine the environmental fate of a chemical
substance have been studied in enough detail to predict dominant pathways
or rates of change in the "real world." Each possible transformation or
transport pathway can be studied in the laboratory and rates may be deter-
mined under a given set of "environmental conditions" but how actual
environmental conditions influence rates is largely unknown.
A few years ago it was thought that microcosms, controlled laboratory
systems that attempt to simulate some selected portion of the real world,
might be used to fill this information gap. In a terrestrial microcosm
environmental parameters such as temperature, light level, water content,
and organism diversity can be controlled and varied by the investigator.
A common approach to microcosm development has been to organize some elements
of a selected ecosystem in a container so as to resemble some aspects of the
ecosystem of interest and then systematically to change the complexity by
adding new trophic levels of organisms. Parameters of the system as they
affect chemical fate are monitored (Gillette and Witt, 1979). Unfortunately,
development of the physical models of these microcosms has not always kept
pace with conceptual development. More importantly, complete understanding
of the discrete chemical, physical, biological, and climatological processes
that control chemical fate in a terrestrial system is not yet directly
obtainable from these studies.
Another approach has been to use mathematical models. The development
and use of mathematical models usually involves a systems approach; the
formalized analysis of any system or of the general properties of systems.
xxii

-------
The analysis of complex systems as systems, and the modeling of these
systems, is contrary to reductionist trends in science. Laboratory experi-
ments which isolate and control very small components of nature have up to
the present been the most powerful investigative tools aiding man in under-
standing nature, but this "science by isolation" has its drawbacks as the
larger environment may influence the components of interest to such an
extent that an isolated laboratory experiment may exclude some critical
components, or the behavior of a system may not be simply the sum of the
behaviors of its parts in isolation (Hall and Day, 1977). Mathematical
modeling of environmental fate should be based on both reductionist labora-
tory experiments and key observations of the fate of chemicals in the
environment to include the rates of transport and transformation as a
function of climatic and other environmental variables. Only a handful
of models have been developed which describe chemical environmental fate
in general. The present "state of the art" model is EXAMS, a model of fate
of toxic organic chemicals in aquatic ecosystems, developed by the EPA's
Athens, Georgia Research Laboratory (Lassiter et al., 1978). This model
was used to develop environmental assessments for eleven chemicals in
aquatic systems based on laboratory measurements. The model has since
been modified to include a library of canonical environments and is now
in use on a trial basis in several laboratories and EPA offices.
The work presented here is one portion of an effort to characterize
the dynamics and effects of an example compound in the terrestrial environ-
ment, utilizing primarily field measurements and the methodology of systems
modeling and simulation. Data collection, model refinement, and revised
experimental design were done iteratively, yielding a model which is
parameterizable and data which are relevant to the problem being attacked.
The study of pesticide dynamics through iii situ field studies is
difficult due to the lack of natural or planned experiments (inability to
control much of the variance, i.e., climatic conditions) and the relatively
high levels of error associated with field data. Modeling techniques were
employed to aid in the understanding of the necessarily large amount of
field data needed to construct a "meaningful" picture of the pesticide's fate.
The field experimental program used to investigate the distribution,
attenuation and movement of the organophosphate insecticide azlnphosmethyl,
0,0-dimethyl-5-(4-oxo-l,2,3, benzotriazin-3(4H)-ylmethyl) phosphorodithioate
(Guthion^), in a Michigan apple orchard is given in Chapter I. The compound
was followed from its spray application through the orchard vegetation/
litter/soil environment and into aquatic systems. The form of the model
describing azlnphosmethyl movement and attenuation, as well as data handling
procedures and the derived rates, are presented in Chapter II. Observations
were made concurrently within the same orchard to examine the effects of
azlnphosmethyl on several ground-dwelling invertebrates, including detailed
studies of the isopod Trachelipus rathkel (Snider, 1979; Snider and Shaddy,
1980). Field and laboratory data collected on T^. rathkei were used to
develop a model describing its ecobiology and temporally-distributed
pesticide-induced mortality. The output of the fate model described in
Chapter II was used to determine the time-course of azlnphosmethyl exposure.
In Chapter III the field experimental program used to determine
azlnphosmethyl airborne residues is presented. A multi-component kinetic
model used in the assessment of the contribution of airborne loss to the
overall attenuation of deposit residues is also described.
xxiii

-------
In Chapter IV, movement and attenuation of azinphosmethyl are examined
as a function of environmental conditions. A computer simulation is des-
cribed which allows the user to predict the fate and effects of aziophosmethyl
on several types of organisms.
Chapter V describes the results of the field sampling program for
invertebrates in the orchard plots, providing information on the effects
of aziophosmethyl spraying (additional material on the isopod T. rathkei
is found in Chapter IX). Chapter VI contains the results of a laboratory
assessment of the toxicity of asinphosmethyl and diazinon to various
invertebrates. Chapter VII briefly describes the models developed for
spiders, earthworms, and springtails. Chapter VIII presents a detailed
description of the ecobiology of the isopod Trachelipus rathkei, while
Chapter IX describes the effects of the aziophosmethyl spray program on
the T. rathkei field population. Chapter X presents the model for
T rathkei. including both its general life cycle and its response to
pesticide exposure.
Appendix A documents the data analysis procedures employed locally
at MSU to parameterize the model. Appendix B is the users' guide for the
Pesticide Orchard Ecosystem Model (POEM) described in this report.
xxiv

-------
CHAPTER I
EXPERIMENTAL DATA FOR THE AZINPHOSMETHYL FATE MODEL
Reported here is the field experimental program used to investigate the
distribution, attenuation, and movement of the organophosphate insecticide
azinphosmethyl, 0,0-dimethyJ-5(4-oxo-l,2,3, benzotriazin-3(4H)-ylmethyl)
phosphorodithioate (Guthion ), in a Michigan apple orchard. Studies were
carried out to gather data on both initial distribution of azinphosmethyl
within the orchard and vertical movement of the pesticide under the influ-
ence of rainfall, as well as the attenuation of the pesticide in various
situations. In addition, plots were designed such that each was a separate
watershed, allowing runoff collection from them individually. The compound
was followed from its spray application through the orchard vegetation/litter
soil environment and into aquatic systems, to determine possible exposure
of the orchard ecosystem biota to the compound. The data presented here
were used to parameterize a model (presented in Chapers II and IV) for azin-
phosmethyl attenuation and movement in an orchard ecosystem.
ANALYTICAL METHODS
Organic solvents were glass-distilled and the water was distilled-de-
ionized. After washing, the glassware was rinsed with acetone and hexane
followed by overnight heating at 250°C.
Dislodgeable Residues. Dislodgeable residues were determined by the proce-
dure of Gunther et al. (1973). Leaf discs were extracted with 50 ml portions
of water containing two drops of Triton X-100 in water (1:50,v/v). Samples
were then agitated on a wrist-action shaker for 15 minutes. Depending on
the bulkiness of the grass, litter, or moss, more water was used to insure
good contact with the sample during extraction. One drop of Triton-X 100
solution was added for each additional 25 ml of water used. All samples were
given a final rinse by hand with a third 50 ml portion of water. The com-
bined washes were partitioned three times against 50 ml portions of hexane
which was then concentrated on a rotary evaporator and diluted to a proper
volume for analysis. Using the above method, recovery and standard devia-
tion from the Triton-X 100 water solution fortified with azinphosmethyl
standard was 95 + 3%.
Surface-Penetrated Residues. To estimate the magnitude of the remaining re-
sidues on or within the cuticular matrix, selected samples collected on the
day of application were given an additional extraction in a manner similar
to that described by Steffens and Wieneke (1976). Three 20 ml volumes of
acetone-water extracts were concentrated by rotary evaporation to remove the
acetone. The water left in the rotary flask was partitioned in the flask
against three 30 ml volumes of hexane which were combined and then dried over
anhydrous sodium sulfate and concentrated to about one ml. This procedure
was fleeted as Wieneke and Steffens (1974) were able to recover 100.4% of
the C azinphosmethyl applied to bean leaves as an aqueous formulation, one
1

-------
day following application.
This extract was cleaned up by the silica gel microcolumn described by
Kadoum (1967) with our elutlon system. One and a half grams of unactlvated
Adsorbosll CAB, 100/140 mesh (Applied Sciences, Inc.) was packed as a hexane
slurry Into a Pasteur plpet. The sample was added In hexane and the column
eluted with 15 ml of 10% benzene In hexane. Ethyl acetate (4%) In benzene
then eluted the azlnphosmethyl In 15 ml. Recovery and standard deviation
of 2 ml of a 5 ppm azlnphosmethyl standard solution added to the column was
94 + 3%.
Soil. Extraction of soil was based on Schulz et al. (1970). Approximately
50 g of soil (largest roots removed) were mixed with 10 ml of water, If dry,
and acetone (1.5 inl/g) on a wrist-action shaker for 15 minutes. The solvent
was poured Into a rotary flask through a funnel lined with Whatman No. 1
filter paper; as much soil as possible was retained in the extraction flask.
A second extraction followed and the combined extracts were treated exactly
as the surface-penetrated residues of the previous section, including the
clean-up column. Recovery and standard deviation from soil fortified with
azlnphosmethyl standard was 84 + 9%.
Filter Paper Targets. Targets, consisting of two circles of Whatman No. 1
paper, 18.5 cm in diameter (attached one atop the other, by a single staple
to a cardboard backing), were extracted for two hours in a soxhlet extractor
using a mixture of hexane:acetone, 100 ml:25 ml. Recovery and standard de-
viation from targets fortified with azlnphosmethyl standard was 104 + 5%.
Pesticide Runoff Studies. Azlnphosmethyl In runoff samples collected during
the 1977 season was examined by extracting one liter subsamples following
centrifugation to remove sediment. The sediment was soxhlet extracted for
four hours with a 1:1 hexane:acetone mixture, air dried and weighed. The
supernatant was extracted with hexane and both sediment and supernatant ex-
tracted with hexane and both sediment and supernatant extracts were concen-
trated by rotary evaporation. Recovery and standard deviation from runoff
water fortified with azlnphosmethyl standard was 93 + 4%.
Quantitation of azlnphosmethyl was accomplished using a Tracor 560 gas
chromatograph having a flame photometric detector in the phosphorus mode.
A six-foot glass column (2 mm i.d.) packed with 3% SE-30 on Gas Chrom Q,60/
80 mesh, was operated at 195 C; N_, 40 ml/mln; air, 90 ml/mln; H., 60 ml/
min. This system was Interfaced with a Digital PDP 8 Pamlla PDP 11/40 RSTS
computer for integration of the area under the single peak produced.
THE EXPERIMENTAL SITE
The experimental apple orchard was located In the vicinity of Grand
Rapids, Michigan in Kent County. The twelve-year old trees were a mixture
of semi-dwarf duchess and wealthy cultlvars. In 1976 the orchard had been
left unattended for approximately five years. This was a necessary require-
ment to assure adequate populations of ground-dwelling invertebrates, un-
affected by previous seasons' pesticide applications (Snider, 1979). In the
early spring the trees were pruned and understory brush was removed. Just
prior to the first spray application the area was mowed to a height of ap-
proximately five cm. Due to dry conditions further mowing was not necessary.
In 1977 light pruning was repeated and the orchard was mowed just prior to
the first, second, and fourth spray applications. A survey of the ground
cover by the line-transect method (Cox, 1974) was carried out in 1976; cover
2

-------
type categories were established and their Importance values calculated.
Above the soil (a marlette sandy clay loam, pH 6.0, 56% sand, 20% silt, 24%
clay, and 6.0% o.m. in top 10 cm) was moss intermingled with litter; the
most recent litter deposits covered the moss. The herbaceous growth cover-
ing the moss and litter was predominantly broad-leaved weeds, with grass be-
ing of secondary importance. An examination of the ground cover in the areas
sampled for residue analysis yielded similar results. This ground cover
composition is not atypical of orchards in the temperate Eastern United
States and Canada, although they vary with the level of secondary succession
(Whlttaker, 1975), climatic, topographic, edaphic conditions, and management
practices (Klingman and Ashton, 1975; Teskey and Shoemaker, 1978; Schubert,
1972).
PREPARATION OF ORCHARD FOR RUNOFF COLLECTION
To facilitate the study of azinphosmethyl loss via runoff, four experi-
mental plots and one control plot were established, each a separate water-
shed (Figure 1-1). Slopes in plots 1 and 2 and the control were between 6 and
12%, slopes in plots 3 and 4 were between 18 and 25%. All the runoff was
collected from a given plot for each event and subsampled for pesticide anal-
ysis. A 600-liter galvanized steel livestock tank (61 cm x 61 cm x 183 cm)
was placed in an excavation at the natural drainage point for each plot, at
a level to collect the runoff. A smaller tank (40 liter capacity) was placed
inside the larger tank so that it would fill and any overflow would be con-
tained in the larger tank. Also, an overflow diversion pipe was Installed
on the collection tank for plot 1; in the event the large tank overflowed,
the runoff water would be diverted to a point below plot 4 (see Figure 1-1).
Runoff water was finally channeled into the tanks by galvanized steel collec-
tion chutes, 61 cm long with sides 15 cm high, 61 cm wide at the top and 30.5
cm wide at the bottom. A 61 cm by 61 cm area of the tank directly below the
chute was covered with a 6.4 mm mesh screen and the rest of the tank was
covered with plywood covered with plastic, to keep out rainfall, animals,
and debris. The chute and screen were also covered with plastic. Both the
tanks and chutes were coated with epoxy paint, as tests showed that after
seven days, the residues remaining from a five ppb solution of azinphosmethyl
In galvanized steel containers were 30% less than the residues remaining in
glass, stainless steel, or epoxy-painted containers. Azinphosmethyl hydrolysis
experiments (see Part IV) indicated little degradation over the seven-day
period, at solution pH values below 7.0 (first order rate constant of .008 at
25°C). As the pH of the runoff water collected averaged 5.9 + 0.2, losses
by this mechanism prior to sample collection and extraction were thought to
be small. Loss via volatilization was also congidered to be negligible as
azinphosmethyl has a vapor pressure < 7.5 x 10 mm Hg at 20°C (Mobay Chemical
Co., personal communication) and the maximum concentration of azinphosmethyl
measured in the runoff collected was 22 ppb, well below its water solubility
of 33 ppm.
Plots were enclosed and separated from one another by a 30.5 cm, 20
gauge aluminum core fence buried approximately five cm in the ground. This
fence was made continuous with the collection chute to aid in the final chan-
neling of runoff at the bottom of the plot slope. Fencing was installed with
minimal disturbance of the ground cover and in such a manner that potential
runoff would have as little contact as possible with the fencing. This col-
lection scheme was designed and implemented during the 1976 season, but due
to dry conditions, no runoff was collected that year.
3

-------
t
CONTROL
(.158 ha)
A




PLOT 3
PLOT 1

(.052 ha)
(.099 ha)
A (
PLOT 4
/ PLOT 2

(.052 ha)
/ (.060 ha)



A
a RUNOFF COLLECTION POINTS
» <
8 M
Figure I-l. Orchard Plots
4

-------
SPATIAL STRUCTURE OF THE ORCHARD PLOTS
The orchard plots (Figure 1-1) were subdivided both vertically and
horizontally in order to represent adequately the variations which affect
the behavior and impact of the pesticide in the orchard ecosystem.
Horizontally, each plot was subdivided into four regions potentially
different with respect to inital pesticide distribution (see Figure 1-2).
Vertically, each alley region was divided into three compartments: grass-
broadleaves, litter-moss, and soil. The canopy regions contained these com-
partments plus a compartment for the leaves of the tree.
Region 4 (under the canopy) was determined to be 32% of the area in
plot 1, 34% in plot 2, 54% in plot 3, and 45% in plot 4. The remainder of
the plot area was divided among the three alley regions in a ratio of 2:3:1
for plots 1 and 2, and 1:2:1 for plots 3 and 4. The difference in the ratios
was due to the closer spacing of the rows along the North-South axis in plots
3 and 4. The number of trees in each of the plots was as follows: plot 1,
24; plot 2, 15; plot 3, 10; and plot 4, 11. Tree heights average 3.0 meters.
SPRAYING AND SAMPLING SCHEDULE
The size of the plots and their inaccessibility due to the aluminum
fences made the use of commerical spray equipment impossible. What was needed
was a plot sprayer that could: 1) be pulled by a small tractor, such as a
garden tractor, and 2) match as closely as possible the coverage of a com-
mercial spray unit. To meet these criteria, a low pressure, low volume
sprayer similar to one developed by Howitt and Pshea (1965) was constructed.
The sprayer was mounted on a trailer and was pulled by a John Deere model
200 garden tractor. Mounted on the trailer was a double inlet, high volume,
low velocity blower manufactured by Dayton, model 3C011. The blower was
powered by an eight horsepower Briggs and Stratton gasoline engine. Blower
output was approximately 10,000 cubic feet per minute (at the highest engine
speed setting). The air flow was directed by a fiberglass deflector, con-
structed for use on Agtech crop sprayers. One side of the deflector was
widened to accommodate three Beecomlst model 350 mini-spin spray heads. The
deflector was further modified between 1976 and 1977 seasons to direct more
of the spray into the trees. Pesticide was delivered to the spray heads by
a Masterflex variable speed tubing pump.
The pesticide application rate was determined by the rate at which the
pesticide mixture was delivered to the spray heads, tractor speed, and dis-
tance covered to spray each plot. The actual application time was recorded
in the field for each plot and the actual amount of pesticide applied was
determined by tank residual volume after spraying. In addition, during the
first spray period of 1978, samples of spray mixture were taken, at the spray
heads, before and after spraying each plot. Results from this method were
comparable to those of the comparable to those of the residual volume method.
The average application rate and standard deviation was 1.62 + .33 kg/ha 50%
w.p./100Jl/ha in 1976 and 1.64 + .38 kg/ha 50% w.p. /100Jl/ha in 1977. Plots
were sprayed in sequence between 7 a.m. and 9 a.m.
Deposit Residue Samples
To determine initial horizontal distribution of azinphosmethyl, a series
of 15 to 20 ground-located filter paper targets was deployed among the four
horizontal regions (see Figure 1-2) in each plot on the date of spraying.
5

-------
CANOPY
REGION (4)
CANOPY
REGION (4)
ALLEY REGION (2)
ALLEY REGION (II
ALLEY REGION (3)
ALLEY REGION M>
CANOPY
REGION (4)
CANOPY
REGION (4)
ALLEY REGION 12)
Figure 1-2. Overhead View of Horizontal Regions o ' rchard Plots

-------
Targets were placed in the center of the alley regions, and at the in-row
midpoint between the tree trunk and the edge of the canopy region. As soon
as a plot had been sprayed, the filter papers were removed from the backing,
packaged, and transported in picnic coolers cooled by dry ice. Storage was
at -20°C until analysis.
The initial distribution of azinphosmethyl was also determined among
the three ground layers: grass-broadleaves, litter-moss, and soil. In each
plot, two samples of each layer were taken from the tree and alley regions.
A square of sod (15 cm x 15 cm) was cut to a depth of about four cm and the
grass-broadleaf plants, litter, and moss were separated into three discrete
samples for residue analysis. From the bare sod remaining, three cores (5
cm x 7 cm deep) were taken for analysis providing a sample of 40 to 60 g.
For analysis of leaf deposits a 2.0 cm diameter disc was taken from the
center of each leaf. Three trees were sampled in each plot, taking 30 or more
discs per tree. Samples were taken at shoulder height (distributed within
the canopy easily reached) around the entire circumference of the tree.
Sampling dates for the 1976 season are shown in Table 1-1. Infre-
quent rainfall during this season yielded little information usable to assess
movement due to rainfall. In 1977, lnterspray sampling was optimized for
movement data as samples were collected primarily after rainfall events (see
Table 1-2).
Residues in Runoff
Azinphosmethyl movement out of the orchard with runoff was also deter-
mined. Runoff water was removed from the orchard in four-liter, dark glass
bottles and stored at 5°C until analysis. Excess runoff was measured volu-
metrically in the field and discarded.
Rainfall was measured with static rain gauges and this Information was
correlated with data from recording rain gauges maintained 3/4 mile from the
orchard by the Michigan State University Institute of Water Research. Tem-
perature and humidity data shown in Figures 1-5, 1-6, 1-7 and 1-8 were ob-
tained from the Michigan Weather Service, Grand Rapids, Michigan.
RESULTS AND DISCUSSION
Initial Horizontal Distribution
Table 1-3 and 1-4 show the results of the analysis of the filter paper
targets for the first spray day of the 1976 and 1977 seasons. Tl^average
amount and standard error (SE) of pesticide are expressed In wg/cm ground
area and proportions shown are proportions of dose applied. The grand average
across plots 1 through 4 for both seasons shows that alley region 2 and the
canopy region received the majority of the residues reaching the ground; alley
regions 1 and 3 received less. Pesticide distribution was more uniform in
the 1977 season, and residue levels were proportionally lower across all re-
gions. The authors believe this was due to modification of the sprayer be-
tween seasons, after which more of the spray was directed into the atmosphere
and away from the ground.
Initial Vertical Distribution
Azinphosmethyl Initial vertical distribution for the 1976 and 1977
seasons is given in Table 1-5. Each plot is represented by only two regions;
alley and canopy. In 1976 the alley was randomly sampled. In 1977 the average
7

-------
TABLE 1-1. SUMMARY OF SIGNIFICANT EVENTS FOR THE ORCHARD, 1976 SEASON
Rainfall	Days
Spray	Sample	Rainfall	Amount	Since
Date	Date	Date	mm	in	Spray
2 July
15 July
28 July
17 August
Orchard Mowed 28 June
30 June
36.3
1.40
2 July



0
6 July



4
9 July



7
14 July



12
15 July



0
20 July
a


5

20 July
22.9
.90

22 July



7

26 July
06.4
0.25

27 July
h


12

28 July
11.4
0.45
0
29 July



1

31 July
04.6
.18

2 August



5
4 August



7

5 August
10.2
.40

10 August



13

14 August
27.9
1.10

16 August



19
17 August



0
20 August



3
24 August



7

28 August
06.4
0.25

30 August



13
aRainfall started immediately after spraying.
^Rainfall occurred after sampling.
8

-------
TABLE 1-2. SUMMARY OF SIGNIFICANT EVENTS FOR THE ORCHARD, 1977 SEASON
Rainfall	Days
Spray	Sample	Rainfall	Amount	Since
Date	Date	Date	mm	in	Spray
Orchard Mowed 23	May
25 May	5.1 0.20
26 May 26 May	0
39 May	20.3 0.80 4
1 June	6
4 June	25.4 1.00 10
6	June	11
11	June and
12	June	12.7 0.50 17
13 June	18
Orchard Mowed 15	June
16 June 16 June	0
17	June and
18	June	6.4 0.25 2
20 June	4
27 June	5.1 0.20 11
28 June	12
30 June	25.4 1.00 14
1 July	15
4 July	12.7 0.50 20
7	July8	5.1 0.20 22
7 July 7 July	0
18 July	22.9 0.90 11
19 July	12
Orchard Mowed 20	July
22 July 22 July	0
24 July	2.5 0.10 2
25 July	3
29 July	2.5 -.10 6
1 August	10
3	August	2.8 0.11 11
4	August	13.7 0.54 13
6 August	3.8 0.15 15
7	August	16
8	August	2.5 0.10 17
10 August	22.9 .90
12 August 12 August	0
13	August	12.7 0.50 1
15 August	3
18 August	6
Rainfall occurred just prior to spraying.
9

-------
TABLE 1-3. AZINPHOSMETHYL INITIAL HORIZONTAL DISTRIBUTION FROM THE ANALYSIS OF TARGETS, 1976 SEASON
Plot
Alley(l)
Alley(2)
Alley(3)
Canopy(4)
2	b
Average Amount (yg/cm + SE)
Sample size
£
Proportion
2
Average Amount (Vg/cm + SE)
Sample Size
Proportion3
2
Average Amount (yg/cm + SE)
Sample Size
Proportion3
2
Average Amount (yg/cm + SE)
Sample Size
Proportion
2.48 + 0.22 3.35 + 0.45 2.76 + 0.34 2.89 + 0.51
10
.071
9
.050
5
.009
9
.042
14
.131
7
.026
6
.079
9
,026
5
.076
5
.023
4
.111
3
.065
13
.093
1.36 + 0.31 2.46 + 0.44 1.65 + 0.39 3.20 + 0.37
11
.136
,75 + 0.20 3.81 + 1.00 1.36 + 0.50 5.10 + 0.77
7
.226
2.51 + 0.95 3.06 + 1.28 2.14 + 0.65 2.90 + 1.10
10
.175
Grand Average Amount (yg/cm + SE)
1.78 + 0.44 3.7 + 0.28
1.98 + 0.31 3.5Z + 0.53
Proportion of amount applied
^Standard error

-------
TABLE 1-4. AZINPHOSMETHYL INITIAL HORIZONTAL DISTRIBUTION FROM THE ANALYSIS OF TARGETS, 1977 SEASON
Plot

Alley (1)
Alley (2)
Alley (3)
Canopy (4)
1 Average Amount
(pg/cm2 + SE)b
2.75 + 0.63
3.12 + 0.77
2.63 + 0.50
3.55 + 1.14
Sample size

7
7
13
6
0
Proportion

.070
.114
.023
.095
2 Average Amount
(ug/cm2 + SE)
1.34 + 0.24
1.96 + 0.12
1.24 + 0.26
1.83 + 0.44
Sample Size

9
3
6
4
3
Proportion

.047
.109
.026
.103
3 Average Amount
(pg/cm2 + SE)
1.03 + 0.22
2.51 + 0.13
1.34 + 0.28
1.39 + 0.45
Sample Size

7
5
8
6
Q
Proportion

.105
.064
.020
.088
4 Average Amount
(jjg/cm2 + SE)
.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
2
Grand Average Amount (|ig/cm + SE)
1.51 + 0.42
2.47 + 0.25
1.55 + 0.37
2.05 + 0.51
Proportion of amount applied
^Standard error

-------
TABLE 1-5. INITIAL VERTICAL DISTRIBUTION OF AZINPHOSMETHYL DISLODGEABLE
RESIDUES (ug/cm + SE)
Year	Tree Grass-Broadleaves Litter-Moss Soil
1976	Canopy
Average Amount 2.92 + .41	1.49 +	.40 .18 + .04 .20 + .05
Sample Size	11 9	9 8
Alley
Average Amount	1.72 +	.33 .15 + .08 .19 + .02
Sample Size	8	5 8
1977	Canopy
Average Amount 2.88 + .42	.55 +	.04 .29 + .04 .29
Sample Size	12 6	8 1
Alley
Average Amount	1.15 +	.23 .27 + .05
Sample Size	8	8
a	2
Soil residues were determined on a whole basis, value represents pg/cm in the
top 10 cm.
12

-------
amounts for the alley represent weighted averages of alley regions 1, 2, and
3, weighted by percent of plot area. These data indicated that for both
years the majority of the dislodgeable pesticide residues deposited in the
orchard was initially distributed vertically to the tree and the grass-
broadleaf layer. Azinphosmethyl residues distributed to the litter-moss
layer and soil were roughly ten times lower than leaf residues.
Using the vertical distribution of the dislodgeable residue data along
with the pesticide application rates and the plot characteristics, the pro-
portion of the applied pesticide reaching each of the vertical regions was
calculated on a plot basis and then averaged across plots (see Table 1-6).
To determine the amount of pesticide distributed to the tree leaves, the
tree leaf surface area was estimated by removing all the leaves from two
trees, one in plot 2 and one in plot 4, following the 1978 sampling season
(an exhaustive search for other methods of determining tree leaf surface
was unsuccessful). The leaves from each tree were contained in six 50 1 bags.
Each bag was weighted and sub samples of ten leaves were randomly selected
for each 500 g of leaves. Weights and areas were determined for each sub
sample. Total leaf area was determined from these measurements and the to£al
leaf weight. The average for these two trees was approximately 400,000 cm
(one side of leaf only). This is a crude estimate due to sampling limitations
and is specific for the tree size and vigor found in this orchard. As the
pesticide was not evenly distributed vertically in the tree, estimates based
on leaf residue data taken from the lower half of the tree in the 1976 and
1977 seasons would tend to overestimate the proportion distributed to the
entire tree. In 1978, the upper half of the tree was sampled separately.
From these data, a ratio was determined by dividing the average amount of
dislodgeable residue found in the entire tree (upper and lower) by the average
amount of dislodgeable residue found in the lower half of the tree. This
ratio was determined using an across-plot average for each of the sample
dates for the first spray period of the 1978 season. A least squares linear
regression of these ratios with time gave an intercept of .7027 and slope of
-.0063T, T being the time since spray in days (r = .836). The value for the
intercept was then multiplied by the average lower leaf values for the 1976
and 1977 seasons to arrive at an estimate of the average leaf residue level
for the entire tree (data in Table 1-6 and Figures 1-3 and 1-4 have been ad-
justed accordingly). Due to pre-1977 sprayer design the 1976 data are prob-
ably still^slightly overestimated. This estimate was then multiplied by
400,000 cm /tree and by the number of trees per plot. The result was divided
by the leaves. The proportion of pesticide distributed to the tree included
residues deposited on both leaves and bark. The proportion distributed to
the bark was estimated at 15% of that going to leaves, based on the work of
Steiner (1969).
The proportion of the pesticide applied that is distributed as surface-
penetrated residues was estimated from matched samples from the first spray
of the 1978 season. The surface-penetrated to dislodgeable residue ratios
and standard errors were as follows; leaves, .081 + .015; grass, .190 + .002;
litter, .421 + .018; moss, .272 (one sample). These ratios were multiplied
by the appropriate dislodgeable residue proportions to estimate the surface-
penetrated residue proportion. All soil residues were determined on a whole
sample basis. To estimate the amount of pesticide deposited in the orchard
during spray application, all the dislodgeable, surface-penetrated and soil
residue proportions were summed across both the alley and canopy regions.
In 1976, 66.8% (SE .093) of the pesticide applied was estimated to be initially
deposited to the various orchard layers. Confidence in this conservative
13

-------
Table 1-6. Estimation of Initial Vertical Pesticide Distribution as a Proportion of the Amount Applied (+SE),
from the Analysis of Samples
1976
Alley
Dislodgeable
Surface-
Penetrated
Canopy
Dislodgeable
Surface-
Penetrated
Total
Tree
Grass
Litter-Moss
SoilC
.112 + .015	.021
.014 + .005	.006
.015 + .002
376+082	.027
.060 +.014	.011
.013 + .004	.033
.010 + .006
,661 + .093
1977
Tree
Grass
Litter-Moss
Soilc
.070 + .016
.018 + .002
.013
.007
.337 + .020
.033 + .009
.014 + .004
.022
.005	.556 + .011
.005
.015
.015
^ata averaged across plots for the first application of the season.
^The proportion distributed to the tree includes both leaves and bark.
p
Soils residues were determined on a whole sample basis.
^1976 value used as there are insufficent data for 1977.

-------
	1	I	
June	July	August
Figure 1-5. Daily Temperature Range, 1976 Season
_Ld

-------
estimate Is subject to the following sources of error; accuracy of amount
applied, homogeneity of application, representative sampling of residue
distribution, and the tree surface area estimate. The 33.2% not accounted
for is assumed to be due to airborne loss, primarily as drift at application,
but also as volatilization and wind erosion in the four hours between appli-
cation and sampling. Another possible pathway of residues not accounted
for is foliar penetration and subsequent unavailability to surface strip ex-
traction. Numerous studies have compared dislodgeable to whole-leaf tissue
residues (Winterlln et al., 1975; Elliot et al., 1977; Gunther et al., 1977).
Their applicability to the present study is uncertain, as the degree of
pesticide leaf uptake may vary with pesticide, formulation, mode of applica-
tion, amount applied, leaf type, and environmental conditions (Hull, 1970).
The efficiency of the solvent strip procedure used here, as compared to
whole-leaf extraction procedures, in removing freshly deposited azinphos-
methyl residues is not known. Use of the solvent strip procedure was based
on the results of a detailed study conducte^by Wieneke and Steffens (1974).
These researchers found that 100.4% of the C azinphosmethyl applied to
bean leaves in an aqueous formulation could be recovered with a water follow-
ed by a benzene strip, for samples taken one day following application. In
1977 only 55.4% (SE .001) was estimated to be deposited in the orchard with
45.6% as airborne loss. A similar pesticide mass balance shown in Table 1-7
uses the target data to estimate the proportion of the pesticide applied
that reaches the orchard floor. Tree proportions were estimated in the same
manner as for Table 1-6. This treatment of the data shows that 74.1% (SE
.099) of pesticide applied in 1976 was deposited in the orchard, on the ave-
rage, with 25.9% as airborne loss. In 1977 61.0% (SE .034) was estimated
to be deposited in the orchard with 39.0% as airborne loss. Analysis of the
plot totals (Table 1-7) using targets to estimate the proportion of the ap-
plied dose distributed to the orchard floor suggests that the amount remain-
ing in the orchard following application is greater in 1976 than in 1977
(paired t-test of plot total proportions gave a P<.15). A similar analysis
indicate that this difference was not due to amounts distributed to the trees
but rather that proportion distributed to the orchard floor. The authors be-
lieve that this difference may be due to the between-season sprayer modifi-
cation that directed more of the spray upward into the atmosphere and away
from the ground. Wind speed and atmospheric stability during application
may also have contributed to this difference. There is also some evidence
that target residues as a proportion of dose applied were greater than the
combined residues distributed to the various orchard floor layers (paired t-
test of plot averages across both years gave a P<.15). As these data were
normalized for extraction efficiency, this result would indicate a substantial
loss in residues during the approximately four hours between spray application
and sampling. The most likely sources of this loss are thought to be volati-
lization and/or wind erosion (Gunther and Blinn, 1955; Taylor et al., 1977).
Losses In Runoff
Azinphosmethyl levels found in runoff collected during the 1977 season
are given in Table 1-8. No runoff was found in the orchard plots as a result
of the rainfall events that occurred on June 27, July 5, 7, 24, 28, and
August 2, 4 (Table 1-2). These rainfall events averaged 5.8 + 3.5 mm (inten-
sity, 2.6 + 1.2 mm/hr) which is considerably lower than the average runoff-
producing rainfall event of 21.2+4.7 mm (intensity, 6.5 + 1.8 mm/hr) shown
in Table 1-8. Rainfall events that occurred on June 11-12 and 17-18 produce
16

-------
Table 1-7. Initial Vertical Pesticide Distribution, as a Proportion of the Amount Applied, using Target
Data to Calcualte Proportion Reaching the Orchard Floor
	13.26	
Orchard Floor	Tree
Surface-
Plot	Alley	Canopy	Dislodgeable	Penetrated	Total
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

-------
Table 1-8. Azinphosmethyl in Runoff, 1977 Season
Date
Days Since Rainfall
Spray	mm mm/Hr.
Runoff
Plot
Control
30 May
4
20.3
5.1
-2
mm x 10
0.80
5.00
0.60
1.00
a




Liters
6.0
20.0
3.5
4.0





PPbC
10.50
0.70
21.70
18.80

A,5 June
10
25.4
9.8
-2
mm x 10
2.00
2.50
6.70
10.00
1.00




Liters
14.0
9.8
40.0
40.0
20.4




PPb
0.74
0.08
0.26
0.25
0.03
30 June
14
25.4
5.4
-2
mm x 10
1.00
5.60
1.70
0.10
a




Liters
8.0
22.5
10.0
0.5





PPb
0.40
0.03
0.24
0.73

18 July
11
22.9
7.1
-2
mm x 10
1.30
5.50
2.20
1.00
0.70




Liters
10.7
22.0
13.0
4.0
10.0




PPb
3.90
0.15
4.10
2.25
0.20
3,4,6 August
13
20.3
5.1
-2
mm x 10
0.06
2.60
0.50
3.10
a




Liters
0.5
10.5
3.0
12.5





PPb
b
1.30
3.40
b

13 August
1
12.7
6.4
-2
mm x 10
0.50
0.80
a
a
a




Liters
4.0
3.0







PPb
14.35
b



^o runoff collected.
^Runoff not analyzed.
CPPb of Azinphosmethyl in runoff water.

-------
small amounts of runoff in some of the plots, but samples were not analyzed.
Rainfall intensity for these two events was similar to those that produced
no runoff but the duration of rainfall was slightly longer. The 25.4 mm
(intensity, 8.4 mm/hr) rainfall that occurred on August 8-10 produced runoff
in all but plot 4, but no analysis was undertaken as samples were not col-
lected until after the spray application on August 12. No runoff-producing
events occurred during the 1976 season. The concentration of azinphosmethyl
in the sediment (sediment amounts were less than 0.5 grams per liter of run-
off) was very small compared to levels in the water, and these data are not
reported here. The lack of sediment is to be expected as the dense ground
cover on the orchard floor allows little soil erosion (Asmussen et al., 1977;
Harrold et al., 1970). The amount of runoff is also expected to be less than
might be found from tilled fields of similar slope characteristics. This
may be attributed to effect of ground cover, which increases infiltration
and slows overland flow, thus decreasing total runoff and loss of pesticide.
Many researchers have reported that concentrations of pesticides in runoff
are highest for rainfall events occurring soon after application (Hall et al.,
1972; Bauer et al., 1972; Glass and Edwards, 1974; Ritter et al., 1974; Caro
et al., 1974). This is indeed what was observed for the pesticide levels in
runoff for the orchard studied. The levels of azinphosmethyl in runoff col-
lected on June 6, 10 days after application, were 10 to 100 times less than
the levels found in runoff collected on June 1* just 4 days after application.
Time since spray has an effect on the amount of pesticide in runoff only as
it is related to the amounts of pesticide residue available to runoff. The
amount of rainfall preceding the rainfall event that produces runoff also
influences the amount of pesticide that is lost to runoff. A comparison of
azinphosmethyl concentrations in runoff collected from plots 1, 3, and 4 on
June 6 (10 days following application) with azinphosmethyl concentrations
in runoff from the same plots collected July 19 (11 days following applica-
tion) shows much higher levels in the July 19 runoff. This may be explained
by the fact that although both events occurred approximately the same number
of days following application, no rainfall occurred between application and
the event producing runoff on July 19, whereas 20.3 mm of rainfall occurred
just 4 days prior to the June 6 event. The intervening rainfall event not
only produced runoff, removing pesticide residues from the orchard, but
also redistributed those residues remaining such that they would be less
likely to be removed with successive runoff. The fac£ that tree leaf resi-
dues for June 6 averaged .399, 1.267, and 1.225 ng/cm for plots 1, 3, and 4,
respectively, and leaf residues found on July 19 averaged 1.942, 2.286, and
2.374 Mg/cm for plots 1, 3, and 4, respectively, supports this hypothesis.
In general, increased runoff caused Increased pesticide loss so that the
concentration of pesticide in varying amounts of runoff on a given day re-
mained relatively constant. The data, for the most part, reflected this re-
lationship, except for plot 2. Plot 2 was characterized by consistently
high runoff compared to the other plots, but upon analysis, azinphosmethyl
levels detected were considerably lower than those in the other plots. We
have not yet arrived at a satisfactory explanation for this. One might
hypothesize that collected runoff was coming from outside the plot or possibly
was interflow from upslope that surfaced just prior to the collection point.
Overall it appears that the contribution of runoff to the azinphosmethyl
loss from the orchard was quite small. For example, the 20.3 mm of rain on
May 30 produced only .01 mm of runoff from plot 4, which contained 75 mg of
azinphosmethyl, whereas 44 g were applied just four days prior to the event.
This loss was well under 1% of the residues present at the time of the rainfall.
19

-------
Residue Data
Figures 1-3 and 1-4 represent azlnphosmethyl dislodgeable residues for
the various above ground layers and soil residues, for alley and canopy re-
gions. Leaf residues are shown as entire tree estimates using the method
described for Table 1-6. Prior to averaging across plots, statistical outliers
were determined and eliminated using the test proposed by Grubbs (1969). k
total of 954 samples for 1976 and 637 samples for 1977 are represented, with
the average number of samples per sampling date as follows: leaves, 11,
grass-broadleaves, 14; litter-moss, 14; and soil, 4. Each bar is divided
into two parts. The mean is above the line and its corresponding standard
error below the line. Time is represented by the Julian date with upward
directed arrows indicating spray application dates. Downward directed arrows
above the bars indicate dates and amounts of rainfall in millimeters. (These
data are also given in Tables 1-1 and 1-2.) The concentration scale is de-
signed so that the highest bar for a given layer, over both canopy and alley
regions, is nearly full scale.
The concentration of dislodgeable pesticide residue found in a given
layer at any point in time throughout the season is a function of two pro-
cesses: movement and attenuation. Attenuation is thought to include all de-
gradative processes (chemical, photochemical, and microbial), airborne loss,
penetration into plant subsurfaces, and irreversible soil binding (Gbllng,
1963; Hull, 1970; Katan et al., 1976). Movement mainly redistributes the
pesticide within the orchard. This redistribution, primarily by rainfall,
is confounded with the attenuation processes, making determination of atten-
uation rates in the field difficult. A treatment of this problem using math-
ematical modeling techniques is described in Chapter II. An "eyeball" ex-
amination of these data indicates contrasting rates of both movement and
attenuation among the various vertical strata and horizontal regions of the
orchard. Comparison of leaf pesticide residues for the first spray period
in 1976 (Figure I-3a) with leaf pesticide residues for the first spray period
in 1977 (Figure I-4a) shows that the residues remaining after eleven days in
1977 are roughly two thirds the value of those pesticide residues remaining
after 12 days in 1976. This difference is possibly due to the 20 mm and
25 mm rainfall events occurring in the interval between pesticide application
and the sampling eleven days later in 1977, while no rain fell during the
first spray period of the 1976 season. That rainfall is responsible for the
reduction of foliar residue deposits has been suggested by a number of re-
searchers, many of the earlier studies are summarized by Ebling (1963). In
more recent studies, McMechan et al. (1972) reported that azlnphosmethyl
applied as a wettable powder formulation to apples was lost at a much more
rapid rate from foliage during wet weather as opposed to dry. Similar re-
sults were reported by Williams (1961) for azlnphosmethyl and carbaryl applied
to apples, and by Thompson and Brooks (1976) for dislodgeable residues of
azlnphosmethyl and four other organophosphate insecticides applied as
emulslfiable concentrates to oranges in Florida. Gunther et al. (1977) also
noted the influence of rainfall on the decline of parathion dislodgeable
residues applied as a wettable powder formulation to oranges in California.
Nigg et al. (1977) used multiple linear regression to examine the relation-
ship between residue decline of ethlon dislodgeable residues (applied as E.C.)
and the variables: degree-days, cumulative leaf wetness, and ordinary time.
They found, for the experiment where rainfall occurred^ that residue decay
was most highly correlated with cumulative rainfall (r = .963). In the
present study, the leaf residue data also show that rainfall had a signifi-
20

-------
cant effect on pesticide residues in the tree. This is indicated by the
results of a 23 mm rain occurring six days after the second spray applica-
tion, which reduced the residue levels to less than half their former value.
Similar results were reported by McMechan et al. (1972) for rainfall events
occurring much closer to application. In their study, a 17.5 mm rainfall
(1.8 mm/hr) that started six hours after a 50% WP azinphosmethyl application
(.23 Kg a.i./Ha) to semi-dwarf apple trees, removed 41% of the initial
deposit. A much lighter rainfall of 3.0 mm (0.9 mm/hr) that started five
hours after a similar application, removed 12% of the initial deposit. Van
Dyk (1976) examined the effect of artificial rainfall on parathlon residues
applied as both wettable powder and emulsifiable concentrate to orange,
lemon, and grapefruit leaves and fruit. Rainfall was applied at 33 mm/hr.
Factorial analysis of variance showed significant interaction (P<0.01) be-
tween parathlon residues and the simulated rainfall. The type of formula-
tion was not significant. This is not surprising considering the high rate
of rainfall applied. By contrast, a 28 mm rain occurring 17 days following
the third spray of the 1976 season (present study) had little effect on the
apparent residue decline, suggesting that rainfall events occurring soon
after application have a greater influence on leaf pesticide dlslodgeable
residue levels. This phenomenon has also been suggested by Ebllng (1963),
McMechan et al. (1972) for azinphosmethyl applied to apple foliage, and by
Steffens and Welneke (1975) for C azinphosmethyl applied to bean foliage.
One theory that might explain this occurrence was first proposed by Gunther
and Blinn (1955), also by Ebling (1963), and most recently by Elliott et al.
(1977). These researchers suggest that pesticide deposits on foliar surfaces
are lost at different rates due to the degree at which they adhere or penetrate
the leaf surface. Initial rapid loss is a result of erosion of loosely bound
deposits-, possibly adsorbed to the formulations or dust on the plant. More
tightly bound residues are lost primarily through volatilization, decomposi-
tion, and penetration into subsurface tissues. According to this theory,
the percentage of the deposit that is loosely bound decreases rapidly following
application. It is this fraction of the deposit that is most susceptible
to erosion processes, including rainfall. As the loosely bound residues
are lost, those more tightly bound residues that remain are less susceptible
to loss with rainfall.
Comparison of 1976 canopy leaf residue data (Figure I-3a) with canopy
grass-broadleaves residue data (Figure I-3b) shows little difference in
residue decline, under the no-rainfall conditions of the first spray period.
The effect 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 I-3a). This might indicate that in addition to
pesticide moving downward out of the grass-broadleaves with rainfall, pesti-
cide is moving in from the tree above. Comparison of canopy litter-moss
pesticide residues during the first spray period of 1976 (Figure I-3c) with
those residues found in the grass-broadleaves (Figure I-3b) shows a marked
difference in residue decline, as pesticide levels remain at a rather con-
stant level throughout the spray period. This would suggest that under no-
rainfall conditions, pesticide is moving into this layer at the same rate
that pesticide is moving out and/or being attenuated. An alternative hypoth-
esis would be no movement and no or very slow attenuation, which seems highly
unlikely. Residue levels in the canopy litter-moss layer increased following
the 23 mm rainfall event occurring during the second spray period of the 1976
season, again suggesting movement of pesticide into the litter-moss from layers
above, in amounts greater than movement out and/or attenuation losses. In
21

-------
3.6
t. 7
j
% 1.6
UN
I uO.I
| So.o
A A
a 0.9
i.a
x
36
CANOPY LEAVES 1976
~r~
26
TW
21 II 10
rr
t.4
1.6
j
I
HTio.o
|i., j
M ^ .
« 0.1
l.t .
T
36
CANOPY ORRSS-BROAOLEAVES 1976
"TW
II II 10
T
16
t' ' V ' t
X
36
l.t
0.9
£ 0.6
MM
| So.)
f = o.e
S o.)
0.6
CANOPY LlTTER-riOSS 1976
u»l i j
tj II 10	•
r 1 r 11' ¦' t1 '
t.o
1.6
I ...
WM
I 5 0.5
$ ^0.0
S 0.6
1.0
1
M
1 1 . 1
1
IS
1
•1' 1 "
II 10
II

1
6
' ' t '
'
11
111 ' '
r
'
te
Figure I-3(a-d).
Azinphosmethyl Dislodgeable and Soil Resiues for the
Canopy Region, 1976 Season. Upward directed arrows show
dates of spray application, and downward directed arrows
indicate dates and amounts of rainfall in ram. Each bar
is divided into two parts; the mean (ug/cm ) is above the
line and its corresponding standard error (S.E.) below
the line.
22

-------
RAIN (tin) '
2.4
1.6
5 »-2 -I
U N
M 5 0.6
O N
r a
£ ="0.0 .
5 0.6 .
1.2
36
I7S
flLLET GRflSS-BROADLEflVES 1976
~1	
20
T
T
23 II 10
-L.
—A	I	L_	I	I	1_
Tit
JULIAN OATE
I 1 I I
266
RAIN (ftri)
1.2
0.9
£ 0.6
UN
c c M .
<0 u 0*3
O X
X e
£ =0.0
« 0.3
0.6
T
36
ALLEY LITTER-MOSS 1976
l.vi i 1
29 It 10	6
176
t' 1 r 111 1 11
JULIAN OATC
265
RAIN inn)
2.0
l.S .
% 1-0
WN
c c a e
to 
-------
general, canopy litter-moss residue values remained fairly constant through-
out both the 1976 and 1977 seasons. Under no-rainfall conditions (first
spray period, 1976 season, Figure I-3d), canopy soil residues did not in-
crease with time following application as did the litter-moss residues. As
azinphosmethyl has been shown to degrade faster in non-sterile soils as com-
pared to sterile soils (Yaron et al., 1974), microbial degradation may play
an important role in losses from this layer. In addition, root uptake and
translocation (Al-Adil et al., 1973) may have been partially responsible for
the observed soil residue pattern. Pesticide movement into the canopy soil
with rainfall is again pronounced, as indicated by the 23 mm rainfall event
during the second spray period of the 1976 season (Figure I-3d). The first
spray period of the 1977 season shows canopy soil residue levels increasing
with each rainfall event (Figure I-4d).
Alley grass-broadleaves pesticide residues show a systematic decline
over the first spray period of the 1976 season (Figure I-3e). The decline is
slightly steeper than the decline shown for canopy grass residues during the
same no-rainfall period (Figure I-3b). This difference is probably due to
less pesticide movement from the trees into the alley grass-broadleaves. In-
creased attenuation losses in the alley regions due in part to greater ex-
posure to wind and solar radiation must also be considered. Taylor et al.
(1977) in discussing factors influencing dieldrin volatilization from orchard
grass, cited solar radiation as the most significant. Associated surface
temperatures may also influence rates of chemical and microbial degradation
and foliar uptake (Ebllng, 1963; Hull, 1970). Azinphosmethyl applied to plant
surfaces has been shown in one study (Liang and Lichtensteln, 1976) to be
susceptible to photodegradation by sunlight. This degradative pathway is
directly related to solar radiation. Rainfall effects are clearly shown when
comparing alley grass residue to decline for the first spray period in 1976
(Figure I-3a) with the first spray period of the 1977 season (Figure I-4e).
The difference in rainfall effects between canopy grass-broadleaves and alley
grass-broadleaves is shown in the second spray period of the 1976 season
(Figures I-3b and I-3e) which indicates that the pesticide movement, follow-
ing the 23 mm rainfall event, out of the alley grass-broadleaves is probably
greater due to direct exposure to rainfall, and also that there is little
or no pesticide movement in from the trees. Again, differential attenuation
due to shading cannot be disregarded. Alley litter-moss pesticide residues
show a definite decline during the first and last spray periods of the 1977
season (Figure I-Ac), the alley litter-moss residue values show a decline,
indicating that residue movement out and/or attenuation is greater than re-
sidue movement into this alley layer. Alley soil residues for the first spray
period of the 1976 season (Figure I-3g) are similar to canopy soil residues
for the same period (Figure I-3d, note the difference in scaling). There
appears to be a decrease in alley soil residue levels following the 23 mm
rain during the second spray period of the 1976 season (Figure I-3c) whereas
the canopy soil residues (Figure I-3d) increase following this rainfall event.
The high variability of these data discourages speculation as to the processes
taking place. The general pattern of these residue levels over the season,
characterized by a buildup of residues toward mid-season, is surprisingly
similar to the azinphosmethyl orchard soil residue pattern reported by Kuhr
et al. (1974). A general observation of the data presented in Figure 1-3
indicates that, as the season progressed, the interrelationship between
pesticide movement and attenuation became more difficult to follow. This
complexity is possibly a function of changes in weather patterns in addition
24

-------
to rainfall (temperature, humidity and wind). Gunther et al. (1977) attempt-
ed to relate temperature data to the decline of azinphosmethyl residues
applied to citrus in southern California. They found it difficult to make
any meaningful interpretation of the data, but stated that azinphosmethyl
dissipation was slightly more rapid during warmer weather. No correlations
between residue data and temperature or humidity were attempted in the pre-
sent study, but weather data were provided for possible future use (Figures I-
4 and 1-5). Changes in plant physiology and soil microbial activity or a
buildup of tightly bound or penetrated residues, may also have an effect on
pesticide residue dynamics throughout the season.
However, from the limited analysis of the data, there is some indication
that azinphosmethyl is redistributed throughout the orchard with time follow-
ing application. Redistribution is indicated during periods without rainfall,
but is more pronounced following rainfall events. Little more information
can be gained without the aid of a more sophisticated technique of analysis.
Such an analysis of the data is presented in Chapter II. Using mathematical
modeling techniques, the change in observed residue levels in each layer
and region is estimated as a function of both movement and attenuation.
Information on pesticide concentrations as a function of movement and
attenuation is essential to the estimation of possible exposure to the biota
of this agroecosystem. If pesticides are to be used In an effective manner
in conjunction with biological control techniques, possible exposure to
beneficial populations, as well as the target species, must be better under-
stood under a variety of climatic conditions.
25

-------
CANOPY LEAVES 1977
5.6
1.7
£ 1.8
| 5 o.»
| § 0-0
5 0.9
i .0
1 | ,i | i |,i \a | 11 i|i| a
6 is 23 sen
J	L
f t
f f
JUU«" OAK
iraiM irmi •
l.S
l.J
CANOPY GRASS-BROAOLEAVES 1977
MAM ia i* in -I-I a
s is t) Jew
fr * o.o
S 0.4
o.e
¦U
I t
JUUIM MfC
f ' ' '
I	t40
CANOPY UTTER-flOSS 1977
o.e
o.t
£ 0.4
n uO.t
Mil* HI* I iHJiJ-'-
SI) ts I I K
1*
0.0
o.t
0.4
V * ' * T' ' T
*¦	CfiNOPY SOIL 1977
min (mil •
9.0
1.1
1 1A | \ 11 |l | i 1 m A
I
i 0.9
*0.0
0.9
t II	(9	3 8 1
-M—i	1-4-1	i-l
f t
T T
JW.IAN OflfC
Figure 1-4(a-d). Azinphosmethyl Dislodgeable and Soil Residues for the
Canopy Region, 1977 Season. Upward directed arrows show
dates of spray application, and downward directed arrows
indicate dates and amounts of rainfall in »jm. Each bar
is divided into two parts; the mean (yg/cm ) is above
the line and its corresponding standard error (S.E.)
below the line.
26

-------
rain inn)
1 .6
1.2
.j
£ o.e
Ui CW
« u0<4
© N.
r o
£ =0.0
« 0.4
0.8 J
RLLEY ORflSS-BROflDLEflVES 1977
T
25
1 | i J i [J]3 J"
20	13	6 13	23
T
TTTTT
S J.8 I 13
B 25
J	U
140
' 't ' ' ' 't '
* ' _ ' ' 1 _ ' 1 1 11 ' 1
t	240
i r
JULIRN ORTE
rain (nm ¦
0.6
ALLEY LITTER-MOSS 1977
T
o.e .
% o.«
UJ N
w 5 0.2 -|
O N.
z o
i30.0 J
S 0.2 .
0.4
Mi 1 i Ji |i | 11
20	13	6 13	23	3
5 18 I 13
3 8 26
-U-
' > ' ' ' 't
M0 1	1
i__I	L-
	I.I I	L_J	1	1	1—_i_
t t	1
JULIAN DATE
240
e.
rain inn)
3.6	,
2.7
2 I-B
W C4
n u 0.9
° i
z o
£ =0.0
£ 0.9
1.8
1 ^ ^
5 25 .6
ALLEY SOIL 1977
J IT
26 5
20
13
-I	L
FFT
5 13	23
TpiT
3 5 8
T
13
8 26
-U	1-
-1—1-
I	L,	I	L.
' _ ¦	I	I	I.I I I
140
t—r
JULIAN 0ATE
240
Figure I-4(e-g). Azinphosmethyl Dislodgeable and Soil Residues for the
Alley Region, 1977 Season. Upward directed arrows
show dates of spray application, and downward directed
arrows indicate dates and amounts of rainfall in mm®
Each bar is divided into two parts; the mean (pg/cm )
is above the line and its corresponding standard error
(S.E.) below the line.
27

-------
100
90
80
60
R.H.
50
40
30
20
June
July
August
Figure 1-6. Daily Minimum and Maximum Relative Humidity, 1976 Season

-------
100
90
80
70
N3
VO
60
50
40
30
June
July
August
id
Figure 1-7. Daily Temperature Range, 1977 Sea
son

-------
100
90
80
70
R.H.
60
50
40
30
August
July
June
Figure l-C. Daily Minimum and Maximum Relative Humidity, 1977 Season

-------
CHAPTER II
PARAMETERIZATION OF THE AZINPHOSMETHYL FATE MODEL
INTRODUCTION
In Chapter I a description of the field data collected on the distribu-
tion, movement, and attenuation of azinphosmethyl in an experimental apple
orchard was given. These data were gathered specifically in order to allow
development, refinement, and parameterization of a model describing the spa-
tial and temporal distribution of azinphosmethyl in the orchard in response
to rainfall. While the time series of concentrations observed were reported
in the earlier part, the model and the parameterization process, together
with the parameter values generated, are described in this part. Validation
of this model using a third season's data is presented.
The form of this model was chosen to allow: (1) use of model output to
provide pesticide exposures for models of organisms dwelling in the orchard
floor (Goodman, 1982) and (2) future development to represent the dynamics
of other pesticides and the effects of additional environmental factors.
THE MODEL
In order to structure the experimental program and the data analysis,
a conceptual model for the distribution, movement, and fate of the pesticide
in the orchard was formulated. Information gathered has resulted in continual
refinements of the model structure. The conceptual model utilizes a spatial
subdivision of the orchard into a canopy region and three alley regions
(Figure 2 of Chapter I).
Vertically, four strata (not necessarily all present at a given sampling
location) are identified. These are called tree, grass-broadleaves, litter-
moss, and soil.
The conceptual model (see Figure II-l) describes the dynamics of the
pesticide from its spray application to its ultimate disappearance from the
orchard via drift, attenuation (including airborne loss, photolysis, chemical
degradation, microbial degradation, and penetration of surface residues) and
runoff. Each day, a vector C of concentrations of pesticide in each of
seven regions (two horizontal by four vertical, minus one for non-existent
alley trees) is calculated, based on management actions (spraying, mowing)
and rainfall. Figure II-l describes the processes affecting the pesticide
concentration in only one of the seven regions.
For analysis of the field data, the processes in Figure II-l are lumped
into three categories: inital distribution, movement, and attenuation. Move-
ment is further subdivided according to rainfall intensity (none, light, and
heavy). The mathematical forms of the various components of this model are
described below:
(1) Initial Distribution
The spray rate (in Kg/ha) is supplied as an input. Drift, includ-
ing losses from the orchard during spraying and up to the time of
LIBRARY
M S Environmental	t
200 <5 *

-------
r-^Pasticide^
LXj
Management
Decisions
L*>
Spray Rate
Wind
Characteristics
Airborne
Loss
Spray
Characteristics
Degradation
Products
Drill Loss
<
Agricultural
Alterations
Microbial
Degradation
Non-rain Vertical
Movement
Photolyale
nvironment
and Compound
Characteristics
Rainfall Vertical
Movement
Chemical
Degradation
Orchard Physical
Parameters
W
Lateral
Movement
>
Penetration
weather
Conditions
Figure 11-1. Conceptual Model for the Distribution, Movement, and
Attenuation of Azfnphosmethyl in an Apple Orchard

-------
post-spray sampling, was estimated using mass conservation, based
upon the known application rate, estimated pre-spray residue levels,
and measured residues following application (see Chapter 1). The
proportion drift averaged .376 + .149 over the three seasons. Be-
cause the standard deviation observed among the data was higher
than anticipated, an attempt was made to relate the drift loss to
wind speed at the time of application from the available data.
Mean wind speeds during application were estimated from data of
the Michigan Weather Service, Grand Rapids, Michigan. The follow-
ing relationship was determined:
D = .021 WS + .091 (r2 = .475)
where D is proportion drift 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 vector, in which
each entry specifies the proportion of the spray that is captured
by the corresponding region, and added to any remaining residue
from earlier sprays in the pesticide concentration vector C. All
units are expressed as yg pesticide/cm ground area.
Attenuation
Attenuation is treated in the model as a set of dally proportion
losses—a single proportion for each layer. Thus it is conveniently
representable as a diagonal matrix A which pre-multiplies the
pesticide distribution vector C, a seven-element column vector con-
taining the concentration of pesticde (vg/cm ) in each region at
a particular time. Each day, attenuation is extracted via equation
(1):
C(after) = (I-A)C(before)	(1)
where I is the 7x7 identity matrix.
and C_, the pesticide concentrations in canopy and alley
soils, respectively, represent total soil residues, while
C,, and Cg represent only dislodgeable residues. Thus the attenua-
tion for non-soil layers includes surface penetration of residues.
Movement
Daily redistribution of the pesticide within the orchard is
modeled using three matrices to pre-multiply the pesticide distri-
bution column vector C. Daily movement not attributed to rainfall
is modeled by equation (2):
C(after) = PC(before)	(2)
where P is a 7 x 7 column-stochastic lower triangular matrix known
as the non-rainfall pure movement matrix. The matrix P is restrict-
ed to containing at most 19 non-zero entries, as the tree layer is
the only canopy layer from which movement to alley layers is model-
ed, so P . = 0, for i = 5,6,7 and j = 2,3,4.
Several parameters are necessary to adequately describe a
rainfall event; for example, duration, average intensity, peak in-
tensity, etc. Unfortunately the small number of rainfall events
during a spray season precluded using so fine a description. Rain-
fall events were classified on a daily basis into only two categories
(heavy and light) in order to obtain enough instances of each cate-
gory to parameterize the model. Heavy rain was defined as any event
34

