A Weibull distribution-based method for estimating seed longevity in Solanum rostratum

Introduction Seed longevity is a key determinant of population persistence, spread, and outbreak potential in annual invasive plant species. Understanding longevity of invasive seed bank is crucial for determining colonization timing and assessing invasion potential, thereby supporting sustainable weed management strategies. While soil seed bank fluctuations have become a focus in invasion biology area, efficient and accurate methods for evaluating seed bank longevity in annual invasive plants remain scarce so far. In this study, we focus on a representative annual globally malignant invasive plant Solanum rostratum , investigating seed longevity by accelerated aging test (60°C and 85% relative humidity) across multiple regions and collection years. Methods We used a three-parameter Weibull distribution model to characterize seed aging and applied it to assess S. rostratum seed bank longevity in both grassland and abandoned farmland habitats. Results The results showed that S. rostratum seeds lost viability rapidly within 3 d under accelerated aging condition. Seeds from different regions in the same year exhibited similar aging patterns, while interannual variation led to significantly divergent aging curves. Based on polynomial regression of viability data and germination tests, the upper limit of seed longevity under natural field conditions was estimated to be approximately 8–9.79 years. Discussion This study demonstrates that combining accelerated aging assays with the three-parameter Weibull distribution provides an effective approach for assessing seed longevity and soil seed bank persistence. The method offers a practical, efficient, and reproducible framework for estimating seed bank persistence in annual invasive plants. Our findings highlight the critical role of persistent seed banks in facilitating the invasion success of S. rostratum , thereby offering a robust analytical basis for evaluating invasion risks. Moreover, the modeling framework developed here can be extended to other annual plant species for seed viability assessment, providing valuable theoretical support for the development of ecologically sustainable weed management strategies.