Oil spills cause long-lasting mangrove loss, threatening their conservation and ecosystem services worldwide. Oil spills impact mangrove forests at various spatial and temporal scales. Yet, their long-term sublethal effects on trees remain poorly documented. Here, we explore these effects based on one of the largest oil spills ever recorded, the Baixada Santista pipeline leak, which hit the mangroves of the Brazilian southeastern coast in 1983. Historical, Landsat-derived normalized difference vegetation index (NDVI) maps over the spilled mangrove reveal a large dieback of trees within a year following the oil spill, followed by a eight-year recolonization period and a stabilization of the canopy cover, however 20-30% lower than initially observed. We explain this permanent loss by an unexpected persistence of oil pollution in the sediments based on visual and geochemical evidence. Using field spectroscopy and cutting-edge drone hyperspectral imaging, we demonstrate how the continuous exposure of mangrove trees to high levels of pollution affects their health and productivity in the long term, by imposing permanent stressful conditions. Our study also reveals that tree species differ in their sensitivity to oil, giving the most tolerant ones a competitive advantage to recolonize spilled mangroves. By leveraging drone laser scanning, we estimate the loss of forest biomass caused by the oil spill to be 9.8-91.2 t ha^-1, corresponding to 4.3-40.1 t C ha^-1. Based on our findings, we encourage environmental agencies and lawmakers to consider the sublethal effects of oil spills on mangroves in the environmental cost of these accidents. We also encourage petroleum companies to use drone remote sensing in monitoring routines and oil spill response planning to improve mangrove preservation and impact assessment.

Mangroves are among the most ecologically valuable ecosystems of the globe. Reliable remote sensing solutions are required to assist their management and conservation at broad scale. Canopy gaps are part of forests' turnover and rejuvenation, but yet no method has been proposed to map their occurrence and recovery in mangroves. Here, were propose an approach based on a deep learning framework called Mask R‐CNN to achieve automatic detection and delineation of gaps using very‐high‐resolution satellite imagery (<1 m). The Mask R‐CNN combines a series of neural network architectures to identify and delineate gaps, determine their recovery stage, and estimate their morphological attributes. The approach was tested on four mangroves from different regions of the globe with high concentration of gaps of various origins (lightning strikes, oil spills, cutting, pests). The Mask R‐CNN performed well to detect gaps, and accurately delineated gap contours (F1‐score of segmentation ≥0.89). The model also succeeded in distinguishing among five recovery stages of gaps, from their onset to closure (Overall Accuracy = 91.4, Kappa = 0.89). Accurate retrieval of gap area, eccentricity, and compactness – three relevant morphological attributes – were obtained (R² ≥ 0.83, NRMSE ≤10%). Several sources of confusion and misdelineation were identified. Our approach shows promising transferability to other mangrove sites and optical sensors and could help monitor canopy recovery in mangroves. It also opens promising perspectives for identifying the origin of gaps (natural or human‐induced). It is intended to assist environmental managers and field experts in the management and conservation of these fragile ecosystems.