Want to reduce herbicide resistance? Spray more herbicides!

Herbicide resistant kochia

Herbicide resistant (right) and susceptible (left) kochia after being sprayed.

A while ago, I wrote a post summarizing the pros and cons of using a regulatory framework to slow the evolution of herbicide resistant weeds. As a part of that post, I suggested there exists a “resistance management paradox.” In a nutshell, to reduce the problem of herbicide resistant weeds, one important strategy may be to actually use more herbicide. From my previous post:

“The only way to assuredly prevent herbicide resistance from evolving is never to apply the herbicide; but if we are going to use a herbicide, the best way to minimize herbicide resistant weed evolution may be to increase herbicide use. Research from Hugh Beckie and Xavier Reboud has shown that mixtures of two herbicides are far more effective at slowing the evolution of herbicide resistant weeds compared to an annual rotation of herbicides.”

When Beckie & Reboud used a herbicide by itself just one time over a 4 year period, they observed a nearly 8-fold increase in resistant weed seed production. However, if they applied the same herbicide every year, but mixed it with another effective herbicide, resistant weed prevalence remained statistically the same as if they had never used the herbicide at all.

Resistant seed produced per square meter averaged over 4 years Percent of weed seed in soil that was resistant after 4 years
Never used ALS herbicide 155 4
Used ALS herbicide once over 4 years 1210 29
Used ALS herbicide every year, mixed with 2nd herbicide 178 8
Data from Beckie & Reboud (2009)

A new study published earlier this year in the journal Pest Management Science lends additional evidence for the resistance management paradox. The goal of this new study from Illinois was to “identify risk factors associated with resistance.” Their work is pretty unique in the herbicide resistance literature because they used an epidemiological approach. They used a case-control design common in the medical literature to determine whether exposure to various factors was associated with development of herbicide resistance. For this study, the researchers used glyphosate-resistant common waterhemp as their model species, and obtained field management data for a few hundred crop fields over a 300 square mile area over an 8 year period. They evaluated 66 different environmental and landscape conditions and various management practices that were thought to potentially influence resistance development in some way.

Unsurprisingly, their results suggested that management practices, especially herbicide choices, were the most predictive factors in the development of glyphosate-resistance. Glyphosate-resistance was most prevalent in fields where glyphosate was used frequently, and where few herbicide modes of action were used. But at least one of the results has surprised some weed scientists. From the study (emphasis mine):

“Contrary to prevailing wisdom, both binary and proportional models indicated that herbicide turnover increased the frequency of resistance. Logistic regression models supported this result, with positive or non-significant effects of several metrics of herbicide MOA turnover…”

That’s an overly science-y way of saying that rotating herbicides from year to year increased the likelihood of finding herbicide resistant weeds. For many years, rotating herbicides annually has been a primary recommendation from weed scientists; if you use Roundup Ready with glyphosate this year, you should use something else next year. Rotating herbicides from year to year will decrease the frequency of any single herbicide, thereby reducing selection for those resistant weeds. But herbicide rotation, on its own, isn’t enough and we’ve known this for a while. From a review article written by Jason Norsworthy and others regarding best management practices (page 40, emphasis mine):

“Annual herbicide rotations are useful but not sufficient because they subject a weed population to a single [herbicide mode of action] at a time…”

I think there are a few important takeaways from the Illinois research. First, rotation of herbicide modes of action simply isn’t enough to prevent (or, perhaps, even significantly delay) the appearance of herbicide resistance. Weed scientists and extension specialists need to be more clear about this in our education efforts. It seems that farmers in the Illinois study may have had a false sense of security from rotating herbicides, since they were actually more likely to develop herbicide resistant weeds compared to farmers who weren’t rotating. Rotating herbicides from year to year is only effective if it is also accompanied by other effective weed control measures.

