Applications of thermal imaging in avian science · Thermal imaging, or infrared thermography, has...
Transcript of Applications of thermal imaging in avian science · Thermal imaging, or infrared thermography, has...
Enlighten – Research publications by members of the University of Glasgow http://eprints.gla.ac.uk
McCafferty, D.J. (2013) Applications of thermal imaging in avian science. Ibis, 155 (1). pp. 4-15. ISSN 0019-1019 Copyright © 2012 The British Ornithologists’ Union Wiley OnlineOpen http://eprints.gla.ac.uk/73507/ Deposited on: 21 December 2012
Review article
Applications of thermal imaging in avian scienceDOMINIC J. MCCAFFERTY*
Institute of Biodiversity, Animal Health and Comparative Medicine, University of Glasgow, Glasgow, Scotland, UK
Thermal imaging, or infrared thermography, has been used in avian science since the1960s. More than 30 species of birds, ranging in size from passerines to ratites, havebeen studied using this technology. The main strength of this technique is that it is anon-invasive and non-contact method of measuring surface temperature. Its limitationsand measurement errors are well understood and suitable protocols have been developedfor a variety of experimental settings. Thermal imaging has been used most successfullyfor research on the thermal physiology of captive species, including poultry. In compari-son with work on mammals, thermal imaging has been less used for population counts,other than for some large bird species. However, more recently it has shown greater suc-cess for detection of flight paths and migration. The increasing availability and reducedcost of thermal imaging systems is likely to lead to further application of this technologyin studies of avian welfare, disease monitoring, energetics, behaviour and populationmonitoring.
Keywords: bird surveys, infrared thermography, thermal physiology, welfare.
Technology has an important role in answeringecological and behavioural questions in avian sci-ence. For example, our understanding of birdmigration has been transformed by the use ofsatellite tracking devices, seabird diving behaviourby the deployment of time depth recorders, andenergy expenditure by the use of heart rate loggers(Takahashi & Yoda 2010, Green 2011). Thesenew technologies are highly valued because theyallow measurement of the behaviour of free-rang-ing animals. One technology that is increasinglyused in avian research is thermal imaging, other-wise known as infrared thermography. Althoughthis technique was first used on birds as far backas the 1960s to identify sites of heat loss in Arcticspecies (Veghte & Herreid 1965), it is now widelyused to study thermoregulation, energy expendi-ture welfare for monitoring animal populations(Kastberger & Stachl 2003, Stewart et al. 2005,McCafferty 2007, Tattersall & Cadena 2010), andalso on a small scale in ecological and behaviouralstudies. Thermal imaging is one of a number oftechnologies applicable to nocturnal studies of
birds, including spotlighting, radar and imageenhancement, all of which have strengths andweaknesses (Allison & Destefano 2006). Cost islikely to have been one of the most important rea-sons why this technology has not been widely usedin field research. However, thermal imaging cam-eras are becoming more affordable, ranging inprice from hand-held detection cameras (< £1000)to high level research and development systems(>£30 000).
This review is aimed at researchers who may beinterested in using thermal imaging in laboratoryor field studies of birds, and builds on earlier workwhich focused on mammals and the use of ther-mal imaging in biophysical modelling (McCafferty2007, McCafferty et al. 2011). A literature searchwas undertaken using knowledge of existing litera-ture, together with a recent reference search usingWeb of KnowledgeTM. This was not intended toproduce an exhaustive review of the literature butto examine critically the use of thermal imagingfor avian research, provide guidelines for its appro-priate use, and identify research questions that canbe tackled using this technology.
*Email: [email protected]
© 2012 British Ornithologists’ Union
Ibis (2013), 155, 4–15
THERMAL IMAGING: THEORY ANDMEASUREMENT
Thermal imaging involves the detection of infraredradiation (heat) emitted from an object (Speakman& Ward 1998). Any object that has a temperatureabove absolute zero (�273.15 °C, or 0 °K) emitsinfrared radiation and the temperature of theobject will determine the wavelength of the radia-tion emitted. At biologically relevant temperatures(�40 to 100 °C) radiation peaks in the infraredwavelength (longwave) range of approximately 8–12 lm. Infrared radiation is readily absorbed by waterand therefore thermography can only be used in airand not for objects underwater. Thermal cameras workby measuring infrared radiation R (W/m2) and com-puting temperature T (°K), from the Stefan BoltzmannLaw: R = erT4, where e is the emissivity of the sur-face and r is the Stefan Boltzmann constant(5.67 9 10�8 W/m2/K4). Emissivity is defined as theability of a surface to emit and absorb radiation (Mon-teith & Unsworth 1990). The emissivity of biologicaltissues is close to a blackbody (perfect emitter) andfor plumage has been recorded in the range 0.96–0.98(Hammel 1956, Best & Fowler 1981). Plumage colourdoes not influence emissivity but may influence solarabsorbance and hence dark plumages may have higher
surface temperature than pale-coloured plumages instrong sunshine (McCafferty 2007).
A thermal imaging camera produces a grey orcolour-scaled image made up of pixels that repre-sent individual temperatures (Fig. 1). With mostsystems there is a choice of colour palette avail-able. Some palettes are only suitable for colourdisplay/printing, whereas proportional intensitypalettes can be used for both colour and greyscaleprinting. Thermal imaging records only surfacetemperature and for un-feathered areas this is thetemperature of the skin surface or tissue. How-ever, plumage is not a smooth solid surface andradiation will be emitted from different layersfrom within the feather matrix. Depending onspacing between outer feathers, the temperaturerecorded is effectively the temperature of theplumage a few millimetres below the physical sur-face of the plumage (McCafferty et al. 2011).
Selection of thermal camera will depend onmeasurement requirements of your study. Themain choices relate to whether the researcherwishes to make temperature measurements or sim-ply detect the difference between warm and coldsurfaces, what temperature and spatial resolution isrequired, what the measurement distance is, andwhether the researcher wishes to record still or
a
b c
Figure 1. Thermal images showing (a) sequence of female Barn Owl Tyto alba in flight (AGA 782), (b) Zebra Finch Taeniopygia gut-tata clutch in nestbox (FLIR E300), and (c) thermal survey of Feral Pigeons Columba livia (FLIR E300).
