Short and long-term impacts of ultra-low-volume pesticide and biopesticide applications for locust control on non-target arid zone arthropods
While locust control is necessary to avoid the high cost of locust damage to agriculture, land managers are increasingly seeking to minimize the environmental impact of pesticide spray treatments used. The comparative impacts of different locust control treatments on non-target arid zone fauna are rarely studied in the field, leading to uncertainty as to which treatments represent the lowest hazard to the sprayed ecosystems. A phenyl pyrazole pesticide, fipronil, and a fungal biopesticide, Metarhizium acridum (Green Guard®) were applied aerially in either a barrier or blanket ULV treatment at replicated sites which mimicked the techniques employed for locust control operations in Australia. Effects of the two pesticide treatments were compared in the absence of dense locust populations. We measured the abundance and community composition of non-target arid-zone arthropods at control and treatment sites before and after pesticide applications using a large field-based pitfall trapping experiment. Arthropod community composition was not significantly affected over time by either locust control treatment. However, significant short-term times x treatment interactions were found for 6 of 11 most common taxa at family or higher taxonomic level (collembolans, acarians, coleopterans, psocopterans, gryllids, and dipterans). We also compared unsprayed and sprayed areas within fipronil and Metarhizium treatment sites, and found 2 of the 10 most common ant species (Formicidae: Rhytidoponera mayri and Iridomyrmex purpureus) showed significant time x treatment interactions for fipronil but none for Metarhizium, indicating that ants were more severely affected by fipronil within sites than between the three treatments. One year post-treatment, significant time x treatment interactions persisted for only two taxa (dipterans and blattodeans) at Metarhizium treatments, indicating full recovery of most taxa. The suppression of the ant R. mayri in fipronil sprayed areas within treatment sites persisted after one year, while I. purpureus had fully recovered. Relative arthropod abundance and community assemblage changed over time in control and treatment sites, probably reflecting changes in patterns of local rainfall over the study period. Most of the statistically significant treatment effects recorded for different taxa in our study were not long lasting, suggesting that the two locust control methods studied represent a relatively low and transient hazard to most arthropod taxa. The pronounced temporal variation in arthropod abundance across all sites indicated that climate and environmental factors are likely to be stronger drivers of arid zone arthropod abundance and community structure than single aerial applications of low-dose aerial pesticide treatments used to control locusts in arid and semi-arid regions of Australia.