Synonyms

Ladar; Laser radar; Optical radar

Definitions

Lidar. LIght Detection And Ranging.

Ladar. LAser Detection And Ranging.

Introduction

The term lidar (i.e., the optical analogue of radar) embraces a broad collection of active optical probing techniques for remotely determining the physical and chemical characteristics of diffuse or condensed matter targets, both natural and otherwise. Although lidar methods have been practiced since before the dawn of the laser era, their classification here as an emerging technology reflects the fact that they have only in relatively recent years begun to see systematic exploitation from space-based vantage points (most especially with the advent of robust, high-performance laser systems).

Some remarks on terminology are in order as a guide to those unfamiliar with this field. Although the terms lidar and ladar are essentially interchangeable (except for exceedingly rare instances in which an incoherent source is used and which would therefore exclude consideration of ladar), different application communities have preferentially adopted one over the other. Hence, the atmospheric science community uses the form lidar, while those working in areas that involve remote characterization of hard targets have adopted laser radar or ladar (McManamon, 2012). It is also possible to find instances where ladar and lidar appear in upper case, reflecting their origin as acronyms. This is becoming less common now that these terms (like radar before them) have entered general parlance, but the hybrid form LIDAR is nevertheless commonplace albeit used exclusively by the airborne laser terrain mapping community.

A number of textbooks are available that treat lidar systems and methodology from a variety of standpoints. Measures (1992) provides a thorough grounding in the basic theory and practice of laser remote sensing, while Fujii and Fukuchi (2005), Weitkamp and Walther (2005), and Kovalev and Eichinger (2004) collectively offer a comprehensive overview of the wide array of measurement applications that have benefitted from lidar. Grant et al. (1997) provide an anthology of seminal lidar papers from the earliest days, when arc lamps were the brightest sources available, up to the mid-1990s.

Lidar basics

A canonical lidar system comprises the set of subsystems depicted in Figure 1. Depending on the intended application, the laser transmitter may require stringent control of wavelength, spectral purity, polarization, pulse duration, or some combination of all of these properties. By corollary, the receiver system may also be required to analyze one or all of these same properties of the return radiation.

Figure 1
figure 14

Functional schematic of a generic lidar system.

Full size image

Following the detection stage, processing electronics convert the signals to a range-gated (i.e., time-of-flight registered) data stream which is then logged for further processing to extract the desired observable quantity. This is accomplished through an understanding of the often complex interaction phenomenology of the laser radiation with the target and frequently leads to a suboptimally constrained retrieval process which must be augmented by ancillary measurements, modeled parameters, or a combination of both. The optical power P(R) returned to the lidar detection subsystem by a scattering volume located at range R is given by the standard basic lidar equation:

$$ P(R)=\left[ {E\eta O(R)\frac{cA }{{2{R^2}}}} \right]\left[ {T(R){T}^{\prime}(R)\beta (\lambda, R)} \right], $$
(1)

where E is the pulse energy emitted into the atmosphere (outward transmission T, return transmission T′), A is the receiver collecting area, and β is the backscatter from the range element of interest at the specified emission wavelength, λ. η is the overall optical efficiency of the receiver subsystem and O(R) is the range-dependent transmit/receive overlap function. (In general, the optical efficiency represents the optical transmission. However, for a coherent detection lidar system, this parameter will also include a component due to heterodyne mixing efficiency (Zhao et al., 1990).) The distinction made between outgoing and return atmospheric transmission recognizes that the incident radiation in some cases will be inelastically scattered, as explained below. Terms grouped together in the first set of square parentheses in Equation 1 are lidar system-dependent parameters, while those contained in the second set of square parentheses characterize the atmospheric response. This expression also assumes single scattering. Many lidar systems are subject to multiple scattering to a greater or lesser extent, and this will complicate the data inversion procedure. Eloranta (1998) provides a discussion of these issues and techniques for their resolution.

While pulsed operation is implicitly assumed in Equation 1, other forms of range gating have used modulation of continuous wave (cw) laser sources, such as pseudorandom noise encoding (Takeuchi et al., 1986), in conjunction with synchronous detection of the return to extract the time-dependent signal behavior.

