Cm/l Prorecmn Vol 14. No. 4. pp.

335-340, ,995
0261-2194(95)00002X Elaewer Saence Ltd
Printed in Great Rritain.
OZhI-2194195 $10.00 + 0.00

Weed control in sugar beet by precision guided

implements
R. P. van Zuydam,* C. Sonneveld* and H. Naber+
“DLO-Institute of Agricultural and Environmental Engineering (IMAG-DLO), P.O. Box 43, 6700 AA
Wageningen, The Netherlands and ‘Wageningen Agricultural University, Deptaltment of
Theoretical Production Ecology, P.O. Box 430, 6700 AK Wageningen, The Netherlands

A field experiment was carried out in sugar beet with a wide-span (12.2 m) tractor and laser guided
implements. By means of a side-shift facility, implements were mounted on this vehicle for seedbed
preparation, drilling, fertilizing, spraying and hoeing. Automatic laser guidance was possible with an
accuracy of 0.6 cm on a track length of 220 m on arable land.
No inputs (soil cultivation, fertilizer, herbicide) were made at places where they were not needed, or
even would potentially pollute the environment. The aim of the experiment, which was carried out in 2
successive years on fields of the same farm, was to investigate the influence on weed occurrence and
efficiency of weed control.
Leaving out seedbed preparation between the future crop rows left already germinated weeds alive. In
1 year pre-emergence application of paraquatdiquat was necessary to stop growing of well. developed
weed plants. A crumbling operation had to be carried out to break the clods, otherwise inter-row hoeing
was impossible. From the viewpoint of weed control restricting seedbed preparation to the future
sugar beet rows was not of advantage.
Precision guidance allowed enlargement of mechanical weed control, i.e. interrow hoeing to 80%
(40 cm wide at a row distance to 50 cm). Savings on herbicides were 75%, because little overlap was
necessary of chemically and mechanically treated areas. The absence of fertilizers in these inter-row
bands did not diminish the number of weeds, and speed of emergence of weeds. The effect of seedbed
preparation and drilling the sugar beet crop in complete darkness (at night), made possible by the
automatic guidance, on weed infestation was not different from daylight treatments. In these
experiments this so called photo-control of weeds was only demonstrable after hoeing.

Keywords: weed control, precision guidance, photo control

Many agricultural crops are grown in rows with inter- the rows. Both ways require some form of intelligence
row distances that vary between 0.25 and 0.75 m. and automation, the first to detect the spots where the
Intra-row spacing is often far less, varying from 0.05 to sugar beet plants are growing, the second to guide the
0.5 m. This means that on both sides of a crop row, hoe very closely along the sugar beet rows. Bontsema,
there are bands of soil which serve as spacer bands Grift and Pleyzier (1991) did research on how to
only. Nevertheless, these spacer bands are crumbled, distinguish the sugar beet plants from weeds, to make it
fertilized, sprayed and treated as part of the crop. possible to swing a hoeing knife through the row of
To lower the dependency of farmers on the use of sugar beet plants. Laser guidance of the hoeing
chemicals, and to decrease the overall use of chemicals implement, and prior to that of the sowing machine
by half in the year 2000 (MJP-G, 1992), research has to used in this experiment, make it possible to approach
be done on alternative weed control methods. Band the row of sugar beet plants very closely without risk of
spraying is an option by which a reduction in the damaging them.
chemically treated surface of a sugar beet field can be The aim of this research was to investigate the
achieved by enlarging the mechanically treated part. possible economy on fertilizer and herbicide inputs by
The normal practice nowadays is an inter-row treat- applying them only where needed, aided by the use of a
ment with a hoe that covers 6047% of the total surface precision guidance system. Also the effect of creating
(Kouwenhoven, Wevers and Post, 1991). different situations inside the row and between the
There are two possibilities of enlarging the mechanic- rows, related to the absence or availability of fertilizer
ally treated area: one is an intermittent weed killing in the inter-row bands and the absence or presence of a
action inside the sugar beet row, and the other, a germination bed at the same place, is of influence on
broadening of the mechanically treated area between the emergence and development of weeds. The option
of doing the various activities in complete darkness (at
night), or in the daylight, is of influence on the
“To whom correspondence should be addressed. germination of the weed seeds in the soil (Hartmann

