Nocturnal flight call analysis as a method for monitoring density and species composition of migratory songbirds

(Order Passeriformes) across southern Vancouver Island, British Columbia in 2004.

Photo: Ralph Hocken

Prepared for – TD Friends of the Environment Foundation

Prepared by – Jessica Murray, Director Rocky Point Bird Observatory

1

TABLE OF CONTENTS PAGE

PROJECT OVERVIEW................................................................................................. 2 INTRODUCTION .......................................................................................................... 2

METHODS .................................................................................................................... 4 Site location............................................................................................................ 4

Microphone design ................................................................................................. 5 Flight call analysis ................................................................................................. 5

RESULTS....................................................................................................................... 6 Number of detections.............................................................................................. 6

Weather and detection ability ................................................................................. 6 Time of night .......................................................................................................... 7

Highlights................................................................................................................ 7 DISCUSSION .............................................................................................................. 10

Methodology ......................................................................................................... 10 Swainson’s Thrush............................................................................................... 11

Considerations for analysis .................................................................................. 11 Species complexes................................................................................................. 12

Infrequently banded species ................................................................................. 12 Weather ................................................................................................................ 12

CONCLUSION ............................................................................................................ 13 ACKNOWLEDGEMENTS........................................................................................... 13

LITERATURE CITED................................................................................................. 14

2

PROJECT OVERVIEW

Rocky Point Bird Observatory (RPBO) is a registered, non-profit organization located west of Victoria, British Columbia. The primary role of RPBO is to monitor the long-term population trends of migratory songbirds (order Passeriformes) through the method of bird banding. Banding provides critical information about the productivity, survivorship, arrival and departure dates, and long-term population trends of many songbird species. In addition to banding, an emerging method for monitoring the migratory movements of songbirds is nocturnal acoustic monitoring. Acoustic monitoring used in conjunction with banding at RPBO would provide a valuable data set on migration trends to support the conservation of songbird species. The main objective of this study was to determine whether nocturnal acoustic monitoring could be used as a tool at RPBO to estimate migration and population trends of songbirds over time. The nocturnal flight calls of migrating songbirds were recorded each night between August 6

and September 20, 2004. 38 nights were successfully monitored and allowed for 572 individual flight calls of 12 different songbird species to be recorded. The most frequently recorded species were Swainson’s Thrush, Yellow Warbler, Wilson’s Warbler, and Hermit Thrush. This project demonstrated the potential of using nocturnal acoustic monitoring as a conservation tool for many migratory songbird species, especially the Swainson’s Thrush.

INTRODUCTION

Rocky Point Bird Observatory (RPBO) is one of 23 member stations composing

the Canadian Migration Monitoring Network and is the only bird observatory situated on

the Canadian Pacific coast (Figure 1).

Rocky Point

Bird Observatory

FIGURE 1. The 23 banding stations that compose the Canadian Migration Monitoring Network. Source – Bird Studies Canada

23

North

CANADA

3

The observatory is uniquely situated on Department of National Defence land,

located at the southern-most tip of Vancouver Island, British Columbia. The observatory

encompasses a mosaic of habitat types including mixed coniferous and deciduous forests,

freshwater ponds and marshes, marine tidal flats and one of the few remaining areas of

intact Garry oak (Quercus garryana) habitat in Canada.

In 1994, the potential of Rocky Point as a songbird migration station was

recognized and the Rocky Point Bird Observatory Society officially formed a non-profit,

volunteer organization in association with the Victoria Natural History Society and the

Canadian Wildlife Service. Since RPBO’s inception, over 24,650 birds of 95 different

species have been banded, and the master checklist of all birds observed at RPBO

includes 293 species. Over the past 11 years we have learned a great deal about the

migratory habits of many songbird species including information on arrival and departure

dates, the relationship between weather conditions and migration, and the degree of

regularity with which individual birds return. This has provided RPBO and the Canadian

Wildlife Service with a valuable data set to enable the long-term monitoring and

conservation of many avian species.

For numerous songbird species, banding is a valuable tool for obtaining

information about morphological features and timing of migration. However, at RPBO

certain bird species that forage high in the forest canopy are infrequently observed at

ground level. For these species, population trends cannot be accurately assessed through

the traditional method of banding because their capture rate in mist-nets is so infrequent.