© 2012 British Ornithologists’ Union
Thermal imaging in avian science 5
moving images. An un-cooled camera capable ofresolving 0.1 °K and with image resolution of320 9 240 pixels or greater is appropriate formost applications in avian science where tempera-ture measurement is required. Where greater tem-perature resolution, < 0.05 °K, and higher imagecapture frequency is needed, a cooled camerashould be used. Many high level infrared camerascan use a variety of lenses including wide-angle,telephoto and macro. This will influence the spa-tial resolution; for example, each pixel with awide-angle lens at a given distance will represent alarger spatial area compared with a lens with smal-ler angle of view. Most cameras will record stillimages in jpeg or similar format. These are radio-metric files that contain thermal information foreach pixel that is translated into colour/greyscale.This means that radiometric images can be re-scaled post hoc without losing any measurementprecision or content. Higher level cameras willallow thermal images to be recorded to memoryor streamed to a computer as a radiometric videofile. Again, this file format contains temperaturedata for each pixel rather than a standard videofile. For detecting or counting birds, temperaturedata are not necessarily required and a surveillance
system that simply identifies a ‘hot body’ against acool background may be the most appropriate.
The strengths and limitations of thermal imag-ing for measuring surface temperature or locatinganimals are well understood (Table 1). Thermalimaging is highly suited for work on captive birdsthat can be approached to within a few metres. Itis possible to take images through metal aviarycages where the distance between bird and cage isrelatively small relative to the distance betweencage and camera. The cage wire will block some ofthe infrared radiation emitted from the bird andthe temperature of the wire cage will introduceinaccuracies in measurement if the camera is heldclose to the cage. Calibration can be made usingan object of known temperature in these circum-stances. Glass is transparent to visible wavelengthsbut opaque to infrared. Specialised transparentinfrared windows composed of a number of metal-based compounds allow transmission of infraredradiation of different wavelengths (see: http://iriss.com/) A thin polyethylene film (e.g. a Ziploc©freezer bag) allows the partial transmission ofinfrared radiation, for example for studies in meta-bolic chambers. These materials selectively absorbcertain wavelengths, giving inaccurate tempera-
Table 1. Summary of advantages, limitations and constraints in study design for thermal imaging. Further details in McCafferty(2007) and McCafferty et al. (2011).
AdvantagesNon-invasive, remote measurement of surface temperatureLong-distance detection in open habitats at night, avoiding disturbance from spotlightsHigh spatial (e.g. 320 9 240 pixels) and temperature (< 0.1 °C, accuracy ± 2% of full temperature range) resolutionPortable devices, still or moving thermal images recordedAnalytical software for image analysis and summary statistics
LimitationsSurface temperature dependent on environmental variables (solar and reflected radiation), precipitation and evaporation. IRradiation absorbed by water and cloud coverLimited detection probability with high environmental temperatures and daytime detection poor due to thermal heterogeneity ofenvironmentTemperature of bird influenced by time of day, age, stress and range of physiological factorsLow temperatures recorded at edge of curved surface (Watmough et al. 1970)Expensive (although prices have fallen)Low spatial resolution of temperature at distanceThermal signals blocked by vegetation or terrain
Constraints in study designTake measurements in controlled conditions and allow animal to equilibrate with surroundingsMeasure temperature of reference blackbodyRecord air temperature and relative humidityMinimized environmental variation by recording in dark/low light. Control for physiological variablesFor detection work, thermal contrast greatest at night against cold substrate or clear skyFor diagnosis of disease or injury look for asymmetry in thermal patterns (e.g. adjacent limbs)Use area analysis tools to avoid curvature effects or take readings from centre of distant objects
© 2012 British Ornithologists’ Union
6 D. J. McCafferty
tures, but it is possible to calibrate for this usingobjects of known surface temperatures. The envi-ronmental temperature on the camera side of theplastic film should remain constant because someinfrared radiation is reflected from the room ontothe plastic film (Tattersall & Milsom 2003, Scottet al. 2008). Most cameras are designed for work-ing in industrial environments which also makesthem suitable for fieldwork with high humidityand extremes of temperature. Thermal camerascan be used for bird detection at considerable dis-tances. Desholm (2003) was able to detect a duckat distances of up to 3000 m and migrants havebeen recorded from the ground at 3000 m altitude(Zehnder et al. 2001).
Depending on the choice of system, images canbe recorded manually, programmed to record atset time intervals or recorded as moving imagesonto computer/storage device. It is recommendedthat visual images are recorded along with thermalimages for subsequent analysis. Image analysis canbe undertaken with proprietary software availableto derive summary statistics from different regionsof the image (McCafferty et al. 2011). This allowspost hoc adjustment of environmental parameterssuch as temperature and humidity that influenceradiative properties. Statistical data (e.g. mean,minimum, maximum) can be obtained quickly byfitting polygons from regions of interest in theimage. Analytical software is becoming increasinglycomplex, with functionality for time sampling andtarget tracking for automated analytical proce-dures.
OTHER INFRARED TECHNOLOGIES
Other technologies that may be confused withthermal imaging are (1) infrared photography,which uses photographic film/digital sensor sensi-tive to near-infrared light (0.7–0.9 lm), (2) night-infrared cameras that are sensitive to visible lightand near-infrared where the visual field is illumi-nated by an infrared beam or (3) night visiondevices that rely on the collection and amplifica-tion of small amounts of light in the visible andnear-infrared spectrum (Kastberger & Stachl2003). Night cameras are effective where filmingor detection of nocturnal animals from a fixedpoint is required (e.g. camera traps or CCTV).Night vision devices (e.g. night vision goggles) areuseful for navigation or searching for birds. Thesethree systems do not measure temperature but are
cost-effective for behavioural observations. Finally,recent mobile phone and computer cameras pro-duce pseudo-thermal images that simply substitutecolours to give the appearance of a thermal image.
APPLICATIONS
Thermal physiology
The main emphasis of early thermography studiesof birds was to identify variation in heat loss andinsulative capacity of different areas of plumage(Table 2). Traditionally, solid sensors such asthermocouples and thermistors have been used tomeasure plumage or skin temperature (Goldsmith& Sladen 1961). These are inexpensive but a largenumber of sensors would be required to give goodspatial coverage and solid sensors may disrupt andinfluence the insulation properties of the plumage(Cena 1974) as well as interfering with movementof the bird. Infrared thermometers that measureinfrared radiation emitted from an object can alsobe used to obtain single spot/area temperaturemeasurements (Wilson et al. 1998). It is also possi-ble to measure body conductance from coolingcurves of dead birds (Herreid & Kessel 1967) andthe thermal conductivity of plumage samples on aheated flat plate or cylinder (Scholander et al.1950, Ward et al. 2007). More realistic measuresof metabolic heat loss and operative temperatureof birds can be made using heated taxidermicmounts (Bakken et al. 1999). The obvious advan-tage of thermal imaging is that it provides mea-surement of the temperature across the entirebody of a live bird without the need for restraintor capture.
