To attain more sustainable aviation, it is required to lower its environmental impact. The corresponding effects typically appear in exhaust plume and contrail. Hence, investigating the interaction of turbine exhaust and aircraft wake is of prime importance. To this end, an accurate simulation of the exhaust-to-air mass fraction's distribution allows to detect the area invested by the plume and the amount of pollutant released in the atmosphere. Accurate plume models require a multi-species flow simulation, which can be hard and costly to solve. Indeed, if a reasonable grid convergence is not reached, the area of the plume will be drastically underestimated. Similarly, the interface between plume and outer flow must be resolved. However, meshes for aircraft applications are optimized to capture the flow around the aircraft, not downstream of it. Adaptive mesh refinement (AMR) constitutes a cost-efficient strategy to avoid re-meshing, as showcased by this work on industrially available configurations and corresponding grids. This contribution compares feature-based indicators with classical refinement techniques, such as global refinement and the more general residual-based refinement, and it investigates its effectiveness at capturing exhaust plumes. This work tests the procedure on an exemplary aircraft configuration and refines a predetermined ratio of cells over multiple solution-refinement iterations. Our heuristic refinement indicators are feature-based. Tailored to our application, we employ the gradient of the exhaust-to-air mass fraction, scaled by the local mesh size. As solver, we employ a two-species simulation with the flow solver HyperCODA, an extension of the modular flow solver "CFD by ONERA, DLR and Airbus" (CODA) that accounts for multi-species flows and chemical reactions. The code is used in the larger software integration framework FlowSimulator, thus enabling the combination with a library for hierarchical mesh adaptation, FSAdaptationNG. [1] Huismann I., Fechter S., Leicht T., HyperCODA – Extension of Flow Solver CODA towards Hypersonic Flows, 22nd STAB/DGLR Symposium on New Results in Numerical and Experimental Fluid Mechanics XIII, Vol. 151, pp. 99–109, 2021. [2] Hoechel M., Hafemann T., Orlt M., Sanchez Ortiz R.I., Algorithms for Geometry-Adapted Mesh Refinement with Cell Stack Preservation in the FlowSimulator, Conference for Adaptive Mesh Refinement and Applications 2025 (AMR2025), 2025.