-------
of more than 10 mm rainfall or more than 5 mm/hour in a day, with
other measurable rain classified as light. Equations (3) and (4)
show the pesticide redistributions caused by heavy and light rain-
fall, respectively:
C(after) = HC(before)
C(after) = LC(before)
(3)
(4)
where H and L are 7x7 column-stochastic lower triangular matrices.
As in equation (1), H and L do not model movement between alley
and canopy ground layers; thus pesticide movement with overland
runoff is not included in these equations.
Attenuation and non-rainfall movement were modeled (and para-
meterized) as dally phenomena. Thus, to represent the natural
changes in pesticide distribution from one day to the next, (C(k)
to C(K+1)), exactly one of the following relationships is used:
This section describes the techniques used to estimate the model para-
meters from the data base of field determinations. The routines described
are designed to be used repeatedly, i.e., as entries are added to the data
base, new parameters including their effects can be quickly generated. This
capability for dynamic reparameterlzation of the model gives it the flexi-
bility to begin with relatively few crude data, producing preliminary outputs,
and to produce more accurate results as the data base grows.
The data base consists of a large number of sequential disk files, which
are updated and utilized by various programs. Rainfall data, orchard plot
physical characteristics, and records of samples analyzed for pesticide are
stored in these files. Figure 11-2 shows how sample and target data, entered
as "raw" outputs from gas chromatograph measurements, are transformed into
new files of "processed" sample and target data. In this stage, appropriate
corrections for analytical technique, sample weight, averages across samples,
etc. are made, yielding files suitable for use in calculating the model para-
meters. Figure II-3 is a flowchart of the process of calculating the para-
meters using the "processed" files. Files used by a given routine are shown
with a dashed arrow into the routine, and files produced are shown with a
dashed arrow to the files. Flow of the program is via the solid arrows.
Subroutines LEAF and AREA perform needed conversions of some residue ^
measurements (for example, from yg pesticide/cm leaf area to pg pesticide/cm
ground area). TARGETS uses "processed" target values to determine a proportion
of pesticide reaching each plot and region.
The process of parameterizing this model continues with FINDEV, a sub-
routine which searches the data files for sets of samples useful for calculat-
ing redistribution and losses of the pesticide. For estimating loss and
movement in the absence of rain, it prepares a file of sample sets each of
which consists of a vector of concentrations (in each region and layer) at
the beginning of a non-rainfall interval and a similar vector at the end of
the interval. Of course, either rain or a spray intervening will exclude a
C(k+1) - P(I-A)C(k)	(no rain)
C(k+1) » L P(I-A)C(k) (light rain)
C(k+1) = H P(I-A)C(k) (heavy rain)
(5)
(6)
(7)
METHODS FOR PARAMETER ESTIMATION
Data Base
35

-------
INPUT FILES
OUTPUT FILES
OMFim 4
(DATA A
TARGETS J
c
NONRAIN
TARGETS
(CORRECTED V.
SAMPLES J
r
DAYAVOS

CNONRAIN A
TARGETS H
c
DATA
SAMPLES
/ INPUT

| N«w

1 OataCard


^ ii
ADD
1
1
I k
1 1'
rl
1
I
DUP

H1LO AVG
DATA
SAMPLES
CORRECT
DAY AVOS
T
OUTLIER

r~^r\
\ SAMPLES J *
f OATA A
I TARGETS J
CNONRAIN A '
TARGETS J
* Updated fllM
1
AVERAGE

v
AVERAGE

) *

1
1
T
)
AVERAOE

) '

/processed^
I SAMPLES J
fPROCESSED I
Va"qIts J
(processed\
NONRAIN 1
TARGETS J
Figure 11-2. Flow Chart Showing Data Processing to Yield Files
Suitable for Use in Model Parameterization
36

-------
(MEASURED \
LEAF AREAS J
LEAF
c
(
PLOT AREAS
SPRAY RATE
>-|
>~lt
/TREl\
	W INTERCEPTION 1
V INDEX J
	~	
AREA
tH-
i
I _y PSEUDO
' A. areas J
TARGETS
I—1
!	J TARGET \	
^	^^DISTRIBUTION J
r	yNON-RAINFALL\
*	INTERVALS J
D --	y
(PROCESSED ^	
TARGETS J
(PROCESSED \
SAMPLES J \
~	v X. I T	U PINO	Y' y-
( RAIN DATES		» EVENT	RAINFALL A
V	/ | | |	^ INTERVALS J
f SELECTED \
	« NON-RAINFALL)
A INTERVALS J
SELECTED ^
*( RAINFALL )
V INTERVALS J
'-if NON-RAIN \_.
MOVEMENT J
^attenuation)
MANUAL
SELECTION
ADAPT 1
ADJUST
ADAPT 2
r*c*
k-rr ^
i v
IMMEDIATE
PRE/POST RAIN
SAMPLEVAL
PEN RATE
Lfcrz
RAINFALL
MOVEMENT
)
r INITIAL PEN
attenuation J
A[ drift )
	L J
SPR DIST
I L_
I	
—¦>
TT
V SPRAY ^
\ DISTRIBUTIONJ
SIT REP
T~
< SITUATION
REPORT J
Figure II-3. Flow Chart of the Process of Calculating the
Parameters Using the "Processed Files"
37

-------
pair of sequential sample vectors from being used for this purpose. F1NDEV
also prepares similar files of vectors representing samples before and after
rainfall events of various Intensities to allow estimation of the effects
of these rains on the pesticide distribution. As shown in the diagram,
manual selection is utilized to determine rain events to be classified as
similar, and to exclude sample dates which are unacceptable for the task to
be performed (when numerous data are missing for a given sample date). The
matrices A and P (equations (5), (6) and (7)) are determined from a matrix
Q = P(I-A). After Q is found (see equation (9) below), the entries A are
calculated as:
7
Ay ° 2! Q^, i=j; = 0,1 / j, 1 = 1|• • •) 7; Js 1,....,7
Then P = Q(I-A)"1	(8)
From equation (8), Q should relate the pesticide vector C(k) to C(k+1)
for each day k^ on which no rainfall occurred. Thus, given matrices B =
[C(k.) C(k0)...C(k )] and A = [C(k. + 1) C(k. + l)...C(k +1)], k, e days
l z	n	i	i	n	l
without rainfall, the following relationship should hold:
A = QB.	(9)
So long as at least seven sets of before/after concentrations are avail-
able, Q should be uniquely determinable using Gaussian elimination. However,
owing to both random and systematic variations in the data sets, such a proce-
dure did not produce a feasible solution. It is necessary to introduce additional
constraints that a "best fit" must satisfy. Entries in Q must be restricted
to the range [0,1], so that pesticide removed from a region is limited to
what is available for movement. Also, because too few soil samples were
analyzed to allow for a reliable solution for the pesticide attenuation rate
in soils, a rate determined from the incubation of the orchard soil in the
lab was introduced into the solution matrix before the remainder of the 19
were determined. Solution for the entries in Q to optimize the fit to all
available data for non-rainfall intervals was carried out in program ADAPT1,
using the adaptive optimization technique of Holland (1975), also described
in DeJong (1980).
The optimization was done on a broadened version of equation (9) allowing
varying intervals between sampling days, since:
C(k^ + m) ° QmC(k^), m > 1, so long as no rain falls on days K^,...,Km_^.
Once Q was determined, it was used, together with Equation (6) and data
surrounding light rainfall events, to calculate L, the light rain movement
matrix. Because samples were typically collected at intervals of three to
five days, the effects of degradation and non-rainfall movement were removed
from the data surrounding each rainfall before the rain effect could be de-
termined. Using equations (5) and (6), we obtained:
C(k^ + p + q) = Qq L QP C(k^), where C was known at days and k^^^
and light rain occurred on day k^+p. Thus (Q *)qC(kj+p+q) = L Qp C(k^),
and matrices of before-and after-rainfall estimated concentration vectors
were assembled as
AL = [(Q'Vl C(kx + Pl + qi)...(Q-1)qn C(kn + Pn + qn)]
B = [(QP1 C(k.) ... (QPn C(k )]
LI	n
38

-------
and the matrix L in A^ = L was optimized by ADAPT2, a program similar to
ADAPT1. It also solves, in an analogous fashion, for the heavy rainfall
movement matrix H of equation (8).
Once the movement matrices are parameterized by the adaptive routines,
the initial pesticide distribution for spray applications subsequent to the
first application of the season are estimated. The attenuation and movement
matrices are used, along with the residue distribution of the last sampling
day of the preceding spray period, to estimate the dislodgeable residues
present just prior to application. PENRATE (Figure II-3) calculates the
estimate of the amount of surface-penetrated residues present at the time of
application based upon its estimate of the dislodgeable residue and a ratio
of surface-pentrated to dislodgeable residues. This ratio was determined
from matched samples taken the last sampling day for selected spray periods
over the 1976, 1977 and 1978 seasons. The surface-penetrated to dislodgeable
residues and standard errors were as follows: leaves, 0.110 + .003; grass,
1.075 + .195; litter, 2.367 + .263; moss, 1.701 + .184. These ratios were
multiplied by the appropriate dislodgeable residues to estimate the surface
penetrated residues present just prior to application. To estimate the ini-
tial pesticide distribution, the estimates of dislodgeable and surface-pene-
trated residues present at the time of application were subtracted from the
total residues measured the afternoon following application. Assuming these
estimates are conservative (see Chapter I), drift is determined as the dif-
ference between the estimate of the amount deposited in the orchard at appli-
cation and the amount applied. Using the method described above, the average
proportion and standard deviation of the dose initially distributed to the
orchard, for all spray applications over the three seasons (except two appli-
cations for which there was rainfall following the last sampling day of the
preceding spray period) was .624 + .149, giving an average drift estimate
as reported earlier.
The final routine, situation report, produces a daily record of both
measured and predicted residue values. The data are provided both tabular
and graphical form, as shown in Figures II-8a-g.
RESULTS AND DISCUSSION
Parameterization of the Attenuation and Movement Matrices
The model was parameterized with azinphosmethyl residue data, collected
over two seasons (1976, 1977) as previously described in Chapter I, plus a
third season (1978). The matrices generated (for equations (1) to (4)) are
shown in Figures II-4, II-5, II-6, and II-7. The matrices describe the daily
attenuation and movement of pesticide under three specified rainfall conditions;
none, light, and heavy. The no-rainfall movement matrix (P) shows some pesti-
cide movement from the tree, possibly due to wind erosion, dew or guttation,
with the majority of the residue being trapped in the litter-moss layer.
Also of Interest is the fact that the model estimates that equal amounts of
residues (approximately 0.9%) move from the tree to the canopy litter-moss
and alley litter-moss layers. Gunther et al. (1977) estimated the movement
of parathion from orange trees to the ground over a five-day period to the
less than 1% of the applied dose. In order to directly detect this movement
in the orchard used in this study, filter paper targets were placed on the
orchard floor (as described in Chapter I) for periods without rainfall of up
to four days. Recovery of approximately 0.5% of initial tree residues con-
firmed the existence of some pesticide movement. The unkown rate of loss
39

-------
CANOPY CANOPY	CANOPY CANOPY ALLEY	ALLEY	ALLEY
LEAVES GRASS- LITTER-MOSS SOIL GRASS- UTTER-MOSS SOIL
ATTENUATION	broaoleaves	broadleaves
.049







.041







.167







.079







.067







.223







.079
Figure 11-4. Azinphosmethyl Attenuation Matrix
40

-------
FROM
NON-RAINFALL
MOVEMENT MATRIX
CANOPY CANOPY CANOPY CANOPY ALLEY	ALLEY	ALLEY
LEAVES GRASS- UTTER-MOSS SOIL	GRASS- LITTER-MOSS SOIL
BROAOLEAVES	BROADLEAVES
TO
CANOPY
LEAVES
CANOPY
GRASS-
BROAOLEAVES
CANOPY
LITTER-MOSS
CANOPY
SOIL
ALLEY
GRASS-
BROAOLEAVES
ALLEY
LITTER-MOSS
ALLEY
SOIL
.983






0.0
.935





.008
.065
.907




0.0
0.0
.093
1.0



0.0



.973


.009



.027
.996

0.0



0.0
.004
1.0
Figure 11-5. Azinphosmethyl Non-Rainfall Movement Matrix
41

-------
PROM
CANOPY CANOPY CANOPY CANOPY ALLEY	ALLEY	ALLEY
LEAVES	GRASS* UTTER-MOSS SOIL	ORASS- LITTER-MOSS SOIL
LIGHT RAINFALL	broaoleaves	broaoleaves
CANOPY
leaves
.932






CANOPY
GRASS-
BROAOLEAVES
.030
.747





CANOPY
LITTER-MOSS
.002
.071
1.0




CANOPY
SOIL
.036
.182
0.0
1.0



ALLEY
GRASS-
6ROAOLEAVES
0.0



.869


ALLEY
UTTER-MOSS
0.0



0.0
1.0

ALLEY
SOIL
0.0



.131
0.0
1.0
Figure I1-6. Azlnphosmethyl Light Rainfall Movement Matrix
42

-------
FROM
HEAVY RAINFALL
CANOPY
LEAVES
CANOPY CANOPY CANOPY CANOPY ALLEY	ALLEY	AUEY
LEAVES GRASS- LITTER-MOSS SOIL	OR ASS- UTTER-IAOSS SOIL
BROAOLEAVES	BROAOLEAVES
CANOPY
GRASS-
BROAOLEAVES
CANOPY
UTTER-MOSS
Trk CANOPY
L*	SOIL
ALLEY
GRASS-
BROAOLEAVES
ALLEY
LITTER-MOSS
ALLEY
SOIL
.822






.020
.613





.096
0.0
1.0




.008
.387
0.0
1.0



.027



.280


0.0



.447
1.0

.02?



.273
0.0
1.0
Figure 11-7. Azinphosmethyl Heavy Rainfall Movement Matrix
43

-------
from the targets precludes a direct assessment of the movement rate based
only on these data.
Figure II-4 shows an attenuation rate for azinphosmethyl dislodgeable re-
sidues in the tree of 4.9% day. Hall et al. (1975) examined the loss of
azinphosmethyl dislodgeable residues applied to apples using two types of
air blast sprayers. Trees were sprayed on one side only and each tree was
divided into nine sites for sampling. The average residues remaining for
sites one to five (which most closely correspond to the tree area sampled
in the present study) over a 14-day pe^log were used to determine the loss
rate. Rates determined wjre ^.6% day (r = .914) for the high flow-rate
application and 5.5% day (r <= .887) for the low air flow-rate application.
These loss rates are higher than the attenuation rate alone as determined
in the present study, but are in excellent agreement with the overall ljss
rate of azinphosmethyl dislodgeable residues from the tree of 6.7% day
(obtained by summing attenuation and movement from the tree under dry con-
ditions). These results suggest that approximately 25% of the daily loss
of azinphosmethyl dislodgeable residues from the tree is redistributed with-
in the orchard, under dry conditions.
Figure 11-5 shows that less movement occurred from the grass-broadleaf
to litter-moss layer in the alley region than in the canopy, suggesting that
conditions under the tree canopy may be more favorable to movement, in the
absence of rainfall. The attenuation rate for the canopy grass-broadleaves
is similar to that observed in the tree. However the attenuation rate for
the alley grass-broadleaf layer is considerably higher. Increased exposure
to solar radiation resulting in higher surface temperatures, which in turn
influence losses by volatilization, degradation, and plant uptake (Ebling,
1963; Hull, 1970; Bukovac, 1976) may be responsible for the observed difference.
Movement out of the canopy litter-moss layer in the absence of rainfall was
similar to that from the grass-broadleaves, with virtually no movement from
the alley litter-moss. Attenuation rates observed in the litter-moss layer
are higher than might be expected, but little is known about pesticide at-
tenuation on these surfaces. Again the rate observed in the alley was greater
than that in the canopy. The reason for this difference is thought to be
the same as for the grass-broadleaf layer. The laboratory-determined rate
for azinphosmethyl attenuation in soil (7.9% day ) is only an approximation
of the field.rate. However this rate is in rough agreement with the rate
of 6.3% day determined from the field data of Kuhr et ali (1974), in which
samples were taken from beneath apple trees in an upstate New York orchard
(ignoring any movement of azinphosmethyl into the soil). A loss rate of 5.8%
day was determined from the field data of Schulz et al. (1970), in which
azinphosmethyl was applied as an emulsiflable concentrate to the soil surface.
The experiment was run in the early spring in Wisconsin (soil temperature
5-15°C). Only a general comparison can be made with these field studies,
as soil type, pH, available moisture, ground cover, and temperature, as well
as amount and mode of application, may influence azinphosmethyl loss from
soil (Hamaker, 1972;Schulz et al., 1970).
Rainfall-induced losses of azinphosmethyl and other deposit residues in
orchards have been indicated by a number of researchers (McMechan et al., 1972;
Williams, 1961; Thompson and Brooks, 1976; Gunther et al., 1977; Nigg et al.,
1977). The influence of rainfall on the removal of azinphosmethyl from apple
foliage is discussed in Chapter I. The present treatment of the data cal-
culates azinphosmethyl loss as a function of heavy or light rainfall (as de-
fined above). No attempt was made to relate susceptibility to rainfall re-
moval with the age of the residue deposit. Figures II-6 and II-7 show that
44

-------
approximately 7% and 12% of the residues are moved out of the tree by light
and heavy rainfall, respectively. The light rain moves all these residues
into the canopy region, while the heavy rainfall moves two-thirds to the
canopy and one-third to the alley region. Values for proportions moved to
the 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 the largest proportion of the canopy litter-moss layer.
Canopy and alley grass-broadleaves and alley soil receive lesser amounts.
Eighteen and 38% of the residues deposited on the canopy grass-broadleaf
layer are moved to the soil following light and heavy rainfalls, respectively.
Thirteen percent Is moved from the alley-grass-broadleaves to the soil as a
result of light rainfall, whereas 27% is moved as a result of heavy rainfall.
The largest proportion of alley grass-broadleaf residue (45%) is moved to the
litter-moss following a heavy rainfall. No movement is indicated from the
litter-moss layer as a result of either light or heavy rainfall.
Certain types of movement are difficult to distinguish based on the
available data. For example, a large amount of movement from the canopy grass-
broadleaves to the soil, accompanied by a similar amount of movement from
the tree to the canopy grass-broadleaves, can produce the same result as a
direct movement from the tree to the soil. Only a widely varying set of
initial pesticide distributions would enable the parameterization routes to
distinguish these two processes definitively. However, in the operation of
the model, the net resulting distribution will be similar in either case, so
long as the initial distribution is similar to the one used to parameterize
the model. More credibility should be attached to the model-generated dis-
tributions than to the individual matrix entries.
Comparison of the Model Outputs with the Field Data
Figures II-8(a-g) show a test of the model's predictions as compared to
the field data obtained during the 1978 season. The continuous lines (solid
or dashed) represent the model's daily prediction of azlnphosmethyl dislodge-
able foliar and soil residues in each region for the 1978 season and are
directly comparable with the vertical lines showing measured residug levels
(and standard errors) in 1978. All residues are expressed as yg/cm ground
area, including the tree residues, to maintain a material balance in the
transfer of pesticide between layers and regions of the orchard. The models
were driven by 1978 rainfall data (shown as descending arrows in the figures)
and the 1978 spray application rates (shown as ascending arrows). The matrices
used to generate the solid line in Figure II-8a-g were calculated using 1976
and 1977 data only, while the dashed line was generated from the matrices
(Figures II-4, II-5, II-6, and II-7) calculated using data from all three
years. The 1978 sample data therefore represents an independent set of
field data, collected from the same orchard plots, by which to judge the
predictive capability of the model parameterized using data from the 1976
and 1977 seasons (continuous solid line).
In the tree leaves and grass-broadleaf layers, to which over 80% of the
residues are Initially deposited, the trends in the data are predicted quite
well. Lower residue levels and the high variability in the amount and com-
position of the litter-moss layer are most likely responsible for the dif-
ficulty in predicting attenuation and movement in and out of this layer. In
addition, the overall predictive capability of the model should improve with
a better estimate of drift loss.
45

-------
CflNOP* LtflVES 1078
is.?
t.l
5 i.<
TTTrPT" W TTfT
7 17 II &l
P
0.0
).)
6.8
.V.
"bt
-r*	i	*i—
CPNOPY ORflSS-BROROlERVES 1978
¦•nip n ¦
t 17 II 81 8
IvJW-j
W H 1
1 1 8
NJ
¦ i
K


CfiNOPT LIT TER-flOSS 1978
' ™ if If | i I PI* I *1 *
1.2 .	t 17 II 81	8	* I t • l«
s
8
w
ISO
4 rsin imi •
CANOPY SOIL 1978
T
i •' I II'! •' I •'! »
* i.O
* o.o
09

fn
/!
MT'


lip"'


i


1' 1
-	r	^	r
•ji rat pair
Figure II-8 (a-d)
Actual and Predicted Azinphosmethyl Dislodgeable and Soil
Residues for the Canopy Region, 1978 Season, using pre-
dicted initial distribution for each spray period. Upward
directed arrows show dates of spray application, and down-
ward directed arrows indicate dates and amounts of rainfall
in mm.^ Each bar is divided into two parts; the mean
(yg/cm ground area) is above the line and its corresponding
standard error (S.G.) below the line. The solid line
running through the bars represents the model's prediction
of the residues using 1976 and 1977 data only, while the
dashed line shows the predicted residues based on all three
years (1976-1978) data.
46

-------
« rrin iwu
1.6
1.2
r 0.0
Ul N
n u 0>i
m
FILIEY GRflSS-BROflDLEflVES 1978
T
T
c o
£ 3 0.0
0.«
o.e
2 n n si
wrr
3	s ip3 | e
4	I 2
V
1
Vf-
j | . -i	



ISO
—t—r
julIpn onrc
' '	i—i—<—i—i-
t
2SS
I rain inrti —-
ALLEY LITTER-nCSS 1978
I !TT W * I
1.2
0.9
£ 0.6
U N
w 5 0.3
O N
r o
£ =0.0
a 0.3
0.6
JM V i 6
2 n it t\
*
m j
* l 2
71
TT~7
rl" T:"'i-IT'N*
L	l_
ISO
¦ ' ' » ' 	I	1	L_l	1	1	1	1—J	J
T	T
JULIA* Oflff
2SS
RLLEr SOIL 1978
Figure II-8(e-g). Actual and Predicted Azinphosmethyl Dislodgeable and
Soil Residues for the Alley Region, 1978 Season.
Upward directed arrows show dates of spray applica-
tion, and downward directed arrows indicate dates
and amounts of rainfall in mm. -Each bar is divided
into two parts; the mean (yg/cm ground area) is
above the line and its corresponding standard error
(S.E.) below the line. The solid line running through
the bars represents the model's prediction of the change
in residues using 1976 and 1977 data only, while the
dashed line shows the predicted change in residues
based on all three years' data.
47

-------
canopy t.rnvr:i r.r/H
• Mia imi
.'FIT IT]	W|.' I TT
» it ii «i	t	i i ?	• i«
UifcJhJiStk
"S*

CANOPY GftflSS-BROflOLtftVES 1978
ill# I * 1 #!•' I •'! ¦'
£ e.t
s-
*0.8
sr o.«
o.«
t if ii ti
• I I	•	14
-M
t ' 1 1 t
« M|« (Mil
crnopt urrER-noss 197a
~ »m 1 •' 1 .wii 1 •'! i
¦ o.<
*io.»
Sfl.o
S 0.1
> iv 11 11
111 ¦ ¦ t
f r
JUUM Qfltl
1 imi |
CflNOPt SOIL 1978
\ £o.«
f 30.0
¦'Mi 1 •' I I !| •'
¦	" 1 I	• |«
i^msi
	"f
T
Figure II-9 (a-d) Actual and Predicted Azinphosmethyl Dislodgeable and Soil
Residues for the Canopy Region, 1978 Season, using the
observed initial distribution for each spray period. Up
ward directed arrows show dates of spray application, and
downward directed arrows Indicate dates and amounts of
rainfall in^nnn. Each bar is divided into two parts; the
mean (yg/cm ground area) is above the line and its cor-
responding standard error (S.E.) below the line. The line
running through the bars represents the model's prediction
of the residues using 1976 and 1977 data only, while the
dashed line shows the predicted residues based on all three
years (1976-1978) data.
48

-------
1.6
1.2
£ o.e
UN
w u0<4
© V
z o
£ 3 0.0
a 0.4
o.e
w
17 11
ISO
Jk
ALLEY GRASS-BROADLEAVES 1978
WW
T
61

3N3
4 1
T
rr~r
t ' ' ' I
i i.i »
JULIAN OATE
265
ALLEY L I TTER-flOSS 1978
g rain inn)•
2.0
1.6
£ 1.0
UJ W
c c n e
to (j O >5
o \
s o
& => 0.0
0.6
1.0
ALLEY SOIL 1978
T
Hf 11 1 #11 I >|
2 17 II SI	S	4 12	8 14
Ti r
150

-I	J	1	I	I	L_j_l	I-
JUL[AN DATE
Figure II-9(e-g). Actual and Predicted Azinphosmethyl Dislodgeable and
Soil Residues for the Alley Region, 1978 Season, Using
the Observed Initial Distribution for Each Spray Period.
Upward directed arrows show dates of spray application,
and downward directed arrows indicate dates and amounts
of rainfall in Mm. Each bar is divided into two parts;
the mean (ug/cm ground area) is above the line and its
corresponding standard error (S.E.) below the line.
The solid line running through the bars represents the
model's prediction of the residues using 1976 and 1977
data only, while the dashed line shows the predicted
residues based on all three years' (1976-1978) data.
49

-------
It is often difficult to determine an appropriate criterion for goodness
of fit of model predictions to field determined values. In this study, the
heterogeneity of the samples, lack of uniformity of application, variation
in the microclimate, and other factors combined to yield large errors in
residue estimates. In Table II-l, column 3 indicates the root mean square
error (RMSE) of the daily mean for each layer from the values on which it
is based. Clearly no model prediction could be expected to have a smaller
RMSE from the measured residue values than the RMSE for the means shown in
column 3.
Validation of the model is most convincingly demonstrated by comparing
model predictions with a set of data which were not used in model parameter-
ization. Therefore, for this purpose the movement and attenuation model
was parameterized with 1976 and 1977 data only. In column 1 model predictions
using these matrices together with 1978 application rates and environmental
conditions are compared with measured 1978 field residues. Thus a comparison
of column 1 with column 3 indicates the predictive performance of the model
relative to the best possible prediction of each day's residue distribution.
Model predictions based on matrices parameterized with data from all three
seasons (1976-1978) are not independent of the 1978 residue data being used
for model validation (see Figures II-4, II-5, II-6, II-7). However, these
matrices are still of interest as our best estimate of the orchard residue
dynamics. Column 2 shows RMSE's between measured 1978 field residues and
predictions based on these matrices together with 1978 application rates and
environmental conditions.
Comparison of the RMSE's in column 3 of Table II-l shows the expected
direct relationship of RMSE to sample mean. The similarity of the RMSE's in
columns 1 and 2 indicate little difference in the predictive capability of
the matrices, based on 1976 and 1977 data, with or without the 1978 data.
Comparison of columns 1 and 2 with column 3 shows the best fit of the pre-
dicted values was In the Grass-Broadleaf layer. The RMSE's for this layer
averaged 14% greater than the mean RMSE for both predictions. The RMSE's
for the canopy litter-moss layer, based on both predictions, averaged 24%
and the canopy soil averaged 22% greater than the mean RMSE, while the tree
layer RMSE based on both predictions averaged 21% greater than the mean RMSE
for this layer. The poorest fit was in the Alley litter-moss and soil layers.
RMSE's for these layers based on predicted values averaged well over 100%
greater than the mean RMSE. This may be due to the greater heterogeneity in
composition of the alley litter-moss layer and low residue levels distributed
to these layers.
The accuracy with which the model predicts post-deposition residue move-
ment and attenuation is partially masked by the errors in prediction of
initial distribution and drift. While sophisticated drift models can be de-
veloped, the parameterization of these models for a particular field situation
requires an intensive measurement effort. For some spray applications, our
simple drift model, based on limited measurements of drift, caused predicted
residue levels on spray days to differ considerably from the measured levels.
These differences affect the model's accuracy throughout the remainder of
the season. In order to eliminate the effects of drift and to better eval-
uate the model's ability to track residue movement and attenuation, a separate
test of the model was made. In this test, on each day of spray application,
the residue distribution in the model was set to the distribution observed
in the field on that day. The model then was used to predict only the sub-
sequent residue attenuation and movement. Figures II-9(a-g) shows these
predicted residues versus the measured values for 1978, in the same format as
50

-------
TABLE II-l. COMPARISON OF THE MODEL PREDICTIONS (INCLUDING DRIFT)3 WITH AZINPHOSMETHYL RESIDUE DATA FOR THE
1978 SEASON
Region	Layer	Root Mean Squared Error*5 From Sampled
1978 Residue Data
12	3


1976-1977
Prediction Matrices
1976-1978
Prediction Matrices
1978 RMSE'S
Dally Mean
Sample
Size
Canopy
Leaves
2.699
2.387
2.102
164

Grass-broadleaves
0.327
0.290
0.262
131

Litter-moss
0.300
0.293
0.239
136

Soil
0.740
0.832
0.642
114
Alley
Grass-broadleaves
0.253
0.249
0.228
122

Litter-moss
0.148
0.126
0.017
144

Soil
0.480
0.338
0.130
105
Mean squared errors for the relationship between the predicted and sample values shown in figures 6a-6g.
Column 1 shows the relationship of the 1978 data to the model parameterized with data from 1976-1977, while
column 2 shows the 1978 data versus the model parameterized with 1976-1978 data. Column 3 shows the root
mean squared error determined for the individual 1978 samples about each sample day's mean.
Spray distribution was predicted using a model for drift and deposition. Therefore errors in drift prediction
effect the RMSE values shown..

-------
Figures II-8 (a-g). In Table II-2, the fit of the predictions to the measured
residues is quantified. As in Table II-l, column 3 shows the root mean square
errors of the daily mean for each layer from the values on which it is based.
The sample numbers and most means are lower than in Table II-l because the
spray date measurements have been excluded. Columns 1 and 2 in Table II-2
are analogous to those columns in Table II-l, with the errors due to initial
spray distribution and drift eliminated.
Comparison of columns 1 and 2 with column 3 in Table II-2 shows the best
fit of the prdicted values was in the soil layer. The RMSE's for this layer
averaged 17% greater than the mean RMSE for both predictions. The RMSE's
for the grass-broadleaf layer, based on both predictions, averaged 25% greater
than the mean RMSE, while the tree leaf layer RMSE based on both predictions
averaged 39% greater than the mean RMSE for this layer. The poorest fit
was in the litter-moss layer. RMSE's for this layer based on predicted values
averaged 61% greater than the mean RMSE for the canopy region, and well over
100% greater than the mean RMSE for the alley region. Again, this may be
due to the greater heterogeneity In composition and low residue levels dis-
tributed to this layer.
The effects of rainfall on the model predictions are quite striking.
In some layers a net movement out is seen while in others more pesticide is
carried in. Of major importance Is the fact that throughout the season the
residues in the tree are continually redistributed to the orchard floor.
This movement counteracts the attenuation mechanisms, resulting in a slower
net loss rate of pesticide and, in some situations, a net Increase In residue
deposits to layers of the orchard floor. Such information on pesticide dynamics
within the orchard ecosystem has in the past been given little consideration
in assessing crop protection. Some pest species of apples and some of their
natural -enemies have life stages which Inhabit the orchard floor, subjecting
those populations to the pesticide residue dynamics of these layers. More
information is needed on the possible Impacts of pesticides on these popula-
tions and their interactions. Future pest management strategies, involving
the integration of chemical control techniques with biological, cultural,
and other control measures, must consider pesticide fate throughout the
entire orchard ecosystem.
Although the estimation of pesticide fate Is an end in itself, the over-
all goal of this research was to model the effect of azinphosmehtyl on ground
dwelling invertebrates which inhabit the orchard. Field studies conducted
concurrently within the same orchard, along with laboratory data, collected
on the isopod Trachellpus rathkei (Snider, 1979; Snider and Shaddy, 1980),
were used to develop a model describing its ecobiology and temporally dis-
tributed mortality (See Chapter X and Goodman, 1982). A form of the fate model
presented here was used to determine the time-course of azinphosmethyl exposure.
The model in the form reported here predicts azinphosmethyl redistribution
and attenuation as a function of time and rainfall data only. In order to
account for the effects of other environmental factors on pesticide dynamics,
a more sophisticated model is required. Further model development should
involve the decomposition of both the movement and attenuation matrices to
represent their component physical, chemical, and biological processes, each
a function of the relevant environmental variables (see Chapters III and IV).
This decomposition requires controlled experiments, many of which can only
be performed In a laboratory setting. However, the data and parameters reported
here remain as a set of field measurements against which such lab studies
may be compared or calibrated.
52

-------
TABLE II-2. COMPARISON OF THE MODEL PREDICTIONS (EXCLUDING SPRAY DATES)® WITH AZINPHOSMETHYL RESIDUE DATA FOR
THE 1978 SEASON
Region
Layer
Root
1978
Mean Squared Error**
Residue Data
From Sampled



12 3
1976-1977 1976-1978 1978 RMSE's
Prediction Matrices Prediction Matrices Dally Means
Sample
Size
Canopy
Leaves
2.699
2.387
2.102
164

Grass-broadleaves
0.327
0.290
0.262
131

Litter-moss
0.300
0.293
0.239
136

Soil
0.7A0
0.832
0.642
114
Alley
Grass-broadleaves
0.253
0.249
0.228
122

Litter-moss
0.148
0.126
0.017
144

Soil
0.480
0.338
0.130
105
Mean squared errors for the relationship between the predicted and sample values shown in the figures 6a-6g.
Column 1 shows the relationship of the 1978 data to the model parameterized with data from 1976-1977, while
column 2 shows the 1978 data versus the model parameteized with 1976-1978 data. Column 3 shows the root
mean squared error determined for the individual 1978 samples about each sample day's mean.
^o better evaluate the model's ability to track residue attenuation and movement and to eliminate the error
associated with drift estimates. The residue distribution was set to the measured values after each spray.
Therefore spray dates were excluded in the calculation of RMSES's.

-------
CHAPTER III
AIRBORNE LOSS FROM THE ORCHARD
INTRODUCTION
Recent field studies have indicated that airborne loss plays a major
role in the attenuation of field-applied pesticide residues (Taylor et al.,
1976; Taylor et al., 1977; White et al., 1977; Woodrow et al., 1977; Claith
et al., 1980). Taylor et al. (1977) report that attenuation of dieldrin
applied to orchard grass was due solely to volatilization. In-depth reviews
of both laboratory and field studies have been prepared by Spencer et al.
(1973) and Spencer and Claith (1975, 1977). Volatilization of pesticides
under field conditions is discussed by Taylor (1978). Techniques for measur-
ing volatilization rates in the field are described by Caro et al. (1971) and
Parmele et al. (1972). Orchard sampling of airborne pesticides bas been for
the most part limited to studies of worker re-entry inhalation or dermal
contact with airborne residues (Iwata et al., 1977; Gunther et al., 1977).
Airborne loss consists of spray droplet drift, evaporation and volatili-
zation at application, and post-application losses by volatilization and wind
erosion (Ebling, 1963; Leonard et al., 1976). The work described here was
designed to assess the contribution of airborne loss to the overall attenuation
of the organophosphate insecticide azlnphosmethyl, 0,0 dimethy1-S-(4-oxo-l,
2,3, benzotriazin-3(4h)-ylmethyl) phosphorodlthloate (Guthlon), in a Michigan
apple orchard ecosystem. Previous field work (Chapter I) designed to estimate
the distribution, attenuation, and movement of azlnphosmethyl in a deciduous
orchard environment suggested that airborne loss was the major pathway of pest-
icide attenuation soon after application. In the present study airborne re-
sidues of azlnphosmethyl were measured directly, accompanied by meteorological
data, over a season of periodic low-volume spray applications. These measure-
ments along with orchard deposit residue data were used to estimate pesticide
airborne flux from the orchard as a part of the total attenuation of pesticides
over the season.
ANALYTICAL METHODS
Procedures for the residue analysis of all sample types are described
in Chapter I. Briefly, dislodgeable residues were determined by the procedure
of Gunther et al. (1973), surface-penetrated residues procedure of Steffens
and Wieneke (1976), and soil residues by the procedure of Schulz et al. (1970).
Targets, consisting of two circles of Whatman No. 1 paper, 18.5 cm diameter
(attached one atop the other by a single staple to a cardboard backing), were
extracted for two hours in a soxhlet extractor using a mixture of hexane:
acetone, 100 ml:25 ml.
Airborne Residues. The porous polyurethane foam (PPF) plugs (ester form),
five cm long and 4.5 cm in diameter, were used as the trapping medium for
sampling azlnphosmethyl airborne residues. The PPF plugs were rinsed with
distilled, deionized water in a Nalgene pipet washer for six hours followed
54

-------
by soxhlet extraction with 500 ml acetone for six hours and then 500 ml hexane
for six hours prior to use. This method Is similar to that used by Turner
and Glotfelty (1977). After sampling, the FPF plugs were soxhlet extracted
for four hours with 500 ml of a 4:1 hexane-acetone solution. Extracts were
further concentrated by rotary evaporation followed by air evaporation prior
to analysis by GLC.
Trapping ability and extraction procedures were tested in the laboratory.
A closed glass system consisting of a volatilizing chamber (U-tube) followed
by two plugs In series, followed by a cold trap (dry ice in acetone) was used.
Air flow was drawn at 1.0 m /hr for 60 minutes. At this flow no pressure
drop could be measured with a mercury barometer. No azlnphosmethyl was found
in the second plug or the cold trap (limit of detection was 1.0 ng). Recovery
and standard deviation was 91.7 + 1.6 % from the volatilization of 42.8 + 13
pg(three repetitions)^ The detection limit for azlnphosmethyl in air was
approximately .Olyg/m
Quantitation of azlnphosmethyl was accomplished using a Tracor 560 gas
chromatograph having a flame photometric detector in the phosphorus mode.
A six-foot glass column (2 mm i.d.) packed with 3% SE-30 on Gas Chrom Q, 60/80
mesh, was operated at 195 C;N„, 40 ml/min; air, 90 ml/mln; H^, 60 ml/mln.
This system was interfaced with a digital PDP 11/20 RSTS computer for Inte-
gration of the area under the single peak produced.
EXPERIMENTAL SITE AND PESTICIDE APPLICATION
The apple orchard was located in the vicinity of Grand Rapids, Michigan
in Kent County. The 12 one-year old trees were a mixture of semi-dwarf duchess
and wealthy cultlvars. In the spring of 1976, the orchard received a moderate
pruning and cleaning to counter several years of disuse. The orchard plan
was mowed to a height of approximately 5 cm prior to the first, second, and
third spray applications. Four experimental plots and one control plot
(Figure III-l) were established by a method designed to facilitate the study
of azlnphosmethyl loss via surface runoff study conducted in 1977 are reported
in Chapter I.	^
Azlnphosmethyl (Guthion , 50% W.P., Mobay Chemical Company) was applied
with a custom-built, low pressure, low volume sprayer similar to one developed
by Howitt and Pshea (1965).. The across plot^average application rates and
standard deviations (Kg ha , 50% w.p. 100H ha ) were 1.3+0.2 on June 9.
1.6+0.6 on June 29, 1.0 on July 21, and 1.6 + 0.2 on August 10. A more
detailed description of the experimental site, sprayer design, and applica-
tion techniques can be found In Chapter I.
SPATIAL STRUCTURE OF THE ORCHARD
The orchard plots were subdivided both vertically and horizontally.
Horizontally, each plot was subdivided into four regions potentially differ-
ent with respect to initial pesticide distribution (see Figure III-2). Region
4 (under the canopy) was determined to be 32% of the area in plot 1, 34% in
plot 2, 54% in plot 3, and 45% in plot 4. The remainder of the plot was
divided among the three alley regions in a ratio of 2:3:1 for plots 1 and 2,
and 1:2:1 for plots 3 and 4. The difference in the ratios was due to the
closer spacing of the rows along the North-South axis in plots 3 and 4. The
number of trees was 24 in plot 1, 15 in plot 2, 10 in plot 3, and 11 in plot
4. Tree heights averaged 3.0 meters. Vertically, each alley region was
divided into three compartments; grass-broadleaves, litter-moss, and soil.
55

-------
Figure III-l. Orchard Hots Showing Air Sampling Locations
CONTROL 7
+
~ 1
PLOT 3
PLOT 5
~ 6
PLOT 6
PLOT 4
A 4
A 5
+• runoff collection points
A air collection points
¦ wind measurement points
56

-------
Ln
1 > * » »
2.0 1.0 4.0 t.O 0.0
UO/M*
-J fi.O
*1 ' ¦ ¦ 1 1 1 1 1
UO/tt)
Figure 111-2. Diurnal Azlnphosniethyl
Airborne Residues at Sampling
Heights Between 0.5 and 6.0
Meters Above the Orchard,
Measured at the Center
Location During the First
Spray Period of the 1978
Season
_j	i	i	i	i	i	i
30 ao «.o i.o «.o t. o #.o
wo/«s
DAT It
4.0
iio/m)
go/at'

-------
ORCHARD SAMPLING PROCEDURES
Deposit residue sampling was carried out in a manner similar to the two
previous seasons, as described in Chapter I. An abbreviated description of
sampling procedures is given here. To determine azinphosmethyl initial
horizontal distribution, a series of 15 to 20 ground-located, filter paper
targets was deployed among the four regions (Figure III-2) in each plot prior
to application. Targets were placed in the center of the alley regions and
at the ln-row mid point between the tree trunk and the edge of the canopy
region. The distribution of azinphosmethyl residues was also determined
for the trees and among the three ground layers (grass-broadleaves, litter-
moss, soil). In each plot, three trees were sampled and four samples of
each ground layer were taken, two from the canopy and two from the alley
regions. Trees were subdivided for sampling into lower (1-2) and upper (2-3 m)
regions. Tree heights averaged 3 m. Thirty to 35 leaf discs were taken per
sample^ Block sampling was employed for both ground samples and trees. One-
half m ground area sites and trees were selected randomely for consecutive
sampling in the first and second spray periods; areas at the end of rows
and close to adjacent plots were avoided. Similarly, new sites were chosen
for consecutive sampling during the third and fourth spray periods. Plots
were sampled In sequence, between 1300 and 1500 hr at various times following
application, as shown in Table III-l.
AIRBORNE RESIDUE SAMPLES
Two air sampling masts were employed, one located at the center of the
sprayed plots and one at the downwind edge (Figure III-l). The PPF plugs
were positioned in glass cylinders, tapered at one end, and located at heights
of 0.5, 1.0, 3.0 and 6.0 meters above the orchard floor. Air was drawn through
each plug at a rate of 8.33£/mlnute giving a combined flow rate of 33.33£/
minute measured by flowmeter, controlled by an adjustable needle valve. Flow
rate was maintained with an electric vacuum pump powered by a gas generator.
On June 9 the orchard air was sampled during application at the downwind
edge (Location 2, Figure III-l) and also at Location 1, as the wind direction
had a slight northern component. Samples were taken from 0700 to 1000.
During the June 29 application samples were taken at the downwind edge (Location
2) and 20 meters from the downwind edge (Location 7). The .two remaining
spray applications (July 21, August 10) were sampled at the downwind edge
only. The orchard air was sampled in the afternoon following each application
and on successive dates for two hours, between 1300 and 1500 (see Table III-l).
Wind speed was measured at three meters (Location 12) during application and
at three and six meters (Location 13) during afternoon sampling periods. Wind
direction was continuously recorded from a wind vane mounted on the same mast.
Additional measurements of temperature, humidity, solar radiation, and rain-
fall were continuously recorded throughout the season (Location 9). Rainfall
amounts and wind speed data are included in the summary of significant events
for the 1978 season shown in Table III-l. Temperature and humidity data are
shown in Figures III-7 and III-8.
RESULTS AND DISCUSSION
Table III-2 shows the results of the analysis of targets for tt^first
spray day of the 1978 season. The average amount is express in pg/cm ground
area and proportions shown are proportions of the dose applied to each plot.
58

-------
Table III-l. Summary of Significant Events for the Orchard, 1978 Season
Spray
Date
Sample
Date
Days Since
Spray
Rainfall
Date
Rainfall
Amount
(mm)
3 m
Windspeeds
(m/s) during
Air Sampling
6 m
Orchard Mowed 1 June
9 June
29 June
9 June
12 June
15 June
22 June
27 June
29 June
3 July
7 July
10 July
19 July
0
3
6
13
18
0
4
8
11
20
7	June
8	June
12 June
16	June
17	June
18	June
25 June
Orchard Mowed 28 June
1 July
9 July
2.0
6.1
17.3
3.6
11.2
3.0
51.3
5.6
5.1
2.4C, 2.2
2.7
1.6
1.3
1.6
e
1.3
2.1
0.9
4.4
4.5
3.9
2.2
2.1
e
2.0
3.6
2.2

-------
Table III-l continued
Spray
Date
Sample
Date
Days Since
Spray
Rainfall
Date
Rainfall
Amount
(mm)
Windspeeds
(m/s) during
Air Sampling
Orchard Mowed 20 July
21 July
10 Aug
21 July
24 July 3
28 July
4 Aug
9 Aug
10	Aug
11	Aug
14 Aug
0
21 Aug
25 Aug
15
20
0
1
4
11
15
20	July.
21	July
22	July
24 July®
26 July
29 July
2 Aug
9 Augk
16 Aug
19 Aug
3.0
4.3
5.6
0.8
3.0
1.5
7.6
8.1
3.6
14.2
N.D.
0.5d.
N.D.
1.6
1.8
0.7
0.91
0.9
0.7
0.5
l.r
N.D.
0.9
N.D.
3.0
3.1
1.3
.3
1.5
0.9
1.5
aTwo hr mean, measured at 3 m and 6 m above the orchard floor (Figure 1, Location 6) between 1300 and 1500 hours.
^Rainfall occurred before sampling.
°Six hr mean, measured at 3 m above the orchard floor (Location 9) between 0600 and 1200 hours.
^Estimated value determined from the average ratio of wind speeds at 3.0 m and 6.0 m for the entire season
and the wind speed measured for this date.
Recorder malfunction
^None detected, limit of detection 0.5/ms.
^Rainfall occurred after sampling.

-------
100
Figure 111-7. Dally Temperature Range, 1978 Season
90
80
70
60
50
40
30
- 0
June
July
August
61

-------
Figure II1-8. Daily Minimum and Maximum Relative Humidity, 1978 Season
100 r-
90
80
70
60
R.H.
50
40
30

June
July
August
62

-------
Table III-2. Azinphosmethyl Initial Horizontal Distribution from the Analysis of Targets, First Spray
Period of the 1978 Season.
Plot
Alley(l)
Alley(2)
Alley(3)
Alley(4)
w
1	Average Amount (yg/cm + SE)
Sample size
Proportion
2
2	Average Amount (ug/cm + SE)
Sample Size
Proportion
2
3	Average Amount (yg/cm + SE)
Sample Size
Proportion
2
4	Average Amount (yg/cm + SE)
Sample Size
Proportion
2
Grand Average Amount (yg/cm + SE)
1.40 + .15
3.
.052
0.62 + .16
3.
.021
1.09 + .35
5.
.018
.018
1.
.004
0.2 + .27
1.16 + .33
3.
.052
1.37 + .34
5.
.096
1.94 + .28
4.
.053
1.24 + .27
3.
.059
1.431 + .18
0.86 + .27
4.
.010
1.52 + .25
4.
.030
0.95 + .4
5.
.014
0.57 + .15
6.
.016
0.98 + .20
1.04	+ .15
4.
.044
1.09 + .30
5.
.081
1.05	+ .35
3.
.062
1.07 + .22
3.
.094
1.06	+ .01

-------
The grand average across plots shows that alley region 2 and the canopy
region receive the majority of the residues reaching the ground. This agrees
with the pesticide horizontal distribution data for the 1976 and 1977 seasons
(Chapter I).
Vertical pesticide distribution for the first spray period of the 1978
season is given in Table II1-3. The across plot average is represented by
only two regions; alley and canopy. The average amounts for the alley re-
present weighted averages of alley regions 1, 2 and 3, weighted by % plot
area. These data agree with the results of the 1976 and 1977 field studies
in that the majority of the residues are initially distributed vertically to
the tree leaves and the grass-broadleaves layer. Orchard soil received ap-
proximately 1.5% of the dose applied.
The vertical distribution of pesticide residues as proportion of dose
applied is shown in Table III-4. Proportions were determined from the dose
applied t02each plot and then averaged. Tree surface area was estimated as
460,000 cm /tree (see Chapter I). The proportion of the pesticide applied
that was distributed as surface penetrated residues was estimated from matched
samples from the first spray day of the 1978 season. The surface-penetrated
to total dlslodgeable residue ratios and standard errors were determined as
follows: leaves, .081 + .015; grass-broadleaves, .190 + .002; litter, .421
+ .078; moss, .272. These ratios were multipled by the appropriate dlslodge-
able residue proportions to estimate the surface-penetrated proportion. Soil
residues were determined on a whole sample basis. Summing all dlslodgeable,
surface-penetrated and soil residue proportions for the first spray day of
1978, across both alley and canopy regions, yielded the estimate that 72.1%
of the pesticide applied was initially deposited in the various orchard
strata. This estimate is considerably higher than the 1977 estimate of 55.6%.
The 1976- estimate was 66.1% (Chapter I). These are only crude estimates which,
after accounting for extraction efficiency, assume that all residues In the
samples analyzed were accounted for (a more detailed discussion of the point
can be found in Chapter I). The most likely sources of error are: (1) under-
estimation of the amount applied and (2) uneven application (due to steep
slope In plot 4) and sampling not representative of residue levels present.
The estimated proportion of pesticide applied initially distributed to the
trees in plot 4 for the first spray of the 1976 season was also greater than
the estimates for the other plots resulting in a total deposit estimate of
84.6% of the dose applied for that plot. The 1977 data shows no great varia-
tion in tree proportion (31.5%, S.D., 6.2%) among plots.
On the premise that all the residues deposited in the orchard were ac-
counted for, 27.0% of the dose applied to the orchard in the first spray
period of the 1978 season is assumed to be airborne loss, primarily
as drift at application, but also as volatilization and wind erosion in the
four hours between application and sampling. Table III-5 shows estimates by
plot of airborne loss using target data to determine proportion distributed
to the orchard floor. The across-plot average airborne loss estimate of
26.7% is in close agreement with the estimate made using the orchard floor
deposit residue samples.
Azinphosmethyl concentrations in the air measured during application at
sampling heights between 0.5 meters and 6.0 meters above the orchard floor
are given in Table II1-6. For those samples taken at the downwind edge,
concentrations measured are roughly uniform from 0.5 to 6.0 meters. Differ-
ences in mean amounts sampled between the four spray dates are thought to be
primarily a function of wind speed. Temperature and relative humidity (R.H.)
also determine the amount of pesticide reaching the downwind edge sampling
64

-------
Table III-3. Initial Vertical Distribution of Azinphosmethyl Dislodgeable Residues,3 First Spray Period
of the 1978 Season
	Leaves	
Upper	Lower	Grass-Broadleaves Litter-Moss	Soil
Canopy
Average Amount (ug/cm^+SE) 1.69 + .17 4.64 + .71	0.59 + .03	0.11 + .05	0.21 + .04
Sample Size	11	12	4	6	8
Alley
2
Average Amount (yg/cm +SE)	1.00 + .56	0.12 + .04	0.09 + .05
Sample Size	4	7	5
aExcept soil residues which were determined on a whole sample basis.
b	2	2
Across plot average and standard errors expressed as Ug/cm ground area, except tree leaves which are Ug/cm
leaf area. Lower (l-2m) and upper (2-3m) regions of the tree were sampled separately.