Crop Protection 1995 Volume 14 Number 4 335

Weed control in sugar beet: R.P. van Zuydam et a/.

and Nezadal, 1990). Band fertilizing not only saves spiked roller. A V-shaped clod mover was mounted in
costs (De Wit, 1953), but also has an effect on the front of the teeth.
germination of weeds in the non-fertilized spacer bands Fertilizing was done in the first year with a Nodet-
(Bouwmeester, 1990). Gougis pneumatic fertilizer spreader with 12 double
The objectives of the experiment requires a very outlets, one outlet for each row. By lifting the machine
accurate guidance system for all the implements used, the full-field treatment was effected, because the
to be able to re-find the growing regions or bands in the fertilizer grains bounced in ali directions when hitting
field each time, and to make it possible to create a field the ground resulting in an even distribution. As the
build-up as a collection of alternating growing bands fertilizer grains were also bouncing sideways in row
and spacer bands. application, because of the airstream from the pneu-
matic transport system of. the spreader, this machine
was replaced by a Stegsted cereals drill in the second
Materials and methods year. On the latter, every fourth outlet was used to
obtain a row distance of 0.5 m. Because this machine
Airborne navigation systems such as the Global uses coulters instead of an airstream, the fertilizer was
Positioning System (GPS) using satellites as a reference dosed in neat bands of about 7 cm wide, with the
are unlikely to provide the accuracy and the speed coulters just touching the surface during application.
needed for guiding an implement over the field in a row The amount of fertilizer was 130 kg N ha-’ when dosed
crop (Tillett, 1991). Shmulevich, Zeltzer and Brunfeld full field and 35 + 60 kg N ha-’ at the row application,
(1987) described a laser scanning positioning system on given in two times: the first 1 week before sowing,
the subject of automatic guidance of vehicles in the the second 11 weeks after sowing. The first dose was, as
field. Accuracy of this method (0.15 m) is inadequate was the full field application, mixed with the soil top
with respect to the objective required for this applica- layer when preparing the seedbed, the second dose was
tion . given in a 2 cm deep furrow about 5 cm next to the crop
Therefore, a field-bound guidance system was row.
developed using a laser transmitter which emits a laser Drilling the sugar beet was done with a 24-element
beam rotating in a vertical plane. The transmitter was Vicon-FBhse (Monopill) precision drill, which is almost
positioned at the end of the field and had to be a standard implement for this purpose.
repositioned for each next double pass of the imple- In 1991 band spraying of herbicide was carried out by
ment. adjustment of the permanent spray boom on the gantry
The research project was carried out in 1991 and 15 cm above ground level. In this position the spraying
1992 on the ‘Oostwaardhoeve’ experimental farm in width per nozzle was limited to approx. 12.5 cm, with
Slootdorp. The soil type, though slightly varying over the flat fan nozzles (Teejet 4003E; 40 degrees top
the field, was mainly a highly calcareous loamy sand angle, even spray variety, spray pressure 2.1 bar) being
type of approx. 20% silt, 1.5% organic matter and a pH spaced 0.5 m apart and thus spraying right over the
(KCl) level of 7.6. The 20 ha field was split up into two future crop rows. Due to the very low height of the
10 ha fields, each with a 5-year rotation. The 5-year spray boom, the drift of droplets by wind was also
rotation was made up with cereals, potatoes, sugar beet, limited to almost zero. Two post-emergence sprayings
broccoli and iceberg lettuce. This resulted in two X five were carried out with so called low dosage mixtures of
plots, each of 73.2 m X 200 m. All tests were restricted phenmedipham, metamitron, ethofumesate and
to sugar beet because of the availability of mechaniza- mineral oil (0.5 1 or kg ha-’ of the commercial products,
tion for this crop on the special-type vehicle. in 200 1 water). In 1992 one full field application
The vehicle used was a Dowler gantry, a wide-span pre-emergence was carried out with metamitron
(12.2 m) vehicle with four wheels. On each side there is (2.8 kg ha-’ a.i., in 300 1 water using nozzles Albuz
a main wheel with hydrostatic drive and a pivoting rear green at 2.3 bar). In both years the fixed quadrats used
wheel in a parallelogram-type suspension. The vehicle for counting the number of weeds were not treated with
has a 70 kW (90 hp) turbo diesel engine that makes it herbicides.
suitable for secondary tillage and nursing activities When weed control became necessary in the ‘spacer
only, the implements having a 12 m working width and bands’ between the crop rows, hoeing was executed
working in between the tracks of the wheels on 12 m with a Kongskilde Vibra-Beta hoeing equipment
wide beds. (23 elements). These elements each existed of a
The main tillage was carried out in November of the parallelogram-mounted frame, that carried three spring
previous year with a four-furrow reversible plough tines, each equipped with a duckfoot chisel. By sliding
drawn by a 108 kW (145 hp) tractor on wide tyres with the outer teeth to the side, the working width of the set
a maximum air pressure of 1 bar. The ploughing depth of three could be adjusted by increasing or decreasing
was 0.28 m. the overlap of adjacent chisels. Each crop row was
For preparing the full field seedbed, a 81 kW protected by teethed, concave protection discs that
(110 hp) tractor was used, driving a 4.5 m wide rotary shielded the crop from moving soil and clods when the
tiller of the type ‘Lelyterra’ for full-field plots. The chisels passed.
tractor was equipped with wide-section tyres (0.42 m) The two opposite oriented trial fields were approx.
at a tyre pressure of 1.0 bar. The row-type seedbed 1.5 ha each and separated by a field road and a ditch.
preparation was carried out with the gantry and a 24- Along this road the adaptors for the laser transmitter
element Stanhay Webb row crumbler. This implement pedestal were present. The fields were ploughed in
exists of 24 parallelogram-mounted elements, each autumn and cultivated once with a rigid tine cultivator
carrying three vertical round teeth and a cast iron with duckfoot chisels during wintertime.

