Wed, Aug 17, 2022: 2:15 PM-2:30 PM
518A
Background/Question/MethodsMany crops benefit from natural enemies of pests, such as predatory arthropods. However, frequent harvesting can disrupt arthropod communities and diminish natural enemy populations. Nearby refuge habitats may help sustain predator populations and reduce pest pressure, but little is known about how the quality of refuge habitat and its spatial arrangement can modify this effect.We addressed several research questions, including: how do insect abundances vary across seasons? What are the qualities of a good insect refuge? Does the proximity and spatial arrangement of refuge habitat matter? And, how can we best use satellite imagery in the analysis of agricultural landscapes? To address these questions, we studied alfalfa (Medicago sativa) farms across northern Nevada, focusing on predators of the dominant aphid (Aphididae) pests.We combined field measurements of arthropod abundances and vegetation characteristics with satellite imagery to identify landscape features that are associated with robust communities of beneficial arthropods. In addition, we used predator addition treatments to confirm whether increased predator abundances actually reduce aphid abundance. We used machine learning to identify land cover characteristics and used structural equation models to relate these characteristics to arthropod abundances. In addition, we compared several different analytical approaches and assessed their utility.
Results/ConclusionsThe abundances of aphids and predators fluctuated across seasons. In the spring, Acyrthosiphon aphids dominated, and were associated with high densities of ladybugs (Coccinellidae). In the fall, the dominance of Acyrthosiphon aphids was supplanted by other alfalfa aphids, and these were accompanied by a distinct community of predators, including parasitoid wasps (Ichneumonoidea), spiders (Arachnida), and minute pirate bugs (Anthocoridae). Across seasons, predator addition sometimes – but not always – reduced aphid densities. These results suggest that targeted efforts to support the right mix of arthropod predators may help managers tailor their pest management efforts to a particular species and season. Three types of satellite imagery were classified using different algorithms, and land cover classes were weighted by proximity to fields using several different weighting functions. These factors all had strong effects on model performance. Our results show that remotely-sensed data must be used with care in order to create informative models of arthropod distributions across agricultural landscapes. Spatial analyses are most likely to succeed when they account for the relevant biological factors and are supported by field-collected data. Enhanced spatial and temporal resolution may not improve models, especially when insect densities are driven by stable, large-scale landscape features.
Results/ConclusionsThe abundances of aphids and predators fluctuated across seasons. In the spring, Acyrthosiphon aphids dominated, and were associated with high densities of ladybugs (Coccinellidae). In the fall, the dominance of Acyrthosiphon aphids was supplanted by other alfalfa aphids, and these were accompanied by a distinct community of predators, including parasitoid wasps (Ichneumonoidea), spiders (Arachnida), and minute pirate bugs (Anthocoridae). Across seasons, predator addition sometimes – but not always – reduced aphid densities. These results suggest that targeted efforts to support the right mix of arthropod predators may help managers tailor their pest management efforts to a particular species and season. Three types of satellite imagery were classified using different algorithms, and land cover classes were weighted by proximity to fields using several different weighting functions. These factors all had strong effects on model performance. Our results show that remotely-sensed data must be used with care in order to create informative models of arthropod distributions across agricultural landscapes. Spatial analyses are most likely to succeed when they account for the relevant biological factors and are supported by field-collected data. Enhanced spatial and temporal resolution may not improve models, especially when insect densities are driven by stable, large-scale landscape features.