Planetary boundary layer height retrieval from spaceborne lidar
The planetary boundary layer governs how pollutants disperse and how heat moves between surface and sky. Spaceborne lidar can estimate its depth globally, but sun-synchronous orbits, elevated aerosol layers and cloud contamination impose hard limits every analyst must understand.
Sensors
- CALIOP / CALIPSO: Two-wavelength (532 nm and 1064 nm) polarisation lidar. Vertical resolution 30 m below 8.2 km, horizontal averaging typically 5 km for boundary-layer products. Near-polar sun-synchronous orbit gives a 16-day exact repeat; local overpass near 13:30 and 01:30. Archive from June 2006 to August 2023.
- ATLID / EarthCARE: High-spectral-resolution lidar at 355 nm on ESA's EarthCARE satellite, launched May 2024. Separates molecular and particulate backscatter without a separate depolarisation assumption, improving aerosol-layer-top discrimination. Vertical resolution 100 m, horizontal 1 km along-track.
- CATS / ISS: Cloud-Aerosol Transport System operated on the International Space Station from 2015 to 2017. The ISS non-sun-synchronous orbit (51.6° inclination) sampled multiple local times over the same location within days, making it uniquely useful for studying diurnal PBL growth cycles that CALIPSO cannot resolve.
- Aeolus ALADIN: UV Doppler wind lidar at 355 nm, operational 2018 to 2023. Primarily a wind profiler, but its co-polar Mie and Rayleigh backscatter channels yield aerosol-layer-top estimates as a by-product. Horizontal resolution ~87 km; useful for synoptic-scale PBL height climatology rather than city-scale work.
What the backscatter gradient actually measures
The planetary boundary layer (PBL) is the lowest part of the troposphere, typically 200 m to 3 km deep, where turbulent mixing driven by surface heating and friction dominates. Aerosols, water vapour and heat are well-mixed within it and drop off sharply above. That sharp drop in aerosol concentration produces an equally sharp drop in the attenuated backscatter signal returned to a downward-pointing lidar. Algorithms identify this gradient minimum, usually by wavelet covariance transform or gradient methods applied to the 532 nm attenuated backscatter profile, and call it the PBL top.
The physics is sound in principle. In practice, the aerosol-layer top and the thermodynamic boundary-layer top coincide reliably only when the atmosphere is well-mixed and aerosol loading is moderate. The gradient method detects where aerosol concentration changes, not where turbulence stops. Over clean oceanic air, where backscatter contrast is weak, detection uncertainty can exceed 500 m. Over land after sunrise, when convective thermals are still forming, the gradient is diffuse and retrievals scatter widely.
The diurnal problem that one overpass cannot solve
CALIPSO crosses the equator at approximately 13:30 local solar time on its ascending node. That single snapshot catches the PBL near its daily maximum depth over most continental regions, which is useful for pollution-dispersion modelling but gives no information about the shallow nocturnal boundary layer or the morning growth phase. A city's peak-hour commute emissions disperse into a PBL that may be only 200 to 400 m deep at 08:00, not the 1,500 m the afternoon overpass records.
CATS on the ISS partially addressed this by cycling through local times over a given latitude band within roughly two weeks, producing multi-time-of-day samples that revealed the diurnal amplitude of PBL depth. That dataset ended in 2017. EarthCARE, also sun-synchronous, inherits the same single-overpass limitation. Honest use of spaceborne lidar PBL products therefore requires pairing them with mesoscale model output or surface meteorological data to reconstruct the diurnal cycle; the satellite constrains the model, not the other way around.
Elevated aerosol layers: the retrieval's persistent adversary
Saharan dust, biomass-burning smoke and volcanic ash frequently form aerosol layers that are decoupled from the surface mixed layer, sitting at 2 to 6 km altitude. A gradient algorithm that simply finds the strongest backscatter gradient in the profile will often return the base of an elevated dust layer rather than the true PBL top. CALIOP's depolarisation ratio helps distinguish non-spherical dust from spherical boundary-layer aerosols, and its 1064/532 nm colour ratio provides additional aerosol-type information, but the discrimination is imperfect when layers overlap vertically.
The consequence for users is systematic positive bias in PBL height over dust-affected regions, including the Sahel, the Arabian Peninsula and downwind oceanic areas. Published studies comparing CALIOP retrievals with collocated radiosonde profiles over the Sahara report biases of several hundred metres to more than 1 km during active dust events. ATLID's high-spectral-resolution capability should reduce this confusion by cleanly separating molecular from particulate backscatter, but the mission is still in its commissioning and calibration phase as of mid-2025.
Cloud contamination and the effective sampling fraction
Lidar is an active optical instrument: it cannot see through cloud. A single opaque cloud layer anywhere below the satellite extinguishes the beam and renders the profile below it useless for PBL retrieval. CALIOP's global cloud-free fraction at the surface level is roughly 30 to 40% of profiles, varying strongly by region and season. The tropics, where convective cloud is persistent, have particularly poor sampling. This is not a criticism of the instrument; it is a physical constraint any user must account for when building climatologies or near-real-time products.
