Hydrothermal plumes play a crucial role in vent fields by injecting significant buoyancy flux from centimeter-scale vents and rising hundreds of meters, yet their near-field dynamics remain poorly understood. Using a Large-Eddy Simulation approach with adaptive mesh refinement, we study these plumes at centimeter-scale resolution within a 6 m domain above the vent. We first study a typical black smoker in the forced plume regime to quantify the mean flow and spatial variability, investigating the link between turbulent structures and the entrainment rate $\alpha$ of surrounding water. Significant vertical and temporal variability in $\alpha$ is observed, with an overall value of $\alpha = 0.19$. The results are compared with in-situ data and plume theory. Next, we investigate the sensitivity of the flow field to source parameters, characterizing the transition between forced or lazy plume regimes and the pure plume regime. In the far-field, plumes achieve self-similarity and the flow field is consistent with theoretical plume scalings, showing a dependence on source buoyancy flux for predictions in this region. The extent of the transition region where plume self-similarity breaks down is defined, and its importance in the context of in-situ observations is highlighted. Finally, we show that extreme temperatures above 100{degree sign}C occur in the first two meters of the column, but exposure times for a proxy tracer are short, suggesting that hydrothermal plumes could serve as viable transport vectors for biological materials.