Communicating the uncertainty in epidemic forecasts is particularly challenging due to the stochastic nature of infectious disease dynamics. In the 2023-2024 influenza projections from the Scenario Modeling Hub (SMH) in the United States (US), modeling teams transitioned from submitting quantiles to providing full sample trajectories, offering new opportunities to analyze the shape and variability of epidemic trajectories. However, existing visualization methods often fail to effectively convey the full range of uncertainty and key dynamic features, such as trajectory peaks. To address this limitation, I implement a curve-based visualization pipeline that applies ranking statistics to represent ensembles of projected trajectories at the state level and across age groups. While this method effectively summarizes overall trajectory uncertainty, it does not explicitly capture key peak characteristics. To bridge this gap, I extract and visualize peak magnitudes and timings across models to enhance the representation of uncertainty in model outputs. To complement these visual tools, I develop a quantitative evaluation pipeline that employs proper scoring rules and pairwise model comparisons to assess peak prediction accuracy and inter-model agreement. My findings reveal that although ensemble projections can encompass the true peak observed in surveillance data, their widespread dispersion results in low confidence in peak predictions across models. This work establishes a structured pipeline for improving peak-specific evaluation and uncertainty communication in multi-model infectious disease scenario modeling.