Use of Radar for Wind Power-Related Avian Research
by
Brian A. Cooper, ABR Inc.
Radar has been an important tool in ornithological research for nearly five decades (Eastwood 1967). Radar was first used in wind power-related avian research during the mid-to-late 1970s in Ohio and California (Rogers et al. 1977; McCrary et al. 1984), but in the last five years it has been used widely for wind power-related studies of birds in North America and Europe (Pedersen and Poulsen 1991; Cooper and Ritchie 1994; Cooper et al. 1995a,b). Radar was used in these studies mainly because many species of birds (e.g., songbirds and ducks) migrate largely at night, when they are impossible to study with standard visual techniques. Radar also is useful during periods when fog or clouds restrict visibility during daytime, for observations over large areas that cannot be covered by a single visual observer, and to help visual observers detect and locate birds that otherwise would be missed (Kerlinger and Gauthreaux 1984, 1985; Cooper and Ritchie 1995). This is not to say that radar detects all birds in an area; it also is a sampling tool with its own biases and limitations. In fact, none of the sampling tools we have at our disposal today can detect all birds in an area at all times. Fortunately, radar and visual techniques complement one another well for avian studies relevant to wind power developments.
The purpose of this paper is to familiarize a general audience with the practical aspects of using radar for wind power-related avian studies, discussing both radar's benefits and its limitations. I will discuss briefly some principles and a history of radar, then will explain some of the benefits and limitations of some of the most commonly used types of radars, and next will describe in detail the marine radar laboratory we have used, including some practical aspects of its operation. The paper concludes with some ideas on how radar could be used for avian research during the pre- or post-construction phases of windfarm development, and lists some of the future needs for radar studies.
Introduction to Radar.-Radar stands for RAdio Detection And Ranging. Pulses of electromagnetic radiation are transmitted many times per second. During the brief intervals between pulses, the radar receives the echoes that are reflected back from objects within the radar's beam (e.g., a bird, plane, ship, or hill). At any given time, the pulses are transmitted toward a particular range of directions and elevation angles, determined by the antenna design and orientation. Objects from which echoes are received are normally within that same range of directions and angles, so antenna orientation provides information on target position. The radio waves travel at the speed of light, so the time interval between trans-mission of a pulse and reception of an echo is directly related to the distance to the object. The maximum detection distance for a particular object depends on many factors, including the radar power output, radar wavelength, and size and composition of the object. Object size and composition determine its "radar cross-section" at the wavelength in question. For birds, maximum detection distance varies from a few hundred meters for single small birds studied by the smallest marine radars to over 150 km in the case of bird flocks studied by long-range surveillance or tracking radars.
Radar was developed during World War II. Some of the first radar operators saw what they called "angels" on their radar screens over areas that they knew were devoid of aircraft. Most of these "angels" turned out to be birds, and the field of radar ornithology was born. Many studies of migratory and local movements of birds have been conducted with radar since the 1950s; the general principles and the early work were reviewed by Eastwood (1967).
We have used radar for several different studies during the past nine years, including wind power-related studies of birds for the Niagara Mohawk Power Corporation in upstate New York and for Kenetech Windpower in Spain. I will focus on the use of marine radars, the type of radar that we use for bird studies, but first I will describe some of the other types of radar that have been used by ornithologists.
Common Types of Radar Used for Bird Studies.-Large weather radars (e.g., WSR-57, NEXRAD or WSR-88D) and air surveillance radars are excellent tools for studying patterns of bird migration over extensive areas (Eastwood 1967; Gauthreaux 1975; Rich-ardson 1979; Able 1985; Buurma and Bruderer 1990; Buurma 1995). These types of radars could be useful for broadarea preliminary site selec-tion surveys. However, they are not useful for collecting high-resolution data over small areas such as wind sites. Further, they are usually stationary and may not always be available near a particular wind site. Finally, some of these radars are equipped with devices that filter out and remove echoes of some birds (Rich-ard-son 1972). For these reasons, the following discussion will focus on smaller, mobile types of radar that could provide high-resolution data from a desired location.
Tracking radars are designed to lock onto and follow targets such as aircraft or missiles, providing continuous data on their positions and movements in three dimensions. Small military surplus tracking radars can provide good information on the flight behavior of birds (including altitude, speed, and direction), provide some limited identification ability via wingbeat signature, and have MTI (Moving Target Indicator) circuitry to reduce ground clutter (echoes from the ground and other stationary objects). The disadvantages of these systems are that they do not provide a broad picture of migration over a site unless they are used in a surveillance mode (e.g. Bruderer et al. 1995), are not readily available, require fairly extensive training to operate, and are difficult and expensive to repair. Although a mobile military tracking radar may be the ultimate system for many ornithological studies, it probably would get only a "good" rating for most types of wind power-related avian research because of these limitations. For more details on the specifications, past use, and merits of tracking radar for ornithological studies, see Eastwood (1967), Blokpoel (1971), Brud-erer and Steidinger (1972), Griffin (1973), Larkin et al. (1979), Richardson (1979), Kerlinger (1980, 1982), Able (1982, 1985), Buurma and Bruderer (1990), and Bruderer et al. (1995).