-------
Table III-4. Estimation of Initial Vertical Pesticide Distribution as a Proportion of the Amount Applied
(+ S.E.) from the Analysis of Samples for the First Spray Period of the 1978 Season.
Alley	 	Canopy

Dlslodgeable
Surface-
Penetrated
Dlslodgeable
Surface-
Penetrated
Total
Tree


.513 + .093
.036

Grass-Broadleaves
.077 + .036
.015
.028 + .002
.005
.721 + .100
Litter-Moss
.009 + .002
.004
.008 + .003


Soil
.009 + .005

.014 + .004



-------
Table III-5. Estimation of Initial Vertical Pesticide Distribution as a Proportion of the Amount
Applied, Using Target Data to Calculate the Proportion Reaching the Orchard Floor, for
the First Spray Period of the 1978 Season.
Orchard Floor			Tree	
Surface-
Plot	Alley	Canopy	Dislodgeable	Penetrated	Total
1
.114
.044
.391
.028
.607
2
.147
.081
.483
.029
.670
3
.085
.061
.549
.032
.637
4
.079
.094
.789
.056
1.018
.733 + .096

-------
3
Table III-6. Azinphosmethyl Concentrations (yg/m ) at Sampling Heights between 0.5 and Six Meters Over Orchard.
Date


Center


Location3

Downwind Edge
Location

Days Since




Height Above Orchard
Floor




Spray
0.5
1
3
6

0.5
1
3
6

9 June^






17.70
27.70
29.30
22.80
2
9 June






12.70
3.40
15.80
15.10
1
9 June
0
6.90
5.20
4.80
4.40
6
5.70
4.60
4.40
5.30
2
12 June
3
1.58
0.23
2.10
1.00
6
0.35
0.92
2.40
1.13
4
15 June
6
0.36
0.17
0.78
0.59
6
0.13
0.09
1.47
0.80
1
22 June
13
0.17
0.44
0.37
0.93
6
0.07
0.13
0.27
0.31
1
27 June
18
0.07
0.07
0.23
0.17
6
0.05
N.D.c
0.29

2
29 June






26.40
25.00
24.10
27.60
2
29 June






6.67
7.69
27.70
9.54
7
29 June
0
1.00
1.92
2.88
1.38
6
1.94
2.53
0.91
0.64
2
3 July
4
1.55
1.26
0.35
0.67
6
1.77
.96
0.59
2.30
5
10 July
11
0.09
0.62


6
0.32
0.24
1.00
0.05
3
19 July
20
0.17
0.11
0.11
0.22
6
0.25
1.14
0.11
0.16
2
21 July






11.82
18.18
18.18
13.33
1
21 July
0
1.02
0.19
0.57
0.82
6
1.30
1.51
0.87
1.15
1
24 July
3
0.63
0.35
0.91
0.38
6
0.36
0.38
0.35
0.53
1
28 July
7
0.23
0.09
0.22
0.16
6
0.13


0.25
2
4 Aug
15
0.09
0.05
0.05
0.09
6
0.16
0.32
0.28
0.25
3
9 Aug ,
20
0.06
0.21
0.14
0.22
6
0.57
0.09
0.07
0.10
1
10 Aug






14.06
15.22
18.33
16.67
4
10 Aug
0
0.86
1.29
3.13
0.95
6
0.57
0.27
0.43
0.25
3
11 Aug
1





0.73
0.40
0.61

2
14 Aug
4





0.34
0.70
0.77

2
21 Aug
11





0.32

0.67

2
25 Aug
15

0.22


6





^Location of sites shown in Figure 1.
Azinphosmethyl concentrations measured during application, all others are 2 hr mean sample concentrations taken
between 1300 and 1500 hours.
None detected, limit of detection .01 Mg/m -

-------
sites. Conditions favoring rapid evaporation tend to reduce the droplet
spectrum towards smaller droplets which travel farther before being deposit-
ed out. Temperature, through its effect on atmospheric conditions, also in-
fluences spray drift. Inversion conditions favor a smaller drift cloud by
discouraging vertical diffusion. Turbulent atmospheric conditions accelerate
both vertical and lateral dissipation of spray droplets (Yates and Akesson,
1973; Maybank and Yoshida, 1977). The mea^ pesticide concentration measured
during application on June 9 was 22.9 pg/m . The mean wind speed at 3.0
meters was 2.2 m/s. The temperature ranged from 3.9°C (0600) to 11.9°C
(1200) and the R. H. ranged from 92% (0600) to 35% (1200). On July 21^the
mean pesticide concentration measured during application was 15.4 yg/m. The
mean wind speed measured at 3.0 meters during sampling was 1.1 m/s. The tem-
perature ranged from 17.2°C (0600) to 18.4°C (1200) and the R. H. ranged from
89% (0600) to 78% (1200). Comparison of wind speeds between these two dates
suggests that the higher wind speed during application on June 9 was respon-
sible for the higher concentrations measured at the downwind edge. If the
airborne pesticide concentrations measured during application on these two
dates are scaled by the average amount of pesticide applied the difference
is more pronounced.
On June 29 a second mast was located 20 meters from the downwind edge.
Air samples taken during application indicated that there was rapid dissipa-
tion of airborne residues between the two masts at all heights except 3.0
meters. This persistence at 3.0 meters may be the result of pesticide moving
through the air as a plume generated by the closer passes of the sprayer.
Table III-6 also shows azlnphosmethyl concentrations in air at sampling
heights between 0.5 meters and 6.0 meters over the orchard at two locations;
the center of the sprayed plots (Figure III-l, Location 6) and at the down-
wind edge. Samples were taken between 1300 and 1500 hr on the day of appli-
cation and on selected days following application (Table III-l). Mean wind
velocities at 3.0 and 6.0 meters for the two-hour sampling period are also
given in Table III-l. Graphical representation of airborne pesticide residues
for the first spray period is shown in Figures III-2 and III-3. These fig-
ures show a systematic decline in airborne pesticide residues at all heights
above the orchard floor over the 18-day sample period. Data for the remain-
ing spray periods (Table III-6) also show this trend. As might be expected,
the decline in airborne pesticide residues agrees fairly closely with the
decline in orchard deposit pesticide residues (Figure III-4). Variation in
airborne pesticide residues between the heights sampled is the greatest on
the early sample days following application, with the highest airborne pesti-
cide at 0.5 meters the day of application (day 0) and at 3.0 meters 3 and 6
days following application. Lower levels of azlnphosmethyl airborne residues
were measured in the afternoon of the second, third, and fourth application
dates as compared to the first. In discussing reasons for observed diurnal
variations in vertical vapor flux of dieldrin and heptachlor applied to
pasture, Taylor et al. (1977) concluded the phenomenon was directly related
to diurnal variation in solar radiation input. The authors qualify this
statement with the hypothesis that solar radiation is one factor that may
be singled out as a source of vapor flux variation, as solar radiation input
affects all other parameters of the complex relationship between the crop
microclimate and the adjacent atmosphere. One parameter which solar radia-
tion most definitely affects is the temperature at the leaf surface, which
has a direct effect on the rate of volatilization. A similar conclusion
was reached by Phillips (1974) from the results of laboratory experiments
in which the volatilization rate of dieldrin from glass and cotton leaf
69

-------
Figure 111-3. Diurnal Azlnphosmethyl Airborne Residues at Sampling
Heights Between 0.5 and 6.0 Meters Above the Orchard,
Measured at the Downwind Locations During the First
Spray Period of the 1978 Season
\
I	1_
i I I
0.0 1.0 1.0 to 4.0
UG/M>
1 ' »
0.0 1.0 1.0 S.0 4.0 1.0 4.0 TjO to
UG/M>
1
J	I	I	L
¦ ¦ ¦ ' ¦
UQ/M*
UO/M1
70

-------
CANOPY LEAVES 1978
if
f
(f 1
11 II 81
II . .
! 1 «
6 4
III i
1 t 1
1 1
•1
14
1
I 4 _1_ 1	1	1	1	1	1	1	1_

CANOPY ORASS-BROAOLEAVES 1978
!||l 1
I IT II SI
Jj 1 ^
> J !
in ,
'PI*!
1 1 1
•l
14
I
L	1	1	


CANOPY UTTER-flOSS 1978

If .!•' 1 <
I | , 1 | l i
1
|! 1
t •
| |
•'I
14





*•	CANOPY SOIL 1978
""mijf | i | |
l.< -	I 17 II II	8	4 1 I	• 14
Figure 111-4(a-d). Azinphosmethyl Dislodgeable and Soil Residues for
the Canopy Region, 1978 Season. Upward directed
arrows show dates of spray application, and down-
ward directed arrows indicate dates and amounts of
rainfall in mm.o Each bar is divided into two parts;
the mean (ug/cm ) is above the line and its corres-
ponding standard error (S.E.) below the line.
71

-------
rain inn) ¦
1.2
0.9
-j
£ 0.6
UN
C E A .
co u 0*3
o v
| § o.o
N „ -
a 0.3
0.6
ALLEY GRASS-BROAOLEAVES 1978
H'it M i m* i n
Z IT II 51	6	4 1 2	8	14
¦ 1 I	1 ¦ ¦	x	I	I i i.i	1—I	I.I i i	1	1
160 *	1	T	1	246
JULIAN OATE
RAIN crni¦
0.8
ALLEY LITTER-MOSS 1978
wi i •* i w i »
0.6
i ... j
hJM
M 5 o.z
© ^
| § o.o
a 0.2
0.4 J
2 IT II 61
4 I 2
-M-
J—L
I I , .1
J	u
160
t ' ' ' t
JULIAN OATE
246
g.
RAIN (Hill •
1.6
t .2
-j
£ 0.8
UM
a 5 0.4
o x
r a
£ => 0.0
M . .
a 0.4
0.8
ALLEY SOIL 1978

¥
2 IT 11
T
L.
ISO
61
4 1
O-J—I || I 1 '
TT
14
-UU-4-

JULIAN OATE
246
Figure III-4(e-g). Azinphosmethyl Dislodgeable and Soil Residues for
the Alley Region, 1978 Season. Upward directed
arrows show dates of spray application, and down-
ward directed arrows indicate dates and amounts of
rainfall in mm.2 Each bar is divided into two parts;
the mean (pg/cm ) is above the line and its corres-
ponding standard error (S.E.) below the line.
72

-------
surfaces was measured at various temperatures and wind speeds. Phillips
stated that "clearly the factor having the most dramatic effect on loss was
temperature." These observations would suggest that a possible partial ex-
planation for differences in airborne concentrations measured in the after-
noon on successive spray dates may be tt^differences in solar radiation in-
put. The average solar radiation (cal/cm /min) measured during the after-
noon sampling periods on each of the four application days was as follows:
1.11 on June 9, 1.06 on Jyne 29, 0.57 on July 21, and 1.00 on August 10.
The value of 1.11 cal/cm /min on June 9 was 100% of the 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 the average
airborne pesticide concentration measured at the center location (Location
6, Figure 111-1) for these application days (Table I1I-1) shows a nonlinear
trend with the highest airborne residues on June 9, which had the highest
solar radiation input, and the lowest airborne residues on July 21, which
had the lowest solar radiation input. In addition to solar radiation, the
influence of wind speed on the airborne pesticide concentrations measured
was also examined. In the comparison of the mean wind speed at 6.0 meters
(see Table III-l) measured during the afternoon sampling period (1300-1500 hr)
with the average airborne pesticide concentration measured at the center lo-
cation for the four application dates (Table III-l), a nearly linear relation-
ship was observed. Again the June 9 application date, on which the highest
airborne pesticide concentrations were measured, also had the highest mean
wind speed. The July 21 application date, on which the lowest airborne
pesticide concentrations were measured, had the lowest mean wind speed (none
detected, < 0.5 m/s). Phillips (1971) demonstrated using wind tunnel ex-
periments that there was a marked increase in the rate of loss of dieldrin
deposits on glass surfaces at wind speeds of 3.2 Km/hr (.89 m/s) as compared
to still air. Dieldrin was deposited as a thin film in a manner designed to
prevent mechanical loss from wind. We feel that increased wind speed is
responsible for increased loss in azinphosmethyl deposits in the orchard.
Both volatilization and wind erosion are mechanisms thought to be affected
by wind speed. What is not clear is whether increased airborne loss of de-
posits due to increased wind speed should be reflected in the increased air-
borne pesticide concentrations measured, as increased wind speed, although
removing more pesticide, would have a diluting effect on concentrations
measured in the air. It is possible that wind erosion of freshly deposited
residues is responsible for the marked difference between airborne residues
measured the afternoon of June 9 and those measured the afternoon of July 21.
The combined effect of wind speed and solar radiation input must also be
considered. Further data collection under a variety of environmental con-
ditions (actual and/or simulated) is needed to better describe the processes
contributing to pesticide airborne loss.
On sampling dates following the application date, more often than not,
airborne pesticide residue levels measured at 6.0 meters exceeded those
measured at 0.5 or 1.0 meter. Six meters is twice the average tree height.
One wonders how far above the orchard airborne residue levels are measurable.
The data recorded did not allow estimation of horizontal airborne residue
flux by direct methods. Estimation of either pesticide vertical airborne
flux or mass horizontal flow of pesticide vapor by aerodynamic methods re-
quires knowledge of wind speed profiles. Reliable interpretation of these
profiles requires minimal fluctuation In atmospheric turbulence. While this
criterion may be satisfied by uniform stands of short crops (Parmele et al.,
1972; Taylor et al., 1977), the orchard profile does not satisfy it. An
73

-------
indirect estimate utilizing both deposit pesticide residues and airborne
residues is presented as an alternative approach.
Hartley (1969) proposed that the volatilization rate (F^) of one pesti-
cide, from a non-adsorbing surface, could be predicted from that (F ) of
another pesticide, given the vapor pressures (P) and molecular weights (m)
of the two compounds, by the following equation:
Pb(mb)1/2
F = b , F	(III-l)
b P (m )1/2 3
a a
This relationship assumes that vapor flux is solely a function of molecular
diffusion through the stagnant air closely surrounding the surface. Taylor
(1978) points out that when examining vapor flux from plant surfaces, this
equation is valid only during the time when there is complete coverage of
the plant surface, i.e., when the pesticide vapor pressure is not reduced
due to adsorption and is In effect volatilizing from itself. As the deposits
become increasingly smaller the vapor flux is no longer strictly a function
of molecular diffusion, but also a function of the degree of adsorption to
the plant surface. In addition residue may penetrate leaf tissues or accu-
mulate in cracks and fissures in the epicuticular wax superstructure or leaf
specialized structures. With these limitations in mind, equation (III-l)
was used to predict the airborne loss of azinphosmethyl from the orchard
using dieldrin vapor flux over orchard grass determined by Taylor et al. (1977),
along with vapor pressure data found in the literature. Using the aero-
dynamic model described by Parmele et al. (1972), Taylor et al. (1977) esti-
mated the dieldrin vapor flux to be 80.4 g/ha/hr during the 1300 to 1500
hours sampling period 3.5 hours following application^
The vapor pressure reported dieldrin is 2.6 x 10 mm Hg at 20°C (Spencer
and Clalth, 1969). Only an upger limit for the vapor pressure of azinphosmethyl
has been reported; < 7.5 x 10 mm Hg (no temperature given, Schrader, 1963).^
The vapor pressure of the ethoxy analog, azinphos-ethyl is given as 2.2 x 10~
mm Hg at 20°C by Ouellette and King, 1977. To better estimate the vapor
presuure of azinphosmethyl from the data available, the relationship
between the possible crystal structures of these two analogs was examined.
Rohrbaugh et al. (1976) examined, using xray diffraction crystallography,
the crystal and molecular structure of azinphosmethyl. The unit cell
stereograph depicted indicates that crystal structure is dictated, in
part, by overlapping of the nearly planar ring systems and by intermolecular
repulsion effects of the methoxy groups; the repulsion working against
the packing forces of the crystal. The increased size of the ethoxy
group of azinphosethyl should result in weaker packing forces by Increasing
the allowed distance between molecules. This effect is In turn reflected
in the heat of sublimation, which varies directly with the magnitude of
the packing forces, and the observed vapor pressure, which varies inversely
with the magnitude of the heat of sublimation (Barrow, 1966). In addition,
the melting point of azinphosmethyl (7374°C) is approximately 18° higher
than azinphosethyl (56°C), indicating stronger packing forces and
therefore a lower vapor pressure. Based on these observations, it was
estimated tha£ the vapor pressure of azinphosmethyl was slightly less
than 2.2 x 10 mm Hg at 20°C, the vapor pressure of its ethoxy analog.
This value was used in equation (III-l) as P. , along with dieldrin vapor
pressure of 2.6 x 10 mm Hg at 20°C as P , and 80.4 g/ha/hr, the field durnal
vapor flux (1300-1500 hr) as ,a, to give an upper limit estimate for azin-
phosmethyl diurnal airborne flux (1300-1500 hr) of 6.2 g/ha/hr. To examine
74

-------
the sensitivity of equation (III-l) to vapor pressure data, the value of
1.0 x 10" mm Hg, reported as the maximum for azinphosmethyl vapor pressure
by Schrader (1963), was also used, resulting in an estimated azinphosmethyl
diurnal airborne flux of 2.82 g/ha/hr. These predicted values of azinphos-
methyl diurnal flux (1300-1500 hr) from the orchard are by no means precise
estimates, but rather crude approximations, due in part to the general nature
of the azinphosmethyl vapor pressure data. Other sources of variation
are the differences in application rate (5.6 Kg/ha a.i. for dieldrin vs.
.65 Kg/ha a.i. for azinphosmethyl), and the increased influence of wind
that tree deposits may receive as compared to grass deposits. In theory,
if there were complete coverage in both experiments the application rate
should have little effect, as the vapor flux is independent of the
amount of pesticide applied. As both pesticides were applied as water
base sprays, which are deposited on plant surfaces as droplets, complete
coverage is doubtful (Taylor, 1978). Under these "nonideal" conditions,
vapor flux is no longer Independent of the amount of pesticide present.
As stated earlier, one explanation for the dependency of vapor flux on
pesticide deposit residues is the hypothesis that incomplete coverage
affects the apparent vapor pressure of the pesticide due to adsorptive
effects of the foliar surface. Depending on the degree of coverage,
part of the pesticide deposit will be unencumbered by external adsorptive
effects and will evaporate at the maximum allowable rate (governed only
by molecular diffusion through the stagnant layer); the remaining portion
of the pesticide deposit will volatilize at some slower rate, depending
on the degree of adsorption. To approximate the effect of foliar adsorption
(i.e., Incomplete coverage) on the vapor flux predicted in equation
(III1), it was assumed that the higher the application rate the greater
the percentage of the pesticide deposit that is affected by foliar
adsorption. This is a necessarily simple approximation of a complex
phenomenon, as additional factors including formulation, spray droplet
size, climatic conditions during application, and characteristics of the
foliar surface may influence the state of the pesticide deposit and the
volatilization rate (Ebling, 1963; Hartley, 1969; Hull, 1970). With
these limitations in mind, equation (III-l) was modified in the following
manner
1/2
W	. . °b	(III-2)
b = n ,1/2	a D~
where D is the application rate of pesticide a and D, is the application
rate ofapesticide D^. As 75% of the pesticide deposited in the orchard was
applied to the trees, tree surface area must be considered when comparing air-
borne loss from the orchard with that of a grass field. Therefore the appli-
cation rate of azinphosmethyl was based not only on the .263 ha of ground
area but also the .276 ha of tree surface area. The average application rate
for the first spray period of 1978 was 0.65 Kg/ha a.i. Three fourths was
applied to .276 ha tree surface area and one fourth to .263 ground surface
area. If the tree and ground surface areas are weighted by the proportion
of the pesticide they receive and averaged, the adjusted application rate is
0.63 Kg/ha a.i. This value was divided by 5.6 Hg/ha a.i. of dieldrin applied
to the grass pasture to give a D^/D^ ratio in equation (III-2) of .112. In
using this ratio as a crude approximation of foliar adsorptive effects on
vapor flux, it must also be assumed that both application rates represent
75

-------
pesticide deposits that are less than or equal to complete coverage. The
azinphosmethyl flux estimates determined from equation (III-2)^were 0.697 g/
ha/hr, using 2.2 x 10 mm Hg and 0.317 g/ha/hr using 1.0 xlO mm Hg for
the vapor pressure of azinphosmethyl. Diurnal airborne loss during this
sample period, for the remaining dates samples, was estimated by assuming
this loss is directly proportional to the measured horizontal flux (g*m *s )
at 3.0 meters, determined from airborne pesticide concentrations (Table 111-6)
and wind speed measurements at 3.0 meters taken during sampling (Table 1II-1).
Dally airborne loss was estimated on the work of Taylor et al. (1977), who
reported as marked diurnal change in volatilization rates of dieldrin and
heptachlor, when applied to orchard grass. Similar diurnal variations were
reported by Taylor et al. (1976) for soil incorporated dieldrin. These re-
searchers measured a peak flux early in the afternoon with virtually no vola-
tilization measured before 0500 or after 2300 hr. Based on these findings,
daily diurnal airborne loss was estimated assuming peak flux occurred during
the two-hour sampling period between 1300 and 1500 hr, that no airborne loss
occurred before 0500 or after 2300 hr, and that the loss rate varied linearly
between the end points and the peak. Integration of the area under the tri-
angle formed gives a daily airborne loss of 9.0 times the peak hourly flux
measured between 1300 to 1500 hours. To estimate the loss occurring during
the afternoon of the first day (from 1300 to 2300) one-half the daily estimate,
or 4.5 times the peak flux, was used. Table III-7 shows the estimated daily
airborne loss of azinphosmethyl from the orchard for the first spray period
of the 1978 season. Loss is-shown both as an absolute loss in g/ha and also
as a percentage of the total dlslodgeable residues and soil residues present
on a given sample date. No dally loss estimate on a percentage basis was
made for the application date as absolute loss shown is the estimated loss
from 1300 to 2300 hr, rather than the entire day. These estimates suggest
a marked decline in the daily airborne loss over the 18-day sample period.
As a part of the parameterization of the model presented in Chapter II, first-
order rate constants were determined for the disappearance of azinphosmethyl
foliar dlslodgeable residues and soil residues in each canopy and alley layer
(accounting for movement). If these Individual rate constants are weighted
by the proportion of the total dlslodgeable residues and soil residues which
they attenuate, the average attenuation rate is 5.1%/day. The estimated air
loss rate on day 3, based on measurements made at the downwind edge (Table III-7,
Column 1), Is 1.86%/day. This loss rate represents approximately 36% of the
total attenuation. By day 6 air loss is 15%, day 13 it is 5%, and on day 18
air loss represents 8% of the total attenuation. These observations indicate
that initially airborne loss plays a major role in the overall attenuation
dlslodgeable residues, however, this contribution crops off rapidly between
3 and 6 days following application. The change in the rate of airborne loss
is believed to be a function of the state of residue deposits with the loosely
bound residues being rapidly lost to the atmosphere soon after application.
The remaining more tightly adsorbed and/or protected dlslodgeable residue
deposits volatilize at a slower rate. The effects of rainfall on the state
of the pesticide deposits within the orchard must also be considered (see
Chapter I). Certainly rainfall events that occurred during the first spray
period (see Figure III-A) had an effect on the pesticide airborne concentra-
tions and deposit residues measured. Again, loosely bound residues should
be most susceptible to rainfall effects redistributing these residues to
other parts of the orchard or as runoff leaving the orchard. These redistri-
buted residues may no longer be loosely bound, as they may be more evenly
distributed over a larger foliar surface than at application. By contrast,
76

-------
TABLE III-7. ESTIMATED AZINPHOSMETHYL DAILY AIRBORNE LOSS FROM THE ORCHARD, FIRST SPRAY PERIOD OF THE
1978 SEASON
Days After
Application
Mean Horizontal
Flux at 3.0 m During
Sampling2(lJ00-1500)
(yg m s )
Estimated Loss
During.Sampling
g ha hr
, -1
g ha
Estimated
da'1
Daily Loss**
96 day
¦1


Downwind
Edge




0

1
2
1
2
1
2
3
10.6
.697
.317
. 313°
1.43C


6
2.4
.158
.072
1.43
0.65
0.4
0.34
13
0.4
.026
.012
0.24
0.11
0.24
0.11
18
0.4
.026
.012
0.24
0.11
0.43
0.20


Center




0
11.5
.697
.317
3.13c
1.43°


3
5.7
.345
.157
3.11
1.41
1.51
0.68
6
1.2
.073
0.33
0.65
0.30
0.34
0.16
13
0.5
.030
.014
0.27
0.12
0.27
0.12
18
0.4
.024
.011
0.10
0.10
0.39
0.18
^ased on the estimated airbome7loss at peak flux (1300-1500 hr) on day 0, calculated using 2.2 x 10 mm Hg
at 20°C (column 1 and 1.0 x 10 mm Hg (column 2) for the V.P. of azinphosmethyl In equation (2). Diurnal
loss for the remaining sample dates was estimated by assuming this loss Is directly proportional to the
measured horizontal flux at 3m.
^Dally diurnal loss was estimated assuming peak flux during the 2 hr sampling period (1300-1500 hr), that no
airborne loss occurred before 0500 and after_J300, and that the loss rate varied linearly between the end
points and the peak. The daily loss in % da was determined by dividing the daily loss in grams by the
total g ha~ dislodgeable residues (Including soil residues) remaining.
c0ne-half day estimate.

-------
residues that were initially distributed to protected plant areas could be
redistributed with rainfall so as to increase their susceptibility to air-
borne loss. Rainfall may also enhance volatilization; if the soil moisture
level is low prior to a rainfall event, the Increased soil water may displace
adsorbed pesticide, increasing its effective vapor pressure at the soil
surface (Spencer et al., 1973). Heavy rains on unsaturated soils may also
move initially displaced pesticide to lower soil depths (Spencer and Claith,
1977; Helling et al., 1971).
In the model described in Chapter II, movement is separated from overall
attenuation only, and this attenuation process is assumed to be first order.
Modeling pesticide disappearance using first-order kinetics assumed that the
individual attenuation processes are first order and strictly additive, giving
an exponential decay of pesticide residues with time. As early as 1955,
Gunther and Blinn proposed that the first-order loss curve was an approxima-
tion of a bilinear or trilinear loss curve. This hypothesis was again pro-
posed by Gunther et al. (1969, 1977), by Hill (1971), and Van Dyk (1974, 1976).
Taylor et al. (1977) indicated that the disappearance of heptachlor and
dieldrin from orchard grass-and soil could be attributed solely to airborne
loss. They represented pesticide loss as a bilinear process with two regres-
sion equations; one for days 1 to 5 and another for days 5 to 107. Stamper
et al. (1979) proposed an alternative to first-order kinetics for foliar ap-
plied insecticides, showing for a number of foliar applied organophosphate
insecticides, that In concentration versus In time gave a linear relationship
with a better correlation coefficient than the linear relationship established
by plotting In concentration versus time (which indicates first-order kinetics).
The author suggests the reason the ln-ln plots give a better linear relation-
ship is that one form of the equation describing the fitted line is in agree-
ment with equations describing molecular diffusion from small volumes. Their
solution is interesting, but the use of In time would not be compatible with
the algorithms used in the model presented In Chapter II.
A third approach to modeling airborne loss was proposed by Phillips (1971).
He suggested that a double exponential equation
T = A "kt + B ~k,t	(III-3)
e	e
may apply when two volatilization processes are occurring simultaneously at
different rates. This model was later proposed by Fopendorf and Leffingwell
(1978) to explain the observed bilinear loss of parathlon dislodgeable residues
on citrus foliage. The approach used here is an extension of the double
exponential equation. At the present, residues deposited in the orchard are
classified as either dislodgeable, surface-penetrated, or soil. This class-
ification is based on analytical procedures for the recovery of these residues
as outlined in the experimental section. Dislodgeable residues are pictured
as residing on the foliar surface, whereas surface-penetrated residues are
embedded on or in the cuticular matrix. Further penetration, resulting in
tissue bound residues, Is also thought to occur (Wleneke and Steffens, 1974).
To model airborne loss, it was assumed that a portion of the dislodgeable
residues are loosely bound. The loss of these residues to the atmosphere
therefore being only slightly affected by the adsorptive effects of the leaf
surfaces. The remaining dislodgeable residues are more tightly bound and
volatilize more slowly. A similar theory for the decrease in the daily air-
borne loss rate of foliar applied pesticides is presented by Taylor (1978).
Volatilization of surface-penetrated and tissue-bound residues is thought
to be neglible and the kinetics of the disappearance of these residues is
78

-------
not treated here. Making these assumptions, pesticide airborne loss from
foliar surfaces Is represented In the differential form as:
^ = -aiD(t)-a2L(t)-a3T(t)	(III-4)
where t Is time In days, D Is dlslodgeable residues, L Is loosely bound re-
sidues, T Is total residues (dlslodgeable plus loosely bound residues, and
excluding penentrated residues), a. Is the air loss rate of the tightly bound
dlslodgeable residues, a^ Is the air loss rate of loosely bound residues,
and a. is the attenuation rate of the total residues (other than air loss
and assumed to be Independent of binding state). The constant a^ Is thought
to be primarily a function of degradation at the foliar surface and penetration
to subsurface tissues. As T = D + L, equation (III-4) can be reduced to:
fjl = _(ai + a3)D(t)-(a2 + a3)L(t)	(III-5)
and therefore:
^ = -(Sl + a3)D(t)	(III-6)
and DL
^7 = -(a2 + a3)L(t)	(III-7)
As loosely bound residues are what their name Implies, these residues are
more easily lost to wind erosion and volatilize at a rate governed only by
diffusion across the stagnant air layer. Consequently it is assumed that
airborne loss of these residues Is responsible for the observed early rapid
decay of total residues; this loss Is represented by Gunther et al. (1958)
and more recently by Popendorf and Lefflngwell (1978) as the first phase of
a bilinear loss pattern. Ulth the disappearance of these residues from the
foliar surface the rate of decline of remaining dlslodgeable residues Is
primarily responsible for the observed rate. The loss rate of these more
tightly bound residues 1s determined by the combination of airborne loss (a^),
degradation and penetration (a3) of the remaining dlslodgeable residues.
This represents the second phase of an observed bilinear loss pattern. As
current sampling or analytical techniques cannot distingusih between loosely
bound and more tightly adhered dlslodgeable residues, D, Land the rate con-
stants a^, a^ and a3 were determined indirectly. In determining the two rate
constants in the double exponential equation, Popendorf and Lefflngwell (1978)
used dlslodgeable residue data and a constrained optimization procedure that
compared predicted and observed values. As the hypothesized mechanism for
the observed bilinear loss of foliar applied pesticides presented here is
based on the assumption that early pesticide disappearance is primarily due
to airborne loss, the change in the estimated daily airborne loss rate was
used to estimate when the loosely bound residue deposits approached zero.
If the dally air loss rates (average of center and downwind edge estimated
in Table III-7, Column 1) are examined, it is noted that there is a sharp
decrease in the rates between day 3 (1.69%) and day 6 (0.54%). The average
rate and standard deviation for days 6, 13, and 18 is 0.40% + 0.18%. It was
therefore assumed that the loosely bound residues disappeared between 3 and
6 days following application and that the average daily air loss rate of 0.40%
represented the air loss rate (a^) of tightly bound dlslodgeable residues.
The total loss rate (a^ + a^) of the tightly bound disdodgeable residues was
79

-------
determined from a least squares linear regression of log g/ha total residues
versus time since applications, for days 3, 6, and 13 of the first spray
period. Total residues included foliar dislodgeable residues and soil re-
sidues. Soil was Included to give a conservative estimate with respect to
movement of residues present in the orchard following rainfall events. Day
18 was omitted from this analysis as it was felt that the 51.3 mm (2.1") of
rainfall the orchard received on day 16 redistributed the residues in such
a manner as to not allow a conservative estimate of residues present. Fjom
this analysis the daily loss rate (a^ + a.) was determined to be 7.5% (r =
.958).	rate is considerably faster than the dally attenuation rate of
4.2% (r = .985), determined from total residues for days four to 20 of the
second spray period of the 1978 season. Only one rainfall event occurred
during the second spray period (5.1 mm on day 10), whereas 17.3 mm of rain-
fall fell just prior to day 3, and 16.8 mm fell between days 3 and 13 of
the first spray period (see Table III-l and Figure III-2). This would suggest
that the disappearance of these residues was influenced by rainfall (although
other environmental parameters such as wind speed, solar radiation, humidity,
or some combination of the above also may have contributed). The results
may have been that the pesticide residues measured during the first spray
period were not conservative (i.e., movement of pesticide to unsampled soil
depths, pesticide carry-in with moisture uptake by litter and foliar surfaces,
Increasing the penetrated residue pool). Also, as stated earlier, rainfall
may have redistributed foliar residues as to allow increased volatilization.
An influx of moisture at the soil surface may have also caused an Increase
In volatilization. That this may be the case is suggested by the increase
in the estimated average dally air loss rate between day 13 (0.26%) and day
18 (0.41%) (Table III-7, Column 1). Until the effects of rainfall and other
climatic conditions on pesticide disappearance can be adequately different-
iated, of the daily attenuation rate (a^ + a~) to this set of environmental
conditions must be assumed. Given a^ to 0.40% and a^ + to be 7.5%, then
a^ is 7.1%. L and the amount ana rate of airborne loss of the loosely
bound residues, were estimated using an optimization routine similar to that
employed by Popendorf and Leffingwell (1978). Equation (III-3) was rearranged
to solve for T (total residues) at some sample date L:
T(kt) = D(ki_1) (l-(ax + a3))(ki"ki-l) +
L(k1-;L) (l-(a2 + a3))(ki"ki-l)	(III-8)
At T is known at sample dates k. = 0,3,6,13, and a^ and a^ have previously
been estimated, values for L ana a. were tried in equation III-8 to give a
minimum value for the test statistic:
Ti - (Li + V2
T±	(HI-9)
Using this statistic, It was estimated that 37.5% of azinphosmethyl dislodge-
able residues, Initially deposited in the orchard, were loosely bound. The
daily aij loss rate (a„) of the loosely bound fraction was estimated to be
90% day . This indicates a rapid loss of approximately one third of the
residues to the atmosphere, resulting in the disappearance of this fraction
in approximately 3 days. It is also suggested that loss of the remaining
dislodgeable residues is governed by a much slower air loss rate and atten-
80

-------
uation at the foliar surface. Figure III-5 shows the solution to equation
(III-5) as a line through the observed residue values (total dislodgeable
residues and soil residues) for the first spray period of the 1978 season.
The proportion of azinphosmethyl deposit residues estimated to be loosely
bound and the rate of loss of this fraction Is thought to be a function of
factors associated with application, environmental conditions, and orchard
characteristics. This is suggested by the difference in the decline of
total dislodgeable and soil residues between the first and second spray
periods as shown In Figures III-5 and III-6. Parameterization of equation
(111-4) using the residue data from the second spray resulted in an estimate
of 7.7% for the loosely bound fraction (L) with an estimated daily air loss
rate (a^) of 30% day for this fraction. The rate constant a^ + a. was
determined from linear regressslon of the total dislodgeable and soil residues
for days 4 to 20 to be 4.2% day (r ° .985). The air loss rate	of
the dislodgeable fraction (D) was_|ssumed to be the same as determined for
the first spray periods, 0.4% day . The lack of an initial period of rapid
loss during the second spray period may have been due solely to the lesser
amount of rainfall received, as compared to the first spray period. That
rainfall may affect airborne loss has been discussed earlier. In addition,
rewetting of the residue deposit may enhance penetration and plant uptake
(Hull, 1970; Bukovac, 1976), resulting in an increase in the observed loss
rate of dislodgeable residues. The effects of other environmental parameters
(i.e., solar radiation, ambient temperature, relative humidity, wind speed
and atmospheric stability) should also be considered.
To better understand the Influence of those factors associated with the
formulated pesticide, its application, and environmental conditions on the
attenuation of residue deposits in orchards, a more thorough sampling pro-
gram of both airborne and deposit residue is required. The pesticide should
be applied and its disappearance monitored under a variety of properly char-
acterized "natural conditions." The multi-component kinetic model presented
here provides a tool for investigating those parameters that influence
pesticide airborne loss and the overall attenuation of deposit residues.
81

-------
Figure 111-6. Decline In Total Azlnphosmethyl Dlslodgeable and
Soil Residues Measured During the Second Spray
Period of the 1978 Season
*
500 1- 5.6
400<
RAINFALL (mm)
+
5.1
300
200
g/ha
100
12 15 18 21
DAYS
82

-------
Figure II1-5. Decline in Total Azinphosmethyl Dislodgeable and
Soil Residues fteasured During the First Spray
Period of the 1978 Season
500i
17.3
RAINFALL (mm)
2.6* 3.0
11.2
u
51.3
400
300
200
g/ha
100
12 15 18 21
DAYS
83

-------
CHAPTER IV
GENERALIZATION OF THE PESTICIDE FATE MODEL
INTRODUCTION
It would be unrealistic to believe that a model of pesticide fate in
any terrestrial ecosystem, driven by time and rainfall conditions only,
would be capable of explaining all the variation observed in the field data.
To better explain the variation observed, both movement and attenuation
should be decomposed to represent their component physical, chemical, and
biological processes, each a function of the relevant environmental para-
meters (i.e., solar radiation, temperature, humidity, wind speed, atmospheric
stability, etc.).
The conceptual model shown in Figure II-l of Chapter II was developed
to represent those processes which determine the distribution and fate of
a pesticide applied to an orchard ecosystem. The initial mathematical form
of the model describes azlnphosmethyl fate in terms of attenuation and move-
ment, based only on field data from three seasons. The attenuation rate
constants determined represent the combined effects of a number of individual
processes as indicated in the conceptual model. These processes are indivi-
dually influenced by different combinations of environmental parameters.
In addition, because these processes occur simultaneously, the use of field
experiments to study their individual contribution to the overall attenuation
rate is difficult.
Laboratory experiments, designed to Isolate and study Individual pro-
cesses under controlled environmental conditions, are at present the preferred
method for studying environmental fate. Such experiments, while giving the
researdher much greater control of the variables involved, often do not give
results that can be extrapolated to the real world. However, much progress
has been made recently in determining rate constants for a number of attenua-
tion processes under various sets of environmental conditions (Zepp et al.,
1975; Smith et al., 1977; Freed et al., 1979).
Laboratory experiments should be designed to isolate and examine the
influence of selected environmental parameters on the individual processes
that in combination result in the observed field attenuation and movement
rates. In addition, in situ field rate data should be collected under a
variety of "natural conditons" to allow the partial isolation of individual
environmental effects. Further model development and experimental design
should address the following questions:
(1)	How are various processes influenced by individual environmental
parameters and compound properties, and
(2)	What is the contribution of each process to the overall attenuation
and/or movement rates under a given set of environmental conditions.
The purpose of the present study was to refine and develop further the
model described in Chapter II, using in situ field measurements, data from
laboratory experiments, and results of other laboratory experiments reported
in the literature, to describe the influence of selected environmental para-
meters and compound properties on pesticide attenuation and movement.
The mathematical form used to represent each process in the model is
described below, together with its default parameters for azlnphosmethyl in
our experimental orchard. The user can alter these values to describe other
compounds or situations. The present model can be greatly improved in both
form and parameter values as the results of future field and laboratory
studies become available.
84

-------
The Attenuation Submodel
In Chapter II a matrix A of daily attenuation rates of azinphosmethyl
in various orchard layers was developed. The orchard is treated as four
vertical layers (tree, grass-broadleaves, litter-moss, and soil) by two hori-
zontal areas (canopy, alley). Seven regions (no alley tree region exists)
appear in each of the matrices and vectors below, beginning with canopy trees
and proceeding to alley soil. Each diagonal element A^ represents the frac-
tion of azinphosmethyl lost from region i in one day, excluding any movement
to or from another region (which is accounted for in matrices P, L, and H,
the nonrain movement, light rain movement, and heavy rain movement matrices,
respectively).
To investigate the influence of rainfall and other environmental para-
meters of pesticide loss, the attenuation matrix A may be decomposed into
the diagonal matrices:
A = A + A + A + A + A
p c v m u
where the summand matrices A , A , A , A , and A represent photolysis,
chemical degradation, volati?iza£iony microbial degradation, and plant up-
take, respectively, and are called the attenuation component matrices. Each
of these matrices is a 7 x 7 diagonal matrix with one non-zero entry for
each region. Each attenuation component matrix is decomposable into a dia-
gonal matrix of constants (the rates under "standard conditions") and a set
of functions which modify those rates based on environmental conditions.
The "standard" rates for each process should sum to the field-determined
values. To examine the feasibility of this approach to further model devel-
opment, the relationship between azinphosmethyl degradation kinetics and a
number of environmental parameters was determined from laboratory data and
the results of other laboratory experiments reported in the literature. The
volatilization kinetics of azinphosmethyl were discussed in Chapter III.
Pesticide degradative mechanisms have been reviewed by a number of re-
searchers (Ebling, 1963; Crosby, 1973; Leonard et al., 1976). Pesticide
degradation has been traditionally divided into three major areas:
Chemical, photochemical, and biological. Pesticides may undergo a number
of chemical transformations in the environment to Include: hydrolysis and
other nucleophilic reactions, oxidation, isomerization, reduction and free
radical reactions (Goring et al., 1975). For the organoph'osphates, including
azinphosmethyl, hydrolysis and oxidation are thought to be the most commonly
occurring. Biological degradation is a major pathway for the disappearance
of many pesticides in the soil (Kearny and Helling, 1969). The diverse micro-
bial populations of most soils are capable of degrading pesticides with
little difficulty, either by adaption, or more commonly, by co-metabolism
(Matsumura, 1975). Microbial degradation on plant surfaces must not be ex-
cluded in assessing possible causes of pesticide disappearance (Steffens and
Wieneke, 1975). Photochemical degradation of pesticides has been demonstra-
ted in water and air and on soil and foliar surfaces (Crosby, 1969; Nilles
and Zablk, 1975; Liang and Lichtenstein, 1976; Zepp and Cline, 1977).
Chemical Degradation
As with the overall attenuation rate, the rates for the individual pro-
cesses are assumed to be first-order or pseudo first-order. Chemical de-
gradation of azinphosmethyl is assumed to occur primarily by hydrolysis,
85

-------
but oxidation is also possible (Eto, 1974). Oxidation may occur in all
regions. The oxidation rate is represented by
nr
— (oxidation) = . c R = K C
Dt	def ox ox
The most likely oxidation pathway is through the reaction with free radicals,
assuming an excess of free radicals available for interaction (regeneration
> K ) then K is rate limiting and the reaction is pseudo first-order
(Smilfh et al.°x1977). Photooxidation, as a result of reaction with photo-
chemically formed free radicals, the oxygen triplet diradical, or the more
reactive singlet oxygen, may be more responsible for many pesticide non-
biological oxidations (Crosby, 1973; Khan, 1976). Soil free radicals may also
be Important in the oxidation of pesticides in this medium (Plimmer et al.,
1967; Armstrong and Konrad, 1974). Spear et al. (1978) indicated that para-
oxon production may be related to both ozone and dust levels on citrus in
central California. Spencer et al. (1975) also noted paraoxon formation on
dust and dry soil beneath citrus trees in southern California citrus foliage
has also been indicated (Gunther et al., 1977). However, azinphosmethyl-
oxon levels never exceeded 1.0% of the azinphosmethyl present. The oxon
formed was more stable, but dissipated rapidly following rainfall. The like-
lihood that oxidation would contribute significantly to the degradation of
azinphosmethyl during the relatively wet and humid summer months normally
experienced in the temperate eastern United States is doubtful.
Hydrolysis has been shown to be an important mechanism of organophosphate
degradation in both soil and aqueous environments (Freed et al., 1979).
Hydrolysis resulting from reaction with moisture on foliar surfaces must
also be considered. A discussion of hydrolytic mechanisms for the organo-
phosphates in water can be found in Faust and Gomaa (1972) and Smith et al.
(1977). The reaction is a function of pH, and can be either neutral, acid,
or base-catalyzed. Again, the reacting species 0^0, H , OH) were assumed
to be in excess of the pesticide, and the reaction first order, at a given
pH.
Azinphosmethyl aqueous hydrolysis as a function of pH was determined
by the procedure of Freed et al. (1979). Buffers used were as follows:
pH 1.0, 0.0 m KC1 and 0.01 m HC1; pH 3.0 and 5.0, 0.01 m potassium hydrogen
phthalate 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. Azinphosmethyl concentration was determined by analysis of residual
parent compound by GLC (see Chapter I). First-order rate constants at 25°C
were determined from linear regression of log concentrations versus time,
over a 20-day incubation^period. Hydrolysis rate constants at pH 1, 3, 5,
and 7 averaged 0.8% day (standard deviation, 0.1%). The observed stability
under neutral and acid conditions is in agreement with the observations of
Liang and Lichtensteln (1972). Faust and Gomaa (1972) report that many
organophosphates are stable under acid conditions. (Azinphosmethyl hydrolysis
as a function of pH is shown in Figure IV-1). The relationship between pH
and the base-catalyzed reaction rate (pH 7.5-9.0) is represented by the
linear regression equation:
= 0.095 pH - .713 (r2 = .996)	(IV-1)
where K^ is the aqueous hydrolysis rate constant (time ^).
The rate of hydrolysis as a function of temperature was represented
using the Arrhenius equation:
= Ae~Ea/RT	(IV_2)
86

-------
.16
.14
.12
.10
RATE
CONSTANT
DAY-1 -08
.06
.04
.02
J	L
J	I	L
4 5 6
pH
Figure IV-1. Azinphosmethyl Hydrolysis Rate Constant Versus pH
87

-------
where is the aqueous hydrolysis rate constant (time ) at temperature T,
A is a constant depending on the chemical and other non-thermal factors, R
is the gas constant, and Ea is the energy of activation. If 1C is deter-
mined over some temperature range, Ea can be calculated from the slope of
the line, for a plot of log K, versus 1/T. To estimate temperature effects
on azinphosmethyl hydrolysis the data of Liang and Lichtenstein (1972) was
analyzed using equation IV-2 as shown in Figure IV-2. The value determined
for Ea, over the temperature range 5-50°C, was 12.5 Kcal/mole. This is an
approximate value and will vary with pH and the temperature range used.
Buffer composition and strength may also influence hydrolysis rates and
associated values for Ea (Smith et al., 1977).
Hydrolysis of azinphosmethyl is assumed to be important in all regions.
Although little is known about reactions occurring on foliar surfaces, it
is assumed that, due to transpiration, there is an environment with suffici-
ent moisture to allow hydrolysis to occur (Wieneke and Steffens, 1974). A
neutral pH is assumed in all but the soil regions. Hydrolysis is treated as
a function of air temperature only. This again is an approximation as tem-
perature. In the soil regions hydrolysis is treated as a function of both
soil temperature and pH. Properties of the soil which influence hydrolysis
are discussed by Freed et al. (1979). Adsorption to clay and organic matter
is thought to play an important role in degradation, as adsorbed organo-
phosphates may be protected from hydrolysis or, in some cases, may result
in Increased reaction rates due to surface catalysis (Crosby, 1970). The
use of the Arrhenius equation (IV-2) to represent the effects of temperature
on hydrolysis and other non-biological degradative mechanisms in the soil
environment is discussed in Hamaker (1972). This author suggests that the
heterogenous nature of soils as a reaction medium may not allow the use of
equation (IV-2) as it has traditionally been applied to reactions in homo-
genous solutions. This is indicated by the fact that the distribution co-
efficient, Kd (pesticide adsorbed/pesticide in solution), may change as the
pesticide is transformed, and also by the exothermic nature of the adsorption
process resulting in an equilibrium shift towards sorption to organic matter
with increasing temperature (Felsot and Dahm, 1979). Because both these
phenomena will influence the concentration of pesticide in the soil solution
and, theoretically, the rate of hydrolysis, this suggests that the "A" term
in equation (IV-2) is not a constant, but a function of both pesticide con-
centration and soil temperature. Until the influence of soil properties on
hydrolysis is better understood, and for the purposes of this study, "A" is
assumed to be constant. The data of Yaron et al. (1974) were used to de-
termine the relationship betwen temperature and the azinphosmethyl hydrolysis
rate constant in soil. Degradation rates for wet soil (50% of saturation)
incubated at 6, 25, and 40°C, were used to parameterize equation (IV-2)
as shown in Figure IV-3. The Ea determined from this data was 13.5 Kcal/mole.
The soil used in this study was a silty loam, pH 8.4 and <1% organic matter.
Both pH and soil type should be considered when comparing rates determined
from the data in this particular study with other research. Soil in the
orchard used in the present study was a marlette sandy clay loam, pH 6.0,
56% sand, 20% silt, 24% clay, and 6.0% organic matter in the top 10 cm. The
results of degradation studies in the laboratory, using soil from the orchard
(moisture content 30% = 48% of saturation) sterilized with sodium azide,
fortified with 10 ppm azinphosmethyl, and incubated at 25 + 1°C for 20 days,
showed no degradation. This is not surprising considering the acid pH and
the clay and organic matter content of the soil.
88

-------
0.2 i—
J	I	I	I	I	I	I	I	I
60
TEMP
Figure IV-2. Azinphosmethyl Hydrolysis Rate Constant Versus Air Temperature
89

-------
0.2
DEG
RATE
day"1
0
50
TEMP
Figure IV-3. Azinphosmethyl Hydrolysis Rate Constant Versus Soil Temperature
90

-------
0.1
0
10
% O.M.
Figure IV-4. Azinphosmethyl Microbial Degradation Versus Soil Percent
Organic Matter
91

-------
Microbial Degradation
Although microbial degradation cannot be ruled out as contributing to
the overall attenuation of residues deposited on foliar surfaces, microbial
degradation is thought to be a major degradative pathway in the litter-moss
and soil layers. In some instances, pesticides can be used as the sole food
source of microorganisms, but more often they are co-metabolized with other
organics. If the pesticide is used as a primary nutrient by the microorgan-
isms of the soil, then a lag period may be observed following application
while the soil microorganisms population adapts to the new food source. This
period may become shorter with successive applications (Hamaker, 1972). With
co-metabolism, no lag period should occur. This Is the situation assumed to
be present in the orchard soil and litter, with regard to azinphosmethyl
blodegradatlon. The rate of biological degradation will therefore vary with
the total available food source (i.e., organic matter content of the soil),
temperature, and a moisture (Hamaker, 1972). To determine the microbial
degradation rate, a 20-day incubation at 25 + 1°C, of the non-sterilized
orchard soil (moisture content 30% = 48% of saturation) fortified with 10
ppm azinphosmethyl, was performed. The first-order rate constant, determined
from linear regression of concentration of azinphosmethyl remaining versus
time, was 7.9% day (r = .969).
The influence of organic matter content of the soil on microbial de-
gradation of azinphosmethyl was based on the data of Iwata et al. (1975).
Soils in this study were passed through a 100 mesh sieve and moisture was
added to 40% saturation. The soils were fortified at 450 ppm azinphosmethyl
and incubated at 30eC. Characteristics of these soils and the corresponding
azinphosmethyl degradation rates over a 20-day period are as follows.
cl
TABLE IV-1. Azinphosmethyl Degradation when Incubated with Various Soils .
X Organic Mechanical Analysis, %
Soil Matter	Sand Silt Clay pH K ^	r^
1
0.8
53.6
31.0
15.4
6.9
-.010
.992
2
1.8
56.0
33.0
11.0
7.6
-.033
.996
3
2.1
12.5
50.7
36.8
7.3
-.050
.991
4
2.3
22.4
34.5
43.1
7.3
-.061 *
.969
5
6.0a
56.0
20.0
24.0
6.0
-.079
.969
Data for soils 1-4 reported in Iwata et al. (1975).
Soil was taken from the orchard used in the present study.
^Proportional daily loss determined from the data of Iwata et al. (1975)
(Soils 1-4) and the orchard soil used in the present study (soil 5).
Figure IV-4 shows a plot of percent organic matter versus the azinphos-
methyl degradation rate. Degradation is assumed to be primarily microbial,
due to the neutral or acidic pH values for these soils. However, hydrolysis
and other forms of chemical degradation may also have contributed to the
observed rates derived from the data of Iwata et al. (1975). A modified
form of the Verhulst-Pearl logistic equation (Pielou, 1969) was fit to the
data to give the following relationship:
Km = .079 [1 + e_1*85(Z 0,n1, " i'87)]-1	(IV-3)
92

-------
The correlation of azinphosmethyl degradation with a single soil factor
(organic matter) must be interpreted with some caution, as the organic matter
content may be correlated to other soil conditions effecting degradation.
High organic matter is usually accompanied by a low soil pH, and adsorption
has been shown to be positively related to organic matter content (Saltzman
et al., 1972). Even though the data present here shows a positive correlation
of degradation rate to organic matter content, a number of researchers have
found that at very high organic matter contents of peat and muck soils,
degradation is decreased, presumably due to adsorption (Beynon et al., 1966;
Hamaker, 1972; Kaufman, 1964). In addition, the relationship between the
data of Iwata et al. (1975) and that of the present study must be viewed in
light of the differences in initial concentrations used. Hamaker (1972) cites
a number of studies whifch indicate that the degradation rate, on a percentage
basis, increases with decreasing concentration. The azinphosmethyl microbial
degradation rate of 7.9% day determined using an initial concentration 10
ppm, may have been lower if the Initial concentration of 450 ppm, employed
by Iwata et al., (1975) was used. Temperature differences between the two
studies are not thought to be significant, as indicated by the temperature
relationship shown in Figure IV-5.
The relationship between temperature and soil microbial degradation was
determined from the data of Yaron et al. (1974). The difference in degrada-
tion rates between sterile and non-sterile wet soil (50% of saturation, in-
cubated at 6°, 25°, and 40°C, was used to determine this relationship. The
data was fit to the following polynomial equation:
K (t) = .00298T - .00005T2 - .013	(IV-4)
m
where K is the microbial degradation rate at a given temperature, T. This
fit shows the optimum temperature to be approximately 30°C. This is in
agreement with the data of Day et al. (1961) for degradation of amitrole in
soil. This temperature relationship is thought to reflect a response to
temperature by the microorganism population. Changes in population size dis-
tribution, or adaption to degradation of the chemical may all be responsible
for the observed temperature effect. The response may also be due in part
to an Arrhenius type mechanism. At temperatures above the optimum (30°C)
the rate no longer shows a positive relationship to temperature, possibly
due to the heat lability of the microorganisms.
Soil moisture has also been shown to be an important factor in soil de-
gradation of pesticides. The data assembled by Hamaker (1972) show much slower
rates in dry as compared to moist soils and that rates tend to level off at
higher moisture levels (>30% of saturation). At saturation levels rates
often drop due to the lack of oxygen necessary for aerobic degradation. Many
pesticides are capable of being degraded anaerobically, so this possible
mechanism must not be neglected. Soil moisture in the orchard may remain at
saturation levels only briefly. Under these conditions the development of
an anaerobic microorganism population able to significantly alter pesticide
concentration is not likely.
The data of Yaron et al. (1974) shows little or no degradation of
azinphosmethyl for both sterile and non-sterile dry soils. Increased adsorp-
tion, due to the dry conditions, may be partially responsible for the observed
results. At 50% of saturation, the degradation rate for the non-sterile soil
at 25®C was twice that of the sterile soil. No other data was found in the
literature on the relationship between azinphosmethyl degradation and soil
moisture. As microbial degradation has been shown to be the primary pathway
93

-------
0.5
DEG
RATE
day-"!
0
50
TEMP
Figure IV-5. Azinphosmethyl Microbial Degradation Versus Soil Temperature
94

-------
of amitrole loss in soils (Kearny and Helling, 1969), the treatment by Hamaker
(1972) of the data of Day et al. (1961) for amitrole is presented as a crude
estimate of the influence of soil moisture on azinphosmethyl microbial de-
gradation (Figure IV-6).
The effect of both soil moisture content and temperature on the degrada-
tion of a number of herbicides was examined by Walker (1974, 1976a, 1976b,
1976c), Smith and Walker (1977), and Walker and Smith (1979). In these
studies no distinction was made between chemical and biological degradative
pathways. In all cases a positive correlation was observed betwen the de-
gradation rate and both soil moisture content and soil temperature. These
data were used to develop and test a model for herbicide persistence. Assum-
ing first-order kinetics, herbicide half-life as a function of soil moisture
was represented by the following empirically derived equation (Walker, 1974):
H = am k	(IV-5)
where H is the herbicide half-life, m is the soil moisture content, and a
and b are constants. The Arrhenius equation was used to represent the effect
of temperature on the degradation rate. Herbicide persistence in the field
was estimated from these two equations using both laboratory determined
values and simulated seasonal and diurnal soil temperature and moisture flux
(Walker, 1974; Walker and Barnes, 1981). The model was tested against
herbicide loss when applied to bare soil, and the model worked best when the
herbicide was incorporated. Applicability to cropped fields and particularly
perennial crops such as orchards lies in the ability of the model to represent
the effects of the crop and/or ground cover on soil temperature and moisture
flux. The model in its present state does not have this capability. Further
development to represent the effect of soil moisture and temperature on
chemical and biological degradative pathways individually would also be de-
sirable.
pH and soil texture may also be important factors influencing microbial
degradation. Soils with extremes in pH will most likely have developed micro-
bial populations adapted to these conditions. The influence of pH might be
more important, however, if a soil amendment which drastically alters the
soil pH is used, thereby requiring adaptation by the microbial community. Soil
texture describes the soil aggregate size and the pore space of the soil,
which may influence the availability of moisture and air to the microbial
population. The interrelationship between all soil properties and the micro-
organism population must be considered in assessing pesticide degradation.
Photodegradation
Solar radiation is known to be important to the attenuation of pesti-
cides in the environment, as it supplies thermal energy which influences
pesticide volatilization and the rate of many degradative reactions. The energy
of the photons may also be adsorbed by the bond (electronic) energy of the
molecule, which may result in transformation. Much research has been done
in the laboratory on the photolysis of pesticides and the photoproducts formed
(Zabik et al., 1976). Only recently have attempts been made to determine
rates of photodegradation under natural conditions. The most promising approach
uses computer modeling techniques to extrapolate laboratory findings to field
conditions (Zepp and Cline, 1977). These authors have attempted to predict
photolysis of a number of compounds in the aquatic environment. The study
of photolysis in solution is preferred as the homogeneity of this medium
allows the researcher to accurately monitor and vary the chemical environment.
95

-------
100
60
% MAX.
RATE
60
40
20
• Vista Sandy Loam
OChino Silt Loam
1
J
10
20
30
40
50
60
70
% FIELD CAPACITY
80
90 100
Figure IV-6. Amltrole Degradation Versus Soil Moisture for Two Soil Types
96

-------
Experiments to determine the photochemistry of compounds in the solid or
sorbed state, such as pesticides applied to the soil and plant surfaces, are
far less manageable (Zabik and Ruzo, 1981). Relatively few studies have
attempted to determine the rate of photolysis on plant and soil surfaces under
natural conditions. One of the major difficulties associated with this kind
of study is differentiation between photodegradation, volatilization, and
metabolism. In addition, the presence of sensitizers and quenchers may great-
ly alter the rate of photolysis determined in the absence of these substances.
For Example, Liang and Lichtenstein (19|£) reported that following an eight-
hour exposure to sunlight, 2.8% of the C azinphosmethyl applied to bean
leaves was determined to be the oxygen analog, whereas no oxygen analog was
found for applications to corn leaves or glass plates. No oxygen analog
was present in the dark controls. The author suggest that a component of the
bean leaf may have enhanced the formation of azinphosmethyl oxon in the pre-
sence of sunlight. Ninety-three, 88, and 94% of the radiocarbon was recovered
from the glass, corn, and bean leaf dark controls. In all cases, the radio-
carbon recovered was determined to be unaltered azinphosmethyl. Sixty-six,
86, 72% was recovered from the glass, corn, and bean leaves following the
eight-hour sunlight exposure with 5.7, 8.9, and 4.3% of that recovered de-
termined to be photoproducts, based on differential extraction and thin layer
chromatography. Substantial loss to volatilization of azinphosmethyl on glass
and foliar surfaces as a result of the sunlight exposure is indicated. The
greater volatilization of the azinphosmethyl on exposed surfaces, as compared
to the control (covered with black cloth), may have been due in part to
higher temperature and greater air exchange at the foliar surface. Volatili-
zation of photoproducts must also be considered. It it is assumed that only
a small amount of the photoproducts are lost to volatilization, then a crude
estimate of the initial rate of photolysis can be made from the percent of
photoproducts formed.
As the eight-hour exposure period (0900 to 1700 hr) used by Liang and
Lichtenstein (1976) represents only a portion of the daily solar radiation
exposure during the summer months in North America, the relationship between
the photolysis rate of carbaryl and time of day (July) reported by Zepp and
Cline (1977) was used to estimate the initial daily loss rate. Based on the
relationship reported by these authors it was assumed that the peak rate of
photolysis occurred at approximately 1300 hr, that no photolysis occurred
before 1600 hr or after 2000, and that the loss rate varied linearly between
the end points and the peak. Integrating the area under the triangle formed
yields an estimate that exposure to sunlight between 9099 and 1700 hr is
responsible for approximately 82% of the daily loss due to photolysis, during
the summer months. The amount of photoproducts formed (estimates of azinphos-
methyl loss due to photolysis during the eight-hour exposure), as a percent
of dose applied to each surface, was adjusted accordingly to arrive at an
estimate of azinphosmethyl Initial daily loss due to photolysis. The Initial
daily loss rates calculated were 6.9, 10.8, and 5.2% day for azinphosmethyl
on glass, corn, and bean leaves, respectively. The azinphosmethyl degradation
in Chapter II was approximately day . xne initial daily losses due to
photolysis, as determined above, would indicate that photolysis may contribute
significantly to the disappearance of azinphosmethyl foliar deposits.
The data of Liang and Lichtenstein (1976) also indicated that azinphos-
methyl on the soil surface undergoes photodegradation. Estimated daily
initial rates of photodegradation for the three soil types used (sand, loam,
muck) were slightly faster than those estimated for the glass and foliar
rate for dlslodgeable
orchard grass determined
97

-------
surfaces.
Factors that must be considered when estimating direct photolysis under
natural conditions are discussed by Zepp and Cline (1977) and Smith et al.
(1977). These include: the incident light intensity, as a function of
season, latitude, time of day, cloud cover, percent of light adsorbed by
the pesticide (as compared to its surroundings) and qunatum yield (fraction
of photons adsorbed that results in transformation). As many pesticides
show maximum adsorbance in the ultra violet region, outside the range of
wavelengths reaching the surface of the earth, sensitized reactions are often
important to pesticide photodegradation. A sensitizer adsorbs, light at a
wavelength present, followed by an energy transfer to the pesticide (Khan,
et al., 1974). The components of a plant surface which may act as
sensitizers are largely unknown.
Rate of Photolysis in the Field
Rates of photolysis on foliar surfaces and soil, as a function of com-
pound properties and environmental parameters, are estimated in the predictive
model using the approach for predicting direct photolysis in natural waters
proposed by Zepp and Cline (1977). They report that for those situations
where the substrate concentration is low and the substrate plus this matrix
adsorb less than 5% of the incident light the rate of direct photolysis can
be represented by the following equation:
^ = Ka = Quantum yield (m/Glnstein or molecule/photon)
s = Substrate concentration in m/1
K is determined by summing the product	f°r the ads^rptjon range
of the substrate. I, is the light intensity (photons ciji s^ ) and wave-
length X and is tne molar extinction coefficient (m cm ) of the substrate
at wavelenth X. J is a conversion factor making the units of 1^ and
compatible (see Zepp and Cline, 1977, for a more detailed discussion).
In natural systems where the substrate concentration is low, K is
independent of substrate concentration and the daily photolysis rate can
be estimated from the following equation.
K = K •  • A • D	(IV-7)
P a
The A term takes into account the effects of adsorbers or quenchers
which may produce the observed rate of photolysis in natural systems and
sensitized reactions. At the present this term is set at unity. D is a
factor that converts the instantaneous photolysis rate at midday to a daily
rate, based on the diurna^ variations in solar radiation intensity. D is
estimated to be 2.52 x 10 for the summer months at a latitude of 35°N
(based on the data of Lelghton, 1961). This value for D is used when the
midday rate is represented as SEC- .
To estimate K for azinphosmethyl the data for parathion presented
by Smith et al. (1§77) and the I^(co^) data givej by Zepp and Cline (1977)
were used. This yields an estimate for K = 10 . In the predictive model
it is assumed that only a portion of the tally overall attenuation rate is
98

-------
due to photolysis. For azinphosmethyl applied to the tree leaves, the daily
photolysis rate is arbitrarily set at .018 or 37% of the overall daily at-
tenuation rate. Substituting this value into ^quation (2), the remaining un-
known term, the quantum yield , must be = 10 . It must be understood
that Eq. 2 is a simplified form representing a very complex phenomenon. It
is hoped that the crude estimates used in this equation are in error by no
more than two orders of magnitude. Nevertheless, understanding its limita-
tions, equation (2) allows the user to vary the rate of photolysis based on
both environmental parameters (1^, A, 0) and compound properties (<(i,e^).
Volatilization
In the present form of the predictive model the rate of pesticide air-
borne loss is fixed and may not vary with environmental conditions. This
is certainly not the case in the environment as it has been suggested that
windspeed, rainfall, and solar radiation Influence the rate of volatilization
and erosion of pesticide deposits (see Chapter III). Too little is known
about the influence of these parameters to propose even the crudest of
equations. Proper representation of airborne loss is further complicated as
the rate of loss is far from being first order, for as the surface
deposits are depleted the airloss rate decreases rapidly. A more detailed
discussion of this phenomenon may be found in Chapter III. Airborne loss
is presently represented as an average rate over the sampling period.
This badly underestimates the initial loss rate within the first few
days after application, and overestimates the rate towards the end of
the sampling period. The volatization rate for the tree leaves is 40%
of the overall attenuation rate. The user is allowed to vary the volatiliza-
tion rate based on compound properties using Hartley's equation:
FB ¦ PB("B)11^ FA	(IV-8,
PA(mA)
where: FB = volatilization rate of new compound 'B';
PB = vapor pressure of compound B;
mB = molecular weight of compound B;
FA = volatilization rate of reference compound 'A';
PA = vapor pressure of compound A;
mA = molecular weight of compound A.
For best results the vapor pressure of the new compound "B" should as close
as possible to the reference compound A. This will give a rough estimate
of the mean rate over A the sampling period. Unfortunately the rate at
which the new compound dissipates due to other loss processes may have a pro-
found effect on the actual field airloss rate.
Plant Uptake
14
Greenhouse studies on plant uptake and metabolism of C azinphosmethyl
applied to bean leaves (Steffens and Wleneke, 1976) suggest that this
loss mechanism may be an Important pathway in the attenuation of foliar
deposit residues in an apple orchard. The influence of environmental factors
such as solar radiation, temperature, humidity and other conditions influenc-
ing leaf wetness on plant uptake of pesticides is discussed by Hull (1970),
Steffens and Wieneke (1975) and Bukovac (1976). However, no quantitative
99

-------
relationships are presented. A computer model for foliar uptake of pesticides
was reported by Bridges and Farrington (1974). This model represents plant
uptake by both diffusion and mass flow through the observable structure of a
wheat leaf. The model structure does not allow for the influence of environ-
mental conditions on the rate of foliar uptake. A general lack of data in
the literature points to the need for further research to investigate the con-
tribution of plant uptake and metabolism to the overall attenuation of foliar
residues as a function of environmental conditions. Plant uptake is not re-
presented as a separate attenuation process in the current predictive model,
due to the scarcity of data.
The Pesticide Movement Matrices
In Chapter II, a set of matrices was derived to describe the movement
of azinphosmethyl among various orchard regions. Matrix P described movement
during 24 hours of dry conditions, L described additional movement due to
light rain at some time during the 24 hours, and H, movement due to heavier
rain. All of these matrices were parameterized using field data collected
during 3 seasons of azinphosmethyl spray applications. However, the proportion
of the residue in one layer moving to any other layer obviously depends on
many factors, including age of residue, formulation, method of application,
properties of the active ingredient, material on which the residue is deposited,
and environmental conditions. Much laboratory work will be required in order
to learn the influences of all of these factors, although they have been
studied in many laboratories. However, the mathematical technique by which
the influences of these factors, as they are learned, may be incorporated
into the model framework presented here is described below.
The movement matrices P, L, and H are each lower-triangular and column-
stochastic. An entry H ., for example, is the proportion of pesticide that
remains in region i during one day (a first-order loss under fixed conditions
is assumed). H , j>i, represents the proportion of pesticide that enters
(and remains in; region j from region i during one day. If laboratory studies
indicate that a different formulation,for example, washes off region 1 at a
rate r times faster than the currently modeled formulation under heavy rain-
fall, an adjustment to the ith column of the H matrix can be done based on
the assumption that the continuous process underlying the daily loss propor-
tion is:
dCi
= -k1Ci(k) t>0,	(IV-9)
where C^(t) = pestlcj.de residue in (on) region i at time t, and k^ = instanta-
neous loss rate (da ) of pesticide from region i to the lowjr regions.
Then, converting H.. into an instantaneous loss rate, k.(da ),
-k
Hii o e i, or kt = -lne(H1±)	(IV-10)
Thus we can write
ki = rki' where is the new loss rate- Then	(IV-11)
Hj^ = (H^^)r, where H' is the new movement matrix.	(IV-12)
100

