River embankments offer potential semi-natural habitats for grassland species, but their large-scale reconstruction leads to vast vegetation loss. To restore both ecosystem services and biodiversity on reconstructed river embankments, we tried to achieve rapid cover by native grasses and to enhance the richness of species typical of semi-natural grasslands. We conducted a 4-year experiment to clarify whether season and plant life history stage in revegetation determine competitive interactions among foundation species (Imperata cylindrica), subordinate species (eight species typical of semi-natural grasslands), and weeds (any other species). Square frames measuring 1 m × 1 m × 30 cm depth were filled with soil used for the construction of river embankments. Two I. cylindrica life history stage modalities (seeds, seedlings) and a control were established at three timings (March, May, July) with three replications. Fifty seeds of each subordinate species were sown in each plot. In the March and May timings, annual weeds were dominant in year 1. In the July timing, I. cylindrica became dominant because high temperatures inhibited the germination of weeds and several subordinate species. Planting seedlings in any season was successful in achieving dominance of I. cylindrica, but only in the absence of rhizomes of invasive Solidago altissima. Many individuals of subordinate species were recorded in the May timing. Subordinate species with higher competitive ability or with less niche overlap with I. cylindrica performed better. Even so, overgrowth by I. cylindrica or weeds generally emerged, leading to low survival rates of subordinate species in all treatments. Season or life history stage had no effect on richness. Since rapid coverage of river embankments by native grasses is essential, planting in July is desirable for restoration. A shortage of dominant weeds in year 1 and relatively slow dominance by sown I. cylindrica led to less severe competitive exclusion of subordinate species. Thus, sowing in July would enhance the dominance of I. cylindrica and the establishment of subordinate species. Since the germination rate in July sowing was the smallest, further studies of successful introduction methods in the warm season are needed.

河流堤防为草地物种提供了潜在的半自然栖息地，但是它们的大规模重建导致大量植被丧失。为了在重建的河堤上恢复生态系统服务和生物多样性，我们试图通过原生草实现快速覆盖，并提高半自然草原典型物种的丰富度。我们进行了一项为期4年的实验，以阐明植被恢复的季节和植物生命史阶段是否决定了基础物种（Imperata cylindrica），从属物种（半自然草原典型的8种）和杂草（任何其他物种）之间的竞争性相互作用。深度为1 m×1 m×30 cm的方形框架填充了用于修建河堤的土壤。两个I. cylindrica在三个时间点（3月，5月，7月）建立了3个重复的生活史阶段模式（种子，幼苗）和对照。在每个样地中播种了每个下属物种的五十个种子。在3月和5月的时间中，一年杂草在第一年占主导地位。在七月的时间中，由于高温抑制了杂草和一些下属物种的发芽，圆柱状伊勒菌成为优势。在任何季节都可以成功地播种幼苗，从而实现了I. cylindrica的优势，但前提是没有侵略性Solidago altissima的根茎。五月份的时间记录了许多下属物种的个体。次要的物种具有较高的竞争力或具有更少的生态位重叠白茅表现更好。即便如此，一般也会出现长岛I或杂草的过度生长，导致在所有处理中次生物种的存活率较低。季节或生活史阶段对丰富度没有影响。由于必须用原生草迅速覆盖河堤，因此，7月播种对于恢复很重要。第一年主要杂草的短缺和播种的圆柱状伊蚊的相对较慢的优势导致次要物种的竞争性排斥较轻。因此，在七月播种将增强的主导地位白茅，并建立下属的物种。由于7月播种的发芽率最低，因此需要进一步研究在暖季成功引入方法。