Marine radar, which typically is used on boats for navigational purposes, is an excellent tool for many types of wind power-related avian research. The advantages of marine radar systems are that they are inexpensive, are available off-the-shelf, require little modification or maintenance, are dependable, have repair personnel readily available worldwide, are easy to operate, have very high resolution, and can be modified to collect altitude information (Williams et al. 1972; Korschgen et al. 1984; Williams 1984; Gauthreaux 1985a,b; Cooper et al. 1991). Largely because of these factors, almost all avian-wind power research based on radar has been done with marine radar systems (McCrary et al. 1984; Pedersen and Poulsen 1991; Cooper and Ritchie 1994; Cooper et al. 1995a, b); the exception was Rogers et al. (1977), who used long-range air surveillance radar. The disadvantages of marine radars as compared with tracking radars are that marine radar systems have more problems with ground clutter interference, have a very limited ability to identify birds to the family let alone the species level, and have shorter range than many of the tracking radar systems. Small marine radars (10 kW peak power) can detect small, individual songbirds to range 1 km and large, individual hawks (e.g., Red-tailed Hawk) up to 4 km (Cooper et al. 1991). With the Flycatcher tracking radar system, single thrushes can be detected to 7 km (Buurma and Bruderer 1990). I believe that the disadvantage of the more limited range of the marine radar is compensated for by its ease of operation, convenience, cost, simpler surveillance capability, and high resolution.
Components of a Mobile Marine Radar Laboratory
Radar Equipment.-Our mobile laboratory (Cooper et al. 1991) consists of two small marine radars mounted on a cab-over camper on a four-wheel-drive pick-up truck (Fig. 1). One of the radars (surveillance) is used to scan the entire area around the lab, gathering information on flight paths, movement rates, and ground speeds of flying birds. A second radar (vertical) has been modified to measure altitudes of flight. A description of a similar radar laboratory can be found in Gauthreaux (1985a,b). The lab can be powered by a generator or by deep-cycle bat-ter-ies; when fully charged, four 6-V golf cart batteries can power the lab continuously for ~14 h.
Surveillance radar: The surveillance radar (Furuno Model FCR-1411, Furuno Electric Co., Nishinomiya, Japan) is a standard X-band marine radar transmitting at 9410 MHz (i.e., 3 cm wavelength) through a slotted wave guide 2 m long. The peak power output is 10 kW; how-ever, Furuno now makes a similar model that operates at 25 kW. The radar can be operated at a variety of maximum range settings, from 0.5 km to 133 km. Pulse length can be set at 0.08, 0.6, or 1.0 ms, depending on the range setting used. At shorter pulse lengths, echo definition is improved (giving you more accurate information on target location and, hence, distance), whereas at longer pulse lengths echo strength is improved, increasing the probability of detecting a target. An echo is a picture of a target on the video display screen. A target that is of interest here consists of one or more birds flying so closely that the radar presents them as one echo on the display screen.
FIGURE 1. The mobile radar lab.
Our surveillance radar has a digital, color display with several scientifically useful features. These include color-coded echoes (to differentiate the strengths of return signals), on-screen plotting of the sequence of echoes obtained during different anten-na revolutions (to depict flight paths), and True North correction for the display screen. A plotting function records the location of a target at selected time intervals (0.25, 0.5, 1, 3, or 6 min) (Fig. 2). Because these time intervals are fixed, ground speed is directly proportional to the distance between consecutive echoes and can be measured with a hand-held scale. In addition, an alarm function can be set to sound when echoes above a certain signal strength appear on the screen.
Vertical radar: The vertical radar (Furuno Model FR-8100) is a standard marine radar that was modified by replacing the slotted wave guide with a 0.6-m-diameter parabolic dish. This radar also transmits at 9410 MHz with a peak power output of 10 kW, can be operated at various maximum ranges from 0.5 to 89 km, and has a digital, eight-shade, monochrome display. Pulse length can be set at 0.08, 0.3, 0.6, or 1.0 ms, depending on the range setting used. A plotting function records the position (in this case, altitude) of a target, either continuously or at intervals of 0.5, 1, 3, or 6 min. An alarm function can be set to sound when echoes above a certain signal strength appear on the screen. In addition, interference rejection circuitry allows simultaneous operation of both this and the surveillance radar. Because of the vertical orientation of its beam, the vertical radar cannot detect birds flying below an altitude of approximately 25 m above ground level. In contrast, the surveillance radar can (depending on terrain, antenna angle, range, etc.) detect some birds that are only ~1 m above ground level, but cannot detect birds within a horizontal distance of ~25 m.
