Researchers found that on plots where herbicides have never been used, yield losses to weeds have been consistently increasing since the 1960s.
They used data from the world’s longest-running experiment, the Rothamsted Broadbalk wheat trial in Hertfordshire.
According to the study, less than a third of the harvest was lost to weeds in the first ten years of the dataset, but between 2005-2014, this had risen to more than half.
This is due to weeds doing better than crops in a warming climate, coupled with a shift towards shorter crop varieties that get shaded out by the taller weeds, the team from Rothamsted Research say.
They add that just like the crops, many weed species have also benefited over this period from increased use of nitrogen fertilisers.
The research further suggests more than half a century of consistent chemical spraying has also led to the rise of herbicide-resistant weeds.
This threatens the ability to protect the gains in crop productivity achieved since the 1960s.
Lead author Dr Jonathan Storkey, said: “Reducing yield losses from weeds is increasingly challenging because of the evolution of herbicide resistance and we can no longer rely on herbicides alone to counter the increasing threat from weeds.
“If we compare yield lost to weeds in the first ten years of the dataset, weeds ‘robbed’ on average 32% of the yield compared to 54% in the last ten years of data.
“Our results demonstrate that weeds now represent a greater inherent threat to crop production than before the advent of herbicides, and integrated, sustainable solutions to weed management are urgently needed to protect the high yield potential of modern crop varieties.”
The data also indicated that weeds reduced wheat yields proportionally more on plots with higher rates of nitrogen fertiliser.
On comparable plots where weeds were controlled with chemicals, higher wheat yields were achieved when more nitrogen fertiliser was added – meaning herbicides are most beneficial on the highest-yielding wheat fields.
Therefore, modern crop varieties with high fertiliser inputs are more vulnerable to higher yield losses if herbicide resistance becomes a greater problem.
The Broadbalk winter wheat trials were established in 1843 to contrast different amounts and combinations of inorganic fertilisers compared with farmyard manure on wheat yields.
Researchers looked at data from 1969 onwards – around the start of the so-called Green Revolution when higher yielding wheat cultivars, and increased nitrogen fertiliser use, became the norm worldwide.
These increased global yields were only made possible with a concurrent increase in the use of pesticides.
A total of 41 weed species were recorded in the plots, with farmers’ main nemesis, blackgrass, being ever present – and especially abundant on plots with high fertiliser rates.
Dr Storkey said going forward “it is likely that plant breeders will need to consider traits that reduce yield losses from weeds as well as those that optimise yield potential”.
The final factor potentially responsible for the increase in weed pressure is climate change.
Between 1969 and 2014, average air temperatures measured at the meteorological site local to the Broadbalk experiment have risen consistently.
As the growth of the weeds responds more to warming temperatures than the crop, this has given the weeds a competitive advantage, researchers suggest.
Dr Storkey said: “Management and climate change have combined over the past 45 years to increase the threat from weeds.
“If we could no longer rely on herbicides, it could be argued that, in terms of weed pressure, the situation is now worse than before their widespread introduction in the 1960s.
“Our results highlight the need to diversify weed control strategies by complementing herbicides with non-chemical options including increasing crop competition and disrupting weed life cycles using fallows or more diverse cropping rotations.”
The study is published in the journal Global Change Biology.