| dc.description.abstract | The awareness of biodiversity loss stimulates amounts of studies to explore the ecosystem consequence of biodiversity loss in the past two decades. However, forecasting the consequence of biodiversity loss in natural ecosystems is still challenging because the effects of biodiversity on ecosystem functioning is embedded in a complex network of ecosystems components with mutual interactions and these are poorly understood. On the one hand, although the experimental studies have successfully demonstrated the positive influence of biodiversity on ecosystem functioning, these experiments can only reveal incomplete unidirectional causality of biodiversity under highly simplified environmental context. On the other hand, the studies based on empirical data analyses face the problem that linear analyses (correlation, regression, structure equation modeling, etc.) cannot reveal causation in nonlinear dynamical systems, such as ecosystems. In this thesis, I apply empirical dynamical modeling (EDM) to construct the causality network among species diversity, ecosystem functioning, and environmental factors in natural phytoplankton communities. Through the cross-system analysis of the networks, I firstly recognize the prevalence of the diversity-mediated mutual interactions in natural ecosystems (Chapter 2). To explore the ecological roles of these mutual interactions, I propose a novel theoretical dynamical model which reveals the dynamical stabilizing effects of biodiversity via the mutual interactions between diversity and the other parts of dynamical ecosystem (Chapter 3). Moreover, I empirically test this dynamical stabilizing effects in natural aquatic ecosystems and demonstrate how these stabilizing effects weakened under climate warming (Chapter 4). Finally, by acknowledging the complexity of diversity-associated interaction network, I propose a new EDM method to evaluate the realistic consequence of biodiversity loss that account for environmental context (Chapter 5). In brief, I highlight the reality that diversity is one of the key components in dynamical ecosystems, and its dynamics, responses, and impacts are time-varying and context dependent. Therefore, studying the diversity effects embedded in dynamical interaction network is more helpful to understand the stabilizing mechanisms of biodiversity than studying the diversity effects in isolation. In conclusion, a new biodiversity-ecosystem functioning paradigm emerges from my thesis concerning the dynamical interaction networks. | en_US |