Effects of pesticides on soil invertebrates in model ecosystem and field studies: a review and comparison with laboratory toxicity data.

A systematic review was carried out to investigate the extent to which higher-tier (terrestrial model ecosystem [TME] and field) data regarding pesticide effects can be compared with laboratory toxicity data for soil invertebrates. Data in the public domain yielded 970 toxicity endpoint data sets, representing 71 pesticides and 42 soil invertebrate species or groups. For most pesticides, the most frequent effect class was for no observed effects, although relatively high numbers of pronounced and persistent effects occurred when Lumbricidae and Enchytraeidae were exposed to fungicides and when Lumbricidae, Collembola, and Arachnida were exposed to insecticides. No effects of fungicides on Arachnida, Formicidae, or Nematoda or of herbicides on Lumbricidae, Formicidae, or Nematoda were observed in any studies. For most pesticides, higher-tier no-observed-effect concentration or lowest-observed-effect concentration values cannot be determined because of a lack of information at low pesticide concentrations. Ten pesticides had sufficient laboratory data to enable the observed higher-tier effects to be compared with 5% hazardous concentrations (HC5) estimated from acute toxicity laboratory data (atrazine, carbendazim, chlorpyrifos, diazinon, dimethoate, gamma-hexachlorocy-clohexane, lambda-cyhalothrin, parathion, pentachlorophenol, and propoxur). In eight cases, higher-tier effects concentrations were within or below the 90% confidence interval of the HC5. Good agreement exists between the results of TME and field tests for carbendazim, but insufficient information is available for a comparison between TME and field studies for other pesticides. Availability and characteristics (e.g., taxonomic composition and heterogeneity) of the higher-tier effects data are discussed in terms of possible developments in risk assessment procedures.