-------
Remaining entries in column i of H' can be calculated if the pesticide
moving out of region i is assumed to be distributed among the lower layers
in the same relative proportions as it was in H:
(IV-13)
Specification of the Attenuation Parameters
The field-determined attenuation matrix A is based upon field conditions
encountered during three seasons, and represents average total daily attenua-
tion for each layer under those conditions. These daily attenuations, as
developed in Chapter II, are presented in Table IV-2, re-arranged from a
diagonal matrix into two rows (canopy and alley). The user of the model can
change these entries at will. It is possible to further decompose these
losses according to the individual chemical and biological processes which
were described in the preceding sections. Table IV-3 shows the attenuations
separated by process, where each row represents a process. Thus, the entries
in each row correspond to the diagonal entries of one of the attenuation
component matrices A , A^, A , A^, and Aq. (Surface penetration and plant
uptake are considered separately.) Table IV-3 is used only as a starting
point for calculating daily losses in the model. If the user alters the
overall attenuation matrix A (Table IV-2), entries in Table IV-3 directly,
the entries are simply summed to produce a new Table IV-2.
Each process in Table IV-3 is affected by various environmental and
compound-specific properties. The model contains factors which can increase
or decrease the rates shown in Table IV-3 on a daily basis, depending on
environmental conditions (temperature, pH, soil moisture, and soil percent
organic matter). These factors are based upon the equations described earlier
in this chapter. The user can also modify any of the coefficients used in
these equations to incorporate additional information or represent other
compounds or conditions. Changes in these multiplicative factors affect the
daily rates of attenuation, but are not shown in Table IV-3 which represents
only rates under standard conditions. The standard conditions for Table IV-3
are: Air temperature, 23°C; soil temperature, 23°C; soil pH, 7; soil organic
matter, 6%; soil moisture, W/W or 50% field capacity; azinphosmethyl as
50% w.p.; vapor pressure, 10 mm Hg; molecular weight, 317.
Values presented in Table IV-3 were not directly determined experimentally.
Rather, they represent our best estimates of the relative contribution of
each of these processes to the observed attenuations of azinphosmethyl of
Table IV-2. Clearly, intensive laboratory and field work aimed at estimating
under natural conditions is required. Such information is now largely un-
available. As values are determined in the future, they may be entered
directly into the model.
TABLE IV-2. DAILY AZINPHOSMETHYL ATTENUATION PROPORTIONS, UNDER
STANDARD CONDITIONS, SUBJECT TO USER ALTERATION.
LEAVES
GRASS
LITTER
SOIL
CANOPY .0487
ALLEY
.0412
.0670
0412
0670
.0790
.0790
101

-------
TABLE IV-3. DAILY AZINPHOSMETHYL ATTENUATION PROPORTIONS, SEPARATED
BY PROCESS, UNDER STANDARD CONDITIONS, SUBJECT TO USER
ALTERATION.
CANOPY	ALLEY

LEAVES
GRASS
LITTER
SOIL
GRASS
LITTER
SOIL
MICROBIAL DEG.
0.000
0.000
.010
.070
0.000
.017
.070
HYDROLYSIS
.012
.012
.015
.009
.019
.025
.009
PHOTOLYSIS
.018
.006
.003
0.000
.010
.004
0.000
VOLATILIZATION
.019
.024
.013
0.000
.038
.021
0.000
OXIDATION
0.000
0.000
0.000
0.000
0.000
0.000
0.000
Penetrated Residues
Surface-penetrated residues (as defined in Chapter I) consist of residues
on or within the cuticular matrix of tree leaves, grass/broadleaves, and
litter/moss layers. These residues, together with all pesticide residues
in soil, will be called "penetrated" residues in the remainder of this report.
All dislodgeable pesticide residues which enter soil become "penetrated,"
but there is no daily penetration process, and attenuation of soil residues
is included in the attenuation matrices already presented. Three processes
are modeled in representing all other penetrated residues: (1) initial
penetration on the day of pesticide application, (2) daily penetration of a
proportion of the dislodgeable residues in each layer and region, and (3)
daily attenuation of a proportion of the current penetrated residue in each
layer and region.
Initial penetration was estimated from field data as presented in Chapter
I. Little is known about the time-course of penetration of dislodgeable
residues following pesticide application, although it is known that pesticide
leaf uptake may vary with pesticide, formulation, mode of application, amount
applied, leaf type, and environmental conditions (Hull, 1970). Our field-
determined ratios of dislodgeable to surface-penetrated residues in each layer
showed a large daily increase in penetrated residues in the litter/moss layer.
From these limited field data, the estimated daily penetration rates for the
litter/moss layers were .126 and .156 for canopy and alley litter/moss, re-
spectively. In contrast, no daily trend in penetration within tree and grass/
broadleaf layers was discernable, and the model's default daily penetration
rates for these layers are set to zero.
Due to the dearth of information on attenuation of penetrated residues,
the default daily attenuation proportion of penetrated residues was arbitrarily
set at 1%/day for all (non-soil) penetrated residues.
Results and Discussion
A series of runs of the model were made in order to examine the effects
of various modeled processes on the dynamics of the pesticide azinphosmethyl.
Each of these runs is compared graphically with a standard run in which all
parameters are at their default values, and wind speeds, temperatures, rain-
fall, and spray applications are set to represent the 1978 season in our
102

-------
orchard (see Chapters II and III). Daily temperature and soil moisture
effects on the various attenuation rates are included in the solid lines
(the "base line run") of Figure IV7, but are exluded from the dashed
lines, which are based on the standard condition attenuation rates of
Table IV2.
Each graph ^hows one layer's daily azinphosmethyl dislodgeable or soil
residue as yg/cm ground area versus Julian date for the 1978 spray season
(for a definition of the layers and region^, see Chapter I). Ijote that leaf
surface residues were converted from yg/cm leaf area to Ug/cm ground area
based on estimates of leaf surface area/tree described in Chapter I. Dates
and amounts of rainfall are indicated by arrows above the graph, whil pesti-
cide application dates are shown by arrows below (see Chapter III).
In Figure IV-7(a), the characteristic decline of pesticide level follow-
ing each spray is observed in the tree. After heavy rains, an even steeper
decline is evident. The relative lack of effect of temperature on attenuation
in the tree is visible in the overlapping of the solid and dashed lines.
This reflects the lack of temperature effects in our photolysis and volatiliza-
tion submodels, and the relative unimportance of hydrolysis in this layer.
Similar patterns are observed in the grass/broadleaf layers (Figures IV-7
(b) and (e)). Note that while rainfall washes residues from trees to grass,
it washes even more out of the grass layer, resulting in a net decline in
grass/broadleaf residues after rainfall. Again, temperature effects in the
model are minimal on this layer.
The litter/moss layers (Figures IV-7(c) and (f)) are net recipients
of pesticide residues following rain events. Temperature effects are more
pronounced in this layer, although they are still not striking.
In the soii layers, the Influence of daily temperature changes on the
microbial degradation rate is evident, as shown in Figure IV-7(d) and (g).
The warmer-than-standard temperatures later in the season resul in lower
levels of soil residues than under standard conditions.
Figure IV-8(a-d) compares residues from the run using field-determined
daily temperatures (the baseline run of Figure IV7) with residues from a
run in which daily temperatures are lowered by 10°F (dashed lines).
Temperature effects on the residue levels in the tree and grass/broadleaf
layers were barely dlscernable, and are not shown. Greater effects are
observable in the litter/moss and soil layers, which show higher residue
levels at the lower temperatures, attributable to the temperature effects
on microbial degradation in these layers.
Figure IV-9(a-d) shows the effect (dashed ^ines) og increasing the
volatilization rate by a factor of 10 (from 10 to 10 mm Hg). This change
in volatilization rate is small relative to the range of volatilization rates
commonly observed for pesticides. However, the model predicts a pronounced
decrease in residue levels in all layers. Especially interesting is the in-
direct effect on soil residues, from which the model does not postulate any
volatilization (using its default parameters). Thus, in the soil, the re-
duction is solely attributable to lowered availability of residues from upper
layers.
Figure IV-10 shows the effect (dashed line) of altering soil pH from
7 to 9 in the soil layers. This resulted in a marked reduction in residue
levels, attributable to an increased hydrolysis rate. Under these postulated,
strongly alkaline conditions, the model predicts that hydrolysis in soil is
changed from a relatively unimportant process to the dominant cause of
azinphosmethyl loss.
Figure IV-ll(a-g) shows the predicted effects of a postulated two-fold
increase in the daily rates of movement of pesticide residues under rainfall
103

-------
CHMflPY LEHVE; 197fl
imj
CHNOP* ORfiSS-BROnOLEflVES 1976
CANOPY UTTER-MOSS 1978
CANOPY SOIL 1970
TTTTT1 T
i. o
i*
I 1.0
5o.»
< o.s
1.0
ii i« !•' i wiirn
.......	-	• | t	I 14
I	V
t II II SI
4AIW ?*'*
Figure IV-7(a-d).
Predicted 1973 azinphosmethyl residues in orchard canopy region
layers. Dashed lines represent predictions based on standard
condition attenuation rates (Table IV-2), while solid lines in-
clude predicted effects of daily temperature and soil moisture
changes (the "base line run").
104

-------
e	ALLEY GROSS-BSOADLERVES 1978
™"—"firfFi"t wr^n
1.6	i 17 M SI	S	4 12	6	14
1.2
>
* o.e
ImJ *4
n o 0 • 4
O N
£ O
% 3 0.0
*4
« 0.4
0.Q
150
JULIAN OAtC
0.9
£ 0.6
WM
a 5 0.J
O N
s o
* 3 0.0
0.3
0.6
ALLEY LlTTER-flOSS 1978
RAIN (rin) «
1.2
i»l# 11 I fPI• I i
rrr
2 I? II 6)
4 I 2
l l I l I
150
JULIAN 0ATC
2S5
RAIN (M)
2.0
I.S
I t.o
U N
c c M .
w u 0.5
O Si
r o
* 3 o-o
a 0-5
1.0
tut
t ii m
RLLET SOIL 1978
T
SI
wrnrrr
4 12	t 14
150
julian onrt
2SS
Figure IV-7(e-g).
Predicted 1978 azlnphosmethyl residues 1n orchard alley region
layers. Dashed lines represent predictions based on standard
condition attenuation rates (Table IV-2), while solid	in-
clude predicted effects of daily temperature and soil moisture
changes (the "base line run").
105

-------
CRNOPY LEAVES 1978
rain mm
9.2
TT
if i •' i m1- i •'!;
E «-6
UN
C B. .
n u 2.3
e s
£ O
£	3 0.0
a 2.9
4.6
2 I? 11 SI
» 1 ¦ ¦ ' ¦
1 1 1 ¦
ISO
JULIAN CATC
CANOPY GRflSS-BROflOLEflVES 1970
RAIN IIUII •
1.6
U2
•'in j i j .fi-'u I n •'
* 0.8
UM
S U0.4
o s
s o
£ =» 0.0
N ^
a 0*4
o.e
2 1? It 61
4 \ i e m
» * i	* 1 * ' * 1 -
150	JULIAN 0ATC
255
255
CANOPY L[TTER-M0SS 1978
n mm
1.2
J 0-6
u (V
£ uO.l
a 0.3
0.6
*m w i m
i n u si
¦ i ¦ ¦ ¦ « ¦ « *	* '
160
wULlflN OflTC
CANOPY SOIL 1978
ITTT
2.0
1.5
1 1.0
UN
5 uo.s
0.5
1.0
z n ii si
wrn ^
255
'	¦«—1 ' 1 ' * i * 1 ¦ ' i ¦ i »
JUMflN OAie
2SS
Figure IV-8(a-d). Predicted effects of lowering daily temperature by
1o°F on azinphosmethyl residues (dashed lines)
versus the base line run (solid lines).
106

-------
* M|« I Ml
IH'IQI'? UflVCK 1*7*
—:lfIf : i I flMl I i] I
».r	} II ll li	I	• • » «t ««
:: i< , i I j\
fd*. f J j' J
• ?.i l«
** I » k. i t . t. , i
I Ml
L. . 1 k. I
14

!•*

$ 04
8
f r*4
( ®04

• 0.4
*5
04


IM
UMMM GKItrs-bHIMUirNVCO 1978
if If; i j wu i M •'
I •» II Cl	I	4 I • • l«
k.. I . i * t_ ^ t. . »
mim eiir
( i. * •.
Ct-NOPr LltTCB-nOSS ISVB
it
5 o.i
04
	ffl t -I W1H M f
IJ ,	I l» II II	I	« I t • ¦«
-isL
	b U i--l • •*- l.-fc
C.HN0PC SOIL 1978
I i-o
Ite
"TTjirrrwi^ r 7| j
I IT II II	I	4 1 9	• l«
//V
! fi:„ A; cr
¦ 	-
Figure IV-9(a-d). Predicted effects (dashed lines) of increasing pesticide
volatilization rate by 10* (from 10* to 10" on Kg).
Solid lines are base line run.
107

-------
CANOPY SOIL 1978
RAIN (tin)
2.0
1.5
¦E	i .0
Ul CM
c	c n c
co	o U • 5
O	v
x	O
£	=» 0.0
M	_ _
a	0.5
1.0 J
^ J""	^	|r
3 5	4 3	6
W H U	H
2 17	U SI
ffiTT
3 5 3
v w
4 1 2
8	14
UJ
.
OT
I I	L.
J	¦ ¦ ¦	I	I	I	I	I	I	I	I	I	I	I	l—l
ISO
JULIRN OflTE
255
Figure IV-10. Predicted effects (dashed line) of altering soil pH from 7 to 9.
108

-------
CM OPT UriVCS 1978
K	CANOPY ORASS-BROAOLERVES 1976
1H 1 W II I—iTTTTTTl T
i.i
l.t
£ °*»
0.4
04
« 0.4
o.t
l»H I i I f|i|i I .'I !
t IT II fl	I	4 I f	I l«
CflNOPr LtrrcR-noss 1978
c
i .*
0.1
I 0->
11::
M
tu
ISO
CANOPY SOIL 1978
i.i
iU lU II 	Mill i II I
> » v •	»I • • '
I ,i.i
0::
< it ii fi
Figure IV-ll(a-d). Predicted effects (dashed lines), In canopy layers, of
doubling dally rates of pesticide movement due to heavy
and light rainfall. Solid lines represent the base line
run.
109

-------
fthlN inn)
1.6
* 0.8
UN
n o 0.4
o \
e o
£ =>0.0
N
a 0.4
0.0
flLl.rr GRHSS-BROHOLEHVES 1978
n.
2 17 II
.]_r	.
SI	S	4 1 2	B 14

ISO
JULfAM oatc
0.9 .
£ 0.6 .
UN
m u 0.3 .
ox
s o
* 3 0.0
0.3 .
0.6
ALLEY LITTER-flOSS 1978
T
'Ml I • I Ml* i fJ s
2 17 11 SI
1 ¦ ' ¦ ¦
ISO
JULIRN OflTC
255
kJC4
r c
M U
'0.0
0.6 .
1.2 .
ALLEY SOIL 1978
RAIN I Mil •
2.4
1.8
1 .2
0.6
m i pi * i
2 17 II SI	5	4 12	8
ISO
JUIIPN DOTE
25S
Figure IV-ll(e-g). Predicted effects (dashed lines), in alley layers, of
doubling daily rates of pesticide movement due to heavy
and light rainfall. Solid lines represent the base
line run.
110

-------
conditions (both heavy and light rainfall). Such changes may represent dif-
ferences in compound properties, formulation, or foliar characteristics.
Tree residues are reduced, while grass/broadleaf and litter/moss residues
show only slight changes. Canopy soil residues are noticeably Increased,
while alley soil residues rise only slightly. The relative lack of effect
on the grass/broadleaf and litter/moss layers results from the counter-
balancing of Increased residue movement into and out of these layers, while
the soil layers are only recipients of the residues.
Figure IV-12(a-g) shows the effect (dashed lines) of postulating no
pesticide movement under non-rainfall conditions. The tree, as expected,
shows Increased residue levels. The canopy grass/broadleaf residues are
higher than "normal," demonstrating that under normal conditions, this layer
was a net loser of residues due to non-rainfall movement. Similarly, but
to a lesser degree, the canopy litter/moss layer shows an Increase in residue
levels. Canopy soil, on the other hand, shows a decrease in residue levels,
as would be expected, but of a relatively large magnitude. Alley grass/broad-
leaves shows a smaller Increase than the corresponding canopy regions, while
alley litter/moss actually shows a considerable decrease in residue levels,
since this layer is a net recipient of residues under normal non-rainfall
conditions. Alley soil shows little change, as very little pesticide moves
into this layer even under normal non-rainfall conditions.
Conclusions
The literature reviewed and the results of the preliminary experiments
reported provide a basis for additional work. Further model development,
as a method for better understanding of pesticide environmental dynamics,
will require the formulation of laboratory and field experiments that are
specifically designed to provide data to be used in the quantitative analysis
of pesticide fate as a function of environmental conditions. The conceptual
model presented identifies the key processes in the assessment of pesticide
fate in an orchard ecosystem and the working model provides a framework for
further investigation.
Understanding pesticide fate throughout the entire orchard ecosystem
Is now gaining importance as a result of the interest in integrated pest
management as a long-term strategy for pest control. In orchards and many
other crops, where chemicals are still heavily relied upon,, the success of
biological control as a major component of IPM may depend largely on the
judicious use of pesticides. The trend in pest management research has been
to use modeling techniques to better understand the ecobiology of pest, host,
and beneficial species so as to maximize the effectiveness of biological
control measures. A comparable effort in modeling pesticide fate and effects
on both harmful and beneficial species is essential to the development of
effective and efficient orchard integrated pest management programs.
The use of modeling techniques to better understand pesticide fate in
orchards may also provide a basis for the development of models describing
the fate of organic toxicants in terrestrial -ecosystems in general. Ecosystems
use energy from the sun to constantly recycle their water, air, mineral,
plant and animal resources. Natural mechanisms recognized by ecologlsts to
be essential to these systems include: blogeochemial cycling, host-parasite
relations, competition for food and habitat, predator-prey relationships,
food chains, symbioses, community-diversity and succession, and natural se-
lection (Odom, 1971; Pimentel and Goodman, 1974; Southwick, 1976; NSF/RA, 1976).
The effect of the introduction of chemical, substances Into the environment
111

-------
:swp» -i^.ES ;S78
1.1
I 1.0 . 1
n»!'

-------
rain mm —
1.6
1.2
-j
I o.e
ijm
ft o 0.4
D \
c 2
fc 3 0.0
St 0*4
0.8
Tfjyjl
2 I? 11 SI
ALLEY GRASS -BR0AOL EAVES 1978
TTT
\
ill
i

3
5 /
e

¦ * ' ¦ * *
'
160
JULIAN OAre
2SS
DAIN Mil)•
1.2 .
0.9 .
0.6 .
ALLEY LITTER-MOSS 1978
JJW M l fI'J• I .'1 i
.fc S0.0 .
2 0.3 .
0.6 .
2 I? 11 SI
4 I 2
'
ISO
»***«¦- 1 1 ¦¦ 1 I .1 ¦
2SS
JULIAN DATE
g	ALLEY SOIL 1978
		 iju 1i i ijiii i ij i
2.0	9 11 ii Kl	c	* i *	o	i j
1.6
* 1.0 .
UN
g 5 o.s
fe §0.0 .
a O.S .
1.0 .
Figure IV-12(e-g). Predicted effects (dashed lines), in alley layers, of
postulating no Inter-layer pesticide movement under
non-rainfall conditions. Solid lines represent the
base line run.
S
(T
Iti
C
113

-------
on the above-mentioned natural mechanisms of ecosystems is largely unknown.
No tests are currently available which can address ecosystem effects at this
level. The EPA admits that proposed tests under FIFRA and TSCA are only
screening tests and that there is an urgent need for development of higher
level ecosystem tests. The question now arises as to whether these screening
tests can adequately select those compounds which will most likely pose no
hazard to ecosystems. Does a negative result in an acute toxicity test pre-
clude behavioral effects on a population influencing its ability to compete
for food and/or habitat? There is no evidence that even chronic toxicity or
life cycle studies can give the answer to this question. It is disturbing
that negative results for all of the proposed "ecosystem effects" tests may
allow continuous low level exposure of a chemical substance when nothing
is known about the effects of this exposure on any of the aforementioned
"essential natural mechanisms" of ecosystems. A large effort by the scientific
community is needed to better understand ecosystem effects above the species
level. This effort seems paramount in light of the apparent inability of
present screening tests to preclude higher level effects. Rapid development
of ecosystem effects test standards should Include the use of mathematical
modeling techniques to deal with the complexity of the interactions involved.
These ecosystem effects models can then be coupled with chemical exposure
models producing a very powerful tool in ecosystem hazard assessment. The
development of ecosystem models appears to be much farther off than exposure
models. In the near future, much emphasis must be placed on rapid development
of exposure models, in screening chemicals for ecosystem effects. Accurate
predictions of exposure may offset much of the uncertainty In ecosystem effects
screening tests.
114

-------
CHAPTER V
FIELD STUDIES OF AZINPHOSMETHYL
EFFECTS
ON SOIL/LITTER INVERTEBRATES
INTRODUCTION
This chapter summarizes the results of a program of sampling of several
Invertebrate species in sprayed and control plots over a four-year period.
Additional field data for the 1977-78 sampling years for Trachellpus rathkel
(Isopoda) are presented In Chapter IX, while the 1979 T. rathkel field data
are Included here.
Orchard plots;
The following general description of experimental plots Is Intended to
Introduce some of the considerations that will be brought up In the Inter-
pretation of field sampling data given In subsequent sections.
Plot size varied from 0.04 to 0.11 hectares; slopes ranged from 6-12%
(plots 3 and 4) to 18-25% (plots 1 and 2), and slope direction from North
(control 8) all the way to South (plot 4). Soli types were relatively uni-
form, belonging to the Marlette loam, sandy subsoil series (Soil Conservation
Differences in tree population can be exemplified by plot 2, where large,
dense trees prevailed, and plots 3 and 4 which contained small trees with
sparse to medium canopy. Average canopy areas over ground varied from 27%
to 54% of plot area. In 1976, the distribution of ground cover types was
analyzed in all plots (line intercept technique, Cox, 1974), and a link be-
tween tree population and vegetation type became apparent: Importance values
for ground cover categories (ranging from "bare ground" to more complex cate-
gories such as "moss covered by litter and broadleaves") were similar In all
plots except plot 2, which harbored the largest, densest tfees.
All this implies that vertical and horizontal temperature profiles in
soil and litter were equally unevenly distributed. In 1976, temperatures
under canopy cover and in alleys were compared on a number of dates: litter
surface temperatures differed by up to 15°C, soil surface temperatures by up
to 8°C; at a depth of 3 cm, soil temperatures In alleys were up to 5°C
higher than under canopy cover.
The orchard was well Isolated from commerlcal orchards in the area.
Surrounding habitats consisted of oldflelds to the North, an alfalfa field
to the South, and small wooded areas and shrubs to the East and West. Al-
though 12" aluminum fencing (burled about 4") partially surrounded the sprayed
plots for runoff channeling, the fencing was not continuous around all plot
peripheries, and probably presented only a minor obstacle to Invertebrate
migration.
While most of these physical variables can be taken Into account in a
model of pesticide distribution and movement, their combined effects on
invertebrate populations (Including edge habitats as a source of Immigrants)
matter.
115

-------
often have to remain speculative.
Surface-active Invertebrates (1977 to 1979):
Pit-traps containing ethylene glycol were set out for 72 hours at a
time, on 8 sampling dates In 1977 and 1979, and on 5 dates In 1978. The
total number of traps per plot were assigned equally to locations under
canopy cover and in alleys.
Due to priority shifts in the spray program, pit-trap sampling in a
given plot was not consistent through the 3 years. Table V-l provides an
overview of the plots sampled.
The trapped faunal material was sorted to order level (Collembola being
further subdivided into Arthropleona and Symphypleona), and each year's data
for selected groups were subjected to a nested analysis of variance. The
statistical models used for analysis were modified to fit the specific
sampling plan for each year.
Table V-2 shows results of analysis of variance for those faunal
groups that were trapped most consistently and in large enough numbers (at
least in control plots) to merit analysis. Significance levels varied con-
siderably from year to year. Seasonal activity or density peaks were re-
flected in highly significant date effects for all groups. Treatment effect
were generally, and unexpectedly, more prominent in 1979 than In previous
years, although dlazlnon application in 1979 was technically hot successful.
Spiders, which prey on Collembola, may have been affected indirectly:
smlnthurlds, which constituted the overwhelming majority of all trapped
Collembola, were significantly decreased In all years.
Effects of other factors and their interactions were so variable as to
make valid interpretation virtually Impossible. As has been mentioned re-
peatedly, migration of invertebrates to and from plots could not be prevented,
and probably was the major factor responsible for inconsistent results. Over-
all, the data indicate that this type of investigation would be worthwhile
in other locations (such as large, commercially managed orchards) where small-
scale population movements would have relatively little effect.
TABLE V-l PLOTS SAMPLED BY PIT-TRAPPING IN 1977-79
Plot no.
3
4
7
8
1977
not sampled
not sampled
2nd year
Guthlon
2nd year
Guthlon
control
control
1978
3rd year guthlon,
but spraying stopped
after June 9
3rd year Guthlon
not sampled
not sampled
control
control
1979
Not sampled
dlazlnon after
3 yrs Guthlon
not sampled
not sampled
dlazlnon
control
116

-------
TABLE V-2. Results of analysis of variance of selected groups of Invertebrates found In pit-trap samples
(NS " not significant; dashes • factors not Included In the statistical model in 1979).
SYMPHYPLEONE	ARTHROPLEONE
SPIDERS	COLEOPTERA	HEMIPTERA	COLLEHBOLA	COLLEMBOLA

'77
*78
•79
•77
'78
*79
'77
"78
'79
•77
•78
•79
'77
•78
*79
Tmc
.1
.05
.01
.1
NS
.001
NS
NS
.001
.1
.05
.001
NS
NS
.001
Plot(tmt)
.005
NS
-
.025
.001
-
.001
NS
-
.001
.001
-
NS
.05

Location
NS
NS
.005
NS
NS
.001
NS
NS
.1
.1
.01
.001
NS
NS
.001
Tmt-loc.
NS
NS
NS
NS
NS
.001
NS
NS
.001
.1
.05
NS
NS
NS
.1
Plot-loc.
NS
.01
-
NS
.05
-
NS
.1
-
NS
NS
-
NS
NS
-
Date
.001
NS
.005
.001
.1
.001
.001
.001
.001
.05
.005
.001
.001
.001
.001
Tint-date
.1
NS
NS
NS
NS
.001
.005
NS
.001
NS
.01
.001
NS
.05
.001
Plot-date
NS
NS
-
.025
.001
-
NS
NS
-
.001
.1
-
.025
NS
-
Loc.-date
NS
.1
NS
.1
.005
NS
NS
NS
.001
.1
NS
.001
.001
NS
.001
Tmt-loc-da
.001
NS
NS
NS
.025
NS
NS
NS
.1
.1
NS
.025
NS
NS
NS
Plot-loc-da
NS
NS
-
NS
NS
-
NS
.1
-
NS
.001
-
NS
NS
-

-------
LUMBRICIDAE:
Methods: Block samples, 25x25 cm, were taken In 10 cm depth Increments down
to 40 cm In 1976 and to 30 cm In 1977. Each Increment was handsorted for
worms and coccoons. A preliminary comparison between handsorting and formalin
extraction showed that the shallow-burrowing species found In the upper 30
on (predominantly Allolobophora callginosa Savlgny) did not respond to
formalin. The deep-burrowing, surface-feeding forms could have been extracted
by formalin, but the concurrent disturbance of large plot areas prohibited
use of this technique.
A minimum of 3 block samples were taken per plot, twice a year, in June
and October. The sampling locations were chosen such that data analysis
could take in to account: slope location (up-, mid- and down-slope); location
with respect to canopy (under canopy cover vs. in alleys); and sprayed vs.
control plots.
Results: The number of worms recovered from samples varied considerably even
Within one given type of location. Table V-3 shows results of the 1976 season,
in which all plots were sampled (June » pre-spray), and of the 1977 season,
In which only 4 plots were sampled, and plot 8 (as a second control) was
sampled for the first time. Analysis of sample counts from both seasons,
combined according to the parameters mentioned above (slope, canopy cover,
pesticide and season), did not reveal any significant differences in worm
densities, (variances equal, two-sample t-tests, P > 0.2 in all cases).
Table V-A gives a more detailed summary with respect to subsample
depth. Two-way analysis of variance showed a significant interaction between
Guthion. application and depth distribution only in the June 1977 samples (f =
11.13, P < 0.0002): the uppermost layer in sprayed plots harbored a greater
proportion of the total number of worms found than the upper 10 cm in control
plots. One should consider, however, that the samples were taken over a
period of 2 weeks, under slightly varying weather conditions. In addition,
soil texture was not.entirely uniform throughout all control and sprayed
samples. Statistical significance of this difference is therefore open to
several possible interpretations.
Under our experimental conditions, no significant conclusion could be
drawn with respect to any of the parameters anlysed, including effects of
Guthion in the field. If the pesticide had any effects on- the topsoll-dwelllng
worm species, variability introduced by other habitat features masked them.
Thus a reduction in worm populations in June 1977 included those In the
control plot (Table V-4); In this case, periodic compaction by mowing and
spraying would provide, a more plausible explanation for worm mortality (In
yiew of the work by,Morris, 1968, Ghapell et al., 1971, and Arltajat et al.,
1976) than would pesticide effects;
Furthermore, since A. callginosa is a topsoll-dweller, but not a surface
feeder (classified as endog£e by Bouchl, 1972), contact with contaminated
orchard floor/strata must be minimal, especially in dry years such as 1976.
In laboratory experiments, it has been shown that prolonged exposure of A.
callginosa to Guthion will reduce worm weight and coccoon production (van Rhee,
1976). In commercial orchards, which characteristically have heavy spray
programs, pesticide effects may be noticeable: van Rhee (1977) documented
drastic reductions In A. callginosa populations, but in his study Guthion was
only one of several compounds used. Under the comparatively light spray re-
gime used in the present study, Guthion application had no discernible Impact
on the earthworm population.In.uppersoil-layers.
118

-------
TABLE V-3. Mean number of A. callglnosa per sample In 1976 and 1977. Three
samples were taken per plot. The figures Include all worms found In the
upper 30 cm of soil (means ± S.E.)
Plot	June	October	June	October
number	1976	1976	1977	1977

1
56.3
+
0.9
40,0
±
4.6
38.7
+
10.8
70.3
±
12.8

2
55.3
+
7.1
51.3
±
9.6







3
53.3
±
4.7
27.7
+
4.4
25.7
±
1.7
26.7
±
7.9

4
32.3
+
9.4
49.0
±
18.4







5
25.7
±
5.9
22.3
±
2.3







6
34.6
+
5.0
38.0
±
13.3






7
CON
53.6
±
13.7
47.3
t
11.4
23.3
±
6.9
59.7
±
20.6
8
CON






39.5
±
9.3
48.3
±
3.3
TABLE V-4. Average percent of worms found in 3 depth Increments (mean ± S.E.,
number of samples In parentheses) In the 1976 and 1977 sampling seasons.
SUBSAMPLE DEPTH

0-10 cm
10-20 cm
20-30 cm
June 1976
30.4 ± 3.9
53.0 ± 3.3
16.6 ± 2.2
(all plots)
(21)
(21)
(21)
October 1976
43.3 ± 6.1
44.5 ± 4.6
L2.2 ± 2.2
(all plots)
(21)
(21)
(21)
June 1977
61.2 ± 7.9
32.8 ± 6.7
6.0 ± 2.1
(all plots)
(15)
(15)
(15)
October 1977
94.1 ± 1.4
4.6 ± 1.2
.1.3 ± 0.9
(all plots)
(11)
(11)
(11)
June 1977
83.2 ± 6.4
14.1 ± 5.6
2.7 + 2.0
(sprayed)
(6)
(6)
(6)
June 1977
46.4 ± 9.8
45.4 ± 8.3
8.2 ± 3.2
(control)
(9)
(9)
(9)
October 1977
93.7 ± 1.7
5.9 ± 1.8
0.4 ± 0.3
(sprayed)
(6)
(6)
(6)
October 1977
94.6 ± 2.6
3.1 ± 1.6
2.3 ± 1.8
(control)
(5)
(5)
(5)
119

-------
Trachellpus rathkel (Brandt): field populations:
In 1977 through 1979, data on T. rathkei were obtained by field sampling,
and substantiated by laboratory observations on growth, reproductive biology,
and pesticide susceptibility. For data pertaining to 1977-78, and the pop-
ulation model derived from them, the reader may be referred to Chapters VIII-
X. Results from the 1979 season will be discussed in the following section.
The 1979 (diazlnon) spray program was not successful due to technical
inadequacies, which resulted in unknown, but certainly low rates of diazlnon
being applied in the orchard. The first spray event took place in early July,
which is also unusually late in the season.
Isopods were collected from cryptozoan boards, as in previous years.
A brief history of the plots sampled in 1979 is given here: Plot 8: used
as control in 1977 through 1979. Plot 7: was used as second control plot
in 1977-78; sprayed with diazlnon in 1979. Plot 3: had received Guthlon
for three years (1976-78), was sprayed with diazlnon in 1979.
A minimum of 12 boards (usually 20) were sampled per plot and date.
Fig. V-l shows mean numbers of T. rathkei found per board in the three plots.
The catches were divided into two weight classes: £ 12 mg and > 12 mg.
In plot 8 (control), the isopod population behaved much as in previous
years. The conspicuous peak of small animals in August was due to the cumu-
lative effect of two brood releases, which typically take place in late June
and late July to August (Snider and Shaddy, 1980).
Plot 3, in May and June, showed the characteristic depletion of small
animals which can be attributed to Guthlon application (see Chapter X).
The relatively large number collected in mid-June is likely to be due to
immigration: in three consecutive years of pit-trapping (where any catches
are solely the result of active animal movement), spring and fall samples
were the only ones that yielded isopods in significant numbers. From July
on, the population in plot 3 actually underwent a recovery. Ih August, the
season's recruits were plainly evident. Their presence can be ascribed to
three factors: a) the first brood release had taken place prior to diazlnon
application, and the young had time to grow just beyond the most vulnerable
stages; b) diazlnon, In laboratory experiments, was not as toxic to small
animals as Guthlon (at temperatures up to 21°C); and c) in all likelihood,
much less diazlnon reached the orchard floor than Intended, so that even
those young stemming from the second brood release may not-have been affected
by the pesticide at all.
Plot 7, except for the relatively low numbers captured in early July,
showed strong similarities to the control plot. In fact, population structure
Indicated that diazlnon application was Ineffective: rather than being cur-
tailed, recruitment was similar to that observed in the control plot in 1979,
as well as to that found in plot 7 itself in 1977 and 1978, when it was used
as control.
Although T. rathkei populations had been affected by Guthlon application
in previous years, one must conclude that under a subsequent light spray
regime (or none at all) the populations would recover rapidly. Even if no
immigration were taking place, it is likely that normal population levels
would be attained within no more than two years.
120

-------
Figure V-l. T. rachkei sampled In 1979 In three orchard plots
6-
2"
CONTROL PLOT 8
I1
N
H
o> A
Ml
P
•P

6-
3
g ¦
Q
H
£ 4-
a
2
s
a 2-
u
5/18 6/13 6/28 7/6 8/7 8/28 11/7
I]
PLOT 3
n hi H"

T——i——r
5/18 6/13 6/28 7/6 8/7
8/28 11/7
4-
2-
"P
5/18
PLOT 7
-F
J
w
6/13 6/28 7/6 8/7
SAMPLING DATE, 1979
8/28 ll/7
121

-------
CHAPTER VI
LABORATORY TOXICITY STUDIES ON SELECTED SOIL/LITTER INVERTEBRATES
INTRODUCTION
A number of Invertebrates were tested for Guthion susceptibility in a
series of preliminary screening experiments. Some species were chosen be-
cause they were available in mass cultures, and their biology was known,
while others were selected because they occurred in the orchard and were
potential candidates for organism models.
Resulting preliminary data,, as well as the first season's sampling ex-
perience in the field, provided direction for further testing. Some taxa
(chllopods, diplopods and oribatld mites, for Instance) were difficult to
track in the field, as well as being tolerant of Guthion. Others (i.e.
lycosid spiders) could be neither collected nor reared in numbers large
enough for testing, given the time and resources available. Thus laboratory
experiments were later reduced to a few species. Selected data on Collembola,
chllopods and lsopods are summarized here. Data concerning Trachellpus
rathkel, the species tested most extensively, are given in detail In Chapter X.
Materials and Methods:
Soil contamination with Guthion and dlazlnon:
Procedural details for Guthion contamination may be found in Chapter X.
The same methods were used to make up dlazinon-contamlnated soil for contact
toxicity observations on T. rathkel.
Contamination of yeast with guthion:
A slurry of powdered yeast, Guthion (50% W. P.) and acetone was stirred
continuously until all acetone had evaporated, to yield yeast mixtures con-
taining 1000, 2000 and 3000 ppm. By serial dilution of a 1000 ppm stock
mixture with pure yeast, lower concentrations were obtained. Control yeast
was prepared with acetone alone. All mixtures were kept frozen at -20°C,
and small portions were removed from these frozen batches when needed. When
feeding experimental cultures, yeast was placed in the center of a hardened
plaster-charcoal substrate, in cylindrical plastic jars with tight lids, and
was replaced daily throughout a given observation series.
Field-collected litter:
On four dates in 1978, surface litter was raked up In one Guthlon-sprayed
and one control plot and chopped, mixed, and brought up to 200% moisture in
the laboratory. It then served as substrate for various invertebrates, par-
ticularly T. rathkel. For further details, refer to Chapter X.
Results
Collembola:
Adults of Folsomia Candida (Willem) were exposed to Guthion by all
three methods described above. Hypogastrura arroata (Nicolet) was used in
122

-------
soil contact and oral toxicity experiments only.
F. candida (10 Individuals in each of 10 replicate jars) were incubated
at 21°C using field litter as the substrate, and total mortality after 6 days
was recorded. As shown in Table VI-1, the animals incurred 92% mortality
after 6 days exposure to litter collected 2 days after a spray event. A
full month after the last spray period, field-collected litter was still toxic
to F. candida. Judging by our soil toxicity data, showing that higher moisture
content produces higher mortality, litter at less than 200% moisture might
be less toxic to the species (except for rainy periods, litter in the field
would not be that moist).
Table VI-1. Percent mortality of Folsomla Candida after 6 days exposure to
field-collected, Guthion-contaminated litter at 21°C. Replication: 10 indi-
viduals per each of 10 replicate containers.
% mortality in:
ppm residue	days since	control	Guthlon
(GLC analysis)	last spray	litter	litter
25.7	2	4.0	92.0
3.1	30	7.0	15.0
In contrast to litter observations, contact (soil at 10% and 25% moisture)
and oral (yeast) toxicity were assessed by means of 24-hour mortality records
9ver 7 days. These tests were run at 15.5°C, with a total of 200 individuals
of each species per run.
In both H. armata and 1?. candida. contact toxicity was more pronounced
than oral toxicity (Table VI-2). At 10 and 100 ppm, moist soil was distinctly
more toxic than dry soil. Soil at 1 ppm had no effect on either species,
but at 10% moisture egg-laying was Inhibited due to the marginal humidity
regime. At 25% moisture both species produced viable eggs.
Table VI-2. Percent mortality of Collembola after 7 days exposure to Guthlon
In yeast and in soil at two moisture regimes, at 15.5°C. Replication: 40
individuals of each species in each of 5 replicate containers per dosage.
c
o
•H
JS
4J
5
0
a
a


F. candida
H. armata
55
yeast
0.0*
0.3*
O
a
soil, 10%
0.0
0.0

soil, 25%
0.0
1.0

soil, 10%
0.0
0.0

soil, 25%
0.0
0.0
©
yeast
0.5*
1.0
•
o
soil, 10%
5.0
2.0
rH
soil, 25%
26.0
29.0
o
•
yeast
5.0*
9.5
o
H
soil, 10%
20.0
28V0

soil, 25%
39.0
40.0

yeast, 1000 ppm
38.5*
16.0

yeast, 2000 ppm
55.0
15.0

yeast, 3000 ppm
51.8
23.0
*) Averaged results of 2 replicate runs under identical conditions.
123

-------
At termination of the contact toxicity tests (25% moisture) half of the
surviving animals were transferred to control soil for a 4-day recovery
period. In the 1 ppm sample, viable eggs were again produced. The 10 and
100 ppm series, animals which had begun to show slightly uncoordinated move-
ment continued to die. None showed signs of recovery.
In oral toxicity tests, we noted that only yeast with low pesticide
concentration was readily eaten. But apparently even the small amount of
food Initially Ingested by F. Candida produced 55% mortality at 2000 ppm,
and 51.8% at 3000 ppm (Table VI-2); of H. armata, 9.5% died at 100 ppm, com-
pared to 3% of F. Candida. At 3000 ppm, the relatively low mortality of H.
armata (23%) probably Indicates that the food was avoided.
tomlin (1975), using 17 insecticides other than Guthlon, showed that
H. armata was much more tolerant than F. Candida to almost all compounds
tested. In the present study, however, the two species behaved very similarly
when kept on soil at 10 and 100 ppm (Table VI-2). Tomlin's conclusions on
interspecific susceptibility differences were based on 24-hr mortalities.
For comparison, Table VI-3 shows 24-hr Guthlon mortality records. Using
soil, neither 24-hr nor 7-day mortalities (Table VI-2) showed appreciable
Interspecific differences. Using yeast as the pesticide vehicle, £. Candida
did incur higher mortality after 24 hours, but only levels of > 1000 ppm
supported Tomlin's (1975) conclusions concerning the two species.
Folsomla Candida has been used In pesticide screening studies by other
workers (Thompson, 1973; Scopes and Llchtensteln, 1967) and has been subjected
to Guthlon by Thompson and Gore (1972). The latter, authors found that mor-
tality after 24 hours amounted to 70% at 24°G, and 20% at 13°C; the substrate
was Plainfleld sand at 1 ppm Guthlon. The same substrate with 10 ppm Guthlon
resulted in 100% mortality within 24 hours, at both temperatures. In sharp
contrast, we recorded zero mortality at 1 ppm, at 15.5°C, and the production
of viable eggs testified to the health of the animals. Even'after 7 days
on 10 ppm soil (at 25% moisture), mortality In F. Candida was only 26%
(Table VI-2), and 24 hours exposure had no effect at all (Table VI-3).
Table VI-3. Percent mortality of Collembola after 24 hours exposure to Guthlon
In soil (25% moisture) and in yeast, at 15.5°C. Replication: 200 animals
of each species per dosage.
ppm in soil	ppm in yeast

CON
1
10
100
CON
10
100
1000
2000
3000
F. Candida
0.0
0.0
0.0
1.0
0.0
0.0
0.3
8.3
6.2
8.1
H. armata
0.0
0.0
0.0
1.0
0.0
0.5
0.0
0.0
0.0
0.0
Determination of what constitutes "death" may have some bearing on these
discrepancies. Thompson and Gore (1972) counted animals as dead when they
were unable to actively run or spring, while we waited until almost total
inactivity except for-slight twitching'-of- appendages)However, inability
to run after 24 hours would surely have led to total immobility within 7
days or less: the discrepancy between 100% and 26% mortality remains.
Differences in soil type, pesticide formulation Or solvents could fur-
nish more valid considerations in a comparison with Thompson and Gore's
(1972) results. Although this has not been documented for Guthlon, Harris
124

-------
(1967) did show the importance of soil type in toxicity studies: LD.g,s
for crickets kept on fine sandy loam increased by factors of 8.3 (for dlazlnon)
and 9.4 (for parathlon) compared to crickets on a clay soil high in organic
matter (Thompson and Gore, 1972), so that pesticide activity would likely
be much higher than in our soil which contained > 20% clay and _> 3% organic
matter.
Of equal importance, and equally dependent on soil type, is the effect
of the substrate's water content. Harris (1970) documented that the bio-
activity of certain insecticides in sandy loam Increased by a factor of up
to lOOx with increasing moisture content. These effects were greatest in
light mineral soils, but were negligible or even reversed in muck soils
(Harris, 1967). In Brookston clay loam (somewhat similar to our orchard
soil although higher in organic matter), moisture effects were intermediate
(Harris, 1967).
In our experiments, Guthion at 10 ppm was also less active at the lower
soil moisture (Table VIr2): by a factor of 5.2 for F. Candida, and 14.5 for
H. armata. At 100 ppm, the moisture effect was less pronounced (a factor
of < 2.0 for both species).
Chllopoda and Dlplopoda:
The centipede T-ldabius tivlus (Llthobildae) was tested on orchard soil
containing from 3 to 100 ppm Guthion. Table VI-4 summarizes cumulative 7-day
mortalities. As expected, first instars proved to be more susceptible than
adults. Decreasing susceptibility with increasing body size also became
clear in isopods (Chapter X), as well as in oral toxicity tests with the
dlplopod Polydesmus inconstans: yeast at 10 and 100 ppm Guthion doubled
the mortality of instars I over that of adults. However, neither stage
was very susceptible to Guthion: a dosage of 1000 ppm was needed to pro-
duce appreciable effects (> 40% mortality) in both stages.
Table VI-4. Percent mortality of Tldabius tlvius (Chllopoda) after 7 days on
Guthlon-contaminated soil (25% moisture) at 26.7°C. Replication: 20 indivi-
duals per stage and concentration.
ppm Guthion in soil

CON
3.0
5.0
10.0
25.0
50.0
100.0
InBtar I
0.0
5.0
20.0
25.0
100*
100**
-
Adults
0.0
_
0.0
0.0
0.0
100
100**
*) 100% in 48 hourB
**) 100% in 24 hours
Both T. tlvius and I\ Inconstans occurred in the orchard, and were at
one time considered as ciandidates for modeling. The dlplopod population was
extremely sparse, however, and could not be sampled by either pit-trapping
or by the cryptozoan board technique. And, while adults of T. tlvius were
relatively numerous under cryptozoan boards, immature stages of the species
could not be sampled accurately by any means. For these reasons, laboratory
observations on these species were later discontinued.
125

-------
Iaopoda:
Initially, both Traehelipus rathkel and Armadillldium nasutum were used
In toxicity experiments. Once it became clear that only the former species
occurred in the orchard, tests were restricted to T. rathkel, and pertinent
data are reported in Chapter X.
The few test runs performed with A. nasutum are summarized in Table VI-5,
and contrasted to data from a concurrent test of T. rathkel. Somewhat lower
mortality of A. nasutum on soil at 10 ppm may have been due to the fact that
this species tends to be larger and bulkier: in subsequent tests of T.
rat-hfcpj, aninverse-relation between weight and susceptibility to Guthion
was documented (Chapter X)."
Table Vl-5. Toxicity of Guthion to adult isopods at 15.5°C (soil: 25% moisture),
as percent mortality after 7 days exposure.
ppm in yeast	ppm in soil


CON 10
100
1000
CON
1
10
100
A.
nasutum
0.0 5.0
0.0*
100
5.0
5.0
5.0
100
T.
rathkel
- -
-
-
0.0
5.0
15.0
90
*) All animals severely affected (inactive, body lesions evident).
In A. nasutum, some avoidance of contaminated yeast was observed, but
ingestions of yeast with 100 ppm Guthion affected the animals severely,
though slowly; after 7 days exposure, all were inactive, rolled up, and showed
lesions with leaking body fluids, indicating imminent death (Table VI-5).
Because reduced feeding rates were observed in oral toxicity tests, the
possible ability of isopods to detect and avoid contaminated substrates was
investigated. Chopped litter at 200% moisture was distributed in plastic
dishes such that quadrants of contaminated and control litter alternated.
Dosages were 0.5, 5.0 and 10.0 ppm; control dishes contained 4 quadrants of
uncontaminated litter. The covered dishes were kept In the dark at 15.5°C
after two adult isopods had been Introduced into each of 5 replicate dishes
per dosage and species (both T. rathkel and A. nasutum were tested). Three
times dally, for 5 consecutive days, the containers were taken out a few at
a time, and an immediate count was made of the position of the animals with
respect to the four quadrants. Moving versus immobile (resting) numbers
were recorded.
Analysis of variance comparing the numbers of animals found in each
quadrant was performed, both with inclusion and exclusion of moving indivi-
duals. None of the results were significant, indicating that contamination
of the litter did not influence the animals' choice of resting places.
Active avoidance of contaminated strata of the orchard floor was thus made
highly unlikely, and chances of isopods encountering contaminated substrates
in the field probably depend solely on movement of the animals In response
to microclimatic conditions.
126

-------
Table VI-6. Replication in diazinon toxicity experiments with T. rathkel:
numbers of animals observed per weight bin and ppm diazinon, at 26.7°C.
Wt bins:


ppm diazinon in soil


upper limit,







CON
1.0
3.0
5.0
10.0
20.0
0.69
—
1
—
—
_
_
0.97
2
1
-
-
-
—
1.34
3
-
1
-
-
-
1.86
-
2
4
2
1
—
2.59
6
8
3
5
3
—
3.60
4
10
10
15
2
—
5.00
10
19
18
22
5
-
6.22
12
13
14
13
4
_
7.25
4
9
3
8
1
-
9.64
4
4
6
6
5
-
12.00
2
2
2
2
3
—
15.09
2
-
5
7
5
-
18.97
7
-
7
7
9
2
23.86
13
-
10
25
12
6
30.00
19
-
12
25
21
6
31.75
5
-
2
5
6
1
33.60
8
-
4
5
5
3
35.57
5
-
2

3
2
37.65
6
-
5
7
8
2
39.84
4
-
5
3
3
1
42.17
3
-
2
4
8
3
44.63
2
-
-
1
12
2
47.24
2
-
-
-
5
1
50.00
1
-
-
-
5
1
53.03
1
—
—
—
8
—
56.23
1
-
-
-
3
2
59.64
-
-
-
-
3
1
63.25
—
—
_
—
3-

67.07
-
-
—
—
_
1
71.13
-
-
-
-
2
2
75.44
-
-
_
—
1

80.00
. —
_

	


For instars II, a separate series was observed at selected levels of
diazinon (see summary of results in Table VI-11).
For these tests, we used soil from a single large block, taken in an
orchard control plot, sieved, dried, and mixed In large vats before dlazlnon-
contamlnation of weighed subsamples. In spite of repeated, careful efforts
to contaminate soil evenly, the substrates finally used all contained some-
what less than the desired concentrations. (Examples of GLC extraction
results of contaminated subsamples are given in Table VI-7).
127

-------
Trachelipus rathkei: dlazinon toxicity;
In anticipation of sampling results from the last spray season (dlazinon
application) a pilot contact experiment was set up using T. rathkei. orchard
soil contaminated with dlazinon in the laboratory, and constant temperatures
of 15.5, 21 and 26.7°C. Experimental procedures were essentially identical
to previous Guthlon tests (see Chapter X), except for the way in which test
animals were selected and pooled.
Instars II and small individuals up to about 1 mg in weight were obtained
from reproducing laboratory cultures. All other animals were collected
from campus woodlots and Incubated at the three temperatures for at least
one week. They were then weighed individually, assigned to weight bins ac-
cording to those specified in the T. rathkei model, and used -for tests.
A dlazinon "run" at a given dosage and temperature (unlike Guthion runs)
therefore consisted of varying numbers within each weight bin, distributed
such that the most efficient use would be made of the > 2000 animals weighed
at the time. Table VI-6 provides an example of how animals were assigned
to bins in the test series Incubated at 26.7°C; similar distributions were
set up at the lower temperatures. The distinctly uneven distribution of re-
plicate individuals over the entire weight range is a direct reflection of
the weight distribution in the isopod population at the time of collection
(August - September).
Table VI-7. Results of GLC analyses of dlazinon-contamlnated soil.(adjusted
for GLC extraction efficiency) prior to using the soils for toxicity experi-
ments with T. rathkei.
Desired	Detected ppm	n	range
ppm	mean + S.E.	Samples	ppm
3.0	2.81 + 0.09	3	2.60-2.95
10.0	9.02 + 0.59	3	8.21-10.45
20.0	17.97+1.58	3 15.20-21.70
Table VI-8 shows mortality percentages for those weight classes (bins
combined) which contained more than five individuals, after discarding approxi-
mately 15% of the diata because the animals had molted during the observation
period. Both 7-rfay and 10-day mortalities are given. In most cases, and
irrespective of temperature and pesticide concentration (except at 20 ppm)
mortality did not increase from the 7th to the 10th day.
128

-------
Table VI-8. Percent mortality of T. rathkel exposed to dlazinon-contamlnated
soil at three temperatures. Values not In parentheses are cumulative 7-day
mortalities; values In parentheses are 10-day mortalities (same animals,
3 days later). Weight bins as specified In Table VI-6 were combined Into
larger weight classes, to obtain a minimum of 5 replicate animals per class.
Weight



DIAZIHON,
PPM
class, mg
T°C
Control
3.0
5.0
10.0
20.0

a*
4.0(4.0)
0.0(6.3)
8.3(16.7)
— —
100(100)
1.35-5.00
b*
0.0(7.7)
13.5(13.5)
-
- -
- -

c*
0.0(0.0)
42.9(54.3)
47.7(59.1)
90.9(90.9)
-

a
4.2(4.2)
0.0(3.4)
30.8(30.8)
— _
_ _
5.01-12.00
b
0.0(1.1)
3.7(7.4)
-
- -
- -

c
0.0(4.5)
20.0(28.0)
55.2(69.0)
100(100)
-

a
2.8(2.8)
0.0(12.5)
4.5(4.5)
19.0(19.0)
18.2(51.5)
12.01-30.00
b
0.0(0.0)
3.6(3.6)
0.0(0.0)
15.4(30.8)
66.1(81.4)

c
12.2(12.2)
17.6(20.6)
17.2(29.7)
72.3(76.6)
100(100)

a
8.3(8.3)
-
0.0(0.0)
0.0(0.0)
25.0(60.0)
30.01-39.84
b
0.0(0.0)
0.0(0.0)
0.0(5.3)
8.3(3.3)
64.9(78.4)

c
3.6(7.1)
11.1(18.8)
3.8(11.5)
51.9(60.0)
100(100)

a
0.0(0.0)
-
0.0<0.0)
12.5(12.5)
36.4(55.6)
39.85-50.00
b
0.0(0.0)
0.0(0.0)
0.0(0.0)
0.0(0.0)
97.6(97.6)

c
12.5(12.5)
-
20.0(20.0)
53.3(60.0)
100(100)

a
0.0(0.0)
- -
- -
- _
— —
50.01-63.26
b
-
-
-
-
84.6(84.6)

c


"" •
52.9(64.7)
100(100)
*) Temperatures: a
= 15.5°C; b=
21.1°C; c =
26.7°C; all
+ 1°C.