FIGURE 2. The surveillance radar display with plotted echoes of swans flying from southeast to northwest (note that the screen is oriented so that north is up). The large, irregular blotches are ground clutter. The adjustable, dashed ring has a radius of 4.02 km (noted in lower right corner of screen). The dashed, straight line (oriented at 299.5 , see lower left corner of screen) can be moved to determine flight direction (from Cooper et al. 1991).
The vertical radar is mounted on a hinged assembly that allows one to swing the antenna from the vertical position useful for counting birds directly above the laboratory and measuring their flight altitudes to a horizontal position useful for sampling birds crossing a nearly-horizontal line. Excessive scattering of radar energy from the antenna can be prevented by installing a tight-fitting collar of aluminum flashing ~100 cm high around the antenna (Gauthreaux 1985a; Cooper et al. 1991; Beerwinkle et al. 1993).
In a partially modified vertical radar system, the radar display screen shows only a thin, illuminated line that does not move. As birds pass through the radar beam, the targets appear along this line as bright spots; these are easily missed. We modified this system further by moving the antenna motor plate ~10 mm off-center, disengaging the gears between the motor and antenna, and allowing the motor to turn while the parabolic dish remained stationary. With this additional modification, targets form large areas or circles that are not easily missed on the display screen (Fig. 3).
FIGURE 3. The vertical radar display with plotted echo (the broken arc) of a Great Blue Heron flying at 324 m above ground level. Note that the adjustable, dashed ring (here set at 89 m, see lower right part of screen) can be used to measure flight altitude (from Cooper et al. 1991).
A customized data downloading system has been developed for a vertical radar system used to study insect migration (Beerwinkle et al. 1993). We currently are determining if "off-the-shelf" software can be modified to download the vertical radar information automatically into a database.
Equipment for Nocturnal Visual Observations.-Visual observations are an essential complement to any radar study. During the day, observations can be made with binoculars and telescopes. At night, night-vision scopes or Forward-Looking InfraRed (FLIR) devices are more useful. The range of the night-vision scope is positively correlated with the amount of incident light present (e.g., from streetlights, cars, the moon). The performance of this scope can be enhanced dramatically by using a spotlight equipped with an infrared filter as an external source of light. The filter renders the light invisible to the human eye and presumably to birds, thus avoid-ing effects on the birds' behavior while enhancing the images in the eye-piece of the night-vision scope. Using a 5X night-vision scope (second generation) and a 1.25 million candlepower spotlight equip-ped with an infra-red filter, one can detect small songbirds to ~150 m and gull-sized birds to ~400 m. With enough incident light, gull-sized birds can be detected to ~1000 m.
Forward-looking infrared (thermal imaging) devices like FLIR model 2000A (FLIR Inc., Portland, OR) do not require any incident light to work and can detect gull-sized birds to ~800 m. Other FLIR units designed for long-range use can detect birds considerably farther away, but have a narrow field of view (Liechti et al. 1995). Unfortunately, the larger FLIR scopes are not as portable as a night-vision scope, and FLIR scopes are very expensive (~$100,000 US) compared to night-vision scopes (~$4000 to $8000 for 2nd or 3rd generation equipment, respectively).
Limitations of a Marine Radar System
The major limitations of a marine radar system, common to many of the other types of radar as well, are (1) the actual number of birds represented by each target is unknown; (2) identification to the family or species level usually is not possible; (3) insects and bats sometimes make echoes that can be confused with slow-flying birds; (4) ground clutter can obscure large parts of the screen; and (5) birds cannot be detected during periods with moderate-heavy precipitation. An additional limitation of the vertical (but not surveillance) radar system is that, when it is in the vertical mode, birds cannot be detected below an altitude of 25 m. Fortunately, several techniques are available to minimize or mitigate the effects of these limitations.
Number of Birds.-A flock of birds usually appears as one echo on the radar screen. That is why most radar studies report movement rates as targets/h rather than as birds/h. For many types of research, this index of movement can be used without problems, but there are occasions when one wants information on actual numbers of birds crossing a site. The best way to estimate the actual number of birds from radar data is to obtain concurrent visual information on mean flock size. The number of targets can be multiplied by the mean flock size to obtain an estimate of the actual number of birds. The concurrent visual information can be collected with standard optical equipment during the day, but at night it is nec-es-sary to use night-vision (image intensification) devices or FLIR scopes. Because all radar and visual methods are sampling techniques, and as such are likely to miss at least a few birds, a further correction for missed flocks may be desirable based on double counting methods.
