Features Agronomy
Using agronomy to control weeds: what works, what doesn’t

I know that the idea of integrated weed management has overwhelmed producers with complexities and checklists of all the possible control measures. Farmers want to know what works and what doesn’t.

The first practice is harvest seed management. I did my PhD work in Manitoba 23 years ago and looked at chaff collection. We were looking at it as a way of weed control for Group 1 herbicide-resistant weeds. The first thing we noticed with wild oats – and others have found this, too – is that as the harvest timing goes later, most of the seeds fall off. But in my work in Manitoba with wheat, when you swath the crop about 50 per cent of the wild oat seeds would still be in the swath. If you straight cut, most of them would’ve fallen off.

This characteristic of seed drop prior to harvest should write off the idea of harvest weed seed management, but I think a key idea here is wild oats have very little seed dispersal. They don’t have a seed dispersal mechanism other than something mechanical like a combine spreading the chaff and seeds. So, if the wild oat seeds fall down in the exact same spot every year during straight combining that may not be a big problem. My approach is that we have to start looking at herbicide resistance as a proportional argument and not as a yes/no argument – “does this field have it or not?” We have to start looking at it as a spatial argument – i.e. if you have one small patch of wild oat resistance in your field, that’s not a big problem. If we can isolate it and keep it there, that’s a good thing. So, I think having weed seeds that fall off early is a good thing.

One thing that we did realize was how harvest weed seed management – in the case, we were using chaff collection – influences the dispersal in the field. What we found was that chaff collection reduced the spread of wild oats by about a factor of ten. Without chaff collection, the wild oat seeds are moved down the field with the combine, which results in the wild oat patch growing in size.

As part of my PhD thesis I went to Montana in the late 1990s and learned computer modelling from Bruce Maxwell. I wanted to model the dispersal of wild oats in the field, and to test harvest weed seed management and timing of harvest. My idea was that maybe you’re better off letting the weed seeds fall off rather than going through the combine and being spread around the field.

We simulated wild oat combine dispersal with and without chaff collection and compared swathing (early harvest) and straight cut (late harvest). We started off with five little patches of wild oats that were two metres by two metres (six feet by six feet).

After five years of the model running, the patches with chaff collection had spread out somewhat. With a late harvest and chaff collection the patches hardly got bigger at all. When we harvested without chaff collection, the wild oat dispersal when swathing was much worse. With a late harvest, most seeds fell off and there was much less dispersal.

I think with straight cutting in Western Canada, most of our wild oat seeds are falling off and the patches aren’t spreading that much. I would present this as a hypothesis: that Group 1 resistant wild oats has gotten worse, but it hasn’t affected how we manage weeds that much; that in many cases, it’s still isolated patches that haven’t been spread around. I can’t prove this, but I would say that, definitely, allowing those seeds to fall off onto the ground – because there’s probably a big seed bank there already – it’s probably a good strategy, but that should be tested by somebody.

Managing imidazolinone resistant weeds in pulses
In lentils, Pursuit (imazethapyr) herbicide was used off-label for many years to control wild mustard.  When the imi-tolerant lentils were released, Odyssey, (imazamox + imazethapyr) was relied on heavily without any tank mixes. Not surprising within a few years Group 2 resistant weeds were selected. I would look at fields in lentil-growing areas and think, “That’s terrible. I can grow organic lentils that are less weedy than that.”

First thing to look at is seeding rate. If you seed your lentils at what would be considered a normal rate, you get half lentils and half weed biomass. If you increase the seeding rate, you decrease the weed biomass. You’re putting more plants in there to take up more space, essentially increasing the competition. Increased seeding rate increases the yield, as well.

We’ve looked at this in the context of an integrated weed management system, where colleagues Chris Willenborg and Eric Johnson at the University of Saskatchewan looked at using fluthiacet-methyl – which is Cadet in the U.S. – as a possible herbicide on lentils. The lentil-breeding program is still looking at this, trying to find some varieties that are more tolerant to it. But tolerance is a big issue in this crop. As the rates get higher, you get a lot of leaf bleaching.

We used the concept of a dose response curve. As we increased the dose of the herbicide on a log scale, it decreases mustard biomass. But what we also found, as we increased the seeding rate from low to regular to double the recommended rate, we were able to shift the dose response curve down. This means the herbicide could work a lot better with a higher seeding rate than a lower seeding rate. Colleen Redlick, who did this work, likes to refer to this as a tank mix, but instead of adding another herbicide to your spray tank, you’re adding more seed to your air seeder tank.

This came through in the lentil seed yield as well. If we had too low of a dose of the fluthiacet-methyl, the yield was lower. However, when we doubled our seeding rate, we had this nice, flat response. We weren’t relying on the herbicide as much and could have a less toxic dose. So, increasing the lentil seed rate increased our herbicide efficacy. Essentially, we’re not relying on the herbicide as much. About 260 seeds per metre squared was where it resulted in greater efficacy.

Mechanical weed control in pulses
We have used mechanical weed control in field pea using a small flex-tine weed harrow at the third node stage. There’s a sweet spot in the amount of crop burial at around 75 per cent. The one key problem is that most of the research is on fields with almost has no residue. Working on a no-till field with narrow spacings will just plug up the harrow.

Eric Johnson started working with a min-till rotary hoe about 10 years ago. The great thing about a rotary hoe is that it can work in fields with a lot of residue. Eric found a lentil crop to be quite tolerant to the rotary hoe. They’re available in wide widths, and are starting to be used in some areas.

