Lidar Remote Sensing for Ecosystem Studies

Remote sensing has facilitated extraordinary advances in the modeling, mapping, and understanding of ecosystems. Typical applications of remote sensing involve either images from passive optical systems, such as aerial photography and Landsat Thematic Mapper (Goward and Williams 1997), or to a lesser degree, active radar sensors such as RADARSAT (Waring et al. 1995). These types of sensors have proven to be satisfactory for many ecological applications, such as mapping land cover into broad classes and, in some biomes, estimating aboveground biomass and leaf area index (LAI). Moreover, they enable researchers to analyze the spatial pattern of these images. However, conventional sensors have significant limitations for ecological applications. The sensitivity and accuracy of these devices have repeatedly been shown to fall with increasing aboveground biomass and leaf area index (Waring et al. 1995, Carlson and Ripley 1997, Turner et al. 1999). They are also limited in their ability to represent spatial patterns: They produce only two-dimensional (x and y) images, which cannot fully represent the three-dimensional structure of, for instance, an old-growth forest canopy. Yet ecologists have long understood that the presence of specific organisms, and the overall richness of wildlife communities, can be highly dependent on the three-dimensional spatial pattern of vegetation (MacArthur and MacArthur 1961), especially in systems where biomass accumulation is significant (Hansen and Rotella 2000). Individual bird species, in particular, are often associated with specific three-dimensional features in forests (Carey et al. 1991). In addition, other functional aspects of forests, such as productivity, may be related to forest canopy structure. Laser altimetry, or lidar (light detection and ranging), is an alternative remote sensing technology that promises to both increase the accuracy of biophysical measurements and extend spatial analysis into the third (z) dimension. Lidar sensors directly measure the three-dimensional distribution of plant canopies as well as subcanopy topography, thus providing high-resolution topographic maps and highly accurate estimates of vegetation height, cover, and canopy structure. In addition, lidar has been shown to accurately estimate LAI and aboveground biomass even in those high-biomass ecosystems where passive optical and active radar sensors typically fail to do so. The basic measurement made by a lidar device is the distance between the sensor and a target surface, obtained by determining the elapsed time between the emission of a short-duration laser pulse and the arrival of the reflection of that pulse (the return signal) at the sensor's receiver. Multiplying this time interval by the speed of light results in a measurement of the round-trip distance traveled, and dividing that figure by two yields the distance between the sensor and the target (Bachman 1979). When the vertical distance between a sensor contained in a level-flying aircraft and the Earth's surface is repeatedly measured along a transect, the result is an outline of both the ground surface and any vegetation obscuring it. Even in areas with high vegetation cover, where most measurements will be returned from plant canopies, some measurements will be returned from the underlying ground surface, resulting in a highly accurate map of canopy height. Key differences among lidar sensors are related to the laser's wavelength, power, pulse duration and repetition rate, beam size and divergence angle, the specifics of the scanning mechanism (if any), and the information recorded for each reflected pulse. Lasers for terrestrial applications generally have wavelengths in the range of 900–1064 nanometers, where vegetation reflectance is high. In the visible wavelengths, vegetation absorbance is high and only a small amount of energy would be returned to the sensor. One drawback of working in this range of wavelengths is absorption by clouds, which impedes the use of these devices during overcast conditions. Bathymetric lidar systems (used to measure elevations under shallow water bodies) make use of wavelengths near 532 nm for better penetration of water. Early lidar sensors were profiling systems, recording observations along a single narrow transect. Later systems operate in a scanning mode, in which the orientation of the laser illumination and receiver field of view is directed from side to side by a rotating mirror, or mirrors, so that as the plane (or other platform) moves forward, the sampled points fall across a wide band or swath, which can be gridded into an image. The power of the laser and size of the receiver aperture determine the maximum flying height, which limits the width of the swath that can be collected in one pass (Wehr and Lohr 1999). The intensity or power of the return signal depends on several factors: the total power of the transmitted pulse, the fraction of the laser pulse that is intercepted by a surface, the reflectance of the intercepted surface at the laser's wavelength, and the fraction of reflected illumination that travels in the direction of the sensor. The laser pulse returned after intercepting a morphologically complex surface, such as a vegetation canopy, will be a complex combination of energy returned from surfaces at numerous distances, the distant surfaces represented later in the reflected signal. The type of information collected from this return signal distinguishes two broad categories of sensors. Discrete-return lidar devices measure either one (single-return systems) or a small number (multiple-return systems) of heights by identifying, in the return signal, major peaks that represent discrete objects in the path of the laser illumination. The distance corresponding to the time elapsed before the leading edge of the peak(s), and sometimes the power of each peak, are typical values recorded by this type of system (Wehr and Lohr 1999). Waveform-recording devices record the time-varying intensity of the returned energy from each laser pulse, providing a record of the height distribution of the surfaces illuminated