AEcol
A comparison of the spatial range of three bat detectors
___________________________________________________________________________________
A comparison of the spatial range of three bat detectors Henry Andrews MSc MIEEM Tom Staton BSc MSc Mark Latham BSc MSc In order to provide a range estimate for use in deployment and data-analysis of AnaBat, SM2 and Duet bat-detectors we performed a simple field-test at a roost holding brown long-eared Plecotus auritus and lesser horseshoe bats Rhinolophus hipposideros. Detectability of calls varied according to detector type, distance and species.
INTRODUCTION In order to provide an estimate of the detection range of Titley Electronics AnaBat SD1 CF, Wildlife Acoustics Songmeter SM2BAT (384kHz), and Batbox Duet in relation to brown long-eared and lesser horseshoe bats we performed a simple fieldtest in September 2011. We tried to assess the effectiveness of the detectors on two levels: Test 1. Can a detector deployed within a sealed building register brown longeared and lesser horseshoe bats outside? Test 2. What is the effective range of detectors for brown long-eared and lesser horseshoe bats? This included a consideration of microphone catchment area or ‘beam-pattern’ based on their specifications. The likelihood of the detection of a bat via its echolocation call alone is the product of a number of variables which centre around: The type, spatial range, intensity and directionality of the bats call; The bat’s behaviour and its movement in relation to the detecting microphone; The attributes of the receiving microphone; The probability that the spatial ranges of the bat’s call and the microphone’s area of sensitivity will coincide within space; The impact of environmental and structural aspects of the habitat upon the rate of sound absorption; and The type of post-processing utilised by the detector system. (Brigham et al. 2002) Some microphones will therefore be better suited to the detection of certain bats and certain bats’ behaviours and physiology will increase their likelihood of detection. This variability in the probability of detection potentially under-represents species presence and therefore biases data sets. ___________________________________________________________________________________ -1-
©AEcol – Andrews Ecology Ltd 2011
AEcol
A comparison of the spatial range of three bat detectors
___________________________________________________________________________________
Brown long-eared bats belong to the genus Plecotus which are typified by their multiharmonic echolocation calls (Dietz et al. 2011). The lower harmonic ranges between 55-20 kHz whilst the second occurs between 80-40 kHz (Dietz et al. 2011). Calls are quiet and short, c. 2 milliseconds, and the pulse repetition rate is high, the softness of the call and the relatively high frequency of the second harmonic results in the species displaying the shortest echolocation range of all the UK’s bat fauna (Russ 1999). Lesser horseshoe bats of the genus Rhinolophus, utilise a long constant frequency pulse between 108-114 kHz, with shorter frequency modulated sections at the beginning and end of each call, which in total can last up to 60 ms (Dietz et al. 2011). The high frequency of the lesser horseshoe bat’s call suggests that it would quickly be absorbed within the atmosphere and as such the effective range of the species’ call in typical atmospheric conditions might be as little as c. 10 m, the call is also significantly directional (Dietz et al. 2011). Logically, the soft and short FM calls of the brown long-eared bat and the contrastingly loud and highly directional CF calls of the lesser horseshoe bat will pose unique challenges to the design of bat detectors and, when the distribution, population numbers and conservation significance of both species are considered in parallel, the ability of a bat detector to deal with both types of call will go some way to highlight their usefulness as a reliable and unbiased piece of field equipment for the collation of species inventories, the quantification of activity levels and the documentation of roosts. We wanted to understand the potential pitfalls of using ultrasound detectors to perform unmanned surveillance. Firstly, we wanted to know whether a detector deployed inside a building might register a brown long-eared and/or lesser horseshoe bat passing outside. Secondly, we wanted to have an idea of the approximate range the two species might be detected by the ultrasound detectors at our disposal, both in terms of distance and also area of sensitivity.