Most, if not all, lidars are operated in the backscattering or quasi-backscattering configuration in which the transmitter and receiver are collocated and form part of an integral instrument. When the transmit and receive optical axes are coaxial, the lidar is said to be monostatic, whereas if they are offset, the lidar is described as bistatic. Note that this nomenclature differs from that applied in the radar field in that a bistatic radar system comprises transmitter and receiver elements that are geographically separated, whereas a monostatic radar implies collocated transmitter and receiver systems that may or may not be coaxial. Multistatic lidar configurations comprising two or more receiver stations have been considered for certain specialized applications (e.g., Olofson et al., 2008), but thus far, none have been implemented beyond the experimental stage.

An additional distinction that is frequently used to classify lidar systems is the mode of photon detection employed. In a direct (incoherent) detection lidar, the received photons are converted directly to electrical impulses by the detector with the loss of all phase information, and this is the detection method found in the overwhelming majority of lidar systems. In a heterodyne (coherent) detection lidar, the received photons are mixed with a reference oscillator signal and detection is performed on the intermediate frequency. Such a configuration is technologically more complex than a direct detection system, and also makes considerably more stringent demands on laser frequency stability and beam quality, but preserves phase knowledge and is essential if the ultimate goal is to directly retrieve motion-induced Doppler signatures, such as for atmospheric wind measurement (Huffaker and Hardesty, 1996) or object tracking (Osche and Young, 1996). A few coherent detection lidars have been described that are non-Doppler capable (e.g., Grant et al., 1987; Menzies and Tratt, 1994; Gibert et al., 2006). In these cases the additional system complexity is tolerated because of the much greater noise rejection that is afforded by coherent detection, which can be as much as an order of magnitude above that of a comparable direct detection system (Menzies, 1976).

Inelastic backscatter lidar

Inelastic backscatter lidars make use of wavelength conversion properties of the target to infer details of its physical and chemical makeup. The chief downshifting processes that are commonly employed in the lidar field are Raman and fluorescence scattering, both of which require excitation in the UV-visible spectral range.

Although the weak cross sections for Raman scattering tend to limit its usefulness for long-range remote sensing purposes, this disadvantage is offset by the high specificity that it affords (because the observed wavelength shift is diagnostic of the molecule being probed). Hence, many users tolerate the need for a high-power transmitter, large collection aperture, and long accumulation intervals in order to gain the benefit of unambiguous measurements, which is clearly of central importance when such systems are used to validate satellite data products (Wessel et al., 2000; Tratt et al., 2005). Raman lidar has met with significant success in the ground-based remote sensing of atmospheric constituents and structure (Turner and Whiteman, 2006), atmospheric dynamics (Koch et al., 1991), and hazardous substances (Sharma et al., 2005) and has also seen limited airborne application (Heaps and Burris, 1996).

Fluorescence lidars are generally used for the investigation of chemical and biological targets. For example, laser-induced fluorescence has been successfully used in the remote assessment of plant health (Saito et al., 2000), airborne monitoring of phytoplankton populations and ocean productivity (Hoge et al., 2005), and the diagnosis of biodeterioration in building fabrics (Weibring et al., 2001). These techniques rely on measurement of the 690 nm chlorophyll-α fluorescence signature (sometimes in association with other features) and an understanding of the way it is modified by environmental disturbances. Fluorescence lidar has also been proposed for monitoring of marine oil spills (Karpicz et al., 2006) and the identification of airborne chemical agents or pathogens (Buteau et al., 2007).

So-called resonance fluorescence techniques, as the name implies, resonantly excite a target atomic species in order to directly probe the atmospheric properties at the residence altitude of the species in question, which are typically metal ions entrained in the middle and upper atmosphere (80–100 km altitude) that are derived from, and continually replenished by, meteoritic deposition. Hence, there is a long heritage of using sodium lidars to track atmospheric dynamics in these difficult-to-access regions (Gardner, 1989) as well as to retrieve upper atmospheric temperatures through measurement of the Doppler-broadened atomic lineshape, which can then be related directly to the ambient temperature. Other atomic metal species that have been studied for similar purposes are potassium (Papen et al., 1995) and iron (Kane and Gardner, 1993). More recently, iron Boltzmann lidars have been used to infer atmospheric temperature by measuring the spectral content of radiation emitted by resonantly excited mesospheric iron layer atoms (Gelbwachs, 1994; Raizada and Tepley, 2002). An extensive treatment of this field has been given by Chu and Papen (2005).