by the laser pulse (Harding et al. 1994, 2001, Dubayah et al. 2000). By analogy to chromotography, the discrete-return systems identify, while receiving the return signal, the retention times and heights of major peaks; the waveform-recording systems capture the entire signal trace for later processing. Conceptual differences between the two major categories of lidar sensors are illustrated in Figure 1. Both discrete-return and waveform sampling sensors are typically used in combination with instruments for locating the source of the return signal in three dimensions. These include Global Positioning System (GPS) receivers to obtain the position of the platform, Inertial Navigation Systems (INS) to measure the attitude (roll, pitch, and yaw) of the lidar sensor, and angle encoders for the orientation of the scanning mirror(s). Combining this information with accurate time referencing of each source of data yields the absolute position of the reflecting surface, or surfaces, for each laser pulse. There are advantages to both discrete-return and waveform-recording lidar sensors. For example, discrete-return systems feature high spatial resolution, made possible by the small diameter of their footprint and the high repetition rates of these systems (as high as 33,000 points per second), which together can yield dense distributions of sampled points. Thus, discrete-return systems are preferred for detailed mapping of ground (Flood and Gutelis 1997) and canopy surface topography, as in Figure 2. An additional advantage made possible by this high spatial resolution is the ability to aggregate the data over areas and scales specified during data analysis, so that specific locations on the ground, such as a particular forest inventory plot or even a single tree crown, can be characterized. Finally, discrete-return systems are readily and widely available, with ongoing and rapid development, especially for surveying and photogrammetric applications (Flood and Gutelis 1997). The primary users of these systems are surveyors serving public and private clients, and natural resource managers seeking a cheaper source of high-resolution topographic maps and digital terrain models (DTMs). A potential drawback is that proprietary data-processing algorithms and established sensor configurations designed for commercial use may not coincide with scientific objectives. A detailed technical review of the various sensors can be found in Wehr and Lohr (1999). Baltsavias (1999) reviews a directory of sensors and lidar remote sensing firms. The advantages of waveform-recording lidar include an enhanced ability to characterize canopy structure, the ability to concisely describe canopy information over increasingly large areas, and the availability of global data sets (the extent of their coverage varies, however). Examples of waveform-recording laser altimeters include MKII (Aldred and Bonnor 1985) and a similar system described in Nilsson (1996), as well as a series of airborne devices developed at NASA's Goddard Space Flight Center, starting with a profiling sensor described by Bufton and colleagues (1991) and including SLICER (Scanning Lidar Imager of Canopies by Echo Recovery; Blair et al. 1994, Harding et al. 1994, 2001), SLA (Shuttle Laser Altimeter; Garvin et al. 1998), LVIS (Laser Vegetation Imaging Sensor; Blair et al. 1999), and VCL (Vegetation Dubayah et al. 1997) One advantage of these waveform-recording lidar systems is that they record the entire time-varying power of the return signal from illuminated surfaces and are of information on canopy structure the most dense of lidar In addition, waveform-recording lidar canopy structure information over a large footprint and is of that information from the of both data and data Finally, only waveform-recording lidar in the near be collected from waveform-recording lidar have been by the Laser et al. 1998), which were to topographic data and to and from These data were collected along a single of in which limits their for the measurement of vegetation canopy structure, especially in areas (Harding et al. The and for in 2001, will the Laser which will make measurements along a single with diameter which the size to characterize vegetation in and The Vegetation Lidar to be is the designed with the of vegetation inventory in VCL is a waveform-recording to diameter canopy height and structure over of the Earth's land surface between during et al. 1997). with the VCL is the Lidar Vegetation Imaging System an mapping system developed at NASA's Goddard Space Flight that is used to LVIS can with diameter across over in VCL is a sampling will make waveform measurements along a of three to the ground of the resulting in a of the Earth's These will not images of canopy structure, they be with images from other sensors as Thematic data from Landsat a number of with VCL data or even the by data et al. Dubayah et al. 2000). as discrete-return and waveform-recording lidar are The between data from each is illustrated in Figure data collected with a lidar at the a the three-dimensional distribution of data from a footprint on a tree the distribution of these points as a of height. Blair and (1999) that this vertical distribution of the data is related to the recorded by waveform-recording devices are most a high of collected a small footprint the of so that data can be collected from small in the canopy structure. a lidar the vertical distribution of the discrete return would have to be for the spatial and distribution of energy the lidar pulse and receiver as described in Blair and (1999). a areas of for lidar remote sensing have been other applications are generally they have not been in lidar remote sensing are so that is to which applications will be in applications of lidar remote sensing in fall into three remote sensing of ground topography, measurement of the three-dimensional structure and of vegetation canopies, and of forest structure as aboveground of topographic features is the and area of for lidar remote of use in commercial land (Flood and Gutelis 1997). are also in which often has a on the structure, and of ecological and photogrammetric for