The Promise of GMOs: Herbicides

This is a part of the series The Promise of GMOs. Do GMOs live up to the promises of the biotech industry? In the case of decreasing herbicide use and improving the environmental impact, there is data to back up industry claims.

Reducing herbicide use

Data from the EPA.
Data from the EPA.

BIO’s claims here are that “Biotech is helping to feed the world by: Lowering volumes of agricultural chemicals required by crops-limiting the run-off of these products into the environment [and] Using biotech crops that need fewer applications of pesticides.”
Verdict: Promise met (for herbicides specifically, insecticides will be covered in another post) .
Looking at the most up to date EPA data on agricultural pesticide use, we see that use of herbicides, insecticides, fungicides, and other conventional pesticides (such as miticides) all have a downward trend between 1998 and 2007. The slopes of the linear trendlines are in the legend of the chart to the right. Strangely, other pesticidal substances (such as sulfur, petroleum, and other chemicals used as pesticides) has a strong upward trend, which caused the total to increase. If you were to look at just the total, you wouldn’t get the overall picture.
This dataset from the EPA includes “one from the USDA and others from private pesticide marketing research companies”. I wish there was more recent data, but this is what we have to work with. If you extrapolate into the future without adequate data you might not get the right picture. Brookes and Barfoot have published results from slightly more recent datasets (such as a herbicide use decrease of 75 million lbs from 1996 to 2010) but since part of the dataset is proprietary and I couldn’t look at it, I didn’t feel comfortable using it here.
So we know that herbicide use seems to be decreasing at the same time that GM crops increased. But is that correlation or causation? And is total herbicides even the right way to look at this?
Let’s turn again to the USDA and EPA researchers in their Conservation Tillage, Herbicide Use, and Genetically Engineered Crops in the United States: The Case of Soybeans (paraphrased):

Most researchers measure herbicide use by the total pounds of active ingredients applied. While this is informative, these of these studies can be biased by unobservable conditions prevailing in the year of the study. Also, when different types of active ingredients are grouped together, it covers up the fact that their characteristics (potency, toxicity, etc.) vary widely.

Therefore, any studies that look at volume or mass of herbicides is not telling the whole story. So the small decrease in herbicide use shown above may not be relevant if there was a switch to much more toxic herbicides. They continue (paraphrased):

Some studies suggest that herbicide use on HT soybeans may be slightly higher than herbicide use on conventionally grown soybeans in the United States. However, glyphosate (the herbicide used on most HT crops) is less toxic to humans and not as likely to persist in the environment as the herbicides it replaces. Consequently, increased herbicide use on HT soybeans is not necessarily indicative of worse environmental outcomes.

Data from Fernandez-Cornejo et al, 2012.
Data from Fernandez-Cornejo et al, 2012.

The table to the right summarizes herbicide use on soybeans in 1996, before HT crops were widely adopted, and in 2006, after HT crops were widely adopted. The data is again from Conservation Tillage, Herbicide Use, and Genetically Engineered Crops in the United States: The Case of Soybeans.
For the column headings, Use = % of this pesticide compared to the total of all pesticides; Rate = pounds of active ingredient applied per acre of soybeans; Toxicity = chronic toxicity score (higher is safer); and Half life = number of days it takes for 1/2 of the herbicide active ingredient to degrade in the soil.
I added the EIQ (Environmental Impact Quotient, lower is safer) for each herbicide because I’m not familiar with the chronic toxicity score. EIQ is a numerical representation of the risks a pesticide poses for the environment, consumers, and farm workers. The EIQ for 2,4-D is an average of the EIQs for 5 different formulations of the herbicide.
In 1996, soybean farmers were using a variety of pesticides, and at least 68% of herbicides used had toxicity scores that were worse than glyphosate. In 2006, the majority of farmers had switched to glyphosate, the least toxic of the bunch, making up 85% of herbicide use.
The total amount of herbicides is trending down and the total environmental impact due to herbicides has decreased. While we can’t be sure if HT crops were the cause of the total decrease in herbicide use, we can be reasonably certain the change in herbicide types was due to HT crops. So, the verdict for BIO’s claim is promise met. However BIO’s use of “volumes” in their claim specifically is a poor measure for pesticides, it makes much more sense to look at the types of herbicides.


  1. The claim is about pesticides, so wouldn’t it be good go also look at bt crops? You likely see a much stronger effect there (if there is appropriate data). I believe that was the case for bt cotton.

  2. Interesting. Wouldn’t a lot of the substances used in organic agriculture fall under “other pesticidal substances”? (Afaik at least petroleum can be used on organics.) This is all correlation, but if true, at the same time that GMO and organic acreages increased, the pesticides used in conventional/GMO agriculture went down, while the substances used in organic agriculture went up? Anybody any insights on this?

  3. I’m curious about this too but didn’t have time to track down what EPA terms “other” and how much is allowed on organic farms. Organic acres are such a small percent of farm land so I doubt it’d be able to sway the total so much but it’d still be interesting to see.

  4. Fewer pounds of herbicide in the environment, and less environmental risk from the herbicides out there, are both benefits. On the other hand, with glyphosate accounting for 85% of herbicide use these days, I’d be concerned about increased selection pressure for resistant weeds developing.

