Mechanical management of weeds - Ecological Management of Agricultural Weeds - page 152

A second strategy is to delay seedbed preparation and planting until after
many of the weeds have germinated and are susceptible to destruction by soil
disturbance. This approach appears to be most successful for spring-planted
row crops, particularly maize and soybean, and for autumn-planted grains.
Delaying planting from 25 April until 15 May in Wisconsin reduced in-row
weed density in rotary hoed maize by an average of 55% (Mulder & Doll,1994).
However, late planting also reduced maize yield in one of two years.Similarly,
late-planted soybean often has lower weed densities and lower percentage
yield loss to weeds relative to early-planted soybean, but may suffer some
reduction in yield due to shortened growing season (Weaver, 1986; Gunsolus,
1990; Buhler & Gunsolus, 1996).A substantial percentage of fall-germinating
weeds, including Alopecurus myosuroides , Bromus secalinus , Avena ludoviciana ,
Veronica persica , and Lamium purpureum , can be eliminated by delaying the
planting of winter wheat, but late planting often reduces yields (Moss, 1985 a ;
Koscelny et al ., 1991; Christensen, Rasmussen & Olesen, 1994; Singh et al .,
1995; Cosser et al ., 1997).
Delayed planting can also be useful for weed management in some spring-
seeded vegetable crops (Wellbank & Witts, 1962), but will generally not be
useful in summer-seeded vegetables. The latter are typically infested with
weeds like Galinsoga ciliata and Portulaca oleracea that germinate over a wide
range of dates in response to tillage. Note, however, that this type of weed is
highly susceptible to stale and false seedbed techniques (see below).
In all the studies discussed above, the seasonal cycle was dominated by
changes in the temperature regime. Whether similar planting date strategies
exist for some crop-weed combinations in the wet-dry seasonal tropics
remains to be determined.
Forcella and others have addressed question (2) regarding determination of
the optimum planting date through a series of models that predict the per-
centage emergence of individual weed species based on soil temperature and
moisture data (Forcella, 1993, 1998; Harvey & Forcella, 1993; King & Oliver,
1994; Wilen, Holt & McCloskey, 1996).The models predict relative weed pres-
sure at successive potential planting dates as a proportion of the density that
would occur without seedbed preparation.The time to plant occurs when the
expected weed density falls below a threshold the grower believes he/she can
control with the weed management tools available, or, alternatively, when the
expected yield loss due to further delay in planting exceeds the expected yield
loss due to weeds. A problem with these models is that they predict the per-
centage of ultimate emergence that has occurred by a given date rather than
the actual density. The model of Reese & Forcella (1997) is presently paramat-
erized for 15 weed species common in the central USA.