Thermal imaging has shown that areas aroundthe head and in particular the eye-auricular regionhave the highest temperature, while well–featheredregions are closer to ambient temperature. Rela-tionships between surface temperature and airtemperature for different parts of the body havebeen used to estimate coat conductance and meta-bolic heat loss (Veghte & Herreid 1965, Hill et al.1980). For species with large feather-free areas,such as the ratites, regions of the bill, leg and neckwere estimated to dissipate 40% of metabolic heat,even though these structures represent < 20% ofsurface area (Phillips & Sanborn 1994). Similarlyin the Humboldt Penguin Spheniscus humboldti,flippers and feet are important thermal windowsfor heat dissipation in warm conditions (Despin
© 2012 British Ornithologists’ Union
Thermal imaging in avian science 7
et al. 1978). Domestic Fowl Gallus domesticusshowed warmest areas on bare skin (bill, face, wat-tles and feet) and cooler regions on well-insulatedfeathered areas. Average surface temperature wasshown to decrease with chick age (due to
increased insulation from feather development),but there was greatest variability in temperatureacross the body at late stages of development(Cangar et al. 2008). Posture (bill, leg and feetcovering) and piloerection (raising of plumage)
Table 2. Applications of thermal imaging for research on the thermal physiology and welfare of birds.
Species Research topic Thermal imaging system Reference
Anna’s Hummingbird Calypteanna
Modelling metabolic rate Fluke, Everett, WA Evangelista et al. (2010)
Arctic birds (four species) Heat loss and air temperature Barnes Model I-8A IRradiometer
Veghte and Herreid (1965)
Barn Owl Tyto alba Temperature variation andheat loss
AGA 782 Thermovision McCafferty et al. (1998)
Black-capped Chickadee Parusatricapillus
Heat loss and air temperature AGA 680/102B Thermovision Hill (2009), Hill et al. (1980)
Domestic Fowl Gallus domesticus Heat and water stress – Yahav et al. (2005)Domestic Fowl Gallus domesticus Naked neck and feathered
birdsIRTIR, Inframetrics 760 Yahav et al. (1998)
Domestic Fowl Gallus domesticus Bumblefoot screening ThermaCAM PM695, FLIRSystems
Wilcox et al. (2009)
Domestic Fowl Gallus domesticus Feather cover score ThermaCAM S60, FLIRSystems
Cook et al. (2006)
Domestic Fowl Gallus domesticus Spatial patterns and age AGEMA Thermovision 570 Cangar et al. (2008)Domestic Fowl Gallus domesticus Metabolic heat production and
dietTesto® 880 Ferreira et al. (2011)
Domestic Fowl Gallus domesticus Skin temperature variation Inframetrics, model 760, FLIRSystems
Tessier et al. (2003)
Domestic Fowl Gallus domesticus Egg temperature and coldexposure
Model PM545, FLIR Systems Shinder et al. (2009)
Domestic Fowl Gallus domesticus Facial and cloacal temperature PM545; FLIR Systems Giloh et al. (2012)Domestic Fowl Gallus domesticus Temperature response to
distressE4 ThermaCAM, FLIRSystems
Edgar et al. (2011)
Domestic Fowl Gallus domesticus Temperature response topalatable reward
Thermovision A40, FLIRSystems
Moe et al. (2012)
House Finch Carpodacusmexicanus
Effect of wind AGEMA, 782 Thermovision Zerba et al. (1999)
Humboldt Penguin Spheniscushumboldti
Temperature responses tocold
AGA 680 Thermovision Despin et al. (1978)
Mallard ducklings Anasplatyrhynchos
Prediction of bodytemperature
ThermaCAM PM 575, FLIRSystems
Bakken et al. (2005)
Mallard ducklings Anasplatyrhynchos
Heat loss during swimming ThermaCAM PM 575, FLIRSystems
Banta et al. (2004)
Mallard ducklings Anasplatyrhynchos
Effect of radio-transmitters Model 525, Inframetrics Bakken et al. (1996)
Nests (12 species) Cooling rate of eggs THI-300, Tasco, Japan Lamprecht and Schmolz(2004)
Ratites (three species) Heat loss andthermoregulation
Inframetrics 525 Phillips and Sanborn(1994)
Common Starling Sturnusvulgaris
Metabolism and heat loss inflight
AGEMA, Thermovision 880 Ward et al. (1999, 2004)
Toco Toucan Ramphastos toco Heat loss from bill Mikron Instruments Model7515
Tattersall et al. (2009)
Waterfowl (three species) Responses to hypoxia Mikron Instruments Model7515
Scott et al. (2008)
White Peking Duck Anasplatyrhynchos
Bill thermoregulation AGA Thermovision 680 Hagan and Heath (1980)
Zebra Finch Taeniopygia guttata Egg and brood patch temperature E300, FLIR Systems Hill (2009)
© 2012 British Ornithologists’ Union
8 D. J. McCafferty
influenced exposure of regions of high heat lossand increased thickness of the insulative layer(Veghte & Herreid 1965, Hill et al. 1980).