  5. Benbrook, among others (including the US Gov.), have continually, and incorrectly, insisted on perpetuating this bogus concept of lbs or kgs of pesticides used. It should be blatantly obvious that a) not all compounds weigh the same, and b) the weight of a compound has no relationship to its potential toxicity. Anastasia is kind above when she says this is “not telling the whole story”. I’d say its not telling anything at all. It is completely meaningless. The EIQ is one means of attempting to equivocate differing compounds (and one that has apparently been rejected by Benbrook). My understanding is Mr Benbrook has been developing an alternative measure for some time, but we yet to see it. Until that time, I’d have a hard time believing anything he produces on the subject.

  6. Anastasia,
    I would add that the HT traits created by GE were made that way out of convenience and that HT crops would have still existed without GE tech through “traditional” breeding methods. Such HT crops, in fact, do exist. The multiple line of BASF Clearfield crops, for example, or glyphosate resistant wheat developed by Colorado State University (was never released). Blaming herbicide resistant weeds or increased herbicide use solely on GE is a red herring.

  7. I would agree that the total herbicide usage in pounds doesn’t mean anything really. What the data really shows is that glyphosate use went up while use of most other popular herbicides have gone down- something already known. Thanks for including the EIQ. I’ve never seen that before.
    What I would like to see is specific herbicide (glyphosate) use in pound per acre for a crop that is not GE compared to it’s herbicide resistant GE equivalent. Then we can get an idea of how GE crops affect herbicide use under controlled conditions. I’m sure a simple google search would pull up studies on this and I’d bet you guys have already written about it on this website. Does anyone know how glyphosate resistant crops compare in glyphosate use to the older (pre-GE) spraying regimes?
    Anti GMO proponents often claim that farmers spray more herbicide on GE crops (just because they can), so it would be nice to have some numbers. And ideally the numbers would consider what farmers are currently spraying to get the most yield or profit per acre. Anyone have a link handy?

  8. One thing to note for field crops (corn, cotton and soybean) is that insect protection is mainly used for the former two crops. Interestingly, soybean production has been faced with the influx of 3 invasive and serious pests in the past 10 years: soybean aphid, brown marmorated stink bug and the kudzu bug. There are no transgenic options for control of these pests, but insecticide use in soybean has likely increased dramatically. Indeed, insecticide use in soybean for soybean aphid control alone has led to a 130-fold increase in the past 10 years (see: My educated guess is that a larger decrease would be seen in corn and cotton because that’s where transgenic insect control is used.

  9. Benbrook also enjoys axis manipulation to make his point.
    I recall at one point he had a hectare of crop vs lbs of active ingredient *type* graph which showed that while acreage didn’t change the quantity of active ingredient went through the roof.
    When the Y axis was separated an utterly different picture was seen – I haven’t been able to take Benbrook at all seriously since I deconvoluted that particular figure. (I believe the work was done here on biofortified, but I’m far too lazy to recall where or indeed when)

  10. I always thought that was a strange argument. Every ounce of pesticide, fertilizer, water, and other inputs that a farmer uses cuts into profit. Farmers won’t use as much as they can but as little as they can. Unfortunately, it’s tough to get raw data for herbicide use, which is why I didn’t include specific data in this post. However, there are sources out there. Brookes and Barfoot seem to be the go-to guys for this:

  11. That’s a great point. If we didn’t have GE HT traits, we’d have more non-GE HT traits. I’ve argued before that if Benbrook et al want to complain about HT then they should complain about HT, not about GE.

  12. Weeds will become resistant no matter what herbicide you’re using so resistance isn’t a GE problem, and there’s no such thing as “superweeds” but it is concerning when a relatively safe, cheap herbicide loses its efficacy. I think that more diverse rotations needs to be considered more broadly across the US and elsewhere, but farmers have to make financial decisions so if they are going to profit most with corn/soy then that’s what they’ll plant. We need more creative ag policies to incentivize growing other crops. But that’s probably not going to happen. In the meantime, there are a few other herbicide tolerant traits coming out and if those are added into rotations it’ll reduce weed resistance without changing the farmer’s bottom line.

  13. I’d really like to take the graph above and divide by cropland to see if the amount of herbicide per acre changed in the US. Unfortunately, I can’t find a dataset that really works here. The USDA has this nice Trends in Major Land Uses but the types of land uses aren’t divided conveniently for this purpose. The graph above is all pesticide used for the purposes of agriculture. But the land use data subdivides agricultural land into cropland, pasture, other farmland, and special (which also includes farmland) and I’m not really comfortable just picking one or just adding them together. If any one has any ideas, they would be appreciated.

  14. What do you think about replacing a herbicide that literally weighs one ounce per acre with a better, more environmentally friendly herbicide that weighs 26 ounces per acre?
    This year I used a few gallons of imazethapyr and over a thousand gallons of glyphosate. I will have more collateral damage with just the few gallons of imazethapyr than with all of the glyphosate.
    Glyphosate replacing imazethapyr is acknowledged within the study you cited.
    As an example: I raise potatoes. Did you know that I must wait three seasons to raise a crop of potatoes when I use imazethapyr because of the potential tuber damage? However, I can plant any crop immediately after spraying glyphosate.
    Anastasia is exactly correct with this article, it isn’t the “weight” of what is used, it is what that product does and the other factors that are affected.

  15. One glaring area of distinction is that the data in this article is actual measured data from the EPA. A lot of the data in the Benbrook piece (your link) is using a data model to project what they think the application amounts are. Toward the end of that piece there is quite a bit written about estimating herbicide use on certain types of acres. In addition, Benbrook seems to be making the assumption that all glyphosate sold is attributed to GMO acres when in fact, glyphosate is used on many types of non-gmo crop acres.
    It seems of all the papers written on environmental impacts of GMO crops, Benbrook is the only one claiming and large increase.

Comments are closed.