An emerging approach for species that cannot be monitored by traditional banding is the

method of nocturnal acoustic monitoring. Each year millions of songbirds migrate across

North America throughout the night. Several species of warblers (Parulinae), sparrows

(Emberizinae) and thrushes (Catharus) are well known for giving reliable nocturnal flight

calls while migrating (Evans et. al 2000), and on a good migration night numerous bird

calls have been heard by listeners on the ground at RPBO. The majority of songbird

species each have a unique nocturnal flight call and therefore a uniquely associated

spectrogram. Spectrograms are visual representations that illustrate the frequency,

modulation, and duration of each flight call. By recording and analyzing these

4

spectrograms, we can identify the species composition and numbers of birds migrating

over Rocky Point.

The objectives of this study are to determine whether acoustic monitoring can be

used as a conservation tool at RPBO to estimate migration and population trends of

songbirds over time, and to determine whether acoustic monitoring can be used to

monitor species that are infrequently encountered through traditional banding methods to

support the long-term monitoring of songbird species.

METHODS

Site location An open meadow located west of the banding station (Figure 2) was determined

to be the most conducive area to record nocturnal flight calls.

FIGURE 2. An aerial view Rocky Point Bird Observatory, the extreme southernmost tip of Vancouver Island. 1 = Rocky Point Bird Observatory banding station; 2 = west meadow (selected location of microphone for nocturnal monitoring in 2004); 3 = proposed microphone locations for future monitoring. Source – Department of National Defence - Rocky Point

2 1

3

North

5

The site was free from any nearby interference noises (frogs at marshes, ocean waves,

human activity) and overhanging canopy which could potentially muffle flight calls

passing overhead.

Microphone design

A single, pressure zone microphone was used to record the flight calls of

nocturnal migrants. Microphone construction followed the guidelines laid out by Old

Bird Inc., a non-profit organization dedicated to facilitating acoustic monitoring and

providing expert advice to acoustic monitoring stations (www.oldbird.org). A pre-amp

was added to the design to increase detection sensitivity; the microphone recorded an

approximate 1,000 m2 cross section of sky. The microphone was waterproofed with two

layers of plastic wrap and was mounted on the roof of a storage trailer located in the west

meadow. A wooden pallet and bungee cords were used to secure the microphone to the

roof in case of strong winds. An audio cable was run from the rooftop to inside the trailer

and a laptop computer where the flight calls were recorded directly to the hard drive each

night.

Based on known migration chronology of songbirds at RPBO from previous

banding seasons (Derbyshire 1999 and 2000, Gibson 2001, Woodward 2002, Jantunen

2003), nocturnal flight calls were recorded each night from August 6 until September 20,

2004, from sunset to sunrise.

Flight call analysis

Flight calls were detected using ‘Tseep’ and ‘Thrush’ software – programs

specifically designed by OldBird Inc. for detecting the short, high-pitched call notes of

most North American songbird species. Flight calls were then run through ‘GlassOFire’

software, also designed by OldBird Inc., which produced an accompanying spectrogram

for each call. Resulting spectrograms were closely examined and identified to species

where possible through comparison with a flight call spectrogram library provided by Old

Bird (Evans and O’Brien, 2002), and also by comparing nocturnal spectrograms to

diurnal spectrograms for each species, as the general structure of some diurnal and

nocturnal call notes are similar.

6

RESULTS

Number of detections

Monitoring was conducted over 38 nights resulting in 572 individual flight calls

of 12 songbird species to be recorded. Of these flight calls, 340 calls were identified to

species and the remaining 232 were identified to either a ‘complex’ (a group of species

which were problematic to identify due to almost identical nocturnal flight calls,) or

‘unidentifiable’ (more sophisticated analysis software needed). The most frequently

recorded species were Swainson’s Thrush, Yellow Warbler, Wilson’s Warbler and

Hermit Thrush (Appendix 1). Detection of flight calls generally peaked twice in the

season: once in mid-August and once in mid-September (Figure 3).

0

20

40

60

80

100

120

6-Aug-

04

10-Aug-

04

14-Aug-

04

18-Aug-

04

22-Aug-

04

26-Aug-

04

30-Aug-

04

3-Sep-

04

7-Sep-

04

11-Sep-

04

15-Sep-

04

19-Sep-

04

Date of night

Nu

mb

er f

lig

ht

call

s re

cord

ed p

er n

igh

t

FIGURE 3. Nightly abundance of migratory flight calls detected throughout the monitoring season (August 6 – September 20, 2004), (no flight calls recorded August 21, 24, 26, 27, 31, September 1, 2, and 3 due to weather interference).

Weather and detection ability

No flight calls were detected on August 21, 24, 26, 27, 31 and September 1, 2,

and 3. These nights had strong winds and/or rain, which interfered with recording.

Among the remaining nights of monitoring, there was considerable variation in weather

and the number of migrants recorded per night.