An intriguing new class of inelastic backscatter lidar is the so-called femtosecond white-light lidar. This employs an ultrashort-pulse transmitter operating in the near-IR at around 800 nm whose high intensity gives rise to nonlinear self-phase modulation that results in broadband upconverted continuum emission across the UV-IR spectral range (Kasparian et al., 2005). The critical factor in such systems is that the broadband continuum is emitted predominantly in the backward direction and is thus a particularly strong and effective backscattering mechanism (Yu et al., 2001). Applications for the femtosecond lidar center on investigations requiring multiple wavelength channels to constrain multivariate retrievals, such as high-resolution atmospheric absorption spectrum measurement and aerosol composition and microphysics determination (Kasparian et al., 2003).

Elastic backscatter lidar

The majority of lidar systems in use can be broadly classed as elastic backscatter systems. Elastic backscattering processes conserve the centroid wavelength and consist of Rayleigh (or molecular) scattering from air molecules and particle scattering from suspended particulate material (aerosols and clouds). (The term Mie scattering is frequently used to implicitly denote particle scattering; however, the Mie descriptor strictly speaking only applies to scattering from spherical particles, which in general account for only a small percentage of the total atmospheric aerosol population.) Rayleigh scattering magnitude is inversely proportional to the fourth power of wavelength, which in practice limits the usefulness of Rayleigh lidar approaches to wavelengths not exceeding the visible spectrum. Rayleigh lidar systems tend to be used for profiling fundamental atmospheric properties such as temperature and density (Chanin and Hauchecorne, 1984).

By far the commonest application of lidar is concerned with cloud and aerosol (i.e., particulate) detection and characterization. In its simplest form, this can entail measurements at a single wavelength to infer the broad characteristics of scattering structure in atmospherically entrained dust and cloud features. However, in the absence of ancillary data, such minimally constrained measurements can only acquire estimates of attenuated backscatter. To properly retrieve aerosol optical properties and physical parameters requires a multiwavelength lidar approach and calls for backscatter coefficients at a minimum of three wavelengths and aerosol extinction coefficients for at least two wavelengths (e.g., Müller et al., 1998; Althausen et al.,, 2000). Aerosol extinction has typically been obtained from collocated, contemporaneous column optical depth measurements (using sunphotometers) or from auxiliary Raman lidar channels, which measure the Raman backscatter from a well-mixed atmospheric constituent (usually nitrogen) and observe its departure from expected values to ascribe the aerosol extinction (Ansmann et al., 1992). However, in recent years there has been increasing interest in the high spectral resolution lidar (HSRL) technique because of its ability to separate and directly measure the aerosol and molecular scattering components with a single, self-contained multiwavelength instrument that avoids recourse to ancillary measurements (Eloranta, 2005; Hair et al., 2008).

In addition to the wealth of ground-based and airborne cloud/aerosol lidar systems that are today operated around the world, some functioning as components of autonomous, long-term monitoring networks, three multiwavelength systems have flown in Earth orbit, one on the Space Shuttle (Winker et al., 1996) and two on dedicated free-flying platforms (Spinhirne et al., 2005; Winker et al., 2007). There has even been a cloud/aerosol lidar operated on the surface of Mars (Whiteway et al., 2008).

Polarimetric lidar

The shape of cloud/aerosol particles can also modify the polarization state of scattered radiation. Polarization lidars exploit this phenomenon by separately measuring the unaltered (p) and depolarized (s) components of the return radiation, for which a range-dependent backscatter depolarization ratio, δ(R), can then be defined:

$$ \delta (R)=\frac{{{P_s}(R)}}{{{P_p}(R)}}. $$
(2)

(Note, however, that Gimmestad (2008) has offered a recasting of the traditional polarization lidar formulation that is more descriptive of the microphysical processes involved and which may in time supersede this notation.)

The magnitude of the depolarization ratio can be used to deduce the degree of nonsphericity of the particles. The relative polarimetric properties of various aerosol and cirrus particles show sufficient diversity that polarization lidar can provide for some level of discrimination between the different classes of such particles (Sassen, 1991; Sassen, 2007). However, care must be taken when inferring scattering or attenuation coefficients from lidar measurements of cirrus, as the preferential horizontal orientation of cirrus platelets has been shown to dramatically enhance lidar returns through the addition of a specular reflection component to the backscatter signal (Platt et al., 1978). In order to separate the intrinsic cirrus scattering properties from this effect, it is therefore advisable to offset the lidar viewing angle by ∼3° from the vertical.