determining ground elevations are limited in several The primary of surveying are time and and associated for determining elevations from aerial or images collected by other sensors are an established alternative to field 1999). However, they are in areas, where the ground is not and in areas of and such as areas and In these airborne laser can be an accurate and applications most often use discrete-return When information from the lidar is with position and the result is a series of data or the of the surfaces in three-dimensional the accuracy of these points can in the and in the However, the elevations recorded in these will be associated with including the ground, clouds, or in the path of the laser pulse. a topographic surface from these a series of be to points not on the ground for this generally they highly with some et al. and commercial data use proprietary they are often to describe in a potential for Examples of topographic applications of lidar include mapping of for et al. 1999), mapping of and shallow water and 1999), and high-resolution mapping of under forest for and (Harding and 2000). The mapping of features such as and et al. is one for which lidar is to be well The ability of airborne lidar to of the is by the Lidar of a of the and Center, the for and the and Space et al. 2000). the Mapper developed at NASA's Flight detailed maps are for areas along the and of of the areas, measurements and images of are The resulting data are designed for accurate and mapping of and be to understanding of the for instance, and vegetation in ecosystems. In the single most in lidar mapping of the of data points returned from vegetation and, in areas, However, for most ecological applications, is the from the vegetation canopy that will be of primary in and including the and of the aboveground of 1995, a amount of information the of of plant et al. and canopy and and and for wildlife (Hansen and Rotella 2000). The canopy structure measurements are of canopy height and cover canopy heights have been with accuracy and of to maximum and tree height in and et al. 1997, et al. and and et al. In addition, and colleagues found between lidar measurements of height in both forests and The is as that vegetation height measurements can be made accurately even on vegetation of at in There are two in determining vegetation height lidar the of the ground surface for both discrete-return and waveform-recording In complex canopies, elevations returned from to be the ground in may be from the the is dense to the ground In addition, each type of lidar system in the of the plant canopy. discrete-return high footprint are to that the of tree is waveform sampling a large footprint is increasing the that will be illuminated by the However, the of the may not be of area to as a significant return signal and may not be In either the height of the canopy may be of canopy cover have been made both discrete-return and waveform-recording lidar sensors. These estimates are made the fraction of the lidar measurements that are to have been returned from the ground surface et al. et al. 1995, et al. 1994, 1997), where the measurements are the number of discrete or the power of a In some a is to for the reflectance of ground and canopy surfaces at the of the laser 1997, et al. 1999). with the measurement of canopy height, the of the ground surface is a of cover the number (or of the measurements to the ground return is the of the ground surface is will be and the height and cover of the canopy surface are canopy structure are detailed measurements that can better describe canopy and structure. The height distribution of canopy surfaces which such features as light et al. et al. has been in several and and These maps were devices such as and with the is et al. et al. The vertical distribution of the canopy the canopy may be the (MacArthur and for use with waveform-recording lidar as the 1997, Harding et al. of these height on the of of canopy surfaces that are not to they have been shown to yield a in forests et al. Harding et al. Lidar data have been used to the of light as a of height on a series of the penetration of the laser light into the canopy to the penetration of natural light into the canopy. both the and orientation of typical laser illumination from that of natural a et al. that lidar can accurately estimate the of active absorption and the and of the where the maximum of absorption 1997). Lidar has also been used to the of plant canopies and In over a forest canopy, the is the height at which the speed and used a profiling laser to of complex a of and areas, and found with field The described so use lidar data to make measurements of canopy structure that been made with and ability to measure the three-dimensional structure of canopies the of systems of canopy One such the canopy is the to advantage of the ability of a waveform-recording sensor to directly measure the three-dimensional distribution of canopy structure. lidar and colleagues were to the forest canopy as a of Figure each of which be as canopy or and either in the or of the canopy. information used to describe the and differences in canopy structure between classes to a better understanding of the structure of the old-growth forest canopy, of the of old-growth development, and estimates of forest structure. Lidar data also have been used to biophysical of plant communities, most forests and 2000). the may not by ecological they the for that use these to map biophysical over large data from sensors such as LVIS and possible a of ecological of forest structure discrete return lidar in the of and a photogrammetric canopy the of lidar The canopy area is the total area between the ground and the canopy surface along a transect. When into the were to of the in (the of the of the and including and in by or of the two have in a of forest et al. the and biomass of and forests several estimates of canopy height and cover from discrete-return between and of in field measurements of these Later by et al. in forests at the obtained similar results for of and They also developed a canopy structure that to understanding of the spatial of field sampling for