Predicting
Longterm Giant Foxtail and Velvetleaf Populations
C.M.
Boerboom and A.J. Bussan*
Introduction
The high cost of
herbicide programs in corn and soybeans has created an interest
in using reduced herbicide rates to lower production costs. The
successful use of reduced herbicide rates in Wisconsin and the
associated management considerations have been summarized
previously (Doll et al. 1992, Proost et al. 1996).
Unfortunately, most of the research in Wisconsin has focused on
the use of preemergence-applied herbicides in corn. While other
states have extensively researched postemergence herbicide use
in soybeans (Defelice and Kendig 1994), less research on reduced
rates of postemergence herbicides has been conducted in
Wisconsin, especially in corn. Therefore, we conducted a two
year study to determine how effective reduced postemergence
herbicide rates were in Wisconsin corn and soybeans (Bussan and
Boerboom 1997). Still, many farmers are concerned that using
reduced herbicide rates may leave a few escaping weeds and that
these escapes will lead to increased weed problems in the
future. Therefore, we also wanted to be predict what the
long-term impact of using reduced rates of postemergence
herbicides would be on future weed populations.
Computer
Simulation Model
To make these
predictions, we needed to develop a computer simulation model
that would, in essence, be able to accurately "grow" a
weed through an entire season and then repeat it for several
cycles into the future. This requires weed information such as
the percent emergence from the seedbank, the percent surviving
the herbicide, the seed production and biomass of the survivors,
the effect of weed density on survival, seed production, and
biomass, and the survival of seed returning to or already in the
seedbank. These factors vary by weed, tillage, herbicide, and
crop. Therefore, we focused specifically on modeling giant
foxtail, controlled with Accent and Poast Plus, and velvetleaf,
controlled with Banvel and Pursuit, in corn and drilled
soybeans. The herbicide rates ranged from 0 to the full labeled
rate (the proportion of the full rate is described below as 1x,
1/2x, etc.). Other details about the study and single year
results were reported at the conference last year (Bussan and
Boerboom 1997). Information on factors such as emergence, seed
decay, and predation were obtained from the literature.
All of the following
results are based on the scenario of a corn-drilled soybean
rotation under reduced tillage and that the same rate of
herbicide is used in both crops. The postemergence herbicides
were applied at an early timing as recommended when using
reduced herbicide rates. Because it is also recommended that
mechanical weed control should be used when herbicide rates are
reduced, we also included mechanical treatments in some
predictions. The mechanical treatments were cultivation in corn
and rotary hoeing in the drilled soybean. Each model simulation
was run for a 20-year time period and all of the results are
based on the total of 20 separate runs. The results were based
on 20 runs because each run results in a unique prediction. This
variation exists because the model can randomly use results from
the range of weed survival, seed production, etc. that we
measured in the field experiments. For example, we had more
giant foxtail survivors and seed production in one year than the
other year. In addition, the model selects from a range of
values within that year. This variation makes the predictions
more realist because the model may use several poor years in a
sequence which gives a better perspective on the potential
variation or risk.
Computer
Simulation Predictions
Seedbank
populations
The model starts with
an initial seedbank level and the effect of the herbicide rate
used on the weed population can then be predicted. Without any
herbicide use, the giant foxtail seedbank quickly rises from
1,000 seeds/m 2 to a level that fluctuates between 5,000 and
10,000 seeds/m 2 (data not shown). In contrast, when
postemergence herbicides were used at 1x rates, the seedbank was
quickly depleted by 95% by 4 years (figure
1). As expected, the use of 1/2x rates alone were not as
successful as 1x rates and had much more variation. Still, the
seedbank was reduced, which suggests that while some giant
foxtail were escaping treatment and producing seed, more seed
was being lost because of seed decay and from controlled
seedlings. However, mechanical treatments are recommended with
reduced herbicide rates. When mechanical treatment was combined
with 1/2x rates, the predicted result is very similar to the
prediction with the 1x herbicide rate (figure
2). This supports the recommendation for using mechanical
treatments with reduced herbicide rates and removes much of the
risk. |
| Figure
1. Predicted giant foxtail seedbank in a
corn-drilled soybean rotation. |
|
|
|
|
| Figure
2. Predicted giant foxtail seedbank in a
corn-drilled soybean rotation. |
|
|
|
| When the velvetleaf
seedbank started at 100 seeds/m2
, the model predicted that the seedbank would increase to a
level of about 4,000 seeds/m2
in 3 years if not controlled (data not shown). Besides herbicide
rate, verticillium also significantly affected velvetleaf seed
production. Because verticillium reduced seed production, the
model predictions shown below were simulated without