Differential absorption lidar

The differential absorption lidar (DIAL) evolved as a means for identifying and quantifying gaseous atmospheric constituents. The technique entails transmitting two different wavelengths, one tuned to a characteristic spectral feature of the target species and the other offset by a small but sufficient amount that it is unaffected by the target gas and so acts as an atmospheric transmissivity reference. The usual practice is to contrive the on-and off-resonant wavelengths to be as narrowly separated as possible (in order to eliminate dispersive effects from other atmospheric constituents that could confuse the measurement), so that the following close approximation for the number density N(R) of absorbing molecules can be applied:

$$ N(R)=\frac{1}{{2|\sigma |}}\frac{{d\ln \left[ {{P_{on }}(R)/{P_{off }}(R)} \right]}}{dR }, $$
(3)

in which σ is the differential absorption cross section between the on-and off-resonant wavelengths.

A large number of DIAL systems are in operation worldwide to measure a wide variety of atmospheric trace gases from the ground and from the air. Because of their climatic impacts, water vapor and ozone profiling are two common DIAL measurement objectives (Browell et al., 2004), while a plethora of systems operating from the UV to the thermal-IR are in use to track the transport and dispersion of plumes containing common pollutants such as sulfur dioxide (Weibring et al., 1998), ammonia (Zhao et al., 2002), hydrocarbons (Murdock et al., 2008), and many other compounds.

The ongoing debate over the role of carbon dioxide in climate change has driven a resurgence in interest concerning DIAL techniques for monitoring this atmospheric constituent. The requirement is for mixing ratio measurement globally throughout the lowermost levels of the troposphere with a precision of ∼0.5 %, which is considerably more challenging than the typical ∼20 % DIAL precision specification. Conventional DIAL is one of two approaches that are under evaluation for this application (Koch et al., 2004; Gibert et al., 2006), the other being integrated path differential absorption (IPDA), which entails cw or quasi-cw illumination of the Earth’s surface at on-and off-resonant probe wavelengths (Menzies and Tratt, 2003). The retrieved differential absorption return can then be transformed into a path-averaged column content (Abshire et al., 2010; Spiers et al., 2011). Detailed trade studies will ascertain the optimum wavelength combination and detection configuration for acquiring high-precision gas mixing ratios by this technique.

Doppler lidar

Doppler lidars are used to measure target motion, either for atmospheric wind measurement (Baker et al., 1995) or object tracking (Osche and Young, 1996), and can be divided into two distinct classes. Coherent detection Doppler lidars directly measure the Doppler shift imposed on the optical return by the atmospheric target by mixing the return signal with a well-defined frequency reference laser, termed a local oscillator. The heterodyne mixing efficiency in such systems is a sensitive function of frequency purity, wave front quality, and optical aberration factors, any one of which can be the source of significant performance degradation if not adequately managed (Zhao et al., 1990; Chambers, 1997). For hard target coherent Doppler lidar systems, where backscatter is high, these considerations are less of an issue, but for atmospheric applications, where performance is contingent on aerosol and cloud backscatter, one is effectively driven to operating wavelengths in the IR spectral region.

By contrast, direct detection Doppler lidars operate in the “photon bucket” mode and thus do not mandate the wave front quality requirements of coherent systems. In a direct detection system, the Doppler shift is inferred indirectly by measuring the differential transmission of the return signal through a high-resolution filter mechanism whose dispersion characteristics must be known to high accuracy and very well controlled. Freed from the IR wavelength constraint to which coherent systems are subject, direct detection Doppler lidars are thus designed to take advantage of the omnipresence of molecules in the atmosphere by operating in Rayleigh backscatter mode (usually in the eye-safe UV). Direct detection systems are less photon efficient than their coherent counterparts, and their retrieval accuracy can be compromised by the individual and collective stability of optical components within the receiver subsystem, but are preferred for operation against atmospheres that have little or no aerosol content. The combining of both Doppler lidar approaches in a hybrid instrument has been suggested as a means to make best use of the advantages of both types of system (Emmitt, 2004).

There is a significant body of experience with both surface-based (Huffaker and Hardesty, 1996; Grund et al., 2001; Henderson and Hannon, 2005) and airborne (Bilbro et al., 1986; Rothermel et al., 1998; Werner et al., 2001; Hannon et al., 1999) coherent Doppler wind lidars. In addition, there have been several ground-based direct detection Doppler wind lidars (von Zahn et al., 2000; Gentry et al., 2000; Irgang et al., 2002) and also a few airborne systems (Dehring et al., 2006; Durand et al., 2006). All require some means of scanning the transceiver in order to obtain the multiple perspective line-of-sight retrievals that are necessary to resolve 2D or 3D wind vector fields.

