Recordings were made of adult male and female Weddell seals on the ice during the breeding seasons of 1990 and 1997. The recordings were made near Davis, Antarctica in the Vestfold Hills. The vocalisations made with both the mouth and nostrils closed were classified into call types. These call types are also produced by the seals when underwater. The call classifications were based on those described by Thomas, J.A. and Kuechle,, V.B. (1982, J. Acoust. Soc. Amer. 72: 1730-1738) and Pahl, B.C., Terhune, J.M. and Burton, H.R. (1997, Aus. J. Zool. 45: 171-187). Nineteen call types were identified. Of these, males made 18 and females made 15. Trills are only made by males and it is likely that a stepped ascending whistle is only made by females. A roar and mew are also potential male-only call types. The data suggest that the Trill vocalisations can be used to indicate the presence of males. This will be useful when recording underwater where the calling seals cannot be observed directly. A description of the types of calls made by Weddell Seals is listed below. SymbolNameDescription OToneConstant-frequency, predominantly sinusoidal call. LGrowlConstant-frequency, broad bandwidth, long call. QWhoopConstant-frequency call with a terminal upsweep. SSqueakBrief call with constant frequency or rising frequency and an irregular waveform. WAWhistle AscendingAscending frequency, sinusoidal waveform. TCTrill Constant-FrequencyNarrow bandwidth trill with a constant-frequency beginning, sinusoidal or frequency-modulated waveform. TTrillNarrow to broad bandwidth, containing a frequency downsweep, greater than 2 seconds. WDWhistle DescendingDescending frequency, sinusoidal waveform (less than 2 seconds). MMewAbruptly descending frequency followed by a long constant-frequency ending. CChugAbruptly descending frequency followed by a brief constant-frequency ending. GGuttural Glug (Grunt)Descending-frequency call that was lower than a Chug and had a brief duration. WAGWhistle Ascending - GruntBrief Ascending Whistle followed by a Guttural Glug (Grunt), the two types alternate in a regular pattern. KKnockAbrupt, brief-duration broadband sound

The underwater and in-air recordings were used to derive a technique to classify the call types. The in-air recordings demonstrated that both males and females vocalise and often a single seal will string up to 6 call types together in a variety of orders. No 'Trills' were heard by males or females on the ice. The seals lengthened the duration of multiple-element calls when they were 'interrupted' by another calling seal. This suggests that the seals are listening for the calls of conspecifics while they themselves are calling. A pilot project indicated that almost none of the calls are completely masked by other calling seals. The recordings are being used (in association with recordings obtained in later years) to address other aspects of Weddell seal vocal communication. See the link below for public details on this project.

From the abstract of the attached paper: Vocal recognition may function as a critical factor in maintaining the phocid mother-pup bond during lactation. For vocal recognition to function, the caller must produce individually distinct calls that are recognised by their intended recipient. Mother-pup vocal recognition has been studied extensively in colonial otariids and appears to be characteristic of this family. Although less numerous, empirical studies of phocid species have revealed a range of recognition abilities. This study investigated whether Weddell seal (Leptonychotes weddellii) females produce individually distinct 'pup contact' calls that function during natural pair reunions. Fifteen calls from each of nine females recorded in the Vestfold Hills, Antarctica were analysed. One temporal, nine fundamental frequency and five spectral characteristics were measured. Results of the cross-validated Discriminant Function Analysis revealed that mothers produce individually distinct calls with 56% of calls assigned to the correct individual. The probability of achieving this level of discrimination on novel data by chance alone is highly improbable. Analysis of eight mother-pup reunions recorded near McMurdo Sound, Antarctica further demonstrated that these 'pup contact' calls function during natural pair reunions. Behavioural analysis also revealed that pups were chiefly responsible for establishing and maintaining close contact throughout the reunion process. Our study therefore demonstrates that Weddell seal females produce calls with sufficient stereotypy to allow pups to identify them during pair reunions, providing evidence of a functioning mother-pup vocal recognition system. Column A - Row 1: Gives the tag number of the female. - Rows 3-33: The list of acoustic measurements recorded from the spectrograms. - Rows 3-5: Temporal measurements recorded in milliseconds. - Rows 7-12: Frequency measurements recorded from the fundamental frequency. Rows 9-11 were measured at the 1/4, 2/4 and 3/4 temporal positions along the fundamental frequency respectively. - Rows 13-17: Give the number of the frequency band with the most energy at the temporal positions stated (i.e. fundamental frequency band=1, first harmonic=2 etc). - Rows 19-29: List the fundamental frequency measurements, taken at the temporal positions stated, used to calculate Mean frequency (Row 31) and the Coefficient of Frequency Modulation (Row 33) using the formula listed in the publication. - Rows 35 and 36: List the cursor error margins of the acoustic analysis program I used. Columns B-P - Give details of the above mentioned acoustic characteristics for 15 replicate calls from each of the 9 females sampled.

