Abstract. With the rapidly growing number of automated single-wavelength backscatter
lidars (ceilometers), their potential benefit for aerosol remote sensing
received considerable scientific attention. When studying the accuracy of
retrieved particle backscatter coefficients, it must be considered that most
of the ceilometers are influenced by water vapor absorption in the spectral
range around 910 nm. In the literature methodologies have been proposed to correct for this
effect; however, a validation was not yet performed. In
the framework of the ceilometer intercomparison campaign CeiLinEx2015 in
Lindenberg, Germany, hosted by the German Weather Service, it was possible to
tackle this open issue. Ceilometers from Lufft (CHM15k and CHM15kx, operating
at 1064 nm), from Vaisala (CL51 and CL31) and from Campbell Scientific
(CS135), all operating at a wavelength of approximately 910 nm, were
deployed together with a multi-wavelength research lidar (RALPH) that served
as a reference. In this paper the validation of the water vapor correction is
performed by comparing ceilometer backscatter signals with measurements of
the reference system extrapolated to the water vapor regime. One inherent
problem of the validation is the spectral extrapolation of particle optical
properties. For this purpose AERONET measurements and inversions of RALPH
signals were used. Another issue is that the vertical range where validation
is possible is limited to the upper part of the mixing layer due to incomplete
overlap and the generally low signal-to-noise ratio and signal artifacts
above that layer. Our intercomparisons show that the water vapor correction
leads to quite a good agreement between the extrapolated reference signal and
the measurements in the case of CL51 ceilometers at one or more wavelengths
in the specified range of the laser diode's emission. This ambiguity is due
to the similar effective water vapor transmission at several wavelengths. In
the case of CL31 and CS135 ceilometers the validation was not always
successful. That suggests that error sources beyond the water vapor
absorption might be dominant. For future applications we recommend monitoring
the emitted wavelength and providing “dark” measurements on a regular
basis.