Edge Effects On Forest Understory Vegetation And Soil Seed Banks

Human-induced forest edges are expanding in worldwide forest landscapes. Their ecological consequences have attracted many attentions from ecologists and conservationists. This study is focused on determining changes in species composition in understory vegetation and soil seed banks at human-induced forest edges and quantifying human disturbance and forest edge ecosystem degradation. The understory vegetation and soil seed banks were sampled at three forest edges from a continuous mid-mountain moist evergreen broad-leaved forest on the northern crest of the Ailao mountains, and at three edges from three tropical seasonal rain forest fragments in the Bubeng area of Xishuangbanna, Southwest China.The pattern of species composition in understory vegetation along an edge- interior gradient revealed the presence of the three ecological species groups. For the Lithocarpus xylocarpus forest, the degree (abundance, richness, diversity, relative abundance and relative richness) of invasion by nonconstituent species and intrusion by secondary species was relatively high at the immediate edge (0-15 m) and decreased sharply in the interior (>15 m) till these species disappeared. The abundance of primary species decreased along the edge to interior gradient, while the richness and diversity did not vary significantly. Thus, there existed three areas with distinct species assemblage which could be quantitatively detected (0-15 m, 15-30 m, 30-60 m). For the Shorea wantianshuea forest, the degree of invasion by nonconstituent species was low. The degree of intrusion by secondary species decreased significantly at the immediate edge (0-10 m) and kept relatively stable at the succedent edge. The degree (abundance, richness and diversity) of existence of primary species was high and did not vary significantly along the edge-interior gradient. Thus, there did not exist distinct species assemblage that could be quantitatively detected at the edge zone (0-50 m).The species composition in soil seed banks at the edges was relatively homogeneous whether in Lithocarpus xylocarpus forest or in Shorea wantianshuea forest. The dominance pattern of the three ecological groups did not vary significantly along the edge-interior gradient. Secondary species was always the most dominant ecological species group. The invasion of nonconstituent species in understory vegetation lagged behind that in soil seed banks for both penetration depth and occurrence period. We concluded that there was not a lack of propagule pressure for invasion by nonconstituent species at forest edges. The indicators of nonconstituent species in soil seed banks could provide early warning signals for the invasion. We constructed a framework model for predicting and warning the invasion of nonconstituent species in forest ecosystems.The edge responses of understory vegetation and soil seed banks reflected a strong damage in forest edge ecosystems. Although the degree of damage varied with forest ecosystems, the degradation of forest edge ecosystems was prevalent. The result of quantitative diagnosis showed that, the degradation did not exceed the threshold of ecosystem security and was reversible through the natural restoration in forest edge ecosystems. We constructed an innovative conceptualized model of the dynamic processes of forest edges.Combining ecological attributes with edge preference, we identified 5 functional response groups for human disturbance: invading, flourishing, proliferating, degrading and indifferent functional response groups in understory vegetation at forest edges. They might be applicable for other forest ecosystems.28 indicators based on 7 indicator species groups selected from ecological species groups, edge response species groups and functional response groups, were used in principal component analysis (PCA) to ordinate the plots of understory vegetation at forest edges. The integrating indices derived from the principal components showed that there were three orders of disturbance intensity from intense to light along the edge to interior gradient, which coincided with the suitability of our methodology for quantifying human disturbance. We strongly suggest that quantitative assessment of human disturbance in terms of species responses is essential and feasible in forest management.