Quantifying the transport and fate of plastic debris in rivers is vital for implementing effective strategies to mitigate plastic emissions to the ocean. We need strategies to dynamically monitor plastic pollution sources and sinks. Models used to estimate plastic litter exports from global rivers to the ocean presently lack sufficient empirical data on riverine plastic transport. Tracked floating macroplastic items can provide insight into the retention and re-mobilization phases of litter in rivers. We provide the first cross-site study using identical methodology that measures riverine macroplastic transport and retention at a continental scale. We measured transport along 30 river reaches across 7 regions or watersheds in the conterminous United States and Canada. Cohorts of 10-12 trackers were monitored for multiple months between autumn 2022 and spring 2024. Mobilization events and distances were related to instantaneous discharge observations at nearby stream gauges and numerous independent variables describing reach geomorphology and upstream land-cover. We find that mobilization events can be predicted by specific stream power and the antecedent time a tracker remained lodged prior to a flow event. We leveraged gradient boosted regression trees to infer common metrics of importance in describing the distance macroplastic trackers moved following re-mobilization. Mobilization distance appears far more variable between sites than probability of re-mobilization. Average discharge while in transit was a significant predictor, upstream forest cover had significantly less importance, and other predictors describing upstream land-cover and channel characteristics exhibited diminishingly less importance. Building on the evident parameter dependencies, we are building a probabilistic model of litter transport to estimate measures of expected travel distance and residence time under prevailing hydrologic conditions with the goal of informing communities on appropriate cleanup strategies.