Content uploaded by Heping Zhu
Author content
All content in this area was uploaded by Heping Zhu on Nov 13, 2020
Content may be subject to copyright.
Transactions of the ASABE
Vol. 56(6): 1263-1272 2013 American Society of Agricultural and Biological Engineers ISSN 2151-0032 DOI 10.13031/trans.56.9839 1263
SPRAY DEPOSITION INSIDE TREE CANOPIES FROM A NEWLY
DEVELOPED VARIABLE-RAT E AIR-ASSISTED SPRAYER
Y. Chen, H. E. Ozkan, H. Zhu, R. C. Derksen, C. R. Krause
ABSTRACT. Conventional spray applications in orchards and ornamental nurseries are not target-oriented, resulting in
significant waste of pesticides and contamination of the environment. To address this problem, a variable-rate air-assisted
sprayer implementing laser scanning technology was developed to apply appropriate amounts of pesticides based on tree
canopy characteristics including tree height, width, volume, foliage density, and occurrence. The new sprayer perfor-
mance was evaluated in an apple orchard by quantifying spray deposition inside canopies at three different growth stages
(leafing, half-foliage, and full-foliage) with three sprayer treatments: the new variable-rate sprayer (S1), the same sprayer
without the variable-rate function (S2), and a conventional air-blast sprayer (S3). Their spray coverage and deposits inside
canopies were measured and compared with water-sensitive papers and nylon screens. The three sprayer treatments pro-
vided fairly consistent spray coverage and deposits in the spray direction (or canopy depth direction) at the leafing stage.
The variations in spray coverage and deposits in the spray direction increased considerably for S2 and S3 at the half-
foliage and full-foliage stages. S1 produced better uniformity in spray coverage and deposits across the tree height direc-
tion than S2 and S3 at all growth stages. Compared to constant-rate sprayers, the new variable-rate sprayer only con-
sumed 27% to 53% of the spray mixture while still achieving adequate spray coverage inside the canopies. In addition, the
spray deposition from the new sprayer was very consistent regardless of the canopy growth stage. Therefore, the new
sprayer increased spray efficiency and improved spray accuracy by greatly lowering the possibility of overspray, resulting
in reduced spray costs and potential reduction of environmental pollution.
Keywords. Canopy sensing, Laser sensor, Orchard sprayer, Precision farming, Spray coverage.
rees in nurseries and orchards have great varia-
tions in shape, size, foliage density, and spacing
between in-row trees. This variability requires
future sprayers to be flexible in their operation
and spray an amount of chemical that can match each tree
structure. Conventional sprayers for nursery or orchard
applications do not have this flexibility, and their spray
deposition quality inside canopies varies greatly with tree
growth conditions (Hoffmann and Salyani, 1996; Farooq
and Salyani, 2002; Pergher et al., 1997; Salyani et al.,
2007; Zhu et al., 2008). Applicators using these sprayers
typically spray the entire field with a constant rate during
the entire growing season. Consequently, crops are either
oversprayed or undersprayed (Salyani et al., 2007; Zhu et
al., 2008), causing a significant portion of the spay mixture
to be lost to the air and ground (Derksen et al., 2007; Fox et
al., 1993; Zhu et al., 2006a). Unnecessary overuse of pesti-
cides not only results in economic loss but also is a poten-
tial source of environmental contamination that may affect
the safety and health of applicators, workers, and nearby
residents.
To reduce chemical use in nurseries and orchards, re-
searchers have designed several variable-rate sprayers with
different types of sensors. Giles et al. (1989), Escola et al.
(2003), Solanelles et al. (2006), and Gil et al. (2007) devel-
oped variable-rate sprayers with integration of ultrasonic
sensors and documented savings in spray mixture for these
variable-rate sprayers. Jeon et al. (2011) and Jeon and Zhu
(2012) also used 20 Hz ultrasonic sensors in a vertical
boom variable-rate spraying system developed for nursery
liner-size trees and reported that the sprayer could reduce
the spray volume by over 70% compared to conventional
constant-rate sprayers.
However, variable-rate sprayers with ultrasonic sensors
have several disadvantages. The sensors have low detection
accuracy because of their low measurement resolution. In
addition, their response can be easily influenced by tractor
operating speed and environmental conditions such as tem-
perature and humidity. Studies by other researchers con-
cluded that laser scanning sensors were able to characterize
crop structures with higher accuracy and reliability than
ultrasonic sensor technology (Tumbo et al., 2002; Wei and
Salyani, 2004, 2005). Because of their advantages, laser
Submitted for review in July 2012 as manuscript number PM 9839;
approved for publication by the Power & Machinery Division of ASABE
in November 2013.
Mention of proprietary product or company is included for the reader’s
convenience and does not imply any endorsement or preferential treatmen
t
by the USDA-ARS and The Ohio State University.
The authors are Yu Chen, ASABE Member, Postdoctoral Research
Associate, USDA-ARS Application Technology Research Unit (ATRU)
and Department of Food Agricultural and Biological Engineering (FABE),
The Ohio State University/OARDC, Wooster, Ohio; H. Erdal Ozkan,
ASABE Member, Professor, FABE, The Ohio State University,
Columbus, Ohio; Heping Zhu, ASABE Member, Agricultural Engineer,
Richard C. Derksen, ASABE Member, Agricultural Engineer, and
Charles R. Krause, Research Plant Pathologist, USDA-ARS ATRU,
Wooster, Ohio. Corresponding author: Heping Zhu, USDA-ARS ATRU,
1680 Madison Ave., Wooster, OH 44691; phone: 330-263-3871; e-mail:
heping.zhu@ars.usda.gov.