verticillium to present a worst case scenario. The control
provided by the 1x herbicide rates are predicted to reduce the
velvetleaf seedbanks, but not as rapidly as the 1x rates reduced
the giant foxtail seedbanks (figures 1 and
3). This is expected because velvetleaf can remain viable in
the seedbank longer than giant foxtail. The 1/2x herbicide rates
only gradually reduced the seedbank (figure
3). However, when this reduced rate was supplemented by
mechanical treatment, the seedbank was reduced more rapidly (figure
4). Still, there was more variation in the predictions among
simulations, suggesting that using reduced rates with high
velvetleaf densities may be more risky than with giant foxtail. |
| Figure
3. Predicted velvetleaf
seedbank in a corn-drilled soybean rotation. |
|
|
|
|
| Figure
4. Predicted velvetleaf seedbanks in a
corn-drilled soybean rotation. |
|
|
|
Annualized net
returns
The computer model
also predicted the effect of any surviving giant foxtail or
velvetleaf on crop yield. Therefore, the net returns (crop
income less all weed control costs) could be determined for each
year of a simulation. To simplify comparisons, the net returns
were summarized as annualized net return (ANR), which accounts
for interest over the 20 years. For giant foxtail, the ANR for
the 1/2x herbicide rates either alone or with mechanical
treatments was higher than the 1x herbicide rates because of the
lower herbicide cost (figure 5). The
additional cost for the mechanical treatments with the 1/2x
rates was offset by slightly higher crop yields so that the ANR
were nearly identical for 1/2x rates with or without
supplemental mechanical treatments. |
| Figure
5. Predicted annualized net return of selected
giant foxtail treatments. |
|
|
|
| The 1/4x rates
combined with mechanical treatments also had high ANR because
Accent and Poast Plus were quite effective when applied at early
growth stages as in our field trials. For giant foxtail, the ANR
for the 1/2x rates were also greater than for the 1x rates
regardless of the initial foxtail density. In contrast, the
results for velvetleaf varied depending on the starting density.
At lower densities, the ANR for the 1/2x rates were greater than
the ANR for the 1x rates (figure 6). At
high densities, this advantage is eliminated. |
| Figure
6. Predicted annualized net return of selected
velvetleaf treatments. |
|
|
|
Treatment
thresholds.
The simulation model
also allowed thresholds to be studied. After several years of
good control, there would be years when densities would be below
threshold levels. For example, if giant foxtail was not sprayed
at densities below 1 weed/m2
, the foxtail density would call for treatment in most years.
However, when not treated, the seed produced would increase the
seedbank and require herbicide treatment in subsequent years (figure
7). This shows the long-term consequences that need to be
considered for thresholds. The threshold that maximizes ANR over
time is called the economic optimum threshold (EOT). For giant
foxtail, the predicted EOT is very low and in the range of 0.1
to 0.2 seedlings/m2
when mechanical treatments are never used and about 0.7
seedlings/m2
when mechanical treatments are always used (figure
8). The EOT was about 0.5 velvetleaf/m2
and was similar whether or not mechanical treatments were used (figure
9). For either weed, the EOT was similar for 1x and 1/2x
herbicide rates. |
Figure
7. Giant foxtail seedbank with a threshold of 1
giant foxtail
per m2
while using 1x herbicide rates. |
|
|
|
|
| Figure
8. Predicted annualized net return as giant
foxtail threshold density increases. |
|
|
|
|
| Figure
9. Predicted annualized net return as velvetleaf
threshold density increases. |
|
|
|
Conclusions
The predictions made
by this computer simulation model should provide added
confidence to those wishing to use reduced postemergence
herbicide rates when supplemented with mechanical treatments in
fields with low or moderate weed densities. The model predicts
that if some giant foxtail or velvetleaf escape control with
reduced rates, the weed populations will not explode in
subsequent years. Therefore, if management allows for proper
herbicide timing and follow-up cultivation, the risk should be
minimal. Of course, if the reduced herbicide rate application
can not be made to small weeds, the herbicide rate can be
increased and not increase the risk for the farmer. However, the
model also realistically predicts that reduced rates are not
economically justified for higher velvetleaf densities. This
confirms previous recommendations that reduced herbicide rates
should not be used on high weed densities or difficult to
control weeds.
Literature
Cited
Bussan, A.J. and C.M. Boerboom. 1997. Effect of reduced herbicide rates on velvetleaf
and giant foxtail survival and seed production. Proc. 1997 Wisconsin
Fert., Aglime and Pest
Management Conf. pp. 84-87.
Doll, J., R. Doersch,
R. Proost, P. Kivlin. 1992. Reduced herbicide rates: aspects to
consider. Univ. Wisconsin Extension
Bull. A3563.
Defelice, M. and A. Kendig. 1994. Using reduced herbicide rates for weed control in
soybeans. Univ. Missouri Extension
Bull. MP686.
Proost, R.T., P.T. Kivlin, K.B. Shelley, and K.A
Talarczyk. 1996. A summary of six
years of on-farm demonstrations
focusing on reduced preemergence herbicide rates. Proc. 1996
Wisconsin Fert., Aglime and Pest
Management Conf. pp. 179-182.
|
|