Species Identification.-Another reason that radar ornithologists often report their data in terms of "targets" rather than "ducks" or "geese", for instance, is that the identity of most echoes is unknown. With the marine radar system, it often is possible to separate targets into "songbirds", "raptors", "shorebirds", and "waterfowl", based on flight speed, target strength, target size, and flight behavior. In areas where a particular species has unique flight characteristics, it is possible to identify targets to species. For example, Marbled Murrelets were identified to the species level on radar with an accuracy >95% at inland nesting locations in northern California (Hamer et al. 1995). These results were verified with auditory and visual observations. This type of confirmation is needed for any studies using marine radar that wish to report species-specific information.
A promising new auditory technique for obtaining information on species identification of nocturnal migrants is being developed (B. Evans, Cornell Laboratory of Ornithology, pers. comm.). This technique uses special microphones to record the call notes of nocturnal migrants. (Most landbirds, except for tanagers, vireos, and flycatchers, emit call notes at least occasionally when they are migrating at night.) The calls are then analyzed by ear or with spectrographic analysis. The Cornell Bioacoustics Research Program is developing software that could be used in the future to automatically detect and classify warbler and sparrow calls. During fall 1994, we used this auditory technique concurrently with radar observations in upstate New York and obtained a strong correlation between the number of birds detected by the two methods. Because acoustic monitoring allows one to determine the species of many calling birds, this technique might have potential for monitoring the abundance of vocalizing species during their night migrations. In any case, it certainly is a very useful complement to the radar technique because of the ability to identify many night-migrating birds to the species level.
Insects and Bats.-Large, fast-flying insects sometimes can be confused with birds on X-band marine surveillance radars, but it often is possible to identify and disregard insect targets based on flight speed and target strength. These criteria are not valid for bats, which often are indistinguishable from birds on radar, except when the bats are foraging. Again, visual sampling is recommended to assess the extent of this potential problem. There may be periods when the problem (especially with insects) is severe enough that radar sampling should be discontinued until conditions improve. Insects and bats are more problematic for the vertical radar because, on that system, one does not receive information on flight speed or behavior, only altitude. In fact, a vertical radar system nearly identical to the one I describe has been used by entomologists to study insect migration (Beerwinkle et al. 1993). The extent of insect and bat activity can be assessed by making vertical, visual observations. A nice system for making concurrent visual and vertical radar observations is described in Gauthreaux (1985a). Alternatively, the level of insect activity can sometimes be assessed with a surveillance radar operating concurrently. Observations should be discontinued if large insects (e.g., large moths, beetles) are abundant above the minimum range of the vertical radar (25 m). For additional information on the separation of bird and insect targets, refer to Larkin (1991) and Vaughn (1985).
Another possible solution to the insect problem would be to use S-band marine radar instead of X-band radar. S-band radars transmit at a lower frequency (2000-4000 MHz) and have a longer wavelength (~10 cm) than do X-band radars (~9000 MHz and 3 cm). When targ-et size is much smaller than the wavelength, radar cross-section diminishes very rapidly with decreasing ratio of target circumference to wavelength (Skolnik 1980). Thus, most insects produce far less echo on S-band (10 cm) than on X-band (3 cm) radars. S-band radars prob-ably also do not detect small songbirds as well as do X-band radars, which is a problem if songbirds are of interest. Almost all bird studies to date that have used small marine radars have used X-band radars. However, S-band and even L-band (~23 cm) radars of other higher-powered types have been widely used to study bird move-ments (East-wood 1967; Rich-ard-son 1979).
Ground Clutter.-Whenever energy is reflected from the ground, surrounding vegeta-tion, and other objects that surround the radar unit, a ground clutter echo appears on the display screen. This can obscure bird targets. Ground clutter can be caused by, for example, stubble as low as 0.5 m high in a newly harvested agricultural field. Ground clutter can be minimized in some cases by elevating the forward edge of the antenna and by using a ground clutter reduction screen (described in Cooper et al. 1991). Ground clutter also can be reduced by selecting radar sites that are surrounded closely by trees, buildings, or low hills, or sites in a low depression, such as a shallow quarry or pit. These objects act as a radar fence that shields the radar from low-lying objects farther away from the lab and produces only a small amount of ground clutter in the center of the display screen. For further discussion of radar fences, see Eastwood (1967), Williams et al. (1972), Skolnik (1980) and Williams (1984).
