|
Background and Objective:
With increasing availability and decreasing cost of GPS and GIS
technology, there is intense interest in the development of site-specific farming.
Site-specific farming caters management practices to unique areas within a field,
encouraging the application of appropriate input rates. This technique reduces both
environmental and economic risk.
Two components of site-specific farming are of interest in this study. Site-specific
weed management uses weed infestation maps created through ground reconnaissance or remote
sensing to facilitate spot spraying. This allows the farmer to target each weed species
individually and on a spatially specific basis in one pass. This would provide more
effective weed control on a whole field basis, leading to increased yields and more
prudent use of herbicides. The efficacy of spot spraying and the benefits associated with
it are reliant upon the accuracy of the weed infestation map and its functional lifespan.
We can only determine the useful lifespan of a weed map if we understand how, why, and at
what rate weed patches move. If weed patches move unexpectedly, maps made in one year may
be useless for the next year.
Site-specific fertilizer application constitutes a large portion of efforts in
precision agriculture. To apply this technique, areas of differing yield potential are
mapped according to crop response to landscape variability. Farmers then typically boost
fertility in high yield potential areas. These areas are often at relatively low
elevations where increased moisture can support higher yielding crops. The higher moisture
levels not only support the crop, but encourage weed growth as well. The effect of
site-specific fertilizer application on weeds is currently unknown.
Researchers have found that fertilization increased number of wild oats present in an
oat field and early spring applications of nitrogen broke dormancy of greater number of
wild oat seeds, increasing wild oat germination. Various researchers have also reported
that increased N at high weed densities provided little improvement in crop yield, but
tended to increase weed growth. This research suggests that high rates of applied nitrogen
may make weeds more competitive. The weeds may thus produce more seed, resulting in an
unpredicted weed patch spread. This could prove costly for producers in terms of
unexpected yield loss, and will be of particular concern to producers who are dealing with
herbicide resistant weed patches. In addition, it will hinder the development of
site-specific weed management by decreasing the predictability of weed patch spread and
lowering the value of weed maps.
The objective of this project was to determine if the practice of
site-specific fertilizer application influenced the competitive ability of wild oat or
wild buckwheat in spring wheat.
Procedure and Project Activities:
Experiments were conducted at two sites (Birtle and Carman, Manitoba) and in two
separate seasons (1998 and 1999).
Birtle Site Description
The Birtle site was located on the farm of Ron Bell northwest of Miniota, Manitoba at SE
32-14-25W. This site is classified as a gently undulating glacial till soil of the Newdale
association. The relief difference between the knoll position and the foot position was
approximately 4 m, and the gradient did not exceed 4%. Prior to this study, the field had
been cropped for more than 50 years. The field was cropped in a wheat fallow rotation
until 1976, when a continuous cropping rotation was adopted. In 1978, a minimum tillage
management system was established, and by 1988 tillage was eliminated and a
cereal-broadleaf rotation was initiated. The field had been under a zero-tillage
management system for 9 years before experimental trials were established on the site. The
crop rotation sequence on the experimental site was peas (Pisum sativum): canola (Brassica
napus): wheat (Triticum aestivum): wheat: wheat from 1995 to 1999.
Carman Site Description
The Carman site was located on the University of Manitoba Carman Research Farm. This site
is classified as a localized depression of the La Salle soil type, ranging from sandy clay
loam to clay loam in texture. Relief between the knoll and foot positions differed by
approximately 1m, and the gradient did not exceed 5%. The experiment was a field scale
trial. Prior to this trial, the field had been cropped in an annual crop rotation for at
least the previous 4 years. The field is considered to be conventionally tilled.
Experimental Design
Treatment factors included landscape position (Foot slope and Knoll), target weed density
(0 to 100 plants m-2), and nitrogen application treatment (0 versus 90 kg/ha actual N).
Nitrogen application treatments were replicated twice at the Birtle site, three times at
the Carman site in 1998, and twice at the Carman site in 1999. Target weed density
treatments were replicated three times at all site-years, but these were not considered
true replicates because the actual weed densities in each plot were never exactly the
target densities. Landscape position could not be replicated. The wild buckwheat and wild
oat experiments were conducted at the same site-years, but were analyzed separately .
Treatment Application
At the Birtle site, the experiments were applied within an existing experiment being
carried out by the Department of Soil Science, University of Manitoba. They had created 9
m by 212 m transects which cut across an undulating landscape. Nitrogen treatments (0kg/ha
versus 90kg/ha actual N) were broadcast prior to seeding each year and treatment transects
were replicated once. This experiment was conducted at two landscape positions (Foot and
Knoll) on 4 separate transects.
At the Birtle site, hard red spring wheat (CHRSW, cv Teal) was direct air-drilled into
the seedbed at a seeding rate of 128 kg ha-1 on May 4 and June 8 in 1998 and
1999, respectively. At the Carman site, hard red spring wheat (cv AC Barrie) was seeded
into a tilled seedbed at a rate of 103 kg/ha on May 8 and May 28, in 1998 and 1999,
respectively using a double disk press drill. At both sites, nitrogen was broadcast on the
surface (34-0-0) prior to seeding. This nitrogen was incorporated by pre-seeding tillage
at the Carman site only. Weed seeds were sown by hand and raked into plots. After crop
seeding plots were sprayed with herbicides that would remove unwanted weeds.
Field Data Collection
Measurements from plots used to determine the influence of treatment factors on wheat
yield included weed density, wheat grain yield, and wheat and weed dry biomass.
Characterization of landscape to explain any variation in wheat yield by treatment factors
included measurement of soil fertility, gravimetric soil moisture, soil profile
characterization and site topographical characterization. Weed density counts were taken
when weeds were at the 2-3 true leaf stage using two 0.1 m2 quadrats per plot.
