Science in Society Archive

GM Contamination At 21 km and Farther
No Co-Existence Possible

Researchers find GM pollen cross-pollinated non-GM plants at 21 km and predict much worse. Dr. Mae-Wan Ho

GM contamination rampant

Everyone knows by now that GM contamination of non-GM crops and produce is inevitable. There have been142 contamination incidents recorded worldwide since 1997, according to the GM Contamination Register [1]. This is an underestimate, as not every shipment of non-GM produce has been tested, and not every incident registered.

Many wary consumers are buying organic to avoid eating GM food. But GM contamination of organic produce is no longer a rarity.

A recent case of serious contamination involved a shipment of organic soybeans to a processor in the United States. The processor had the shipment tested after being tipped off by a buyer. The lab result showed up a massive GM contamination of 20 percent. The organic certifier was unable to prosecute the supplier, also in the US, who sent a different sample for testing. The processor lost $100 000 in the incident, but the supplier was still selling his crop [2].

Perhaps in anticipation of widespread contamination of organic produce, the European Union Council of Agriculture Ministers voted in June 2007 that organic produce could contain up to 0.9 percent GM [3], despite the fact that, in March the same year, the European Parliament passed a directive setting the contamination threshold at 0.1 percent, which effectively maintained the organic industry’s insistence on “zero tolerance” of GM contamination.

Current separation distances derisory

The current regulatory regime where GM and non-GM crops are allowed to grow in adjacent fields separated by tens or hundreds of metres is based on the assumption that the separation distances are sufficient to reduce cross-pollination levels to the acceptable minimum.

Pollen flow is not the only means of GM contamination. Other means involve GM seeds: impurities in the seed stock, volunteers from a previous crop, seeds dropped during transport, seeds inadvertently mixed by suppliers and during processing. Seeds persist much longer and can travel much farther. When the same machinery is used in several fields for harvesting, cultivation and spraying, seeds will be readily moved around from one field to another.

But even the extent of cross-pollination is greatly underestimated, as has been pointed out, pollen can remain airborne for hours and a 25 miles-per-hour wind speed is not unusual [4], which is why extensive contamination of certified seed stocks had been detected as far back as 2003 (Transgenic Contamination of Certified Seed Stocks, SiS 19).

Two recent scientific studies have now confirmed that the extent of cross-pollination has been greatly underestimated.

GM bentgrass pollen spreads 21 km or more

A research team led by scientists at the US Environment Protection Agency in Corvallis, Oregon, used an atmospheric model of wind blowing above fields planted with GM bentgrass to look at GM pollen dispersal, combining modelling with actual analysis of cross pollination with non-GM plants of the same or related species [4]. The GM bentgrass carried the glyphosate tolerance trait, which provided a ready selectable marker for cross-pollination.

During extensive greenhouse and laboratory testing, glyphosate-tolerant progeny of non-GM test plants were found up to 21 km from the GM fields.

This was consistent with the model of wind direction and speed, which showed movement of pollen up to 15 km from the GM fields by the first hour; and maximum travel distances increased to 40 and 50 km after two and three hours respectively. The three-hour cut off period was based on previous findings that the viability of the grass pollen dropped to zero within three hours.

These findings were at odds with previous small-scale experiments, involving hundreds of GM plants in small plots, which showed pollen dispersal limited to a few kilometres, basically because the source of GM pollen was too small. It is like putting a drop of ink in an ocean, which soon gets diluted.

In the present experiment, GM bent grass was planted in 162 hectares, at about 2.8 million seeds per hectare. This provided realistically high pollen concentration for the estimation of pollen dispersal. The maximum potential spread of 21 km observed was an underestimate because pollen trapping plants were not set much further than the distance observed. According to the model, GM contamination could be as far as 75 km downwind of the GM field.

Realistic modelling shows highly variable and unpredictable GM pollen spread, but most studies underestimate the real extent

In a similar theoretical study, researchers at Exeter University in the United Kingdom used records of wind direction and speed from 27 weather stations across Europe to predict pollen dispersal and wind-borne cross-pollination in maize, oilseed rape, sugar beet and rice [5].  Their results showed that cross pollination rates vary greatly according to the relative orientation of the GM and non-GM fields, and substantially from year to year. The main determining factor is wind direction, which accounts for most of the variation, 75 percent in the case of maize in the UK.

