We employed the trophic upsurge hypothesis as an analytical framework to describe the response of Neotropical fish communities to terrestrial inundation associated with river damming. Data were collected for the impact assessment of the Peixe Angical Dam, Tocantins River, Brazil. Monthly surveys were conducted at nine sites (seven upstream, two downstream) from 15 months before to 20 months after dam closure and a follow‐up survey 60–83 months after closure. Fish responses differed in timing and magnitude across sites. In the new reservoir, populations peaked 8–13 months after flooding, demonstrating increases in richness (84 ± 52%), biomass (329 ± 266%) and abundance (681 ± 348%) with the maxima for richness and abundance tending to precede biomass maxima. Populations of all families increased but were greatest for Engraulidae, Hemiodontidae, Serrasalmidae and Characidae, and lower for the benthic catfish: Sciaenidae, Dorididae, Pimelodidae. Downstream populations peaked 5–12 months after closure or showed consistent decline. Five years after dam‐closure richness and abundance were lower compared to predam levels, with populations at all sites demonstrating an ongoing decline. Reservoir creation triggers reproductive recruitment and otherwise pools spatially disparate populations from across the flooded valley. If the rise in richness is largely explained by the increased catchability of species otherwise present prior to flooding, it may be more appropriate to estimate species loss with reference to upsurge data. As peak biomass coincides with declining richness, modelling fisheries production from annually aggregated data may risk overestimating the potential for sustainable harvests. Upsurge–response curves can help identify the timing of critical ecological thresholds for flood‐managed fisheries.