The Impact of Small-Amplitude Perturbations to the Temporal Scales of Tracer Predictability in the Surface Layer over Urban Environment

The predictability horizon of dispersion of passive tracer is of both theoretical interest and practical importance. Despite extensive research on urban dispersion and atmospheric predictability, significant knowledge gaps still exist in the predictability of passive tracer dispersion in an urban surface layer. Using an obstacle-resolving large eddy si­mulations, we investigated the predictability of tracer by conducting twin experiments under both neutral and convective conditions. Passive tracer is released in the domain at a constant rate. Volume error between the twin experiments increases with time, indicating reduced predictability. The rate of error increase is also computed to characterize how predictability evolves at different stages. We further analyzed the empirical probability distribution function (PDF) of the tracer concentration. The change of PDF of tracer with time at different locations are assessed based on different metrics to characterize when the twin experiments diverge. Based on these objective metrics, we developed a novel framework to characterize the predictability horizon. With increasingly high spatiotemporal resolution in modeling tracer dispersion in the urban environment, this work highlights the importance of quantifying the predictability horizon.