Atmospheric metal burden across urban land-use settings: Laboratory calibration and field evaluation of living-moss multimodal biosensing

Atmospheric metal monitoring remains constrained by the sampling and analytical requirements of conventional methods. This study developed MossSense-HM, a living Hypnum cupressiforme-based multimodal sensing platform combining electrical impedance, optical reflectance, chlorophyll fluorescence, and microclimatic measurements to estimate accumulated metal burden. Laboratory exposures to Pb, Cd, Cu, Zn, Ni, and Cr at 25–200 μg cartridge􀀀 1 over 72 h produced exposure-dependent responses. Electrical changes were detectable within 6–12 h at the two highest doses, providing an indication of stress before endpoint chemical analysis. Including temperature, relative humidity, and moss moisture improved XGBoost prediction of ICP-OES-derived metal burden, achieving test R2 = 0.91 compared with 0.63–0.76 for individual sensing modalities. The Moss Biosensor Pollution Index (MBPI) correlated with measured metal burden (r = 0.89) and showed 86.5% category agreement (109/126) across laboratory observations, including calibration data. Independent evaluation at 30 sites during a 30-day deployment in Çankırı yielded R2 = 0.84 (95% bootstrap CI: 0.73–0.91). Field MBPI categories agreed with training-derived ICP-OES pressure categories at 83.3% of sites (25/30; Wilson 95% CI: 66.4–92.7%). Traffic-, railway-, and industrial-influence sites showed higher responses and metal burdens than background environments. These findings support living-moss sensing as a complementary screening approach for prioritizing chemical analysis. Model performance and MBPI boundaries remain specific to the evaluated species, calibration conditions, and deployment period; transfer across cities, seasons, and species requires further validation.