Predicting impact of a biocontrol agent: integrating distribution modeling with climate-dependent vital rates

Species distribution models can predict the suitable climatic range of a potential biological control agent (BCA), but they provide little information on the BCA's potential impact. To predict high population buildup, a prerequisite of biocontrol impact, studies are needed that assess the effect of environmental factors on vital rates of a BCA across the environmental gradient of the BCA's suitable habitats, especially for the region where the BCA is considered for field release. We extended a published species distribution model with climate-dependent vital rates of Ophraella communa, a recently and accidentally introduced potential BCA of common ragweed, Ambrosia artemisiifolia in Europe. In field and laboratory experiments, we collected data on climate-dependent parameters assumed to be the most relevant for the population buildup of O. communa, i.e., temperature driving the number of generations per year and relative humidity (RH) determining egg hatching success. We found that O. communa concluded one generation in 334 cumulative degree days, and that egg hatching success strongly decreased from > 80% to RH drops from 55% to 45% during the day. We used these values to spatially explicitly project population densities across the European range suitable for both A. artemisiifolia and the beetle and found that the present distribution of the beetle in Europe is within the range with the highest projected population growth. The highest population density of O. communa was predicted for northern Italy and parts of western Russia and western Georgia. Field observations of high impact on A. artemisiifolia with records of 80% aerial pollen reduction in the Milano area since the establishment of O. communa are in line with these predictions. The relative importance of temperature and RH on the population density of O. communa varies considerably across its suitable range in Europe. We propose that the combined statistical and mechanistic approach outlined in this paper helps to more accurately predict the potential impact of a weed BCA than a species distribution model alone. Identifying the factors limiting the population buildup of a BCA across the suitable range allows implementation of more targeted release and management strategies to optimize biocontrol efficacy.