Publication bias
The funnel plots show the asymmetry in the data (discussed further
below), particularly in the case of land-use intensification (Figure
S5). In addition, in the data for the main model, as well as the data
for the models focused on the stressors of land-use intensification,
nutrient enrichment, and to a lesser extent climate change, there is a
lack of small effect sizes that have high precision (i.e., the tip of
the funnel) (Figure S5).
Discussion :
We examined a wide array of GCs across all soil fauna, and in doing so,
we identified that pollution and land-use intensification had the
greatest negative impacts on soil fauna communities. Climate change also
had a negative, albeit smaller, impact, although after adjusting for
biases within the publications, this significant effect was lost. For
GCs where environmental stressor was a significant predictor, there was
no strong indication that press stressors had greater impact on
biodiversity, except in the case of land-use intensification. Overall,
the effect of the GCs did not vary with context, as there was no effect
of habitat type on the estimate, and rarely was there an effect of the
different body size classes. Studying the impact of pollutants may not
be possible with aboveground organisms, due to the lack of primary
studies focussed on pollutant impacts , but by focusing on soil
biodiversity, we are able to understand how detrimental environmental
pollution is relative to all other GCs. The only GC that comes close is
land use intensification.