The European Space Agency (ESA) plans to deploy the first space-based Doppler lidar. The Atmospheric Dynamics Mission (ADM-Aeolus) will carry a direct detection Doppler wind lidar and an integral high spectral resolution lidar (HSRL) with two receiver channels for analysis of aerosol/cloud backscatter and molecular backscatter, respectively (Stoffelen et al., 2005). This instrument will address only a single off-nadir line-of-sight perspective but will nevertheless represent the first on-orbit demonstration of Doppler wind lidar techniques.

It is worth noting here that the first lidar wind measurements from space were made by a non-Doppler elastic backscatter system using optical scatterometry of the ocean surface (Menzies et al., 1998). These measurements were necessarily indirect, scalar in character, and restricted to sea surfaces only, but more recent measurements indicate that this approach can provide significant new information on ocean winds and associated air–sea interaction processes (Hu et al., 2008).

Laser ranging and geodetic imaging

Laser ranging to hard targets was among the earliest of lidar applications and has seen by far the most commercial exploitation. Apart from the proliferation of laser-augmented rangefinding binoculars and surveying instruments, the technique is now in high demand for airborne high-density terrain mapping (Slatton et al., 2007) and shallow-water hydrographic surveying (Banic and Cunningham, 1998), for which a significant commercial services capability has arisen. Laser altimetry was also the first lidar application to be implemented in space (Sjogren and Wollenhaupt, 1973; Bufton, 1989; Garvin et al., 1998), and global-scale geodetic mapping by laser altimetry can in many respects be regarded as commonplace, having now been demonstrated with moderate horizontal resolution (10–100 m) on Earth (Schutz et al., 2005), Mars (Smith et al., 2001), Mercury (Zuber et al., 2008), the Moon (Smith et al., 1997; Smith et al., 2010), and the irregular asteroids Eros (Cheng et al., 2002) and Itokawa (Mukai et al., 2007). Most of these were first-return time-of-flight instruments, the exception being the Geoscience Laser Altimeter System (GLAS), which was designed to transcribe the entire return waveform (Abshire et al., 2005). Full waveform recovery makes it possible for laser altimetry to resolve volumetric vegetation structure and sub-canopy topography by taking advantage of the optical porosity of vegetation cover (Blair et al., 1999; Lefsky and McHale, 2008).

Some measurement scenarios call for a level of areal density of individual ranging measurements that pushes horizontal resolution toward the meter range. For these applications, limited pulse energy and the need for rapid scanning mean that “photon starvation” becomes a significant issue (Abrams and Tratt, 2005). To address this concern, single-photon counting detector arrays with integrated readout circuitry have been developed to acquire 3D imagery of laser flood-illuminated scenes in what is often termed a flash lidar scheme (Albota et al., 2002). In addition, work is currently in process to extend active image formation techniques from the established synthetic aperture radar domain into the optical spectral region (Beck et al., 2005). This technique faces nontrivial technical challenges but is nevertheless being pursued for tactical applications (Ricklin and Tomlinson, 2005) and has also been proposed for planetary remote sensing (Karr, 2003).

High-density laser ranging is also used in space-to-space scenarios where the uncompromising demands of autonomous rendezvous and proximity operations rule out less precise instrumental approaches (Nimelman et al., 2006).

Summary

Active optical remote sensing (lidar) techniques have provided demonstrable enhancements in spatiotemporal resolution and reduced dependence on the diurnal cycle in comparison to other remote sensing methods used for Earth and planetary science applications. While the breadth and diversity of lidar methodology attests to its fundamental versatility, there are certain measurements that are enabled uniquely by active optical sensors (e.g., aerosol and cloud vertical structure, global-scale tropospheric wind measurement, 3D vegetation structure, high-resolution sub-canopy topography mapping). The intrinsic benefits of lidar now drive a steadily increasing number of fielded systems on the ground, in the air, and in space. While not exhaustive, this entry describes the basic characteristics and associated applications of the commonest types of lidar likely to be encountered and also provides an introduction to some promising emerging concepts that have yet to fully mature into accepted field-capable systems.

Cross-references

Cryosphere, Measurements and Applications

Ocean, Measurements and Applications

Optical/Infrared, Atmospheric Absorption/Transmission, and Media Spectral Properties

Optical/Infrared, Radiative Transfer

Optical/Infrared, Scattering by Aerosols and Hydrometeors

Radiation, Multiple Scattering

Radiation, Volume Scattering