Thermal imaging can provide new insights intothe functions of different parts of the body. Haganand Heath (1980) showed that the temperature ofthe bill decreased with air temperature and varia-tion across the bill was associated with arterial vas-culature. More recently, Symonds and Tattersall(2010) revealed that the bill of the tropical TocoToucan Ramphastos toco also acts as a thermalradiator. On average, 30–60% of body heat is dissi-pated by blood flow to this region. Adults havethe ability to control heat loss from the bill from5% in the cold to up to 100% of total heat loss inwarm conditions (Tattersall et al. 2009). In hyp-oxic (low oxygen) conditions, birds control periph-eral heat loss by warming the bill to bring about areduction in body temperature. Bar-headed GeeseAnser indicus, a high-altitude species, toleratedgreater hypoxia before body temperature wasreduced compared with lower altitude species(Scott et al. 2008). Thermal imaging can be avaluable tool to examine energy expenditure, andenergy costs of different behaviours that cannot bemeasured on restrained birds. For example, therelationship between activity and metabolic heatproduction has been investigated in a number ofthermal imaging studies. Wind tunnel experimentswith House Finches Carpodacus mexicanus indi-cated that there was no difference between thesurface temperature of resting and active birdsexposed to the same convective conditions, indi-cating that exercise-generated heat is likely to be acommon mechanism of thermoregulation in natu-ral conditions (Zerba et al. 1999). McCaffertyet al. (1998) showed that considerable amounts ofheat are released from flight muscles. For a BarnOwl Tyto alba the temperature was greatest over-lying the main flight muscles, from both the pecto-ralis and the ventral wing region. Heat generatedfrom the leg muscles was also obvious during take-off (Fig. 1a). Ward et al. (1999) took thermalimages of Common Starlings Sturnus vulgaris in awind tunnel and using a heat transfer model dem-onstrated that during flight, heat loss from radia-tion accounted for only 8% of total heat loss,whereas convection accounted for 81%. Most heatloss occurred from the wings and the trailing feetcould increase heat loss by up to 12%. These esti-mates of total heat loss agreed well with metabolicrates measured by doubly labelled water and oxy-
gen consumption (Ward et al. 2004), giving confi-dence in the use of heat transfer modelling forstudying avian energetics. In Anna’s Humming-birds Calypte anna, Evangelista et al. (2010) com-pared heat loss by convection and radiationderived from thermal images, metabolic rate usingflow-through respirometry and aerodynamic powercalculated from wing beat kinematic data andfound they gave similar estimates of power output.However, at higher temperatures heat transfermodels underestimated metabolic heat output andindicated that evaporative and heat storagechanges as well as enhanced convective heat lossfrom the wings may have accounted for this differ-ence.
Thermal imaging can also be used to estimateindirectly the internal body temperature of birds.Bakken et al. (2005) showed that by removing asmall area of plumage (� 1 cm2) from the scalpof 2–3-day-old Mallard Anas platyrhynchos duck-lings, cloacal temperature could be estimated towithin 1 °C using only scalp temperatureand ambient temperature at wind speeds up to2.5 m/s. In this study there was concern that con-vective cooling from the shaved area would occur.However, model tests using a polyethylene con-vective shield (0.025 mm) over the measurementsite increased measurement error. The investigatorssuggested that the relationship between scalp tem-perature and internal temperature may declinewith body size as tissue thickness increases. It isunlikely to be a substitute for direct measurementsof core temperature in birds but where appropri-ate may minimise effects on behaviour and physi-ology associated with instrument attachment,surgery or handling. As yet, this technique has notbeen widely used but it has provided a non-inva-sive method of measuring the internal temperatureof ducklings, estimating that a minimum of 70% ofmetabolic heat production is lost to water (Bantaet al. 2004).
Thermography and heat transfer modelling hasbeen extended to examine the energy cost of dif-ferent behaviours. Ward and Slater (2005) usedthis approach to estimate that male Canaries Seri-nus canaria increased their metabolic rate by 14%when singing. Previously the energy cost of differ-ent behaviours had been confined to studies incaptivity; however, this approach could also beapplied to estimate the energy cost of differentbehaviours in free-ranging birds (e.g. courtship dis-plays, fighting) or different stages of the lifecycle
© 2012 British Ornithologists’ Union
Thermal imaging in avian science 9
such as moult. There is therefore considerablescope for estimating the relative energy cost ofdifferent behaviours or environmental conditions(e.g. precipitation, wind) that are either difficult orimpossible to measure with respirometry or doublylabelled water methods. Further details onapproaches to heat transfer modelling are given inMcCafferty et al. (2011).
There has been relatively little use of thermalimaging for examination of egg temperature.Lamprecht and Schmolz (2004) evaluated theinsulating effect of different types of nests bydetermining the cooling rate of eggs. Thermalimaging revealed strong thermal gradients withinthe clutch between relatively warm inner contactregions and cooler outer regions (Fig. 1b). Clearly,thermal imaging cannot measure the temperatureof the clutch while a bird is incubating and there-fore solid sensors attached to eggs or model eggsmust be used. However, thermal imaging can pro-vide further information on temperature variationwithin a clutch. Hill (2009) successfully used ther-mography in the laboratory to monitor egg incuba-tion temperature of the Zebra Finch Taeniopygiaguttata by taking into account the time incubatingbirds were off the nest. This approach could beapplied in field studies by using an automatedthermal imaging system to record egg temperaturewhen birds left the nest. Moreover, thermal imag-ing may be able to answer questions related to theenergetic cost of incubation. Hill (2009) showedthat female but not male Zebra Finches had higherabdominal temperatures when incubating enlargedclutches, suggesting that females showed a greaterphysiological response to the demands of incuba-tion. Surprisingly, thermal imaging has rarely beenused to study the temperature of nestlings: Ovadiaet al. (2002) found that the energy cost of begginginferred from surface temperature was not affectedby nestling rank in House Sparrows Passer domesti-cus. There appears to be considerable scope forstudies on the development of thermogenic capac-ity, social thermoregulation and energetic conse-quences of nest position (see comparative studieson Rabbit Oryctolagus cuniculus pups: Gilbert et al.2012).
Welfare
In contrast to the extensive use of thermographyfor detection of disease and injury in farm, zooand domestic mammals (Eddy et al. 2001, Stewart
et al. 2005, Hilsberg-Merz 2007), there have beenfew studies on birds. Wilcox et al. (2009) foundthat thermal imaging was successful in detectingcases of bumblefoot (chronic inflammation: poder-matitis), and better at detecting cases at the sub-clinical infection stage than by eye. Thermalimaging was also an effective method of assessingfeather cover of laying hens (Cook et al. 2006).Thermography has the potential to examine thehealth and welfare of wild birds, especially toassess research impacts. Bakken et al. (1996) usedthermography to reveal disruption of feathers byradio-transmitters mounted by harnesses on Mal-lard ducklings (although this did not increasemetabolic rate). Thermal imaging would be suitedto further studies examining the possible impactsof other externally mounted sensors or assessingrecovery from surgery. Validation of using eye andfacial temperature to study stress-induced hyper-thermia may also allow the development of a use-ful tool of monitoring avian welfare (Fig. 2). Forexample, when hens were exposed to a minorstressor, eye and comb temperature decreased andwhen their chicks were exposed to the same stres-sor, hens showed an empathetic decrease in eyetemperature associated with an increase in heartrate and time spent standing alert (Edgar et al.2011). Animals respond to stress or pain by rapidchanges in their pattern of blood flow, from theperiphery to the core via sympathetically mediatedvasoconstriction, which reduces skin temperature(Stewart et al. 2008). The skin is the first body
Figure 2. Head of Domestic Fowl Gallus domesticus showingspatial variation of head, comb and wattles (FLIR SC640, highresolution 640 9 480 pixels).