7

Time of night

Detections tended to build in frequency throughout the night, and the highest

frequencies of calls were detected approximately 30-60 minutes before sunrise (Figure 4).

0

20

40

60

80

100

120

8:00:00

PM

9:00:00

PM

10:00:00

PM

11:00:00

PM

12:00:00

AM

1:00:00

AM

2:00:00

AM

3:00:00

AM

4:00:00

AM

5:00:00

AM

6:00:00

AM

Time of night (hrs)

Nu

mb

er o

f fl

igh

t ca

ll d

etec

tio

ns

FIGURE 4. Average time of night nocturnal flight calls were detected.

Highlights

Due to the fact that very little acoustic monitoring has been conducted on the west

coast of North America, this exploratory research resulted in a number of interesting

observations being made.

Hermit Thrush set a new earliest arrival record at RPBO on August 9; the

historical arrival record was August 27, 1998.

MacGillivray’s Warbler was one of the target species to record, as very few of

them are banded each season at RPBO. Two calls were recorded: one on August 7 and

the other on August 17. There were no reference recordings available for the nocturnal

flight call of this western warbler species, so unidentified spectrograms were compared to

nocturnal flight calls of its eastern counterpart, the Mourning Warbler. These two species

have nearly identical vocalizations, and as expected the spectrograms for both species

8

matched very closely. This detection/discovery and recording of the MacGillivray’s

Warbler nocturnal flight call has provided essential baseline data to allow for future

monitoring of this western species.

Additionally, a Northern Waterthrush, which is a rare bird for Rocky Point

(usually occurring less than five times per year) was recorded on August 12.

Spectrograms

The majority of songbird species have a unique nocturnal flight call and therefore

a uniquely associated spectrogram. Spectrograms are visual representations that illustrate

the frequency, modulation, and duration of each flight call. By recording and analyzing

these spectrograms, we can identify the species of birds migrating over Rocky Point.

MacGillivray’s Warbler – August 18th, 0421 hrs

Photo credit: Ralph Hocken

0 100 Call duration (mS)

10 F r e q u e n c y (kHz) 0

9

Yellow Warbler – August 15th, 0235 hrs

Photo credit: Ralph Hocken

Chipping Sparrow – September 10th, 0455 hrs

Photo credit: Ralph Hocken

0 100 Call duration (mS)

10 F r e q u e n c y (kHz) 0

0 100 Call duration (mS)

10 F r e q u e n c y (kHz) 0

10

Lincoln’s Sparrow – September 7th, 0500 hrs

Photo credit: Ralph Hocken

(*Note – all nocturnal flight calls are “presumed”, even if 99.9% confident of the species identification because there is no visual confirmation of the bird).

DISCUSSION

Methodology Initially, an open, grassy meadow (Figure 2) was thought to be the most

conducive location to recording nocturnal flight calls. However, it appears that a more

productive location may be a transect running parallel to the ocean shoreline. Because

landbirds are generally reluctant to cross large bodies of water, birds tended to call much

more frequently at this land-water interface before embarking on their flight across the

Straight of Juan de Fuca. The majority of birds followed the shoreline for hundreds of

meters, before seeming to reluctantly cross over the water. This significant increase in

flight call numbers was not detected from our microphone location which was only 1 km

away. If multiple microphones were put in series running along the ocean shoreline a

few hundred metres apart, many more flight calls could be detected in this zone of

increased call activity.

0 100 Call duration (mS)

10 F r e q u e n c y (kHz) 0

11

Swainson’s Thrush

In 1995, the Canadian Wildlife Service (CWS) conducted a four-night pilot study

at RPBO to determine if the number of nocturnal migrants was comparable to the

banding data. CWS detected 495 migrating Swainson’s Thrush during this four night

period, whereas only 20 Swainson’s Thrush were banded the entire fall migration period

from July 21 - October 18. This nocturnal acoustic migration project conducted by

RPBO in 2004 revealed similar results. Using the acoustic technique, we detected 251

Swainson’s Thrush over 20 nights; whereas only 148 were banded over the entire three

month season.

Over the acoustic monitoring season (August 6 – September 20, 2004) the highest

numbers of Swainson’s Thrush were recorded on the night of September 11. The

microphone recorded 86 flight calls within the last 20 minutes (0556 – 0616 hrs) of

recording. At the land-water interface over the same time period (0556 – 0616 hrs) 2,326

Swainson’s Thrush calls were counted by observers. However, this does not imply that

each call was a distinct bird (i.e.-this does not imply 2,326 individual birds). Of the

2,326 individual calls, a conservative estimate would be 300-400 birds on that single

morning. Near dawn, nocturnal migrants descend to the ground to spend the daytime

feeding and resting. During their descent there is a natural inclination for each bird to

call several times (Farnsworth et al., 2004). Due to our numerous detections of

Swainson’s Thrush, nocturnal recording would likely be an extremely useful alternative

to banding as a monitoring tool for this species.