Like Guthlon (Chapter X) and other Insecticides (Harris 1971), biological
activity of diazlnon Increased with Increasing temperature, but magnification
factors were consistent neither within weight class nor within dosage (Table
VI-8).
Using F. Candida as test species, Thompson and Gore (1972) found diazlnon
to be more toxic than Guthlon in a Plalnfleld sand substrate. Since that com-
parison was of interest in the present study as well, our preliminary diazlnon
data were summarized such that they could be contrasted to Guthlon data. Where
10 animals had been observed in selected weight classes (2-3, 5, 15-20,
30-40, and approximately 50 mg), diazlnon and Guthlon toxicities could be
opposed (Table VI-9).
129

-------
Table VI-9. 7-day mortality percentages (corrected for control) for T. rathkel
reared on Guthion- and diazlnon-contamlnated soil. Data Included only if a
minimum of 10 replicate test animals had been observed per weight class and
dosage.
Dosage level
3.0	5.0	10.0	20.0
^?nf1 GUTH DIAZ GUTH DIAZ GUTH DIAZ GUTH DIAZ
Weight, mR			;	
2-3	32.9 0.0
15-20	26.3 6.8 31.6 16.7	15.6°C
30-40	15.0 0.0 26.7 23.6
2-3	70.3 15.4
15-20	10.0 0.0 32.5 16.7
30-40	5.0 0.0	5.0 8.3 100.0 64.9 21.1°C
50	5.0 83.3
2-3	38.9 46.2 55.6 60.0
5	0.0 28.6 60.0 46.2 90.0	88.9 90.0 88.9
15-20	55.0	68.3 55.0 68.3 26.7°C
30-40	20.0 0.0 7.2	55.4 67.8 100.0
50	0.0	84.6
The number of comparisons thus obtained are limited, and require sub-
stantiation with large numbers of equal-weight individuals. However, pre-
sented simply in qualitative terms (Table VI-10), diazlnon appeared to behave
in a way unexpectedly different from Guthion. At lower temperatures, Guthion
was more toxic than diazlnon at 3, 5 and 10 ppm, to animals of all sizes in-
cluding II instars (Table VI-11). At 20 ppm, 21°C, large rathkel were more
susceptible to diazlnon. At 26.7°C (with 2 exceptions at 5 ppm) diazlnon
was more toxic than Guthion, particularly to the larger-sized animals. These
limited data indicate that the distinct Inverse relationship between body
size and pesticide susceptibility (Chapter X) as documented for Guthion,
may not hold true for diazlnon; or may hold true only under certain conditions
(possibly combined functions of dosage and temperature)
130

-------
Table VI-10. Relative toxicity of Guthion and diazinon to Trachellpus rathkei
at three temperatures. For details see Table VI-8 and Chapter X.
Temperature, 9C
15.6	21.1	26.7
3.0	guth > diaz	guth > diaz	diaz > guth
5.0	guth > diaz	guth > diaz	diaz > guth
(for 2-3 mg)
10.0	guth > diaz	guth > diaz	diaz > guth
(for lower weights)
diaz » guth
(for high weights)
20.0	guth - diaz	guth > diaz	diaz > guth
(for 30-40 mg)
diaz » guth
(for 50 ma)
Table VI-11. Toxicity of diazinon-contaminated soils to II. instar Trachellpus
rathkei at 15.6°C, contrasted to Guthion toxicity at the same temperature.
All data corrected for control by Abbott's formula. Replication: minimum of
40 animals per run, (two per replicate container).
Dosage, ppm
1.0	3.0	20.0
diazinon
7-day, %	10.2	34.3	100.0
diazinon,
10-day, %	16.3	34.3	100.0
Guthion,
7-day,%	29.9	55.8	100.0
131

-------
CHAPTER VII
DESCRIPTIONS OF SPIDER, EARTHWORM, AND SPRINGTAIL MODELS
SPIDER MODEL DESCRIPTION
The spider model simulates the population dynamics of a temperate-cli-
mate spider with a two-year life cycle, such as the wolf spider (Lycosidae).
The spider is assumed to grow during its first two summers and breed during
its third. Most female Lycoslds produce a brood In early summer, followed
by a second brood later in the year (Edgar, 1970, Eason and Whltcomb, 1965).
The Lycosidae pass through a number of distinct lnstars separated by molts.
Since most of the demographic data for parameterizing the model are from
Edgar's work with Lycosa lugubris, we have separated the spider population
into eight lnstars, which Edgar found to be a typical number of lnstars for
Lycoaa lugubris, plus adults. The model tracks only female spiders.
The passage of spiders through an individual instar is modeled as a
distributed delay process (see Manetsch, 1976). The delay parameter for each
instar is determined each day by dividing the degree-day threshold for the
Instar (the number of degree-days necessary for completion of the instar)
by the degree-days accumulated on that day, which yields the mean number
of days the spiders should spend in the Instar at the current temperature.
Individuals are removed from the population at each time step to
simulate "natural" mortality, cannibalism, and pesticide-induced mortality.
Values for "natural" mortality rates were taken from Edgar (1971), while
values for cannibalism mortality rates were approximated to keep the popu-
lation size close to observed field values since no data exist concerning
cannibalism rates. It was assumed that spiders of a given instar would be
consumed only by spiders of the same instar (at rates proportional to the
number of spiders in the instar) and spiders of larger lnstars (at rates
proportional to the product of the number of larger spiders and the number
of spiders of the given instar).
For pesticide-induced mortality, a model similar in form to the one
used in the Trachellpus model (described In Chapter X) was used. The cumula-
tive mortality expected at the end of seven days' exposure to the present
day's pesticide concentration is calculated from a regression of mortality
vs. pesticide concentration for first instar spiderllngs and corrected for
temperature and size differences. As in the Trachellpus model, the problt
of the current day's mortality thus calculated is weighted and averaged
with the cumulative mortality experienced in the previous time-step to
yield a new expected cumulative mortality. As in the Trachellpus model,
corrections are made for graduations into and out of an instar.
According to Edgar (1971), spiders hatched in a given year reach,
at most, the fifth instar by winter. Likewise, these spiders do not develop
past the eighth instar (subadult) during the following year. The model
simulates this behavior by storing any spiders which graduate from the fifth
and eighth lnstars from late summer onward and then releasing them after
132

-------
winter into the sixth instar and adult, respectively.
Another developmental phenomenon to be modeled is the rapid growth of
individuals from the second brood. To simulate the increased growth, the
degree-day thresholds for the spiders in the first and second instars are
reduced during the late summer. This treatment is justified since eggs
from the second brood are larger than those from the first (Edgar, 1971)
and would be expected to develop more rapidly.
All adults breed for the first time after accumulating a sufficient
number of degree-days. All adults produce the same number of eggs, but 10%
of the eggs are subtracted as losses to parasitism. After a second degree-
day breeding theshold is passed, 35% of the adults breed for a second time.
Again, all breeding adults produce the same number of eggs, but the second
brood is smaller than the first.
Since the spiders diapause during the winter, the model is turned off
late in the fall and is turned back on in spring after a degree-day
threshold is reached.
Since the model tracks only female spiders, the population size
predicted by the model must be doubled to yield the "true" population size,
except for the adult population from June onward since adult males die in
June while the females live until October (Edgar, 1971).
EARTHWORM MODEL DESCRIPTION
The earthworm model simulates population dynamics of the lumbricld
earthworm Allolobophora callglnosa. This earthworm commonly resides in the
upper ten centimeters of soil, although it may venture deeper than 30
centimeters when conditions become unfavorable (e.g., too cold or hot, or
too dry). Its life expectancy is normally less than two years, and growth
is hermaphroditic; all adults are Included in the reproductive process.
The model tracks the number of cocoons developing and the biomass of the
juvenile and adult populations.
The development of the cocoons produced by the adults is temperature
dependent, and juveniles hatch from the cocoons after a degree-day theshold
is passed. Cocoon mortality is assumed to be zero since cocoons are
apparently very resistant to temperature and moisture changes.
New hatchlings are added to the juvenile population after converting
their population from numbers to biomass. The juveniles gain weight at a
rate which diminishes as they grow, in accordance with the observations of
Murchle (1960). The growth rate Is adjusted downward during any time-step
if soil moisture or temperature are outside the ranges found to be optimal
for growth. In addition, the growth rate is decreased as the population
biomass density Increases, utilizing a carrying-capacity term which multiplies
the growth rate by a factor ranging linearly from unity (when the biomass
is zero) to zero (when the carrying capacity is exceeded). To simulate
mortality, biomass is removed each time step at a rate which Increases
exponentially with decreasing soil moisture and exponentially with tempera-
tures outside the optimal growth range. Other sources of mortality have
been found to be insignificant (Van. Rhee, 1961).
After juveniles have matured for 55 weeks, they are added to the adult
population. The biomass of the adult population is increased each time-step
at a rate equal to the maximum growth rate (taken from the data of Barley
(1959), Satchell (1967), and Waters (1955)) scaled by factors for tempera-
ture and soil moisture, and the carrying-capacity term used to express
blomasss density limitation as in the case of the juvenile growth rate.
133

-------
The mortality rate used is the same used for the juvenile population.
Since it was assumed that the reproduction rate is proportional to
adult biomass, the number of cocoons produced during any time-step is deter-
mined by multiplying the adult biomass by the cocoon production rate. The
cocoon production rate, in turn, is determined by multiplying the maximum
production rate (taken from Evans and Guild, 1948) by factors which scale the
production rate for moisture, temperature and crowding effects. It should
be mentioned that the scaling factor for moisture effects used here is the
same one used throughout the model whenever corrections must be made for
moisture conditions. The function used to determine the scaling factor was
generated from the data of Evans and Guild (1948). Their data were also used
to generate the temperature scaling factor for cocoon production. However,
the scaling factor used to adjust the growth rate of juveniles and adults
for temperature effects was derived from a function fabricated to yield an
increasing growth rate with increasing temperature which levels off at a
constant value of unity for temperatures above 20°C.
FOLSOMIA MODEL DESCRIPTION
The Folsomia model simulates the population dynamics of a hypothetical
springtail of the family Isotomldae, which will hereafter be referred to only
by the generic name Folsomia. Almost all the data used stem from a study of
Folsomia quadrioculata by Gregoire-Wibo (1975), but pesticide-induced
mortality rates are those recorded for Folsomia Candida in our laboratory.
Folsomia has a life span of nine months, but three separate, overlapping
generations exist during the year. That the life expectancies of the three
generations are not the same. Individuals of each generation pass through
three life stages: prejuvenile (which covers the egg stage up to hatching),
juvenile, and adult (which begins when an individual begins egg-laying).
Folsomia reproduce parthenogenetically, so all individuals are tracked by
the model.
The progression of individuals through the life stages is simulated as
a distributed delay process, as in the other organism models. The rate of
advancement is based on the growth rate, which is determined from a degree-
day model of growth. The variability in the durations of the corresponding
life stages for the three generations is a reflection of the Influence of
temperature on the growth rate.
Adults produce eggs continuously in the model at a rate proportional
to the growth rate, and lay eggs at a faster rate in the warmer months than
in the cooler months. In addition, the total number of eggs laid by adults
in the warmer months is greater than the egg production of adults living
in the cooler months.
Sources of mortality considered by the model include predation, other
"natural" mortality and pesticide-induced mortality. Natural mortality
rates were taken from the data of Gregoire-Wibo (1975), while the predation
rates were fabricated to keep the population from growing without bound.
Pestlcide-lhduced mortality rates are determined from a linear regression
of probit mortality vs. log concentration established in the lab. The raw
mortality rate (in probits) obtained from this regression is treated as in
the Trachelipu8 model (including correction for graduations) to yield a
more accurate estimate of the expected mortality.
134

-------
CHAPTER VIII
THE ECOBIOLOGY OF TRACHELIPUS RATHKEI
After extensive Investigation in both laboratory and field, McQueen
(1976c) was able to construct a successful demographic model for a population
of TrachelipuB (Tracheoniscus) rathkei (Brandt 1833) in Ontario, Canada.
The work presented here was undertaken in a similar effort to model a
Trachelipus population, and the information given by McQueen provided useful
reference points. While McQueen's sampling site was a gravel-covered field,
the present study was carried out in an orchard in mid-Michigan; it formed
part of a project concerned with assessing the fate and effects of
azinphosmethyl spray application. Preliminary results pertaining to the
effects of the pesticide on Trachelipus have been published by Snider (1979)
and are given In Chapter IX. The data stemming from unsprayed control
plots form the basis of this report. Together with some of the laboratory
results given here, they have led to the construction of a model of the
Trachelipus population In the orchard (see Chapter X). In accordance
with the structure of that model, the bulk of the data we discuss concerns
only the females of the species.
MATERIAL AND METHODS
In 1976, experimental plots were established near Grand Rapids, Michigan,
in an apple orchard which had been left untended for 20 years. The two control
plots in use during 1978 contained a total of 24 boards, 35 x 15 cm and 2 cm
thick. Half of them were placed on the ground underneath tree canopies, the
others out In the unshaded alleys between trees. Isopods were collected
periodically from the underside of the boards as well as from the surface-
soil and litter beneath them. They were taken in the laboratory, measured,
sexed, checked for breeding condition, and returned to the-boards they came
from.
Field-collected animals were not weighed; instead, multiple regressions—
using head width and width of the seventh thoracic segment—were run to
develop equations that would give the best estimates of weight. For animals
with a head width exceeding 1.0 mm, the regression equation:
Wt o 3.514a3 + 0.454b3 - 1.617,
2
where a = head width and b D thoracic width, gave an r of 97%. For animals
with a head width <1.0 mm, a second equation had to be developed:
Wt = 1.501a3 + 0.578b3 - 0.167,
2
with a r of 94%. For first and second instars, which lack a seventh segment,
mean weights were obtained in the laboratory and used for field-caught
Individuals: 0.5 ± 0.05 mg (n = 15), and 0.7 ± 0.13 mg (n ° 15) respectively.
When validated against size and weight gain of 65 Individuals, observed
135

-------
through at least 2 consecutive molts, both equations predicted a "mid-instar
weight"; i.e., the values fell between the low weights immediately after a
molt and the maximum weight obtained prior to the next molt.
Assignment of field-caught females to weight-related developmental
classes is likely to be more accurate if estimated rather than actual
weights are used: females carrying brood gain a large amount of weight,
but return to their true size-related weight after release of the progeny.
For laboratory observations, Trachelipus were collected at various
times of the year, beginning in early April, in woodlots around MSU campus
as well as in the Grand Rapids orchard site. Depending on the use for which
they were intended, females were kept in isolation, temporarily paired with
males, or held in communal stock cultures. Isolates and pairs were kept in
cylindrical plastic jars with snapon lids (35 mm diameter by 25 mm high);
stocks were reared in larger plastic containers.
The substrate consisted of a layer of plaster of Paris and charcoal,
mixed with water, poured, set and remoistened; and above it a layer of duff,
rotting wood and woodland soil. This material was replaced periodically to
provide cover and food in addition to the powdered yeast used as staple
diet. Distilled water was added when necessary to keep the substrate moist,
but not wet. The cultures were kept at constant temperatures of 15.6, 21.0
and 26.7°C. Observations were made at intervals of 1 to 3 days.
Throughout this publication, reference will be made to weight classes.
The Trachelipus model, (Chapter X) simulates growth as temperature-dependent
advancement through these classes, which can be characterized as follows:
0.5-5.0 mg: following McQueen (1976c), this class is termed "prejuvenile."
The sexes cannot be distinguished; where appropriate, total numbers of pre-
juveniles captured in the field were halved in the calculation of results
(assuming a 1:1 sex ratio).
5.01-12.0 mg: juveniles; the sexes are dinstinguishable, but the
females are not yet capable of reproducing.
12.01-30.0 mg: young adults; the lower limit of this class was defined
according to field evidence of breeding potential (details given later on).
30.1-50.0 mg: adult females, which form the bulk of the breeding
segment of the population at certain times of the season.
>50.0 mg: females nearing or in their second year of life; they never
constituted a conspicuous segment of the population. Individuals exceeding
80 mg were rarely encountered in the field, but laboratory animals could be
reared to weights beyond 100 mg. With respect to laboratory data, these
females will therefore be classified as "50-80 mg class."
Preliminary feeding experiments were carried out to determine whether
yeast would be an adequate food source for Trachelipus in culture. Yeast was
compared to Purina Rabbit Chow which, since Merriam's (1971) determination
of its superior nutritive value, has been used for isopod cultures by other
authors (McQueen and Carnio 197A; McQueen 1976a,c). Newly hatched individuals
were raised in isolation for 130 days at a constant temperature of 26.7°C.
Males were discarded once the sexes could be distinguished.
Fig. VIII-1 shows mean weight gain of females on the two diets, summa-
rized at 10-day intervals. Yeast out-performed Purina slightly, and was
henceforth used as the main diet in all cultures, especially since it had a
lesser tendency to mold during non-feeding periods of isopods (e.g. before
a molt). Mortality was nil in the Purina-fed series, and was also nil in
the yeast series after the first 21 days. Up until then, three of the in-
itial thirty-six replicates had died, but their sex was not known at the time.
136

-------
A ge in days x 10
Figure VIII-1. Growth of Trachelipus females on two diets (mean
weights at 10-day intervals, plus or minus one
S.D. at selected ages).
137

-------
RESULTS: LABORATORY
BREEDING PERIODICITY
As shown in Table VIII-1, the duration of gravidity is temperature-
dependent, ranging from about 18 days at 26.7°C to 51 days at 15.6°C. The
rest interval between brood release and the next brood molt is not as clearly
influenced by temperature (Table VIII-2). The incidence of a sterile molt
intervening between two brood molts was most common in females weighing less
than 30 mg at the onset of breeding. It allowed them to undergo a brief
period of growth, and they frequently reached weights of 30 mg or more before
producing their second brood.
The production of two or three consecutive broods appears to be the
rule in laboratory-grown individuals of other species of isopods (Verhoeff
1920, Bakker 1956, McQueen and Carnio 1974). For Trachelipus, McQueen (1976c)
reported that 83% of mated adults produced one brood, and 42% a second, with
some indication that the reproductive rate might be temperature-dependent.
The data in Table VIII-3 were classified by female weight prior to the
first brood molt, in order to assess the reproductive capabilities of a given
weight class. Roughly half of these females were mated in the laboratory,
and their pre-breeding weights, ranging from 18 to 81 mg, were known. All
of them produced eggs within a short period of time (i.e. 100% of these fe-
males, at all temperatures, had a first brood). The remainder of the ob-
servations in Table VIII-3 stems from individuals collected in the field in
late May, already carrying their first brood of the season. For these, pre-
brood weights could be estimated after assessing the size gain connected
with the brood molt of 28 females in the 30-50 mg range (mean gain of 0.1
± 0.02 mm in head width, and 0.2 ± 0.07 mm in thoracic width). After sub-
traction of these means from the measurements of the females, pre-breeding
weights were estimated by regression equation.
At least at constant temperatures, the production of second broods
appeared to be influenced by initial weight, since a larger percentage of
the heavier females tended to breed twice at all temperatures. It also
seemed temperature-dependent, since the incidence of second broods increased
with increasing temperature. Third broods were common at 26.7°C, and only
at that temperature; one female even produced four consecutive broods.
In long-term observations it was found that reproduction then ceased,
but that the same pattern was repeated after varying lengths of time
(Table VII-4). By then, females weighed about 50 mg or more, and again pro-
duced one to three broods. Since none of the animals were allowed to mate
since their very first brood, observations by Heeley (1941) concerning the
sperm storage capability of isopods were confirmed. However, sperm viabil-
ity was affected by temperature: while at 15.6 and 21°C over 90% of all
broods were successful, 89% of the females kept at 26.7°C prematurely ejected
clusters of undeveloped eggs. The occurrence of third and fourth broods
at 26.7°C is reminiscent of the pattern observed in the first breeding
period, with its evidence of temperature-dependency.
Since all individuals had been field-collected in the spring, those
with initial weights exceeding 50 mg had almost certainly reproduced the
preceeding year. For these,the second breeding period observed in the
laboratory actually represented a third, suggesting a reproductive potential
that is continuous throughout the life span of Trachelipus. Its rhythmicity,
determined in the field by seasonal temperature evolution, is retained to a
certain degree at constant laboratory temperatures.
138

-------
FECUNDITY
As has been reported by other authors (Paris and Pitelka, 1962,
Brereton, 1956, McQueen and Carnio, 1974, McQueen, 1976c) the number of
young produced by lsopods is linearly related to size or weight of the
female. In Trachelipus, it ranges from about 10 young for females less
than 20 mg to over 30 young for those more than 50 mg in weight. The
highest number produced in culture was 60 young, by each of two females
weighing 70 and 90 mg.
TABLE VIII-1. DURATION OF GRAVIDITY, IN DAYS, OF Trachelipus FEMALES KEPT AT
THREE TEMPERATURES.	
Mean
T°C
N
duration
S.D.
Range
15.6
28
51.4
±1.5
48-54
21.0
54
30.7
±1.3
28-35
26.7
87
17.6
±1.0
16-20
TABLE VIII-2. DURATION, IN DAYS, OF THE REST INTERVAL BETWEEN BROOD RELEASE AND
THE NEXT BROOD MOLT OF Trachelipus.
T°C
N
Mean
duration
S.D.
Range
15.6
20
16.9
± 3.4
13-27
21.0
18
17.2
± 4.3
8-25
26.7
50
13.4
±5.2
7-27
TABLE VIII-3. NUMBER OF CONSECUTIVE BROODS PRODUCED BY Trachelipus FEMALES IN
	THE FIRST BREEDING PERIOD (Percent italics).	

Pre-brood
No.
obs. No. ^roducing^
T°C
weight (mg)
(1st
brood) 2nd brood 3rd brood
15.6
<30
12
2
16. 7
0


30-50
25
11
44.0
0


>50
10
8
80.0
0

21.0
<30
17
5
29.4
0


30-50
15
10
66. 7
0


>50
18
14
77.8
0

26.7
<30
17
17
100
2
11.8

30-50
20
20
100
16
80.0

>50
11
11
100
5
45.5
Note: Assignment to weight classes was based on the weight prior to the
first brood, regardless of whether they graduated to a higher weight
class during the breeding period.
139

-------
50-
o30-
10-
-T"
10
T
Female weight, mg
Figure VII1-2- First broods of Traohelipua: relationship between the
post-parturient weight of females and the number of
young released (y = 0.859 wt - 6.434; r = 0.77).
40-
o
Z
20-
~20~
40
Weight, mg
60
Figure VI11-3- Second broods of Trachelipus: relationship between
post-parturient female weight and7number of young
released (y = 0.538 wt + 1.154; r = 0.47).
140

-------
TABLE VIII-4. SECOND BREEDING PERIOD OF Trachelipus FEMALES REARED AT CONSTANT
	TEMPERATURES.	
Temperature C	
Days since 1st breed, period 15.6	21.0	26.7	
(Mean ± S.D.)	136.3+30.8 115.2+29.7 140.5±35.8
No. alive	33	20	23
No. producing brood 1	31	17	13
brood 2	11	12	8
brood 3	0	0	5
brood 4	0	0	1
Note: No. alive: total number of females, of approximately the same
weights as the breeders, that were alive through the second
breeding period without reproducing.
2
McQueen (1976c) obtained an r of 37% for this correlation. Fecundity
is indeed variable, but some of that variability may be due to observational
error. At least in culture, females frequently cannibalize their young
during the first 24 hours following brood release. For example, when females
kept at 21°C were observed twice a day, and were removed as soon as the last
progeny had emerged,£the relation between number of young in the first brood
and weight gave an r of 91%. The correlation was not as good for second
broods, nor was it the same at all temperatures.
Second broods yielded fewer young than the first at 15.6°C (t = 2.433,
d.f. = 30, P < 0.05) as well as at 21°C (t = 2.866, d.f. - 36, P < 0.01).
At 26.7°C mean number of young in first and second broods were virtually
equal, due to females that underwent a sterile molt between broods and were
able to produce more young the second time. Heeley (1941) and Paris and
Pitelka (1962) reported smaller second broods in other lsopod species as well.
After summarizing over all temperatures, regression lines were obtained
for first (Fig. VIII-2) and second broods (Fig. VIII-3). The slopes differ
significantly (t = 2.024, d.f. = 179, P = 0.05).
TABLE VIII-6. PERCENT UNSUCCESSFUL BROODS IN THE FIRST PERIOD OF Trachelipus,
PER WEIGHT CLASS AND TEMPERATURE*


Pre-brood
Brood 1 Brood 2
Brood 3
T°C
weight (mg)
n % unsucc. n % unsucc.
n % unsucc.
<30
22
18.2
1
0.0
0
—
30-50
33
9.1
12
8.3
0
—
>50
11
18.2
10
20.0
0
—
<30
18
5.6
0
—
0
—
30-50
20
0.0
11
0.0
0
—
>50
18
5.6
10
10.0
0
—
<30
31
9.7
9
11.1
0
—
30-50
27
7.4
36
11.1
20
10.0
>50
7
14.3
12
25.0
5
0.0
*VJeights are those prior to each of the consecutive broods, so that success
of a brood relates to the weight of the female carrying it.
141

-------
Weighty gain_^ mg ^
c* o ol o
j	i	1	L
Figure VIII-4. Relationship between weight gained during gravidity
and the number of young produced. Weight gain =
maximum recorded weight - initial gravid weight
(y = 0.264x + 1.154; r = 0.47).
142

-------
Brereton (1957) noted that viability of eggs of Porcellio scaber was
about 82%, and Paris and Pitelka (1962) found a brood pouch mortality of
7 to 8% in Armadillidium vqlgare. In Trachelipus» no checks were made for
the presence of undeveloped eggs in the brood pouch after release. Any
brood pouch mortality would already be included in estimates of fecundity
which, strictly speaking, are estimates of natality (the number of live
young actually released by a female). But there were incidences of
unsuccessful broods in which the entire batch of eggs failed to develop.
These eggs were actively expelled after having been carried for varying
lengths of time, or remained in the pouch until being case off with the post-
parturient molt. As observed by Hatchett (1947), morbid ova were almost
always eaten by the female.
Table VIII-6 shows that the occurrence of faulty broods in the first
breeding period was lowest at 21°C, and that, overall, females in the 30-50 mg
class bred most successfully.
During gravidity, Trachelipus females gained weight rapidly as the eggs
developed. Weight increases ranged from 313 to 22.1 mg (or 8 to 28% of the
weight immediately after a brood molt). Maximum weights were usually reached
in the early fourth quarter of the brood-carrying period. A linear relation-
ship between weight gain and number of young could be shown to exist
(t = 6.183, d.f. = 44, P < 0.001), with a correlation coefficient of
0.68 (Fig. VIII-4).
TABLE VIII-7. NUMBER OF DAYS NEEDED BY Trachelipus FEMALES TO REACH THE FINAL
	MEAN WEIGHTS (+S.D.) OF FIVE WEIGHTS CLASSES1.	

Temperature °C




Weight class (mg)
15.6

21.0

26.7

0.5-5.0
167
5.0±1.3
10
79
5.Oil.3
12
43
5.0±1.6
44
5.01-12.0
57
12.0±4.0
10
39
12.0±3.2
22
19
12.0+2.2
44
12.01-30.0
151
30.0±9.7
10
61
30.0+6.5
22
-40
30.0+6.3
44
30.01-50.0
170
50.0+5.3
6
92
50.0+4.6
22
76
50.0±5.6
22
50.01-80.0
2602
65.0+2.9
7
1612
71.0±6.1
7
138
80.0±5.5
14
1In italics: number of replicates alive at the time final mean weights
were attained.
Estimated from growth part way through the weight class; 15.6°C: 15 mg
gain in 130 days; 218C: 21 mg gain in 110 days.
143

-------
GROWTH
The time spent in gravidity, when no size increase through molting is
possible, represents a delay of normal growth which is resumed with occurrence
of the last post-parturient molt. Figure VIII-5 shows the long-term growth
of breeding females. Almost all individuals happened to fall with in three
specific weight ranges after their first brood release, and were grouped
according to these weights: <30 mg (very small adults, 30-40 mg, and 50
to 60 mg. The data were synchronized such that day 10 was the day of the
first brood molt for all replicates. The temporary weight gain during
gravidity was ignored, and all 10-day mean weights were obtained by linear
interpolation between actually recorded weights (at varying intervals of
8 days to 3 weeks).
The delaying effect of reproduction on weight increase became most
obvious at 21°C (Fig. VIII-5b): The first breeding period was followed by
a time of rapid weight gain, a pattern repeated in the second breeding period.
At all temperatures, growth following the first breeding period was most
rapid in the smallest females. Only mean weights without confidence limits
are given In these figures, because the individuals In each group were not
all of the same weight Initially. The largest standard deviations were re-
corded toward the end of the observation period, and ranged then from 12 to
19 percent of the means; standard deviations of mean weights on day 0 were
only slightly lower.
Attempts were made to rear groups of females through each of the four
major weight classes in order to obtain temperature-dependent growth rates
without breeding interference. But many of the post-parturient females
isolated for growth observations, whether field-collected in the fall or
laboratory-reared, produced eggs again before having reached the upper limit
of a given weight class. And the time needed to rear prejuveniles all the
way to large adults was found to be prohibitive, especially at lower
temperatures.
Finally, records of Individual females were selected from a number of
different observational series. Females were chosen only if they had reached
the lower limit of a weight class after the last post-parturient molt, and
had come at least close to the upper limit before breeding again. It was
then possible to calculate the number of days needed to pass through each of
the weight classes, i.e. to determine the day on which mean weights of 5.0,
30.0, 50.0 and 80;0 mg were attained (Fig. VIII-6). In the more detailed
data presented in Table VII-7, standard deviations of final mean weights
were included since all replicates per weight class had the same Initial
weight.
144

-------
eo-
60-
w
s
1 «0-
3)
i
20

KXX«*
X**
(to)
.X"
.x**:
XXX**
*XX*
#»«*xx
...•XX
,xx**xx.
CXXK*'
**xX
<23)
X** .*
|X*'
<9)
(X*
¦X*
15.6 °C
5 10	20	30
Time (days * 10)
80-
60
j= 40-
20-

x«*
*x*
(6)
x«
X**

.••*x*
vX*
vxx*'
(12)
x*'
21 °C
^xx**
(9)
-i	1	1	1	' 3~
5 10	20	30
Time (days x 10)
5 b
60-
£
2 40
20' Jl'
•xx . • ,

(W)
X* •
(17) X
26.7 *C
X
(17)
-i	1	r
5 10
20	30
Time (days x 10)
30
25
20
z
«
» 10
50-80 mg
—> 30-50 mg
5 -30 mg
r0.5 -5 mg
—i	,	-t
15.6	21	26-7
Temperature °C
i;
Figure VIII-5a, b, c. Mean weights at 10-day intervals of breeding
Tradhelipus females reared at three temperatures. Crosses
replacing dots indicate that 40"percent or more of all indi-
viduals were carrying broods./vAt each temperature, the repli-
cates were grouped according to female weights after the
first brood release: <30 mg, 30-40 mg, and 50-60 mg. Numbers
in parentheses: number of replicates at the beginning of ob-
servation.
Figure VIII-6. Number of days spent by Trachelipus females in each of
four weight classes, at three constant temperatures.
145

-------
TABLE VIII-9 PERCENT MORTALITY OCCURRING EITHER DURING GRAVIDITY OR DURING
THE POST-PARTURIENT MOLT OF BREEDING Trachelitis FEMALES.
T°C
Weight
(mg)
Percent Mortality

Brood 1
Brood 2
Brood 3

<30
18.2
0.0
_ _
15.6
30-50
9.1
8.3
—

>50
18.2
20.0
—

<30
0.0
—
—
21.0
30-50
0.0
0.0
—

>50
5.6
10.0
—

<30
0.0
11.1
—
26.7
30-50
3.7
5.6
10.0

<50
0.0
8.3
0.0
Note: As in Table VIII-6, weights are those prior to each of the consecutive
broods.
TABLE VII1-8. CUMULATIVE TOTAL MORTALITY OF NON-BREEDING Trachelipus, IN
PERCENT OF ALL INDIVIDUALS OBSERVED THROUGH EACH OF THE FOUR
WEIGHT CLASSES (ONLY THE LOWEST CLASS INCLUDES MALES AND
FEMALES, IN UNKNOWN PROPORTIONS.
Wt. class (ma)
Temperature °C




15.6

21.0

26.7

0.5-5.0
20.0
40
10.0
20
7.0
100
5.01-30.0
19.0
21
8.7
23
11.5
52
30.01-50.0
29. 4
34
12.1
33
14.0
43
50.01-80.0
—

40.0
15
48.9
45
Note: In italics: number of replicates.
146

-------
TABLE VIII-10. SUMMARY OF BOARD COLLECTIONS OF Trachellpus IN 1978.

Collection
date











4/7
4/25
5/2
5/18
5/30
6/16
6/27
7/10
7/28
8/8
8/30
9/28
11/1
Total
Number
Collected
238
132
89
114
53
114
193
85
114
106
274
180
150
No.
Females
109
70
28
39
33
65
62
30
41
26
69
58
72
%
Females
52.4
55.5
39.4
57.4
66.0
61.3
61.4
60.0
61.2
54.2
52.7
60.4
52.9
No.
Females
> 12 mg
87
63
24
21
30
57
51
24
39
20
22
50
65
Note: Total numbers Include prejuveniles; numbers of females and percent females are based
only on animals In which sexes could be distinguished [% females = no. of females/
(females + males)].

-------
The necessity to discontinue observations arose at a time when animals
kept at 15.6 and 21°C had not yet reached 80 mg. The duration of these
classes was therefore estimated based on the number of days needed to attain
weights of 65 mg (15.6°C) and 71 mg (21°C). These were linear estimates;
actual durations would probably have been longer, since the non-linearity
of growth in large adults has been shown by McQueen (1976c).
MORTALITY
A distinction was made between brooding and non-brooding female
survivorship rates. The data for non-reproductive Trachelipus were derived
from a number of different observational series, all using isolated
individuals. Records of adults were selected if the animals had entered
a size class after a post-parturient molt.
Table VIII-8 shows cumulative mortality percentages per temperature and
size class. The data are based on weight only—e.g. they concern animals
that died before having reached the upper limit of their class, Irrespective
of how many days they were actually alive. For all but prejuvenlles, 21°C
was the most favorable temperature. The data for adults agree well with
McQueen's (1976c) results, which determined maximum survivorship to lie
between 20 and 25°C. But at 15.6°C prejuvenlles Incurred less mortality,
and juveniles greater mortality, than the populations studied by McQueen
at a comparable temperature of 15°C.
Mortality increased with increased female weight. At 15.6°C very few
animals ever reached 80 mg, so that reliable percentages could not be
calculated. At 21 and 26.7°C some individuals continued to grow to weights
of about 100 mg, but almost none of the field-caught females exceeded
estimated weights of 80 mg. This may indicate that field conditions do not
allow survival beyond 80 mg, or if so, that those animals form only a minute
segment of the population.
As previously observed by Hatchett (1947), Trachelipus females frequently
died during a molt, especially while trying to shed the anterior half of the
exoskeleton. Paris and Pitelka (1962) agreed with Geiser (1934), who
explained the preponderance of adult males over adult females by a greater
female mortality during brood molts. In the present study, "breeding
mortality" included females that died while carrying brood, those that died
while expelling morbid ova, and those that did not survive "the post-
parturient molt.
Table VIII-9 shows percent mortality connected with each brood. At
15.6°C all deaths were connected with non-developing eggs; at the two higher
temperatures, some females survived faulty broods (see Table 6) and produced
a second, successful, brood. Thus brooding deaths occurred most commonly
during or immediately after unsuccessful broods. Only 16 percent of all
brooding mortality was connected with unsuccessful post-parturient molts.
RESULTS: FIELD OBSERVATIONS
NUMBERS IN THE FIELD
Table VIII-10 summarizes the capture totals and the data pertinent to
the female population. As in 1977 (Chapter IX or Snider 1979) the sex ratio
during the 1978 sampling season was generally around 60% females, and not
nearly as variable as that observed by McQueen (1976c). Also shown are the
number of females estimated at ^ 12 mg in weight, which was the lower limit
set for breeding ability in this study. McQueen (1976c) suggested that
148

-------
sexual maturity is reached in the 20 to 30 mg range; the present estimate
therefore appears conservative, but individuals estimated at 12 to 24 mg
were found to make up a sizeable segment of the reproductive population in
the orchard.
As pointed out by McQueen (1976c), numbers of animals captured from
boards vary greatly, mainly as a function of weather conditions prior to
the collection day, and do not reflect true density. In order to obtain
a seasonal point-estimate, population density was assessed in late October,
1978; 20 block samples, 15 x 15 x 15 cm, from under tree canopies and 20
from alley areas were handsorted for isopods. Mean density ± S.E. was
estimated at 107 ± 26.8 and 60 ± 19.7 individuals per m respectively.
These differing densities are also reflected in the number of individuals
found under boards in the two habitat types: boards under canopy cover
harbor significantly more animals than those in alleys (Snider 1979).
Population s£ze of Trachellpus seems quite low when compared to an
estimate of 538/m for Armadillldium vulgare in California grassland (Paris
and Pitelka 1962). Aside from species specific differences, one reason for
low density could be found in the degree of disturbance at the orchard site—
mowing by tractor and human trampling being the most obvious.
Using the number of animals2collected from boards on November 1 (a
board covering an area.of 525 cm ), one would arrive at a mean density ±
S.G. of 112.8 ± 50.9/m for the area under trees and 125.3 ± 72.8/m for
alleys between trees. The estimate of density under canopy is close to that
derived from handsorting. That in alleys is about twice the actual density.
The very large standard errors reflect the fact that large aggregations of
Trachellpus often form under individual boards.
Further handsorting at different times of the season is clearly indicated.
It would allow the determination of the degree of "attractiveness" of the
boards as it relates to weather conditions, characteristics of the habitats,
and to the true density of the population in the surrounding areas.
POPULATION SIZE STRUCTURE
In Figure VIII-7 weight distribution over the four major classes is
expressed as percent of total catch of females per collection date. Ad-
vancement from one class to the next is clearly defined in only a few in-
stances; e.g., movement into the highest weight class in mld^summer;
disappearance of these > 50 mg animals in late summer and their reappearance
in the fall during the brief period of growth before the onset of winter.
McQueen (1976a), using mark-recapture methods on a population of
Porcellio spinicornis, noted that capture efficiency varied with sex as well
as with size of individuals. In the present study, collections from April 7
to June 16 yielded variable numbers of prejuveniles, with an apparent peak
on May 18. Since no recruitment took place at that time, increasing pre-
juvenile captures may best be explained by differential response of large
and small animals to temperature and moisture conditions, at least in April
and early May. Beyond that time, it is obvious that rapid growth is respon-
sible for removing individuals from the prejunvelle contingent and shifting
them into the juvenile and young adult classes.
The growth pattern can be shown more clearly if each of the four classes
is broken down to three subclasses (Fig. VIII-8). The prejuvenile peak
apparent in Fig. VIII-7 in May can now be shown to consist of overwintered
animals belonging in the heavier prejuvenile groups, which have outgrown the
5 mg limit by the time brood release begins. New recruits appear in late June
149

-------
60-
40-
20-
c 60
ti
u
41
a.
40
20
1 2 34
11-17-77
(65.5)
¦fl
4-7-78
(12?.S>
4-

fl
4-25
(68-V
5-7
(36;
5-18
(62.5J
5-30
(34.5;
6-16
(684)
I-
1-
1 ri
q
-i
-i
p
ci




-k—
¦Or—
M.
1
6-77 7-10	7-28	8-8 8-30	9-28 11-1
(107; (47.5; (64.5; (5}> (130-SI (98) (7«l
Collection date
Figure VI11-7. Weight distribution of the Trachelipus field popula-
tion, in percent of all females captured per date.
Total numbers (in parentheses) include half of the
prejuveniles plus all identifiable females caught.
Weight class definitions: 1 = 0.5-5.0 mg; 2 =
5.01-30.0 mg; 3 = 30.01-50.0 mg; 4 = >50.0 mg.
150

-------
°/o
30-
in
La
T^-
r i i—r-i— n—i • i i—r
11-17-77	4-7-7E	<1-25
/

5-18
10-
i
fUj
¦ffly {
i 'i i' i—i 'i i i 'i—i i ¦ 1 i
5-30	6-!6	6-?7	1-S

7-26
in
L j
"fii, ,ft

-i—i i i i i—r ¦ ¦ 'i
8-8	8-30	9-28
Collection date
-Pi-
Figure VII1-8. Distribution of the Traohelipus population over four
weight classes, each divided into the following sub-
classes: Class 1: 0.5-0.89; 0.90-2.11; 5.0 mg.
Class 2: 5.01-9.79; 9.80-19.17; 19.18-30.0 mg.
Class 3: 30.01-35.57; 35.58-42.17; 42.18-50.0 mg.
Class 4: 50.01-59.28; 59.29-70.29; >70.29 mg.
Data are given as percent of all females caught.
Total numbers as in Fig. 7.
151

-------
and graduate through the prejuvenile class during July. A second pre-
juvenile peak then occurs, with a subsequent rise in juvenile numbers
apparent in late August.
In analogy to the population studied by McQueen (1976c), cohorts of
recruits can be followed through the year. In April and May, the last
individuals of the 1977 cohort graduate to juveniles and adults. The
heaviest females found in April may be the remainder of the 1976 cohort
(2 year-olds), but individuals also grow into the >50 mg class during June.
In July, both medium-sized and large adults begin to dwindle, indicating
that the majority of the 1977 cohort (1 year-olds) and probably all of the
1976 cohort (2 year-olds) die at that time. The 1978 cohort replenishes
the heavier weight classes during fall growth. By early November, the
bulk of the population consists of young adults and adults, with an over-
all weight distribution reminiscent of that found in April.
BREEDING PERIODICITY
Collections in April and early May never yielded gravid females.
They first appeared on May 18; by late May, over 70 percent of the females
were gravid.
During June the population rapidly reached its full reproductive
potential. The solid line in Fig. VIII-9 includes all individuals "in breed-
ing condition," i.e., both those carrying brood and those which had recently
released their progeny (oostegites still present). For comparison with
McQueen's (1976c) Canadian Trachelipus population, Fig. 9 also shows
breeding percentages based only on females actually carrying brood; the
two peaks are about one month apart, reaching 86 and 60 percent of all
females of potential breeding weight. McQueen's field counts of gravid
females were thus closely replicated, concerning both the timing of repro-
duction and the percentage of breeders in the population.
By assigning brood-carrying females to three weight classes, it can be
shown that reproductive activity is initiated by the heavier females, and
that individuals weighing <30 mg gain in relative importance toward the
latter part of the season (Fig. VIII-10). Larger females exhibit a shift in-
to a second brood. Smaller females appear to come into breeding condition
gradually, as individuals attain reproductive size during summer growth.
By July, their importance begins to exceed by far that of the heavier
females.
This breeding pattern is substantiated by the pattern of brood release.
Table 11 differentiates between two states of breeding condition: broods
developing in the pouch, and empty oostegites prior to the post-parturient
molt. The first release-pulse (late June to early July) is due mainly to
females >30 mg in weight; by late July, these again carry brood. In small
females, first and second broods are not clearly distinguishable. Those
that were able to breed relatively early in the season probably breed a
second time, but may by then have entered the next higher weight class.
Table VIII-11 also shows that 100 percent of all females >30 mg in weight
carried broods by late May; and that the small females reached a repro-
ductive peak in July, when 93 percent of all captured females in this
class were in breeding condition.
152

-------
100-]
I
H
I
I
I
60-j
i
I
|
20-
i
_r	r , •• r •• ~r	1	1	1—
5-18 5-30 6-H'. 6-27 7-10 7 28 8-8 8-30
0 a t e
Figure VI11-9- Breeding periodicity of Traohelipus in the field:
of all females >_ 12 mg, percent in breeding con-
dition. Solid line: females with empty oostegites
included; broken line: only individuals actually
carrying brood are counted.
100-

-------
Table VIII-11. Breeding periodicity of Trachelipus females in the field,
categorized by estimated weights.

12-30
mg

30-50
mg

>50 mg



no. in


no. in


no. in



breed.
BR
REL
breed.
BR
REL
breed.
BR
REL
Date
cond.
%
%
cond.
%
%
cond.
%
%
18/5
1
10
100
—
1
10
100
—
0
1
—
—
30/5
4
12
100
	
13
13
100

5
5
100
—
16/6
10
17
100
	
28
28
100
— —
12
12
91.7
8.3
27/6
15
19
64.3
35.7
22
22
27.3
72.7
10
10
20.0
80.0
10/7
16
18
68.7
31.3
3
5
33.3
66.7
2
2
	
100
28/7
27
29
51.9
48.1
7
7
85.7
14.3
4
4
75.0
25.0
8/8
11
18
54.5
45.5
2
2
"
100
0
0
——
—
30/8
1
16

100
2
6
"
100
0
0
— ™

llote: Total numbers (italics) captured in each weight class. BR: broods
carried; REL: broods recently released, oostegites still present.
It is likely that individuals >30 mg had been fertilized the pre-
ceeding fall and were able to set broods as soon as spring temperatures
triggered molting activity. The lag in the appearance of small gravid
females may be partly due to a lesser "inclination" to breed at lower
temperatures. But it becomes evident that overwintered prejuveniles, and
juveniles of low weight, grow to young adults in May and June, and then
begin to breed; and that their contribution to population recruitment is
relatively important, even though each individual may produce only a small
number of progeny.
Paris and Pitelka (1962) observed a very similar distribution of size
(age) classes in vulgare: early breeding was carried out mainly by 2-
and 3-year old females, and one-year old classes contributed increasingly
to recruitment as the season progressed.
DISCUSSION
There are obvious similarities between the Trachelipus population
at the orchard site and the Canadian population studied by McQueen (1976c).
The fate of each cohort, the periodicity and extent of breeding activity,
and the probable maximum life span, are virtually identical. Where there
are discrepancies, they may well be attributable to disparities in methodol-
ogy rather than to differences inherent in the populations.
A case in point can be found in the weight distribution of breeding
females and the determination of potential reproductive size in field
populations. McQueen (1976c) weighed captured animals directly, while
orchard Trachelipus were measured and their weights estimated. Since
154

-------
gravidity increases weight, weighed gravid females would be assigned to
heavier classes than warranted by size. As a result, the lower weight
limit for potential reproduction would appear to differ in the two
populations.
Survivorship characteristics of the two populations do not seem funda-
mentally different, temperatures around 20°C being the most favorable for
all weight classes. Considering that rearing methods were not identical,
it is not surprising that there are divergent results—e.g., in McQueen's
observations, a higher mortality of prejuveniles at 15 and 25°C and a lower
mortality of juveniles 15°C. If mortality patterns were Inherently
different in either population, the observed similarity in seasonal weight
distribution could only arise if other characteristics differed as well.
None of the pertinent data suggest this to be the case, although the number
of offspring recorded for Michigan Trachelipus was higher than that
observed in Canadian stock. The regression obtained by McQueen (1976c),
applied to the females observed in the present study, would underestimate
natality rates for both first and second broods. Here again, if both
studies had used either gravid or nonbroodlng weights, the regression
lines would probably have been quite similar.
McQueen's (1976c) sampling technique did not enumerate Individuals
weighing <5 mg, but the 0-10 mg class shown in his diagrams was equally as
prominent during May as the corresponding weight classes in the orchard
population. In both cases, the appearance of new recruits was sharply
delineated. Although growth rates in our laboratory stock were slightly
lower than the rates observed by McQueen, growth of the two field
populations seemed comparable, exhibiting similar patterns of cohort
advancement.
Clearly, temperature is a major factor regulating growth, mortality
and reproductive timing of Trachelipus. McQueen (1976c) addressed the
question of population limitation; unlike P_;_ spinicornis (McQueen and
Carnio 1974; McQueen 1976a, b), Trachelipus did not appear to be limited
by climatic conditions in the Ontario region, since the population expanded
over a three-year period. Comparable data are not available for the
Michigan population, since neither mark-recapture programs nor repeated
density estimates were attempted. The different nature and past history
of the sampling sites alone would make such comparisons speculative.
Overall, the populations behaved similarly. What the full -growth potential
of Trachelipus populations might be, would have to be investigated in
contrasting habitats where factors other than climatic conditions could
be quantified and compared. A periodically managed orchard with a relatively
bare floor stratum probably offers less protection and fewer resources
than, for instance, an undisturbed woodlot.
155

-------
CHAPTER IX
THE EFFECTS OF AZINPHOSMETHYL ON A POPULATION OF Trachelipus rathkei(Isopoda)
This Chapter reports the results of the detailed study of azlnphosmethyl
effects on an example soil/litter Invertebrate population. Trachelipus
(Tracheonlscus) rathkei (Brandt, 1833), the only lsopod species present at
the experimental site, was one of the organisms selected for intensive study
in both the field and the laboratory. McQueen (1976c) successfully con-
structed a simulation model of a Trachelipus population in Ontario, Canada,
and provided valuable information and guidelines for the present study. His
concern has been the problem of limitation in terrestrial isopods (reviewed
in McQueen 1976a, b), but he could not pinpoint a definite limiting factor
in the case of T. rathkei.
Although climate patterns appear to regulate some lsopod species (Sutton
1968; McQueen and Camio 1974; McQueen 1976a, b), annual temperature con-
ditions did not limit population growth of T. rathkei, leaving pathogens and
drought as possible alternatives (McQueen 1976c).
In disturbed environments, given that T. rathkei will move until its
microclimatic requirements are satisfied (White 1968), factors connected to
management by man seem to gain in importance. Trampling, for instance, can
play a major role in population reduction (Duffey 1975), and susceptibility
of the species to pesticides may curtail its ability to maintain viable
population levels. The effects of chemical stress due to azlnphosmethyl,
in an orchard system, are the subject of the present report.
THE FIELD SITE
The apple orchard where these investigations were carried out lies in
the vicinity of Grand Rapids, Michigan, in Kent County. Planted over 20
years ago, it had been left untended since then.
In spring of 1976, the trees were trimmed, understory brush was removed,
and the area was mown. Trimming has been repeated every spring; the ground
cover has been mown again two or three times every season, and the clippings
have been left on the ground.
A survey of the ground cover by the line-transect method (Cox 1974) was
carried out in 1976. Rather than determining species composition, "cover
type categories" were established and their importance values calculated.
Moss intermingled with litter, and moss-litter-broadleaf complexes, were
by far the most prevalent in the entire, experimental area.
Seven plots were originally delineated, one of them a control separated
from sprayed areas by an unmanaged buffer strip. Pit-trapping in 1976 had
yielded only few isopods in the control plot, probably because lsopod activity
was curtailed during this particularly dry summer. But in order to safe-
guard against possible low densities and ensuing difficulties in data analysis,
a second control plot was set up in early 1977. A total of eight plots
were than in existence; of these, four were sampled for isopod populations
throughout 1977-1978; plots 2 and 5 were treated with azlnphosmethyl, plots
7 and 8 were control. Plots 2, 5 and 7 were in their second year of use,
plot 8 in its first.
156

-------
The plot boundaries were delineated by aluminium core stock (fencing)
embedded in the ground to a depth of 8 cm, and protruding to a height of
20 cm. The fencing followed the contour of the land and was meant to
enhance channelling of runoff for collection of water samples at the bottom
of the slopes. However, not all plots were completely enclosed. Immigra-
tion of flightless invertebrates, across unfenced boundaries as well as
across the fence, should be kept in mind as a mechanism of population replen-
ishment, but was not quantitatively assessed.
Azinphosmethyl (50% wettable powder) was applied at the rate of 1.1 kg/ha
(1 lb a.i./acre), using a low-volume, low pressure tractor-drawn spray rig
similar to one developed by Howitt and Pshea (1965). Pesticide application
dates in 1976 and 1977 are given below:
1976: July 2, July 15, July 28, August 17.
1977: May 26, June 16, July 7, July 22, August 12.
METHODS AND MATERIALS
Boards were cut from weathered barn wood, to a uniform size of 35x13 cm
and 2 cm in thickness. They were positioned in the field such that the area
underneath them was fairly even, and were weighted down with stones. Plots
2, 5 and 7 each had 10 boards, plot 8 contained 20. In each plot, half were
placed under tree canopies and half out in the alleys between tree rows.
The first collections were made about two weeks after positioning the boards.
Animals were collected both from the underside of the boards and from
the debris and soil crumbs beneath. They were sexed and measured in the
laboratory and returned to the boards they came from.
Size measurements were done under a dissecting microscope with an ocular
micrometer. Head capsule width and .width of the last thoracic segment were
recorded. First and second lnstars are recognizable by absence of the seventh
thoracic segment. Third lnstars possess a seventh segment, but it Is rela-
tively narrower than in older stages. Occurrence of these young stages, and
of the breeding condition of adult females, were noted.
RESULTS
The position of the boards with respect to tree cover had a significant
effect on the number of individuals recovered (Fig. IX-1). T. rathkel
were more numerous under tree canopies on almost all dates.'
In plot 8, not all boards were sampled at all times. The mean number
of animals per board under trees was compared to the mean of those found in
alleys (over all sampling dates): variances being unequal, Behrens' (1929)
t-like test statistic and Welch's (1938) critical values were applied. With
P < 0.001, boards under trees harbored significantly more isopods than those
in alleys. In plot 7, replication being equal on all dates, a paired t-test
on mean number of animals per board per collection date again showed the
same results (P < 0.01). Corresponding data from sprayed plots were not
analyzed because of the low number of individuals collected through the year.
Tree shade with concurrent lower temperatures and higher humidity, as
well as less physical disturbance in the area around the trunk that cannot
be mown, may all have contributed to higher densities under the canopy cover.
SEX RATIOS IN THE FIELD
Young individuals cannot be sexed with certainty until the head capsule
157

-------
u Boards under tree canopies
11 Boards between trees
Ptof 7
10.6. 20.7. 9.8. 25.8 26.9.
PtotS
10.6. 20.7. 9.8 26.8. 26.5.
Collection dote
17.11.
25. >r
Figure IX-1. Mean number of T. rathkei collected from boards
placed underneath tree canopies and from those
in alleys between trees.
158

-------
^ TO A
0
$
8 soA-
So
-S? ¦
1	x\
o ° o
5 x
K »¦
£
—i	1	1—i—i	1—i—i i r
5 6. 7 7 8. 8 9. 11 f *
1977	1978
Collection date (month)
S.
Figure IX-2. Percent females captured, of the total number
of T. rathkei that could be sexed; exact col-
lection dates as in Fig. 1 Symbols:0= Plot 7,
X = Plot 8.
159

-------
measures about 1.1 mm in width. Fig. IX-2 shows the percent females, of
the total number of sexually recognizable animals captured, in plots 7 and
8. The data are not as variable as the percentages given by McQueen (1976c).
However, the recurrence of a sex ratio of close to 50% strongly supports
McQueen's suggestion that differential behaviour rather than differential
mortality, i.e. higher mortality of males (Brereton 1957), is the cause of
the observed sex ratios in the field.
EFFECT OF AZINPHOSMETHYL ON NUMBERS OF INDIVIDUALS
The mean number of animals collected per board and date are summarized
in Table 1. The isopod populations in sprayed plots were apparently dras-
tically reduced during the 1976 spray season and did not recover during 1977.
Paired t-tests were used for all comparisons between plots. The two
sprayed plots did differ significantly (P < 0.05). It should be noted that
the upper edge of plot 2 was not fenced in the usual way, but abutted to a
sparse hedgerow. The nearest boards were at least 6 meters away from it,
but immigration along that edge cannot be ruled out. Plot 5 was completely
fenced in.
Plots 7 and 8 also differed considerably (P < 0.01); the relatively
greater number of animals recovered from plot 8 may be attributable to the
fact that this second control plot had only been established in the spring
of 1977, while plot 7 was in its second year of use. Factors such as com-
paction by human trampling and mowing would not have come into play to the
same extent in these two plots.
Keeping in mind this past managerial history, the sprayed plots were
each tested against plot 7, the older control (plot 2: P < 0.01; plot 5:
P < 0.0i). Thus the mean number of individuals collected was significantly
smaller in the two sprayed plots.
Table IX-1. Mean number of Trachelipus rathkei per board on each collection
date
Date
Plot number



2
5
Z
8
24.5. 1977
0.3
0
1.8
5.1
10.6.
0.3
0
2.4
6.8
20.7.
0.1
0
2.3
5.0
9.8
0.2
0.2
5.6
7.2
26.8.
1.6
0.2
9.0
8.3
26.9.
*
*
11.6
17.6
3.11
1.7
~T.o
*
*
17.11.
0
0.6
~6.3
~5.3
7.4. 1978
3.7
2.4
4.8
10.0
25.4.
3.4
2.2
4.6
6.6
16.5.
1.3
0.6
3.5
6.0
* denotes that no samples were collected on these dates.
Plots 7 and 8 are control, plots 2 and 5 were sprayed; plots 2,5 and 7 con-
tained 10 boards each, plot 8 contained 20.
160

-------
EFFECT OF AZINPHOSMETHYL ON POPULATION SIZE STRUCTURE
Captures from plots 7 and 8 (controls) and 2 and 5 (sprayed) were com-
bined for a comparison of size structure within the populations. In his
model of a Canadian Trachellpus [Tracheonlscus] population, McQueen (1976c)
divided the population into three wieght classes: 0-5 mg (prejuveniles),
5-30 mg (juveniles), and > 30 mg (adults). In the present study, the animals
were measured only. Five size classes according to head width, and based
on life cycle observations of the psecies (to be published elsewhere) were
established:
Class 1: head width 0.5 to 0.7 mm; this class encompasses the first
three instars, approximately; progeny released by females between capture
and return to the field were not included.
Class 2: head width 0.75 to 1.15 mm: sexes cannot be recognized with
certainty In these animals.
Class 3: head width 1.2 to 1.4 mm; sexes are easily distinguishable,
but the females do not contribute the bulk of the breeders (21% of the females
collected in breeding condition had a head width of 1.3 to 1.4 mm).
Class 4: head width 1.45 to 1.8 mm; this class contains the majority
of reproductive adults.
Class 5: head width 1.85 to 2.25 mm; such specimens were found only
rarely throughout the season. This did not appear to be a result of the col-
lection technique used. The orchard population may characteristically lack
these large individuals, although they are relatively common in woodlot
population mid-Michigan.
Fig. IX-3A shows the combined data from plots 7 and 8 (controls).
Breeding starts In late May, and evidence of brood release is found in the
increasing Importance of size class 1 from July through August. As these
young stages grow, they graduate to classes 2 and 3 which provide the bulk
of the juvenile population in April and May of the following year. McQueen
(1976c) and McQueen and Carnio (1974) estimated that prejuveniles-roughly
equivalent to classes 1 and 2 combined - were seldom able to overwinter In
Canadian populations of T. rathkei and Porcellio splnlcornls. In the orchard
population, over-winter mortality seems much less drastic; the survivors of
late summer broods (gravid females were still found in late August) mature
the following year and contribute to classes 3 and 4 during summer growth.
In contrast to this population continuity in control plots, Fig. IX-3B
demonstrates that the sprayed plots, with rare exceptions, harbor an impover-
ished population of only large-sized individuals. Breeding females were
found in May through July, but virtually none of the progeny could be located.
Since the last spray application occurred in mid-August brood release in late
summer and early fall apparently contributed a small number of young, which
appear as classes 2 and 3 in spring of 1978.
During the spray season, however, the majority of newly hatched juveniles
seem to have been eliminated by the presence of azlnphosmethyl in their
habitat. Toxicity experiments in the laboratory strongly support this assump-
tion (see Chapter VIII). Briefly, adults are not susceptible to concentra-
tions of much less than 10 ppm, but individuals belonging to class 1 are
killed by exposure to contaminated soil at concentrations of 1 to 2 ppm.
Soil samples from the experimental site, analyzed by gas-liquid chromatography
generally do not exceed levels of 0.5 ppm; litter on the other hand frequently
contains up to 35 ppm (Dr. Robert Kon, pers. comm.). Thus the impact of the
pesticide would be greatest at the time of peak reproduction, during mid-
summer; but a skeleton population of recruits survives to carry over Into the
next year.
161

-------
60-
bO-
3 ?%¦'' Si "A" & ' «
(238!
R-
W
-ffl
(2Zi
Figure IX-3(A,B) Size class distribution, in percent of total number cap-
tured per date, of T. vathkei populations throughout one
year. A: control plots 7 and 8 combined. B: sprayed
plots 2 and 5 combined. In parentheses: number of ani-
mals collected. See text for description of size classes.
Errata in fig. 3B. Plot-number above the first column
should be read 4 instead of 6. At 07.04.78 61 animals
were collected.
DISCUSSION
Certain aspects of this study should be kept in mind in the discussion
of results. One, that populations in sprayed plots had already been subjected
to azinphosmethyl during the 1976 season. Data on their initial pre-spray
state are not available, but there seems no good reason to assume that the
population throughout the orchard (then not managed and not divided into plots)
would have differed greatly, in density or structure, from area to area.
And second, that the success of collecting individuals, especially very small
ones, probably decreases as a function of decreasing density. No claim can
be made that all animals present under the boards were actually located and
collected.
The significant differences between the control plots might be explained
by the different length of time each had been in use. The differences be-
tween the two sprayed plots are based on very few individuals, which may
lead to an exaggerated view of pesticide effects because of low capture ef-
ficiency. But the absence of juveniles and even subadults (class 3) in sprayed
plots is striking. The maximum life span of the species is two years in this
climatic region (McQueen 1976c), and spring of 1978 represents the beginning
of the third consecutive year of pesticide application. At that time, the
reduced population in sprayed plots must represent survivors of the cohorts
initiated in summer of 1977 (classes 2, 3 and 4) and possibly also in 1976
(large individuals of class 5).
162

-------
It is conceivable that such a low-density population can maintain it-
self under stress of azinphosmethyl application,.aided by heterogeneous dis-
tribution of the pesticide on the orchard floor, washout by rain, and sur-
viving late-summer broods. Whether a lower limit of population density is
ever reached beyond which population persistence is not possible, remains,
unanswered until further data throughout consecutive years of pesticide
application can be obtained.
163

-------
CHAPTER X
A STUDY AND MODEL OF AZINPHOSMETHYL EFFECTS ON A POPULATION OF
T. rathkei
INTRODUCTION
An extensive investigation of the effects of azlnphosmethyl spray ap-
plication in a Michigan orchard was begun in 1976. The field sampling pro-
gram furnished the basis for a model of pesticide distribution and movement
and also provided information on the impact of azphosmethyl on selected
members of the invertebrate fauna of soil and litter.
It became obvious that one of the non-target organisms most affected
was the lsopod Trachelipus rathkei (Brandt), and the species was chosen for
a detailed demographic model - an organism submodel intimately linked to the
orchard-pesticde model mentioned above.
Ecoblological parameters, greatly enhanced by recent work on the demo-
graphy of T. rathkei in Ontario, Canada (McQueen, 1976b), were obtained in
both laboratory and field and are described in detail in Chapter VIII and
in Snider and Shaddy (1980). In addition, successful linkage of the popu-
lation model to the pesticide dynamics model necessitated knowledge of the
susceptibility of the species to azlnphosmethyl. Preliminary field observa-
tions to that effect were reported by Snider (1979) and in Chapter IX: sub-
stantiating data were obtained in laboratory experiments as well as through
additional field sampling, and are included in this report.
T. rathkei Biology
A Summary of the Life Cycle
The biology of the species will be reiterated briefly, based on results
from a two-year investigation carried out partly in the laboratory and partly
In the field. Extensive information on the ecobiology of T. rathkei has re-
cently been given by McQueen (1976c). In addition, McQueen and Carnio (1974),
Carnlo (1974), and McQueen (1976a, b) have compiled reviews encompassing
physiology, behavior, and demographic patterns of terrestrial isopod species.
Unless otherwise stated, the following summary is based on data presented by
Snider (1979) and Snider and Shaddy (1980). Details of rearing and sampling
techniques may be found those publications (Chapters VIII and IX of the report).
As did McQueen (1976c), weight classes were defined for T. rathkei females
as a means of classifying population data. The two classes containing pre-
juveniles extend from a birth weight < 0.5 to 1.34 mg, and 1.35 to 5.0 mg.
Individuals in the 5.01-12.0 mg class, termed juveniles, are not yet repro-
ductive. Breeding is possible, however, for at least some of the females
weighing _> 12.0 mg, and 12.01-30.0 mg class was defined as young adults.
Those termed adult 1 weighted 30.01-50.0 mg; adult 2, which are nearing or
in their second year of life, are brought together in a 50.0-80.0 mg class.
Eggs develop and hatch in the female's brood pouch, from which the young
are released over a period of a few hours to about two days. Development
164

-------
time in the pouch is strongly influenced by temperature (51 days at 15.6°, 31
days at 21°, and 18 days at 26.7°C). Within 24 hours, the newly emerged
young undergo their first molt, then begin to feed.
Through a series of molts, growth is achieved at temperature-dependent
rates. By the time a weight of 5 mg is attained, males are distinguishable
from females by the presence of a faint keel on the carpus of the seventh
pair of legs. Intervals between ecdyses are short at first, but lengthen
and become more Irregular as the animals age; molting frequency also Increases
with rising temperature (unpublished data). As a measure of weight gain
versus time and temperature, Table X-l shows the number of days needed for
T. rathkei to grow through each of the weight classes.
Table X-l. Number of days needed for T. rathkei to grow through each weight
class at three temperatures. In parentheses: final mean weights
± S-E*
Weight Class, mg
T°C	0.5-1.34	1.35-5.0	30.01-50.0	50.01-80.0
15.6	70(1.34+0.11)	97(5+.4)	170(50.0+1.5)	260*(65.0+1.1)
2.0 39(1.34+0.10) 40(5+.4) 92(50.0+1.0) 16*(71.0+2.3)
26.7	23(1.39+0.04)	20(5+.2)	76(50.0+1.2)	138(80.0+1.5)
^Durations estimated linearly, based on the number of days needed to reach
65.0 and 71.0 mg; observations had to be discontinued at that point.
In the field, growth can be followed by analyzing the weight distribution
of the population through the year. In April and May, up to 35 percent of
the population consists of overwintered prejuvenlles, probably late-born
members of the recruits of the preceding year (cohort I). Older individuals
of cohort I (released in June and early July) have reached juvenile and adult
weights.
The bulk of cohort II arises from adults of cohort I, and from juveniles
of cohort I after they have reached reproductive size during summer growth.
During and after the breeding period, surviving cohort I females advance to
weights exceeding 50 mg, overwinter, and enter their second reproductive
period in May of the following year. By late summer, all of them seem to die
out, at a time when the overwhelming bulk of the population consists of pre-
juvenlles and juveniles of cohort III. By the end of autumn, cohort III has
graduated to large prejuvenlles, juveniles, and adults; cohort II has replen-
ished the > 50 mg class and, after overwintering and breeding, will die the
following summer.
Laboratory-raised Individuals do not breed readily at low weights, but
field-collected females of 16-17 mg mate and breed successfully at constant
temperatures ranging from 15.6° to 26.7°C. Heavier females ( > 30 mg),
brought in from the field in late winter, invariably breed within a few weeks
without having been mated. Their ability to store sperm, observed previously
by Heeley (1941), thus permits these individuals to reproduce as soon as warming
temperatures induce molting in the field.
Having released their first brood, females undergo a post-parturient
molt during which a second marsuplum may be developed. The production of
165

-------
second broods appears to be dependent on the weight of the female as well as
on temperature: small females tend to breed only once, but Increased tem-
peratures induce a higher Incidence of second and third broods in individuals
of all weight classes. The rest interval between broods ranges from 7 to
17 days, with mean durations of 16.9, 17.2, and 13.A days at respective tem-
peratures of 15.6°C, 21°C and 26.7°C.
The number of young produced is linearly correlated to female weight,
but first broods are significantly larger than second broods even though
the animals gain size during the intervening molt. (First brood: young =
0.86 wt - 6.43; r = 0.77. Second brood: young = 0.54 wt + 1.97; r = 0.50).
Counts of number of progeny released yield estimates of natality, regardless
of the total number of eggs carried by the female (fecundity). Occasionally,
entire broods consist of non-developing eggs which are eventually expelled
by the female; they contribute to what may be termed "egg mortality." Con-
sidering only female weight, the 30-50 mg class bred most successfully at
all temperatures; as to the effect of temperature, 21°C was most favorable
(no more than 10% faulty broods, as compared to up to 20% at 15.6°C and up
to 25% at 26.7°C).
In the field two peaks of gravidity are discernible within the yearly
breeding period, with 86% and 60% of all potential breeders carrying eggs
(percentages based on all females _> 12 mg). If females with empty marsupia
are included, these peaks rise to 93% and 97%, and are no longer as clearly
separated. Individuals exceeding 30 mg all become gravid as soon as tempera-
ture allows (mld-to late May), and breed a second time. Field evidence also
shows that very small females, (>^ 12 mg), are capable of breeding, but become
gravid later than heavier individuals. Within the 12-30 mg class, breeding
frequency increases gradually through time as juveniles become young repro-
ductive- adults during summer growth. Breeding activity in this weight class
peaks in late July, about one month after the first reproductive peak of
heavier individuals. There are Indications that these small females breed
only once. But they constitute a large proportion of the gravid females
late in the season (older individuals incur heavy mortality losses at that
time), and thus contribute significantly to population recruitment, even
through each individual may produce only a small number of progeny. By the
end of August, no more gravid Individuals are found in the field.
The most extensive data on the mortality patterns of T. rathkei were
gathered by McQueen (1976c) who grew prejuveniles, juveniles and adults at
a wide range of temperatures. Our own data were compiled from observations
on isolated individuals reared through six weight classes, the limits of
which were 0.5-1.34 mg, 1.35-5.0 mg, 12-30 mg, 30-50 mg and 50-80 mg. Com-
parison between these data and McQueen's (1976c) is difficult because culture
conditions and methodology were not identical; but where there Is agreement,
survivorship trends for the species become apparent.
The most favorable temperature for survival seems to lie around 21°C
(both juveniles and adults in McQueen's observations survived best at 20°C).
Higher temperatures become increasingly deleterious for all weight classes;
at 35°C T. rathkei are unable to survive for any significant length of time
(McQueen, 1976 c). Temperatures below 20°C also become unfavorable, prejuveniles
and juveniles both incurring about 20% mortality at 15.6°C (Table X-2).
If one were to accumulate mortality from the time T. rathkei are born
until they reach a weight of 50 mg, a little over 28% would have died at 21°C,
38% at 26.7°C, and 54% at 15.6°C. Beyond a weight of 50 mg, mortality in-
creases drastically (Table X-2). The very few females ever collected in the
166

-------
field that exceeded 80 mg indicate that virtually all of these old Individuals
die before attaining 80 mg.
Females also incur mortality that appear specifically connected to
breeding activity. Of 313 females observed through one to three broods, 21
Table X-2. Percent mortality, within four weight groups, of non-breeding
T. rathkei females at three constant temperatures.
Weight Class, mg
T°C	0.5-5.0	5.01-30.0	30.01-50.0	50.01-80.0
15.6	20.0	19.0	29.4	?
21.0 10.0 8.7 12.1 40.0
26.7	7.0	11.5	14.0	48.9
Table X-3. Percent mortality of T. rathkei females associated with each brood
within one breeding period (three consecutive broods occurred at
26.7°C only). Assignment to weight classes is based on weights
just prior to each brood.
Wt. group
15.68C
BR.	BR.
1	2
21°C
BR. BR.
1 2
BR.
1
26.7°C
BR. BR.
2 3
<30
30-50
>50
18.2
9.1
18.2
0.0
8.3
20.0
0.0
0.0
5.6
0.0
10.0
0.0
3.7
0.0
11.1
5.6
8.3
10.0
0.0
died either during gravidity or during the molt following brood release. The
most favorable temperature was again 21°C, and females in the 30-50 mg range
survived best during the breeding cycle (Table X-3). At 15.6°C all deaths
were linked to non-developing broods. Overall, only 16% of all brooding
mortality was due to unsuccessful post-parturient molts.
The Trachellpus rathkei model
The information discussed above was used to develop and parameterize a
model to simulate the population dynamics of T. rathkei. The model tracks
only females as they pass through seven weight classes, and adjusts their
numbers by mortality as environmental conditions dictate. Individuals in
sexually mature weight classes pass through breeding stages and produce off-
spring according to rates determined by their size and by soil temperature.
Figure X-l shows diagramatically the processes affecting individuals of a
single weight class.
Passage of Individuals through a weight class is modeled as a distributed
167

-------
delay process (see Manetsch, 1976). Each weight class is subdivided into
a set of smaller weight ranges, and individuals advance from one subclass
(or "bin") to the next at a rate proportional to the growth rate. The distri-
buted delay processes are linked so that individuals leaving the last bin of
a weight class enter the first bin of the following weight class.
Growth rate is a function of temperature and is determined by a two-
stage degree-day growth model. First, the number of degree-days spent in
each weight class, using an appropriate degree-day base temperature, was
determined in the laboratory at various temperature, (Table X-l). This in-
formation, and daily temperature, are used by the model to calculate the
proportion of individuals advancing from each bin during the current time-
step, such that mean time spent in a given weight class would allow accumula-
tion of the required number of degree-days at the current temperature.
Table X-4 shows the degree-day span and base temperature for each weight
class (breeding and non-breeding females were found to have different growth
rates, so are modeled separately).
progression from previous
breeding stoge
Soil Temperature
Graduation to
next
weight class
Graduation from
previous weight doss
Cumulative
Natural Mortality
progression to next breeding
stage
natural
mortality
pesticide
induced
mortality
Weight class
subpopulotion
Pesticide
Exposure
Level
Cumulative
pesticide -
induced
mortality
Environmental
Pesticide
Residue
Levels
Figure X-l. Conceptual model showing processes affecting a life stage
of T. rathkei. Solid lines represent flow of T. rathkei;
dasEed lines, flow of information. Valves control rate of
flow, based on incoming information.
168

-------
Table X-4. Degree-day-based weight class durations for breeding and non-
breeding T. rathkei females.
Base	Degree-Days	Order of
Weight Class	Temperature°C to Graduate Distributed Delay
.5-1.34
11°C
319
3 bins
1.35-5.0
11°C
425
4 bins
5-12mg
8°C
432
4 bins
12-30mg (non-breeding)
12°C
560
4 bins
12-30mg (breeding)
9°C
2016
4 bins
30-50mg (non-breeding)
8°C
1303
9 bins
30-50mg (breeding)
7°C
2784
9 bins
50-80mg (non-breeding)
2"C
3147
8 bins
50-80mg (breeding)
4°C
5680
8 bins
Also shown is the number of bins within each weight class (known as the "order"
of the distributed delay).
Individuals in sexually mature weight classes (the last four) move not
only to bins corresponding to increased weight but also advance through the
breeding process (shown as downward movement in Figure X-2). Breeding begins
upon accumulation of a specified number of degree-days for each of the last
four weight classes. First and second breeding periods are treated as dis-
tributed delays, with a mean duration determined from a degree-day model
(similar to the one used for advancement through weight bins). Following
brood release, breeders enter a fixed-duration rest period (16 days); some
will breed again, while others return to the non-breeding population for the
remainder of the season.
Figure X-3 shows in greater detail the bins used to track individuals
in the population. Movement to the right denotes weight gain, while move-
ment downward (in the mature weight classes) represents progress through the
breeding process (other factors suppressed in this figure Tor clarity). The
numbers in each box represent the number of bins down by the number across
in that weight class and breeding stage. Thus the position of an individual
in Figure X-3 indicates not only its weight, but also its breeding status
(carrying first brood, resting after first brood, etc.). Note that a propor-
tion of the individuals completing the first brood return to the top, non-
breeding, row without a second brood. Once a given degree-day accumulation
is exceeded, no further breeding is initiated.
Graduation from the last vertical bin of a breeding stage is followed
by a brood molt, with brood size determined by 2 regression equation. Broods
lost to "egg mortality" are subtracted from the egg population at this time.
The only aspect of the model left to be explained is mortality. Three
sources of mortality (other than egg mortality) are considered by the model;
breeding mortality, pesticide-induced mortality and any other "natural"
mortality. Pesticide-induced mortality deserves special treatment and so is
covered in a later section. Mortality for breeding females, as displayed
in Table X-3 based upon laboratory data, is converted to a table of rates
of mortality at three temperatures, and the model interpolates within the
169

-------
£
a
I
Soil Temperature
breeding rate
breeding rate
Hon-breeding
2nd breeding
period
1st rest
period
Figure X-2. Expanded diagram highlighting the breeding process
and the factors affecting it.
170

-------
TRBIN
TRBR01
TRDRST1
TRBRD2
PREJUV	JUV 1	JUV2	ADULT 1	AOULT 2	OLD
Figure X-3. Expanded view of the life stages of T. rathkei as tracked in
the model. Each box indicates the number of bins it contains.
Individuals may also leave TRBRST 1 and TRBRST 2 (post-
parturient "resting" stages) for the top row of non-breeding
classes, although the arrows were omitted to simplify the
diagram.

-------
table to determine mortalities at the prevailing temperature as each female
completes a brood.
Natural mortality rates determined as a function of temperature under
our laboratory conditions (as shown in Table X-2) are not adequate to de-
scribe the mortalities observed in the field. That is, if the values in Table
X-2 are converted to a table of dally rates at three temperatures, and these
values are used to estimate daily rates at other temperatures, the model will
allow the population to grow without bound. Some form of density - dependent
limitation factor must be introduced in order to allow reasonable dynamics
for (unsprayed) populations. Our laboratory and field studies did not iden-
tify the limiting processes, but did help to establish approximate upper
limits on population densities. However isopod sensitivity to desiccation
suggested that availability of suitable moist microhabitat in soil might act
as a limiting factor, and a crude model embodying this assumption was devel-
oped. This model allows for Increased mortalities under dry field conditions,
particularly when the population density is high.
The model postulates a typical high density biomass of T. rathkei which
can be supported, B^, and a natural mortality which increases as the total
biomass approaches or exceeds that number. The form selected is:
Mort = mort^ab (1 + 0.10 x F-d), where mort at is (corrected) daily
natural mortality, mortj^ is the laboratory-based ?emperature-dependent daily
natural mortality, and
F Btotal
dd B, . . F . „
hi moist	2
where B , is current total T. rathkei biomass (g/m )
total	— 	
B, . is a parameter representing a typical high density of T. rathkei biomass
hi £	2
(g/m ) (currently set at 568 g/m ), and
F . = 1 - (1 + 18 • SM) (e~18 SM) where SM is soil moisture (weight/weight)
moist
in the top 2 cm of soil.
Several observations should be made about these factors:
(1)	with soil moisture ^ .3, F has little effect, and at low
densities, mort ~ mort. , , while at 1 „ = B, ., mort _ " mort, , x 1.1
' nat	lab	total hi	nat	lab
(2)	the soil moisture effect is also weak: with soil-moisture at 0.15,
mort^k is multiplied by only about 1.13 if BtQta^ = B^.
(3)	only when B ^ vastly exceeds B^ and/or soil moisture is extremely
low will mort be mucn nigher than mort. , , and these conditions, unexplored
in the lab, would be expected to have sucn an effect.
Obviously, these soil moisture - and density-dependent factors have been
chosen based only on qualitative observations, and extensive study would be
needed before any confidence could be placed in their numerical accuracy.
However, they do serve to stabilize the simulated populations and produce
reasonable behavior, so they are here adopted in the absence of better in-
formation.
Since the male: female ratio of T. rathkei is approximately 1:1 (Snider
and Shaddy, 1980), total population at any time is determined by multiplying
the female population by a factor of two (recall that the model tracks only
females).
172

-------
AZINPHOSMETHYL TOXICITY
Materials and Methods
T. rathkel were collected in unsprayed areas of the orchard as well as in
woodlots around the Michigan State University campus, and were reared in stock
cultures for varying periods of time. Very young stages were obtained from
isolated gravid females, and the date of their release from the brood pouch
was recorded. All cultures were kept at constant temperatures of 15.6°, 21°
and 26.7°C, and animals from them were used in experimental runs at these
three temperatures. Toxicity tests were arbitrarily terminated after six or
seven days, because at that time fungal contamination often began to hinder
movement of the animals over the substrate.
Individuals to be tested were selected according to weight, with the
exception of second instars which were 2 to 3 days of age. Initially a num-
ber of animals were weighed Individually and those exceeding the chosen weight
limits were rejected. Later on, the good correlation between weight and size,
with an r of 0.97 (Snider and Shaddy, 1980) allowed the more rapid technique
of selecting Individuals by measurement under a dissecting microscope fitted
with an ocular micrometer.
In each weight class, and per dosage level, 10 replicate test containers
were set up with 2 animals per container (20 animals per test run). The
number of times each test run was replicated varied, and will be detailed
where needed.
Two different substrates were used for toxicity experiments:
1)	Soil collected from unsprayed areas of the orchard: pH 5.5 to 5.7,
3.3 to 3.7 percent organic matter, classified as Marlette silt loam with 22
to 25 percent clay (Pregitzer and Stacey, 1973). It was shaken free from
roots, sieved through a 2 mm mesh screen, and dried at 50°C in a vacuum oven.
Azlnphosmethyl stock solutions (100 ppm active ingredient) were prepared
by adding 100 mg Guthion (50% W.P.) to 500 ml distilled-deionized water.
Aliquots of stock in the ratio 5:1 (ml stock: desired ppm in soil) diluted
to 100 ml, were used to make up all spiking solutions. Portions of dry soil
and 5 ml of the appropriate solution were alternately added to a mixing jar
in constant agitation. The final product usually consisted of 200 g soil
and 40 ml solution, to yield a moisture content of 20 percent - adequate to
ensure survival of T. rathkei at all temperatures.
Small portions of prepared soil were then pressed into cylindrical plastic
jars (25 mm diameter) such that no cracks were left in which the animals
could hide. This allowed observation directly through the clear plastic lid,
and major moisture losses could be avoided.
2)	Litter: surface litter was raked up in one of the sprayed orchard
plots, always in the same general area located in an alleyway between trees.
It was chopped into pieces less than 10 mm in length, then mixed and tumbled
thoroughly to reduce the effect of patchy pesticide distribution in the field.
Litter from control plots, processed in the same way, was used for control
runs.
Since all orchard litter was very dry when collected, each batch was
brought up to 200 percent moisture based on litter weight. At this moisture
level, animals of all sizes survived well. Once moistened, small portions
of litter were loosely pressed into plastic jars (25 mm diameter). Two
animals were introduced into each jar; the minimum total number of individuals
per test run was 20, but depending on availability up to 80 individuals were
used per series and weight class.
173

-------
All litter experiments were run at a constant temperature of 21°C.
Samples of each batch of field litter were set aside at the time the
tests were set up, and were analyzed for total azlnphosmethyl residues. Un-
fortunately the sample from one of the collection dates was lost, and pesti-
cide levels had to be estimated by bracketing the missing value with known
values from earlier and later dates. Table X-5 gives details of litter col-
lections and analyses.
In experiments with litter, the animals burrowed into the substrate and
the smaller sizes especially were difficult to locate without severe distur-
bance and moisture loss. Therefore only total mortality after a six-day
incubation period was recorded; the litter could then be spread out on paper
and both live and dead animals could be recovered.
The presence of fecal pellets bore witness to the fact that both soil
and litter were ingested during these experiments. Contact and oral toxicity
could therefore not be kept apart, but the distinction seems unnecessary if
the results are viewed as indications of possible field-effects.
Table X-5. Field-collected litter: results of analyses for total azlnphosmethyl
residues, and relationship of collection date to the last previous
spray application date.
Collection date 1978 July 25 Aug 11 Aug 23 Sept 10
Days since last spray	4	2	14	30
ppm azlnphosmethyl	26.7 25.7 11.0*)	3.1
*) By using the ppm values from the two collections bracketing August 23, the
missing value was estimated: Degradation rate x = ^	 Where 30 is
2577
the number of days be^een August 11 and September 10. The unknown
concentration y = (x) (25.7), where 12 is the number of days elapsed
since August 11.
RESULTS OF EXPERIMENTS
Soil Toxicity:
Daily records of deaths were kept for use in the Trachellpus model, but
Table X-6 only gives cumulative percentages per dosage, weight class and
temperature, so that an overall view of azlnphosmethyl effects might be ob-
tained. (For survivorship on a daily basis see examples given in Table X-7).
An inverse relationship between weight and mortality became clear at
all temperatures. While second lnstars experience up to 37% mortality at
1.0 ppm (26.6°C), animals weighing > 50 mg required dosages in excess of
10 ppm to be appreciably affected (Table X-6). In small animals, the re-
lation between rising temperature and increasing mortality could also be
demonstrated. This effect was expected to hold true for all weight classes,
but appeared masked by highly variable susceptibility of individual test
animals. Given the large number of individuals that needed to be handled
to obtain even these few point estimates of azlnphosmethyl effects, some
sources of error were disregarded. Less variable results could probably have
been obtained if exact weights had been used (rather than estimates from
size), and if the physiological state of the animals had been more closely
174

-------
monitored. The latter consideration requires observation of feeding activity
to exclude those animals which may be fasting prior to an impending molt;
and it requires a larger number of replicates so that those that molt during
an experimental run can be discarded. Repeated exposure of survivors, as
well as monitoring of changing pesticide levels (as a function of time and
temperature) would further enhance the accuracy of these experiments.
Table X-6. Cumulative mortality percentages for T. rathkei reared on con-
taminated soil for 7 days. Data corrected for control by Abbott's
formula. *) Indicates that the values are means obtained from
at least 2 replicate runs (containing 20 animals each).
AZINPHOSMETHYL, ppm
Size
T°C	class	1.0 2.0 3.0 5.0 10.0 20.0 100

Instar
II
29.9*
6.5*
55.8*
94.8*
94.8
100

2-3 mg
0.0

32.9*
0.0
100*

15.6
15-20
mg



26.3
31.6


30-40
mg
5.0



15.0*
26.7 90.0

50
mg




5.0
20.0*

¦Instar
II
20.0*
62.5*
81.7*
86.7*
100


2-3 mg
0.0

70.3
16.7



15-20 mg



10.0
32.5*

21.0
30-40
< 50
mg
mg



5.0
5.0
5.0
100
5.0

Instar
II
36.8*
82.1*
94.9*
100*
100*


Instar
III
12.8

74.4
94.9
100


2-3
mg
0.0

38.9
55.6
100


5
mg
0.0*

0.0
60.0*
90.0*

26.7
10 mg
15-20 mg
0.0


36.8
100
55.0


30-40 mg
3.4*


20.0
7.2*
67.8*

50 mg



5.9*
0.0
45.0
175

-------
Table X-7. Dally mortality records (cumulated percent) for II. instar T.
rathkei reared on azinphosmethyl-contaminated soil at three temperatures
(20 animals per pesticide level and temp).
T°C
Day
after
start
1.0
3.0
5.0
10.0
CON

1
0
0
15
10
0

2
5
5
35
45
0

3
5
10
65
70
0
15.6
4
5
10
75
70
0

5
6
5
5
30
45
80
80
90
90
0
0

7
5
55
95
95
0

1
0
10
15
5
0

2
0
25
40
35
0

3
5
40
60
55
0
21.0
4
15
50
85
80
0

5
20
65
90
90
0

6
20
75
95
95
0

7
25
80
95
100
0

1
0
20
30
45
0

2
10
55
70
60
0

3
15
75
85
65
0
26.7
4
15
80
95
100
0

5
15
90
100

0

6
20
95


0

7
25
95


0
176

-------
Litter Toxicity:
Table X-8 shows mortality percentages for Trachelipus of various weight
classes, confined for six days on field-collected litter, at a temperature
of 21°C.
The largest size class (mean wt + S.E. = 62.3 + 0.98), kept on litter
collected 2 days after spraying, experienced greater control than pesticide
mortality. Of all weight classes, animals that large have the greatest
natural mortality rate (McQueen, 1976c; Snider and Shaddy, 1980); in the
field, additional stress due to azinphosmethyl may not have any significant
impact on survival. This class was then not further tested; neither were
30-40 mg animals tested after a 5% mortality had been recorded at an estimated
level of 11.0 ppm azinphosmethyl.
Within each weight class, the relationship between pesticide level and
mortality was quite clear, especially for the highly susceptible small stages.
Of particular interest Is the fact that second instars still experienced
over 50% mortality 14 days after the last spray event in the field, while
animals weighing 2 to 3 mg were by that time removed from drastic effects.
Even a full month after spraying had ceased, the contaminated litter was still
toxic to the smallest prejuveniles.
PESTICIDE EFFECTS MODEL
The fate model described in Chapters I-IV calculates a dally status of
pesticide residues in the various orchard compartments. This information Is
used by models of various orchard-floor-dwelling organisms to calculate the
effects of the pesticide exposure on their populations. A new model has been
developed which allows estimation of the effects of long-term, rapidly
changing exposures on various organisms. This model, and its application to
T. rathkei, are discussed next.
General Form
Mortality due to long-term toxicant exposure is a very complex process.
Toxicant exposure that might cause 40% mortality In one day will not cause
an additional 40% mortality (of either the original population or the survivors)
in each subsequent day; nor can the exposure level be expected to remain con-
stant over long time periods. A useful model must calculate new mortality
based on both current exposure and past history of exposure (for example,
if the genetically most susceptible Individuals have already been killed,
a new exposure will have less effect). The problem is to capture the most
essential aspects of the population's exposure history in a reasonable number
of variables for a model to track. The model presented here uses a single,
natural, state variable, C^*, for each size class i, to integrate all of the
information about past exposure history that the model will use in calculating
further mortality. It is calculated using relatively simple, straightforward
assumptions about the temporal response of a population to a toxicant. Figure
3 shows a conceptual model for the effects of the pesticide on a single
weight class of the population. (Of course, for a different organism, strati-
fiction by some characteristic other than weight can be used.)
It should be noted that the laboratory techniques employed (i.e., no use
of radiolabeled compounds) did not allow for determination of internal pesti-
cide residues within the isopod population. If such determinations had been
177

-------
Table X-8.
Cumulative percent mortality, over 6 days, of Trachelipus of various weights exposed to field-
collected litter contaminated with Guthion (21°C constant temperature), (n = total number of
animals per run, distributed as 2 per container).
Weight of test animals (mg)
Days	Total	0.5	2-3	15-20	20-30	30-40	> 50
after	residue 	
last spray	ppm	Gu Con	Gu Con Gu Con	Gu Con Gu Con Gu Con
14
30
26.7
n=
25.7
n=
11.0
n=
3.1
n=
93.3 0.0
60 40
100
80
5.0
40
52.5	5.0
80	40
18.3	6.7
60	30
80.0	2.5
40	40
5.0	0.0
40	20
6.7	3.3
60	30
16.7
30
2.5
40
0.0
30
5.0 0.0
0.0
30
16.7 2.5
60 40
30.0	0.0
20	20
5.0	0.0
40	20
6.7 10.0
30 20

-------
possible, then a model based on a direct relationship between internal pesti-
cide concentration (or concentration in some tissue, for example) and cumu-
lative mortality might have been found, and all of the temporal dynamics
could have been included in a compartment-type model for uptake, elimipation,
and breakdown of the pesticide within the organism (see Goodman, et al. 1977).
Instead, the clearly evident lag between Initial pesticide exposure (or ex-
posure to Increased levels) and the resulting mortality must be represented
directly. Examination of the laboratory toxicity results under various con-
ditions shows a general trend of increasing mortality, tapering off as time
approaches 7 days. A useful model must at least perform well in calculating
the proper total pesticide-induced mortality given an exposure to a constant
level of pesticide over a long time period. The rate at which that mortality
is expected can be approximated by a fairly simple model, so long as the sum
of the daily mortalities approaches the proper cumulative value relatively
closely. Thus, a simple expression for incorporating the time dependency
has been chosen, illustrated in terms of soil exposure, of the form:
Ci,s(k) = A(MX	(k-1)). S(k-D) + (1-A) c± s(k-l),k > 1 (1)
where : C. (k) represents the cumulative mortality, in probits, for the
' fh
1— size class at time (day) k, due only to exposure to pesticide
in soil,
CONC(k-l) represents the concentration of pesticide in soil (ppm)
at time k-1,
TEMP (k-1) is the estimated soil temperature (°C) at time k-1, at
at a depth of 1 cm,
is a function wich specifies the cumulative mortality, in probits,
that a population of size-class 1 would suffer if exposed to CONC
(k-1) for many (>>7) days, (i.e., the long-term cumulative mortality
under those constant conditions), and
A is a real number in the range 0
-------
K IN DAYS
Figure X-4. Time response of C- s(k). Vertical axis shows the proportion of the distance between
C. (0) and m. which C. (k) has moved by day k, for various choices of A, the daily
ctafige rate. 1	1,s

-------
is independent of any other sources of mortality, which are lumped together
as "natural" mortality and extracted daily in another portion of this model,
depending on temperature, size class, and breeding condition.
To determine an appropriate value for A and to determine the m^ as
functions of pesticide concentration and temperature, an additional modifi-
cation will be made to the m. functions in this case: because very few or
no lab experiments were run For certain size classes, and because the data
are quite noisy, the discrete functions will be replaced by a single
function m which depends on size as well as on pesticide concentration and
temperature. Once m is fitted, the m.'s a£g easily calculated by substituting
the average size of an individual in Che i— size class into m. So long
as a relatively "smooth" relationship exists between size and susceptibility
to the pesticide, a good fit to m^'s can be produced in this manner. (For
some organisms or toxicants, the m^'s might be quite different from one
lifestage to the next, so separate fits of the m^'s would be mandatory.)
To allow for this size-continuous m function, equation (1) is rewritten
38 •	\
c, (k) - A» max/m(CONC, TEMP, SIZE.),C. (k-l))+ (l-A)C, (k-1), k>l (2)
1,8	\	1 1»S /	S
The use of laboratory data to fit the m function will be described in a later
section.
Two other factors must be taken into account in the model. The first
involves the lenth of the time steps taken by the model. Each time step (DT)
of the model represents .5 day, but equation (2) gives a new cumulative
mortality only once per day. The necessary adjustment is to solve for a new
variable to teplace A (call it B) which will produce a twice-a-day mortality
sequence that agrees with the dally sequence at every other step. In more
general terms, if DT is any fraction of a day, the appropriate value for B is:
DT
B = l-(l-A) ,	(3)
or in our case of DT = .5, B = 1 - /l-A	(4)
Substituting B into the appropriate places in (2) yields:
C (k) = B • max(m(CONC, TEMP, SIZE ), C (k-1)) + (l-B)C (k-l),k>l (5)
®	\	1 X j 6	1|8	'
where k is now in steps of DT, less than one day.
The final correction is used to adjust for the fact that individuals ad-
vance from size class to size class, altering the average cumulative pesticide-
induced mortality experienced by a size class. The desired effect is that growth
between size classes act like a conservative, dilution-type process operating
on the mortalities. The adjustment is performed on the cumulative mortality,
not on C , its probit. Let D, represent the proportion of size class i
that has^een killed by pesticia£Sin soil before correction for "graduations."
That is, D^ g(k-l) = probit (c. (k-1)) . Then D. (k-1) is adjusted at
each time step k by utilizing tn£ number of lndiviadils of size class i-1
advancing to size class i (call this number a. .), the number advancing from
size class i to i+1 (call this a^), the number alive in size class 1 at time
k-1 (call this n.(k-l)), and the cumulative pesticide-induced mortalities
experienced by size classes i-1 and i (denoted D1_1(k-1) and D (k-1), re-
spectively). The appropriate cumulative mortality for size class i prior to
any new kill at time k is thus:
181

-------
^	n. before new pesticide-induced mortality at time k
D (k-1) = 1					
i,s	population^ if no pesticide-induced mortality existed
n (k-1) + a - a
-A, \r
1-D^k-l) l-Di;L(k-l)
Equation (6) is applied after each time step^to adjust the cumulative
mortality of each size class, and the resulting (k-1) = probit (Dj^ g(k-l))
is the value actually utilized in equation (5) in $iace of C. (k-1). *
1,8
Assumptions and Implications
The form of the pesticide-induced mortality model Introduced above for
soil exposure to azinphosmethyl is believed to be applicable to a wide variety
of organisms and toxicants. Since this model has not been presented elsewhere,
it is perhaps worthwhile to highlight a number of the underlying assumptions
and implications of this model:
(i) Toxicant-induced mortality is independent of mortality induced by
other (e.g., "natural") causes. Natural mortality occurs in the
isopod model as a function of temperature, size, and breeding con-
dition. The toxicant-induced mortality can be "overlaid" on
mortality due to predation, climatic extremes, etc.
(ii) It is necessary to track toxicant-induced mortality over long periods
of exposure, in the presence of concentrations of toxicant which
may be changing with time.
(ill) The model should agree (within the bounds of experimental and statistical
fitting errors) with mortalities produced by short- and long-term
exposures to constant levels of a toxicant (i.e., laboratory-type
situations).
(iv) If the organism is followed through discrete life stages (e.g., size,
age, or developmental classes), mortalities should be calculated
for each class, adjusting for advancement from class to class.
(v) In the case of a long-term, constant condition exposure to a fixed
level of toxicant, a given size class will asymptotically approach
but not exceed some level of cumulative mortality. Any subsequent
exposure to the same or lower toxicant levels under those conditions
will not result in additional mortality (except to individuals which
advanced into the size class from a size class with a lower cumulative
mortality). This assumption is justified prima facie whenever the
mortality is determined by persistent (for example, genetic) differences
in the susceptibilities of,the individuals In the population. If
the differences are due to existence of refuges or similar reasons,
and these are not similarly present in the laboratory data used to
parameterize the model, then the model may systematically under-
estimate cumulative mortalities. In this case, an unsatisfying but
easily implemented approach to remedying this problem might be to
establish an "effective exposure level" for the field situation
would be a multiple (>1) of the laboratory exposure, yielding the
appropriately higher mortalities upon long-term exposure in the
field. Of course, the more desirable alternative is more detailed
modeling of the spatial distribution of the pesticide and the
organisms.
182

-------
Combining Mortality Due to Soil. Moss, and Litter Exposure
There are many possible methods for combining estimates of mortalities
due to pesticide in soil, moss and litter (or any habitat subdivisions) in
the face of exposure that is divided among them: (1) at one extreme, separate
C. (k), C. .(k) and C. (k) terms could be calculated. Their inverse probit
equivalents|"l)^ 8(k)» D^m^(k), and D m(k) could be combined to get D^(k)
using weighting^ based oft percent activity^in each layer. Then the increment
in cumulative mortality would be D^(k) - D^(k-l). This approach in effect
subdivides the population into three persistent groups—one in soil, one in
litter, and one in moss. While for some types of organisms and habitat
divisions this might be appropriate, It Is not used here because the organism
is highly mobile among the layers. (2) Anothejj method would calculate the
C (k), C (k)^and C (k^ terms based ujjon C. (k-1), an activity-weighted
average of fcne g(k-I}™ C. -(k-1), and C. (k-1^1. This has the effect of
subdividing (by percent activity) the population into three groups, calculating
a separate mortality for each, then re-comblnlng them for a new, Independent
subdivision at the next step. Th£s Is the method employed below, and results
in the use of a single varlalbe C^(k-l) as the pesticide-induced mortality
state variable for each size class. That Is, the combined C^(k-l) is the only
pesticide-Induced mortality which is "remembered" for use at time k in equation
(5). (3) A further refinement would be to allow individuals to divide their
exposures among the three layers at each time step; however, our laboratory
studies to not provide any help in parameterizing such a model. It would be
very difficult to create an appropriate activity-distribution in the laboratory,
so the approximation in (2) was utilized here.
Once C^(k) and D.(k) ° probit (C^(k)) are calculated, the proportion of
the remaining population to be killed at step k is calculated as:
D.(k) - D*(k-1)
KILL (k) = —			(7)
(1-Dk(k-1))
Laboratory Parameterization of the Model for Soil Exposure
The search for the m function for this particular organism used multiple
regression analysis and began with visual examination of the laboratory-
determined 7-day mortalities to seek their relationships with temperature,
pesticide concentration, and size-class of the population treated. As is
common in toxicity analyses, in which a normal distribution is observed in
relating mortality to toxicant exposure variables, the probit transformation
was applied to the mortality rates, to provide a more nearly linear response.
Because there is no upper or lower limit on the probit variable (i.e., probit
(0) = 00, probit (1.0) = + °°), artificial upper and lower bounds were intro-
duced. Any mortality below 0.1% was assigned a probit transform value 1.9098,
(corresponding to 0.1%), and any mortality above 99.0% was treated as 7.3263
probits. These values were chosen as being close enough to 0 and 100%, re-
spectively, that the difference in model performance would be acceptable,
and that unreallstically high probit values would not be generated by the
model. Transformations were sought for the size, temperature, and concentration
variables to allow a better fit to the dependent (y) variable (cumulative
mortality in probits).
Clearly, many of the independent variables interact non-additively in
determining the value of the dependent variable. For this organism, long-
183

-------
term pesticide-induced mortality was expected to be related positively to
temperature and inversely to size. Further, preliminary analysis suggested
that the degree of concentration-dependency was itself size-related. A
variety of terms, including the traditional log.. Q (concentration), were ex-
plored in various combinations. A form which dia not diverge for CONC = 0
was needed, and the following form for the m function was the best among the
forms tried for this organism:	k
k	k ¦ SIZE r	k
m = k^TEMP • (CONC)	• (SIZE)	(8)
where m = long term (k>>7) probit cumulative mortality resulting from
exposure of organisms of size SIZE (nig) to pesticide at concentra-
tion CONC (ppm) at temperature TEMP°C, in our standard laboratory
environment (see section X. C. 1)
Equation (8) may not be the best representation for the data, and better fits
might be generated if each m^ were fitted separately; however, we wished to
test a form which explicitly incorporates the size-dependence of m. In order
to ease the problem of fitting this equation a log transformation was done,
yielding:
log1Qm = log^ + k2log1()TEMF + k3 SIZE log1QCONC + k^og^SIZE	(9)
In this form, a multiple regression analysis could be versus log^gTEMP, l°g^Q
SIZE, and ceo.3 of the for. SIZE 'lO610C0NC- £or seveIal cholces °£ k4' t0
get a preliminary estimate of the coefficients, for later refinement using
the nonlinear optimizing "COMPLEX" procedure of Kuester and Mize (1973).
The task remained to estimate the values of m, the dependent variable, from
the laboratory measurements of daily cumulative percent mortality observed
each day during a 7-day exposure. This required using the parameter A to
transform each data point DEAD(k), the probit of observed cumulative mor-
tality on day k, to an estimate of m, the long-term probit mortality for
those conditions. The appropriate transformation can be shown to be:
DEAD (k)	(10)
m= 	;—i—r
l-(l-A)
In this way, better use of the daily mortality data available could be
made by including each day's observed cumulative mortality (corrected for
control mortality) rather than using only the observations at the end of 7
days. Of course, this approach suffers from the non-independence of the
cumulative data taken day after day for the same set of animals, but the
departure from the statistical model underlying the analysis was felt to be
acceptable in this case, given the nature of the data involved.
The results of the COMPLEX procedure are given by equation (11) below.
. - 1.3352 TEMP"333W-^"(SIZE-03M>	(u)
SIZE* 7
The A value which gave the best fit was A = 0.6.
Although the data set is large (1477 observations on groups of 20 animals),
the data contain a large amount of noise. It is possible that a better fit
would be obtained if each size-class were treated separately. However, there
is a large amount of variation evident in the data among replicates of the
184

-------
same conditions, which no model based on temperature, time, concentration
and size could explain.
Litter Exposure
The model for mortality in response to litter exposure is identical in
form to the yodel just discussed, and is also based upon a set of laboratory
experiments. Because no separate studies for moss exposure were made, the
litter exposure model is also applied directly to model moss exposure effects.
For the litter experiments, however, the mortality observations were not made
daily for 7 days, but rather once, at 6 days. Thus, it was not possible to
determine from these data the time course of the mortality, so it is assumed
that the A values (determining the rate at which cumulative probit mortality
can rise) were identical with those for the soil exposure model. A regres-
sion analysis was then performed, with the same set of independent variables
as was utilized in the soil toxicity run, but forcing the A value to 0.6,
the value selected in the soil regression. All of the litter toxicity lab-
oratory determinations were run at 21°C, so the temperature-dependence term
found in the soil toxicity study was normalized and incorporated In the litter
equation (12). Finally, the parameters were "fine-tuned" using the COMPLEX
procedure as with the soil mortality equation, yielding:
Ci,litter(k> =0-6 • ",alt [CI,UtterO"1*' " (C0NC' TEMP' 1+0.4 gutter0"1'
(12)
— .18	#33
where m(C0NC, TEMP, SIZE^ = 3.035 • SIZE-"21 . CONC*33 ' SIZE • (~p)
C, .. (k-1) is corrected for growth into the size class in the same fashion
as'equation (6).
Field observation versus predicted mortality under laboratory-based simulation
of field conditions
In the same experimental orchard in which the parameters for the pesti-
cide movement and fate model were derived, an extensive program of sampling
for T. rathkei was conducted, as reported by Snider. The'sampling yielded
only a single point estimate of absolute density of isopods, but could furnish
an estimate of the relative densities in sprayed versus control plots. Figure
X-5 shows the mean numbers of isopods captured per sampling board (625 cm each),
divided into pre-reproductive (42 mg) and reproductive (_> 12 mg) size classes,
in sprayed versus control plots. Table X-9 shows the observed and predicted
ratios in sprayed versus control plots on 20 sampling dates in 1977 and 1978.
Sprayed plots received 4-5 sprays/year in 1976, 1977, and 1978, at approxi-
mately 2 lbs/acre, a typical rate for commercial use. Analysis of the field
data shows almost no surviorshlp of small isopods in the sprayed plots, followed
by a nearly 10-fold recovery of the sprayed population prior to the onset of
reproduction in the spring, which must therefore be attributed to Immigration.
The small size of the treatment plots accentuated this effect beyond what
would be expected in a commercial orchard. Snider also reported that the
size distribution in the sprayed plots was skewed toward the larger sizes,
which may be attributable to differential susceptibility to the toxicant,
sub-lethal effects on reproduction, and size-related rates of immigration.
The reason that the populations in sprayed plots recovered more completely
185

-------
Figure X-5. Mean numbers of Isopods captured/board, In control vs. sprayed plots,
broken into size groups of < 12 mg, >12 mg.
4


,
¦ 100
60
1— 20
"Wl
-100
- 60
- 20
5/30
6/16	6/27
4.0 -
3.0 -
2.0 -
1.0 -
n I
1978
Ik
n
CL
7/10
CONTROL
SPRAYED
-100
- 60
- 20
ex-
es
IS
4->
o
vO
.00
7/28
8/8
8/30
9/28
11/1

-------
in 1978 than in 1977 is not known (see Chapter IX).
Results and Discussion
The pesticide-induced mortality model described above was incorporated
into the T. rathkei life history model. In order to test the model, a three-
year simulation was performed, using as inputs the weather conditions and
spray application rates measured in our experimental orchard in 1976, 1977,
and 1978. Becasue daily vaiues for the pesticide residue levels in each layer
are required for the T. rathkei model, the pesticide fate model must be used
to estimate these. Thus, the behavior of the T. rathkei model is dependent
on the validity of the fate model. Figure X-6 shows the simulated T. rathkei
populations by size class for a control plot, while Figure X-7 represents
simulated sprayed plot populations. The model predicted slightly earlier
onset of reproduction in 1977 than was seen in the field (see Chapter VIII).
The size-dependence of the predicted pesticide-induced mortality is evident
in Figure X-7, as is the longer-term effect of this juvenile mortality on
adult numbers.
Column 4 of Table X-9 shows the model's predicted ratio of control to
sprayed populations. While this measure is designed for comparison with
column 3, the (unquantified) size dependency in the sampling efficiency of
the field measurements, which employed sampling boards, may have introduced
some of the differences between the columns.
Several observations that emerge from examination of Table X-9 appear
to be important:
(1) The model predicted very high mortality of small isopods in '78,
down to 1.8% of the control populations at the end of the season. Observations
in sprayed plots tend to confirm this, with no Individuals <12 mg collected
on 5 of 8 sampling days after the onset of spraying in 1978. The average of
the 8 population observations for <12 mg isopods in the May 31 - November 1,1978
spray season is 1.9% of the control population. As would be expected, the
field measurements were much more variable than the model predictions, but they
are in good general agreement. In both 1977 and 1978, the predicted mortality
for small Isopods was not as high as was observed in the field, which makes
it even more interesting that adult populations were actually higher than
predicted in sprayed plots. These observations indicate that in some way,
the adult populations in sprayed plots are recovering, both within and between
seasons. For example, the sprayed plot adult populations captured went from
2.9% of controls to 33.3% of controls between August 9 and August 26, 1978,
while not a single surviving small isopod was captured in a sprayed plot
that season. This points toward either a large amount of immigration of
adults into the (rather small) sprayed plots, or a sampling bias (e.g., the
possibility that post-spray sampling did not collect surviving adults pro-
portionally to their numbers, due to immobilization by pesticide. This is
considered by the authors to be unlikely). In a sprayed plot large enough
to reduce the effects of immigration, the populations are unlikely to display
the behavior shown in Table X-9. The model prediction of adult sprayed plot
population (last column) increase from 08.4% to 13.3% of control populations
during the over-wintering period is explained by the slight density-dependence
of natural mortality in the model.
The model may not predict high enough mortality of <12 mg isopods. This
may reflect on the transformation of laboratory pesticide exposures in ppm
into field units of yg/cm , a process which obviously depends not only on
the pesticide profile in the orchard layers, but also on the isopod's behavior.
187

-------
CONCLUSIONS
Further work, some of which has already been undertaken, will be necessary
before better tests of the general model presented here can be made. Several
difficulties are evident to the authors:
(1)	Immigration was very significant in the small experimental plots,
but is not represented in the model, which simulates a closed pop-
ulation. This is not a flaw in the mortality model, but a factor
which affects its comparison to the field data.	^
(2)	The process of converting from the fate model's estimate of yg/cm
pesticide in soil to the ppm required in the T. rathkei model needs
further refinement. This is a classical problem—to relate lab
determinations to field measurements. It is confounded with the
need to estimate the depth of soil residues to which the organism
is exposed.
(3)	The distribution of the organism's activity between soil and other
layers is not well established.
(A) Possible effects of the pesticide on fecundity are not incorporated
in the present model. However, the much higher sensitivity of
smaller size classes to the pesticide may make the simplification
acceptable for this organism, since it makes little difference
whether the females subjected to sublethal pesticide levels do not
produce viable off-spring or whether the offspring die shortly
after birth due to direct exposure.
(5) The noisy laboratory data for T. rathkei and the attendant diffi-
culty in deriving parameters for the m function from them mask the
performance of the general model form developed here.
More work is needed to determine a variety of good forms for m functions,
since various organisms (and life stages) and toxicants may behave quite
differently. It would be useful to have a library of forms which could be
selected among for parameterizing a model of a particular organism and toxi-
cant.
A model of this form can be useful even if it is not yet sufficiently
validated to allow direct use of its predicltons for a particular population.
If the model represents the current understanding, based on laboratory studies,
of the effects of a toxicant on a population, and if its predictions do not
agree with field measurements, the model is an excellent tool for discovering
the discrepancies and exploring the possible explanations for the discrepancies.
In order to simulate the effects of a toxicant on an ecosystem, for inte-
grated pest management or other purposes, a model must not only incorporate
single-species population components, but also represent their interactions
with each other and with the environment. However, models which represent
the direct effects of long-term, time-varying exposures of populations to a
toxicant remain as important steps toward the goal of relating the behavior
of toxicants in the laboratory to their effects in the environment.
188

-------
ib.»
"Ha.w
IN ORTb •I01
7».«
MrC (r'oats • 10*
"SV>«	«^.to ,
Tine IN OATS *10
IV*. so
9
Figure X-6. Simulated populations (number/m ) of T. rathkei for three
years in a control (unsprayed) plot: (a) eggs (daily hatch),
(b) 0.5-5 mg, (c) 5-12 mg, (d) 12-30 mg, (e) 30-50 mg,
(f) 50-80 mg, (g) >80 mg, (h) totals, excluding eggs.
189

-------
p
Figure X-7. Simulated populations (number/m ) of T. rathkei for three
years in a plot sprayed with azinphosmethyl (see Pesticide
Fate Model section): (a) eggs (daily hatch), (b) 0.5-5 mg,
(c) 5-12 mg, (d) 12-30 mg, (e) 30-50 mg, (f) 50-80 mg,
(g) >80 mg, (h) totals, excluding eggs.
190

-------
Table X-9. Comparison of field data with model predictions. Columns 3 and
4 are the ratios of numbers of pre-reproductive (<12 mg) and
reproductive (>12 mg) X* rathkei collected per sampling board from
control plots^to number per board from sprayed plots, for each
sampling day. Last two columns are the model's prediction,
given the environmental and pesticide application inputs observed
in the field, of the ratios of <12 mg and 12 mg populations from
unsprayed and sprayed plots on the same day.
SPRAY/CONTROL RATIOS
Observed	Predicted
Day of
Date
Model Run
<12 mg
> 12 mg
<12 mg
>*2 MS
1977





May 24
509
0.0
0.200
0.429
0.343
June 10
526
0.0
0.107
0.230
0.337
July 20
566
0.0
0.32
0.061
0.205
Aug 9
586
0.0
0.029
0.049
0.124
Aug 26
603
0.010
0.333
0.041
0.102
Sept 26
635
0.009
0.525
0.035
0.083
Nov 17
686
0.0
0.118
0.037
0.084
1978





Apr 7
827
0.0
0.633
0.259
0.133
Apr 25
845
0.218
0.648
0.056
0.133
May 2
852
0.0
0.825
0.051
0.132
May 18
868
0.0
0.381
0.063
0.131
May 30
880
0.0
0.451
0.063
0.130
June 16
897
0.0
0.252
0.103
0.128
June 27
908
0.091
0.064
0.125
0.127
July 10
921
0.0
0.056
0.045
0.126
July 28
939
0.0
0.210
0.024
0.107
Aug 8
950
0.009
0.084
0.038
0.086
Aug 30
972
0.0
0.182
0.022
.0.055
Sept 28
1001
0.0
0.220
0.018
0.041
Nov 1
1035
0.053
0.196
0.018
0.040
191

-------
Definitions of Symbols in Chapter X
>etween C.(k-l) and m± (cONC(k-l), TEMP(k-l))
r; i.e., fraction by which mortality (in
Proportion of distance between
that is moved in one day;
probits) can change from its current to its ultimate value under
given conditions in one day.
ai_1(k)
Number/m of individuals "graduating" from size class i-1 into size
class i at time k.
aA(k)
Dumber /m of individuals "graduating" from size class i into size
class i into size class 1+1 at time k.
Ct(k)
C*(k)
CONC(k)
Analogous to A, but corrected for time steps of durations less than
one day.
Cumulative mortality, in probits, for the 1— size class at time k,
due only to pesticide exposure, before correction for growth of
individuals into and out from the size classt.
Corrected cumulative mortality, in probis, for the i— size class
at time k, djje only to pesticide exposure. Mathematically defined
as probit (^00 ) • +
Concentration of pesticide in a layer (litter, moss, or soil) In
parts per million (ppm) at time k.
Dt(k)
D*(k)
Proportion (not in probits) cumulative mortality for size class i
due to pesticide, before correction for mortalltes of_jjjey "graduates"
into the size classt. Mathematically, D^(k) = probit	¦
Proportion (not in probits) cumualtive mortality for size class i
due to pesticide, after correction for mortalities of new "graduates"
into the size class using equalton (10) .
DEAD(k) Laboratory-observed cumulative pesticide-induced mortality on day k
of a 7-day experiment, after correction for mortality of control
population.
Index of time step in the model,
is commonly 1 day or 0.5 day.
Length of time step can vary, but
KILL.(k) New pesticide-induced mortality (a proportion, not a probit) for
size class i on day k due to pesticide exposure.
1 ^21 ^3 * ^ it1 ^5
Coefficients in mortality function, equation (2).
m
= m(C0NC, TEMP, SIZE). Long-term cumulative mortality under constant
exposure with a given CONC, TEMP, and SIZE. Fitted using data for
all size classes, as a continuous function of size.
m.
= m^(C0NC, TEMP). Long-term cumulative mortality for size class i
under constant exposure with a given CONC, TEMP.t
192

-------
2
n^(k-l) Nuraber/m of individuals alive in size class i at time k-1.
SIZE	Weight (mg) of an organism.
SIZE^ Parameter representing average weight (mg) of organisms in the i—
size class.
TEMP(k) Estimated mean soil temperature (°C) for time step k at depth of
1 cm. For laboratory data, TEMP represents the temperature at
which the environment was maintained during the pesticide exposure
of the organisms.
t Additional subscripts 1, m or s specify that mortality is due only to
pesticide exposure in litter, moss, or soil layers, respectively.

-------
REFERENCES
Al-Adil, K. M., E. R. White, W. L. Winterlin, and W. W. Kilgore. 1973.
Uptake and Translocation of Guthion by Beans and Barley. J. Agric.
Food Chem. 21:376-379.
Aritajat, U., D. S. Madge and P. T. Gooderham. 1977. The effects of com-
paction of agricultural soils on soil fauna. I. Field investigations.
Pedobiol. 17:262-282
Armstrong, D. E. and J. G. Konrad. 1974. Nonbiological degradation of
pesticides. In Pesticides in Soil and Water. (Ed. by W. D. Guenzi).
pp. 123-131. SSSA. Madison, WI.
Asmussen, L. E., A. W. White, Jr., E. U. Hauser, and J. M. Sheridan. 1977.
Reduction of 2,4-D load in surface runoff down a grassed waterway.
J. Environ. Qual. 6:159.
Bakker, K., 1956. Feeding activity patterns in Porcellio scaber LATR. Proc.
K. Ned. Akad. Wet. Ser. C, 59-750-757.
Barley, K. P. 1958. Earthworms and soil fertility. I. Earthworm popula-
tions found in agricultural land near Adelaide. Aust. J. Agr. Res.
10:171-178.
Barrow, G. M. (1966) Physical Chemistry. McGraw-Hill, Inc., New York.
843 pp.
Bauer, J. R., Bovey, R. W. and Merkle, M. G. (1972) Concentration of
Picloram in runoff water. Weed Sci. 20, 309-314.
Behrens, W. V. 1929. Ein Beitrag zur Fehlerberechnung bei wenigen
Beobachtungen. Landwirtsch. Jahrb. 68:807-837.
Beynon, K. I., Davies, L., and Elgor, K. (1966) Analysis of crops and soils
for residues of diethyl l-(2,4 dichlorophenyl)-2-chlorovinyl phosphate.
J. Sci. Food Agr. 17, 167-174.
Bouch€, M. B. 1972. Lombriciens de France. Ecologie et syst£matique.
Ed. I.N.R.A., Annales de Zoologie, numero special 72.2, 1-671.
Brereton, J. LeG., 1956. The distribution of woodland isopods. Oikos
8:85-106.
Brown, A. W. A. (1978) Ecology of Pesticides. John Wiley and Sons,
New York, 525 p.
Brown, A. W. A. (1977) How have entomologists dealt with resistance. Proc.
68th Ann. Meet. Am. Phytopathol. Soc. 3, 67-74.
Bridges, R. C. and Farrington, J. A. (1974) A compartmental omputer model
of foliar uptake of pesticides. Pestic. Sci. 5, 365-381.
194

-------
Bukovac, M. J. (1976) Herbicide entry into plants. In Herbicides:Physiology,
Biochemistry, Ecology. (Ed. by Audus, L. J.), pp. 335-364. Academic
Press, London.
Carnio, J. S. 1974. A laboratory study on the effects of some climatic
factors on the demography of the terrestrial isopods, Porcellio
spinicornis Say and Tracheonlscus rathkei Brandt. M. Sc. Thesis,
York University, Toronto, Ontario.
Caro, J. H., Taylor, A.. W. and Lemon, E. R. (1971) Measurement of pesticide
concentrations in air overlying a treated field. International
Symposium on Identification and Measurement of Environmental Pollutants,
Ottawa, Canada.
Caro, J. W., Freeman, H. P., and Turner, B. C. (1974) Persistence in soil
and losses in runoff of soil-incorporated carbaryl in a small watershed.
J. Agric. Food Chem. 22, 860.
Chapell, H. G., T. F. Ainsworth, R. Cameron and M. Redfem. 1971. The
effect of trampling on a chalk grassland ecosystem. J. Appl. Ecol.
8:869-882.
Claith, M. M., Spencer, W. F., Farmer, W. J., Shoup, T. D., and Grover, R.
(1980) Volatilization of S-ethyl n,n-dlpropylthlocarbamate from water
and wet soil during and after food irrigation of an alfalfa field.
J. Agric. Food Chem.. 28, 610-613.
Cox, G. W. (1974) Laboratory Manual of General Ecology. Wm. C. Brown Company,
Dubuque, Iowa, 196 p.
Croft, B. A. (1975) Tree fruit pest management. In Introduction to Pest
Management. (Eds. Metcalf, R. L. and Luckmann, W. H.). Wiley
Interscience, New York.
Croft, B. A. (1978) Potentials for research and implementation of integrated
pest management on deciduous fruit-trees. In Pest Control Strategies
(Eds. Smith, E. H. and Pimentel, David), p. 101-115. -Academic Press,
New York.
Croft, B. A. and Brown, A. W. A. (1975) Responces of arthropod natural
enemies to insecticides. Ann. Rev. Entomol. 20, 285-335.
Crosby, D. G. (1969) Experimental approaches to pesticide photodecompositlon.
Residue Reviews 25, 1-12.
Crosby, D. G. (1970) The non-biological degradation of pesticides in soils.
In Pesticides in the Soil: Ecology, Degradation, and Movement, p. 86-
94, Mich. State Univ., East Lansing, MI.
Crosby, D. G. (1973) The fate of pesticides in the environment. Ann. Rev.
Plant Physiol. 24, 467.
Day, B. E., Jordon, L. J. and Hendrixson, R. T. (1961) The decomposition of
amitrole in California soils. Weeds 9, 443-456.
195

-------
DeJong, K. (1980) Adaptive system design: a genetic approach. IEEE Trans.
Syst. Man Cybern., SMC-10, p. 566-574.
Duffey, E. 1975. The effects of human trampling on the fauna of grassland
litter. Biol. Conserv. 7:255-274.
Eason, R. and W. H. Whitcomb. 1965. Life history of the dotted wolf spider
Lycosa punctulata Hentz (Araneida: Lycosidae). Arkansas Acad. Sci.
Proc. 19:11-20.
Ebling, W. 1963. Analysis of the basic processes Involved in the deposition,
degradation, persistence and effectiveness of pesticides. Residue
Reviews 3, 35.
Edgar, W. D. 1970. Prey and feeding behavior of adult females of the wolf
spider Parodsa parmentata. Netherl. J. Zool. 20:487-491.
Edgar, W. D. 1971. The life cycle, abundance and seasonal movement of the
wolf spider, Lycosa (Pardosa) lugubrls, in Central Scotland. J. Anim.
Ecol. 40:303-322.
Elliot, D. L., Iwata, Y., Carman, G. E., and Gunther, F. A. 1977. Post-
application slough-off of pesticide deposits on orange trees. Bull.
Environ. Contain. Toxicol. 18, 711-718.
Eto, M. (1974) Organophosphate pesticides: organic and biological chemistry.
CRC Press, Cleveland, Ohio. 368 p.
Evans, A. C. and W. J. McL. Guild. 1948. Studies on the relationships
between earthworms and soil fertility. IV. On the life cycles of
some British Lumbricidae. Ann. Appl. Biol. 35:471-484.
Faust, S. D. and Gomaa, H. M. 1972. Chemical hydrolysis of some organic
phosphorus and carbamate pesticides in aquatic environments. Environ.
Lett. 3, 171-201.
Felsot, A. and Dahm, P. A. (1979) Sorption of organophosphorus and carbamate
Insecticides by soil. J. Agric. Food Chem. 27, 557-563.
Freed, V. H., Chiou, C. T. and Schmedding, D. W. 1979. Degradation of
selected organophosphate pesticides in water and soil. J. Agric. Food
Chem. 27, 706-708.
Geiser, S. W. 1934. Further observations on the sex-ratios of terrestrial
isopods. Field & Lab. 3:7-10.
Gillette, J. W. and Witt, J. M., Eds. (1979) Terrestrial Microscosms Proc.
Symp. Workshop on Terrestrial Microscosm, NSF/RA 790034.
Glass, B. L.. and Edwards, W. M. (1974) Picloram in lysimeter runoff and
percolation water. Bull. Environ. Contam. Toxicol. 11, 109-112.
Glass, E. H. and Lienk, S. E. (1971) Apple insect and mite populations
developing after discontinuance of insecticides: 10 year record. J.
Econ. Entomol. 63/23-26.
196

-------
Goodman, E. D. 1982. Modeling effects of pesticides on populations of soil/
litter invertebrates in an orchard ecosystem. Env. Toxicol, and Chem.
1, 45-60.
Goodman, E. D. 1977. Ecosystem responses to alternative pesticides in the
environment: A systems approach. Final Report, EPA Project No.
R-803859, Corvallis, OR. pp. 230-253.
Goring, C. A. I., Laskowski, D. A., Hamaker, J. W. and Meikle, R. W. (1975)
Principles of pesticide degradation in soil. In Environmental Dynamics
of Pesticides (Eds. Haque, R. and Freed, V. H.), p. 135-172. Plenum
Press, New York.
Gr6goire-Wibo, C. 1975. Biodcologie et dynamique de population de Collemboles.
Ph.D. Thesis, University Catholique de Louvain, Fac. des Sciences, 1-217.
Grubbs, F. E. (1969) Procedures for detecting outlying observations in
samples. Technometries 11, 1.
Gunther, F. A. 1969. Insecticide residues in California citrus fruit and
products. Residue Reviews 28.
Gunter, F. A. and Blinn, R. C. (1955) Analysis of Insecticides and Acaricides.
Interscience, New York.
Gunther, F. A., Westlake, W. E., Barkley, J. H., Winterlin, W. and Langbehn,
L. 1973. Establishing dislodgeable pesticide residues on leaf surfaces.
Bull. Environ. Contam. Toxicol. 9, 243-249.
Gunther, F. A., Iwata, Y., Carman, G. E., and Smith, C. A. 1977. The citrus
reentry problem: research on its causes and effects, and approaches to
its minimization. Residue Reviews 67, 1-139.
Hall, A. S. and Day, J. W., Jr., Eds. (1977) Ecosystem Modeling - In Theory
and Practice. Wiley-Interscience, New York. 684 p.
Hall, F. R., Reichard, D. L., and Krueger, H. R. (1975) Dislodgeable
azinphosmethyl residues from air blast spraying of apple foliage in
Ohio. Arch. Environ. Contam. Toxicol. 3, 352-363.
Hall, J. K., Paulus, M. and Higgins, E. R. (1972) Losses of atrazine in
runoff water and soil sediment. J. Environ. Quality 1, 172-176.
Hamaker, J. W. (1972) Decomposition: quantitative aspects. In Organic
Chemicals in the Soil Environment. (Eds. Goring, C. A. I. and Hamaker,
J. W), p. 253-340. Marcel Dekker, New York.
Hansen, J. D., Schneider, B. A., Olive, B. M., and Bates, J. J. (1978)
Personnel safety and foliage residue in an orchard spray program using
azinphosmethyl and captan. Arch. Environ. Contam. Toxicol. 7, 63-71.
Harris, C. R. 1967. Further studies on the influence of soil moisture
on the toxicity of insecticides in soil. J. Econ. Entorn. 60:41-44.
197

