Work

Uncertainty-weighted fuzzy clustering disentangles climate-stress VOC fingerprints and associated stress-dependent temperature sensitivity in urban trees

Year: 2026
Type: preprint
Abstract: Climate change increases the severity and frequency of heat and oxidative stress that plants experience, yet their simultaneous effects on biogenic volatile organic compounds (BVOCs) emissions remains poorly understood, complicating the predictability of future air quality and climate. Here, canopy BVOC emissions from linden and chestnut trees were measured over a continuous two-month chamber experiment without stress, heat stress, and combined O₃ and heat stress. To untangle stress-associated BVOC fingerprints, we developed an unsupervised uncertainty-weighted fuzzy c-means (UWFCM) algorithm that directly incorporates compound- and measurement-specific analytical uncertainties into clustering. UWFCM identified monoterpenes (increased up to 3-fold) and sesquiterpenes (up to 22-fold) as the primary heat-stress markers in both species, while increased diterpenes characterized combined O₃ and heat stress in linden. Stress-induced de novo emissions resulted in carbon loss up to ~ 14% of net daytime carbon fixation, a substantial stress-related carbon reallocation. The temperature sensitivity of terpenoid emissions was strongly stress-dependent. These findings have important implications for global BVOC parameterizations, estimates of daytime NO x -driven ozone chemistry and secondary organic aerosol formation under future climate extremes.
Language: English
Open Access status: green
OpenAlex ID: W7226009781
Cites: 80
Cited by:
Related to: 10