More expensive radars usually have Moving Target Indicator (MTI) circuitry to reduce ground clutter. Although MTI is often extremely useful in bird studies, there are compli-cations: it does not fully suppress echo from moving vegetation, and it has complex effects on the echoes from moving birds as well. For instance, slow-moving bird echoes, or echoes moving perpendicular to the radar beam, may be suppressed by MTI. Further, birds flying over areas with heavy ground clutter echoes are less likely to be detected even though MTI eliminates the ground clutter (Richardson 1972; Buurma and Bruderer 1990).
Effects of Weather.-Rainy or snowy conditions make radar observa-tions of birds difficult to impossible, because the attenuation required to remove the echoes of the precipitation also removes most or all bird echoes. For X-band radars, there is no solution to this problem beyond designing sampling sessions to be short enough (15 to 30 min in length) so that some ses-sions could fit between periods of precipitation. S-band marine radar probably could de-tect birds (or at least larger flocks) in very light precipitation, but field verification is needed to confirm this. L-band radars are less sensitive to pre-cip-ita-tion, and birds flying in light precipitation are often detectable with such radars (W.J. Rich-ardson, LGL Ltd., pers. comm.).
Quantification of Low Flight Altitudes with Vertical Radar.-Despite radar modifications and improvements, data on the heights of birds flying at night below the lowest level sampled by the vertical radar (25 m) remain difficult to collect. The hinged assembly on which the vertical radar antenna is mounted makes it possible to sample lower elevations over water bodies or smooth, snow-covered fields by orienting the antenna horizontally. Over any other surface, however, ground-clutter echoes from vegetation or an uneven ground surface obscures the display screen. To my knowledge, a marine radar system that samples flight altitudes below 25 m over anything but water or a snow-covered field has not been developed (Korschgen et al. 1984; Gauthreaux 1985a,b; Cooper et al. 1991). Until such a system is devised, direct visual observation with night-vision or FLIR scopes, concurrent with vertical radar observations, probably is the best way to obtain altitude data for the lowest 25 m of airspace.
Standards for Equipment, Settings, and Methods
Equipment.-A marine radar system used to monitor birds should be X-band (3-cm wave-length), transmit with 10 to 25 kW of peak power, and have plotting and alarm functions. A color display monitor is an excellent feature to reduce observer fatigue (especially on surveillance radar), but monochrome displays are easier to videotape and are less expensive than are color monitors. The cost of one of these radars, including the modifications for the vertical or surveillance radars, would range from ~$8000 to ~$15,000 (US), exclud-ing installation. Night-vision equipment should be second- or third-generation equip-ment. Goggles can be used for very short distances, but scopes usually are more versatile for data collection. Forward-looking infrared scopes would be useful for bird observations, but the cost may be prohibitive. Image quality through good FLIR units is excellent, but the view through many of the smaller, less expensive infrared scopes is too grainy and small to be of much use for anything except observations within ~25 m.
Equipment Settings.-Once a radar is properly tuned, the gain must be set near the level where a light speckling appears on the screen. The STC (Sensitivity Time Control) should always be in the "off" position. The FTC (Fast Time Constant) should be "off" if possible, but can be applied sparingly to reduce some light clutter on the screen. Range should be set at 0.75-1.5 n.mi. on the vertical radar. On the surveillance radar, range should be set at 0.75-1.5 n.mi. to observe small-bodied species (e.g., songbirds) and at 3.0 n.mi. to observe large species. Some helpful advice on using marine radar for bird observations can be found in Williams (1984).
The importance of training in the use of radar for bird study cannot be stressed enough. It is relatively simple to learn how to operate the radar, but it takes training and extensive experience to learn how to adjust the radar properly under a variety of conditions and to learn how to interpret targets. Radar data gathered by untrained or insufficiently trained personnel are suspect and may be inadequate.
Radar Placement.-One of the most important and difficult-to-learn aspects of using surveillance radar is selection of sampling location. The site one chooses has important implications for data quality and comparability among sites. Basically, one needs to choose a site where ground clutter and shadow zones (e.g., areas behind hills or other objects that shield bird targets from radar) do not obscure or exclude important portions of the study area. Within a particular area, it usually is possible to find a particular site from which observations can be made, especially if "radar fences" are used. One additional technique that could allow greater flexibility in siting would be to mount the radar on a small crane that could be elevated to a desired height. This technique would be particularly useful in flat, heavily wooded areas.
Methods for Data Collection.-Data collection techniques for radar are discussed in Gauthreaux (1985a, this volume) and will not be extensively discussed here. One of the most important aspects of data collection, however, is to collect data in discrete sampling periods no longer than 30 min in length, at a standard range. By collecting data this way, one can standardize movement rates to targets/h/km, which allows compar-ison among studies. Further, time and weather data should always be collected, as these variables can be strongly correlated with movement rates and flight behavior. A sampling design for a visual and radar study to quantify bird movements and flight altitudes at proposed or existing wind-farms can be found in Cooper et al. (1995a).