Research is also looking at inter-row cultivation. Some producers are using inter-row tillage on row widths as narrow as seven or eight inches. The tillage units have camera-guided systems which steer the cultivator between the rows. It can work really nicely, but of course you can’t get the weeds in the row.

Katherine Stanley did some work on the tolerance of field pea to inter-row cultivation.  We found that within the critical period of weed control in peas, between five and 10-node stage, you weren’t damaging the crop at all. Going later, started to damage the potential yield. We were encouraged with the results.

The question I always get is, “What machine should I buy? What works? What doesn’t work?” Some of the decisions are going to come down to timing. The rotary hoe only works when the weeds are just emerging, but it has good tolerance on the crops. The harrow will work with bigger weeds, but it’s hard on the crops, so you tend to use it later. Inter-row tillage doesn’t control the weeds in the crop row, and you actually need the crop to be up so the machine can steer between the rows.

Research was conducted in lentil and field pea looking at all three tillage options and in combination with each other. What we started to see before we analyzed the data was that the rotary hoe was working well. We were getting very nice visual weed control. The rotary hoe followed by inter-row tillage was also giving nice visual weed control.  One of our best treatments was often our rotary hoe followed by inter-row tillage. We were getting about 76 per cent reduction in weed biomass without an herbicide. This was under organic conditions with severe weed competition. But when we combined all three machines, we started to get into issues of crop tolerance – we were beating up the lentils a bit too much from all the mechanical control passes.

We had a seeding rate effect as well – increasing our seeding rate, reduced our weed biomass by 16 per cent. In terms of yield response, increasing our seeding rate gave us a 30 per cent yield boost. Our best weed control, rotary hoe followed by inter-row tillage, gave us about a 55 per cent increase in seed yield, which was getting close to that of the hand weed-control. Overall, the best combinations of increased seeding rate and tillage increased seed yield by about 70 per cent and reduced weed biomass by about 80 per cent.

These tillage treatments are effective on annual weeds. Of course, they’re ineffective on most perennial weeds. The question is whether they have a fit on conventional farms. I don’t know. I’m not a conventional farmer. I’m just throwing it out there. “Are we desperate enough yet” is basically what it comes down to, right?”

Other agronomic techniques to manage weeds
As an extension of harvest weed seed management, there are other ways to reduce weed seeds returning to the field. One way is to target weeds that mature before the crop with weed wiping and weed clipping. There are many weeds that grow through shorter crops like lentils or flax. Most of their flowering bodies and seed production will take place above the crop canopy. Good candidates are wild oat, wild mustard and Canada thistle. Lower growing weeds like wild buckwheat and cleavers would not be good candidates for weed wiping or weed clipping. Crop yield is still going to take a hit, but the idea is to prevent the weed seeds from entering the seed bank and to reduce the problems the following year.

This has been done for years. Farmers have jury-rigged equipment to clip weeds above the canopy. One example is a farmer who modified a swather by taking off the canvas and putting metal over the support frame to allow the swather to clip weeds and let them fall to the ground.

Eric Johnson looked at hand weed-clipping above the crop canopy for wild oat control. He found that either late clipping or clipping three times was as good as the herbicidal weed control in how many wild oats came up the following year.

Some commercial machines have been developed. Out of England, there’s a CTM weed surfer that has rotors that cut off weeds above the crop. There’s also the CombCut machine, which has knife blades and a reel that spins quickly. It allows grasses to go through, or in broadleaf crops, you can clip the weeds above the canopy. Right here in Saskatchewan, Borgault has started building a weed clipper. So, clearly, the industry perceives that there is a need for this technology out there.

We’ve looked at clipping in lentil with the CombCut in an experiment we’re doing with Breanne Tidemann in Alberta. We’ve done two years in Saskatoon and one year in Lacombe [Alta.]. What we’ve seen is that when we do multiple clippings at Saskatoon, we had really good curtailment of the mustard seed production. The results weren’t quite as good in Lacombe but growing conditions in Lacombe are not great for lentils.

If we just went early, it didn’t work very well, which makes sense since not all of the wild mustard would have emerged. We found that one clipping was maybe enough, but that two was probably better. Use the weed stage as a timing guide for clipping. Clip the weeds when flowering is fully extended above the crop but before seed development.  Repeat if necessary.

In cereals, the idea is that you can go beneath the crop canopy because the cereal leaves can slide through the knifes before they’ve elongated while the more rigid stems of a dicot will be cut by that. We did have some success in wheat. We were reducing wild mustard weed seed production by about one-half. The trouble is, though, we were also clipping off a lot of cereal leaves and heads by going in there because our mustard never really got that tall.

Weed wiping with a non-selective herbicide was used in the ’70s before we had herbicide tolerant crops. There are different commercial machines that have been developed to wipe the weeds above the crop canopy. We are starting to use a Garford Weedfoil, which is a better design than a lot of the older ones. It has a wider felt area that is pressure-saturated with herbicide. In our research on wild mustard, glyphosate worked best followed by dicamba. We had some problems with dicamba because it worked well for weed control the first year and didn’t work that well in the second year. We did have dicamba injury with weed wiping with both the normal and the low-volatile formulations. The vapour damaged the lentils a lot, so we threw that treatment out.

In terms of the timing and the seed yield, earlier, but not too early, was often the best. Under the best possible situations, it could work well, but it didn’t always work that well. Weed wiping also was able to reduce the next year’s repopulation that emerged by about one-half, which was much better than we thought it would be because we thought there was enough in the seed bank that reducing the seedbank would take longer.

We are going to scale the research up and do further work with 2,4-D and glyphosate for weed wiping, but given how good the results have been from the clipping, we are going to combine the two projects.