Disturbance has long been recognized as a potent force shaping ecological communities. However, predicting the effects of disturbance requires anticipating how it directly and indirectly affects organisms by causing mortality and modifying resource availability. This body of work addresses how disturbance events and disturbance regimes shape stream communities through both of these mechanisms. I first examine how an extreme flood event altered resource availability and subsequently consumer resource use in montane Colorado streams. I found that flood intensity correlated with higher availability of low-quality detrital resources and that consumers relied more heavily on these highly available resources, regardless of resource quality or their feeding guild. I then propose and test a conceptual framework for predicting functional and taxonomic diversity responses to disturbance events along disturbance intensity gradients. Using a three-year dataset of community composition from prior to and following an extreme flood event in Colorado, I found that stream insect communities that experienced higher intensity flooding experienced higher functional turnover, with a muted taxonomic response. Next, I experimentally tested whether colonization and dispersal strategies of macroinvertebrate communities varied with habitat stability in montane streams in Ecuador. I found that at less stable streams, functional and taxonomic richness of colonizing taxa and background benthic communities was similar, while colonization composition was species-poor compared to the available species pool at more stable sites. However, by the end of the colonization experiment, composition still did not closely resemble background benthic or drift composition, regardless of local habitat stability. Finally, I assess how landscape scale disturbance regimes influence local habitat stability and community stability. By modeling landslide susceptibility in the Quijos River drainage in Ecuador and correlating physical habitat variables with community composition across ten streams and five years of sampling, I find that local habitat stability best predicts interannual changes in multivariate community composition. I also find that while functional community composition remained constant across years, communities became taxonomically and phylogenetically dissimilar over time, regardless of local habitat stability. I conclude that using multiple measures of biodiversity can distinguish the mechanisms driving community responses to specific disturbance events and disturbance regimes.