* Dispersal and colonisation are key processes determining species survival, and their importance is increasing as a consequence of ongoing habitat fragmentation, land-use change and climate change. Identification of long-distance dispersal events, including upstream dispersal, and of the dispersal mechanisms and resulting spatial dispersal patterns involved provides much-needed information for conservation in an era of rapid environmental change. * However, quantifying contemporary dispersal among populations is far from straightforward. We used the relatively well-defined, typically linear, spatial structure of streams, rivers and their associated riparian and aquatic plant populations to illustrate this. We performed a literature review on studies where dispersal and its directionality (upstream versus downstream) were explicitly quantified. * Upstream dispersal was detected in the majority (75%) of examined stream and riparian plant species and mediated mainly by waterfowl, but also by other animals and wind. However, upstream movements are generally less frequent than downstream. Upstream dispersal can occur in excess of tens and sometimes even hundreds of kilometres. * Most of the reviewed studies suffer from important methodological limitations that generate difficulties in detecting uncommon dispersal events. Major limitations include use of molecular ecological analyses based on unrealistic assumptions, and the inability to separate seed from pollen flow. On the basis of these findings, we outline a flexible research design using DNA-based assignment methods that allows quantification of contemporary dispersal in future studies. We suggest four key improvements: (i) assignment of propagules and/or seedlings; (ii) use of spatial models to inform sampling design; (iii) reducing the influence of unsampled populations and (iv) combined use of nuclear and uniparentally inherited DNA markers to separate gene flow (including pollen and sperm) in general from propagule-mediated dispersal. In combination with direct measurements of seed dispersal these facilitate empirical quantification of dispersal and the detection of uncommon dispersal events, allowing more realistic assessment of spatial population dynamics, relevant for sedentary and relatively immobile organisms.