Comparing inward and outward strategies for delimiting non-native plant pest outbreaks

The delimitation of outbreaks is an essential step in the containment and eradication of non-native plant pests. Outbreaks are habitually delimited by sampling around the initial finding, moving away from this locus in several directions as long as infestations are found (outward strategy). An alternative, inward, strategy would entail starting delimitation with an initial estimate of the location of the frontier and then sampling inward until the first infestations are found or outward until no more infestations are found. We used individual-based modelling to compare the effectiveness and sampling effort of the two strategies. Both successfully contained > 99% of infested plants within the delimited zone. Yet, both had a low probability (< 15%) of containing all the infested plants within the delimited zone. The number of samples of the inward strategy depended greatly on the size of the initially hypothesized infested zone. Best performance of this strategy was obtained with an accurate initial estimate of the infested zone width, while sample size increased strongly when the estimated frontier was far beyond the true location of the frontier. Consequently, the outward strategy uses fewer samples on average than the inward strategy when the position of the frontier is uncertain. Both strategies were prone to error when delimiting outbreaks caused by pests with fat-tailed dispersal. Whether the inward or outward strategy is more effective depends on the certainty about the true position of the leading frontier of the outbreak. Possibilities are discussed for maximizing the cost-effectiveness of sampling for outbreak delimitation.