Abstract Long-lasting overpopulation of deer can prevent or retard the recovery of ground vegetation of forests even after overpopulation would have abated. This persistent impact, or legacy effect of deer, would be explained by hypothetical mechanisms such as limitation of seed sources, dominance of competitors that are unpalatable to deer, and negative feedback from soil conditions degraded by direct and indirect effects of deer. So far, however, little evidence is available for existence of such legacy effects. To detect the legacy effects caused by sika deer (Cervus nippon) and reveal their possible mechanisms, we studied ground vegetation and soil condition of Japanese natural broad-leaved forests and cedar plantations, which have records of vegetation surveys of the past 11 years. Foliar cover and species composition of ground vegetation were analyzed together with the deer density index (DDI), which represents mean deer density of the last 9 years before the survey. The relationship between vegetation cover and deer density had significantly changed between the survey years, as was proved by a regression analysis. The cover of ground vegetation did not increase during the 11 surveyed years even at most study sites whose DDI value had decreased, thereby indicating the existence of legacy effects. This delay of recovery was not attributable to the dominance of unpalatable plants, which had slightly decreased during the 11 surveyed years. Palatable plants did not increase in most study sites where DDI decreased, suggesting seed source limitation. Negative indirect effects of DDI on soil physical property in the present, such as increased bulk density and decreased coarse porosity, were detected through structural equation modeling (SEM) analysis. In the case of broad-leaved forests, the SEM model that included the 2005 DDI values showed better performance than that using present DDI values, indicating a strong legacy of past deer impact. These facts suggest the existence of legacy effects, which possibly have been caused by seed source limitation and degraded soil conditions. Moreover, above statistical analyses showed that canopy closure, which can be caused by delayed management practices (e.g., gap creation and thinning), further limited recovery of vegetation. Ground vegetation of broad-leaved forests were suggested to show lower resilience than those of cedar plantations, requiring conservation activities such as canopy gap creation.

中文翻译：

---

梅花鹿种群过多对地面植被和土壤物理性质的影响

摘要 长期的鹿群过剩会阻止或阻碍森林地表植被的恢复，即使是在鹿群过剩已经缓解的情况下。鹿的这种持久影响或遗留效应可以通过假设机制来解释，例如种子来源的限制、鹿不喜欢的​​竞争者的主导地位以及鹿的直接和间接影响导致土壤条件退化的负面反馈。然而，到目前为止，几乎没有证据表明存在这种遗留效应。为了检测梅花鹿（Cervus nippon）造成的遗产效应并揭示其可能的机制，我们研究了日本天然阔叶林和雪松人工林的地面植被和土壤条件，这些植被具有过去 11 年的植被调查记录。地面植被的叶面覆盖和物种组成与鹿密度指数 (DDI) 一起分析，该指数代表调查前过去 9 年的平均鹿密度。回归分析证明，植被覆盖度与鹿密度之间的关系在调查年份之间发生了显着变化。即使在大多数DDI值下降的研究点，在11个调查年中，地面植被覆盖率也没有增加，从而表明存在遗留效应。恢复的这种延迟不是由于难吃植物的主导地位，在调查的 11 年中略有下降。在 DDI 下降的大多数研究地点，可口的植物没有增加，表明种子来源受到限制。目前 DDI 对土壤物理性质的负面间接影响，通过结构方程模型 (SEM) 分析检测到诸如堆积密度增加和粗孔隙率降低等。在阔叶林的情况下，包含 2005 年 DDI 值的 SEM 模型显示出比使用当前 DDI 值更好的性能，表明过去鹿影响的强大遗产。这些事实表明存在遗留效应，这可能是由种子来源限制和土壤条件退化造成的。此外，上述统计分析表明，由于管理措施延迟（例如，间隙产生和变薄）可能导致冠层关闭，进一步限制了植被的恢复。建议阔叶林的地面植被显示出比雪松人工林低的恢复力，需要进行保护活动，例如形成冠层间隙。