Optimal schedule for monitoring a plant incursion when detection and treatment success vary over time

Abstract

Management of an invasive plant species can be viewed as two separate and successive processes. The first, survey, aims to find infested areas and remove individuals. The second, monitoring, consists of repeated visits to these areas in order to prevent possible re-emergence. As detection probability may vary over time, the timing and number of monitoring visits can dramatically impact monitoring efficacy. We explore the optimal timing and number of monitoring visits, by focusing on one infested site. Our decision-analysis framework defines an optimal monitoring schedule which accounts for a time-dependent probability of detection, based on the presence/absence of a flower. We use this framework to investigate the optimal monitoring schedule for Hieracium aurantiacum, an invasive species in the Australian Alps and many other countries. We also perform a sensitivity analysis to draw more general conclusions. For H. aurantiacum eight monitoring visits (compared to 12 visits in the current program) are sufficient to obtain a 99% monitoring efficacy. When four or fewer visits to a site are allowed, it is optimal to visit during the high season, when the weed is likely to initiate flowering. Any extra visits should be scheduled in the early season, before the plants flower. The sensitivity analysis shows that increasing the detection probability early in the season has a greater impact than increasing it late in the season. An effective treatment method increases the value of site visits late in the season, when the detection probability is higher. Our decision-analysis framework can assist invasive species managers to reduce or reallocate management resources by determining the minimum number of monitoring visits required to satisfy an acceptable risk of re-emergence.