The average of both density counts served as a measure of the actual weed density in each
plot.
Analysis
All data were first treated to separation on the basis of treatment factors using analysis
of variance. Factors which proved significant were further analyzed using regression with
weed density as the independent variable.
Results and Discussion:
Effect of Nitrogen Application and Landscape Position on
Weed-Free Wheat Yield
Weed-free wheat yield did vary significantly within experimental sites, but the variation
was poorly accounted for by either landscape position or nitrogen application rate.
Landscape position had no significant effect on weed-free wheat yield for any site-year.
There was a consistent trend of higher yields at the knoll versus the foot landscape
positions, but both 1998 and 1999 were relatively wet years. In only 50% of the site-years
was there a significant effect of nitrogen application on weed-free wheat yield. This is
surprising given that the high rate of nitrogen application was 90 kg/ha actual N.
Available nitrate levels were relatively high for all site-years. This would help to
explain the lack of consistent response of weed-free wheat yield to N application rate.
Effect of Nitrogen Application and Landscape Position on
Wild Oat Competitiveness
Wild oat density had a significant effect on wheat yield. As wild oat density increased,
wheat yield decreased. The competitiveness of wild oat was unaffected by landscape
position for all site-years. Under wild oat infested conditions, wheat yield always
responded positively to the application of nitrogen. This result is different from the
lack of wheat yield response to nitrogen application under weed-free conditions. It
suggests that the presence of wild oat caused available nitrogen to become more limiting,
leading to a more consistent yield response of wheat to the application of nitrogen.
Under high densities of wild oat (over 150 plants/m2), the rate of wheat
yield loss per wild oat plant was significantly greater when nitrogen was applied versus
when it was not applied. This suggests that wild oat is a preferential consumer of applied
nitrogen in comparison to wheat. This is supported by the fact that wild oat biomass was
significantly greater when nitrogen was applied and its relative biomass (relative to
wheat plants it was competing with) was also significantly greater when nitrogen was
applied versus when it was not applied. Wild oats ability to preferentially consume
nitrogen in the presence of spring wheat may be related to the fact that its rooting depth
is similar to that of wheat and its early season emergence and growth rate, both above and
below ground, can be greater than spring wheat. Wild oat biomass and relative biomass
where never found to be significantly affected by landscape position.
Effect of Nitrogen Application and Landscape Position on
Wild Buckwheat
Competitiveness
Wild buckwheat density did not consistently have a significant effect on wheat yield. The
competitiveness of wild buckwheat was not significantly affected by landscape position.
For most site-years, wild buckwheat caused no significant yield loss in wheat, even when
it was present at densities of up to 150 plants/m2. The application of nitrogen
fertilizer did not significantly affect the competitiveness of wild buckwheat, but in 3 of
4 site-years the presence of wild buckwheat significantly reduced the yield increase in
spring wheat caused by the application of nitrogen fertilizer. When no nitrogen fertilizer
was applied, the presence of wild buckwheat caused a slight (but not significant) increase
in wheat yield.
The relative biomass of wild buckwheat (relative to wheat plants it was competing with)
was unaffected by landscape position and by the application of nitrogen fertilizer. The
latter result suggests that wild buckwheat is not a preferential consumer of nitrogen in
comparison to wheat. This may be related to the fact that wild buckwheat is relatively
shallow rooted compared to wheat.
Landscape position did not appear to affect spring wheat yield, although there was a
trend towards greater weed-free wheat yields at the knoll versus the foot landscape
position in years when there was adequate (and sometimes excessive) moisture. The
application of nitrogen (at rates of 90 kg/ha of actual N) caused no consistent increase
in weed-free spring wheat yield, and there was no interaction of nitrogen and landscape
position (the application of nitrogen was not effective at increasing wheat yield at only
one landscape position, for example). These results suggest that the available nitrogen
levels at these experimental sites was high to begin with and that the addition of
nitrogen fertilizer under these conditions may not be economically warranted. It also
suggests that spring soil tests for nitrogen can be valuable. The results also suggest
that spring wheat yield potential cannot be reliably predicted on the basis of simple
landscape position classification.
Wild buckwheat is relatively uncompetitive in spring wheat, although it does cause
harvesting problems because of its twining nature, tough stems and late maturity.
Landscape position appears to have little effect on the competitiveness of wild oat or
wild buckwheat in spring wheat. The application of nitrogen, however, increased the
relative competitiveness of wild oat, and in the presence of wild buckwheat the potential
yield benefit to spring wheat of applied nitrogen was lost. These results suggest that
good weed control is more important in fertilized versus unfertilized spring wheat. In
cases where no weed control is applied (organic or pesticide-free production (PFP)),
additional care should be taken to assure that nitrogen is made available to the crop and
not to the weeds. In a PFP system for example, where fertilizer application is allowed,
fertilizer placement to favour the crop (such as side-banding) would be a better strategy
than broadcasting fertilizer.
Acknowledgments:
This project was made possible due to funding from the governments of Manitoba and
Canada through the Canada-Manitoba Agri-Food Research and Development Initiative (ARDI),
and funding from Novartis Crop Protection and the Canadian Fertilizer Institute (who
supplied a partial scholarship to D. Ross). Norwest Labs helped to reduce the costs of
soil sample analysis, Grant Manning and the Department of Soil Science, University of
Manitoba shared their data and Mr. Ron Bell for allowing the work to be conducted on his
farm and provided extensive assistance. Technical and statistical assistance was provided
by Mr. B. R. Oree and Mr. L. Friesen, respectively.
|
HomeContact
UsAbout
ManitobaDepartmentsLinksPrivacy