© 2012 British Ornithologists’ Union
10 D. J. McCafferty
region to be affected by any injury and the short-term metabolic requirements of the skin are notessential, so that blood flow can be used for othermetabolic requirements. In some species, percep-tion of psychosocial threats may also illicit cutane-ous vasoconstriction (Blessing 2003). Recently,thermal imaging showed that comb temperaturemay also change in response to positive anticipa-tion and consumption of palatable rewards (Moeet al. 2012). These recent findings indicate thepotential for applying methods developed on poul-try to monitor physiological stress in laboratory orwild bird populations, for example to detect theimpact of disturbance or handling on wild birds asa comparative approach to heart rate monitoring(Nimon et al. 1996).
Behaviour
Thermal imaging is advantageous for studying thebehaviour of birds at night (Table 3). Kuwae(2007) monitored the nocturnal feeding behaviourof Kentish Plovers Charadrius alexandrines, as thethermal camera was able to pick up warm excretawhen deposited on the ground. Defecation ratewas therefore used as an index of feeding rate.Similarly, Tillmann (2009a) used thermal imagingto study the relationship between defecation andescape behaviour in roosting coveys of GreyPartridges Perdix perdix, and also found that anti-predator behaviour was different at nightcompared with day in terms of roost location, cir-cadian shifts in behaviour, tightness of groups andflight distances to avoid predation (Tillmann2009b). Further insights into the behaviour ofboth nocturnal species and the nocturnal roostingbehaviour of diurnal species may be possible withthermal cameras. Although the cost of thermalcameras is high compared with conventional infra-red CCTV, no infrared source is required for illu-mination and therefore greater viewing distance ispossible (Kuwae 2007).
Field surveys
Relatively few studies have used thermal imagingfor field surveys. This may be because more cost-effective alternatives are available for counting andlocating birds. Thermal surveys can be used tolocate nests, individuals and flocks of birds in avariety of open habitats (Fig. 1c, Table 3). Thesecan be conducted on foot, from a boat or, where
greater spatial coverage is required, even from air-craft. Early studies were not encouraging for thistechnique because of vegetation blocking the ther-mal signal, especially for cavity nesters, and practi-cal considerations such as the large size and cost ofcameras (Boonstra et al. 1995). Since then, therehave been a number of successful thermal aerialsurveys of some large species (turkeys and cranes).More recently, portable hand-held cameras havebeen used to locate nests and nestlings (Galliganet al. 2003, Mattsson & Niemi 2006). Most suit-able thermal surveys are for ground-nesting specieswhere incubating birds can be located in relativelyopen habitats. A study of Malleefowl Leipoa ocella-ta used an airborne thermal scanner to count incu-bator nest mounds in Eucalyptus scrub(Benshemesh & Emison 1996). Malleefowl opentheir incubation mound on most mornings toadjust the incubation temperature and aerate themound, and the surveys were able to locate up to36% of active mounds.
Thermal imaging from a boat has increased thecatch rates of some species at night (Mills et al.2011). Small hand-held devices produced forhunting show considerable scope for studies inwhich birds must be caught at night and mayavoid some of the disturbance effects of spot-light-ing. There is interest in the use of thermal surveysfor the purposes of environmental impact assess-ments. Desholm (2003) used an automated systemmounted on a wind turbine for monitoring colli-sion rates of birds at an offshore wind farm andfrom thermal signatures was able to differentiatebetween species ranging in size from passerines toducks. However, due to the considerable researchand development costs of this system, there hasbeen limited application. Nevertheless, EuropeanNightjar Caprimulgus europaeus flight altitude anddirections have been tracked readily with a ther-mal camera on a tripod to assess possible collisionsfrom wind turbines (Calbrade & Henderson 2009).Combined thermal imaging and radar systems havealso been deployed to count large numbers ofbirds on nocturnal migration (Zehnder et al. 2001,Gauthreaux & Livingston 2006). Advanced imagedetection systems have been successfully used tocount large bat roosts (Hristov et al. 2005, Betkeet al. 2008) and similar approaches could be takenfor population counts of birds at night in openhabitats. Some analytical methods of image detec-tion based on wing beat frequency have beenexplored (Lazarevic 2009) but there is a need for
© 2012 British Ornithologists’ Union
Thermal imaging in avian science 11
Table3.
Field
survey
san
dbe
haviou
rals
tudies
ofwild
birdsus
ingthermal
imag
ing.
Spe
cies
Res
earchTop
icHab
itat
Max
imum
distan
ce/altitude
(m)
The
rmal
imag
ingsy
stem
Autho
r(s)
Ten
spec
ies(ran
geof
body
size
s)Bird
detectionsy
stem
Offs
hore
wind
farm
3000
The
rmov
isionIRMV32
0VDes
holm
(200
3)
Ninesp
ecies(ran
geof
body
size
s)Cou
ntsan
dne
sts
Trees
,op
enha
bitat
andArctic
tund
ra
–The
rmov
ision21
0Boo
nstra
etal.(199
5)
Ammod
ramus
sparrows(two
spec
ies)
Nes
tsGrass
land
3–5
The
rmaC
AM
PM57
5,FLIR
Sys
tems
Galligan
etal.(200
3)
Clapp
erRailR
allus
long
irostris
Cap
ture
succ
ess
Tidal
marsh
es–
The
rmal-Eye
250D
and
X20
0xp,
L-3
Com
mun
ications
Infrared
Produ
cts
Millset
al.(201
1)
GreyPartridg
ePerdixpe
rdix
Night-tim
eroos
tingan
dan
ti-pred
ationbe
haviou
rArableland
250
Palm
IR-250
D,Ray
theo
nTillman
n(200
9a,b)
Ken
tishPlove
rCha
radrius
alex
andrinus
Noc
turnal
beha
viou
r:de
feca
tionrate
Intertidal
flat
–The
rmaC
AM
SC-300
0with
telesc
opic
lens
,FLIR
Sys
tems
Kuw
ae(200
7)
Malleefow
lLeipo
aoc
ellata
Aerials
urve
yof
nests
Fores
tan
dsh
rubs
305
Dae
dalus12
40/60
Sca
nner
Ben
shem
eshan
dEmison
(199
6)Europ
oean
Nightjar
Cap
rimulgu
seu
ropa
eus
Bird
flight
heightsfor
windfarm
survey
Fores
t10
0–20
0P64
0,FLIR
Systems
Calbrad
ean
dHen
derson
(200
9)Not
give
nMigratio
nco
unts
Ope
nsky
2800
Rad
ianc
e1,
Ambe
r,Ray
theo
nGau
threau
xan
dLiving
ston
(200
6)Not
iden
tified
Noc
turnal
migratio
nco
unts
Coa
stal
3000
LORIS,IRTV-445
,Infram
etric
sZeh
nder
etal.(200
1)
Ove
nbird
Seiurus
auroca
pillus
HermitThrus
hCatha
rusgu
ttatus
Nes
tsan
dfled
glings
Fores
t30
NIG
HTSIG
HTTMPalmIR
250,
Ray
theo
nSystems
Mattsso
nan
dNiemi
(200
6)San
dhillCrane
Grusca
nade
nsis
Aerials
urve
y:floc
kdistrib
utionan
dde
nsity
River
300–
1200
Mitsub
ishi
IR-M
600
Kinze
letal.(200
6)
Wild
Turke
yMelea
gris
gallopa
voAerials
urve
yUplan
ds39
820
00A/B
FLIR
Systems
Garne
ret
al.19
95
© 2012 British Ornithologists’ Union
12 D. J. McCafferty
further development of automated avian surveil-lance systems. Development of appropriatesystems for bird monitoring may benefit fromcommercial testing of thermal surveillance systems.For example, applications to automatically detectanimals during harvesting (Steen et al. 2012) andas early warning systems for bird strike prevention(Müenzberg et al. 2011) show promise in thisfield.
