Can population modelling predict potential impacts of biocontrol? a case study using Cleopus japonicus on Buddleja davidii.

As weed biological control comes under much closer scrutiny from legislators and risk managers, increasingly we are asked to provide evidence on the potential impacts a biological control agent will exert on the weed we want to control. Invariably this evidence is required well before the potential biocontrol agent can be released from quarantine. When quantitative data cannot be readily accessed from the country of origin, tools such as population modelling become invaluable. From laboratory studies, the biology of Cleopus japonicus (Curculionidae) had already been ascertained in relation to temperature and day length. Additional studies were undertaken on the leaf area consumption by larvae and adults. Using these data, we simulated the population dynamics of the weed biocontrol agent as if it was being released in a non-limiting monoculture of its host plant, the weed buddleia ( Buddleja davidii; Buddlejaceae), in the central North Island of New Zealand. The results are useful for predicting the potential impacts on the weed, the rate of population build-up, and how many generations can be expected per annum in the likely distribution of the agent. The model predicts that only two generations of Cleopus japonicus can be expected per year and that overwintering survival is critical to population build-up. Experiments that ascertained the consumption of leaf area by larvae and adults showed that the leaf area index (LAI) for buddleia will be significantly reduced only from mid-summer until midwinter, leaving the spring flush undamaged. The extent to which population modelling such as this will be utilized and accepted as a predictive tool before the release of weed biological control agents will depend upon the verification of predictions such as these.