Review of studies on crop-weed interactions reveals a lack of methodological standardization

Intensive management practices have been conducted since the 1950's to reduce the impact of weed competition on crop production. It is now widely recognized that the intensive use of herbicides has led to air, water and soil pollution, as well as massive biodiversity loss. Several public policies have been implemented to reduce herbicide use. However, is it possible to reduce herbicides while keeping sustainable agricultural production? We addressed this question by reviewing studies investigating the relationship between crop yield, weeds and herbicides from 1970 to 2016. A query to Web of Science surprisingly showed a mere number of 54 articles. We considered methodological choices such as the variety of metrics used to measure crop-weed competition or the herbicide used. Among the 40 articles for which full text was available, the dominant crops were wheat and corn. Most studies were conducted in experimental fields, 12% in actual farms and only 5 % in greenhouses. 70 % assessed the effect of a single herbicide, and only 12 % three or more herbicides and only 16 % tested multiple doses of herbicides. Similarly, only 10 % considered more than ten weed species. Forty eight different metrics were used to quantify the impact of weeds on crop. 55 % of studies reported a negative impact of weeds on grain yield, compared to weed-free controls. One of the two studies reporting economic analysis, one mentioned an increase in net benefits. In this talk, we will present how studies results are closely linked to methodological choices.