336 Crop Protection 1995 Volume 14 Number 4

 Weed control in sugar beet: R.P. van Zuydam et a/.

The experimental plots were 12 m wide and had a made ‘full-field seedbed’ = seedbed present, and
length of 45 m each. Four factors were investigated in ‘row-type seedbed’ = untouched soil.
two variations. In 1991 all variants and factors were In 1991 the number of weeds was significantly lower
combined in one experiment. This gave 24 combina- on the prepared soil compared to the untouched soil,
tions; in three replications this means a total of 48 plots. but only during the first 4 weeks after sowing (Table I).
In 1992, variants A and B were combined (see below), The increase of the number of weeds on the prepared
and so were variants C and D, resulting in 24 plots for 6 soil was much faster than on the untouched spots. The
repetitions. The variants were: higher amount of biomass on the untouched soil is the
result of bigger sized (older) weeds,
Band application of fertilizer, compared with full-
field application. In 1992 weeds present on the ploughed land (before
Seedbed preparation in bands, compared with seedbed preparation) were rather well developed and
full-field cultivation. had therefore been treated with a full-field dose of
Seedbed preparation in the dark (on moonless paraquat-diquat applied one week before sowing. This
nights; co.1 Ix), compared with seedbed prepara- resulted in a lower number of weeds on the untouched
tion in daytime. soil compared to the soil cultivated for seedbed
preparation, indicating also a faster development of the
Sowing and hoeing in .the dark, compared with
weeds on prepared soil as on untouched soil. The
the same in daytime.
weeds were very small, resulting in low amounts of
All treatments were done with the 12.2 m wide biomass.
gantry, except the full-field seedbed preparation, which Also in 1991 the two implement types for making a
was done with a normal tractor with a rotary cultivator. germination bed were compared in relation to the weed
The demands for the light level when working in development on the treated surface. The number of
darkness corresponded with previous research by weeds in the row were counted. Difference in the
Hartmann and Nezadal (1990). This made the use of emergence of weeds was not measured. The weeds that
an automatic guidance system, to guide the drilling grew in the rotavated seedbed were larger: biomass was
machine, the hoe, etc., accurate over the field in almost double compared to the row type seedbed.
complete darkness, so obligatory. The sugar beet
variety was Lucy, the row distance was 0.5 m and the Presence and dose or absence of fertilizing
seed spacing in the row was 0.2 m.
In 1991 the numbers of weeds were counted 7 times For this aspect, the availability of fertilizer for the
at intervals of 1-2 weeks at 4 fixed places in the rows of germination of weeds, the results of countings in the
sugar beet and on four fixed places between these rows rows and between the rows are used for analysis. Here,
to determine the effect of the various treatments. There the comparison can be made ‘full-field fertilized’ =
were three repetitions of each treatment, resulting in 12
countings per treatment. Counting was done with the
help of a rectangular steel bar framework of 0.125 X Table 1. Number of weeds me2 between crop rows. Average
countings on 3 X 4 (1991) and 10 (1992) fixed spots. Biomass in
0.8 m (0.1 m’), on fixed quadrants. When the canopy g dry matter per m2, at 10 weeks after sowing
closed, the countings were terminated. In 1991 the
statistical software used to analyze the data was 1991 1992
GENSTAT. Weeks after Germination Germination
In 1992 the number of weeds was counted three sowing bed Untouched bed Untouched
times, 4 and 8 weeks after sowing and 6 weeks after
3 1.8** 10.3
hoeing. The countings were done at 10 fixed places in 4 8.6” 14.1
the row of sugar beet, and at 10 fixed places between 5 17.5 19.1 15.5** 3.8**
the rows. The rectangular framework used to indicate 7 25.0 22.8
the area to be counted was the same as in 1991. 8 24.7 23.4
10 38.0 47.7 20.0** 5.3**
Statistical analysis of the data was done with SPSS this Biomass (g) 338 446 0.7 0.2
year.
*P < 0.10; **p < 0.05