-------
Harris, C. R. 1970. Laboratory evaluation of candidate materials as potential
soil insecticides. III. J. Econ. Entom. 63:782-787.
Harris, C. R. 1971. Influence of temperature on the bilogical activity
of insecticides in soil. J. Econ. Ent. 64 (5): 1044-1049
Harrold, L. L., Triplet, G. B., Jr., and Edwards, W. M. (1970) No-tillage
corn-characteristics of the system. Agric. Eng. 51, 128.
Hartley, G. S. (1969) Evaporation of pesticides. In Pesticide Formulations
Research. (Ed. Gould, R. F), p. 115-134. Am. Chem. Cos. 'Advances in
Chemistry Series,' 86.
Hatchett, S. P. 1947. Biology of the Isopoda of Michigan. Ecol. Monogr.
17:48-79.
Haynes, R. J. (1980) Influence of soil management practice on the soil agro-
ecosystem. Agro-Ecosystems 6, 3-32.
Heeley, W. 1941. Observations on the life histories of some terrestrial
lsopods. Proc. Zool. Soc. London, B, 111:79-149.
Helling, C. S., Kearney, P. C., Alexander, M. (1971) Behavior of pesticides
in soils, Adv. Agron. 23, 147-241.
Hill, K. R. (1971) Report of half-life working party. J. Ass. Offic. Anal.
Chemists 54, 1316-1317.
Holland, J. H. (1975) Adaptation in natural and artificial Systems. The
University of Michigan Press, Ann Arbor. 183 p.
Howitt, A. T. and Pshea, A. (1965) The development and use of ultra low
volume ground sprayers for pests attaching fruit. Mich. Quart. Bull.
48:144-160.
Hoyt, S. C. and Burts, E. C. (1974) Integrated control of fruit pests. Ann.
Rev. Entomol. 19, 231-252.
Hull, H. G. (1970) Leaf structure as related to absorption of pesticide and
other compounds. Residue Reviews 31, 1-150.
Iwata, Y., Ousch, M. E., Westlake, W. E., and Gunther, F. A. 1975. Behavior
of five organophosphorus pesticides in dust/derived from several soil
types. Bull. Environ. Contam. Toxicol. 14, 49.
Iwata, Y., Westlake, W. E., Barkley, J. H., Carman, G. E., and Gunther, F. A.
(1977) Behavior of phentohoate (cidial) deposits and residues on and
in grapefruits, lemons and lemon leaves, oranges and orange leaves,
and in the soil beneath orange trees. J. Agric. Food Chem. 25, 362-368.
Kadoum, A. M. (1967) A rapid micromethod of sample cleanup for gas
chromatographic analysis of insecticidal residues in plant, animal,
soil, and surface and ground water extracts. Bull. Environ. Contam.
Toxicol. 2, 264.
198

-------
14
Katan, J., Fuhrmann, T. W. and Lichtenstein, E. P. (1976) Binding of C
parathlon In soil: a reassessment of pesticide persistence. Science
193, 891-894.
Kaufman, D. D. (1964) Microbial degradation of 2,2-dichloro-propionic
acid in five soils. Can. J. Microbiol. 10, 843-852.
Kearney, P. C. and Helling, C. S. (1969) Reactions of pesticides in soils.
Residue Reviews 25, 25-44.
Khan, M. A. Q., H. M. Khan and D. J. Sutherland. 1974. Ecological and
health effects of the photolysis of insecticides. In Survival in
Toxic Environments, M. A. Q. Khan and J. P. Bederka, eds. Academic
Press, pp. 333-355.
Khan, M. A. Q., Khan, H. M., and Southerland, D. J. (1976) Ecological and
health effects of the photolysis of insecticides. In Survival in Toxic
Environments (Eds. Khan, M. A. Q. and Bederka, J. PI, Jr.), p. 333-355.
Academic Press, New York.
Klingman, G. C. and Ashton, F. M. (1975) Weed Science: Principles and
Practices. Wiley Interscience, New York, p. 32-33.
Kuester, J. L. and J. H. Mize. 1973. Optimization Techniques with FORTRAN.
McGraw-Hill.
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. Contam. Toxicol. 11, 224-230.
Lasslter, R. R., Baughman, G. L. and Burns, L. A. 1978. Fate of toxic organic
substances in the aquatic environment. In State-of-the-Art in
Ecological Modeling (Ed. S. E. Jorgensen). pp. 219-245. Inter. Soc.
Ecol. Mod., Copenhagen.
Leonard, R. A., Bailey, G. W., Swank, R. R., Jr. (1976) Transport, detoxifi-
cation, fate and effects of pesticides in soil and water environments.
In Land Application of Waste Materials, 49-78. Soil Conservation
Society of America, Ankeny, Iowa.
Liang, T. T. and Lichtenstein, E. P. (1972) Effect of light, temperature,
and pH on the degradation of azinphosmethyl. J. Econ. Entomol. 65,
315-321.
Liang, T. T. and Lichtenstein, E. P. (1976)^Effects of soils and leaf
surfaces on the photodecomposition of [ C] azinphosmethyl. J. Agrlc.
Food Chem. 24, 1205-1210.
Manetsch, T. J. 1976. Time-varying distributed delays and their use in
aggregate models of large systems. IEEE Trans. Systems, Man, and
Cybem., SMC-6, pp. 547-553.
Matsumurs, Fumio. 1975. Toxicology of Insecticides. Plenum Press, New
York, 503 p.
199

-------
McMechan, A. D., Morgan, C. V. G. and Wardle, G. A. (1972) Erosion of
azinphosmethyl from apple leaves by rain and overtree irrigation. J.
Entomol. Soc. Brit. Columbia, 69, 5-8.
McQueen, D. J. 1976a. Poroellio spinicomia SAY (Isopoda) demography. II.
A comparison of field and laboratory data. Can. J. Zool. 54: 825-842.
McQueen, D. J. 1976b. Porcellio spinicomia SAY (Isopoda) demography. III.
A comparison between field data and the results of a simulation model.
Can. J. Zool. 54:2174-2184.
McQueen, D. J. 1976c. The Influence of climatic factors on the demography
of the terrestrail isopod Traoheoniscus rathkei BRANDT. Can. J. Zool.
54:2185-2199.
McQueen, D. J. and J. S. Carnio. 1974. A laboratory study of the effects
of some climatic factors on the demography of the terrestrial isopod
Porcellio spiniaornis SAY. Can. J. Zool. 52:599-611.
Menzie, C. M. 1972. Fate of pesticides in the environment. Ann. Rev.
Ent. pp. 199-221.
Merriam, H. G. 1971. Sensitivity of terrestrial isopod populations
(Armadillidiitm) to food quality differences. CAn. J. Zool. 49: (5),
pp. 667-674.
Morris, M. G. 1968. Differences between the invertebrate faunas of grazed
and ungrazed chalk grasslands. II. The fauna of sample turfs. J.
Appl. Ecol. 5:601-611.
Murchie, W. R. 1960. Biology of the oligochaete Bimastos zeteki Smith and
Gittins (Lumbricidae) in Northern Michigan. Amer. Midi. Natur. 64:
194-215.
National Academy of Sciences. 1969. Insect Pest Management and Control.
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 of ethion in Florida citrus
and relationships to weather variables. Arch. Environ. Contam. Toxicol,
6, 257-267.
Nilles, G. P. and Zabik, M. J. (1975) Photochemistry of bioactive compounds.
Multiphase photodegradation and mass spectral analysis of basagran. J.
Agric. Food Chem. 23, 410-414.
NSF/RA. (1976) Ecosystem processes and organic contaminants, research needs
and an interdisciplinary perspective. NSF/RA 760008.
Odom, E. P. 1971. Fundamentals of Ecology. 3rd Edition. W. B. Saunders Co.
Philadelphia. 547 pp.
Ouellette, R. P. and J. A. King. 1977. Chemical week Pesticides Register.
McGraw-Hill Book Co. New York. 346 pp.
200

-------
Paris, 0. H. and F. A. Pitelka. 1962. Population characteristics of the
terrestrial isopod Armadillidium vulgare in California grassland.
Ecol. 43(2). pp. 229-248.
Parmele, L. H., Lemon, E. R. and Taylor, A. W. (1972) Micrometeorological
measurement of pesticide vapor flux from bare soil and corn under field
conditions. Water Air Soil Poll. 1, 433-451.
Phillips, F. T. T. (1971) Persistence of organochlorine insecticides on
different substrates under different environmental conditions. I. The
rates of loss of dieldrin and aldrin by volatilization from glass
surfaces. Pestic. Sci. 2, 255-266.
Phillips, F. T. (1974) Some aspects of volatilization of organochlorine
insecticides. Chem. Ind. 5, 197-201.
Pielou, F. C. (1969) An Introduction to Mathematical Ecology. Wiley
Interscience, New York, p. 19-22.
Pimentel, D. and Goodman, N. (1974) Environmental impact of pesticides. In
Survival in Toxic Environments. (Eds. Khan, M. A. Q. and Bederka, J.
B., Jr.) 25-52.
Plimmer, J. R., Kearney, P. C., Kaufman, D. D. and Guardia, F. S. (1967)
Aniitrole decomposition by free radical generating systems and by soils.
J. Agric. Food Chem. 15, 996-999.
Popendorf, W. J. and Leffingwell, J. T. (1978) Natural variations in the
decay and oxidation of parathion foliar residues. J. Agric. Food Chem.
26, 437-441.
Pregltzer, K. E. and J. R. Stacey. 1973. Soil Survey, Alpine Township,
Kent County, Michigan. Publ. Soil Conservation Service, U.S. Dept. of
Agriculture.
Ritter, W. F., Johnson, H. P., Lovely, W. G. and Molnau, M. (1974) Atrazine,
propachlor, and diazinon residures on small agricultural watersheds,
runoff losses, persistence and movement. Environ. Sci. Technol. 8,
38-42.
Rohrbaugh, W. J., Meyers, E. K. and Jacobson, R. A. 1976. Crystal and
molecular structure of organophosphorus insecticides.
Saltzman, S., L. Kliger and Yaron, B. (1972) Adsorption-desorption of parathion
as affected by soil organic matter. J. Agric. Food Chem. 20, 1224-1226.
Satchell, J. E. 1967. Lumbrlcldae. In: Burges and Raw, eds., Soil Biology.
Acad. Press, London, pp. 259-322.
Schrader, G. (1963) Die entwichlung neuer insecktizider phosphazaure-ester.
Weinhein/Bergstr.: Verlag Chemie GmbH.
Schulz, K. R., Llchtenstein, E. P., Liang, T. T. and Fuhrmann, T. W. (1970)
Persistence and degradation of azinphostnethyl in soils, as affected by
formulation and mode of application. J. Econ. Entomol. 63, 432-438.
201

-------
Scopes, N. E. A. and E. P. Lichtenstein. 1967. The use of Folsomla fimetarla
and Drosophlla melanogaster as test insects for the detection of
insecticide residues. J. Econ. Entorn. 60 (6):1539-1541.
Smith, R. F. 1976. Insecticides and integrated pest management. In the
Future for Insecticides, Needs and Prospects. Wiley Intersclence,
New York, pp. 489-506.
Smith, A. E. and Walker, A. (1977) A quantitative study of asulam persistence
in soil. Pestic. Sci. 8, 449-456.
Smith, J. H., Mabey, W. R., Bohonos, N., Holt, B. R., Lee, S. S., Chou, T. W.,
Bomberger, D. C. and Mill, T. (1977) Environmental Pathways of Selected
Chemicals in Freshwater Systems. EPA Report 600/7-77-113.
Snider, R. M. (1979) The effects of azinphosmethyl (Guthioi^) on a popula-
tion of Trachelipus rathkel (Isopoda) in a Michigan orchard. Pedoblol.
19, 99-105.
Snider, R. M. and Shaddy, J. W. (1980) The ecobiology of Trachelipus
rathkei (Isopoda). Pedoblol. 20, 394-410.
Wouthwick, C. H. (1976) Ecology and the Quality of Our Environment. 2nd Ed.,
D. Van Nostrand, London. 426 p.
Spear, R. C., Y. Lee, J. T. Leffingwell and D. Jenkins. 1978. Conversion
of parathlon to paraoxon in foliary residues: effects of dust level
and ozone concentration. J. Ag. Food Chem. 26:434-436.
Spencer, W. F. and M. M. Clalth. 1969. Vapor density of dieldrin. Environ.
Sci. Tech. 3:670-674.
Spencer, W. F. ans M. M. Clalth. 1975. Vaporization of chemicals. In
Environmental Dynamics of Pesticides. (R. Haque and V. H. Freed, Eds.)
p. 61-78. Plenum Press. New York.
Spencer, W. F. and Claith, M. M. (1977) The solid-air Interface: transfer
of organic pollutants between the solld-alr interface. In Fate of
Pollutants in the Air and Water Environments, Part I. (Ed. Suffet, I.
H), P. 107-126. John Wiley and Sons, Inc., New York.
Spencer, W. F., Farmer, W. J. and Claith, M. M. (1973) Pesticide Volatiliza-
tion. Residue Reviews 49, 1-47.
Stamper, J. H., H. N. Nigg, and J. C. Allen. 1979. Organophosphate insecti-
cide disappearance from leaf surfaces: an alternative to first-order
kinetics. Environ. Sci. Tech. 13:1402-1405.
Steffens, W. and Wieneke, J.^1975) Influence of humidity and rain on up-
take and metabolism of C-azlnphosmethyl in bean plants. Arch. Environ.
Contam. Toxicol. 3, 364-370.
202

-------
Steffens, W. and J. ^eneke. ^®7£) Studies on the uptake, metabolism and
degradation of C-labelled Gusathion in kidney beans. Pfanzenschultz-
Nachrichten 29, 1-17.
Steiner, P. W. (1969) The distribution of spray material between target and
non-target areas of a mature apple orchard by airblast equipment. M.S.
Thesis, Cornell Univ.
Sutton, S. L. 1968. The population dynamics of Trichoniscus pusillus and
Philosoia musoomm (Crustacea: Oniscoldea) in limestone grassland. J.
Anim. Ecol. 37:425-444.
Taylor, A. W. (1978) Post-application volatilization of pesticides under
field conditions. J. Air Poll. Control Assoc. 28, 922-927.
Taylor, A. W., Glotfelty, D. E., Glas£, B. L., Freeman, H. P. and Edwards,
W. M. (1976) Volatilization of dieldrin and heptachlor from a maize
field. J. Agric. Food Chem. 24, 625-631.
Taylor, A. W., Glotfelty, D. E., Turner, B. C., Silver, R. E., Freeman, H. P.
and Weiss, A. (1977) Volatilization of dieldrin and heptachlor
residues from field vegetation. J. Agric. Food Chem. 25, 542-548.
Teskey, B. J. E. and J. S. Shoemaker, 1978. Tree Fruit Production. AVI
Publishing Co., Inc. Westport, CT. pp. 19-25.
Thompson, A. R. and F. I. Gore. 1972. Toxicity of 29 insecticides to
Folsomia Candida. Laboratory studies. J. Econ. Entom. 65:1255-1260.
Thompson, A. R. 1973. Persistence of biological activity of seven in-
secticides in soil assayed with Folsomia Candida. J. Econ. Entom. 66
(4):855-857.
Thompson, N.P. and Brooks, R. F. (1976) Disappearance of dislodgeable
residues of five organophosphate pesticides on cirtrus leaves and fruit
during dry and wet weather in Florida. Arch. Environ. Contam. Toxicol.
5, 55-61.
Tomlin, A. D. 1975. Toxicity of soil applications of insecticides to three
species of Springtalls (Collembola) under laboratory conditions. Can.
Ent. 107:769-774.
Turner, B. C. and Glotfelty, 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 some pesticides on South African
citrus. Environ. Qual. Safety, suppl. Ill, p. 330-335.
Van Dyk, L. P. (1976) Parathion persistence on South African citrus. Arch.
Environ. Contam. Toxicol. 4, 2389-311.
Van Rhee, J. A. 1977. Effects of soil pollution on earthworms. Pedobiol.
17:201-209.
203

-------
Van Rhee, 1961.
Verhoeff, K. W. 1920. Zur Kenntnis der Larven, des Brutasackes und der
Bruten der Onlscoidea-Isopoden. AufsMtze 28. Zool. Anz. 51:169-189.
Walker, A. (1974) A simulation model for the prediction of herbicide
persistence. J. Environ. Quality 3, 396-401.
Walker, A. (1976a) Simulation of herbicide persistence in soil. I. Slmazine
and Prometryne. Pestle. Sci. 7, 41-49.
Walker, A. (1976b) Simulation of herbicide persitence in soil. II. Simazlne
and Linuron Long-Term Experiments. Pestic. Sci. 7, 50-58.
Walker, A. (1976c) Simulation of herbicide persistence in soil. III. Propyza-
mlde in Different Soil Types. Pestic. Sci. 7, 59-64.
Walker, A. and A. Barnes. 1981. Simulation of herbicide persistence in
soil: a revised computer model. Pest. Sci. 12:123-132.
Walker, A. and A. E. Smith. 1979. Persistence of 2,4,5-T in a heavy clay
soil. Pest. Sci. 10:151-157.
Waters, R. A. S. 1955. Numbers and weights of earthworms under a highly
productive pasture. N. Z. J. Sci. Techn., Sect. A: Agrlc. Res.,
36:516-525.
Welch, B. L. 1938. The significance of the difference between two means
when the population variances are unequal. Biometrika 29:350-362.
White, A. W., Jr., Harper, L. A., Leonard, R. A. and tumbull, J. W. (1977)
Trifuluralin volatilization losses from a soybean field. J. Environ.
Qual. 6, 105-110.
White, J. J. 1968. Bio-energetics of the woodlouse TraaheoniauBus rathkei
BRANDT in relation to litter decomposition in a deciduous forest.
Ecology 49:695-704.
Whittaker, R. H. (1975) Communities and Ecosystems. MacMillan Publishing
Co., New York.
14
Wieneke, J. and Steffens, W. (1974) Degradation of C-azinphos on glass
plates as compared to the metabolism on bean leaves. Environ. Qual.
Safety, Suppl. III., 544-548.
Williams, K. (1961) Note on the effect of rain, and sprinkler irrigation, on
the persistence of spray residues of Guthion and Sevin on apple leaves.
Can. J. Plant Sci. 41, 449-451.
Wlnterline, W., Bailey, J. B., Langbehn, L. and Mourer, C. (1975) Degradation
of parathion applied to peach leaves. Pest. Mont. J. 8, 263-269.
204

-------
Woodrow, J. E., Seiber, J. N., Crosby, D. G. , Moilanen, K. W., SoderqulsC,
C. J. and Mourer, C. (1977) Airborne and surface residues of parathion
and its conversion products in a treated plum orchard environment.
Arch. Environ. Contam. Toxicol. 6, 175-191.
Yaron, B., Hever, B. and Birk, Y. (1974) Kinetics of azinphosmethyl losses
in the soil environment. J. Agric. Food Chem. 22, 439-441.
Zabik, M. J. and L. 0. Ruzo. (1981) Factors Affecting Pesticide Photo-
decomposition Studies. In Test Protocols for Environmental Fate and
Movement of Toxicants, p. 20-27. Proc. Symp. 94th Ann. Meeting Ass.
of Offic. Anal. Chemists, Washington, D.C.
Zabik, M. J., Leavitt, R. A. and Su, G. C. (1976) Photochemistry of bioactive
compounds: a review of pesticide photochemistry. 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 esters in surface waters hydrolysis, and vaporization.
Environ. Sci. Tech. 9, 1144-1150.
Zepp, R. G. and Cline, David M. (1977) Rates of direct photolysis in aquatic
environment. Environ. Sci. Tech. 11, 359-366.
205

-------
APPENDIX A
Running the Process and Orcdata Data Analysis Routines
This appendix is intended only to document the process in which data
were prepared for analysis and use in parameterizing the models described in
this report. These routines are written to run only on MSU's CYBER 750 Com-
puter, and are not intended for use by others. The appendix consists of brief
descriptions of programs utilized and the control card structures needed to
run them. It does not address the data used in parameterizing the organism
models.
Process
The binary for processing the data (adding of data, duplicate check,
error correction and statistics) is on file PROCESSBIN.
IRUNMOD gives the user choice of routines to be used.
IRUNMOD
1	= Addition of data to existing data files (ADD)
2	= Search for duplicates and delete oldest copy of duplicates found (DUP)
3	= Processing of samples (PROCESS)
4	= Processing of targets (rain and non-rain) (PROTARG)
It is suggested that If IRUNMOD (1) is chosen, IRUNMOD (2) should also be
chosen.
ADD and DUP use these files:
Tape 2 = DATASAMPLES
Tape 3 = DATATARGETS
Tape 4 = NONRAINTARGETS
and put out these files
Tape 20 = DATASAMPLES
Tape 21 = DATATARGETS
Tape 22 = NONRAINTARGETS
PROCESS uses these files:
Tape 20 = DATASAMPLES
Tape 48 « SPRAYDATES
Tape 49 » RAINDATES
and puts out this file:
Tape 23 - PROCESSEDSAMPLES
PROTARG uses these files:
Tape 21 = DATATARGETS
Tape 22 = NONRAINTARGETS
and puts out these files:
Tape 24 = PROCESSEDTARGETS
Tape 25 = PROCESSEDNONRAIN
To run this program attach the files needed for the routines that are desired
A-l

-------
INPUT rues
OUTPUT rues
Old files
Hti Card
NOKRAIH TARGET
T«pe »
DATA OAKK.ES
/t*p« V \
( CORRECTED r
\SJWPLES J
/upT*\ "N
( OATA TARGETS J
Up*
NONRAIH
" Y
>1* TARGE Ty


MY A»SS it))

\
....
•	f	
i

OAT AVGS (27)
)

1
OUTLIER
MEUGE (22)
iM N
i saw.es J
(upTzi X
DATA TARGETS )
HILO AVG


CORRECT
,/
(
/tap, V \
1 CORRECTED I
Vsahpies J
AVERAGE (23)
1
(Ttpe 26	X
PKOCESSEO	I
SAMPLES	J

AVERAGE (21)
/
)
\



-------
and then attach the binary on file PROCESSBIN. If the ADD and DUP parts of
the program are being used, this card must be Included before execution of
the program:
LDSET.LIB = FORTRAN/CRM.
This card is necessary any time sort/merge is used.
After execution of the program, purge off the previous files and catalog
the new files created.
After the control cards, the first card of the data is IRUNMOD in a 1011
format with a "1" for each part of the program desired and a "0" input other-
wise.
After the IRUNMOD card, place the data to be added (if the ADD and DUP
routines are selected).
Example:
If execution of the entire program is desired, use the following sequence
of cards:
ATTACH, TAPE 2, DATASAMPLES, PW=...
ATTACH, TAPE 3, DATATARGETS, PW=
ATTACH, TAPE 4, NONRAINTARGETS, PW=
ATTACH, TAPE 48, SPRAYDATES, PW=
ATTACH, TAPE 49, RAINDATES, PW=
ATTACH, BIN, PROCESSBIN, PW=
LDSET, LIB°FORTRAN/CRM.
BIN.
PURGE, TAPE 2.
CATALOG, TAPE 20, DATASAMPLES, PW=	 RP=999.
PURGE, TAPE 3.
CAT0L0G, TAPE 21, DATATARGETS, PW=..., RP=999.
PURGE, TAPE 4.
CATALOG, TAPE 21, DATARGETS, PW=..., RP=999.
PURGE, TAPE 4.
CATALOG, TAPE 22, NONRAINTARGETS, PW=..., RP=999.
ATTACH, A, PROCESSEDSAMPLES, PW=
PURGE, A.
CATALOG, TAPE 26, PROCESSDSAMPLES, PW=..., RP=999.
ATTACH, B, PROCESSEDTARGETS, PW=
PURGE, B.
CATALOG, TAPE 24, PROCESSED TARGETS, PW=...,RP=999.
ATTACH, C, PROCESSEDNONRAIN, PW=
PURGE, C.
CATALOG, TAPE 25, PROCESSED NONRAIN, PW=...,RP=999.
1111111
Data cards to be added
A-3

-------
Whenever the raw data files are changed be sure to check the files to assure
that they are correct. If something went wrong with the program so that
the data are lost, there are backup files for the raw sample and raw target
files called DATASAMPLESBACKUP and DATATARGETSBACKUP respectively. After
making sure the data were added correctly and is on permanent files, purge
the backup files, make copies of the new data files and catalog these copies
under the name DATASAMPLESBACKUP and DATATARGETSBACKUP. These files contain
all the data upon which the orchard model is based so be careful not to lose
them.
ALSO, back up the file NONRAINTARGETSBACKUP in a similar manner.
Subroutine ADD
Sorts the new data into samples, targets, and nonrain targets and then
merges the new data with the appropriate file (DATASAMPLES, DATATARGETS or
NONRAINTARGETS).
Targets are distinguished from samples by checking column seven of the
data card. All target data have a 6 in column 7 while sample data have a
number less than 6 in column 7.
Nonrainfall targets have a 1 in column 25 while spray targets have a
blank in column 25.
A description of the Fortran - Sort/Merge Interface can be found in
section 111-16 of the CDC Cyber Fortran Reference Manual.
Inputs to ADD
TAPE 2
TAPE 2
TAPE 3
TAPE 4
Outputs
TAPE 11
TAPE 12
TAPE 13
INPUT:	New Data Cards to be added
:	DATASAMPLES
:	DATATARGETS
:	NONRAINTARGETS
files to be updated
Raw Samples with new samples added
Raw Targets with new targets added
Raw Nonrain targets with new non-rain targets added
Subroutine DUP
DUP checks the data files for duplicates. Any two cards which have the
same first 13 characters are considered duplicates.
When duplicates are located the copy with the greatest number of blanks
is deleted. (In other words, the card with the most data is retained.)
DUP prints out which cards are deleted from the file and which cards are
kept. This listing should be checked to make sure the proper cards were
deleted.
Inputs to DUP
TAPE 11
TAPE 12
TAPE 13
Raw Samples
Raw Targets Output files from ADD containing duplicates
Non-rain Targets
A-4

-------
Outputs
TAPE 20 : DATASAMPLES
TAPE 21 : DATATARGETS	These files should be catalogued
TAPE 22 : NOURAINRAGETS
Subroutine HILOAVG
HILOAVG computes the ratio of the average amount of dislodgeable residue
per leaf surface area in the entire tree to amount of residue found in the
bottom portion of the trees. It does this by analyzing leaf samples from
year 3 (1978) as this was the first year that samples were taken from the
upper portion of the tree.
Leaf sample data cards have a 1 In column 10 if the sample was taken
from the upper portion of the tree. Samples taken from the lower portion of
the tree have a blank or a 2 in column 10 of the card.
To find the ratio of average tree to bottom tree amounts HILOAVG finds
the average for both upper and lower tree values for each plot on a single
day. The average tree value Is then taken to be the average of the value
found for the upper tree and the value for the lower tree.
ZTOP/Nj + IB0TT0M/N2
Avg. Tree Residue =	z	
where is the number of upper tree samples and N, is the number of bottom
tree samples. The Ratio is then calculated by dividing this average tree
value by the average value found for the lower tree.
HILOAVG then computes a rate at which the ratio decreases per day during
rainfall and non-rainfall intervals.
The ratio and the rate of change of the ratio are eventually going to
be used to calculate the average tree residue amounts from bottom tree
samples for years 1 and 2 (1976 and 1977).
Inputs to HILOAVG
TAPE 20 : PROCESSEDSAMPLES
Output: Currently output to printer only.
Subroutine CORRECT
CORRECT checks for soil samples and then makes corrections for:
(1)	weight extracted .LT. total weight
(2)	% efficiency of extraction
(3)	% efficiency of column cleanup
TFRWT = WTEXT + WTCOR
PENRES = TFRWT * PENRES/WT EXT
RESIDUE = PENRES * 1.1765 * 7.008
or
1.3072
7.008 =
2
210.25 cm	(area of sample)
2	2
3 x (10 cm	(10 cm = area of core sample))
A-5

-------
1.1765 = 1/.85 (.85 = Extraction efficiency)
1.3072 = l/(.85*.9) (.9 = column cleanup efficiency)
CORRECT also makes similar corrections for grass, litter, and moss.
Samples which, in comparison with others of the same type (e.g., region and
day) are determined to be statistical outliers, and discarded before the
PROCESSED files are created. Only grass and soil samples are subjected to
outiier procedures, because tree and litter/moss samples are already assembled
from multiple samples (e.g., upper/lower tree leaves, sums of litter and moss)
so the outlier procedure was not easily employed.
Example of Outlier processing
SSE = (SS„ - SST) = |
Yx - .200
Y2 = .387
*3 = .209
t = 1 ni=n=3
ry = .796
Y± = .265
EZY2 = (.200)2 + (.387)2 + (.209)2 = .233
ij
EY2 /n, = (.796)2/3 = .211
X • 1
SSE = .233 - .211 - .022
MSE - .022/(3-1) - .011
e. - (Y,, - Y, ) = .387 - .265 = .122
L	Xj X.
OUTLIER TEST eT¦// MS_ =¦ .122/ /.Oil
L	Zt
= 1.16
Y
.387 is OUTLIER with a = 2.01, because
1.160 > 1.155 (TEST STAT. FROM TABLE A -1)
A-6

-------



Table A-l.
Critical Values
for




Testing an
Outlying Observation

11
a=0.01
0.05
0.10
n
a=0.01
0.05
0.10


• * •
• • •
32
3.135
2.773
2.591

• • •
• • •
• • •
34
3.164
2.799
2.616
3
1.155
1.153
1.148
36
3.191
2.823
2.639
4
1.492
1.463
1.425
38
3.216
2.846
2.661
5
1.749
1.672
1.602
40
3.240
2.866
2.682
6
1,944
1.822
1.729
42
3.261
2.887
2.700
7
2.097
1.938
1.828
44
3.282
2.905
2.719
8
2.221
2.032
1.909
46
3.302
2.923
2.736
9
2,323
2.110
1.977
48
3.319
2.940
2.753
10
2.410
2.176
2.036
50
3.336
2.956
2.768
11
2.485
2.234
2.088
55
3.376
2.992
2.804
12
2.550
2.285
2.134
60
3.411
3.025
2.837
13
2.607
2.331
2.175
65
3.442
3.055
2.866
14
2.659
2.371
2.213
70
3.471
3.082
2.893
15
2.705
2.409
2.247
75
3.496
3.107
2.917
16
2.747
2.443
2.279
80
3.521
3.130
2.940
17
2.785
2.475
2.309
85
3.543
3.151
2.961
18
2.821
2.504
2.335
90
3.563
3.171
2.981
19
2.854
2.532
2.361
95
3.582
3.189
3.000
20
2.884
2.557
2.385
100
3.600
3.207
3.017
21
2.912
2.580
2.408
105
3.617
3.224
3.033
22
2.939
2.603
2.429
110
3.632
3.239
3.049
23
2.963
2.624
2.448
115
3.647
3.254
3.064
24
2.987
2.644
2.467
120
3.662
3.267
3.078
25
3.009
2.663
2.486
125
3.675
3.281
3.092
26
3.029
2.681
2.502
130
3.688
3.294
3.104
27
3.049
2.698
2.519
135
3.700
3.306
3.116
28
3.068
2.714
2.534
140
3.712
3.318
3.129
29
3.085
2.730
2.549
145
3.723
3.328
3.140
30
3.103
2.745
2.563




Values were extracted from Grubbs and Beck, Technometrics 14s847-854 (1972).
A-7

-------
Record Formats (Card Images) for Raw Corrected, and Processed Flies.
Table A-2. Raw Flies - data samples, data targets, non-rain targets
Format
Column	Description	I or F
1	year	I
2	spray periods	I
3	sample period within spray	I
4-6	Julian date	I
7	layer	I
8-9	plot //	I
10	(for yr. 3 leaves) upper or lower leaf sample	I
11-12	lab. code 0	1
13	region	I
14-17	yr., date of extraction	I
18-21	yr., date of analysis of dlslodgeable	I
22-27	extracted fresh wt. In gms.	F
28-33	extracted dry wt. In gms.^	F
34-39	penetrated residue ( g/cm ) /soils	F
I acetones
2
40-45	dlslodgeable residue ( /cm )}lvs., moss,
\grass, soil,	F
2 jlltter.targetg
46-51	area of sample in cm (if area ^ 210.25 cm )	F
52-57	SOILS / moisture sample wt. before dry	F
58-63	SOILS / moisture sample wt. after dry	F
64	set // for targets	I
65-70	total sample fresh wt. in gms.	F
71	° 1 if residue analyzed	I
72	column clean up flag	I
73	fleaker flag	I
74	sample 0	I
75-78	yr., data acetone g.c.	I
79	if soil	I
80	if sample cup empty	I
A-8

-------
Table A-4. Processed Files - targets, samples, non-rain targets.
Column	Description	Format
I or F
1	year
2	spray period
3	sample period
4-6	Julian date
7	layer
8-9	plot #
13	region
21-22	# samples
23-31	avg. penetrated residue	F
32-40	std. dev. of penetrated samples	F
41-49	std. error of penetrated sampled	F
50-58	sum of squares of penetrated samples	F
61-62	# dislodgeable samples	1
63-71	avg. dislodgeable residue	F
72-80	std. dev. of dislodgeable samples	F
81-89	std. error of dislodgeable samples	F
90-98	sum of squares of dislodgeable samples	F
ORCDATA, The Parameterization Routines
Figure A-2 shows the data flow through the parameterization routines.
The control card structure to run the ORCDATA program is shown below:
ATTACH, BIN, ORCDATABIN, PW = APPLE.
BIN.
\
Data input
«7
9
Input card sequence
1)	Beginning and ending year of span to be studied	(212)
2)	IRUNMOD (1 if routine is desired, 0 otherwise)	(2011)
Example: If spray distribution from analysis of samples and targets for
years 1 and 2 is desired, use these cards.
ATTACH, BIN, ORCDATABIN, PW - APPLE.
BIN.
7g
9
1 2
0000000000000100000
67
9
A-9

-------
c
'r-4
r_r^
cz>
CZ>-J
CZ>-T-f
1 I I
(
r=v 1
(	i
J	---*	
^ „	) Targets
Leaf
M
f—t-)
ll _L.
I !

f—j Findev
JZ
Adapt I
~T~
Adjust
Adapt I!
h.ra
T~
Ratepen
[-0-"
h-
-------
Unit numbers for ORCDATA - these files are attached, purged, and recatalogued
Inside the program itself.
TAPE
PERMANENT FILE
1
PROCESSED SAMPLES
2
PROCESSED TARGETS
3
PLOTAREAS
4
SPRAYRATES
5
TREEINTERCEPTINDEX
7
SAMPLEDATES
8
SPRAYDATES
9
RAINDATES
14
PESTDIST
16
PENETRATIONRATES
L&
NRMOVE -
18
SITUATIONREPORT
19
PAIREDDATASAMPLES -
IRUNMOD
SUBMODEL
1

2

3

4
Penetration
5

6

7
Leaf Areas
8

9

10

11

12

14
Spraydlstribution Analysis & targets
15
Spray 1 Analysis of samples
16

17
Alternate Spray 1 calculations
18
Situation Report
19

20
Non-rain & rain intervals
A-11

-------
Subroutine Leaf
Leaf calculates the tree Intercept index for each plot from measured
leaf area per tree, number of trees per plot, and areas of canopy regions
In each plot.	^ 2
Measured leaf area per tree Is a constant equal to A x 10 cm. , calcu-
lated by removing all the leaves from a representative tree, counting them,
and multiplying by the average area per leaf.
The leaf area index "LAI" is calculated by
NTREES. x LAPTREE
T AT o	i ¦ —¦ ¦	—
i CANOPY AREA1
where: NTREES. = # of trees in plot i
5 2
LAPTREE = leaf area per tree = 4 x 10 cm.
CANOPY AREA^ = ground area of canopy region for plot i
The tree intercept index "Til" is equal to the LAI multiplied by 1.15.
This is based on the assumption that the trunk region is 15% of the leaf
area index.
Input files used
tape 3: plot areas
Output files
tape 5: tree intercept index, format = 6F10.3
Subroutine Targets
Targets use target samples to calculate the average amount of pesticide
(in ug/cm ) for a given year, region (alleys 1 throguh 3, canopy, trunk), plot
(3 through 6) based on the distribution of targets for the first spray period
of each spray (i.e., only spray dates). It also calculates the standard
deviation, standard error, sample size for each average, and HPESTD", the
proportion of pesticide allocated to each region of a plot per dose applied
to that plot; along with "HCANOPY", the proportion of spray not accounted
for, or allocated to canopy leaves.
Input files used
tape 2: processed targets
tape 3: plot areas
tape 4: sprayrates
Output files - none
HPESTD, the proportion of pesticide allocated to a given region of a
plot per dose of pesticide applied to that plot for a specific year is cal-
culated by
A-12

-------
HPESTD.
vc
AVGT,,. x AREA..
1.1k 	1.1
AMTSPRAY,,
	lk
lj	NSPRAYS
where: NSPRAYS = # of sprays for that year
2
AVGT,,. »• average pesticide in yg/cm for plot 1, region j, on
Ijk
spray number k
AREA = proportion of aea of region j of plot 1
2
AMTSPRAY^ = amount of pesticide in Mg/cm sprayed to plot 1 on
spray number k.
HCANOPY, the proportion of spray not accounted for is calculated by
NREGIONS
HCANOPY1 = 1.0 - DRIFT - \ ^ HPESTD^
where: DRIFT = average proportion of drift for the specific year, from
subroutine ALLOC, assumed to be the average proportion
for each plot
NREGIONS = # of horizontal regions (alleys 1-3, canopy, trunk) ° 5.
SUBROUTINE SPRAY 1
SPRAY 1 analyzes the first spray of each year £rom data samples. It
calculates the average amount of pesticide in pg/cm for each year, region
(including trunk, canopy, and alley, where alley regions 1 through 3 merged
into one region) and plot (3 through 6). It also gives the standard deviation,
standard error, and sample size for each average, and "VPESTD," the proportion
of pesticide applied to each region and layer of a plot per dose of pesticide
applied to that plot.
If IRUNMOD (17) is turned off (I.e., IRUNMOD (17) °» 0), some additional
spray calculations are made which use initial penetration rates.
(1) the vertical proportions of pesticide in each plot, layer, and region
per dose of pesticide applied, "VPESTD,11 is averaged over plots to
give "PROPD," the proportion of pesticide (dislodgeable) for each
layer (1 through 5) and each region (alley and canopy; here trunk
Is assunted^^Q^ included in canopy region) per dose of pesticide
applied. V-*
PROPDjk - NPLOTS
PROPP, the proportion of pesticide (penetrated) for each layer (1 through
5) and region (alley, canopy) per dose of pesticide applied, Is estimated
using initial penetration rates for each layer
INITPEN * PROPD..
PROPP,. 		-	J*
Jk (1.0 - INITPEN, )
k
where INITPEN^ is ratio of penetrated residue to total residue (penetrated-f
dislodgeable) for layer k on the first spray of a season.
Next Airloss is estimated as the proportion not accounted for
A-13

-------
Airloss - 1.0 - x i x (PROPD^ + PROPP^)
NLAYERS / NREGIONS
El E
i=i \ j=i
(2) Airloss Is calculated on a plot by plot basis using targets to
calculate the proportion reaching the ground and leaf samples to
calculate the proportion going to the tree.
where
AIRLOSS^ = 1.0 -	HPESTD^ - VLEAF1
AIRLOSS^ Is the airborne loss from plot^
HPESTD Is described In subroutine TARGETS description
VLEAF, = (VPESTDi41 + VPESTD^)
ITis
Q1
INITPEN(l)
(1 - INITPEN(l))
Q2
Ql, quantity 1 Is the vertical proportion of pesticide allocated
to the canopy leaves plus trunk leaves regions per dose applied
to that plot as previously calculated by SPRAY1. Since the trunk
region Is summed In with the canopy, the sum must be divided by
1.15 which moves the 15% assumed trunk area when the tree Intercept
Index was applied, Q2, accounts for penetrated residue of leaf
values
(3) Finally some calculations are made which have to do with how much
pesticide is in the ground layers. The proportion of proportion of
dislodgable residue in the ground layers out of the total dislodgable
residue (all layers) is calculated for each ground layer out of
the total dislodgable residue In all ground layers (layers 1 through
4) is calculated for the canopy and alley regions.
If IRUNMOD (17) is turned on (i.e. IRUNMOD (17) ° 1),-averages and vertical
proportions, "VPESTD," will be calculated based on the sum of measured dis-
lodgable and penetrated averages.
Input Files Used
tape 1:	processed samples
tape 2:	processed targets
tape 3:	plot areas
tape 4:	spray rates
tape 16:	penetration rates
Output Files - none
A-14

-------
Subroutine Sprdist
Sprdlst Is used to estimate drift and spray distribution. It uses only
first spray dates of a season and other spray dates that have a non-rain
interval between the preceding sample date and that spray date. If the
interval between the sample date and the spray date is a rainfall period, it
is flagged in output saying it cannot be used.
The following steps describe the procedure for estimating the drift
and spray distribution.
(1)	The non-rain movement matrix is applied to sample values of the
sample date preceding a spray date to give the amount of pesticide
present immediately before the spray occurs. If a spray date is
the first of the season these values are all zero.
(2)	The amounts of pesticide before the spray are subtracted from the
sample values of the spray date In each region and laye£. This
gives the contribution of dlslodgable residue (in yg/cm ) on that
date giving the proportion of pesticide in each compartment to
the total amount of pesticide applied.
(A) Proportion drift for this spray is estimated to be 1 minus the sum
of the proportions of pesticide applied to each compartment.
Input files used
tape 1
tape 4
tape 3
tape 7
tape 8
tape 9
tape 17
processed samples
spray rates
plot areas
sample dates
spray dates
rain dates
nrmove
Output files
tape 14: pest dlst
The values of spray distribution used to write the output file (tape 14)
pestdist are the average spray distributions for each year.
To calculate drift for the forward model, a regression was run on drift
vs. windspeed. Windspeeds for each spraydate were estimated from Grand Rapids
weather data, and drifts used were the proportion drift of a spray calculated
by SPRDIST. The regression yielded
DRIFT = 0.03409 * WSGR + 0.09314
Penetration Subroutine
For litter: The following calculations are performed by the penetration
subroutine:
(1) dP
Kin D. - KP PT(t) + K,_tT(0)
dt	2 L L	iPL
where:
dt
Rate of change in penetrated residues
A-15

-------
and
(2) dD.
Kin = Penetration rate of D , litter DSL residues
Li
KP^ = Attenuation rate of P^, litter Pen residues
= Instantaneous penetration of total residues applied (Day 0)
= Dislodgeable residue in layer L
P^ = Penetrated residue In layer L
TI = -K1" »!<*> - KdL DL(t> - Kmo\(t) + £ Kei Di(t)
were: dD^
» Rate of change of litter DSL residues
Kd_ B Attenuation rate (other than penetration) of DT
L	L)
DSL residues
K ° Rate of movement out of DSL residues (D.)
mo	L
= D^ (t) = Sum of DSL residues transfered from above
layers to Layer L
(3)	£Ktl Dl(t) - Kt „ DLV(t) + Ktg Dg(t)
For moss:
(4)	IKtl Dl(t) - KtLy DLV(t) + Ktg Dg(t) + KtL DL(t)
where:
Kt^v = Transfer rate of DSL residues in tree, D , to litter
(a separate	for transfer to alley litter and canopy litter)
K ° Transfer rate of grass DSL residues. D , to litter
tg	6	gl
^tL = 
-------
d. Then the penetration rate, K , for litter or moss can be estimated
from the difference between the litter (or moss) DSL residue
attenuation rate (from ADAPT) and the grass DSL residue attenua-
tion rate (from ADAPT).
The following Kin's were determine for litter and moss
Canopy	Alley
Litter	.026	.055
Moss	.227	.498
was estimated from the observed ratio of penetrated to total residues
for samples taken 4 hrs. following application (day 0). (Use matched pairs
on spray dates).
KlpL = .297 (78)
KJT1_ = .214 (78 data only)	.232 (76 + 78 data)
lrra
P(t) - P(t-l) (l-Kp) + (DSL^(t-l) . (l"kin))
For P(t-l) - P(tQ), P(t-l) = Kip . DSL^o)
NOTE: Penetrated residues on day t = penetrated residues on day t-1
times daily degradation rate of DSL residues present on day t-1
times dally penetration rate,
ex: (for litter)
DSLL(t) = DSI^ (t-1) • (1-Kd) - DSI^ (t-1) • (1-K^) +
DSLLV(t"1) * (1_KtLV) + DSLg (t_1) * (1_Ktg)
NOTE: Litter DSL residues at day t = litter DSL residues on day t-1 times
daily deg rate - litter DSL res on Day t-1 times dally transfer rate
+ grass DSL res on day t-1 times daily transfer rate.
Subroutine Allyest
Allyest uses target proportions to estimate missing alley samples from
alley samples that are present. If all alleys (1, 2, and 3) are missing it
cannot estimate anything. Therefore, while this routine is operating we
have either all three alley values (if one or two different alley regions are
missing we estimate what's missing from what is present) or we have no
alley data at all for a given plot and day.
To fill in missing data we set the missing value equal to the value
present weighted by the ratio of alley area for value present to alley area
of missing value and weighted by the ratio of target proportion of missing
alley to target proportion of alley present.
Next, the 3 alley region averages are averaged across alley regions
weighted by alley region area and stored Into alley region #3 (which now
represents all alley regions).
A-17

-------
The POEM Model User's Guide
Introduction
The Pesticide Orchard Ecosystem Model was constructed for two major
reasons:
(1)	to assist in presenting the results of a study of the fate
of a particular compound under field conditions and its
effects on some significant organisms using laboratory and
field data, and
(2)	to allow users to explore the implications of changes in the
model's parameters on the model outcomes (both fate and
effects), including limited extrapolation to other compounds
and/or field conditions.
The various submodels included in POEM and the methods used to parame-
terize them are described in the body of this report and in a series of
publications. This guide is intended only to supplement them, as an aid
to persons desiring to run the associated computer programs. No explana-
tion is offered here of the methods used to derive the parameters included
as default values, nor is there any presentation of results (see the body
of the report).
POEM is a modular FORTRAN IV program, composed of many subunits which
may be invoked or left idle during any particular run. In general, POEM
units not needed should be "turned off" to minimize computer costs. POEM
may be used to explore the behavior of the fate model using "average" values
independent of temperature and relative humidity, or may be instructed to
re-calculate certain parameters daily based on environmental conditions.
B-l

-------
B-2
Many of POEM's parameters for azinphosmethyl may be altered in order to
incorporate new data or determine the effects of such changes on pesticide
distribution or effects on the modeled populations. Other alterations
allow the user to enter information to simulate a different compound.
Weather data may be entered by the user, or the program may be used to
generate simulated weather data. The user must supply pesticide applica-
tion information, unless the default regime used in our field studies is
desired.
In its simplest use, POEM can be used with minimal input data and
will produce predictions of the pesticide distributions and populations to
be found in our study orchard under our measured environmental conditions
from 1976-1978. POEM produces output files suitable for plotting, and two
plotting programs are included (one for pesticide residues, one for popu-
lations). Alteration of compound characteristics, weather data, or process-
specific parameters may require the user to enter much larger amounts of
data, although it is relatively easy and often enlightening to explore the
effects of changing one or a few parameters at a time.
The next few sections describe the general nature of the units in the
fate model, followed by directions for and examples of reparflmeterization.
A brief description of the organism submodels is included under Organism
Graphs; however, the automatic alteration of organism model parameters has
not yet been completely implemented. See the relevant sections of the body
of the report for more information.

-------
B -3
Attenuation of Pesticide
The attenuation model simulates attenuation, other than penetration,
in the various compartments. Each compartment has an attenuation propor-
tion—the proportion of pesticide within or entering that compartment
during one day that is attenuated (see below) from that compartment during
a one day time step at standard conditions. The standard conditions
approximate the average of those prevailing during collection of the field
data used to parameterize this model. These conditions are:
1)	environmental conditions
-	soil PH » 7.0
-	% organic matter of soil = 6%
2)	compound properties
-	molecular wt. =317
-	vapor pressure = 10"^ atm.
Each compartment's attenuation proportion is further broken down into
5 attenuation processes: hydrolysis, photolysis, oxidation, volatilization,
and microbial degradation. Therefore each compartment has an attenuation
proportion for each of the five attenuation processes. The sum of the
attenuation proportions for each process of a given compartment is equal
to the overall attenuation proportion for that compartment.
The user has the option of using the overall attenuation proportions
for each compartment directly as the proportion of pesticide lost each day
due to attenuation. Or the user may allow individual attenuation propor-
tions of the attenuation process to vary with environmental conditions.
Thus, the sum of the attenuation proportions of the five attenuation pro-
cesses over a compartment, which yields the day's overall attenuation pro-
portion for that compartment will vary daily with environmental conditions.

-------
B -4
In addition, the user may alter certain default parameters to represent
properties of a compound other than azinphosmethyl, the compound used in
the field studies. The overall attenuation proportions, whether constant or
varied with daily conditions, are applied each day and represent the propor-
tions of pesticide attenuated daily from each compartment.

-------
B-5
Non-rainfall Movement of Pesticide
Subroutine Nonrain simulates the daily movement of pesticide within
the orchard independent of rainfall during the period (i.e., it is applied
every day). It operates conservatively; that is, it moves residues without
any losses, which must therefore be accounted for elsewhere in the model
(see Attenuation description in previous section). It operates only on
dislodgeable residues, except that residues entering the soil layers are
considered to be "penetrated residues." Pesticide is assumed to move only
from a higher layer to lower layers, and pesticide moving to the alley region
from the canopy region may only come from the canopy leaves compartment.
The order of layers from highest layer to lowest layer is leaves, grass,
litter-moss, and soil. The pesticide is moved via a lower triangular
matrix "NRMOVE".
CANOPY
ALLEY
FROM

LEAVES
A


.ttOSS




£ GRASS
B
H





B LITTER-MOSS
C
I
K
soi^-




SOIL
D
J
L
M

>


TO
LEAVES







-flOSS
GRASS
>-
E




N

Zl LITTER-MOSS
<
F




0
Q
sott-
SOIL
G




P
R
s

-------
B -6
Each entry in the matrix represents a proportion of the total amount
of dislodgeable pesticide residue present on a given day that moves from
one compartment to another or (for diagonal entries) that remains in the
same compartment. The latter (diagonal) entries in the matrix are
A,H,K,M,N,Q,S. A single column in the matrix represents the movement from
an individual compartment over a one day time step. Being proportions of
the total amount of what was in a compartment, all entries of any column
in the matrix must sum to one. Since there is only one entry for each of
the canopy soil and alley soil compartments, entry M=1.0 and entry S=1.0
by the definition that the columns must sum to one (Implying that pesticide
undergoes no non-rain-induced movement out of soil layers).
For example:
B is the proportion of pesticide moving from the canopy leaves com-
partment to the canopy grass compartment during a one day time step. H is
the proportion of the canopy grass compartment that remains in the canopy
grass compartment during a one day time step. Thus the entire amount of
dislodgeable pesticide in the canopy grass on day t+1 (excluding attenua-
tion) is equal to the B times the amount of dislodgeable pesticide in the
canopy leaves on day t plus H times the amount of dislodgeable pesticide
in the canopy grass on day t:
CanGr(t+l) = B * CanLvs(t) + H * CanGr(t)
The absolute amount of pesticide for any compartment of day t+1 can be
represented by the sum of each entry in that compartment's "TO" row of the
matrix, multiplied by the absolute amount of pesticide on day t in the
appropriate compartment (neglecting attenuation and rainfall-induced movement).
It is important that pesticide values be converted from their normal
p
form of ug/cm ground area to an absolute amount (in pg.) of pesticide when

-------
B -7
applying the matrix. This is because canopy and alley regions do not neces-
sarily have the same ground areas and because the canopy leaves column may
move pesticide from the canopy region into the alley region. In order for
2
the system to be conservative, the pesticide values in yg/cm ground area
are converted to amount of pesticide in yg. before the matrix is applied
2
and converted back to yg/cm ground area afterwards.
The non-rain movement matrix is applied every day regardless of rain-
fall. It represents the daily movement, other than by rainfall, of the
pesticide within the orchard system.
The default values of the non-rain movement matrix are those parame-
terized from field data.
It is possible to change the matrix during reparameterization in a
variety of ways as described under the Movement Parameters category of
the Reparameterization section.

-------
B -8
Movement of Pesticide Due to Rainfall
Subroutine Rain simulates movement of pesticide within the system due
to rainfall. There are two rainfall matrices that have the same properties
and are applied similarly to the non-rain movement matrix. One of the two
matrices is applied on each day of measurable rainfall. One represents
light rainfall movement, the other represents heavy rainfall movement. A
heavy rain is classified as greater than or equal to 10mm on a given day.
Less than 10mm rain is considered a light rain. Rainfall movement is defined
as additional movement beyond the nonrain movement extracted daily.
It is also possible to change either or both rainfall matrices from
their default conditions. See Movement Parameters under the Reparameteriza-
tion section for further details.

-------
B-9
Application of Pesticide
Subroutine Sprayer simulates the application of pesticide to the
orchard. The total amount of pesticide (in pg.) that reaches the orchard
is first calculated. It is equal to the amount applied (in Kg./Ha,) multi-
plied times the area of the orchard (in Ha.) times a Kilogram to microgram
conversion factor (10 ) times the compound formulation fraction times the
proportion of pesticide that doesn't drift away from the orchard site.
This yields the absolute amount of pesticide (in micrograms) that reached
the orchard site. The proportion of drift is calculated based on the wind
speed. Once the absolute amount of pesticide reaching the orchard is known,
it is divided among the compartments using a spray distribution vector,
"PDIST", where an entry represents the proportion of pesticide going to a
specific compartment from the total amount of pesticide reaching the orchard.
Thus, the sum of all entries in the spray distribution vector is equal to
one.
However, some of the pesticide that reaches a compartment becomes pene-
trated residue. The spray distribution only gives the amount of pesticide
that reaches a compartment. The pesticide that reaches each compartment,
except for soil layers (all pesticide in the soil is assumed-penetrated),
is further divided into dislodgeable and penetrated residues. Finally,
2
units are converted from pg to ug/cm by dividing the absolute amount of
pesticide in each layer by the area of its region (canopy or alley).
The spray distribution vector may be altered to a new form in the
Movement Parameters section of the Reparameterization section. Likewise,
initial penetrations of a layer may also be changed in the Penetration
Parameters section of the Reparameterization section.

-------
B -10
Penetrated Residue Attenuation
Subroutine Pentrat simulates the penetration of dislodgeable residues
to become penetrated residues and the attenuation of penetrated residues.
There is a daily penetration rate representing the proportion of dislodge-
able residue that penetrates each day, and a penetrated residue attenua-
tion proportion representing the proportion of penetrated residue that
attenuates daily. Each of the proportions is applied daily regardless of
spraying. Also, the proportions may be changed in the Penetration Para-
meters category of Reparameterization.

-------
B-n
Mowing of the Orchard
Subroutine Mower simulates mowing of the orchard. It moves a propor-
tion of pesticide from the grass layer to the litter-moss layer. The pro-
portion of grass moved, "PGRASM", may be changed to any proportion (0.0 to
1.0) in the movement parameters category during reparameterization of the
model. The default value of "PGRASM" for the field site is 0.0. The pro-
portion of the grass layer moving to the litter-moss layer is assumed
linear with the proportion of pesticide moving from the grass layer to the
litter-moss layer because of mowing.
The maximum number of mow dates allowed for any one season is ten.
The user may select any dates he wishes. The mow dates for a season are
entered just before the beginning of that season.

-------
B -12
Leaf Growth of Trees
Subroutine Trees computes the tree interception index "Til", a value
which converts the ground area under the canopy region to tree surface area
(including leaves). Growth in surface area is driven by degree days until
petal fall (1100 degree days, base 43°F), and then is considered linear
until 56 days after petal fall at which time the Til has reached maximum.

-------
B -13
Running the Model
The user has many options in running P0EM--some of these options call
for the use of pre-stored files of weather data. For example, if the
weather simulator is not used, POEM will attach a file called WTHRSTATSNEW,
which should contain the weather information you wish to use (a file with
3 years of data for Grand Rapids, MI is included). Similarly, if the soil
moisture model is not used, POEM will seek soil moisture data from a file
called PWSM0IST2. (On non-MSU systems, subroutine PARMINT should be
altered to conform to local procedures for accessing these files. For
more information, see section entitled "Attaching Tapes".)
Output from POEM is available in two forms—printed tables and graphs.
Graphical output is provided by specifying to POEM that graphical output
is desired. Two types of graphs are available: pesticide levels versus
time and population levels versus time. Pesticide level graphs are
obtained by turning on option (14) (see below) and post-processing using
program DATAPLT (available on the MSU CYBER as 0RCHARDCALPL0T6 for
CALCOMP plotter output or ORCHARDTEKPLOTB for Tektronix terminal display
via PL0T-10 subroutines). Population graphs are produced from files gener-
ated whenever the organism submodel (option 1) is on, using program ORGPLOT
(again, available on MSU's CYBER in CALCOMP and Tektronix versions). For
more information on the plotting programs, see section "Producing Graphs"
below.
Printed output to display pesticide levels and/or the current rates
of pesticide movement and attenuation are available in POEM as described
below. Printed output from organism models is not specifically provided
for, but is available by listing the plotting files generated by POEM.

-------
B -14
Options and submodels turned on by default as execution begins are
listed, followed by a list of other available options. The user then has
the opportunity to turn off any options that are on by default or turn on
any of the options that are off by default, in response to queries from
the program. The submodels that each option represents are:
option 1) Organism submodels. "1" for any one or combi-
nation of the 4 organism models provided (isopod,
collembolan, spider, or earthworm); further
options will be presented, or "2" otherwise,
option 2) Mower submodel. "1" for periodic mowing of
the orchard each season. The maximum number
of mows allowed each year is 10. The mow
dates for a given year are entered at the
start of that year,
option 3) Attenuation submodel. "1" to use attenua-
tion as a lumped process or to use any of
the attenuation submodels (microbial degra-
dation, hydrolysis, photolysis, volatili-
zation, oxidation); "2" to bypass pesticide
attenuation calculations.
(Any further reparameterization concerning these attenuation processes
is done in the general category of attenuation parameters, below.)
option 4) Weather submodel. "1" to turn on simulated
weather. To allow true recorded weather con-
ditions, the weather simulator must be turned
off. In that case, the recorded weather must

-------
B -15
be provided on tape 4 and must contain the
following conditions for each day in a re-
quired format:
-	minimum and maximum air temperature (°F)
-	minimum and maximum percent rel. humidity
-	average wind speed (mi/hr), and total
rain (mm)
See Attaching Tapes for more information,
option 5) Nonrain Movement Submodel. "1" to select
daily movement of pesticide; "2" suppresses
movement except on days with rain,
option 6) Sprayer Submodel. "1" for periodic spraying
of pesticide during the season. The maximum
number of sprays each year is ten. The spray
dates and spray amounts for a given year are
entered at the start of that year, unless de-
fault dates and amounts are desired,
option 7) Rainfall submodel. "1" for additional pesti-
cide movement (other than daily nonrain"move-
ment) due to rainfall if rainfall is present;
"2" suppresses pesticide movement in response
to rainfall.
option 8) Penetration submodel. "1" for pesticide
penetration submodel. Any further reparame-
terization concerning penetration (see text)
is found in the general category of penetration

-------
B-16
parameters. If option 8 is off, the proportions
of daily penetration are assumed to be accounted
for by other attenuation processes, provided
option 3 is turned on.
option 9) Soil moisture model. "1" for simulated soil
moisture based on amount of rainfall. True
recorded soil moisture values may be provided,
but the simulated soil moisture model must be
turned off. If the soil moisture model is off,
recorded soil moistures must be provided on
tape 9. See Attaching Tapes for more details,
option 10) Tree growth submodel. "1" for daily leaf and
tree growth based on degree days,
option 11) Movement Matrix printout. Provides daily printout
of a matrix which combines the non-rain movement
matrix with the attenuation matrix. This matrix
captures the net effects of the environment-
dependent processes acting on the pesticide,
under non-rainfall conditions, as represented
by the model. Beginning and ending dates for
printout of this matrix for a given year are
entered at the start of that year. This option
is provided to see movement and attenuation
that vary daily with environmental conditions,
as influenced by compound properties,
option 12) Soil Transportation model. "1" selects more
detailed model of soil layer.