T
1264 T
RANSACTIONS OF THE
ASABE
scanning sensors have great potential for use in variable-
rate sprayer development.
An experimental variable-rate sprayer implementing a
high-speed laser scanning sensor was developed for or-
chard and nursery applications (Chen et al., 2012). The
sprayer has the capability to automatically adjust the spray
output to match tree characteristics in real time. Laboratory
evaluations of the sprayer performance demonstrated that it
could deliver liquids to different parts of the tree canopy
with satisfactory coverage uniformity. To better evaluate
how the sprayer could respond to complex conditions of
trees, field experiments were conducted to achieve the ob-
jective of this research: to further evaluate the performance
of the variable-rate sprayer under field conditions through
comparison of the overspray possibility and the uniformity
of spray coverage and deposition inside tree canopies with
constant-rate sprayers.
M
ATERIALS AND
M
ETHODS
V
ARIABLE
-R
ATE
S
PRAYER
The newly developed air-assisted variable-rate sprayer
used in this study is able to control the spray outputs dis-
charged from 20 nozzles independently to match target tree
canopy characteristics (Chen et al., 2012). It integrates a
high-speed laser scanning sensor, a custom-designed signal
processing program, an automatic variable-rate controller,
variable-rate nozzles, and a multi-port air-assisted delivery
system. The program in the computing unit (a portable
computer, in this case) calculates sectional canopy volumes
and foliage densities for each corresponding nozzle with a
refresh rate equivalent to 23 mm of canopy width at 3.2 km
h
-1
tractor speed and then calculates the duty cycle for the
pulse width modulation (PWM) signals that synchronize
the spray outputs with the laser sensor detection. After the
desired duty cycle is determined, a valve driver circuit gen-
erates and amplifies the PWM signals with the desired duty
cycle and actuates the solenoid valves to control the flow
rates of the 20 nozzles independently to achieve the auto-
matic variable-rate function.
The base of the multi-port air-assisted delivery system
mainly includes an axial turbine fan, a 400 L spray tank, and
a diaphragm pump from a vineyard sprayer (model ZENIT
B11, Hardi International A/S, Taastrup, Denmark). The laser
sensor is mounted 1.6 m above the ground on the target tree
side. Sprays are discharged from four specially designed
five-port nozzle manifolds mounted on each side of the
sprayer, and each manifold consists of five nozzle tips modi-
fied from flat-fan TeeJet XR 8002 nozzles (Spraying Sys-
tems Co., Wheaton, Ill.) (Zhu et al., 2006b). Variations in
droplet size from the PWM-controlled nozzles are minimal
and insignificant for all modulation rates, including those
used to produce low liquid flow rates (Gu et al., 2011). The
heights of the four nozzle manifolds are 0.85, 1.35, 1.85, and
2.35 m above the ground. The sprayer is 1.52 m wide, and its
spray pattern is designed to cover targets up to 3.2 m high at
1.5 m distance from the sprayer. The nozzles discharge vari-
able flow rates independently to their assigned sections of
the canopy based on each section height.
S
PRAY
D
EPOSITION
T
ESTS
Field tests were conducted to determine spray deposition
quality (spray coverage and deposit) within apple tree can-
opies with three sprayer treatments: the new variable-rate
sprayer (S
1
) (fig. 1a), the same sprayer with the automatic
variable-rate function disabled (S
2
) (fig. 1a), and a conven-
tional air-blast sprayer (model 1500, Durand Wayland, Inc.,
LaGrange, Ga.) (S
3
) (fig. 1b). Among these three treat-
ments, S
1
performed a variable-rate application, and S
2
and
S
3
performed conventional constant-rate applications. Be-
cause only one row of trees was selected for the test, noz-
zles on only one side of each sprayer that were directed
toward the tested tree row were used for all three treat-
ments.
Sprayers S
1
and S
2
were operated at 207 kPa, and the flow
rate from each of the 20 nozzles was 0 to 0.68 L min
-1
for the
variable-rate sprayer (S
1
) and 0.68 L min
-1
for the constant-
rate sprayer (S
2
). Sprayer S
3
had ten TeeJet D5-DC25 disc-
core hollow-cone nozzle tips (Spraying Systems Co.,
Wheaton, Ill.) on the test side, but the bottom and top nozzles
were closed to follow the best management practice for
avoiding excessive sprays discharged to the soil surface and
above the tree height. A radial spray pattern was formed by
(a) Sprayers S
1
and S
2
(b) Sprayer S
3
Figure 1. Newly developed sprayer with variable-rate function (S
1
),
the new sprayer without variable-rate function (S
2
), and conventional
air-blast sprayer (S
3
) used in the test.
56(6): 1263-1272 1265
the droplets discharged from the eight hollow-cone nozzles
with heights ranged from 0.71 to 1.44 m above the ground.
These nozzles were mounted on the circumference of the
1.03 m diameter fan outlet with an equal radial angle of 16°
between nozzles. S3 was operated at 248 kPa, and the flow
rate for each nozzle was 1.36 L min-1.
During the tests, all three sprayers traveled at a ground
speed of 3.2 km h-1, resulting in application rates of 0 to
554 L ha-1 for S1, 544 L ha-1 for S2, and 443 L ha-1 for S3 (it
would have been 544 L ha-1 if all ten nozzles were activat-
ed). The application rates for S2 and S3 were calculated
according to the alternative tree row spraying method
(Lewis and Hickey, 1972) to conform with the best man-
agement practice commonly used by growers.
