Abstract Managing coastal development requires a set of tools to adequately detect ecosystem and water column degradation, but it also demands tools to detect any post-disturbance improvement. Structural seagrass indicators (such as shoot density or cover) are often used to detect or assess disturbances, but while they may be very sensitive to the impact itself, it is unclear if those indicators on their own can effectively reflect recovery at time scales relevant to managers. We used the construction of a harbour affecting a nearby Posidonia oceanica seagrass community to test the ability of a set of indicators (structural and others) to detect alterations and to evaluate their sensitivity to recovery of environmental quality after harbour construction was complete and the disturbance ceased. We used a Beyond Before After Control Impact (BBACI) design to evaluate effects on one impacted and three control meadows where we used structural, morphological, community and physiological indicators (26 in total) to asses disturbance impacts. Additionally, we measured some of the potential environmental factors that could be altered during and after the construction of the harbour and are critical to the survival of the seagrass meadow (light, sediment organic matter, sediment accrual). Harbour construction caused a clear increase in sediment organic matter and in sediment deposition rates, especially fine sand. Light availability was also reduced due to suspended sediments. Sediment and light conditions returned to normal levels 5 and 15 months after the construction began. As expected, seagrass structural indicators responded unequivocally to these environmental changes, with clear reductions in shoot density. Additionally, reduced light conditions quickly resulted in a decline in carbohydrate content in affected meadows. Unexpectedly, we also recorded a significant increase in metal content in plant tissues. No response was detected in the physiological indicators related to eutrophication (e.g. N and P content in tissues) and in morphological (shoot biomass) and community (epiphyte biomass) indicators. More than three years after the completion of the harbour, structural indicators did not show any sign of recovery. In contrast, physiological indicators, mainly heavy metal and carbohydrates content, were much better in detecting the improvement of the environmental conditions over the fairly short period of this study. These results indicate that while structural indicators are critical to evaluate the immediate effect of disturbances and the recovery on impacted systems, specific physiological indicators may be much better suited to determining the timing of environmental quality recovery. The design of impact and monitoring protocols in the wake of coastal developmental projects need to consider the differential effectiveness and time–response of measured indicators carefully.