Validation of EarthCARE-Derived Planetary Boundary Layer Height Using the E-Profile Ceilometer Network and Radiosondes

The Earth Clouds, Aerosol and Radiation Explorer (EarthCARE), launched in May 2024 as a joint ESA–JAXA mission, provides vertically resolved observations of aerosols and clouds with unprecedented sensitivity from space. In this study, we exploit measurements from the Atmospheric Lidar (ATLID), a high-spectral-resolution lidar operating at 355 nm, whose enhanced signal-to-noise ratio and capability to separate molecular and particulate backscatter enable detailed characterization of the lower troposphere (Wehr et al., 2023). These features make ATLID particularly suitable for deriving the planetary boundary layer height (PBLH) at the global scale.

The PBL is the atmospheric layer most strongly influenced by surface forcing through turbulent exchanges of heat, moisture and momentum. Accurate estimates of PBLH are therefore essential for weather forecasting, climate modelling and air quality studies. Previous spaceborne lidar missions, notably CALIPSO, demonstrated the feasibility of PBLH retrievals from aerosol backscatter profiles, although with limitations related to signal attenuation, cloud contamination and retrieval robustness (McGrath-Spangler and Denning, 2012). EarthCARE’s ATLID offers enhanced capabilities to address these challenges.

We validate ATLID-derived PBLH using independent ground-based observations from the E-Profile network, comprising over 400 ceilometers across Europe, along with collocated radiosonde measurements from the University of Wyoming Upper Air Soundings database. A continental-scale reference dataset was generated by applying the STRATfinder algorithm to ceilometer aerosol backscatter profiles. Planetary boundary layer heights from radiosondes were independently estimated using several thermodynamic and dynamical approaches, including the bulk Richardson number, the parcel method, and gradient-based criteria applied to temperature and humidity profiles. Only radiosonde launches collocated with E-Profile stations were considered, ensuring spatial consistency among the reference datasets. The analysis includes 580 collocated cases, defined as EarthCARE overpasses within 20 km of a ground-based station, from which 25 correspond to radiosonde observation, covering the period from August 2024 to August 2025.

Two complementary approaches were assessed to retrieve PBLH from ATLID Level-2 BA baseline products. The first approach used the operational A-ALD product, which includes PBLH as a retrieved variable. The product showed limitations, with misidentification of cloud layers as the PBL and a lack of retrievals under favourable conditions. These results underline current shortcomings of A-ALD for PBL detection, while indicating potential for future algorithm improvements.

The second approach applied combined variance–gradient methods to attenuated backscatter profiles from the A-EBD product, supported by cloud screening using the A-FM product. This strategy allowed more robust and physically consistent PBLH estimates. The comparison with ground-based ceilometer references resulted in a standard deviation of 343 m and a mean bias of 101 m. The nearly symmetric uncertainty distribution highlights the reliability of this approach. Radiosonde-based results showed a clear dependence on the retrieval method, with the best performance obtained for gradient-based approaches, although their statistical representativeness is limited by the small number of available cases.

These findings highlight the capability of EarthCARE’s ATLID to capture the PBL from space for climatological and modeling applications. The validation also emphasizes the importance of networks such as E-Profile, which provide the necessary reference data to evaluate satellite-derived boundary layer products on a continental scale.