Spontaneous vegetation monitoring as key soil bio-indicators in mangrove rice production agroecologies in Guinea Bissau

Abstract Aims Mangrove rice production (MRP) is essential for food security and environmental sustainability among smallholder farmers in Guinea-Bissau. Distinct agroecologies within MRP systems are shaped by specific soil–water conditions and by the spontaneous vegetation (weeds) colonizing rice fields. This study evaluates the potential of dominated weed species as bio-indicators of soil fertility and organic matter dynamics across contrasting mangrove rice agroecologies. Methods Approximately 300 soil samples were collected during 2022 and 2023 and analyzed for carbon (C) and nitrogen (N), and stable isotopes (δ 13 C and δ 15 N). In parallel, 155 weed species were surveyed using random field trials (2022) and transect-based methods (2023). UAV multispectral imagery (Phantom 4) and random forest classification were applied to assess weed spatial distribution and species dominance. Results Weed distribution followed a clear agroecological gradient, with Echinochloa colona dominating Associated Mangrove fields and Blutaparon vermiculare prevailing in Tidal Mangrove fields. UAV classification achieved up to 88% accuracy, although performance decreased in densely vegetated fields. Soil isotopic composition revealed distinct agroecological patterns, with higher δ 15 N values and C ratios in tidal mangrove agroesystems, suggesting enhanced nitrogen mineralization and faster organic matter turnover. δ 13 C signatures reflected differences in organic matter sources and turnover dynamics between agroecologies. Isotopic analyses clearly distinguished C₃ and C₄ vegetation, while C 3 species showed significantly higher nitrogen content than C₄ species, indicating contrasting nutrient-use strategies. Conclusions Spontaneous vegetation provides a robust proxy for assessing coupled carbon and nitrogen dynamics in mangrove rice systems. Integrating isotopic and spatial approaches enable effective discrimination of agroecological conditions, though further refinement is needed to align these tools with farmers’ management practices.