Applications of Radar for Wind Power-Related Avian Research
Pre-construction Studies.-Siting: Locating windfarms in areas with few low-flying birds probably is the best solution for minimizing bird fatalities. Within an area of interest, radar and visual sampling should occur at a number of sites; the resulting data will provide a comprehensive, around-the-clock look at where "windows of movement" exist and will identify areas with heavy concentrations of low-flying birds. Radar also can be used on a microscale level to identify particular spots within a small area that have concentrations of low-flying birds. Both macroscale and microscale information would be useful in planning facility siting to minimize the potential for bird collisions or in reducing the concerns for collisions (either because all birds are high-flying or because few birds use an area).
Once specific sites are identified for structures such as wind turbines, radar can be deployed to measure the number and altitudes of birds passing through these specific corridors of air. This assessment may help identify critical (i.e., maximal) heights and locations for structures.
Identifying Periods of Risk: It may happen that an area is devoid of significant numbers of low-flying birds most of the time, but that there are certain seasons and/or weather conditions when significant numbers of birds do fly low enough to be at risk. Bird migration often is a pulsed phenomenon, and there are huge differences in both numbers of birds and their flight altitudes, depending on weather conditions, time of day, time of year, location, and the species under consideration. Radar and visual studies could be used to develop procedures to predict critical periods of high risk. Once the wind plant is operational, wind turbines could be shut down during periods when large numbers of low-flying birds are expected to pass through a windfarm, or plant operators could be alerted to watch for birds during those periods and shut down turbines if a large number of birds are at risk.
Post-construction Studies.-Monitoring impacts: Radar and visual studies can be used to assess post-construction changes in avian use or behavior over an area (day or night). Combined with ground searches, this type of study can help to establish mortality rates, estimate total flight and collision rates, and identify specific areas of concern at existing sites. If mortality occurred, one could determine if it occurred in proportion to bird use, or identify other factors that were involved.
Assessing effectiveness of collision reduction techniques: To assess the success of collision reduction techniques properly, it is necessary to know the number and altitude of birds flying over the area, in addition to number of collision victims. Radar and night-vision equipment can be used to monitor and measure the success of these techniques during periods of peak night-time use, under low light conditions, or when visual observations cannot cover a large enough area.
Real-time warning system to reduce bird collisions: Visual and radar monitoring could provide information so that schedules for wind power generation could be adjusted to adapt to periods when large numbers of low-flying birds are passing through a windfarm, either during the day or at night. Kenetech Windpower has supported studies to determine the feasibility of this technique at a wind plant in Spain. An automated radar system would be ideal for such a task, if it could be set up with an alarm to alert wind plant oper-ators when high-risk conditions were occurring. In related applications, marine radars have been used to detect the approach of waterfowl to contaminated ponds and to trigger bird scaring devices at those times (Denver Knight Piésold 1992; C. Johansen, Brigham Young Univ., pers. comm.). Also, software has been added to a few large military radar systems in order to monitor numbers of birds aloft; this information is used to help reduce the risk of collisions between aircraft and birds (Buurma and Bruderer 1990; Buurma 1995).
If radar becomes a standard technique for wind power-related avian research, it probably will be used by more than the small number of researchers who currently are familiar with it. To ensure quality of data, standardization of equipment, training, and data collection techniques would become even more critical than they are now. Efforts are being made to develop standards for equipment and data collection (S.A. Gauthreaux, Jr., Clemson University, pers. comm.) but no progress has been made in development of training protocols.
S-band marine radars need to be field tested to determine how well they work for detection of small birds, for reduction of insect echoes, and for monitoring birds during precipitation. The S-band marine radar may prove to be better than X-band radar in locations with many large insects or frequent precipitation, especially if the birds of interest are large-bodied species.
Efforts should be continued to develop and field test inexpensive, easy-to-use software that can automatically download radar information into a database. Such a system would streamline data collection and would decrease study costs. Similarly, an effort should be made to stay abreast of developments in the fields of radar, thermal imaging, night-vision, and computer technology, which promise additional benefits for wind power-related avian research.
Funds for preparation of this manuscript were provided by the National Renewable Energy Laboratory and ABR Inc. I thank Robert H. Day, Stephen M. Murphy and Robert J. Ritchie for their review of this manuscript, and I thank Paul Kerlinger for his insights regarding use of tracking radar. I also thank W. John Richardson for review and editing.
Able, K.P. 1982. Studies of avian nocturnal migratory orientation I. Interaction of sun, wind and stars as directional cues. Anim. Behav. 30:761-767.
