Acoustic Diversity and Activity Patterns of Insectivorous Bats

Acoustic Diversity and Activity Patterns of Insectivorous Bats in a Riverine Forest of Gunung Mulu National Park, Malaysian Borneo Ellen Mc Arthur and Faisal Ali Anwarali Khan Universiti Malaysia Sarawak

Bats are keystone species that perform vital ecosystem and economic services

Forest cover and condition in Malaysian Borneo and Brunei in 2009. Bryan JE, Shearman PL, Asner GP, Knapp DE, Aoro G, et al. (2013) Extreme Differences in Forest Degradation in Borneo: Comparing Practices in Sarawak, Sabah, and Brunei. PLo. S ONE 8(7): e 69679. doi: 10. 1371/journal. pone. 0069679 http: //127. 0. 0. 1: 8081/plosone/article? id=info: doi/10. 1371/j ournal. pone. 0069679 Ø Habitat loss and fragmentation leading to decline of bat populations in Southeast Asia. Ø Malaysia losing forest at alarming rates. Ø Even within protected areas, little is known about population dynamics, distribution of species and which habitats are being used.

Background Ø Urgent need to improve efficiency of surveys to document diversity, distribution and habitat requirements of bat species. Ø Majority of studies in Sarawak focused on species inventories, particularly in protected areas. Ø Lack of information from many localities Ø Studies relied on capture techniques (e. g. use of harp traps and mist nest). Ø Acoustic sampling used in numerous studies worldwide, particularly temperate areas: Nth America, Europe. Monitoring programmes running for many years.

Background Ø Rivers and water bodies important foraging habitats in temperate and dry regions. Ø Rivers used by bats for drinking, feeding on emerging insects as well as flight corridors for commuting between roosts and other foraging sites. Ø Few studies have investigated use of riverine habitat by bats in tropical forests of Southeast Asia. Ø Traditional methods to study bats: capture (harp trap, mist nets)

Acoustic Sampling Has emerged, in recent years, as an essential tool to study the activity of echolocating bats and identify key habitats used by them for commuting and foraging

Bat Echolocation Calls Range of echolocation calls produced by (a) insectivorous bats, CF dominated calls with FM components; (b) FM call; QCF calls; (d) FM dominated calls with QCF component.

Gunung Mulu National Park 9 different forest types Altitude: 28 – 2376 m asl. 40% lowland forest 52, 865 ha GMNP + Buda NP + Extensions: 96, 978 ha

The limestone karst landscape of Gunung Mulu National Park , surrounded by a combination of different types of vegetation, supports high diversity of both cave and forest roosting bat species.

Deer Cave Chaerephon plicatus colony (estimated 2 – 3 million)

Bat Diversity Mulu National Park H. coxi (EN) Ø 54 species Chiroptera (14 spp. NT, VU, EN categories – IUCN Red List) Ø 16 Pteropodidae, 38 Insectivorous species ØAbout 50% of species known to use caves as roosts Ø 46 species recorded in a small area (3 sq. km) of riverine forest between Park HQ and Deer Cave (13 Pteropodidae, 33 insectivores)

Objectives Ø Build an echolocation call library with reference calls that can be used to identify insectivorous bats through acoustic recordings. Ø Determine which species can be reliably identified from their calls. Ø Identify which habitats, in a riverine forest, are used by various species of bats for commuting and foraging. Ø Determine the timing of commuting and foraging and some of the factors that may contribute to differences in nightly activity.

Methods - Acoustic Sampling Rivers, Forest Trail and Forest Interior 3 rivers: different width SM 2 Bat + SMX-US ultrasonic microphone Trail (50 m from river) Forest (50 m from trail) • • 105 recording nights (2016) 35 points x 3 nights each station 4 pm to 7 am continuous with trigger

Methods - Acoustic Sampling Forest Gaps 10 forest gaps, ranging in size from 6 m to 35 m width.

Methods - Acoustic Sampling Sites Figure 4. 1: Map of acoustic sampling points: (a) river, trail and forest; (b) forest gaps and width.

