Natural variability of four ecotoxicological test species in soils-proof of concept at a regional scale and first step towards a national reference response.

Ecotoxicological indicators could be an important tool in soil monitoring programs to help understand the effects and ecotoxicity of potentially contaminated soils. For soil, one of the most complex environmental matrices, they can elucidate questions regarding the bioavailability of contaminants. However, the use of ecotoxicological indicators in retrospective risk assessment, in particular monitoring, struggles with the lack of adequate reference sites and/or responses. In this study we aim to define a reference or normal response range by evaluating the natural variability in soils of 4 ecotoxicological indicators species (Folsomia candida, Enchytraeus crypticus, Brassica napus, Allium cepa) including 3 endpoints (reproduction, shoot length and root length). Two approaches were used to establish the normal response range, one based on variability towards a control response (minimum detectable differences and maximum tolerable inhibition) and a percentile response based on the variability within the natural test soils. For this, 32 natural soils from the Jura region of Switzerland were tested with the 4 ecotoxicological indicators. To ensure responses were not driven by background chemical concentrations samples were analysed for metal concentrations, as regulated by Swiss legislation and for 144 pesticide residuess. Pesticides concentrations in the selected soils were very low and higher metals concentrations were mostly of geogenic origin and did not appear to influence biological responses. Regarding biological variability, it was possible to determine normal operating ranges for all four ecotoxicological test species, and the percentile approach seems better suited for natural soils from the Jura region, for which LUFA 2.2 is not always very representative. The percentile approach shows a high potential as a methodology for setting reference response ranges based on the natural variability in soils. The approach now needs to be compared to potentially contaminated soils and extended to other biogeochemical regions to allow its use in monitoring programs.