Effects of agricultural intensification on soil-associated arthropod population dynamics, community structure, diversity and temporal variability over a seven-year period

While most studies focusing on the effects of agricultural intensification on soil biota are inherently short-term in nature, longterm (multiyear) stu dies are essential in assessing long-term temporal responses of soil biota to agronomic practices. We investigated the effects of three components of agricultural intensification, i.e. cultivation (disturbance), herbicide add ition (modification of floristic composition) and mulching (resource additi on) on soil-associated arthropods in an annual (maize) and a perennial (asp aragus) cropping system over a 7 yr period. An additional treatment (hand-h oeing of weeds during the crop growing season) was used to represent minima l intensification. Many taxa of arthropods responded positively to mulching and to treatments which allowed high weed biomass in the non crop-producti on period, e.g, the hand-hoeing and cultivation treatments in the perennial crop. Herbicide treatments also facilitated high numbers of many taxa in t he annual crop when this coincided with plot invasion by herbicide-tolerant weeds. Generally, arthropod taxa were positively correlated with weed biom ass and negatively with crop plant biomass, probably because of the superio r resource (litter) quality produced by the former. Ordination analyses ind icated that arthropod community structure was often correlated with weed co mmunity structure. Mulching and allowing high weed biomass also promoted a high species richness of soil-associated Coleoptera, but coleopteran divers ity was not related to weed species diversity. Analyses of temporal variabi lity (inversely related to stability) of arthropod taxa across years reveal ed few treatment effects in the annual crop, but showed destabilising effec ts of weed reduction in the perennial crop. In the perennial crop, temporal variability was also positively correlated with crop biomass and negativel y with weed biomass across plots. Our study shows that agricultural intensi fication is not consistently harmful to the soil fauna, that soil-associate d arthropods are most responsive to management practices which affect the n ature and quality of resource input, and that long-term experiments are ess ential for answering questions about how agricultural practices affect soil organisms against the natural backdrop of temporal variation. (C) 1999 Els evier Science Ltd. All rights reserved.