-------
B -17
option 13} Pesticide Output. "1" selects printing periodic
pesticide values; mow, rain, and spray date flags
are also printed,
option 14) Plotting Output. "1" writes formatted file of
pesticide values to tape 1, file of spray dates
to tape 2, and file of rain dates and amounts
to tape 3 for plotting package.
See Plotting Package for more details,
option 15) Event dates. "1" allows the user to enter
dates of events of sprays and mows. If this
option is off, event dates are those from field
data used to parameterize this program,
option 16) Variable attenuation. "1" allows daily attenu-
ation proportions of the various attenuation
processes to vary daily with environmental con-
ditions, and to reflect compound properties
other than those of azinphosmethyl. If option
16 is off, attenuation proportions will be
identical each day, and based on the measured
values for azinphosmethyl.
If the user selects the organism model option, he must then select
one or any combination of the four organism models (isopod, collembolan,
spider, earthworm). He is asked individually for each organism model to
be turned on or off.
If organism models are desired without the effects of pesticide, then
the sprayer (option 6) should be off. Therefore, no pesticide is sprayed
to the orchard.

-------
3 -18
It is suggested that if the sprayer option is off, then options 2, 3,
5, 7, 8, 10, 11, 12, 13, 14 also be off since all of these submodel options
pertain to movement or output of pesticide and no pesticide can be present
without the sprayer submodel. Additionally, option 16 deals with attenua-
tion and has no effect unless option 3 is on.
Sample
THE FOLLOWING OPTIONS ARE ON BY DEFAULT
ORGANISM SUBHODEL
ATTENUATION SUBMODEL
NON-RAIN HOVEMENT SUBMODEL
PESTICIDE APPLICATION SUBMODEL
RAINFALL MOVEMENT SUBMODEL
PENETRATION SUBMODEL
WRITE FILES FOR PLOTTING PACKAGE
VARIABLE ATTENUATION
OTHER AVAILABLE OPTIONS ARE
MOUER SUBMODEL
WEATHER SIMULATOR
SOIL HOISTURE SIMULATOR
LEAF GROWTH SUBMODEL
MOVEMENT MATRIX DAILY OUTPUT
SOIL TRANSPORTATION SUBMODEL
PESTICIDE PRINTOUT
NON-DEFAULT EVENT DATES -SPRAYS.MOWS.ETC
DO YOU WISH TO TURN ON OR OFF ANY OPTIONS?
(1) YES (2) NO 1
ENTER A VALUE FOR EACH OPTION
(1) FOR OPTION ON <2) FOR OPTION OFF
OPTION
1
ORGANISM SUBMODEL 2
OPTION
2
MOWER SUBMODEL 1
OPTION
3
ATTENUATION SUBMODEL 1
OPTION
4
WEATHER SIMULATOR 2
OPTION
5
NON-RAIN MOVEMENT SUBMODEL 1
OPTION
6
PESTICIDE APPLICATION SUBMODEL 1
OPTION
7
RAINFALL MOVEMENT SUBMODEL 1
OPTION
8
PENETRATION SUBMODEL 1
OPTION
9
SOIL HOISTURE SIMULATOR 2
OPTION
10
LEAF GROWTH SUBMODEL 2
OPTION
11
MOVEMENT MATRIX DAILY OUTPUT 1
OPTION
12
SOIL TRANSPORTATION SUBMODEL 2
OPTION
13
PESTICIDE PRINTOUT 1
OPTION
14
WRITE FILES FOR PLOTTING PACKAGE 2
OPTION
15
NON-DEFAULT EVENT DATES -SPRAYSfHOWS.ETC 1
OPTION
16
VARIABLE ATTENUATION 1

-------
B -19
After the correct submodel choices have been selected, other simu-
lation parameters are entered.
First, the user is asked how many years to run the model. The max-
imum number of years is ten, and the model must be run at least one year.
Second, if the user has chosen option 13 of the on-off vector, he is
asked how often in days he would like pesticide values printed to output.
Selecting pesticide output (option 13) automatically causes printing of
mow, spray, and rain dates where they are appropriate, along with pesti-
cide values on spray dates. If the user inputs "10" for the pesticide
value printout increment, pesticide values will also be printed to output
every ten days.
Example
HOW MANY YEARS DO YOU UISH TO SIMJLATEC ONE TO TEN YEARS)1
HOW OFTEN (IN DAYS) MOULD YOU LIKE PESTICIDE VALUES PRINTED TO OUTPUT 1

-------
B -20
Repa rameter i za t i o n
General Categories of Reparameterization
Reparameterization is divided into five general categories. If the
user chooses to reparameterize, he must select one of these general cate
gories. The five categories are:
1)	physical site characteristics
-	basic site characteristics (areas, # trees, % alley
and canopy area)
-	site characteristics needed for soil moisture model
(altitude, latitude, soil moisture, field capacity, etc.)
2)	attenuation parameters
-	attenuation proportions
-	environmental properties (soil pH, % organic matter)
-	coefficients of equations dependent on environmental
conditions
-	compound properties
3)	penetration parameters
-	initial penetration rates
-	daily penetration rates
-	penetrated residue attenuation rates
4)	pesticide movement parameters
-	spray distribution
-	non-rain movement matrix
-	rainfall movement matrix
-	proportion of grass mowed

-------
B-21
5) organism parameters
-	isopod parameters
-	collembolan parameters
-	spider parameters
-	earthworm parameters
When entering options to select from, the user inputs the number of
the option he wishes. When the user is asked for more than one input value
for a given line, each of the values must be separated by blank(s) or a
comma.
Example
DO YOU WANT TO REPARAMETERIZE THE MODEL (DEFAULT VALUES ARE BASED
ON ESTIMATES FOR AZINPHOSMETHYL AS OUTHION SO U.P.)
<1) YES (2 > NO 1
DO YOU WISH TO CHANGE <1> PHYSICAL CHARACTERISTICS OF SITE
< 2 > ATTENUATION PARAMETERS <3> PENETRATION PARAMETERS
(4) MOVEMENT PARAMETERS <5> ORGANISM PARAMETERS 2
After a general category is selected and reparameterization is com-
pleted for that category, the user will be asked if he wishes to continue
reparameterization. If he selects "NO", execution will begin. However, if
he selects "YES", he may select any general category originally offered in-
cluding the one just completed.
All dates asked for should be entered in a Julian date format (a number
from 1 to 365 representing which day of the year it is; i.e., February 1
would be day 32 because it is the 32nd day of the year). Leap years are
ignored, so day 365 is the maximum date for any year.

-------
B-22
1. Physical Characteristics of Orchard Site
This category first asks for basic physical characteristics defining
the site. Values for total area, of site, amount of site that is canopy
region (remainder is alley region), and the number of trees in the site
requested one at a time. When asking for one of these values, the
model will list the default condition. Default conditions are those
of the field study site for which the model was originally parameterized.
If the soil moisture submodel (option 9) is on, additional physical
characteristics (initial percent soil moisture, field capacity for soil
moisture, altitude of site, latitude of site, etc.) necessary for the soil
moisture submodel are asked for in the same manner as the other physical
site characteristics.
Example
DO YOU UISH TO CHANGE (1) PHYSICAL CHARACTERISTICS OF SITE
(2> ATTENUATION PARAMETERS (3) PENETRATION PARAMETERS
(4> MOVEMENT PARAMETERS (5) ORGANISM PARAMETERS 1
HOU MANY HECTARES OF GROUND AREA UOULD YOU LIKE YOUR ORCHARD?
(CURRENT AREA = .0812 HA.) 1.0
HOU MUCH (IN HECTARES) OF THE 1.0000 HECTARE ORCHARD DILI BE CANOPY REGION
(CURRENT CANOPY AREA = .0302 HA.)
(THE REMAINDER OF AREA IS CONSIDERED ALLEY REGION) .40
HOU MANY TREES ARE IN THIS ORCHARD?
(CURRENT NUMBER OF TREES = 40)800
ENTER PERCENT U/U SOIL MOISTURE CONTENT AT THE FIRST DAY
(CURRENT INITIAL SOIL MOISTURE = 30.00 PERCENT U/U > 25.
ENTER PERCENT U/U SOIL MOISTURE FIELD CAPACITY
(CURRENT FIELD CAPACITY = 60.00 PERCENT U/U) 60
ENTER BULK DENSITY (IN GM./CM**3) OF SOIL
(CURRENT BULK DENSITY = 1.500 GM./CM**3) 1.4
ENTER ALTITUDE (IN METERS) OF SITE
(CURRENT ALT I TUI't = 200.000 M) ?4<>

-------
B-23
ENTER LATITUDE OF SITE
(CURRENT LATITUDE = 40.00 DEGREES) 40.
ENTER DIFFERENCE IN MEAN AIR TEMPERATURES (DEGREES C) BETWEEN
WARMEST AND COOLEST MONTHS OF THE YEAR
(CURRENT DIFFERENCE = 27.80 C) 28.
ENTER EVAPOTRANSPIRATION FACTOR DUE TO FLORA
(CURRENT EVAPOTRANSPIRATION FACTOR = 1.00) 1.00
ANY MORE REPARAMETERIZATION? (1) YES (2) NO 2

-------
B-24
2. Attenuation Parameters
The attenuation parameters are divided into four basic groups:
1) lumped attenuation parameters, used only if option (16) is off,
yielding constant daily rates of attenuation; 2) daily attenuation
proportions for various processes (microbial degradation, volatili-
zation, oxidation, hydrolysis, and photolysis) for the various
regions and layers; 3) attenuation parameters describing effects of
environmental conditions; and 4) attenuation parameters describing
compound properties different from those of azinphosmethyl.
When the attenuation parameters category of reparameterization
is selected, the current daily attenuation proportions at standard
conditions are displayed. The standard conditions are approximately
the average of the conditions obtained during the collection of the
field data used to parameterize the model. These conditions are:
1)	environmental conditions
-	soil pH = 7.0
-	% organic matter of soil = 6%
2)	compound properties
-	molecular wt. =317
-	vapor pressure = 10~^ atm.
The daily attenuation proportion matrix represents proportions of
the amount of pesticide in each region and layer that is attenuated under
the above standard conditions during one day.
If option (16) is off, and after the attenuation proportion matrix
is displayed, the user may change any number of entries in the matrix
one at a time, display the new matrix after some entries have been
changed, or exit the daily attenuation proportion matrix change mode

-------
B-25
once he is satisfied with attenuation proportions at standard condi-
tions.
Example
CURRENT ATTENUATION PROPORTIONS AT STANDARD CONDITIONS ARE:
CANOPY
ALLEY
LEAVES
.0487
0.0000
GRASS
.0412
.0670
LITTER
.0412
.0670
SOIL
. 0790
.0790
WOULD YOU LIKE TO (1> CHANGE AN ENTRY IN THE STANDARD
CONDITION ATTENUATION MATRIX
(2)	SEE CURRENT DAILY ATTENUATION PROPORTIONS
<3> EXIT ATTENUATION PROPORTION CHANGE MODE 1
ENTER REGION (1) CANOPY <2> ALLEY
LAYER (1) LEAVES <2> GRASS (3) LITTER-MOSS (4) SOIL
ATTENUATION PROPORTION (VALUE FROM 0.0 TO 1.0) 12,
WOULD YOU LIKE TO (1) CHANGE AN ENTRY IN THE STANDARD
CONDITION ATTENUATION MATRIX
< 2) SEE CURRENT DAILY ATTENUATION PROPORTIONS
(3)	EXIT ATTENUATION PROPORTION CHANGE MODE2
CURRENT ATTENUATION PROPORTIONS AT STANDARD CONDITIONS ARE:
050
CANOPY
ALLEY
LEAVES
.0487
0.0000
GRASS
.0500
.0670
LITTER
.0412
.0670
SOIL
. 0 7 ¥ 0
. 0790
WOULD YOU LIKE TO <1> CHANGE AN ENTRY IN THE STANDARD
CONDITION ATTENUATION MATRIX
<2> SEE CURRENT DAILY ATTENUATION PROPORTIONS
<3> EXIT ATTENUATION PROPORTION CHANGE MODE3
ANY MORE REPARAME TERIZA TI ON? (1) YES (2) NO 2
If option (16) is on, then the above matrix serves only to show stan-
dard conditions totals for daily attenuation. The matrix used will be
calculated daily, based on the environmental conditions. The daily
attenuation proportion matrix is further subdivided, into 5 components.
Each entry in the new matrix represents one of five attenuation pro-
cesses (hydrolysis, photolysis, oxidation, microbial degradation,

-------
b-26
volatilization). This new daily attenuation proportion matrix, broken
down by region, layer, and attenuation processes, is displayed. Its
entries represent daily proportions attenuated under standard condi-
tions, and the daily rates actually used will be these entries multi-
plied by environmentally-dependent factors, described below.
The user then has the option to change this new matrix. If he
decides to change the matrix he has the option to change a single
entry in the matrix, the contribution of an entire process (a row in
the matrix), or each process for a given region and layer (a column
in the matrix). After the user makes any type of change in either of
the two previous matrices (overall daily attenuation proportion matrix
for a given region and layer, and the same matrix also divided by con-
tributions of different attenuation processes), the user again has
the option to change the matrix in the manner just described or keep
it as it is.
Exampl e
CURRENT OVERALL ATTENUATION PROPORTIONS AT STANDARD CONDITIONS ARE!
LEAVES	GRASS	LITTER	SOIL
CANOFY	.0187	.0412	.0412 ~	.0790
ALLEY	0.0000	.0670	.0670	.0790
THE CONTRIBUTION OF INDIVIDUAL PROCESSES TO OVERALL ATTENUATION
AT STANDARD CONDITIONS ARE!
MATRIX AP

C 1
C 2
C 3
C 4
A 1
A 2
A 3
A 4
MICROBIAL DEC.
0.000
0.000
.010
.070
0.000
0.000
.017
.070
HYDROLYSIS
.012
.012
.015
.009
0.000
.019
.025
.009
PHOTOLYSIS
.018
.006
.003
0.000
0.000
.010
.004
0.000
VOLATILIZATION
.019
.024
.013
0.000
0.000
.038
.021
0.000
OXIDATION
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
WOULD YOU LIKE TO CHANGE THE ATTENUATION PROPORTIONS AT STANDARD
CONDITIONS ALLOCATED TO DIFFERENT TYPES OF ATTENUATION
(1> YES < 2 > NO 1

-------
B -27
WOULD YOU LIKE TO CHANGE <1> A SINGLE ENTRY IN THE MATRIX AP
(2> A ROW IN THE MATRIX
(3)	A COLUMN IN THE MATRIX 1
ENTER 4 NUMBERS
FIRST NUMBER- REGION  CANOPY (2) ALLEY
SCEOND NUMBER- LAYER <1> LEAVES (2) GRASS (3) LITTER-MOSS (4) SOIL
THIRD NUHBER- PROCESS <1> MICROBIAL DEG. (2) HYDROLYSIS (3> PHOTOLYSIS
(4)	VOLATILIZATION (5) OXIDATION
FOURTH NUHBER- ATTENUATION PROPORTION FOR THE REGION* LAYER* PROCESS YOU
HAVE SELECTED (PROPORTION MUST BE FROM 0.0 TO 1.0) 14 1 .060
CURRENT OVERALL ATTENUATION PROPORTIONS AT STANDARD CONDITIONS ARE:
CANOPY
ALLEY
LEAVES
.0487
0.0000
GRASS
.0412
.0670
LITTER
.0412
.0670
SOIL
.0670
.0790
THE CONTRIBUTION OF INDIVIDUAL PROCESSES TO OVERALL ATTENUATION
AT STANDARD CONDITIONS ARE:
MATRIX AP
MICROBIAL DEG.
HYDROLYSIS
PHOTOLYSIS
VOLATILIZATION
OXIDATION
C 1
0.000
.012
.018
.01?
C 2
0.000
.012
.006
.024
C 3
.010
.015
.003
.013
C 4
.060
.00?
0.000
0.000
A 1
0.000
0.000
0.000
0.000
A 2
0.000
.01?
.010
.038
A 3
.017
.025
.004
.021
A 4
.07 0
.00?
0.000
0.000
0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
UOULD YOU LIKE TO CHANGE THE ATTENUATION PROPORTIONS AT STANDARD
CONDITIONS ALLOCATED TO DIFFERENT TYPES OF ATTENUATION
<1) YES (2) NO 2
Once the daily standard attenuation proportions are set, the user
has the option to change environmental properties of the site or com-
pound properties. He is given standard conditions of necessary environ-
mental and compound properties such as soil pH, % soil organic matter,
vapor pressure, etc. and asked to input the value of the property.

-------
Example
B -28
DO rou WISH TO CHANGE ENVIRONMENTAL PROPERTIES OR COMPOUND PROPERTIES
(1) YES (2) NO 1
ENTER SOIL PH VALUE (CURRENT PH = 7.000) 6.5
ENTER PERCENT ORGANIC HATTER OF SOIL (CURRENT
PERCENT ORGANIC HATTER = 6.000) 6.0
MULTIPLICATIVE FACTOR FOR VOLATILIZATION IS OF THE FORM
F ¦ (VP « MI4M0.5) / (STDVP * STDHW**0.5)
WHERE VP = VAPOR PRESSURE OF CURRENT COMPOUND
HU i MOLECULAR HEIGHT OF CURRENT COMPOUND
STDVP ¦= VAPOR PRESSURE OF STANDARD COMPOUND (AZINPHOSMETHYL>
= .0000001 ATM.
STDMW = MOLECULAR UEIGHT OF STANDARD COMPOUND (AZINPHOSNETHYL)
= 317
ENTER COMPOUND VAPOR PRESSUftE( IN ATM.)
(STANDARD COMPOUND VAPOR PRESSURE = .0000001000 ATM.) .0000001
ENTER COMPOUND MOLECULAR UEIGHT (STANDARD COMPOUND MOLECULAR UEIGHT = 317) 317
ENTER COMPOUND FORMULATION (IN PERCENT)
(STANDARD COMPOUND FORMULATION = 50.00PERCENT UETTABLE POUDER) SO.
MULTIPLICATIVE FACTOR FOR PHOTOLYSIS IS OF THE FORM
F » (2.303/C0NST) * PHI < SIGMA * AOS * DIURN / DEF
WHERE CONST ° CONVERSION CONSTANT = 6.02 * 10**20
PHI ° QUANTUM YIELD
SIGMA = TOTAL PHOTONS ADSORBED (DIFFERENT FOR EACH COMPARTMENT)
AOS = A MULTIPLICATIVE FACTOR TO ALLOW FOR NATURALLY OCCURRING
LIGHT ABSORBERS t SENSITIZERSt AND QUENCHERS
DIURN = DIURNAL CONVERSION FACTOR TO LOSS RATE AT PEAK
SUNLIGHT INTENSITY TO DAILY PROPORTIONAL LOSS
= 2.52 * 10**4
DEF » ATTENUATION PROPORTION ENTRY FOR PHOTOLYSIS (DIFFERENT
FOR EACH COHPARTHENT)
ENTER QUANTUM YIELD (IN HOLES CONVERTED PER PHOTON ABSORBED)
(CURRENT QUANTUM YIELD " .100000E-02 HOLES CONVERTED PER
PHOTON ABSORBED) .001
ENTER MULTIPLICATIVE FACTOR TO ALLOU FOR NATURALLY OCCURRING LIGHT
ABSORBERS? QUENCHERS* AND SENSITIZERS (CURRENT AOS MULTIPLICATIVE
FACTOR -	1.000) 1.0

-------
B-29
TOTAL PHOTONS ABSORBED  YES (2> NO 2
MULT IPLI CAT IVE FACTOR FOR HYDROLYSIS DUE TO AIR TEMP.
IS OF THE FORM
F » E ** < DELHA/1•986 > * (T - TO) / 
TO = 296 DEGREES K
DELHA = ACTIVATION ENERGY (CALORIES/MOLE)
ENTER ACTIVATION ENERGY FOR ABOVE EQUATION
(CURRENT ACTIVATION ENERGY » 12500.00 CAL/MOLE) 12500.
MULTIPLICATIVE FACTOR FOR HYDROLYSIS DUE TO SOIL TEMP.
IS OF THE FORH
F n E ** (SDELHA/1.986) * (T - T0> / (T * TO) >
WHERE E = NATURAL LOG BASE = 2.718
T = SOIL TEMPERATURE (DEGREES K)
TO = 296 DEGREES K
SDELHA = ACTIVATION ENERGY (CALORIES/MOLE)
ENTER ACTIVATION ENERGY FOR ABOVE EQUATION
(CURRENT ACTIVATION ENERGY * 13500.00 CAL/hOLE)12500.
The user is next asked if he wishes to change coefficients of
attenuation equations based on environmental conditions. If he
selects this option, he is required to select a specific attenuation
process (microbial degradation, photolysis, hydrolysis, volatiliza-
tion, oxidation). The appropriate equations are displayed one at a
time and the user may enter new coefficients if desired.

-------
Example
B -30
DO rOU WISH TO CHANGE COEFFICIENTS OF ATTENUATION EQUATIONS
DEPENDENT ON ENVIRONMENTAL CONDITIONS
(1> rES (2) NO 1
00 YOU WISH TO CHANGE COEFFICIENTS OF EQUATIONS DEPENDENT ON ENVIRONMENTAL
CONDITIONS FOR
C1» MICROBIAL DEGRADATION (2) HYDROLYSIS (3) OXIDATION 1
MULTIPLICATIVE FACTOR FOR MICROBIAL DEGRADATION BASED ON PERCENT ORGANIC HATTER
IS OF THE FORM
F » <1.0 + E**(A * (PERCENT ORGANIC MATTER - B)>>**C
WHERE A " .079
b «» -i.eso
C = 1.870
E ° NATURAL LOG BASE = 2.718
DO YOU WISH TO CHANGE THESE COEFFICIENTS? (1) YES (2) NO 2
MULTIPLICATIVE FACTOR FOR MICROBIAL DEGRADATION BASED ON SOIL TEMPERATURE
IS OF THE FORM
F = A + B * SOILTEMP + C * 
FOR S (PEG. C) < SOILTEMP < 52 (DEG. C
AND F = 0.0 OTHERUISE
WHERE A = -.19520
B = .04480
C = .00080
DO YOU WISH TO CHANGE THESE COEFFICIENTS? (1) YES (2) NO 2
HO YOU TO CHANGE OTHER COEFFICIENTS TO ATTENUATION EQUATIONS?
<1 ) YES (2) NO 1
BO YOU WISH TO CHANGE COEFFICIENTS OF EQUATIONS DEPENDENT ON ENVIRONMENTAL
CONDITIONS FOR
(1) MICROBIAL DEGRADATION (2) HYDROLYSIS (3) OXIDATION 2
MULTIPLICATIVE FACTOR FOR HYDROLYSIS BASED ON PH VALUE
IS OF THE FORM
F = 1.0	»IF PH < C
F = A # PH + B tlFPH = C
WHERE A = 10.470
B = -76.810
C = 7.430

-------
B -31
ANY MORE REPARAMETERIZATION? (1) YES <2) NO 1
DO YOU UISH TO CHANGE (1) PHYSICAL CHARACTERISTICS OF SITE
(2) ATTENUATION PARAMETERS <3> PENETRATION PARAMETERS
(4) MOVEMENT PARAMETERS (5) ORGANISM PARAMETERS 3
CURRENT INITIAL PENETRATION PROPORTIONS ARE:
LEAVES	GRASS LITTER
.081	.240	.255
MOULD YOU LIKE TO CHANGE THE INITIAL PENETRATION PROPORTIONS OF A SPRAY
<1> YES (2) NO 2
CURRENT DAILY PENETRATION PROPORTIONS ARE*.
LEAVES GRASS LITTER
CANOPY	0.000	0.000	.126
ALLEY	0.000 0.000	.156
WOULD YOU LIKE TO CHANGE DAILY PENETRATION PROPORTIONS? <1> YES <2> NO 2
CURRENT DAILY PENETRATED RESIDUE ATTENUATION PROPORTIONS ARE!
LEAVES GRASS LITTER
CANOPY	.010	.010	.010
ALLEY	.010	.010	.010
WOULD YOU LIKE TO CHANGE DAILY PENETRATION ATTENUATION PROPORTIONS
(1) YES (2) NO 2

-------
B -32
3. Penetration Parameters
There are three basic penetration parameters for each region
and layer in a matrix form. They represent 1) the initial penetration
rate for a spray, 2) the daily penetration proportion of dislodgeable
residue that becomes penetrated, and 3) the penetrated residue atten-
uation proportions.
First, the initial (day of spray) penetration matrix is printed.
The user may change any of the entries one at a time. When this is
completed, the new matrix is printed out. Second, the daily penetra-
tion rate matrix is printed. The user may again change any of the
entries one at a time. When completed, the new matrix will be printed.
Finally the penetrated residue attenuation matrix is allowed to be
altered in the same manner.
Example
DO YOU WISH TO CHANGE THESE COEFFICIENTS? (1) YES <2> NO 2
[10 YOU TO CHANGE OTHER COEFFICIENTS TO ATTENUATION EQUATIONS?
(1) YES (2> NO1
DO YOU UISH TO CHANGE COEFFICIENTS OF EQUATIONS DEPENDENT ON ENVIRONMENTAL
CONDITIONS FOR
<1> MICROBIAL DEGRADATION (2) HYDROLYSIS <3> OXIDATION 3
MULTIPLICATIVE FACTOR FOR OXIDATION IS OF THE FORMt
F =» A
UHEKE A = 1.000
DO YOU UISH TO CHANGE THESE COEFFICIENTS? <1> YES (2) NO 2
DO YOU TO CHANGE OTHER COEFFICIENTS TO ATTENUATION EQUATIONS?
<1> YES (2 > NO 2

-------
B -33
4. Movement Parameters
A.	SPRAY DISTRIBUTION
The first movement parameter the user is allowed to change is
the spray distribution. The spray distribution represents the
proportion of pesticide reaching the orchard during spraying that
goes to each distinct (region and layer) compartment. The current
spray distribution vector is displayed. The user may use this or
enter a new spray distribution vector. If he enters a new spray
distribution vector, he must enter a proportion for each layer and
region. The sum of the proportions he enters must be 1.0. If it
is not, it is normalized to one. (Note—drift losses are accounted
for elsewhere in the model.) When completed, the new spray distri-
bution vector is printed.
Example
ANY MORE REPARAMETERIZATION? <1> YES <2> NO 1
DO YOU UISH TO CHANGE <1) PHYSICAL CHARACTERISTICS OF SITE
(2) ATTENUATION PARAMETERS (3) PENETRATION PARAMETERS
<4> MOVEMENT PARAMETERS <5> ORGANISM PARAMETERS A
CURRENT SPRAY DISTRIBUTION IS!
LEAVES	GRASS LITTER	SOIL
CANOPY	.5392	.1006	.0387	.0299
ALLEY	0.0000	.0942	.0300	.1564
DO YOU UANT TO ENTER A NEW SPRAY DISTRIBUTION? (1) YES (2) NO 2
B.	NON-RAIN MOVEMENT MATRIX
After the spray distribution matrix is complete, the non-rain
movement matrix is printed. The non-rain movement matrix represents
the daily proportions of pesticides that move from one compartment

-------
B-34
to another (see Subroutine Non-Rain description for more details
about how the matrix is applied).
The user has the option to 1) re-enter any single column of
the matrix, 2) display the matrix after changes have been made,
3) increase or decrease the daily proportion of movement of the
entire matrix by any multiplicative factor, 4) increase or de-
crease the daily proportion of a single column by any multiplica-
tive factor, or 5) exit the non-rain movement matrix change mode.
He may make any number of changes (1) through (4) until he exits
the change mode. Note—options (3) and (4) do not act directly
on the matrix entries. They cause the factor entered to be multi-
plied by the appropriate instantaneous rates calculated from the
off-diagonal elements. The (multiplied) rates are then re-
converted to 1-day loss proportions, and the diagonal elements re-
computed by subtraction. Thus a factor of 2 has the effect of
making the pesticide "half as sticky", while .5 makes it "twice
as sticky".
Example
CURRENT NON-RAIN MOVEMENT MATRIX
FROM


C 1
C 2
C 3 C 4
A 1
A 2
c
1
.9804




c
2
.0003
. 9347



c
3
.0088
.0654
.9073


c
4
0.0000
0.0000
.0927 1.0000


A
1





A
2
.0001



. 9732
A
3
.0094



.0265
A
4
.0005



.0003
TO
WHERE C=CANOPY REGION* A=ALLEY REGION
1CLEAVES » 2=GRASS. 3=LITTERMOSS• 4=SOIL
AND COLUMNS MUST SUM TO 1.0
A 3
A 4
.9962
.0038 1.0000

-------
B-35
DO YOU UISH (1) TO RE-ENTER A SINGLE COLUMN
(2)	LOOK AT THE CURRENT NON-RAIN MOVEMENT MATRIX
(3)	CHANGE THE DAILY PROPORTION OF MOVEMENT OF THE ENTIRE HOVEMENT
MATRIX BY A MULTIPLICATIVE FACTOR
(4)	CHANGE THE DAILY PROPORTION OF MOVEMENT OF A SINGLE COLUMN BY A
MULTIPLICATIVE FACTOR
(5> EXIT NON-RAIN MATRIX CHANGE MODE 4
ENTER REGION <1> CANOPY <2> ALLEY
LAYER (1) LEAVES (2) GRASS <3> LITTERMOSS <4> SOIL
OF THE COLUMN VECTOR YOU UISH TO CHANGE 1 1
ENTER FACTOR YOU UISH TO INCREASE OR DECREASE THE RATE OF MOVEMENT BY
(FACTOR MUST BE GREATER THAN ZERO) 2.0
DO YOU UISH (1) TO RE-ENTER A SINGLE COLUMN
(2) LOOK AT THE CURRENT NON-RAIN MOVEMENT MATRIX
<	3) CHANGE THE DAILY PROPORTION OF MOVEMENT OF THE ENTIRE MOVEMENT
MATRIX BY A MULTIPLICATIVE FACTOR
(4) CHANGE THE DAILY PROPORTION OF MOVEMENT OF A SINGLE COLUMN BY A
MULTIPLICATIVE FACTOR
<	5 > EXIT NON-RAIN MATRIX CHANGE H0DE2
CURRENT NON-RAIN MOVEMENT MATRIX
FROM


C 1 C 2
C 3
C 4 A 1 A 2
A 3
A 4
c
1
.9611




c
2
.0006 .9347




c
3
. 0175 .0654
.9073



c
4
0.0000 0.0000
.0927
1 .0000


A
1





A
2
.0002

.9732


A
3
.0186

.0265
.9962

A
4
.0011

.0003
.0038
1 .0000
UHERE C=CANOPY REGION t A=ALLEY REGION
1=LEAVES» 2 = GRASS. 3 = LITTERMOSS r 4 = S0IL
AND COLUHNS MUST SUM TO 1.0
r>0 YOU UISH (I) TO RE-ENTER A SINGLE COLUMN
(2 > LOOK AT THE CURRENT NON-RAIN MOVEMENT MATRIX
< 3) CHANGE THE DAILY PROPORTION OF MOVEMENT OF THE ENTIRE MOVEMENT
MATRIX BY A MULTIPLICATIVE FACTOR
(4) CHANGE THE DAILY PROPORTION OF MOVEMENT OF A SINGLE COLUMN BY A
MULTIPLICATIVE FACTOR
<5) EXIT NON-RAIN MATRIX CHANGE MODE 5

-------
B-36
RAINFALL MOVEMENT MATRICES
(See SUBROUTINE RAINFALL for description of how matrix is applied.)
The user is first asked which matrix to display (light or
heavy rainfall). The changes made to a rainfall matrix are always
to the last rainfall matrix (light or heavy) that was displayed.
When asked to make changes, the user may always display
either the light or heavy rainfall movement matrix.
Other than this, changes to rainfall movement matrices are
the exact same options as those to the non-rain movement matrix.
In the heavy rainfall matrix example below, the movement rate is
first doubled, then cut in half to restore the matrix to its
original values.
Example
DO YOU WISH TO (1) LOOK AT LIGHT RAINFALL MOVEMENT MATRIX
< 2 > LOOK AT HEAVY RAINFALL MOVEMENT MATRIX
(3 > MAKE NO RAINFALL MOVEMENT MATRIX CHANGES 1
CURRENT LIGHT RAINFALL MOVEMENT HATRIX
FROM


C 1
C 2 C 3 C 4 A 1 A 2
A 3
A 4
C
1
.9320



c
2
.0296
.7468


c
3
.0021
.0708 1.0000


c
4
.0362
.1824 0.0000 1.0000


A
1




A
2
.0001
.8687


A
3
0.0000
0.0000
1.0000

A
4
0.0000
. 1313
0.0000
1.0000
UHERE C=CANOPY REGION r A = ALLEY REGION
1=LEAVES i 2<=GRASS. 3=L ITTERMOSS i 4=S0IL
AND COLUMNS MUST SUM TO 1.0
DO YOU WISH TO (1) LOOK AT CURRENT LIGHT RAINFALL MATRIX
<	2) LOOK AT CURRENT HEAVY RAINFALL MATRIX
<	3 > RE-ENTER A SINGLE COLUMN OF THE RAINFALL MATRIX YOUR WORKING
(4)	CHANGE THE DAILY PROPORTION OF MOVEMENT OF THE ENTIRE CURRENT
MOVEMENT MATRIX 6Y ANY MULTIPLICATIVE FACTOR
(5)	CHANGE THE DAILY PROPORTION OF MOVEMENT OF A SINGLE COLUMN BY
MULTIPLICATIVE FACTOR
(6)	EXIT RAINFALL MATRIX CHANGE MODE 4

-------
B -37
ENTER FACTOR YOU UISH TO INCREASE OR DECREASE THE RATE OF MOVEMENT BY
(FACTOR MUST BE GREATER THAN ZERO) 2.
DU YOLI UISH TO (1) LOOK AT CURRENT LIGHT RAINFALL MATRIX
(2) LOOK AT CURRENT HEAVY RAINFALL MATRIX
(3> RE-ENTER A SINGLE COLUMN OF THE RAINFALL MATRIX YOUR UORKING ON
(4)	CHANGE THE DAILY PROPORTION OF MOVEMENT OF THE ENTIRE CURRENT
MOVEMENT MATRIX BY ANY MULTIPLICATIVE FACTOR
(5)	CHANGE THE DAILY PROPORTION OF MOVEMENT OF A SINGLE COLUMN BY A
MULTIPLICATIVE FACTOR
(6)	EXIT RAINFALL MATRIX CHANGE MODE 1
CURRENT LIGHT RAINFALL MOVEMENT MATRIX
FROM


C 1
C 2
C 3
C 4 A 1 A 2
A3 A 4
c
1
. 8686




c
2
.0572
.5577



c
3
.0041
.1237
1 .0000


i:
4
. 0699
.3186
0.0000
1.0000

A
1





A
A
.0002


.7546

A
3
0.0000


0.0000
1.0000
A
4
0.0000


.2454
0.0000 1.0000
UHERE C=CANOPY REGION t A°ALLEY REGION
1=LEAVES» 2=GRASSt 3=LITTERM0SS» 4=S0IL
AND COLUMNS MUST SUM TO 1.0
DO YOU UISH TO (1) LOOK AT CURRENT LIGHT RAINFALL MATRIX
(2)	LOOK AT CURRENT HEAVY RAINFALL MATRIX
(3)	RE-ENTER A SINGLE COLUMN OF THE RAINFALL MATRIX YOUR UORKING ON
(4)	CHANGE THE DAILY PROPORTION OF MOVEMENT OF THE ENTIRE CURRENT
MOVEMENT MATRIX BY ANY MULTIPLICATIVE FACTOR
< 5 > CHANGE THE DAILY PROPORTION OF MOVEMENT OF A SINGLE COLUMN BY A
MULTIPLICATIVE FACTOR
(6> EXIT RAINFALL MATRIX CHANGE MODE 4
ENTER FACTOR YOU UISH TO INCREASE OR DECREASE THE RATE OF MOVEMENT BY
(FACTOR MUST BE GREATER THAN ZERO) 0.5
DO YOU UISH TO (1) LOOK AT CURRENT LIGHT RAINFALL MATRIX
(2)	LOOK AT CURRENT HEAVY RAINFALL MATRIX
(3)	RE-ENTER A SINGLE COLUHN OF THE RAINFALL MATRIX YOUR UORKING ON
(4)	CHANGE THE DAILY PROPORTION OF MOVEMENT OF THE ENTIRE CURRENT
MOVEMENT MATRIX BY ANY MULTIPLICATIVE FACTOR
Cj) CHANGE THE DAILY PROPORTION OF MOVEMENT OF A SINGLE COLUMN BY A
MULTIPLICATIVE FACTOR
(6) EXIT RAINFALL MATRIX CHANGE MODE 1

-------
B -38
CURRENT LIGHT RAINFALL MOVEMENT MATRIX
FROM


C 1
C 2
C 3 C 4 A 1 A 2
A3 A 4
c
1
.9320



c
o
.0296
. 7468


c
3
.0021
.0708
1.0000

c
4
.0362
. 1824
0.0000 1.0000

A
1




A
2
.0001

. 6667

A
3
0.0000

0.0000
1.0000
A
4
0.0000

. 1313
0.0000 1.0000
UHERE C=CANOPY REGION i A = ALLEY REGION
1"LEAVES> 2°GRASS > 3=LITTERMOSS» 4=SOIL
AND COLUMNS MUST SUM TO 1.0
DO YOU WISH TO (1) LOOK AT CURRENT LIGHT RAINFALL MATRIX
(2) LOOK AT CURRENT HEAVY RAINFALL MATRIX
(3 > RE-ENTER A SINGLE COLUMN OF THE RAINFALL MATRIX YOUR WORKING ON
<	4) CHANGE THE DAILY PROPORTION OF MOVEMENT OF THE ENTIRE CURRENT
MOVEMENT MATRIX BY ANY MULTIPLICATIVE FACTOR
(5> CHANGE THE DAILY PROPORTION OF MOVEMENT OF A SINGLE COLUMN BY A
MULTIPLICATIVE FACTOR
<	6) EXIT RAINFALL MATRIX CHANGE MODE 2
CURRENT HEAVY RAINFALL MOVEMENT MATRIX
FROM


C 1
C 2
C 3
C 4
A 1
A 2
A 3
A 4
c
1
.8209







c
2
. 0202
.6126






c
3
.0957
0.0000
1 .0000





c
4
.0082
.3874
0.0000
1.0000




A
1








A
2
.0270




.2798


A
3
.0008




.4465
1.0000

A
4
.0272




. 2737
0.OOOD
1.0000
UHERE C=CANOPY REG ION r A=ALLEY REGION
1=LEAVES» 2=GRASS, 3=LITTERM0SS. 4=S0IL
AND COLUHNS MUST SUM TO 1.0
DO YOU WISH TO (1) LOOK AT CURRENT LIGHT RAINFALL MATRIX
< 2) LOOK AT CURRENT HEAVY RAINFALL MATRIX
(3) RE-ENTER A SINGLE COLUMN OF THE RAINFALL MATRIX YOUR WORKING CIN
<4> CHANGE THE DAILY PROPORTION OF MOVEMENT OF THE ENTIRE CURRENT
MOVEMENT MATRIX BY ANY MULTIPLICATIVE FACTOR
(5)	CHANGE THE DAILY PROPORTION OF MOVEMENT OF A SINGLE COLUMN BY A
MULTIPLICATIVE FACTOR
(6)	EXIT RAINFALL MATRIX CHANGE MODE 6

-------
B-39
D. PROPORTION OF GRASS MOWED
Finally the user is shown the current proportion of grass
mowed for a mow date, and asked if he would like to enter a new
one. If so, the porportion must be >_ 0.0 and <_ 1.0.
Example
WOULD YOU LIKE TO CHANGE THE PROPORTION OF GRASS MOWED
(CURRENT PROPORTION MOWED =0.0000) (1) YES (2) NO 1
ENTER PROPORTION OF GRASS MOWED < 0.0 TO 1.0) 0.40

-------
B -40
Example Run of POEM
In the following sample run, the user opted to use weather data
supplied on a file, but to enter spray dates manually. The user specified
printout of pesticide residues at 5 day intervals, running of only the
isopod (Tracheoniscus) model, and other choices as shown. Files for
plotting pesticide residues and isopod populations were also created by
the run.
Example
EXEC BEGUN.16.18.58.
THE FOLLOWING OPTIONS ARE ON BY DEFAULT
ORGANISM SUBMODEL
ATTENUATION SUBMODEL
NON-RAIN MOVEMENT SUBMODEL
PESTICIDE APPLICATION SUBMODEL
RAINFALL MOVEMENT SUBMODEL
PENETRATION SUBMODEL
WRITE FILES FOR PLOTTING PACKAGE
VARIABLE ATTENUATION
OTHER AVAILABLE OPTIONS ARE
MOWER SUBMODEL
WEATHER SIMULATOR
SOIL MOISTURE SIMULATOR
LEAF GROWTH SUBMODEL
MOVEMENT MATRIX DAILY OUTPUT
SOIL TRANSPORTATION SUBMODEL
PESTICIDE PRINTOUT
NON-DEFAULT EVENT DATES -SPRAYSrMOWS>ETC
DO YOU WISH TO TURN ON OR OFF ANY OPTIONS?
(1) YES (2) NO 1

-------
B-41
ENTER A VALUE FOR EACH OPTION
(1) FOR OPTION ON (2) FOR OPTION OFF
OPTION
1
ORGANISM SUBMODEL 1
OPTION
2
MOWER SUBMODEL 2
OPTION
3
ATTENUATION SUBMODEL 1
OPTION
4
WEATHER SIMULATOR 2
OPTION
5
NON-RAIN MOVEMENT SUBHODM 1
OPTION
6
PESTICIDE APPLICATION SUBMODEL 1
OPTION
7
RAINFALL MOVEMENT SUBMODEL 1
OPTION
8
PENETRATION SUBMODEL 1
OPTION
9
SOIL MOISTURE SIMULATOR 2
OPTION
10
LEAF GROWTH SUBMODEL 2
OPTION
11
MOVEMENT MATRIX DAILY OUTPUT 2
OPTION
12
SOIL TRANSPORTATION SUBMODEL 2
OPTION
13
PESTICIDE PRINTOUT 1
OPTION
14
WRITE FILES FOR PLOTTING PACKAGE 1
OPTION
15
NON-DEFAULT EVENT DATES -SPRAYS,MOWS,ETC 1
OPTION
16
VARIABLE ATTENUATION 1
WHICH ORGANISM SUBMODELS ARE DESIRED (1) YES (2) NO
TRACHEONISCUS? 1
FOLSOHIA? 2
SPIDER? 2
EARTHWORM? 2
HOW HANY YEARS DO YOU WISH TO SIHULATEt ONE TO TEN YEARS) 1
HOW OFTEN (IN DAYS) MOULD YOU LIKE PESTICIDE VALUES PRINTED TO OUTPUT 5
DO YOU UANT TO REPARAMETERIZE THE MODEL (DEFAULT VALUES ARE BASED
ON ESTIMATES FOR AZINPHOSMETHYL AS GUTHION 50 W,P.)
(1) YES (2) NO 2
YEAR 1
ENTER SEASON STARTING DATE 170
ENTER SEASON ENDING DATE 230
HOU MANY SPRAYS DO YOU WISH THIS SEASON( ZERO TO TEN SPRAYS) 2
ENTER DATES CHRONOLOGICALLY
SPRAY HATE 1? 170
DOSAGE( IN KG./HA.) FOR SPRAY 1? 2.0
SPRAY DATE 2? 184
DOSAGE( IN KG./HA.) FOR SPRAY 2? 2.0
SPRAY DATE, DAY 170, SPRAY RATE = 2.00 KG/HA, PERCENT DRIFT OF SPRAY = 50.8

-------
B -42
PREDICTED PESTICIDE LEVELS FOR DAY 170
DISLODGEABLE RESIDUES
LEAVES GRASS LITTERHOSS SOIL
CANOPY 6.559 1.012 .382 0.000
ALLEY	.561 .175 0.000
PENETRATED RESIDUES
LEAVES GRASS LITTERHOSS SOIL
.578 .320 .130 .396
.177 .060 1.226
PREDICTED PESTICIDE LEVELS FOR DAY 175
DISLODGEABLE RESIDUES
LEAVES GRASS LITTERHOSS SOIL
CANOPY 4.697 .601 .384 0.000
ALLEY	.356 .175 0,000
PENETRATED RESIDUES
LEAVES GRASS LITTERHOSS SOIL
.550 .304 .372	.35A
.169 .196	.717
PREDICTED PESTICIDE LEVELS FOR DAY 180
DISLODGEABLE RESIDUES
LEAVES GRASS LITTERHOSS SOIL
CANOPY 3.324 .353 .279 0.000
ALLEY	.221 .132 0.000
PENETRATED RESIDUES
LEAVES GRASS LITTERHOSS	SOIL
.523 .289 .565	.314
.160 .307	.429
RAIN DATE; HEAVY RAIN DAY 182
SPRAY DATE. DAY 184. SPRAY RATE = 2.00 KG/HA? PERCENT DRIFT OF SPRAY = 39.0
PREDICTED PESTICIDE LEVELS FOR DAY 184
DISLODGEABLE RESIDUES
LEAVES GRASS LITTERHOSS SOIL
CANOPY 10.220 1.445 .826 0.000
ALLEY	.777 .363 0.000
PENETRATED RESIDUES
LEAVES GRASS LITTERHOSS SOIL
1.219 .674 .889 .912
.374 .466 1.890
PREDICTED PESTICIDE LEVELS FOR DAY 185
DISLODGEABLE RESIDUES
LEAVES GRASS LITTERHOSS SOIL
CANOPY 9.563 1.303 .798 0.000
ALLEY	.710 .353 0.000
PENETRATED RESIDUES
LEAVES GRASS LITTERHOSS SOIL
1.206 .667 .983 .808
.370 .518 1.712
PREDICTED PESTICIDE LEVELS FOR DAY 190
DISLODGEABLE RESIDUES
LEAVES GRASS LITTERHOSS SOIL
CANOPY 6.765 .765 .582 0.000
ALLEY	.441 .267 0.000
PENETRATED RESIDUES
LEAVES GRASS LITTERHOSS SOIL
1.147 .634 1.374 .719
.352 .736 .983

-------
3-43
PREDICTED PESTICIDE LEVELS FOR DAY 195
DISLODGEABLE RESIDUES
LEAVES GRASS LITTERMOSS SOIL
CANOPY 4.765 .449 .390 0.000
ALLEY	.272 .187 0,000
PENETRATED RESIDUES
LEAVES GRASS	LITTERMOSS SOIL
1.091 .603	1.613	,541
.334	.876	.547
PREDICTED PESTICIDE LEVELS FOR DAY 200
DISLODGEABLE RESIDUES
LEAVES GRASS LITTERMOSS SOIL
CANOPY 3.369 .265 .258 0.000
ALLEY	.169 .130 0.000
PENETRATED RESIDUES
LEAVES GRASS LITTERHOSS SOIL
1.038 .574 1.738 .388
.318 .956 .300
RAIN DATE* HEAVY RAIN DAY 203
PREDICTED PESTICIDE LEVELS FOR DAY 205
DISLODGEABLE RESIDUES
LEAVES GRASS LITTERMOSS SOIL
CANOPY 1.938 .140 .304 0.000
ALLEY	.065 .115 0.000
PENETRATED RESIDUES
LEAVES GRASS	LITTERMOSS	SOIL
.987 .545 1.844	.381
.302 1.010	.227
RAIN DATEf LIGHT RAIN DAY 208
RAIN DATE> HEAVY RAIN DAY 210
PREDICTED PESTICIDE LEVELS FOR DAY 210
DISLODGEABLE RESIDUES
LEAVES GRASS LITTERhOSS SOIL
CANOPY 1.042 .086 .259 0.000
ALLEY	.030 .074 0.000
PENETRATED RESIDUES
LEAVES GRASS LITTERHOSS SOIL
.938 .519 1.890	.370
.288 1.027	.157
RAIN DATEf LIGHT RAIN DAY 214
PREDICTED PESTICIDE LEVELS FOR DAY 215
DISLODGEABLE RESIDUES
LEAVES GRASS LITTERMOSS SOIL
CANOPY .694 .060 .101 0.000
ALLEY	.017 .036 0.000
PENETRATED	RESIDUES
LEAVES GRASS	LITTERHOSS	SOU
.892 .493 1.907	.2V4
.274 1.018	.091
RAIN DATEf HEAVY RAIN DAY 218

-------
B-44
PREDICTED PESTICIDE LEVELS FOR DAY 220
DISLODGEABLE RESIDUES
LEAVES GRASS LITTERMOSS SOIL
CANOPY .408 .031 .077 0.000
ALLEY	.011 .023 0.000
PENETRATED RESIDUES
LEAVES GRASS LITTERMOSS SOIL
.84? .46? 1.872	.220
.260 .??2	.065
PREDICTED PESTICIDE LEVELS FOR DAY 225
DISLODGEABLE RESIDUES
LEAVES GRASS LITTERMOSS SOIL
CANOPY .28? .01? .031 0.000
ALLEY	.007 .012 0.000
PENETRATED RESIDUES
LEAVES GRASS LITTERMOSS SOIL
.807 .446 1.813	.14?
.247 .?57	.038
RAIN DATE* HEAVY RAIN DAY 228
PREDICTED PESTICIDE LEVELS FOR DAY 230
DISLODGEABLE RESIDUES
LEAVES GRASS LITTERMOSS SOIL
CANOPY .170 .011 .02? 0.000
ALLEY	.004 .00? 0.000
END ORCHARD
077000 FINAL EXECUTION FL.
5.048 CP SECONDS EXECUTION TIME.
PENETRATED RESIDUES
LEAVES GRASS LITTERMOSS SOIL
.767 .424 1.744 .100
.235 .?18 .027

-------
B-45
Producing Graphs
The two plotting programs described below have been written to allow
the researchers to conveniently generate Tektronix (CRT) or CALCOMP (pen-
and-ink) graphs of pesticide and population levels as generated by POEM.
Because they must utilize locally-determined procedures for plotting, they
depend on MSU-provided software which is not generally available. Thus,
for other sites, more general-purpose plotting packages should be used,
working from POEM-generated files as input. These descriptions are pro-
vided to illustrate an example working environment in which POEM is con-
veniently utilized.

-------
B-46
Pesticide Graphs
Program ORCHARDCALPLOT allows the user to plot pesticide levels in
any region or layer versus time, producing either CALCOMP or TEKTRONIX
output. It has available multiple output formats, depending on the user's
needs. It can plot the level in a single region based on a single POEM
run, or can plot one solid and one dashed line, based on two runs, for
A, B comparison purposes. It can also display with vertical bars a set of
data (usually measured levels, at irregular intervals) for comparison with
predicted levels.
The program begins by asking whether TEKTRONIX or CALCOMP plotting is
desired. It then presents the user a set of options:
Example
ENTER 1 FOR TEK* 2 FOR CALCOMP...1
PLOTTING OPTIONS	
(1)	MEASURED AVERAGES ONLY
(2)	MEASURED AVERAGES WITH ONE	PREDICTED SEI
(3)	MEASURED AVERAGES WITH TWO	PREDICTED SETS
(4)	ONE PREDICTED SET ONLY = TO BEGINNING YEAR)?1

-------
RRIN [MM)
2.0 _
>-
3:
1 .5
1 -°
UJ CvJ
x: c c
co o U .5
O N
x: &
£ 3 0.0
t—l
r^J _ _
a: 0-5
1 .0
UJ
CO
CRNOPY SOIL 1978
3 6 4 3
\t M \>
2 17 11
T
51
J
6
v f
4 I
T
14
'' i i i i
J	' i	' ¦ ¦ '	I	i	i
ISO
JULIAN DATE
255

-------
B-48
The user needs to make sure that appropriate files are available
for the option he selects. A single file of predicted residues to be
graphed in a solid line must be called TAPE1. An (optional) second file
to apipear as a dashed line must be called TAPE4. If measured values are
to be plotted, they must be called TAPE5. (For plots of outputs of single
POEM runs, the correct file is automatically available as TAPEl.) In addi-
tion, two files which specify the dates of pesticide spraying and the dates
and amounts of rain must be available as TAPE2 and TAPE3, respectively.
If a POEM run has just been made, the appropriate TAPE2 and TAPE3 have
been created by POEM.
For information on file handling on the MSU Cyber system, see the
section on Attaching Tapes.

-------
B-49
Organism Graphs
Program ORGPLOT allows the user to plot population levels versus time
for single life-stages or groups of life-stages. It uses the files gener-
ated by POEM (called TAPE11, TAPEl2, TAPE13, TAPE!4, TAPEl5). They are
created by POEM as follows:
TAPEl1--non-breeding female isopod populations on each day of the model run,
arranged as:
IN ORGPLOT,
LIFESTAGE INDEX
Day Number (Jan. 1 year 1 = day 1)
daily egg hatch (not number of eggs	present) 1
size-class 1 (0.5-1.34 mg)	2
size-class 2 (1.35-5.0 mg)	3
size-class 3 (5.01-12.0 mg)	4
size-class 4 (12.01-30.0 mg)	5
size-class 5 (30.01-50.0 mg)	6
size-class 6 (50.01-80.0 mg)	7
size-class 7 (>80.0 mg)	8
total, classes 1-7 (lifestage indices 2-8)	9
TAPEl2--breeding female isopod populations on each day of the model run,
arranged as:
Day Number
Blank field (egg hatch stored on TAPEl1)
size-class 1-3 0.0 (no small breeders)
size-class 4 (12.01-30.0 mg)	5
size-class 5 (30.01-50.0 mg)	6

-------
B -50
IN 0RGPL0T,
LIFESTAGE INDEX
size-class 6 (50.01-80.0 mg)
size-class 7 (>80.0 mg)
total, classes 4-7 (life-stage indices 5-8)
9
8
7
(Note that in ORGPLOT, the populations displayed are the sums of breeding
and non-breeding populations.)
TAPE13--collembola (also called Folsomia in the model, based on data from
both Folsomia Candida and Folsomia quadrioculata) populations on
each day of the model run, arranged as:
Day Number
Egg hatch on current day (not total	eggs present) 1
Prejuveniles	2
Juveniles	3
Adults	4
Totals (lifestage indices 2-4)	5
TAPE14--earthworm populations (see earthworm model for definitions)
TAPE15--spider populations on each day of the model run, arranged as:
Day Number
viable eggs (not hatch)	1
Instar 1
2

Instar 8
9
Adults
10
Totals (including eggs)
11

-------
B-51
ORGPLOT will expect to find only those files which the user asks to
plot (i.e., for isopods—TAPE11, TAPE12, for collembola--TAPE13, for earth-
worms—TAPE14, for spiders—TAPE15}. The user is asked for the choice of organ-
isms, a label for the y-axis, and for scale factors for each axis. The
x-axis will be scaled at 6 points (with 5 intervals between), and the user
may explicitly enter the starting and ending Julian dates (cumulative--
i.e., 370 is day 5 of year 2), or may enter 0,0 to request auto-scaling
by the program. A maximum of 1200 days may be displayed. Similarly, the
y-(population) axis can be scaled by specifying an upper limit, or the
program will auto-scale it if a 0 is entered.
The user is next asked for a size-class range to plot. This is given
by specifying the number of the beginning and ending size-classes, in
ascending order—i.e., for size classes 2-4, enter 2,4; to see only size-
class 3, enter 3,3.
After a graph is generated, the user may shoot a hard copy (TEKTRONIX)
before the program contaminates the screen with any further questions.
When ready to proceed, the user should type a space and return, to produce
the next question. He may request another graph, or may quit the program.
In CALC0MP mode, the program immediately prompts for further actions, after
indicating that the plot file has been submitted for plotting (an off-line
procedure).
Example
ENTER 1 FOR TEK, 2 FOR CALCOMP! 1
WHICH ORGANISM DO YOU UISH TO PLOT
(1)	TRACHEONISCUS
(2)	FOLSOMIA
(3)	SPIDER
(1) EARTHUGI:m 1
ENTER Y-AXIS LABEL FOR GRAPH SAMPLE ISOPODS
ENTER X LOUERt UPPER LIMITSi OR 0>0 ("OR AUTOMATIC! OrO
ENTER Y UPPER LIMIT. OR 0 FOR AUTOMATIC: 0
ENTER LIFESTAGE RANGE TO PLOT f LOWER tUPF'ER IN&ICES: 2tA

-------
cp.OC
IS0PGD<12 MG RUN
26.00 50.00 75.00
—I	I	:	I	
o
o
I
ro
a'
—I o
•—i O
zz
rn
2®"
o§
3D
-<
CO
PO"
o
o
r
(O
CO"
•
o
o
N)
o
100.00 125.00
_j	i
0)
i
Ul
ro

-------
B-53
Attaching Tapes
Disk files which are to be read or written during the operation of POEM
and its associated plotting programs on MSU's CYBER are locally referred to
as TAPEs. This naming is to agree with local file names specified in the
PROGRAM statement at the beginning of each program. At MSU, there are two
types of disk files: local files and permanent files. Permanent files may
be read and/or written only by "attaching" them to (temporary) local file
names. TAPEl, TAPE2, etc., are examples of such local file names. Thus,
if the user wants to plot data from permanent files stored after earlier
runs (see below), he must first "attach" the permanent files to local files,
as shown below:
ATTACH,TAPEl.PREDICTEDRUNl ,PW = 	.
This command makes the permanent file PREDICTEDRUNl accessible as TAPEl,
assuming that the correct password was specified after "PU = ". Similar
commands will retrieve stored weather data.
In the simplest types of use, the user need not concern himself with
these file issues: POEM automatically attaches the permanent file
"WTHRSTATSNEW" if the weather simulator is turned off, and the file
"PWSM0IST2" if the soil moisture submodel is turned off. POEM also creates
(if plotting option is on) the SPRAYDATES(TAPE2) and RAINAM0UNTS(TAPE3) files
required by DATAPLT to plot the pesticide levels generated by this run (on
TAPEl). Thus, the user need not perform explicit ATTACH commands if he is
using data just generated by the model.
Note--if POEM is to be run after other programs (or POEM) have been run
in the same session, the user must first release the files POEM will try to
attach: namely, type:
RETURN, TAPE4, TAPE9.

-------
B -54
File Conventions in POEM Programs
POEM may read: TAPE4 - weather data (which it attaches as
WTHRSTATSNEW)
TAPE9 - soil moisture data (which it attaches as
PWSM0IST2)
POEM creates (depending on options selected)
TAPE1	- pesticide levels
TAPE2	- spraydates
TAPE3	-	rainamounts
TAPE11	-	isopod non-breeding populations
TAPE!2	-	isopod breeding populations
TAPE13	-	collembolan populations
TAPE14	-	earthworm populations
TAPE15	-	spider populations
DATAPLT (stored on file 0RCHARDCALPL0T6) uses:
TAPE2 - spraydates
TAPE3 - rainamounts
(TAPE1 - predicted pesticide levels, to be plotted as
solid line - optional)
(TAPE4 - predicted levels, to be plotted as dashed
line - optional)
(TAPE5 - measured pesticide levels, to be plotted as
vertical bars - optional)

-------
B -55
ORGPLOT uses:
TAPES11-15 as shown above.
Users wishing to run these programs at other sites should replace the
CDC-specific PROGRAM statements, which also control file access, with appro-
priate file-handling statements. In particular, in F0RTRAN77, the standard
OPEN statements could instead be used.

-------
B-56
Addendum to Users' Manual
Notice
Due to lack of funds and consequent lack of computer and manpower
resources, some features and options of POEM are untested and/or undebugged.
While the bulk of the code required to exercise all of the options des-
cribed here is written, and all subroutines have run at various times as
independent units, the linking of some routines has not been completed.
Known problems are listed below:
(Option 2) Weather simulation—not operable. Weather information
must instead be provided in a file (TAPE4).
(Option 12) Soil transportation model—not operable. Pesticide
levels in soil cannot be broken down by depth.
(Option 9) Soil moisture model--not operable. Soil moisture
information must be provided in a file (TAPE9).

-------