Results
Table 2. Number of weeds m-” in the crop rows. Average
countings on 3 x 4 fixed spots. Biomass in g dry matter per m2,
The most important weeds were Polygonum aviculare,
at 10 weeks after sowing
Matricaria chamomilla, Poa annua and Solanum
nigrum , but also 20 other species were observed, 1991
irregularly spread over the field. Because the number Weeks after Germination bed Germination bed
of individuals was low, it was not useful to present the sowing by rotavator by row crumbler
results of the countings carried out per species. There-
3 0.9 1.4
fore only the numbers of all weed plants are 4 12.3 7.8
mentioned. 5 16.7 20.0
7 26.4 27.4
8 22.6 25.9
Presence or absence of seedbed 10 41.5* 54.4
Biomass (g) 184 105
The number of weeds between the crop rows is of
interest here for analysis. The comparisons can be *P < 0.10

Crop Protection 1995 Volume 14 Number 4 337

Weed control in sugar beet: R.P. van Zuydam et al.

Table 3a. (1991) Fertilizing yes/no. Number of weeds m-‘. Average countings on 3 X 4 fixed spots. Biomass in g dry matter m2

In crop rows Between crop rows

Weeks after sowing Fertilizer 130 kg N had Fertilizer 35 kg N ha-’ 130 kg N ha-’ 0 kg N had

3 0.9 1.4 4.4 4.2

4 4.4 10.8** 9.0* 13.4
5 17.2 19.6 16.6 20.1
7 25.8 28.0 21.5 26.5
8 22.6 26.0 23.1 25.0
10 46.8 48.2 40.6 44.3
Biomass (g) 138. 306 276** 546

*P < 0.10; **p < 0.05

Table 3b. (1992) Fertilizing yes/no. Number of weeds m-‘. the number of weeds in 1991 is lower when sowing is
Average countings on 10 fixed spots. Biomass in g dry matter carried out at night compared to sowing at daytime;
m2
however, this was not confirmed in 1992.
Seedbed preparation in the dark (Figure 4) followed
Between crop rows
Weeks after Fertilizer Fertilizer by sowing at night compared to sowing in the light did
sowing 130 kg N ha-’ 0 kg N had not show any difference in the number of weeds.
This justifies the conclusion that seedbed preparation
5 15.5 13.0 and sowing in darkness have no photocontrol effect on
10 20.0 13.5
Biomass (g) 0.7 0.3
the weeds.
The results of the integrated chemical and
mechanical weed control, outside the area where
countings were conducted, were very good. Without
fertilized and ‘band fertilized’ = fertilized with the
first dose (in the row) or not fertilized (between
the rows). Tables 3a and b show that the number of
weeds is not significantly different between fertilized
and not fertilized treatments. After week 10 a supple-
mentary dose of 65 kg N ha-’ was given (in the row
fertilized object) to complete the 75% dose compared
to full field fertilizing. This reflects on a full field dose,
so actually only 25% of these amounts were supplied,
on 12.5 cm wide bands with 37.5 cm wide inter-row
spacings. The significant higher amount of biomass of
5oo 0

n
light

dark

the weeds on non-fertilized strips can not be made

clear.