Database Description The files represent the 41 different Weddell seal (Leptonychotes weddellii) call types identified at either Mawson, Davis, and/or Casey. They were collected between 60 degrees 49' E and 110o 40' E in longitude, and between 66 degrees 12' S and 68 degrees 34' S in latitude. Each call type name includes two elements. The first is a three-digit number starting at 301 to identify the call type. The second is a one to three-letter code referring to the call category that each type falls into. The 13 different possible call categories are: SymbolNameDescription OToneConstant-frequency, predominantly sinusoidal call. LGrowlConstant-frequency, broad bandwidth, long call. QWhoopConstant-frequency call with a terminal upsweep. SSqueakBrief call with constant frequency or rising frequency and an irregular waveform. WAWhistle AscendingAscending frequency, sinusoidal waveform. TCTrill Constant-FrequencyNarrow bandwidth trill with a constant-frequency beginning, sinusoidal or frequency-modulated waveform. TTrillNarrow to broad bandwidth, containing a frequency downsweep, greater than 2 seconds. WDWhistle DescendingDescending frequency, sinusoidal waveform (less than 2 seconds). MMewAbruptly descending frequency followed by a long constant-frequency ending. CChugAbruptly descending frequency followed by a brief constant-frequency ending. GGuttural Glug (Grunt)Descending-frequency call that was lower than a Chug and had a brief duration. WAGWhistle Ascending - GruntBrief Ascending Whistle followed by a Guttural Glug (Grunt), the two types alternate in a regular pattern. KKnockAbrupt, brief-duration broadband sound (from: Pahl, B.C., Terhune, J.M. and Burton, H.R. 1997). The 41 call types were divided into two sections, the first 33 (301-O to 333-K) being common call types and the last 8 (334-Q to 341-WD) being rare call types. In each call type folder, one to five different samples of each call type are provided. They are identified by a small case letter added at the end of the call type name. Each sample includes both a .WAV audio sample and a .JPG image of the call type spectrogram showing call shape, i.e., changes in call frequency (vertical) over time (horizontal). These call types were used to identify: (a) unique call types or call categories, (b) differences in call type or call category usage (the frequency of occurrence of each call type or category), and (c) differences in call features (number of elements, start frequency, frequency shift and first element duration) among the three stations. The download file also includes a spreadsheet of data and a text file explaining how to interpret the data. Analysis of this dataset is ongoing.

From the abstract of the attached paper: Underwater calling behaviour between breathing bouts of a single adult male Weddell seal (Leptonychotes weddellii) was examined with respect to call type and timing late in the breeding season at Davis Station, Antarctica. Underwater calls and breathing sounds were recorded on 1 and 8 December 1997. Thirty-seven sequences of calls prior to surfacing to breathe and 36 post-submerging sets of calls were analysed with respect to probability of call type occurrence and timing. Dives were 461 plus or minus 259 seconds (mean plus or minus standard deviation). The seal called every 29.7 plus or minus 56.2 seconds throughout a dive. The first call after submerging was usually (n = 29 of 36) a low frequency (less than 0.8 kHz) growl. Three patterns of three- to five-call type sequences were made following 28 of 36 breathing bouts. Call type patterns after submerging exhibited fewer different sequences than those before surfacing (chi-squared = 61.42, DF = 4, p less than 0.000001). The call usage patterns before surfacing were diverse and did not indicate when the seal was going to surface, a time when he would be vulnerable to attack from below. Our findings suggest the hypotheses that territorial male Weddell seals call throughout each dive and use stereotyped call patterns to identify themselves while vocally asserting dominance. This work was completed as part of ASAC project 2122 (ASAC_2122). The fields in this dataset are: Tape number Sequence per tape Sequence entire data Call types Count since last breath Last breathing bout number Count prior to next breath Time in tape (seconds) End time of last breath Start time of next breath Time since dive The 'sequence' relates to the sequence of call types that are given between the end of the last breath of a breathing bout and the beginning of the first breath the next time the seal surfaces to breathe. Essentially the report relates to the stereotyped nature of the call types, especially just after the dominant male dives after finishing breathing. Each time the animal surfaced, that was identified as a breathing bout. They are numbered sequentially. At the very start of the data set the seal had to surface before the breathing bout could be counted (as number 1). This procedure enabled us to identify the order and timing of the calls that occurred immediately before and immediately after each breathing bout. Thus, the 'count prior to the next breath' gives the order of the calls before the seal surfaced to breathe again (third last, second last, last,). The call types were analysed with respect to the following pattern: third last, second last, last, breathing bout, first, second, third, etc. to third last, second last, last, next breathing bout.

Many vocalisations produced by Weddell seals (Leptonychotes weddellii) are made up of repeated individual distinct sounds (elements). Patterning of multiple element calls was examined during the breeding season at Casey and Davis, Antarctica. Element and interval durations were measured from 405 calls all greater than 3 elements in length. The duration of the calls (22 plus or minus 16.6s) did not seem to vary with an increasing number of elements (F4.404 = 1.83, p = 0.122) because element and interval durations decreased as the number of elements within a call increased. Underwater vocalisations showed seven distinct timing patterns of increasing, decreasing, or constant element and interval durations throughout the calls. One call type occurred with six rhythm patterns, although the majority exhibited only two rhythms. Some call types also displayed steady frequency changes as they progressed. Weddell seal multiple element calls are rhythmically repeated and thus the durations of the elements and intervals within a call occur in a regular manner. