Status of invasive plant invasions and their abiotic and biotic habitats on Anijima, Ogasawara Islands, Japan.

Status of invasive alien plant species and their habitat on Anijima, Ogasawara Islands, Japan.

Abstract

Effective and efficient responses are essential for alien species control, especially given limited information and budgets. Early detection of species establishing themselves in undisturbed areas during distribution expansion, as well as evaluating the efficiency of damage reduction measures for already established species, can be achieved through species distribution models. These models help determine whether distribution expansion has reached saturation and estimate potential suitable habitats and growth areas. On Anijima Island in the Ogasawara Islands, where native dry scrub forests of high conservation value remain widespread, large-scale measures against alien plants are being developed. This study analyzed the invasive status and abiotic and biotic growth environments of six alien plant species frequently identified during control efforts on Anijima: Bidens pilosa, Psidium littorale, Leucaena leucocephala, Lantana camara, Casuarina equisetifolia, and Pinus luchuensis. Anijima was divided into approximately 10-meter square grids for the analysis. Two models were developed: a conditional autoregressive (Intrinsic CAR) model, which accounts for spatial autocorrelation, and a logistic regression model, which considers only environmental factors. The presence or absence of each species in each grid was used as the dependent variable, while environmental variables were used as explanatory variables. By comparing the Intrinsic CAR and logistic regression models, the impact of spatial autocorrelation on species distribution was quantified. A greater effect of spatial autocorrelation suggests the presence of key environmental factors not included in the model or that alien species are expanding in suitable environments but have not yet reached saturation. Although this analysis cannot distinguish between these possibilities, species with strong spatial autocorrelation effects may pose a higher risk of distribution expansion, necessitating careful monitoring and preventive control even in the absence of detailed information.The analysis indicates that the wind-dispersed tall tree species Casuarina equisetifolia and Pinus luchuensis may have distributions close to saturation, with limited spatial autocorrelation effects. In contrast, species with the highest spatial autocorrelation effects were Bidens pilosa, Leucaena leucocephala, and Lantana camara. Native forests appear to inhibit the distribution of Bidens pilosa and Leucaena leucocephala, underscoring the importance of conserving native forests to prevent the spread of these alien plants. Additionally, the tendency of Bidens pilosa, Leucaena leucocephala, and Lantana camara to thrive in open abiotic and biotic environments suggests that their potential distribution areas may include native bare lands and grasslands common on Anijima. Careful monitoring is essential to understand future distribution expansion and the characteristics of infested environments.