Considerations for analysis

Several studies have shown that the peak in vocalizations of nocturnal migrants

occur in the hours just before dawn (Farnsworth et al., 2004). One factor to consider was

how to separate the year-round, resident birds that vocalized near sunrise from migrant

birds of the same species. For instance, species such as White-crowned Sparrow and

Song Sparrow are year-round residents, but there are most likely migratory populations

as well. To avoid erroneously counting residents as migrants, I only included flight calls

that were made throughout the night up until 30 minutes before sunrise, which is

approximately when the resident birds at RPBO began to call.

Another consideration was that a single bird can call several times. I determined

that if multiple flight calls were made within a short time span they probably originated

12

from a single bird. However, if the calls were separated by several minutes it was

assumed they came from different birds (Evans and Rosenberg, 2000).

Species complexes

There are some groups of birds that have almost identical nocturnal flight calls,

and therefore are difficult to distinguish to species level. The nocturnal flight calls of

Yellow-rumped Warbler, Orange-crowned Warbler, Townsend’s Warbler, Nashville

Warbler and White-crowned Sparrow are all very similar and often not separable (Evans

and O’Brien, 2002). Approximately 30 % of the flight calls detected over the season fell

into this complex, and therefore were not separated to species. However, with more

study of the slight differences in duration, pitch and character of these calls using more

sophisticated software, the majority of calls should be able to be identified. Additionally,

Fox Sparrow and Song Sparrow were sometimes difficult to separate and some were left

as a ‘sparrow complex’.

Infrequently banded species

One of the objectives of this project was to determine whether acoustic monitoring

could be used as a method to monitor species that are infrequently encountered through

banding. For example, infrequently banded species such as the Black-throated Gray

Warbler, Varied Thrush and Black-headed Grosbeak are target species for monitoring and

should be detected through the acoustic method if the previously mentioned improvements

to protocol such as microphone placement are observed. For species such as the

Townsend’s Warbler this model may be unrealistic without more sophisticated software,

as its nocturnal flight call is too similar to other species to distinguish confidently. The

acoustic method was successful in detecting the infrequently banded MacGillivray’s

Warbler.

Weather

Weather may have played a role in nightly variation in number of nocturnal

migrant detections. Birds tend to gauge weather when initiating migratory movements

(New Jersey Audubon Society, 2004). Bird migration is dependent on a number of

weather factors such as rain, wind direction and speed, and location of frontal boundaries

(Able, 1973). In general, the busiest migration nights tended to be on cold (~8 Co), clear

13

nights. Wind speed was close to zero, and winds were generally south-southwesterly.

Migration is complex and other factors besides weather may be responsible for nightly

variations in migration numbers.

CONCLUSION AND RECOMMENDATIONS

This project effectively demonstrated the use of nocturnal flight call analysis as a

conservation tool for many songbirds, especially the Swainson’s Thrush. Due to our

numerous detections of Swainson’s Thrush, nocturnal recording would likely be an

extremely useful alternative to banding as a monitoring tool for this species.

With a few minor changes in protocol, I believe this method of nocturnal

monitoring could potentially record thousands of nocturnal migrants passing over RPBO.

This project was successful in demonstrating that acoustic monitoring is an effective

method of supplementing the banding data and the technique also demonstrated its use

for detecting infrequently encountered species such as the MacGillivray’s Warbler.

Nocturnal monitoring may be more cost-effective and less intrusive than conventional

monitoring techniques which are subject to numerous biases such as detecting only

understory birds. Continuing this project in future migration periods will enable the

Rocky Point Bird Observatory to accurately assess population dynamics and support the

conservation of songbird species for many generations to come.

ACKNOWLEDGEMENTS

First and foremost, thank you to the TD Friends of the Environment Foundation

for providing the funding and support to this conservation project. Thank you to Bill

Evans and OldBird Inc for their expert advice on spectrogram analysis, microphone

design and analysis software. A special thanks to Bob Chappell for his dedication to

constructing and monitoring the project’s microphone and experience in other electrical

aspects. Thanks to Jeremy Gatten for his help and advice on the project. Thanks to the

board members of Rocky Point Bird Observatory, especially to David Allinson for his

advice, support and assistance in setting up the project, and to Paul Levesque and Katie

Christie for their comments and advice on this report. Thank you to Jukka Jantunen for

the observational flight call data he provided to the project. Thank you to Ralph Hocken

for his excellent bird photographs. Thank you to Bird Studies Canada for their map of

14

the Canadian Migration Monitoring Network. A final thanks to the Department of

National Defence for their aerial photographs and most importantly for providing access

to Rocky Point Bird Observatory.