Methods – Bat Capture

Methods – Call Recording Stationary Enclosure Free Flight (Release) • Stationary – Rhinolophidae, Hipposideridae. • Enclosure (Flight Tent) – Forest Understory species (e. g. Kerivoula, Murina species etc. ) • Enclosure (large room) – Open space species (e. g. Chaerephon plicatus) • Free Flight (Release) – Semi-clutter along trails or over narrow river (All forest understory species) • Free flight (Release) - Open Space (e. g. Chaerephon plicatus)

Methods - Call Analysis Reference Calls Fppeak Fstart Fend Dur IPI Kaleidoscope Viewer (Wildlife Acoustics)

Call files analyzed with Kaleidoscope Viewer Manual ID

Results - Reference Calls (31 species) Emballonura sp. Rhinolophidae Hipposideridae Megadermatidae & Nycteridae Emballonura alecto/monticola QCF (MH) FM-CF-FM Vespertilionidae (13 species) FM (MH) Molossidae & Miniopteridae FM -QCF FM (Broadband)

DFA Analysis Results – CF Calls PERMANOVA (F = 396, p = 0. 0001) Rhinolophidae 5 species Hipposideridae 8 species Emballonura sp. 14 species - 636 calls - 95. 3% Correctly Classified Rhinolophus acuminatus & R. borneensis: 86. 8% correct classification Hipposideros cervinus & H. galeritus: 92. 8% correct classification

DFA Analysis Results – FM & FM-QCF Calls PERMANOVA (F = 71. 68, p = 0. 0001) Chaerephon plicatus Myotis spp. (3) Glischropus tylopus Miniopterus australis Megaderma spasma Kerivoula spp. (5) Murina spp. (4) Nycteris tragata 17 species - 195 calls – 64. 6% Correctly Classified Kerivoula-Murina Group: 64. 6% correct classification 64. 5% Correct Classification Myotis-Miniopterus Group: 73. 47% correct classification (8. 82% overlap between Myotis muricola and Miniopterus australis)

Results – Nightly Acoustic Activity Ø 135 nights: 335. 25 GB WAV files – 131, 695 bat passes. Ø Overall, 82% of passes were recorded at rivers, 13% at forest gaps, 4% were recorded on trails and 1% in forest interior.

Results – Hourly Activity Figure 4. 11: (a)Total hourly activity combined from all sampling nights in each habitat type; (b) habitat types in (a) with lower activity.

Results - Acoustic Diversity Ø Five species of cave roosting Rhinolophidae, seven Hipposideridae, Chaerephon plicatus and Miniopterus australis could be identified and represented 62% of total recorded passes. Ø A further 36% of passes were identified to a Myotis species call group and at least four call types from unknown species were detected. Ø Accurate identification of low intensity, high frequency calls from members of the forest roosting Vespertilionidae subfamilies Kerivoulinae and Murininae proved to be difficult, however this group of species only represented 1% of recorded calls

Figure 4. 12: Comparison of percentage of habitat types that species and species groups were detected in: (a) Rhinolophus borneensis; (b) R. creaghi; (c) R. philippinensis; (d) Hipposideros diadema; (e) H. galeritus/cervinus; (f) H. coxi.

Figure 4. 13: Comparison of percentage of habitat types that species and species groups were detected in: (a) Myotis species; (b) Kerivoula/Murina species; (c) Miniopterus australis; (d) Chaerephon plicatus; (e) Emballonura species and (f) Unidentified FM-QCF calls.

Current Challenges ØBuilding call libraries (species capture and call recording) ØFunding and manpower ØEquipment failure ØData management – storage and transfer ØFaster data processing (Automatic detection and Call ID) ØQuantifying activity and standardizing methods (measurement of number of passes, file length, call sequence duration, call intensity)

Conclusions Ø Rivers appear to be hotspots of activity for commuting and foraging insectivorous bats in the riverine forests of GMNP. Ø Acoustic sampling is not suitable for monitoring low intensity calling species that mainly roost in forest habitats. Capture methods still required to study these species. Ø However, most cave roosting species can be easily detected. Ø Echolocation calls for many aerial insectivores -represented in call libraries. under Ø Acoustic sampling is challenging in tropical environments. Equipment, especially microphones, vulnerable to failure due to high rainfall and humidity. Call Analysis time consuming. However, technology is improving.

Acknowledgements Ø Dr. Ulmar Grafe (UBD) for the generous loan of an SM 2 Bat recorder. Ø Borsarmulu Park Management for support and provision of equipment. Ø Caroline and Michael Schoner, UNIMAS graduate and undergraduate students for collaboration with trapping and collection of calls. Ø SEABCO 2015 participants who took part in the post-conference field trip to Gunung Mulu National Park. Ø Sarawak Forestry Department for research permits.