with profiling lidar of in in of and measurements of maximum and canopy height and the of waveform-recording lidar in forest structure. Nilsson a lidar system for use in forest and for of used the height and the total power of each waveform as and of and colleagues used data from SLICER to aboveground biomass and area in forests from the canopy height particular they found that between height and forest structure area and aboveground be field estimates of the canopy height and directly to the resulting in estimates of forest structure. and colleagues (1999) similar to in forests of and at the They found that accurate estimates of aboveground and biomass be made lidar height and cover A et al. used from the to numerous forest structure including several not from lidar remote were to of structure from both conventional canopy structure and maximum canopy surface height, canopy cover, and such as canopy and and a canopy number of classes per height. of structure are in Figure of the that the values of aboveground biomass and LAI even at large values per of aboveground In addition, the three-dimensional aspects of the canopy lidar were to accurately estimate related to complex aspects of the diameter such as the of diameter at height and the number of The et al. the of waveform-recording lidar to a forest in the LVIS sensor, data were collected near the a of the vertical distribution of the and the fraction of total power associated with the ground they were to and aboveground to and of The resulting map which the in aboveground biomass with and be in and also other in the Lidar remote sensing only has as a and has to widely has been shown to be an accurate for topography, vegetation height, and cover, as well as complex of canopy structure and In addition, the basic canopy structure measurements made with lidar sensors have been shown to highly accurate and estimates of forest structure such as leaf area index and aboveground the basic measurements made by lidar sensors are directly related to vegetation structure and that these will to be in a of biomes, with similar The availability of lidar data will increase with the of several lidar and the use of airborne sensors for topographic data availability a of applications will is that lidar will be in features associated with particular species, including those that are or For instance, the large and associated old-growth that as for and be readily from lidar of also may be to areas of high which be used to such as the and of lidar data is the of forest areas with of that make to especially ability to the spatial pattern as well as the total of a forest canopy would be especially for identifying, at the classes of forest structure that are associated with For instance, lidar enable the detection of which a for to the canopy and In addition, the ability to the size and of canopy of the of large associated with the of those lidar remote sensing potential for with ecological directly the of vegetation canopy structure that are highly with the basic plant measurements of to detailed measurement and of canopies has been the of By the time and associated with canopy structure, lidar can the of a canopy into ecological and vegetation canopy structure at the of to measure and global by a from the of to and of the SLICER by NASA's and the Goddard SLICER data sets for public distribution are at of the SLICER data used by a to The SLICER also by the and to and to at and by the a scientific of the of the and the to Dubayah and an for their detailed reviews of an of this Figure 1. of the differences between waveform-recording and discrete-return lidar the is the of the laser illumination or with a of a tree In the of the figure is a return signal (the lidar that would be collected by a waveform-recording sensor over the the of the the heights recorded by three of discrete-return lidar sensors are lidar devices record only the position of the in the path of the laser lidar devices record the height of the in the path of illumination and are especially for topographic a the height of a small number or of objects in the path of illumination Figure 2. surface of a of the in The data have been lidar would be represented by points in three-dimensional Figure of canopy structure made NASA's SLICER (Scanning Lidar Imager of Canopies by Echo a ground and the vertical distribution of canopy along a in the is the width of one laser pulse and of the canopies of three in and old-growth forest with their ground elevations to a Figure of the potential for lidar from discrete-return lidar a the three-dimensional distribution of discrete-return lidar data from a the vertical distribution of these Figure Lidar measurements of the of on the of a maps the overall for this of the and for a single Figure of the of Figure Conceptual of three canopy a and are a canopy by et al. is a canopy surface the vertical distribution of the canopy is a canopy height the vertical distribution of and the vertical of is a canopy the vertical distribution of classes of canopy structure. Figure from et al. with from Figure Conceptual for the canopy The of the are in diameter and they to a vertical a single waveform is to and a is used to each of the waveform into either or The distribution of the waveform is used to of the into a which the of energy to the sensor, and an of the of the These two are to three canopy structure the canopy the and the is as the between the of each of the and the maximum height in the Figure from et al. with from Figure for and old-growth and forest These for each vertical the of each that to each of the canopy structure are by a canopy surface, and an of the canopy. are by a canopy surface with a large of the canopy. are from by their canopy surface and the broad vertical distribution of each of the canopy structure from in have canopy structure classes in vertical in the old-growth each canopy structure the height range of the has been as a feature old-growth forests from the canopies of and and 1991). Figure from et al. with from Figure of and from at the Figure from et al. with from Figure of aboveground biomass from LVIS data over from et al. with from and