Abstract • Outbreaks of root-feeding scarab larvae in turfgrass are widely managed through preventive applications of imidacloprid. Long-residual activity and application before feasible scouting probably lead to its overuse and overexposure. • Recent investigations revealed a selective impact of imidacloprid (not trichlorfon or halofenozide) on certain nontarget turf arthropods, motivating the present study on the persistence of abundance effects over 6 years of annual applications. • Arthropods were sampled monthly (July to October) in replicated plots using soil core heat extraction and pitfall traps to quantify soil- and surface-active arthropods. Captures were identified to class, order or family. The most represented taxa were analysed to test for cumulative effects and their change over season and year. • Imidacloprid had no impact on pitfall captures, although the abundance of Hemiptera, Thysanoptera, Coleoptera and Collembola was suppressed in soil core captures. Among beetles, impact was expressed in adults (not larvae), and in Carabidae and Staphylinidae (not Chrysomelidae or Curculionidae). Among springtails, impact was expressed in Entomobryomorpha (not Poduromorpha or Symphypleona). Impact did not diminish with year but there was variable recovery between applications. • There may therefore be a diverging response of soil- and surface-active fauna to the nontarget impacts of imidacloprid. The suppression of predaceous (not phytophagous) beetles indicates an indirect effect mediated through declines in prey populations. • The magnitude of abundance effects confirms that the balance between target and nontarget impact should be explicitly examined. Implications are discussed with respect to functional relevance for nutrient cycling and the natural regulation of pests.