Among the multiple research approaches used to study ecosystem dynamics, large-scale experimental manipulations are the hallmark for establishing process-level understanding. Although long-term monitoring research is tremendously important for uncovering key ecological components of complex ecosystem dynamics, establishing the temporal context of processes and detecting the effects of infrequent disturbance events, determining causality for many patterns is impossible because of lacking controlled conditions. Comparative studies are important for unraveling broad spatial patterns, but they are unable to predict how a given ecosystem will respond to fluctuations. Manipulative experiments are designed to measure the effect of altering key variables in the context of other factors controlling ecosystem processes. However, because processes, patterns, and biotic responses are scale dependent, the mechanistic interpretation derived from the experimental manipulation is also constrained by scale dependency. Often, experiments are designed at a temporal and spatial scale with other experiments nested at lower scales, but still accounting for landscape heterogeneity is difficult due to logistical difficulties. Additionally, the manipulations usually apply perturbations at a level of intensity different from the one produced by the real event under investigation. Because of all these factors, ecologists need to asses how effective are large-scale manipulations disentangling the mechanisms behind observed trends which are then going to be used for prediction. This assessment is particularly important in the context of tropical regions where the implementation and maintenance of continuous monitoring activities are constrained by limited financial support in remote areas with difficult access. We anticipate that this symposium will serve as a focal point for ecologists to discuss and evaluate the contributions and limitations of experiments involving large-scale manipulations and their role producing process-level understanding of ecosystem trends. We will cover a range of perspectives from population, community and ecosystem processes, to assess what we learn from this kind of experimentation. The first four talks will present detailed description of large-scale manipulative experiments in terrestrial ecosystems conducted along Panama and Brazil. The next three talks will offer specific examples of uncovered processes from a large-scale experiment in Puerto Rico. The eighth talk will present an example from the perspective of ecologists trying to manipulate mobile organisms. The final talk will focus on aquatic ecosystems to describe how the large-scale, inter-site LINX experiment, contributed to study N saturation in a tropical forest in Puerto Rico. The last 30min will be reserved for a panel discussion and to provide opportunity for audience feedback.