Field tests were conducted at three typical growth stages
in an experimental apple orchard growing Malus domestica
‘Gala’ and Malus domestica ‘Golden Delicious’. The field
was 57 m long and 46 m wide with 4.6 m spacing between
tree rows and 2.6 m spacing between trees within a row
(fig. 2). The first test was carried out in early spring
(12 April 2010) when the trees had just started leafing
(leafing stage), the second test was at about half canopy
growth (3 May 2010) (half-foliage stage), and the third test
was conducted in mid-summer (8 June 2010) when the
trees had fully established foliage (full-foliage stage). Three
Malus domestica ‘Gala’ trees in a random row of the field
were selected to mount artificial targets for documenting
spray coverage and deposit inside the tree canopies. Fig-
ure 3 shows images of tree 1 at the leafing, half-foliage,
and full-foliage stages taken with a digital camera and the
laser scanning sensor during the tests. The maximum tree
widths and heights during the test period were 2.5 and 2.9
m for tree 1, 2.0 and 2.5 m for tree 2, and 3.1 and 3.1 m for
tree 3, respectively.
A 3 m high portable weather station equipped with a
modified CM-6 system (Campbell Scientific, Inc., Logan,
Utah) was placed in an open field 10 m away from the test
site to record air temperature, relative humidity, and wind
speed and direction at 1 Hz frequency. The sensor used in
the weather station to measure the wind speed and direction
was a three-axis ultrasonic anemometer with no moving
parts (model 81000, R. M. Young Co., Traverse City,
Mich.) with a resolution of 0.01 m s-1 for wind speed and
0.1° for wind direction. Only the weather data within the
time of each test run were used. Weather conditions during
the tests for three sprayers are listed in table 1.
ARTIFICIAL TARGET LOCATIONS
Water-sensitive papers (WSP) (26 × 76 mm, Syngenta
Crop Protection AG, Basel, Switzerland) and monofilament
nylon screens (50 × 50 mm, Filter Fabrics, Inc., Goshen,
Ind.) were mounted at different locations inside the target
tree canopies to document the spray coverage (percentage
area of the WSP covered by spray deposits) and deposits
inside the canopies, respectively. The monofilament nylon
screens had a nominal porosity of approximately 56% (or
fiber frontal area percentage of 44%). For this type of
screen, Fox et al. (2004) reported an airborne collection of
50% to 70% efficiency for spray droplets with volume me-
dian diameters of 30 to 45 μm, which was much better than
flat solid collectors. At each target location, a WSP and a
nylon screen were mounted side by side with clips. The
samples were located at the edges of the tree canopies,
halfway inside, and in the middle of the canopies (fig. 4).
In the process of selecting sample locations on real trees,
the first guideline was to mount the samples as close to the
designated locations as possible, given that there were
branches or twigs available for mounting the clips. Howev-
er, as shown in figure 3, random branches in a tree might
block targets from the spray at certain locations. In such a
case, another guideline was to mount the sample clips to a
close location (usually in front of the branches facing the
spray direction) to minimize the blocking by the branches.
Hence, the exact positions of targets to be mounted in the
same distribution plane might be changed slightly to adapt
to the availability of support branches. All sample locations
were measured manually with three rulers placed perpen-
dicularly. In this way, the position of every WSP or nylon
screen could be defined as a set of X, Y, and Z values in a
3-D Cartesian coordinate system (figs. 4a and 4b). The X,
Y, and Z directions correspond to spray direction (horizon-
tal), tractor travel direction (horizontal), and tree height
direction (vertical), respectively. As shown in table 1, dur-
ing the experiments, the wind direction in test 2 (half-
foliage stage) was different from the wind directions in test
1 (leafing stage) and test 3 (full-foliage stage). Therefore,
the spray direction and thus the definition of the X direction
in test 2 was opposite to that defined for tests 1 and 3.
With the defined 3-D Cartesian coordinate system, sam-
ples inside a canopy were divided into four groups in the X
direction with three vertical cross-sections parallel to the Y
Figure 2. Plan view of the field site used for spray deposition tests.
4.6 m
46.4 m
12.9 m
18.2 m
18.2 m
2.6 m
57.0 m
Wind azimuth
Notes:
Dimensions are not to scale
Tree without targets
Tree with targets
N
E
A°
1266 T
RANSACTIONS OF THE
ASABE
direction (fig. 4a). The same samples were also divided into
four groups in the Y direction with three vertical cross-
sections parallel to the X direction, and three groups in the
Z direction with two horizontal cross-sections parallel to
the X and Y directions (fig. 4b). Data from these groups
were used to determine the uniformity of spray deposition
and coverage in the three different directions. The four
groups in the X direction were named front, middle front,
middle back, and back (fig. 4a). The front group had sam-
ples in the canopy section closest to the sprayer, and the
back group was the section farthest away from the sprayer.
The middle front group represented the section located in
the middle part of the canopy closer to the sprayer, and the
middle back group was the middle section away from the
sprayer. Similarly, the four target groups in the Y direction
were named right, middle right, middle left, and left
(figs. 4a and 4b), and the three groups in the Z direction
were named high, middle, and low (fig. 4b).
Spray drift and off-target losses to the ground, beyond
the target trees, and in the air were also measured for S
1
, S
2
,
and S
3
at the same time as the tests to quantify spray depo-
sition inside the canopies, and these results are reported in a
separate publication (Chen et al., 2013).
The spray mixture used in the tests contained 2 g of
Brilliant Sulfaflavine (MP Biochemicals, Inc., Aurora,
Ohio) per liter of water. All artificial targets were collected
15 min after each spray run. Nylon screens were placed in
125 mL glass bottles, which were then stored in opaque
boxes, and WSP were stored in brown paper bags before
transport to the laboratory for analysis. Tank mixture sam-
ples were also collected before and after each run as refer-
ences for calculating the amount of spray deposits on the
nylon screens.
