Dam removal is a direct measure to restore river connectivity, but the assessment of its ecological effects has long been constrained by the lack of continuous monitoring data across multiple trophic levels. Based on eight years (2017–2024) of continuous field monitoring data from the Heishui River, a first-order tributary of the Jinsha River, this study systematically analyzed the spatiotemporal dynamics of phytoplankton, zooplankton, and fish communities before and after the removal of the Laomuhe Dam. Ecopath models were constructed annually for six ecological units to quantitatively assess the restructuring effect of dam removal on the structure and function of the aquatic food web. The results showed that: (1) Plankton communities were highly sensitive to the disturbance of dam removal. In the early stage after dam removal (2019), the diversity of phytoplankton and zooplankton across the whole river suffered a cliff-like decline; the Shannon index of zooplankton in the former reservoir area plummeted from 3.06 to 0.81. During the recovery period, diatoms and rotifers acted as pioneer groups for phytoplankton and zooplankton, respectively, and were the first to recover. By 2024, the proportion of copepods rose to the highest level during the study period, and the dominant group of the community shifted from rotifers to copepods. (2) Fish communities responded rapidly and positively to the restoration of connectivity. In the upstream reach adjacent to the dam site, the number of species increased from 3 before dam removal to 12 afterward; in the downstream reach, it increased from 2 to 15, with diversity index increases of 172% and 246%, respectively. (3) Ecopath model analysis revealed a recovery pattern of “structure first, function delayed” in the food web—species number and community composition could be re-established shortly after connectivity restoration, but the recovery of functional indicators such as the system omnivory index (SOI) and energy transfer efficiency (total primary production/total respiration, TPP/TR) lagged significantly behind. By the end of the study period, the upstream dewatered reach had not yet reached the level of the natural reference reach, and the downstream recovery rate and degree were better than those upstream. This study provides comprehensive time-series evidence for the long-term response of aquatic ecosystems to low-head dam removal and offers valuable references for guiding adaptive management of mountain rivers after dam removal.

Following Yangtze River protection policies, many small dams have been removed. Using the Heishui River as an example, we compared food web structures among areas affected by dams (DAs), areas from which dams were removed (DRAs), and natural reference areas (NRAs) using stable isotopes and a Bayesian mixing model. NRAs had balanced basal carbon sources; DAs relied on periphytic algae (43.6%) and POM (23.5%); and DRAs were dominated by periphytic algae (50.8%), with terrestrial inputs lower in both DAs and DRAs. NRAs showed a compressed food chain (1.89–2.23). DAs exhibited an extended but inverted structure, with benthivorous fish (3.00) surpassing carnivores, perhaps reflecting the accumulation of sedimentary organic matter under the lentic conditions created by damming. DRAs expanded the vertical dimension, with carnivorous fish reaching the highest level (3.70). DAs had the fewest nodes (16) but the highest connectance (0.492). DRAs showed the most nodes (20), highest linkage density (4.00), lowest connectance (0.421), and lowest trophic redundancy, indicating functional differentiation. NRAs had an intermediate number of nodes (17) and the lowest linkage density (3.41). These findings indicate that small dam removal altered basal resource use, triggered food web reorganisation that increased the trophic position of carnivorous fish, and enhanced topological complexity, although terrestrial energy inputs in DRA remained below natural levels.