Various control methods have been developed to manage exotic plant populations and to minimize their impact on native communities. Locally applied treatment methods are often preferable since they can reduce the economic and ecological costs of control. The success of these methods, however, depends on workers’ abilities to find and treat all plants within an infested field. Although the issue of detectability is not new to invasion biology, most attention has occurred at the landscape and regional scale. I propose that at least three factors should be considered when designing a local management program: 1) the degree that a treatment reduces the local density of the target species (= treatment intensity), 2) the spatial distribution of the population to be treated, and 3) how easily the target species can be found and treated. Using Lespedeza cuneata as a model, I developed a simulation program to examine how the above factors could affect occupancy and stem abundance after five years of population spread under treated and untreated conditions.

Results/Conclusions

When left untreated, occupancy was higher in fields with initially random spatial distributions compared to fields with initially patchy distributions. Occupancy was higher in random fields across all levels of treatment intensity when detectability was low, and small patches avoided detection and treatment. At higher detectability levels, the larger patches present in the patchy fields were easily found, but only the most intense treatment reduced occupancy after five years. Stem abundance was 1.2 times higher in fields with initially random spatial distributions if left untreated for five years. Under managed conditions, stem abundance was higher in patchy fields across all treatment intensity and detectability levels. In all cases the most intense treatment overwhelmed the effects of detectability. In actual infestations, changes in annual budgets result in treatment intensities that vary from year to year, thereby reducing the overall effectiveness of control programs. This scenario suggests that detectability could indeed have important management implications. The model, however, assumes that detection of a single, occupied cell is constant and independent from year to year. If managers create weed maps, then they could increase the probability of finding and treating previously detected patches within fields. By maximizing detectability, managers could successfully control local weed populations even when time and budget constraints limit yearly effort.