The nylon screens were washed free of the fluorescence
tracer with purified water. The amount of spray deposits on
the targets was based on the fluorescent intensity of each
wash solution, which was then converted to the volume of
spray mixture per unit area in microliters per square centi-
(a) Leafing (b) Half-foliage (c) Full-foliage
Figure 3. Images taken with a digital camera (top row) and laser scanning sensor (bottom row) of tree 1 at (a) leafing, (b) half-foliage, and
(c) full-foliage growth stages. The color scale represents distance (mm) between the laser sensor and target.
Table 1. Weather conditions during field tests for the sprayer with variable-rate function (S
1
), the sprayer without variable-rate function (S
2
),
and the conventional air-blast sprayer (S
3
).
Growth Stage
Wind Velocity (m s
-1
)
Wind Azimuth (°)
Temperature (°C)
Relative Humidity (%)
S
1
S
2
S
3
S
1
S
2
S
3
S
1
, S
2
, S
3
S
1
, S
2
, S
3
Leafing 2.6 3.2 3.2 43 79 70 15 25
Half-foliage 5.9 3.1 2.5 308 302 277 22 48
Full-foliage 1.8 1.5 1.7 92 120 131 21 43
56(6): 1263-1272 1267
meter (μL cm-2). The fluorescent intensity of each wash
solution was determined with a luminescence spectrometer
(LS 50B, PerkinElmer, Seer Green, U.K.) at an excitation
wavelength of 460 nm. A hand-held business card scanner
(ScanShell 800N, CSSN, Inc., Los Angeles, Cal.) was used
to acquire images of spray deposits on each WSP with
600 dpi imaging resolution. The spray coverage was meas-
ured from the scanned images with the DepositScan pro-
gram (Zhu et al., 2011).
DATA ANALYSIS
Spray deposit and coverage in the X, Y, or Z direction
were first analyzed by one-way analysis of variance
(ANOVA) using statistical software (ProStat version 5.5,
Poly Software International, Inc., Pearl River, N.Y.) to test
the null hypothesis that all groups in each direction for S1,
S2, and S3 had equal means. If the null hypothesis was re-
jected, Duncan’s multiple comparison test was used to de-
termine differences among means. All differences were
analyzed at the 0.05 level of significance.
After significant differences were determined among S1,
S2, and S3, uniformity indexes were calculated to compare
uniformities of spray deposit and coverage among sample
groups in each direction for the three sprayers with the fol-
lowing equation:
2
2
n
C
i
i
U
n
u
I
C
= (1)
where
IU = uniformity index for either spray deposit or cover-
age (0 to 1, with values closer to 0 representing lower
uniformity and values closer to 1 representing greater
uniformity)
i = order of comparisons (1, 2,…, Cn
2)
n = number of groups in a direction to be compared (n =
4 for the X or Y direction, and n = 3 for the Z direc-
tion)
Cn
2 = total possible combinations of two comparisons,
which are denoted with different letters for signifi-
cant difference and with the same letter for no signif-
icant difference (e.g., C4
2 = 12 for either the X or Y
direction, and C3
2 = 3 for the Z direction)
ui = outcome of comparison of the two groups in one
combination (ui = 0 when they are significantly dif-
ferent, and ui = 1 when they are not significantly dif-
ferent).
Similarly, an overspray index was introduced to quantify
the magnitude of overspray to compare the spray deposition
qualities of S1, S2, and S3. Overspray was defined as any
situation with spray coverage greater than 30%, which was
based on WSP samples (fig. 5) provided by a chemical
company (Syngenta, 2004). Calculation of the overspray
index was performed by normalization of the spray cover-
age with the following equation:
30
100 30
O
C
I−
=
−
(2)
where IO is the overspray index, and C is the spray cover-
age on WSP (%); IO ranges from 0 to 1, with 0 representing
no overspray and 1 presenting a saturated spray (100%
coverage).
A quad decision chart was then created with the uni-
formity index (IU) and the overspray index (IO) of spray
(a) Top view (b) Front view
Figure 4. Schematic of artificial target locations inside tree canopies (
X
= spray direction,
Y
= sprayer travel direction,
Z
= tree height).
Left
Right
Middle right
Middle left
1268 TRANSACTIONS OF THE ASABE
coverage, averaged from three trees, to produce a compre-
hensive comparison of spray deposition qualities among S1,
S2, and S3.
RESULTS AND DISCUSSION
SPRAY DEPOSITION UNIFORMITY
IN SPRAY DIRECTION (X)
Table 2 shows the average coverage and deposit on
samples inside three tree canopies grouped in sections
along the spray direction (X direction) for S1, S2, and S3 at
the leafing, half-foliage, and full-foliage growth stages.
Spray deposit distribution in the X direction represented the
effectiveness of spray penetration into the canopies. For the
tests at the leafing stage, the spray coverage and deposit on
targets at the front, middle front, middle back, and back
locations from S1 varied slightly more than those from S2
and S3. The variations in spray coverage and deposit in the
X direction increased slightly for S1 but increased consider-
ably for S2 and S3 at the half-foliage and full-foliage stages.
For example, when the target location changed from front
to back, the average spray deposit from S1, S2, and S3
changed from 1.33 to 1.34 μL cm-2, from 5.57 to 5.92 μL
cm-2, and from 3.95 to 3.49 μL cm-2 at the leafing stage;
from 1.10 to 0.53 μL cm-2, from 3.95 to 3.25 μL cm-2, and
from 2.25 to 0.73 μL cm-2 at the half-foliage stage; and
from 1.81 to 0.38 μL cm-2, from 3.54 to 1.87 μL cm-2, and
from 2.84 to 0.55 μL cm-2 at the full-foliage stage, respec-
tively.
