Commercial glyphosate-based herbicides effects on springtails (Collembola) differ from those of their respective active ingredients and vary with soil organic matter content

Glyphosate-based herbicides (GBH) are currently the most widely used agrochemicals for weed control. Environmental risk assessments (ERA) on nontarget organisms mostly consider the active ingredients (AIs) of these herbicides, while much less is known on effects of commercial GBH formulations that are actually applied in the field. Moreover, it is largely unknown to what extent different soil characteristics alter potential side effects of herbicides. We conducted a greenhouse experiment growing a model weed population of Amaranthus retroflexus in arable field soil with either 3.0 or 4.1% soil organic matter (SOM) content and treated these weeds either with GBHs (Roundup LB Plus, Touchdown Quattro, Roundup PowerFlex) or their respective AIs (isopropylammonium, diammonium or potassium salts of glyphosate) at recommended dosages. Control pots were mechanically weeded. Nontarget effects were assessed on the surface activity of the springtail species Sminthurinus niger (pitfall trapping) and litter decomposition in the soil (teabag approach). Both GBHs and AIs increased the surface activity of springtails compared to control pots; springtail activity was higher under GBHs than under corresponding AIs. Stimulation of springtail activity was much higher in soil with higher SOM content than with low SOM content (significant treatment x SOM interaction). Litter decomposition was unaffected by GBHs, AIs or SOM levels. We suggest that ERAs for pesticides should be performed with actually applied herbicides rather than only on AIs and should also consider influences of different soil properties.