CONCLUSIONS
The development and reduced cost of thermalimaging has led to greater exploitation of this tech-nology in pure and applied research. The strengthsand limitations for research on birds and other ani-mals are well understood. Thermography is mostfrequently used on captive species in studies ofthermal physiology. However, the ability to mea-sure temperature non-invasively and remotely hasconsiderable advantages for further behaviouraland ecological studies on free-ranging birds, espe-cially studies of avian welfare, disease, energetics,behaviour and population monitoring.
I am grateful to John Currie of Edinburgh Napier Uni-versity for sharing his enthusiasm for thermal imagingand making available equipment over many years.Davina Hill kindly provided the thermal image of theZebra Finch clutch and Nigel Peart (FLIR systems) theimage of Domestic Fowl. Thanks to Ruedi Nager for afirst critical reading of the manuscript, and two anony-mous reviewers who also provided helpful comments onthe manuscript.
REFERENCES
Allison, N.L. & Destefano, S. 2006. Equipment andtechniques for nocturnal wildlife studies. Wildl. Soc. B 34:1036–1044.
Bakken, G.S., Reynolds, P.S., Kenow, K.P., Korschgen, C.E. & Boysen, A.F. 1996. Thermoregulatory effects ofradiotelemetry transmitters on Mallard ducklings. J. Wildl.Manage. 60: 669–678.
Bakken, G.S., Reynolds, P.S., Kenow, K.P., Korschgen, C.E. & Boysen, A.F. 1999. Standardization and calibration ofheated mounts illustrated with day-old Mallard ducklings.Physiol. Biochem. Zool. 72: 502–506.
Bakken, G.S., Van Sant, M.J., Lynott, A.J. & Banta, M.R.2005. Predicting small endotherm body temperatures fromscalp temperatures. J. Therm. Biol 30: 221–228.
Banta, M.R., Lynott, A.J., VanSant, M.J. & Bakken, G.S.2004. Partitioning heat loss from mallard ducklingsswimming on the air–water interface. J. Exp. Biol. 207:4551–4557.
Benshemesh, J.S. & Emison, W.B. 1996. Surveyingbreeding densities of Malleefowl using an airborne thermalscanner. Wildl. Res. 23: 121–142.
Best, R.G. & Fowler, R. 1981. Infrared emissivity and radiantsurface temperatures of canada and snow geese. J. Wildl.Manage. 45: 1026–1029.
Betke, M., Hirsh, D.E., Makris, N.C., McCracken, G.F.,Procopio, M., Hristov, N.I., Tang, S., Bagchi, A.,Reichard, J.D., Horn, J.W., Crampton, S., Cleveland, C.J.& Kunz, T.H. 2008. Thermal imaging reveals significantlysmaller Brazilian free-tailed bat colonies than previouslyestimated. J. Mammal. 89: 18–24.
Blessing, W.W. 2003. Lower brainstem pathways regulatingsympathetically mediated changes in cutaneous blood flow.Cell. Mol. Neurobiol. 23: 527–538.
Boonstra, R., Eadie, J.M., Krebs, C.J. & Boutin, S. 1995.Limitations of far-infrared thermal imaging in locating birds.J. Field Ornithol. 66: 192–198.
Calbrade, N. & Henderson, I. 2009. A survey of Nightjar flightheights at Clocaenog Forest in 2009. BTO Res. Rep. 542.
Cangar, O., Aerts, J.M., Buyse, J. & Berckmans, D. 2008.Quantification of the spatial distribution of surfacetemperatures of broilers. Poult. Sci. 87: 2493–2499.
Cena, K. 1974. Radiative heat loss from anaimals and man. InMonteith, J.L. & Mount, L.E. (eds) Heat Loss from Animalsand Man: 33–58. London: Butterworth.
Cook, N.J., Smykot, A.B., Holm, D.E., Fasenko, G. &Church, J.S. 2006. Assessing feather cover of laying hensby infrared thermography. J. Appl. Poult. Res. 15: 274–279.
Desholm, M. 2003. Thermal Animal Detection System(TADS). Development of a method for estimating collisionfrequency of migrating birds at offshore wind turbines. No.440, National Environmental Research Institute Ministry ofthe Environment, Denmark.
Despin, B., LeMayo, Y. & Schmitt, M. 1978. Measures detemperatures peripheriques par thermographic infrarougechez le manchot de Humboldt (Spheniscus humboldti).Oiseau Rev. Fr. Ornithol. 48: 151–158.
Eddy, A.L., Van Hoogmoed, L.M. & Snyder, J.R. 2001. Therole of thermography in the management of equinelameness. Vet. J. 162: 172–181.
Edgar, J.L., Lowe, J.C., Paul, E.S. & Nicol, C.J. 2011. Avianmaternal response to chick distress. Proc. R. Soc. Lond. BBiol. Sci. 278: 3129–3134.