This research supports the potential value of stacked herbicide technologies like Enlist and Roundup Extend as tools for resistant weed management, since they will increase the likelihood that farmers will apply mixtures of effective herbicide modes of action. However, these technologies will be most beneficial only where resistance to glyphosate has not yet evolved. Where glyphosate-resistance is prevalent we will simply be selecting for weeds that are resistant to both herbicide modes of action in Enlist (2,4-D) or Extend (dicamba) systems. This is the “pesticide treadmill” that critics of the technology worry about, which is a valid, and serious, concern. For herbicide mixtures to be most effective, they must be initiated before herbicide resistance is a major problem.

In my previous post on the resistance paradox, I wrote:

“If we consider herbicide-resistant weeds to be an environmental problem that must be addressed, one strategy we must seriously contemplate is increasing herbicide use.”

I still believe that to be true. Unless we plan on completely removing herbicides from modern agriculture (a laudable long-term goal, but unrealistic in the short-term), we need to try and protect the utility of the herbicides we have. Since herbicide rotations are ineffective on their own, the most practical way to maintain the value of current herbicides is to mix them with other effective herbicides. This will almost certainly increase the total amount of herbicides we use.

But “spray more herbicides” is not something anyone wants to hear, including me. (Well, maybe some herbicide manufacturers.) Society wants us to decrease pesticide use. Farmers certainly don’t like spending money on more pesticides. So at some point, we must stop looking to herbicides as the solution to a problem created by herbicides. You know what they say about someone who keeps doing the same thing over and over again while expecting a different result. We need to place more emphasis on non-chemical weed management practices. This means taking some pages from agroecology and the organic weed management playbook. We need to make more regular use of cultural weed control practices like smother crops and stale seedbeds. We need to invest in more mechanical weed management technologies like robotics and seed destructors. Perhaps we’ll even have to reduce some of the benefits that have accrued from adoption of no-till production. I certainly don’t have a magical solution to the herbicide resistance problem. But I can guarantee that a long-term solution will not be simple, and it most certainly won’t be sold in a jug.

UPDATE: if you’re interested in how easy it is to implement the multiple herbicide strategy on an actual farm, see this post.


Evans, J.A., P.J. Tranel, A.G. Hager, B. Schutte, C. Wu, L.A. Chatham, and A.S. Davis. (2015) Managing the evolution of herbicide resistance. Pest Management Science. DOI: 10.1002/ps.4009



  1. If dormancy did not exist, the herbicide rotation model might be more effective. But when you have dormant seeds germinating this year, that were not exposed to herbicide A last year, and you treat them with herbicide B this year, you are essentially using one herbicide for each crop of weeds, not the entire population. Some of each of those crops is likely to have a few resistant plants, which will produce seeds that go dormant and may not germinate for several years. This is a fatal flaw in the old herbicide rotation practice.

    1. Are you sure you understand how the process of natural selection works BEdge? Dormant seed already has all of its traits (including herbicide resistance) embedded in its DNA.

      1. What BEdge is saying, is that because of dormancy, you could theoretically get a lineage of plants that is *only* exposed every other year, thereby never getting the alternative herbicide in the rotation. Which is certainly plausible, and something I’ve often wondered about. In theory, weed species with longer seed dormancy are thought to evolve resistance more slowly, because there is a larger buffer of susceptible seed in the population. I don’t know of any work that has actually tested this experimentally, though. Something I’ve considered writing a grant to study some day…

        1. Yes, Andrew, that is what I meant. And really, it would be random, with some seeds from any given population germinating across multiple seasons. Given that there is already resistance in some weeds to 2,4-D and glyphosate, we are already starting from a weak position with the Enlist Duo system.

        2. Testing this idea experimentally would certainly be interesting, albeit probably quite difficult. However, couldn’t one use demographic modeling approaches to estimate the outcomes of various scenarios (sensu Maxwell et al. Weed Tech 4:2-13)? Perhaps demographic models, when combined experimental approaches to determine life history parameters, could provide a means to address this question?

          1. Hi Greta, Pretty much everything we know about this topic come from modeling approaches, so I think refining models is the likely way forward. But I also think there’s still a lot to be done on the experimental side of things to make sure the models we’re using give sane results. I think integrating models with some better experimental data is probably the best way to get at this question. Over the last several years, we’ve been working on a method that might help us do the experimental side of things. We described it at WSWS a couple years ago (see abstract 069 here: http://www.wsweedscience.org//wp-content/uploads/proceedings-archive/2011.pdf). So far, I’ve been unsuccessful getting funding to develop this idea further. I’m hoping to blog about that project (and seedbanks in general) some time this year.

  2. Oh, will it be cocktails, then?

    As I was reading through, it reminded me a lot of what I’m seeing in the evolution of pharmaceutical strategies. AIDS was the first well-known case of cocktails of drugs working, but increasingly this is going to be the strategy for cancer, too. As we get better at pinpointing the problems, we may have 3 drugs that can chip away at part of the solution, but in combo they all work better.

    Now, this doesn’t mean we shouldn’t also work to prevent cancer. But the fact is, there’s still going to be some. Same for other conditions as well, I just think cancer will be the first one to get it really precise.

    But I can imagine a time when some of those new antisense weed strategies + chemistry + physical solutions end up working out.

    1. Anti-malaria drugs, with artemisinin being associated with another active ingredient to prevent the loss of the first is a better example.

  3. I have wondered whether there has been any thought given to developing a weed resistance management approach that mimics disease response in livestock. In livestock, there are certain diseases that are reportable, and when they appear, livestock health specialists and regulators take steps to confine the disease to the reporting premises, to check neighboring herds, and to attempt to trace the disease, and then to do a cleanup on affected premises (culling animals, disinfecting premises, etc.) Both state and federal government may provide reimbursement or cost share assistance or indemnity to offset the costs to the producer from loss of income from culled or destroyed animals.

    Would something like that be feasible with weed resistance? When a landowner first encounters a weed resistance, a weed resistance specialists and regulators would visit the property to confirm and assess the resistance. Neighboring properties could also be assessed to see if resistance has spread to or from such properties. ONce the pocket of resistance is geographically identified, an intensive eradication plan would be executed, and such plan could include alternative herbicide strategies, manual, mechanical or other means, crop rotation prescriptions, etc. The goal would be to isolate and quickly eradicate isolated pockets when they first occur.

    Perhaps such resistance response protocols could be made a part of the approval and registration requirements for all herbicides and perhaps other pesticides. Potentially, a fee could be added to product registrations that would go into a segregated fund that would be expended in cooperation with state and local weed control entities to support eradication projects as they occur. States could fund their portion with a surcharge on pesticide sales paid by end users. Local resource entities might also contribute. A landowner with a resistant population would be under obligation, similar to discovery of a livestock disease, to follow a plan of eradication, but the funding sources identified could provide a cost share assistance to mitigate the margincal costs to the farmer to employ alternatives. Say the eradication plan for a farm called for hiring manual labor for two years to control weeds, or to plant other crops that are not as lucrative. The fund would be used to mitigate those costs. Once the resistance had been eliminated, the farmer could resume normal weed control choices.

    Maybe it is already too late for some resistance problems if they are already widespread. Perhaps it is something that should have been contemplated clear back in the 1950s or so. But if I am correct, Dow is committing to something along this line with its Enlist products and program.

  4. Interesting, after attending various resistant weed management workshops and presentations we changed things up on our farm this season and in some cases it included more herbicide especially pre-emergent, but also changed up several other factors, (even a bit of tillage) as well, too lengthy of discussion for this page. Bottom line I was very happy overall with our results.
    I guess my only comment on the cover/smother crops, stale seedbed/shallow tillage is loss/use of soil moisture so precious here on the High Plains of NE/WY/CO/KS/TX.

Comments are closed