Agriculture Reference
In-Depth Information
is imperative that these cultivars be used in a
manner that ensures long-term viability of this
herbicide.
with concomitant economic and environmental
benefi ts (Mortensen et al., 1998; Timmermann
et al., 2003).
The potential for herbicide use reduction
with site-specifi c weed management varies
between crops. Larger reductions are possible
in competitive crops like wheat than in less
competitive row crops like sugar beet (
Beta
vulgaris
L.) or maize. In a 4-year experiment
conducted on fi ve fi elds, grass-herbicide use
reductions using a GPS-guided sprayer were
90% in winter cereals (barley and wheat), 78%
in maize, and 36% in sugar beet (Timmermann
et al., 2003). Broadleaf herbicide use was
reduced 60% in winter cereals, 11% in maize,
and 41% in sugar beet. The cost of wild hemp
(
Cannabis sativa
L. ssp.
spontanea
) control in
Hungary was reduced 9%, and herbicide applica-
tion was reduced 34%, with the use of site-specifi c
weed control methods compared with broadcast
spraying of the entire fi eld (Reisinger et al., 2005).
Wild hemp is very competitive against winter
wheat and frequently forms large and stable
patches in agricultural fi elds.
Site-specifi c weed management systems require
accurate detection and location of weeds in fi elds.
Two general approaches to weed detection in cul-
tivated crops are available. One approach uses
morphological differences such as leaf shape,
hairiness, shininess, or plant structure to differ-
entiate between the crop and weeds, while the
other approach uses differences in spectral refl ec-
tance for plant differentiation (Girma et al., 2005).
The detection of multiple weed species within a
crop is challenging, particularly when attempting
to detect grass weeds in a grass crop such as
wheat.
Weed distribution is often associated with
agronomic practices and fi eld characteristics
(Mortensen et al., 1998). Weed presence in Danish
cereal grain fi elds was most dependent on cereal
grain species and clay content of the soil
(Andreasen et al., 1991). Large-seeded broadleaf
species were associated with higher organic matter
and lower elevation in Nebraska maize fi elds,
while annual grasses were found in well-drained,
higher areas (Mortensen et al., 1998). The
correlation between soil properties and weed
Biological control
Examples of commercially successful biological
control are rare in annual cropping systems. Rhi-
zobacteria have been screened for use as possible
biological control agents of downy brome and
jointed goatgrass in winter wheat (Kennedy et al.,
1991; Kennedy and Stubbs 2007). Chemical anal-
ysis of the active fraction complex from strain D7
of
Pseudomonas fl uorescens
, which showed promise
for biocontrol of downy brome in agar studies,
identifi ed chromopeptides and other peptides,
fatty acid esters, and a lipopolysaccharide matrix
(Gurusiddaiah et al., 1994). Separation of any of
the components of the matrix resulted in nearly
total loss of activity. Despite some limited experi-
mental success with rhizobacteria for biocontrol
of weeds in wheat, no commercial application of
this technology is currently available.
Cromar et al. (1999) contend that while most
biological control efforts for weeds have centered
on predators with specifi c feeding habits, inverte-
brates with opportunistic feeding strategies may
provide the most effective broad-spectrum weed
control. They urged use of management practices
that conserve and encourage benefi cial arthropod
populations.
WEED SPATIAL VARIATION AND
PRECISION FARMING
Spatial heterogeneity is common in crop produc-
tion fi elds. Heterogeneity is observed as nonuni-
form plant vigor, crop yield, soil color, and pest
infestation (Mortensen et al., 1998). Weed density
and composition are not uniformly distributed
across fi elds. Many weed populations are aggre-
gated in patches of varying size and shape
(Gerhards and Oebel 2006). Managing weed pop-
ulations on less than a whole-fi eld basis, also
known as site-specifi c weed management, could
result in a signifi cant reduction in herbicide use