Evangelista, D., Fernandez, M.J., Berns, M.S., Hoover, A. &Dudley, R. 2010. Hovering energetics and thermal balancein Anna’s Hummingbirds (Calypte anna). Physiol. Biochem.Zool. 83: 406–413.
Ferreira, V.M.O.S., Francisco, N.S., Belloni, M., Aguirre, G.M.Z., Caldara, F.R., Naeaes, I.A., Garcia, R.G., AlmeidaPaz, I.C.L. & Polycarpo, G.V. 2011. Infrared thermographyapplied to the evaluation of metabolic heat loss of chicks fedwith different energy densities. Braz. J. Poult. Sci. 13: 113–118.
Galligan, E.W., Bakken, G.S. & Lima, S.L. 2003. Using athermographic imager to find nests of grassland birds. Wildl.Soc. Bull. 31: 865–869.
Garner, D.L., Underwood, H.B. & Porter, W.F. 1995. Use ofmodern infrared thermography for wildlife populationsurveys. Environ. Manage. 19: 233–238.
© 2012 British Ornithologists’ Union
Thermal imaging in avian science 13
Gauthreaux, S.A. & Livingston, J.W. 2006. Monitoring birdmigration with a fixed-beam radar and a thermal-imagingcamera. J. Field Ornithol. 77: 319–328.
Gilbert, C., McCafferty, D.J., Giroud, S., Ancel, A. & Blanc,S. 2012. Private heat for public warmth: how huddlingshapes individual thermogenic responses of rabbit pups.PLoS One 7: e33553.
Giloh, M., Shinder, D. & Yahav, S. 2012. Skin surfacetemperature of broiler chickens is correlated to body coretemperature and is indicative of their thermoregulatorystatus. Poult. Sci. 91: 175–188.
Goldsmith, R. & Sladen, W.J.L. 1961. Temperatureregulation of some Antarctic penguins. J. Physiol. Lond.157: 251–262.
Green, J.A. 2011. The heart rate method for estimatingmetabolic rate: review and recommendations. Comp.Biochem. Phys. A Mol. Integr. Physiol. 158: 287–304.
Hagan, A.A. & Heath, J.E. 1980. Regulation of heat-loss inthe duck by vasomotion in the bill. J. Therm. Biol 5: 95–101.
Hammel, H.T. 1956. Infrared emissivities of some arctic fauna.J. Mammal. 37: 375–378.
Herreid, C.F. & Kessel, B. 1967. Thermal conductance inbirds and mammals. Comp. Biochem. Physiol. 21: 405–414.
Hill, D.L. 2009. Sexual conflict and division of labour duringincubation. PhD. Thesis, University of Glasgow.
Hill, R.W., Beaver, D.L. & Veghte, J.H. 1980. Body-surfacetemperatures and thermoregulation in the Black-cappedChickadee (Parus-atricapillus). Physiol. Zool. 53: 305–321.
Hilsberg-Merz, S. 2007. Infrared thermography in zoo andwild animals. In Fowler, M.E. & Miller, R.E. (eds) Zoo andWild Animal Medicine. Current Therapy, 6th edn.philadelphia: Saunders.
Hristov, N.I., Betke, M. & Kunz, T.H. 2005. Assessment ofBrazilian free-tailed bat (Tadarida brasiliensis) populationsusing advanced infrared thermal imaging. Integr. Comp.Biol. 45: 1015–1015.
Kastberger, G. & Stachl, R. 2003. Infrared imaging technologyand biological applications. Behav. Res. Methods Instrum.Comput. 35: 429–439.
Kinzel, P.J., Nelson, J.M., Parker, R.S. & Davis, L.R. 2006.Spring census of mid-continent sandhill cranes using aerialinfrared videography. J. Wildl. Manage. 70: 70–77.
Kuwae, T. 2007. Diurnal and nocturnal feeding rate in Kentishplovers Charadrius alexandrinus on an intertidal flat asrecorded by telescopic video systems. Mar. Biol. 151: 663–673.
Lamprecht, I. & Schmolz, E. 2004. Thermal investigations ofsome bird nests. Thermochim. Acta 415: 141–148.
Lazarevic, L. 2009. Improving the Efficiency and Accuracy ofNocturnal Bird Surveys through Equipment Selection andPartial Automation. Thesis, Brunel University.
Mattsson, B.J. & Niemi, G.J. 2006. Using thermal imaging tostudy forest songbirds. The Loon 78: 74–77.
McCafferty, D.J. 2007. The value of infrared thermography forresearch on mammals: previous applications and futuredirections. Mammal Rev. 37: 207–223.
McCafferty, D.J., Moncrieff, J.B., Taylor, I.R. & Boddie, G.F.1998. The use of IR thermography to measure the radiativetemperature and heat loss of a Barn Owl (Tyto alba). J.Therm. Biol 23: 311–318.
McCafferty, D.J., Gilbert, C., Paterson, W., Pomeroy, P.P.,Thompson, D., Currie, J.I. & Ancel, A. 2011. Estimating
metabolic heat loss in birds and mammals by combininginfrared thermography with biophysical modelling. Comp.Biochem. Phys. A Mol. Integr. Physiol. 158: 337–345.
Mills, W.E., Harrigal, D.E., Owen, S.F., Dukes, W.F.,Barrineau, D.A. & Wiggers, E.P. 2011. Capturing clapperrails using thermal imaging technology. J. Wildl. Manage.75: 1218–1221.
Moe, R.O., Stubsjoen, S.M., Bohlin, J., Flo, A. & Bakken,M. 2012. Peripheral temperature drop in response toanticipation and consumption of a signaled palatable rewardin laying Hens (Gallus domesticus). Physiol. Behav. 106:527–533.
Monteith, J.L. & Unsworth, M.H. 1990. Principles ofEnvironmental Physics. London: Edward Arnold.
Müenzberg, M., Schilling, A., Schlemmer, H., Vogel, H.,Cramer, H. & Schlosshauer, J. 2011. The infrared basedearly warning system for bird strike prevention at Frankfurtairport. In Proceedings SPIE 8012 Infrared Technology andApplications XXXVII, Vol. 8012c, 20 May 2011.
Nimon, A.J., Schroter, R.C. & Oxenham, R.K.C. 1996.Artificial eggs: measuring heart rate and effects ofdisturbance in nesting penguins. Physiol. Behav. 60: 1019–1022.