METHOD To examine the range within which brown long-eared and lesser horseshoe bats might be detected in the field, we devised a test at Rookery Farmhouse, situated within Tarmac Ltd’s Halecombe Quarry. The farmhouse holds colonies of both brown longeared bats and lesser horseshoe bats and is suited to a controlled field study as: The bats have only one entrance/exit; through a custom-built bat-dormer; Their favoured flight-paths follow the outer edge of a 6 m high screening bund; Light-spill from an adjacent workshop deters both species from straying into the test area; and In six emergence surveys performed during post-development monitoring, nearly every bat of the resident brown long-eared colony had been observed crossing an approximately 30 m wide yard in a straight-line with crisp echolocation. ___________________________________________________________________________________ -2-
©AEcol – Andrews Ecology Ltd 2011
AEcol
A comparison of the spatial range of three bat detectors
___________________________________________________________________________________
We deployed one of each of the three detectors at 1.5 m height on two aluminium frames. The frames were deployed as follows: Test 1 Night 1: Control detectors deployed 5 m from the roost exit; Test detectors also deployed 5 m from the roost exit but inside the farmhouse looking out of a single thickness window – this was in order to assess whether a bat detector deployed during a presence/absence survey in a building might record a false positive result from a bat passing outside. Test 2 Night 1: Control 5 m from the roost exit; Test 10 m from the roost exit. Night 2: Control 5 m from the roost exit; Test 20 m from the roost exit Night 3: Control 5 m from the roost exit; Test 30 m from the roost exit. Figure 1 shows the layout of the farmhouse and yard, the location of equipment, and illustrates how light-spill benefits the test. Figure 2 shows the locations of the detector-frames during Test 1 and Figure 3 shows the location of the detector-frames in relation to the bat access-dormer in Test 2.
Figure 1. Layout of Rookery Farmhouse and yard, location of detector-frames and extent of light-spill.
One person was positioned with each of the two frames to visually monitor the flightpaths of bats exiting the roost in order that any bat entering the test distance air-space ___________________________________________________________________________________ -3-
©AEcol – Andrews Ecology Ltd 2011
AEcol
A comparison of the spatial range of three bat detectors
___________________________________________________________________________________
was taken into account. A third person used a Centurion Systems Multitask Gen.2+ night vision scope to conduct a visual emergence count, again monitoring the direction taken by each bat. Light-sampling flights comprising an emergence followed by an immediate return cancelled each other out, and only when the bat came out and stayed out was it counted as an emergence. The test was run from sunset for exactly one hour. The two automated units were set to operate within that time-frame, and the Duet units were manually switched on and recordings made on Sony mini-discs. Sensitivity of AnaBat microphones were set manually to 7 and the gain of SM2 microphones remained set at the supplier standard.
Figure 2. Location of detector-frames in Test 1.
Figure 3. Location of detector-frames in Test 2. ___________________________________________________________________________________ -4-
©AEcol – Andrews Ecology Ltd 2011
AEcol
A comparison of the spatial range of three bat detectors
___________________________________________________________________________________
Anabat SD1 and SM2 recordings were analysed with ANALOOK software. Those of the Duets were analysed using BatScan. Spectrograms were inspected and call peakfrequency, inter-pulse interval, maximum and minimum frequencies, duration, and ‘shape’ were measured or calculated and compared with descriptions described by Russ (1999). Constraints The experimental design was constrained by the orientation of detectors in relation to the bat flight path. While the control detector was orientated facing the emerging bats, the test detectors were orientated perpendicular to the flight path which may have reduced the detectability of bats by the test detectors. Placing the test detectors facing the bat flight path would however, have meant the bats would have ultimately flown directly over the detector, so detection distance could not be established. Perpendicular orientation was therefore unavoidable.
RESULTS General All but one bat exited the roost on the predicted flight-path. The exception was on the third night of Test 2, when an individual lesser horseshoe bat flew south to the 30 m Test-frame. This bat was observed by all three people, its time of emergence noted, and calls for this emergence excluded from the analysis.
Test 1 None of the detectors registered any calls from the Test location inside the farmhouse, as shown at Table 1. Table 1. Results of Test 1 on 26th September 2011. Species Visual Detector Control Inside
Brown long-eared bat 9 AnaBat SM2 Duet 0 4 12 0 0 0
Lesser horseshoe bat 11 AnaBat SM2 Duet 4 3 3 0 0 0
Test 2 The performance of each detector varied according to distance and species, as shown at Tables 2 – 4 (Where Control-counts exceed Visual-counts the excess may be explained by light-sampling flights).
