Evaluating the atmospheric boundary layer height from ceilometer and Doppler lidar: Divergent retrievals across methods and the question of what they truly represent

The atmospheric boundary layer height (ABLH) is a key parameter for understanding turbulent exchange processes, air‑quality dynamics, and land–atmosphere interactions. Radiosonde profiles are traditionally used as a reference for determining the ABLH, but their sparse temporal coverage limits their value for continuous monitoring. Ground‑based remote‑sensing instruments, such as ceilometers and Doppler lidars, offer high‑frequency observations throughout the day, but the derivation of the ABLH from these systems depends strongly on the chosen retrieval method. In this study, we evaluate multiple commonly used algorithms for ABLH estimation, like gradient-based and variance-based methods using a threshold, all applied to co‑located ceilometer and Doppler lidar measurements. The resulting ABLH estimates are systematically compared against radiosonde-derived heights to assess performance under varying meteorological conditions.

Our analysis reveals substantial discrepancies between methods, both within and across instrument types. Ceilometer-based retrievals tend to diverge most strongly during conditions with weak aerosol gradients, at night and during the afternoon transition, while Doppler lidar methods show larger spread during periods with low signal due to weak winds. No single method consistently reproduces radiosonde-derived heights across all stability regimes. Instead, each approach captures different structural aspects of the boundary layer, suggesting that the ABLH is not a single, easily definable quantity, but rather a multifaceted feature of the lower atmosphere.

These findings raise an important question for the boundary layer community: Is the derivation of a robust ABLH from ground-based remote sensing fundamentally limited by the information content of individual instruments and methods, and do we ultimately require a synergistic, multi-sensor, multi-method product to obtain a physically meaningful estimate? This contribution will explore these challenges in detail and discuss pathways towards an integrated ABLH retrieval framework.