A lidar (Light Detection and Ranging) is an active remote sensing instrument which mea- sures range-resolved parameters of the atmosphere. Figure 2.6 illustrates the basic principle of lidar remote sensing: A laser emits very short linearly-polarized laser pulses (typically in the order of several nanoseconds), a telescope collects the backscattered laser light, and electronics is used to sample the collected light in high temporal resolution (typically 10 to 20 MHz). From the travel time of the light t and the speed of light c0, the distance R
2.6 Remote sensing techniques 29 laser telescope electronics aerosol particles pulsed laser beam backscattered laser light scattering lidar air
Figure 2.6.: Basic principle of lidar remote sensing of aerosols.
2.6.1.1. Elastic backscatter lidar
The simplest type of lidar is an elastic backscatter lidar. The wavelengths of the laser and the detector are the same. The lidar signal P(R) is described by the lidar equation (Wandinger, 2005)
P(R) = K·O(R)·R−2·(β(R) +βm(R))·T(R), (2.48)
whereR is the distance to the scattering object, K the system constant, O(R) the overlap function, β(R) and βm(R) the backscatter coefficients of particles and molecules, and
T(R) the transmission term. T(R) is calculated from the extinction profiles of particles and molecules by T(R) = exp −2· Z R 0 (αext(r) +αext,m(r))dr . (2.49)
The overlap functionO(R) determines which fraction of the laser beam is within the field- of-view of the telescope. Typically, lidar signals are used only for R where O(R) = 1 (full overlap). The extinction coefficient αext(R) is the quantity of interest for meteorological
applications. A backward integration scheme, as proposed by Fernald (1984), can be applied to the lidar profile P(R) to get the height-resolved aerosol extinction coefficient
meteorological data (e.g., from radiosondes) for the calculation of βm(R) and αext,m(R),
an atmospheric layer with knownβ(R), as for example an aerosol-free layer, and the lidar ratio of the particles S(R) =αext(R)/β(R). The lidar ratio S of the aerosol particles is
critical becauseS varies approximately from 10 sr to 100 sr, depending on the aerosol type. Additional information about aerosol particles can be obtained by measuring the polar- ization state of the backscattered light. For example, measurements of the linear depolar- ization ratio δl provide information about the sphericity of the particles.
2.6.1.2. Raman lidar and High Spectral Resolution Lidar
Advanced lidar techniques were developed, which observe additional parameters that allow one to derive aerosol extinction coefficients αext without the need to assume a lidar ratio
S. Raman lidars and High Spectral Resolution Lidars provide such parameters.
Like a simple backscatter lidar, a Raman lidar measures the signal at the laser wave- length; but, in addition, it detects photons that were backscattered by air molecules at other wavelengths due to Raman effects (Ansmann et al., 1992). Usually, Raman scat- tering of nitrogen (N2) or oxygen (O2) is detected by Raman lidars because their mixing
ratio in the air is well-known. Only the vertically-resolved air density is needed as an addi- tional parameter for the derivation of the αext-profile from Raman measurements. The air
density can be derived from radiosonde data or from standard atmospheres. A drawback of the Raman lidar is that the backscattered intensity at the Raman-shifted wavelength is approximately three orders of magnitude lower than at the emitted wavelength. Thus, long measurement times are necessary to retrieve accurate atmospheric profiles; moreover, Raman measurements are typically only possible at night.
Another type of lidar that allow one to derive αext without the assumption of a lidar
ratio S is the ”High Spectral Resolution Lidar” (HSRL, Eloranta (2005)). The principle of a HSRL is that it measures the signal from the aerosol particles and the signal from the molecules separately. These signals can be separated because the signal from aerosol particles has a very small spectral width compared to the signal from air molecules. This difference of spectral widths is a result of Doppler effects and differences of the thermal velocities of molecules and aerosol particles: the thermal motion of small air molecules is much faster than the motion of comparatively large aerosol particles. Optical filters or absorption cells with very small bandwidths are used to separate the aerosol signal from the molecular signal. Since both signals are comparatively high, HSRLs can be operated also at daytime.
2.6.1.3. Instrumentation and observed properties
Two lidar systems are available at the Meteorological Institute of the Ludwig-Maximilians- Universit¨at (LMU) in Munich: MULIS (multi-wavelength lidar system, e.g., Freudenthaler et al., 2009) and POLIS (portable lidar system, e.g., Groß et al., 2008). MULIS is a Raman- and depolarization-lidar including channels for elastic backscattering at 355 nm, 532 nm, and 1064 nm, and corresponding Raman channels for the determination of the extinction
2.6 Remote sensing techniques 31
coefficient at 355 nm and 532 nm. The linear depolarization ratio of particles is derived at 532 nm. POLIS is a small low-power two-channel lidar for Raman or depolarization measurements at 355 nm. As a consequence, combining the measurements from both lidars provides depolarization ratios at two wavelengths. The optical design of both lidars is optimized for measurements in the troposphere; the full overlap (O(R) = 1) of MULIS is reached in about 200 m to 400 m depending on field stop adjustments, and full overlap is reached in approximately 70 m for POLIS (Groß et al., 2008). Vertical profiles of the extinction coefficientαextand backscatter coefficientβof the aerosol particles atλ= 355 nm
and 532 nm are derived using the Raman approach (Ansmann et al., 1992). Vertical profiles of the other optical parameters, i.e. the linear depolarization ratio δl at λ= 355 nm and
532 nm, andβatλ= 1064 nm, are derived using the approaches described by Freudenthaler et al. (2009) and Fernald (1984), respectively.
In addition to MULIS and POLIS observations, observations of desert dust aerosols from two other lidar systems have been used for this work: The BERTHA lidar system of the Leibniz Institute for Tropospheric Research in Leipzig is also a Raman- and depolarization- lidar which provides, among other particle properties, the linear depolarization ratio δl at
λ= 710 nm (Tesche et al., 2009; Tesche et al., 2011). The high spectral resolution lidar (HSRL) of the German Aerospace Center, operated onboard the Falcon aircraft provides the lidar ratio S and the linear depolarization ratioδl at 532 nm, as well asδl at 1064 nm
(Esselborn et al., 2009).