LITERATURE CITED

Able, K.P. 1973. The role of weather variables and flight direction in determining the magnitude of nocturnal bird migration. Ecology, Vol. 54. 5: 1031-1041.

Derbyshire, D. 1999. Migration Monitoring at Rocky Point, Fall 1999. Rocky Point Bird

Observatory, Metchosin, B.C. Derbyshire, D. 2000. A Report on Migration Monitoring at Rocky Point: Fall 2000.

Rocky Point Bird Observatory, Metchosin, B.C. Evans W.R. and M. O’Brien. 2002. Flight Calls of Migratory Birds Eastern North

American Landbirds. Old Bird, Inc. [CD-ROM] Evans, W.R. and K.V. Rosenberg. 2000. Acoustic monitoring of night-migrating birds: a

progress report. Proceedings of the 3rd Partners in Flight Workshop; 1995. Farnsworth, A., S.A. Gauthreaux and D. van Blaricom. 2004. A comparison of nocturnal

call counts of migrating birds and reflectivity measurements on Doppler radar. Journal of Avian Biology. 35: 365-369.

Gibson, C.G. 2001. Migration Monitoring at Rocky Point Bird Observatory in 2001.

Rocky Point Bird Observatory, Metchosin, B.C. Jantunen. J. 2003. Migration monitoring at Rocky Point Bird Observatory in fall 2003.

Rocky Point Bird Observatory, Metchosin, B.C. New Jersey Audubon Society. 2004. Preserving oases along the flyway.

http://www.njaudubon.org/Education/Oases/ Woodward, D. 2002. Fall Banding Report for Rocky Point Bird Observatory. Rocky

Point Bird Observatory, Metchosin B.C.

15

APPENDIX 1. Total number of nocturnal flight calls detected per species per night from August 6 to September 20. Note - no flight calls were detected on August 21, 24, 26, 27, 31, September 1, 2, and 3 due to strong winds and/or rain which interfered with recording. SWTH = Swainson’s Thrush (Catharus ustulatus), YWAR = Yellow Warbler (Dendroica petechia), WIWA = Wilson’s Warbler (Wilsonia pusilla), HETH = Hermit Thrush (Catharus guttatus), MGWA = MacGillivray’s Warbler (Oporornis tolmiei), LISP = Lincoln’s Sparrow (Melospiza lincolnii), CHSP = Chipping Sparrow (Spizella passerina), SAVS = Savannah Sparrow (Passerculus sandwichensis), SOSP = Song Sparrow (Melospiza melodia), FOSP = Fox Sparrow (Passerella iliaca), COYE = Common Yellowthroat (Geothlypis trichas), NOWA = Northern Waterthrush (Seiurus noveboracensis), ‘complex’ = a group of birds having almost identical nocturnal flight calls, and therefore difficult to distinguish to species level

DATE SWTH YWAR WIWA HETH MGWA LISP CHSP SAVS SOSP FOSP COYE NOWA 'complex' Totals

06Aug 1 4 1 07Aug 1 1 1 4 3 08Aug 4 3 4 09Aug 1 2 1 10Aug 1 1 13 2 11Aug 1 10 1 12Aug 3 1 1 19 5 13Aug 1 11 1 14Aug 1 3 2 1 11 7 15Aug 9 0 16Aug 16 1 17 17 17Aug 3 1 1 13 5 18Aug 2 1 1 1 19 5 19Aug 4 3 10 7 20Aug 8 1 6 9 22Aug 3 0 23Aug 3 3 25Aug 3 2 3 28Aug 1 1 1 29Aug 4 1 5 30Aug 1 0 04Sep 2 1 9 3 05Sep 28 1 10 29 06Sep 69 2 2 1 1 13 75 07Sep 6 1 1 7 08Sep 1 2 1 09Sep 15 1 1 5 17 10Sep 1 1 2 11Sep 86 4 3 1 18 94 12Sep 5 5 5 13Sep 1 1 1 2 14Sep 1 1 1 15Sep 6 6 16Sep 9 1 10 17Sep 1 1 4 2 18Sep 2 0 19Sep 1 2 1 2 4 20Sep 2 1 2

Totals 251 40 15 13 2 1 6 5 3 1 2 1 171 511 Unidentifiable 61 GRANDTOTAL 572

16