Photocontrol
To determine the influence of light during soil cultiva-
tion on the germination of weeds, the effect of seedbed 1992

preparation and also of hoeing during the day or at Figure 1. Seedbed preparation in light/in dark. Number of
night can be measured between the sugar beet rows. weeds mm2-weeks after sowing.
Due to interactions between ‘seedbed preparation’
and ‘sowing’, carried out sequentially, the results of
these treatments should only be judged together. These
influences can be measured in the rows of sugar beet. 1
The results of these types of spoil cultivations in dark
and in light on the emergence of weeds are already
described in detail by Naber el al. (1992) and are
summarized in this paper in Figures Z-4.
Figure 2 shows that the number of weeds after
cultivation at night is not lower than after cultivation
during the day, so no photocontrol effect on the weeds
of seedbed preparation at night could be established.
The effect on the emergence of new weeds after
hoeing in the dark (Figure 2) was only significantly
different from hoeing in daylight in 1991, although also 0
4 6 7 6
in 1992 a clear tendency towards fewer new weeds
emerging after hoeing in the dark, so photocontrol, 1991 1992

appeared to work. Figure 2. Hoeing in light/in dark. Number of weeds mm2- weeks
After seedbed preparation in the daytime (Figure 3)) after sowing.

338 Crop Protection 1995 Volume 14 Number 4

 Weed control in sugar beet: R.P. van Zuydam et al.

6’31 rows (in the ‘spacer bands’) by applying the fertilizer

only in bands of 12.5 cm wide, is technically no
50
problem with the laser-steered implements, either
before sowing, before emergence or after emergence of
40
the sugar beet crop. Not applying fertilizer to the
Ulight
spacer bands increases the efficiency with which the
30 sugar beet plants use the fertilizer (the uptake) and
n dark
minimizes the chance of leaching. It was expected also
20 that the germination of weeds and their development
would be less when these bands were not fertilized. In
10 fact no difference was found in the numbers of weeds
growing on fertilized or on non-fertilized spots.
0L Theoretically, weed infestation, in number and in
3
biomass, might be reduced by the placement of
1991 fertilizer out of reach of the weed seeds. It cannot be
Figure 3. Seedbed preparation in light, sowing in light/in dark. excluded that residual effects from the previous year
Number of weeds my2 -weeks after sowing. (some nitrate left from the preceding crop) caused a
nitrate level that was high enough to stimulate the
germination of weeds between the rows.
60,-
Research into the effects of photocontrol was the
third item of the field experiment in the 2 years. The
laser-steered implements are ideal for this type of
experiments as seedbed preparation (band cultivation),
sowing and hoeing can be carried out in complete
darkness without difficulty and with high precision.
Jensen (1992) obtained positive effects of photo-
control in his field trials in spring barley and winter
wheat. Also Scopel, Ballare and Radosevich (1992)
report about lower weed infestation after soil cultiva-
tion in the dark, even if they use the tractor’s lights.
In our two year experiments seedbed preparation at
night was not shown to be effective in decreasing the
3 4 5 7 8 10 5 10 number of weed seedlings. Also the combination with
sowing, both during the day and at night, does not
1991 1992
produce a combination that was of advantage. Even if
Figure 4. Seedbed preparation in dark, sowing in light/in dark. all the work is done in the dark, the number of weeds is
Number of weeds m” - weeks after sowing.
the greatest.
It should be concluded here that the germination of
weed seeds depends on more factors than on light
any technical problem herbicide application and hoeing alone. Post (1994) remarks that the need of the weed
could be carried out with the laser guided implements. seeds for light to germinate can be compensated by dry
weather conditions after soil cultivation. The seeds
dry out under these circumstances and germinate
Discussion and conclusions without receiving light when the soil becomes moist
again. The loamy sand soil used in our experiments
The aim of the 2 year field experiment was to easily dries out as was already mentioned under
investigate precision guided implements for crop seedbed preparation.
management, focused on weed control in sugar beet. The only positive effects of photocontrol in our field
The advantage of not making a germination bed experiments were found after hoeing. In both years the
between the rows of sugar beet to be sown is too small amount of new germinating weeds after hoeing was
to justify practical application on this loamy sand soil. lower, when this mechanical weed control was carried
This mechanical effect can hardly be missed in the total out in the dark.
weed control operations, unless herbicides are used to The laser guidance of the implements on the wide
control them. Seedbed preparation has a killing effect span tractor has shown to be completely reliable even
on the weeds that had germinated during winter time in the dark, so a straight course of all implements is
and early spring. When the spacer bands between the guaranteed at any time, and hoeing the weeds close
sugar beet rows are not prepared as the rest of the to the sugar beet plants is possible, increasing the
seedbed, the soil is left in a very rough, cloddy state percentage of mechanically treated surface. Reduction
that is hard to hoe without damaging the small sugar of the chemically treated area for weed control leads to
beet plants by the big clods that are shifted sideways. a minimum band size of 12.5 cm. Combination with
The theory behind leaving out seedbed preparation inter-row hoeing of 0.4 m wide (2.5 cm overlap) results
between the rows to spare energy (fuel) was shown to in an optimum integration of chemical and mechanical
be unrealistic either from the view point of weed weed control and an increase in the mechanically
control and from the point of crop management. treated percentage of the surface from 60-67% in
Avoiding the placement of fertilizer between the normal practice up to 80% in this system.