In addition, for all three sprayer treatments at the leafing
and half-foliage stages, when the tree canopies were not
fully established, both spray coverage and deposit at the
front, middle front, middle back, and back locations had
better uniformity than those at the full-foliage stage. The
data in table 2 show that, in general, S2 had the lowest vari-
ation (or lowest standard deviations) in spray coverage and
Table 2. Average spray coverage on water-sensitive papers and spray deposits on nylon screens grouped in spray direction (
X
direction) inside
three tree canopies at leafing, half-foliage, and full-foliage growth stages for the sprayer with variable-rate function (S1), the sprayer without
variable-rate function (S2), and the conventional air-blast sprayer (S3). Standard deviations are given in parentheses.
Growth
Stage
Tar get
Location
Average Spray Coverage (%)
Average Spray Deposit (
μ
L cm-2 )
S1 S
2 S
3 S
1 S
2 S
3
Leafing
Front 51 (13) 93 (12) 77 (13) 1.33 (0.46) 5.57 (2.47) 3.95 (1.01)
Middle front 39 (21) 86 (13) 77 (17) 1.09 (0.50) 5.97 (3.03) 4.07 (0.78)
Middle back 43 (25) 91 (17) 75 (18) 1.48 (0.93) 7.91 (3.09) 4.11 (1.22)
Back 36 (19) 81 (27) 50 (14) 1.34 (0.79) 5.92 (2.65) 3.49 (1.21)
Half-foliage
Front 49 (17) 90 (17) 75 (14) 1.10 (0.63) 3.95 (1.68) 2.25 (0.87)
Middle front 49 (22) 87 (32) 61 (27) 1.06 (0.57) 4.06 (2.21) 1.81 (0.65)
Middle back 25 (21) 56 (27) 37 (20) 0.61 (0.53) 2.62 (2.13) 0.97 (0.56)
Back 23 (14) 43 (25) 24 (20) 0.53 (0.39) 3.25 (2.56) 0.73 (0.35)
Full-foliage
Front 58 (21) 76 (14) 76 (21) 1.81 (1.11) 3.54 (2.10) 2.84 (0.67)
Middle front 43 (31) 62 (35) 57 (21) 1.19 (1.03) 2.87 (1.50) 1.38 (0.69)
Middle back 35 (26) 52 (32) 32 (22) 0.79 (0.67) 3.42 (2.96) 0.97 (0.70)
Back 22 (19) 48 (28) 16 (15) 0.38 (0.29) 1.87 (1.77) 0.55 (0.49)
4.2%
(underspray)
28.0%
(adequate spray)
39.2%
(overspray)
97.6%
(saturated spray)
Figure 5. Reference WSP images (Syngenta, 2004) with measured coverage values.
56(6): 1263-1272 1269
deposit along the X direction, and S3 had the highest varia-
tion among the three sprayer treatments. Similarly, for the
spray deposit and coverage in the front and middle front
locations, S1 also had the lowest variation along the spray
direction. Thus, better spray uniformity in the X direction
from S1 could be achieved by spraying both sides of the
trees in each row instead of alternating rows.
SPRAY DEPOSITION UNIFORMITY
IN TRAVEL DIRECTION (Y)
Table 3 shows the average spray coverage and deposit
discharged from S1, S2, and S3 on samples grouped for the
travel direction (Y direction) at the three growth stages.
Spray deposition distribution in this direction represented
the effect of canopy surface shape on the spray deposition
quality from the sprayers. In general, S1, S2, and S3 all pro-
duced consistent spray deposit and coverage at the right,
middle right, middle left, and left locations at the leafing
stage. Since canopy depths at the middle right and middle
left locations were normally greater than those at the right
and left locations, which were closer to the canopy right
and left edges, more deposition was expected on samples at
the right and left locations from constant-rate sprayers S2
and S3. Table 3 confirms this expectation, especially with
the spray deposit data for the half-foliage and full-foliage
stages. The data in table 3 show that S1 produced more con-
sistent average spray deposit on samples in the Y direction
than S2 and S3 at the half-foliage and full-foliage stages.
Therefore, S1 was able to automatically adjust the spray
output based on changes in tree canopy shape and foliage
density.
SPRAY DEPOSITION UNIFORMITY
IN TREE HEIGHT DIRECTION (
Z
)
Table 4 shows the average spray coverage and deposit
discharged from S1, S2, and S3 on samples grouped in the
tree height direction (Z direction) at the three growth stag-
es. Spray deposit distribution in this direction represented
the capability of the sprayers to produce uniform spray
deposition inside canopies at different heights. In general,
S1 produced consistent spray deposit and coverage on tar-
gets at the high, middle, and low locations, while the varia-
tions of spray deposit and coverage from S2 and S3 at these
locations were considerably great. This was because S2 and
S3 discharged constant amounts of sprays from all nozzles
at different heights. The top part of trees had less foliage
and thus received greater spray deposits from S2 and S3. In
contrast, S1 was able to overcome this problem by automat-
ic adjustment of the spray output across the tree height to
match the foliage structure measured by the laser scanning
sensor.