Able, K.P. 1985. Radar methods for the study of hawk migration. p. 347-353 In: M. Harwood (ed.), Proc. Hawk Migration Conf. IV. Hawk Migration Assoc. North America, Washington, CT.
Beerwinkle, K.R., J.A. Witz and P.G. Schleider. 1993. An automated, vertical looking, X-band radar system for continuously monitoring aerial insect activity. Trans. Am. Soc. Agric. Eng. 36(3):965-970.
Blokpoel, H. 1971. The M33C track radar (3-cm) as a tool to study height and density of bird migration. Can. Wildl. Serv. Rep. Ser. 14:77-94.
Bruderer, B. and P. Steidinger. 1972. Methods of quantitative and qualitative analysis of bird migration with a tracking radar. p. 151-168 In: S.R. Galler et al. (eds.), Animal orientation and navigation: a symposium. NASA SP262. U.S. Government Printing Office, Washington, DC.
Bruderer, B., T. Steuri and M. Baumgartner. 1995. Short-range high-precision surveillance of nocturnal migration and tracking of single targets. Israel J. Zool. 41(3):207-220.
Buurma, L.S. 1995. Long-range surveillance radars as indicators of bird numbers aloft. Israel J. Zool. 41(3):221-236.
Buurma, B., and B. Bruderer. 1990. The application of radar for bird strike prevention. The Hague, The Netherlands. 75 p.
Cooper, B.A. and R.J. Ritchie. 1994. Wind power and birds: radar techniques for environ-mental assessment. p. 323-326 In: Windpower '94 Proceedings. American Wind Energy Association, Washington, DC.
Cooper, B.A. and R.J. Ritchie. 1995. The altitude of bird migration in east-central Alaska: a radar and visual study. J. Field Ornithol. 66(4):590-607.
Cooper, B.A., R.H. Day, R.J. Ritchie and C.L. Cranor. 1991. An improved marine radar system for studies of bird migration. J. Field Ornithol. 62:367-377.
Cooper, B.A., C.B. Johnson and R.J. Ritchie. 1995a. Bird migration near existing and pro-posed wind turbine sites in the eastern Lake Ontario region. Rep. from ABR Inc., Fairbanks, AK, for Niagara Mohawk Power Corp., Syracuse, NY. 71 p.
Cooper, B.A., C.B. Johnson and E.F. Neuhauser. 1995b. The impact of wind turbines on birds in upstate New York. Windpower '95 Proceedings. American Wind Energy Association, Washington, DC.
Denver Knight Piésold. 1992. Report of investigations/Radar-activated deterrent system/Migratory bird research project/Eureka County, Nevada. Report from Denver Knight Piésold, Denver, CO, for New-mont Gold Co. and Nevada Dep. Wild-life. 39 p. + Tables, Figures, Appendices.
Eastwood, E. 1967. Radar ornithology. Methuen, London, U.K. 278 p.
Gauthreaux, S.A., Jr. 1975. Radar ornithology: bird echoes on weather and airport surveil-lance radars. Clemson Univ. Press, Clemson, SC.
Gauthreaux, S.A., Jr. 1985a. Radar, electro-optical, and visual methods of studying bird flight near transmission lines. RP1636. Rep. from Clemson Univ., Clemson, SC, for Electric Power Res. Inst., Palo Alto, CA. 76 p.
Gauthreaux, S.A., Jr. 1985b. An avian mobile research laboratory: hawk migration studies. p. 339-346 In: M. Harwood (ed.), Proc. Hawk Migration Conf. IV. Hawk Migration Assoc. North America, Washington, CT.
Griffin, D.R. 1973. Oriented bird migration in or between opaque cloud layers. Proc. Am. Philosoph. Soc. 117:117-141.
Hamer, T.E., B.A. Cooper and C.J. Ralph. 1995. Use of marine radar to study Marbled Murrelets. Northwest Naturalist 76:73-78.
Kerlinger, P. 1980. A tracking radar study of migrating hawks. J. Hawk Migr. Assoc. N. Am. 2:34-42.
Kerlinger, P. 1982. Aerodynamic performance and flight speed selection of migrating hawks. Ph.D. thesis, State University of New York at Albany.
Kerlinger, P. and S. A. Gauthreaux, Jr. 1984. Flight behavior of Sharp-shinned Hawks during migration, I: Over land. Anim. Behav. 32:1021-1028.
Kerlinger, P. and S. A. Gauthreaux, Jr. 1985. Flight behavior of raptors during spring migration in south Texas studied with radar and visual observations. J. Field Ornithol. 56:394-402.
Korschgen, C.E., W.L. Green, W.L. Flock and E.A. Hibbard. 1984. Use of radar with a stationary antenna to estimate birds in a low-level flight corridor. J. Field Ornithol. 55(3):369-375.