For maize and rice, cross-pollination rates are relatively high only if the non-GM field is downwind of the GM field with respect to the prevailing winds over the short pollination period. In contrast, contamination rates vary least with field orientation in crops with relatively long flowering periods, such as oilseed rape and sugar beet, because the distribution of wind directions becomes more even as the flowering period lengthens.

“Consequently, even replicated field trials may inaccurately estimate typical levels of cross-pollination, and therefore distort our perception of the separation distances required to achieve sub-threshold adventitious GM presence.”

The best one could do is to predict the likely range in levels of cross-pollination based on limited data typically available from field trials, and to introduce time delays between the peak-flowering periods in adjacent fields to reduce cross-pollination to a specific level.

What the model actually says is that a contamination rate measured in any single experiment without knowing the prevailing winds is unreliable, because the wind changes direction from day to day and year to year. The model gives the mean maximum and minimum relative rates based on prevailing conditions, not the absolute rates. For example, suppose that a rate of 0.001 percent contamination was measured in a single field trial and the necessary meteorological records were not available. For the UK, the mean maximum and minimum relative cross-pollination rates for maize were estimated by the model to be 7 and 0.0005 respectively, in the case of maize grown in fields of dimension 500 x 200m, separated by 1 000 m. The maximum possible rate is given by the ratio of the maximum and minimum rates multiplied by the measures rate, i.e., 7/0.0005 x 0.001 = 14 percent, which is quite substantial. If the prevailing weather and wind conditions were known, then the estimate improves considerably. Suppose that the maize trial was carried out in Leeds in 1998, and the relative orientation of the GM and non-GM fields was 100o, and according to the prevailing wind conditions, the minimum relative rate was 0.5 while the maximum relative rate was 9.5, then a measured value of 0.001 percent would give a maximum possible rate of 9.5/0.5 x 0.001 = 0.019 percent, substantially less than the previous estimate. The moral in the example is that wind speed and direction should be measured during future field trials. No such data currently exist.

The researchers stress that their analysis is conservative, because if a very large quantity of pollen is released in strong gusts of wind, then cross-pollination rates will be even more extreme.

For oilseed rape, sugar beet and rice, contamination rates could be reduced by 50 percent when the lag between the times of peak flowering of the GM and non-GM fields is 13 days and by 90 percent when the lag is 24 days. For maize, similar reductions require lags of only 4 and 8 days respectively because of the shorter flowering period.

Maize, sugar beet and rice are almost entirely cross-pollinated by wind, whereas oilseed rape is cross-pollinated by both wind and insects.

Stop planting GM crops right now   

It is clear that transgene contamination is inevitable and unavoidable if GM crops are planted. We must make the choice to stop planting GM crops right now, not only to avoid the massive economic losses involved in transgene contamination incidents, but also on the basis of the now irrefutable evidence that GM crops are neither safe not sustainable [6, 7] (Scientists for a GM Free Europe, No to GMOs, No To GM Science) 

Article first published 27/06/07


  1. GM Contamination – imports of food and feed at risk, GM Freeze, May 2007,
  2. Questions of fraud raised by GMO-contaminated shipment of organic soybeans. The Organic & Non-GMO Report. June 2007,
  3. “Ministers open door for GMOs in organic food”, European Environmental Bureau, Friends of the Earth Europe, Greenpeace, Press Release, 12 June 2007.
  4. Van de Water PK, Watrud LS, Lee EH, Burdick C, King GA. Long-distance GM pollen movement of creeping bentgrass using modelled wind trajectory analysis. Ecological Applications 2007, 17(4), 1244-56.
  5. Hoyle M and Cresswell JE. The effect of wind direction on cross-pollination in wind-pollinated GM crops. Ecological Applications 2007, 17 (4), 1234-43.
  6. Burcher S and Ho MW. Scientists call for a GM free Europe, Report on European Parliament Briefing, 12 June 2007, Science in Society 35.
  7. Ho MW. No to GMOs, no to GM Science, Science in Society 35.

Got something to say about this page? Comment

Comment on this article

Comments may be published. All comments are moderated. Name and email details are required.

Email address:
Your comments:
Anti spam question:
How many legs on a tripod?