Tables 2 through 4 also demonstrate that spray coverage
and deposit from S2 and S3 were greater than those from S1
at all three growth stages and for almost all sections. Many
target samples sprayed by S2 and S3 in the leafing and half-
foliage tests had coverage of 80% to 100%, which was
more than what was needed for adequate coverage. The
difference in the spray deposition and coverage between
variable-rate (S1) and constant-rate (S2 and S3) applications
decreased as the tree canopies grew larger and denser from
the leafing stage to the full-foliage stage. This was because
the spray volume discharged from S1 was based on the can-
opy foliage volume and density measurement, while S2 and
Table 4. Average spray coverage on water-sensitive papers and spray deposits on nylon screens grouped in tree height direction (
Z
direction)
inside three tree canopies at leafing, half-foliage, and full-foliage growth stages for the sprayer with variable-rate function (S1), the spraye
r
without variable-rate function (S2), and the conventional air-blast sprayer (S3). Standard deviations are given in parentheses.
Growth
Stage
Tar get
Location
Average Spray Coverage (%)
Average Spray Deposit (
μ
L cm-2 )
S1 S
2 S
3 S
1 S
2 S
3
Leafing
High 42 (17) 100 (0) 49 (13) 1.06 (0.65) 6.85 (1.82) 3.10 (0.98)
Middle 55 (15) 100 (0) 82 (16) 1.27 (0.73) 5.87 (1.96) 3.34 (1.16)
Low 35 (23) 80 (23) 69 (18) 0.84 (0.75) 4.16 (3.02) 2.74 (0.95)
Half-foliage
High 45 (13) 85 (18) 59 (16) 0.70 (0.28) 4.13 (2.04) 1.38 (0.40)
Middle 50 (17) 85 (16) 53 (28) 0.74 (0.61) 3.01 (2.37) 1.28 (0.81)
Low 41 (25) 66 (32) 52 (28) 0.66 (0.66) 2.30 (2.44) 1.08 (0.87)
Full-foliage
High 37 (14) 81 (10) 37 (12) 0.62 (0.46) 3.15 (2.09) 1.04 (0.70)
Middle 47 (28) 66 (32) 37 (23) 0.80 (0.91) 2.79 (2.15) 0.79 (0.90)
Low 31 (27) 45 (29) 45 (32) 0.66 (0.98) 1.82 (2.41) 1.00 (1.12)
Table 3. Average spray coverage on water-sensitive papers and spray deposits on nylon screens grouped in travel direction (
Y
direction) inside
three tree canopies at leafing, half-foliage, and full-foliage growth stages for the sprayer with variable-rate function (S1), the sprayer without
variable-rate function (S2), and the conventional air-blast sprayer (S3). Standard deviations are given in parentheses.
Growth
Stage
Tar get
Location
Average Spray Coverage (%)
Average Spray Deposit (
μ
L cm-2 )
S1 S
2 S
3 S
1 S
2 S
3
Leafing
Right 42 (23) 89 (13) 72 (21) 1.52 (0.88) 7.27 (2.28) 4.18 (1.04)
Middle right 44 (21) 86 (23) 67 (23) 1.47 (0.76) 6.54 (2.88) 3.83 (1.24)
Middle left 39 (16) 90 (19) 73 (19) 1.04 (0.45) 5.77 (3.49) 3.90 (1.36)
Left 40 (28) 92 (11) 69 (8) 1.13 (0.84) 6.87 (2.65) 3.73 (0.58)
Half-foliage
Right 44 (25) 85 (20) 65 (15) 0.89 (0.60) 4.37 (2.39) 1.80 (0.60)
Middle right 44 (23) 80 (35) 60 (27) 0.87 (0.54) 3.84 (2.57) 1.73 (0.99)
Middle left 31 (18) 55 (32) 31 (25) 0.91 (0.63) 2.46 (1.58) 0.95 (0.70)
Left 40 (24) 68 (36) 49 (25) 0.91 (0.57) 4.13 (2.89) 1.50 (0.70)
Full-foliage
Right 48 (29) 60 (26) 48 (22) 1.36 (1.01) 4.04 (3.03) 1.30 (0.75)
Middle right 29 (26) 56 (30) 29 (25) 0.61 (0.54) 2.33 (2.12) 0.89 (0.81)
Middle left 46 (21) 60 (28) 58 (30) 1.14 (0.81) 3.35 (2.02) 1.81 (1.15)
Left 22 (21) 42 (31) 47 (32) 0.86 (0.87) 2.64 (2.63) 1.53 (0.90)
1270 TRANSACTIONS OF THE ASABE
S3 used a constant spray volume for all three growth stages.
The amount of spray discharged from S1 was determined
to treat 1 m3 of canopy volume with 0.02 L of spray (Chen
et al., 2012). The spray volume discharged to each tree
from S1 increased as the tree foliage volume and density
increased during the growing season. The application rate
of S1 was 140, 157, and 223 L ha-1 for the tests at the leaf-
ing, half-foliage, and full-foliage stages, respectively. In
contrast, S2 used 526 L ha-1 and S3 used 421 L ha-1 for all
the tests at the three growth stages, both of which were
more than twice the spray volume from S1. The maximum
flow rate from each nozzle was 0.76 L min-1. With all noz-
zles open, S1 discharged 15.14 L min-1. However, even at
the full-foliage stage, none of the nozzles were fully
opened. Therefore, the amount of sprays discharged to the
targets from the nozzles on S1 was lower than those from
the nozzles on S2 and S3 to the same targets.
At the full-foliage stage, leaves closer to the nozzles in-
tercepted a large portion of the spray that was supposed to
reach targets at deeper depths inside the canopy. Thus,
spray coverage and deposit inside the canopies from S2 and
S3 at the full-foliage stage were lower than those at the leaf-
ing and half-foliage stages. However, the spray coverage
and deposit from S1 were consistent regardless of the
change in canopy size and foliage density during the tree
growing season.
