Epidemic models are used during infectious disease outbreaks to estimate disease parameters, examine the effect of interventions, and explore potential future outcomes. Predicting future disease trends can take the form of short-term forecasts or long-term scenario projections. Evaluating the performance of these projections is crucial for understanding their usefulness in guiding public health decision-making. My dissertation aims to enhance methods for probabilistic epidemic model evaluation and scenario analysis. In my first chapter, I introduce the energy score as a robust metric for evaluating predictions in their native stochastic trajectory format. The project introduces the energy score in the context of epidemic modeling, and draws comparison with commonly used scores for model evaluation with applications to scenario projections. In my second project, I propose the scenario ensemble as a novel ensemble method for scenario projections, which combines results across scenario assumptions to easily communicate the range of potential future outcomes. Using 10 rounds of COVID-19 Scenario Modeling Hub data, we show well-calibrated projections and improved performance for the scenario ensemble of multi-model ensembles. In the third chapter, I take a step back to retrospectively evaluate the performance of infectious disease forecasting challenges over the past decade using a relative skill score to assess the evolution of disease forecasting performance over time, which will be essential for improving future infectious disease forecasts. Generating and evaluating epidemic projections becomes more challenging in an emerging outbreak when surveillance is limited and disease properties are unknown. My fourth chapter aims to understand the controllability of global disease outbreaks through a theoretical scenario modeling exercise. I use a metapopulation network framework to describe the global spread of disease, with the goal of examining the impact of various public health interventions and disease characteristics on epidemic outcomes. Exploring the landscape of outbreak control is crucial for identifying effective responses in future emerging disease outbreaks.