Practical workarounds include temporal averaging over monthly or seasonal periods to build statistically representative PBL height maps, and spatial compositing across adjacent cloud-free orbits. Neither approach recovers the true instantaneous PBL height on a cloudy day. For operational air-quality applications that need daily PBL depth estimates everywhere, lidar retrievals serve as a quality-control anchor for numerical weather prediction output rather than a standalone observational product.
Building a usable PBL climatology from the CALIOP archive
Despite its limitations, the CALIOP archive spanning 2006 to 2023 is the most geographically complete observational record of PBL height available. Level-2 aerosol layer products are distributed through NASA's Earthdata portal. The standard product reports layer base and top heights for up to ten detected aerosol and cloud layers per profile, with quality flags for signal-to-noise ratio and feature classification. Extracting PBL height requires selecting the lowest aerosol layer whose base is within a few hundred metres of the surface and whose feature type classification is consistent with boundary-layer aerosol rather than elevated smoke or dust.
Regional climatologies built this way show physically coherent patterns: deep afternoon PBLs over the Sahara and Tibetan Plateau (often exceeding 3 km), shallow marine PBLs (typically 500 to 1,000 m) over stratocumulus regions off the Californian and Namibian coasts, and strong seasonal cycles over monsoon-affected South Asia. These patterns validate well against radiosonde climatologies where the two datasets overlap, which is the basis for using the satellite record in the much larger areas where radiosondes are absent. Satellize's analytics pipeline can ingest CALIOP Level-2 products and apply quality-controlled layer-selection logic to produce gridded monthly PBL height fields for client-specified domains.
EarthCARE ATLID data, once fully validated, will extend and improve this record. The 355 nm wavelength has higher molecular backscatter than 532 nm, improving signal contrast in clean air. The high-spectral-resolution channel removes the need to assume a lidar ratio for aerosol extinction correction, which propagates into more accurate attenuated backscatter profiles and, in turn, more reliable gradient detection near the PBL top.
Typical figures
| Vertical resolution (CALIOP, below 8.2 km) | 30 m native; 60 m in standard Level-1 products |
| Horizontal averaging (CALIOP aerosol layer product) | 5 km along-track (single-shot SNR insufficient for PBL detection) |
| Vertical resolution (ATLID / EarthCARE) | 100 m |
| Horizontal resolution (ATLID) | 1 km along-track |
| Revisit (CALIOP, exact repeat) | 16 days; local overpass ~13:30 and ~01:30 only |
| Cloud-free sampling fraction (CALIOP, global mean) | ~30–40% of profiles reach the surface |
| PBL height retrieval uncertainty (moderate aerosol, cloud-free) | ~100–300 m under favourable conditions; >500 m over clean ocean or during dust events |
| CALIOP archive depth | June 2006 – August 2023 |
| Wavelengths (CALIOP) | 532 nm (polarisation) and 1064 nm |
| Wavelength (ATLID) | 355 nm, high-spectral-resolution |
Analytics Satellize can run
| Monthly gridded PBL height climatology | Wavelet covariance transform or gradient method applied to CALIOP Level-2 aerosol layer products, with quality flags for feature type and SNR | GeoTIFF or NetCDF grid at 1° × 1° or 0.5° × 0.5°, per month, for client-specified domain and period |
| Elevated-layer contamination flag | CALIOP feature classification (depolarisation ratio + colour ratio) to identify profiles where the lowest detected layer is dust or smoke rather than boundary-layer aerosol | Quality-flagged CSV or GIS layer indicating retrieval confidence class per profile |
| PBL height anomaly relative to ERA5 model fields | Collocation of satellite retrievals with ECMWF ERA5 boundary-layer height; bias and RMSE statistics by region, season and aerosol type | Statistical summary report identifying where model and observation diverge most, informing model assimilation priorities |
| Diurnal PBL amplitude estimate (where CATS data available) | Multi-local-time compositing using CATS ISS profiles (2015–2017) to estimate morning-to-afternoon PBL growth amplitude | Regional diurnal amplitude maps for 2015–2017 reference period, delivered as NetCDF with methodology note |
| Air-quality model boundary condition dataset | Satellite-constrained PBL heights merged with NWP output using optimal interpolation; follows published assimilation approaches for CALIOP aerosol profiles | Daily gridded PBL depth fields formatted for WRF-Chem or CMAQ model input |
Who does the work
We can get this done for you. Satellize runs its own analyst desk and a strong science team. You do not buy a data feed and work out what it means; our people source the imagery, run the analysis described on this page, and hand you the answer with its confidence limits stated. Discuss this requirement.