COMPARISONS AMONG THREE GROWTH STAG ES
The average spray coverage and deposit on three trees at
three growth stages for S1, S2, and S3 are shown in figures 6
and 7. At the leafing stage, the average coverage was 41%,
88%, and 70%, and the average deposit was 1.31, 6.61, and
3.91 μL cm-2 for S1, S2, and S3, respectively. At the half-
foliage stage, the average coverage for S1, S2, and S3 was
39%, 74%, and 53%, and the average deposits were 0.87,
3.64, and 1.56 μL cm-2, respectively. At the full-foliage
stage, the average coverage for S1, S2, and S3 was 36%,
56%, and 42%, and the average deposit was 0.92, 2.96, and
1.27 μL cm-2, respectively. Hence, among the three spray-
ers tested, S1 had the lowest variation in spray coverage and
deposits as the trees grew from the beginning of leafing to
fully established foliage.
UNIFORMITY AND OVERSPRAY POSSIBILITY
Figure 8 shows uniformity indexes of spray deposits for
S1, S2, and S3 in the X, Y, and Z directions, averaged from
three trees and three growth stage tests. Among the three
sprayer treatments, S2 had the highest deposit index in the
X and Y directions, and S1 had the highest deposit index in
the Z direction. This means that S2 had the lowest variations
in spray deposits in the spray and travel directions, demon-
strating that the variable-rate sprayer (S1) was developed on
a well-performing spray delivery system (S2).
Figure 9 shows a quad chart to compare the overall
Figure 8. Uniformity index of spray deposition in the
X
,
Y
, and
Z
directions inside three trees at the leafing, half-foliage, and full-
foliage stages for S1 (new variable-rate sprayer), S2 (new sprayer
without variable-rate function), and S3 (conventional air-blast spray-
er).
Figure 6. Comparison of spray coverage on three trees among the
tests at the leafing, half-foliage, and full-foliage stages for S1 (new
variable-rate sprayer), S2 (new sprayer without variable-rate func-
tion), and S3 (conventional air-blast sprayer). Different letters or
symbols on bars for the same sprayer show significantly (p < 0.05)
different means of spray coverage.
Figure 7. Comparison of spray deposits on three trees among the tests
at the leafing, half-foliage, and full-foliage stages for S1 (new variable-
rate sprayer), S2 (new sprayer without variable-rate function), and S3
(conventional air-last sprayer). Different letters or symbols on bars
for the same sprayer show significantly (p < 0.05) different means o
f
spray deposits.
56(6): 1263-1272 1271
spray quality among S1, S2, and S3 using the coverage and
overspray indexes summarized from three trees and three
growth stages. Points in the upper left quadrant represent
better spray coverage uniformity and lower overspray pos-
sibility. All nine indexes for S1 are in the upper left quad-
rant, six out of nine indexes for S2 are in the upper right
quadrant, and three out of nine indexes for S3 are in the two
lower quadrants. The average uniformity indexes for S1, S2,
and S3 were 0.80, 0.85, and 0.69, and the average overspray
indexes were 0.17, 0.64, and 0.37, respectively. Therefore,
S1 had the best spray quality based on the combination of
uniformity and overspray indexes, and S1 minimized the
overspray possibility without reducing spray uniformity.
On the other hand, S2 had the highest overspray possibility,
and S3 had the lowest spray coverage uniformity among the
three sprayer treatments.
CONCLUSIONS
The uniformity of spray deposition inside tree canopies
was compared at three growth stages in an orchard for the
newly developed air-assisted variable-rate sprayer (S1) and
two constant-rate sprayers (S2 and S3). The following con-
clusions are drawn from these tests.
The three sprayer treatments provided fairly consistent
spray coverage and deposit in the spray (or canopy depth)
direction at the leafing stage. The variations in spray cover-
age and deposit in the spray direction increased slightly for
S1 and increased considerably for S2 and S3 at the half-
foliage and full-foliage stages. Moreover, S1 produced bet-
ter uniformity in spray coverage and deposit across the tree
height direction than S2 and S3 at all growth stages.
The application rate of S1 was 140 L ha-1 at the leafing
stage, 157 L ha-1 at the half-foliage stage, and 223 L ha-1 at
the full-foliage stage. On the other hand, S2 used 526 L ha-1
and S3 used 421 L ha-1 at all three growth stages. S1 provid-
ed consistent spray deposition and coverage inside the can-
opies during the three growth stages, while the quantities of
spray deposition and coverage from S2 and S3 varied great-
ly with the growth stage. At the three growth stages, the
canopies received average spray coverage of 36% to 41%
from S1, 56% to 88% from S2, and 42% to 70% from S3.
A quad chart of the spray coverage uniformity index and
overspray index demonstrated that S1 minimized overspray
without reducing spray uniformity, while S2 had the highest
possibility of overspray, and S3 had the lowest spray cover-
age uniformity. Therefore, compared to S2 and S3, sprayer
S1 was able to provide consistent and uniform spray deposi-
tion quality at different growth stages with its automatic
control of spray outputs based on canopy size, density, and
occurrence.
ACKNOWLEDGEMENTS
The authors express their appreciation to A. Clark, K.
Williams, B. Nudd, H. Jeon, J. Gu, A. Doklovic, and L.
Morris for their assistance throughout this research. This
research was supported by the USDA-NIAR SCRI (Grant
No. 2009-51181-06002). Financial support from the
OARDC SEEDS fund is also acknowledged.
REFERENCES
Chen, Y., H. Zhu, and H. E. Ozkan. 2012. Development of a
variable-rate sprayer with laser scanning sensor to synchronize
spray outputs to tree structures. Trans. ASABE 55(3): 773-781.
