Soil degradation hampers stability and functioning of soil ecosystems. Among various soil threats, salinization represents a critical abiotic stress, especially in Mediterranean area, which can compromises land quality and limit plant productivity. To address this issue, the implementation of sustainable nature-based solutions represents an important strategy for soil quality restoration. Organic amendments such as composts can be good soil improver and/or fertilizers depending on their composition. In fact, their effectiveness is highly dependent on feedstock origin (e.g. manure, organic fraction of municipal solid waste and yard trimmings) and production process (e.g. from very small to large scale waste management capacity). This study evaluated four decentralized compost capacity of mitigating salinity stress in a soil microbiome-plant system. Moreover, considering a holistic One Health approach, the potential spread of four antibiotic resistance genes (ARGs) and a mobile genetic element was analysed. Various soil pot experiments were performed with Rosmarinus officinalis and amended with one of the four composts: 1. community compost (CC, waste ca. 20 t/y), 2. decentralized urban compost (DUC, waste ca. 750 t/y), 3. small-scale agro-compost (SA, waste ca. 30 t/y) and 4. medium-scale agro-compost (MA, waste ca. 500 t/y).

Plant biomass, rhizosphere microbial abundance, activity and ARGs were evaluated. Finally, a Soil Quality Index was calculated. Results indicated that salinity stress significantly reduced plant biomass. However, CC and DUC amendments mitigated effectively rosemary biomass decrease, enhancing its growth. Interestingly, the rhizosphere microbial community increased its activity in response to salinity through a pre-existing osmotic defence mechanism, independent from compost presence. Salinity and compost generally increased ARGs in the soil, however the community compost (CC) proved to be the most effective in improving overall soil quality and salt tolerance, confirming that localized processing models (involving a very small number of people) significantly improve compost quality and safety.