Optimising the seasonal deployment of surveillance traps for detection of incipient fruit fly populations

1. Several nations operate risk-based surveillance systems for early detection and proof of freedom from economically damaging fruit flies (Diptera: Tephritidae). In New Zealand, trapping runs from September to June, but there is a need to review the timing in relation to risk and efficiency. 2. Our methods assume that daily temperatures affect the rate of increase of a fruit fly population, which in turn determines the number of trappable adults present each day and therefore the probability of detecting the population. From this we identify which days of the year contribute most to the surveillance sensitivity and therefore the appropriate trapping period. 3. We modelled the likely seasonal dynamics of selected fruit fly species in New Zealand, based on the effect of temperature on population increase and decline. From this we identified which days of the year contribute most to the surveillance sensitivity and therefore the appropriate trapping period. Optimum trapping periods were defined as the minimum contiguous dates which would give a specific proportion (80%, 90%, 95%, or 99%) of the probability of detection compared to year round trapping. 4. The model suggests that Mediterranean fruit fly (Medfly) should be trapped in the upper half of North Island from at least November through May. Additional trapping in October and June may be warranted in the far north. Favourable microclimates or climate change to 2040 may allow Medfly populations to establish as far south as Christchurch and Alexandra, where trapping should run from mid December to early May. 5. Queensland fruit fly (Qfly) and oriental fruit fly are unlikely to establish south of Napier and optimum trapping times for these species are approximately mid November through May. Climate change under moderate scenarios is unlikely to significantly alter the optimum trapping periods until at least 2040. 6. The results were not sensitive to assumptions about trap efficacy and founder population size. However, they are based on simplified relationships between the intrinsic rate of increase and temperature. Validation against overseas populations suggests that the relationships used for Medfly and oriental fruit fly are reasonable, but the Qfly model may be less generally applicable. 7. Synthesis and applications: The analysis suggests that fruit fly trapping in New Zealand could be conducted from October through May without substantially reducing detection efficacy, and further springtime reductions would be justifiable for all but the far north.