Chen, Y., H. Zhu, H. E. Ozkan, R. C. Derksen, and C. R. Krause.
2013. Spray drift and off-target loss reductions with a precision
air-assisted sprayer. Trans. ASABE 56(6): 1273-1281.
Derksen, R. C., H. Zhu, R. D. Fox, R. D. Brazee, and C. R. Krause.
2007. Coverage and drift produced by air induction and
conventional hydraulic nozzles used for orchard applications.
Trans. ASABE 50(5): 1493-1501.
Escola, A., F. Camp, F. Solanelles, S. Planas, and J. R. Rosell. 2003.
Tree crop proportional spraying according to the vegetation
volume. In Proc. 7th Workshop on Spray Application Techniques
in Fruit Growing, 43-49. Turin, Italy: University of Turin.
Farooq, M., and M. Salyani. 2002. Spray penetration into the citrus
tree canopy from two air-carrier sprayers. Tra ns . ASAE 45(5):
1287-1293.
Fox, R. D, D. L. Reichard, R. D. Brazee, C. R. Krause, and F. R.
Hall. 1993. Downwind residues from spraying a semidwarf
apple orchard. Trans. ASAE 36(2): 333-340.
Fox, R. D., R. C. Derksen, H. Zhu, R. A. Downer, and R. D. Brazee.
2004. Airborne spray collection efficiency of nylon screens.
Applied Eng. in Agric. 20(2): 147-152.
Gil, E., A. Escola, J. R. Rosell, S. Planas, and L. Val. 2007.
Variable-rate application of plant protection products in vineyard
using ultrasonic sensors. Crop Prot. 26(8): 1287-1297.
Giles, D. K., M. J. Delwiche, and R. B. Dodd. 1989. Sprayer control
by sensing orchard crop characteristics: Orchard architecture and
spray liquid savings. J. Agric. Eng. Res. 43(4): 271-289.
Gu, J., H. Zhu, W. Ding, and H. Y. Jeon. 2011. Droplet size
distributions of adjuvant-amended sprays from an air-assisted
five-port PWM nozzle. Atomization and Sprays 21(3): 263-274.
Hoffmann, W. C., and M. Salyani. 1996. Spray deposition on citrus
canopies under different meteorological conditions. Trans. ASAE
39(1): 17-22.
Jeon, H. Y., and H. Zhu. 2012. Development of variable-rate sprayer
for nursery liner applications. Trans. ASABE 55(1): 303-312.
Jeon, H. Y., H. Zhu, R. C. Derksen, H. E. Ozkan, C. R. Krause, and
R. D. Fox. 2011. Performance evaluation of a newly developed
Figure 9. Quad chart summarizing coverage and overspray indexes o
f
three trees and three growth stages to compare the overall spray qual-
ity of S1, S2, and S3.
1272 TRANSACTIONS OF THE ASABE
variable-rate sprayer for nursery liner applications. Tr ans .
ASABE 54(6): 1997-2007.
Lewis, F. H., and K. D. Hickey. 1972. Fungicide usage on
deciduous fruit trees. Ann. Rev. Phytopathol. 10(1): 399-425.
Pergher, G., R. Gubiani, and G. Tonetto. 1997. Foliar deposition and
pesticide losses from three air-assisted sprayers in a hedgerow
vineyard. Crop Protection 16(1):25-33.
Salyani, M., M. Farooq, and R. D. Sweeb. 2007. Spray deposition
and mass balance in citrus orchard applications. Trans. ASABE
50(6): 1963-1969.
Solanelles, F., A. Escola, S. Planas, J. Rosell, F. Camp, and F.
Gracia. 2006. An electronic control system for pesticide
application proportional to the canopy width of tree crops.
Biosystems Eng. 95(4): 473-481.
Syngenta. 2004. A user card containing the recommended droplet
density in the target area. Basel, Switzerland: Syngenta Crop
Protection AG.
Tumbo, S. D., M. Salyani, J. D. Whitney, T. A. Wheaton, and W. M.
Miller. 2002. Investigation of laser and ultrasonic ranging
sensors for measurements of citrus canopy volume. Applied Eng.
in Agric. 18(3): 367-372.
Wei, J., and M. Salyani. 2004. Development of a laser scanner for
measuring tree canopy characteristics: Phase 1. Prototype
development. Trans. ASAE 47(6): 2101-2107.
Wei, J., and M. Salyani. 2005. Development of a laser scanner for
measuring tree canopy characteristics: Phase 2. Foliage density
measurement. Trans. ASAE 48(4): 1595-1601.
Zhu, H., R. C. Derksen, H. Guler, C. R. Krause, and H. E. Ozkan.
2006a. Foliar deposition and off-target loss with different spray
techniques in nursery applications. Trans. ASABE 49(2): 325-
334.
Zhu, H., R. D. Brazee, R. C. Derksen, R. D. Fox, C. R. Krause, H.
E. Ozkan and K. Losely. 2006b. A specially designed air-assisted
sprayer to improve spray penetration and air jet velocity
distribution inside dense nursery crops. Tr ans. A SABE 49(5):
1285-1294.
Zhu, H., R. H. Zondag, R. C. Derksen, M. Reding, and C. R.
Krause. 2008. Influence of spray volume on spray deposition
and coverage within nursery trees. J. Environ. Hort. 26(1): 51-
57.
Zhu, H., M. Salyani, and R. D. Fox. 2011. A portable scanning
system for evaluation of spray deposit distribution. Computers
and Electronics in Agric. 76(1): 38-43.