Larkin, R.P. 1991. Flight speeds observed with radar, a correction: slow "birds" are insects. Behav. Ecol. Sociobiol. 29:221-224.
Larkin, R.P., D.R. Griffin, J.R. Torre-Bueno and J. Teal. 1979. Radar observations of bird migration over the western North Atlantic Ocean. Behav. Ecol. Sociobiol. 4:225-264.
Liechti, F., B. Bruderer and H. Paproth. 1995. Quantification of noctur-nal bird migration by moonwatching: comparison with radar and infra-red observations. J. Field Ornithol. 66(4):457-468.
McCrary, M.D., R.L. McKernan, W.D. Wagner and R.E. Landry. 1984. Nocturnal avian migration assessment of the San Gorgonio Wind Resource study area, fall 1992. Rep. for Southern Calif. Edison Co., Rosemead, CA. 87 p.
Pedersen, M.B. and E. Poulsen. 1991. En 90m/2MW vindmølles ind-virkning på fuglelivet. Fugles reak-tioner på opførelsen og idriftsæt-telsen af Tjære-borgmøllen ved Danske Vadehav. [Avian response to implementation of the Tjaereborg wind turbine at the Danish Wadden Sea]. Danske Vildt-undersøgelser, Hæfte 47, Dan-marks Miljø-unders-øgelser, Afdeling for Flora- og Fauna-økologi, Kalø. (Danish, English summary).
Richardson, W.J. 1972. Temporal variations in the ability of individual radars in detecting birds. Nat. Res. Counc. Can. Assoc. Commit. on Bird Hazards to Aircr., Field Note 61. 68 p.
Richardson, W.J. 1979. Radar techniques for wildlife studies. Nat. Wildl. Fed. Sci. Tech. Ser. 3:171-179.
Rogers, S.E., B.W. Cornaby, C.W. Rodman, P.R. Sticksel and D.A. Tolle. 1977. Environmental studies related to the operation of wind energy conversion systems. Rep. from Battelle Columbus Lab., Columbus, OH, for U.S. Dep. Energy, Solar Technol. & Wind Syst. Branch.
Skolnik, M.I. 1980. Introduction to radar systems. McGraw-Hill, New York, NY. 581 p.
Vaughn, C.R. 1985. Birds and insects as radar targets: a review. Proc. IEEE 73(2):205-227.
Williams, T.C. 1984. How to use marine radar for bird watching. Am. Birds 38:982-983.
Williams, T.C., J. Settel, P. O'Mahoney and J.M. Williams. 1972. An ornithological radar. Am. Birds 26:555-557.
Types of Radars.-An attendee asked whether already-existing airport radars could be used for broad-area surveillance, e.g. during site-selection studies. Other attendees noted that airport surveillance radars (ASR) routinely detect birds, as do most other types of radar. Air-port and other air surveillance radars have been widely used for bird studies for many years. However, as previously noted, ASR resolution is lower than that of short-range marine radars, and ASRs usually are not mobile. At most ASR sites, the bird information normally is not recorded or used in any way, variable radar settings may strongly affect bird detectability, and security restrictions often limit access.
Limitations and Calibration.-An attendee asked whether detection biases of marine radars had been studied by attempting to radar track birds that were radio-tagged. This has not been done, but could be a useful approach.
Another question concerned the possible use of an active acoustic sounder, as used by meteorologists, to detect birds aloft. It was noted that such sounders have been used to monitor micrometeorological phenomena relevant to bird flight while radar was used to monitor migrating birds. However, acoustic sounders are not known to be useful in monitoring birds themselves.
There was discussion of the fact that marine radar beams, as normally applied, are wide in the vertical dimension. As a result, when used in a surveillance mode, marine radars do not provide information on flight altitude.
One user of marine radar noted that, in one study, only 1 of 2000 birds seen passing through a radar beam was missed by the radar, and that radar detected many birds not seen vis-ually. Another user of marine radar mentioned that, in a different study, radar detected three bird targets for each one seen visually, and that it would be inefficient not to use radar.
In response to the question, "Would two experts in the use of radar get the same results from a radar study of bird movements", it was stated that they would get very similar results if they used the same equipment at the same site. Although the choice of radar settings is somewhat subjective, experienced radar users select similar settings. One radar user suggested that there is less individual-to-individual variation in bird detections when using radar than when observing visually.
Future Needs.-Several attendees noted that radar and visual studies of bird movements are complementary, and should be done in coordinated fashion. One commenter suggested that assessments of radar limitations and biases usually are done on an ad hoc basis, and need to be more systematic than has been the case in most radar studies.
One attendee commented on the need for automated procedures for digitizing, summarizing and archiving radar data on bird movements. He indicated that he is developing such a capability for a marine radar system.