Ovadia, O., Pinshow, B. & Lotem, A. 2002. Thermal imagingof House Sparrow nestlings: the effect of begging behaviorand nestling rank. Condor 104: 837–842.
Phillips, P.K. & Sanborn, A.F. 1994. An infrared,thermographic study of surface-temperature in 3 ratites –Ostrich, Emu and Double-wattled Cassowary. J. Therm.Biol. 19: 423–430.
Scholander, P.F., Walters, V., Hock, R. & Irving, L. 1950.Body insulation of some arctic and tropical mammals andbirds. Biol. Bull. 99: 225–236.
Scott, G.R., Cadena, V., Tattersall, G.J. & Milsom, W.K.2008. Body temperature depression and peripheral heat lossaccompany the metabolic and ventilatory responses tohypoxia in low and high altitude birds. J. Exp. Biol. 211:1326–1335.
Shinder, D., Rusal, M., Giloh, M. & Yahav, S. 2009. Effect ofrepetitive acute cold exposures during the last phase ofbroiler embryogenesis on cold resistance through the lifespan. Poult. Sci. 88: 636–646.
Speakman, J.R. & Ward, S. 1998. Infrared thermography:principles and applications. Zoology 101: 224–232.
Steen, K.A., Villa-Henriksen, A., Therkildsen, O.R. & Green,O. 2012. Automatic detection of animals in mowingoperations using thermal cameras. Sensors 12: 7587–7597.
Stewart, M., Webster, J.R., Schaefer, A.L., Cook, N.J. &Scott, S.L. 2005. Infrared thermography as a non-invasivetool to study animal welfare. Anim. Welf. 14: 319–325.
Stewart, M., Stafford, K.J., Dowling, S.K., Schaefer, A.L. &Webster, J.R. 2008. Eye temperature and heart ratevariability of calves disbudded with or without localanaesthetic. Physiol. Behav. 93: 789–797.
Symonds, M.R.E. & Tattersall, G.J. 2010. Geographicalvariation in bill size across bird species provides evidencefor Allen’s Rule. Am. Nat. 176: 188–197.
Takahashi, A. & Yoda, K. 2010. Advances in biologgingscience: a review of bird studies. Jap. J. Ornithol. 59: 3–19.
Tattersall, G.J. & Cadena, V. 2010. Insights into animaltemperature adaptations revealed through thermal imaging.Imaging Sci. J. 58: 261–268.
© 2012 British Ornithologists’ Union
14 D. J. McCafferty
Tattersall, G.J. & Milsom, W.K. 2003. Transient peripheralwarming accompanies the hypoxic metabolic response inthe golden-mantled ground squirrel. J. Exp. Biol. 206:33–42.
Tattersall, G.J., Andrade, D.V. & Abe, A.S. 2009. Heatexchange from the Toucan bill reveals a controllablevascular thermal radiator. Science 325: 468–470.
Tessier, M., Du Tremblay, D., Klopfenstein, C.,Beauchamp, G. & Boulianne, M. 2003. Abdominal skintemperature variation in healthy broiler chickens asdetermined by thermography. Poult. Sci. 82: 846–849.
Tillmann, J.E. 2009a. An ethological perspective ondefecation as in integral part of anti-predatory behavior ofthe Grey Partridge (Perdix perdix L.) at night. J. Ethol. 27:117–124.
Tillmann, J.E. 2009b. Fear of the dark: night-time roostingand anti-predation behaviour in the Grey Partridge (Perdixperdix L.). Behaviour 146: 999–1023.
Veghte, J.H. & Herreid, C.F. 1965. Radiometric determinationof feather insulation and metabolism of arctic birds. Physiol.Zool. 38: 267–275.
Ward, S. & Slater, P.J.B. 2005. Heat transfer and theenergetic cost of singing by canaries Serinus canaria. J.Comp. Physiol. A. 191: 953–964.
Ward, S., Rayner, J.M., Moller, U., Jackson, D.M.,Nachtigall, W. & Speakman, J.R. 1999. Heat transfer fromStarlings Sturnus vulgaris during flight. J. Exp. Biol. 202:1589–1602.
Ward, S., Moller, U., Rayner, J.M.V., Jackson, D.M.,Nachtigall, W. & Speakman, J.R. 2004. Metabolic power ofEuropean starlings Sturnus vulgaris during flight in a windtunnel, estimated from heat transfer modelling, doublylabelled water and mask respirometry. J. Exp. Biol. 207:4291–4298.
Ward, J.M., McCafferty, D.J., Ruxton, G.D. & Houston, D.C.2007. Thermal consequences of turning white in winter: a
comparative study of Red Grouse Lagopus lagopus scoticusand Scandinavian Willow Grouse L-l. lagopus. Wildl. Biol.13: 120–129.
Watmough, D.J., Fowler, P.W. & Oliver, R. 1970. Thethermal scanning of a curved isothermal surface:implications for clinical thermography. Phys. Med. Biol. 15:1–8.
Wilcox, C.S., Patterson, J. & Cheng, H.W. 2009. Use ofthermography to screen for subclinical bumblefoot in poultry.Poult. Sci. 88: 1176–1180.
Wilson, R.P., Adelung, D. & Latorre, L. 1998. Radiative heatloss in gentoo penguin (Pygoscelis papua) adults and chicksand the importance of warm feet. Physiol. Zool. 71: 524–533.
Yahav, S., Luger, D., Cahaner, A., Dotan, M., Rusal, M. &Hurwitz, S. 1998. Thermoregulation in naked neck chickenssubjected to different ambient temperatures. Br. Poult. Sci.39: 133–138.
Yahav, S., Shinder, D., Tanny, J. & Cohen, S. 2005.Sensible heat loss: the broiler’s paradox. Worlds Poult. Sci.J. 61: 419–434.
Zehnder, S., Akesson, S., Liechti, F. & Bruderer, B. 2001.Nocturnal autumn bird migration at Falsterbo, SouthSweden. J. Avian Biol. 32: 239–248.
Zerba, E., Dana, A.N. & Lucia, M.A. 1999. The influence ofwind and locomotor activity on surface temperature andenergy expenditure of the Eastern House Finch(Carpodacus mexicanus) during cold stress. Physiol.Biochem. Zool. 72: 265–276.
Received 20 April 2012;revision accepted 7 October 2012.Associate Editor: Stuart Marsden.
© 2012 British Ornithologists’ Union
Thermal imaging in avian science 15