Herbicide resistance can become a problem due to two distinct mechanisms. The first involves independent selection of a resistant biotype from the weed community already present in the field. The other involves movement of the resistant trait (gene flow) from a field with an existing resistance problem to fields without resistance. This spread can occur due to either pollen or seed movement.
Independent selection of a resistant weed biotype
Herbicide resistant biotypes are present at low frequencies within the weed community prior to the discovery and introduction of a herbicide. The initial frequency of resistance is the percentage of the weed population possessing the resistant trait at the time of herbicide introduction. Repeated use of a herbicide results in resistant biotypes becoming more prevalent, and eventually they may become the dominant biotype and thus the weed is no longer effectively controlled by the herbicide. Evolution of resistance within a field is a direct result of the weed management program used within the field and the initial frequency.
The initial frequency of herbicide resistance varies among herbicide classes. With ALS inhibitors, resistance is present at a relatively high initial frequency and it was likely that resistant biotypes of certain weed species were present in most Iowa fields at the time of introduction of these herbicides (mid-1980s). It is believed the spread of ALS resistant waterhemp across Iowa was largely due to selection of resistant biotypes from native infestations within individual fields.
Glyphosate resistance in plants is a rare trait, thus it is likely that GR weeds were not present in most fields when we began to rely on glyphosate for weed control. Even though glyphosate has been used in a way favoring the selection of GR in the majority of fields across Iowa, GR weeds will only appear due to independent selection in those fields 'unlucky' enough to have one of the rare GR biotypes.
Gene flow, movement of resistant trait from one field to another
Now that GR biotypes have evolved in isolated fields across the state, the risk of resistance appearing in 'clean' fields is probably as great due to movement from fields with existing GR problems as it is from independent selection. Gene flow may occur either from pollen or seed movement. The likelihood of these two mechanisms of gene flow varies with characteristics of the resistance trait, biology of the weed and the weed management systems used in adjacent fields.
Although the inheritance of GR has not been characterized in all GR biotypes, where it has been identified, heritability is an incompletely dominant trait. Thus, cross-pollination between a resistant (R) and susceptible (S) weed biotype will result in increased resistance to glyphosate in the hybrid plant (S x R). Both waterhemp and giant ragweed produce large quantities of pollen that move long distances by wind currents thus pollen movement is a threat to spread GR in these two species.
Researchers at the University of Illinois developed a model to help understand gene flow of GR in waterhemp via pollen movement (Liu et al. 2010). Their model predicted that GR could spread approximately three miles per year by wind dispersal of pollen, thus any field within a three mile radius of a field with GR waterhemp is at risk of resistance becoming established.
It is important to understand that GR will not move into every field within this radius due to several factors. The GR trait can only become established in fields where receptive waterhemp plants are present to intercept the pollen. Also, the greater the distance from the source field, the less likely a pollen grain is to pollinate a susceptible waterhemp plant.
Using herbicides with different modes of action thwarts GR
Finally, in order for GR to become established in the field, the weed management program being used must rely heavily on glyphosate. If the weed management program includes herbicides having different mechanisms of action than glyphosate that are effective on waterhemp, there is no advantage for the GR trait and it is unlikely that the resistant biotype would become permanently established in the field.
Herbicide resistant traits can also spread via movement of seed from the source field to fields free of the resistant biotype. Neither waterhemp nor giant ragweed has seed possessing long distance dispersal mechanisms. However, the lack of dispersal mechanisms is not a limitation since seeds can easily be transported by machinery. GR waterhemp was found more than 25 miles from the source field two years after resistance was identified (Liu et al. 2010), thus it moved much farther than the Illinois researchers' pollen transfer model predicted.
The Illinois researchers speculated that seed movement by farm machinery or independent selection was responsible for the appearance of GR a long distance from the source field.
Summary: Diversified weed management is the answer
Glyphosate has been heavily relied upon for weed management since the introduction of Roundup Ready soybeans in 1996. This widespread use has resulted in selection of GR biotypes throughout Iowa. Now that resistant biotypes are present, the spread of GR across the landscape is probably greater due to gene flow rather than the independent selection of resistant biotypes. This is due to the low initial frequency of glyphosate resistance within weed populations.
Does the threat of resistance movement via gene flow mean it is futile for individual farmers to prevent GR from appearing in fields they manage? The answer is no, and in fact, it places greater emphasis on the value of diversified weed management systems.
A diverse herbicide program continually places different herbicide stresses on the weed community. This greatly reduces the benefit of the GR trait to weeds, and therefore reduces the likelihood of resistant biotypes becoming established in fields currently free of the resistant trait. Use of diverse management programs will maintain the value of glyphosate by reducing the spread of GR weeds.
Reference: Liu, J., P.J. Tranel and A.S. Davis. 2010. Modeling the spread of glyphosate resistant waterhemp. Proc. North Cent. Weed Science Society.