Optimizing control of an invasive perennial herb through demographic analysis of a geographically replicated field experiment

The control of invasive plants can be costly and labor-intensive. Mowing provides a simple and tractable control option in many settings, but its efficacy depends on when the mowing occurs in relation to the target species' life cycle. Here we present the results of a multi-year field experiment, replicated over 25 sites across Sweden, testing the effects of different mowing regimes on the invasive large-leaved lupine, Lupinus polyphyllus. Early-season mowing (during the flowering stage) had stronger effects than late-season mowing (during the fruiting stage), but combining both mowing periods was more effective than either on its own. Mowing reduced the number as well as the size of lupine shoots, and especially depressed the rate at which lupine populations flowered, which may help reduce population spread. Demographic modeling indicated that continued twiceyearly mowing would lead to suppressed or reversed population growth at all sites. However, results were strongly dependent on habitat, with strong population reduction in woodlands, weaker effects in roadside environments, and intermediate results in grasslands. The lack of a latitudinal pattern suggests that twice-yearly mowing may be a useful control method for L. polyphyllus across its invasive range, although success is likely to depend on habitat type. More broadly, our results underscore that the within-season timing and frequency of weed control treatments can both greatly affect the outcome.