Crop Protection 1995 Volume 14 Number 4 339

Weed control in sugar beet: R.P. van Zuydam et al.

Acknowledgement Kouwenhoven, J. K., Wevers, J. D. A. and Post, B. J. (1991)

Possibilities of mechanical post-emergence weed control in sugarbeet.
Soil Tillage Research 21, 85-98
The authors want to thank the students Peter Baltus
(1991) and Ilse Leeuwerik (1992) who did the observa- MJP-G (1992) Multi-year Crop Protection Plan, Ministry of
Agriculture, Nature Management and Fisheries, The Hague
tions on weed control.
Naher, H., Van Zuydam, R. P., Baltus, P. C. W. and Leeuwerik,
I. D. (1992) Maximum integrated weed control in sugarbeet with the
help of a laser steered Gantry. Proceedings ANPP- Fifteenth Columa
References conference, Versailles 2, 3 and 4 December 1992, pp. 269-276
Post, B. J. (1994) Invloed van bestrijding van onkruiden op de
Bliss, D. and Smith, H. (1985) Penetration of light into soil and its populatiedynamiek: het belang van zaadvoorraad en kiemingsdyna-
role in the control of seed germination. PI. Cell Env. 8, 475-483 miek. Gewasbescherming 25(5), 1.67-172 (in Dutch)
Bontsema, J., Grift, K. (1991) Mechanical weed
J. and Pleyzier, Scopel, A. L., Ballare, C. L. and Radosevich, S. R. (1992) Promotion
control in sugarbeet growing: the detection of a plant in a row. of weed seed germination by light during soil tillage. Bull. Ecol. Sot.
Proceedings of the IFAC-ISHS Workshop, Matsuyama, Japan Am. (Suppl.) 73(2), p. 340

Bouwmeester, H. J. (1990) The effect of environmental conditions on Shmulevich, I., Zeltzer, G. and Brunfeld, A. (1987) Guidance System
the seasonal dormancy pattern and germination of weed seeds. Ph.D. for Field Machinery Using Laser Scanning Methods. Paper no. 87-
Wageningen Agricultural University, 157 pp 1558 for 1987 Winter Meeting of American Society of Agricultural
Engineers, Chicago, 24 pp
De Wit, C. T. (1953) A physical theory on placement of fertilizers.
Ph.D. Den Haag, Staatsdrukkerij 71 pp N. D. (1991) Automatic Guidance Sensors for Agricultural
Tillett,
Field Machines: A Review. J. Agric. Engng Res. 50, 167-187
Hartmann, K. M. and Nezadal, W. (1990) Photocontrol of weeds
without herbicides. Naturwissenschaften, 17, 158-163
Received 29 December 1993
Jensen,P. K. (1992) First Danish experiments with photocontrol of Revised 29 November 1994
weeds. Z. PflKrankh. PflSchutz, Sonderheft 13, 631-636 Accepted 8 December 1994

340 Crop Protection 1995 Volume 14 Number 4