___________________________________________________________________________________ -5-
©AEcol – Andrews Ecology Ltd 2011
AEcol
A comparison of the spatial range of three bat detectors
___________________________________________________________________________________
Table 2. Results of Test 2 for brown long-eared bat Date Visual Detector Control 10 m 20 m 30 m
AnaBat 6 2
27th 9 SM2 10 3
Duet 13 8
AnaBat 1
28th 4 SM2 0
Duet 6
0
1
1
AnaBat 3
29th 6 SM2 3
Duet 8
0
0
0
AnaBat 6
29th 18 SM2 0
Duet 17
0
0
0
Table 3. Results of Test 2 for lesser horseshoe bat Date Visual Detector Control 10 m 20 m 30 m
AnaBat 4 0
27th 18 SM2 3 0
Duet 12 2
AnaBat 33
28th 19 SM2 0
Duet 16
0
0
0
Table 4. Results of Test 2, detection capabilities of the three detectors DETECTOR
BAT SPECIES Brown long-eared bat Lesser horseshoe bat
10 m
20 m
30 m
SM2
Brown long-eared bat Lesser horseshoe bat
Duet
Brown long-eared bat Lesser horseshoe bat
AnaBat
Detector microphone specification In order to assess not only the range but also the area of sensitivity or ‘beam-pattern’ of the three detectors, we attempted to look up the pattern diagrams for their microphones. These are freely available for both AnaBat and SM2 at: http://titley-scientific.com/AnaBat-bioacoustics-AnaBat-bat-detector; and http://www.wildlifeacoustics.com/support-resources/documentation. The Duet manufacturer does not provide a manual showing the diagram, or the specification or model of the microphone, and did not provide the information when requested. Having opened up one of our own Duets however, we have concluded that the microphone is an electret microphone similar to microphones supplied by Knowles (http://www.knowles.com/search/products/m_ecm.jsp). Beam patterns indicate they are omnidirectional (http://www.avisoft.com/usg/ep3.htm).
___________________________________________________________________________________ -6-
©AEcol – Andrews Ecology Ltd 2011
AEcol
A comparison of the spatial range of three bat detectors
___________________________________________________________________________________
Anabat The AnaBat records and processes bat calls via an electrostatic capacitance microphone. This microphone is directional and, from inspections of the polar diagrams, appears to have a beam-pattern sensitive enough to detect typical brown long-eared bat echolocation calls within a 60° arc spanning between 330° and 30°. This arc is much narrower for the typical echolocation calls of lesser horseshoe bats with studies conducted by Smith (Bath Spa University pers com.) identifying the Anabat microphone as sensitive to typical lesser horseshoe calls through an arc of c.20°, between 350°-10°. SM2 Wildlife Acoustics Songmeter SM2BAT utilises the SMX-US omnidirectional microphone and whilst the sensitivity can only be manually adjusted with difficulty, the detector has a beam pattern of 360°. Duet The Batbox Duet microphone appears to be equipped with a high sensitivity electret condenser microphone, which is less sensitive in the mid kHz range (40kHz), with higher sensitivities, when compared to Piezo and Myca microphones at the lower and upper kHz levels (20-35kHz and 100+kHz). The sensitivity cannot be adjusted, but the microphone is omnidirectional with a beam pattern of 360o. Having found what we concluded to be a reasonable representation of the beampattern of each detector we produced Figure 4; what we think may be a broad approximation of the pattern diagram for the three detectors.
Figure 4. The beam-patterns (sensitivity of detection in 360o for a bat echolocating c.40-50 kHz) of the Anabat SD2, SM2 and Duet microphones, showing the directional qualities of the AnaBat and the omnidirectional qualities of the two other recorders.
In order to better illustrate each detectors performance during the survey we produced the following representative schematic diagrams: Figure 5 shows what we think the ___________________________________________________________________________________ -7-
©AEcol – Andrews Ecology Ltd 2011
AEcol
A comparison of the spatial range of three bat detectors
___________________________________________________________________________________
side view of a surveyor with an Anabat detector might look like, whilst Figure 6 shows what we think might be a broad approximation of the side view of a surveyor with a SM2 or Duet. Using data derived from our field test we have illustrated a representational ‘window of sensitivity’ (in red) which illustrates a generalised spatial window in which the long-eared bats could be detected using the three detectors. Our surveys positioned the detectors facing the emerging bats, as such we can only predict an approximate window of sensitivity for brown long-eared bats based upon the known polar diagrams of the detectors microphones, coupled with the distances at which we detected the species. We can however be confident that the light-spill from the adjacent workshop prevented bats from passing behind the detectors and thus unfairly biasing the results in favour of the SM2 and Duet.
___________________________________________________________________________________ -8-
©AEcol – Andrews Ecology Ltd 2011
AEcol
A comparison of the spatial range of three bat detectors
___________________________________________________________________________________
Note: If we’re significantly misrepresenting the situation with any or all the detectors then we apologise to the manufacturer and stress that this is not intentional. If they would like to produce diagrams we would be very pleased to see them.
DISCUSSION This test was rudimentary and could not take into account some aspects that may have impacted upon the results. First, the directionality of the echolocation calls of the two bat species might have impacted upon the results of the test involving the detector inside the building. The brown long-eared bat has a broad sound cone whilst the lesser-horseshoe bat has a highly directional, narrower sound cone (Dietz et al. 2011). This is important because it was anticipated that the emerging bats would all be flying directly away from the test-frame deployed in the house. Furthermore, the detectors were deployed at 1.5 m height and the bats were emerging at c. 6 m height. However, in the event bats of both species were seen to emerge and descend to the level of these detectors immediately outside the window, and although they were recorded by the test-frame 5 m from the test frame outside, they were not recorded by the test frame inside, despite the fact that they were often within 1 m of it. Furthermore, the bats were flying across the detectors rather than toward or away, so it may be that both species would be detected at greater distances if they were flying toward the detectors, and common sense would indicate this is almost certainly the case with lesser horseshoe bats. When comparing the control-counts with those counts made at 20 m, the SM2 testframe positioned at 20 m recorded a brown long-eared bat when the Control-frame did not. This suggests that the bats were not echolocating with an equal intensity during their emergence flights. In most general terms, there is a trade-off between the directionality and the sensitivity of a microphone. An omnidirectional microphone will therefore be less sensitive to sound across its 360° sphere than a more directional microphone with a smaller cone of reception. As such, it is quite possible that upon leaving the roost the long-eared bat in question was echolocating at too low an intensity for the SM2 microphone to detect its call, which only registered on the more sensitive unidirectional microphone of the AnaBat. If the bat had then increased the intensity of its calls as it crossed the more open yard at 20 metres this might explain the bat’s call subsequently registering on the SM2, whilst the failure of the call to register on the AnaBat suggests the bat’s call had passed outside the AnaBat’s narrower unidirectional beam-pattern. The results of this test suggest the AnaBat has a narrow beam-pattern with an effective range of at least 10 m for brown long-eared bats and 5 m for lesser horseshoe bats. Our investigations suggest that the SM2 and Duet have a somewhat ‘spherical’ beam-pattern and both may have an effective range of up to 20 m for brown long-eared bats in open habitat. However, in relation to lesser horseshoe bats, the SM2 may have a typical effective range of up to 5 m and the Duet up to 10 m. ___________________________________________________________________________________ -9-
©AEcol – Andrews Ecology Ltd 2011
AEcol
A comparison of the spatial range of three bat detectors
___________________________________________________________________________________
Due to the relatively narrow beam-pattern of the AnaBat, it is possible that its range is greater in a straight line than our results indicate. Its efficacy for inventory studies might be negatively influenced by its focussed microphone alone, which gives a far narrower range of airspace coverage. For example, it is unknown how long a bat flying at an average speed would take to pass through a transverse section of the AnaBat beam at 20 m in comparison to the omnidirectional ‘orb’ of the SM2 and Duet. It is only logical to assume that the longer the bat is in the spatial range of the detector, the higher the probability that it will be detected. It is apparent that there are differences in performance between detectors, these differences merit further investigation at other sites and using other species.
CONCLUSIONS The different detectors have subtly different performance which should be considered when deploying them for specific tasks. Our findings indicate that it’s not that one detector is better than another, but that their varying specifications make some of them better suited to meeting specific objectives than others. The findings of this study highlight the fact that as consultants it is imperative we consider the specifications of our equipment in line with the objectives of our surveys, choosing the right detector for the right job; selecting omnidirectional detectors to meet objectives such as long-term surveillance or species inventory collation (i.e. presence/absence) in open habitat, whilst selecting highly directional detectors for finer scale surveys of specific features.
ACKNOWLEDGEMENTS We would like to thank Tarmac Ltd and in particular Halecombe Quarry Manager Daniel Brailsford and Vaughn Gray for arranging access to Rookery Farmhouse and allowing us to perform this study. In addition, Dr. Paul Chanin advised us with substantial edits, and George Bemment generously offered comments and suggestions.
REFERENCES Brigham M, Kalko K, Jones G, Parsons S and Limpens H 2002. Bat Echolocation Research: tools, techniques and analysis. Bat Conservation International Dietz C, Helversen O von and Nill D 2011. Bats of Britain, Europe & Northwest Africa. A & C Black Publishers Ltd. London. Russ J 1999. The Bats of Britain & Ireland: Echolocation Calls, Sound Analysis, and Species Identification. Alana Books, Alana Ecology Ltd ISBN 0 9536049 0 X.
__________________________ ___________________________________________________________________________________ - 10 -
©AEcol – Andrews Ecology Ltd 2011
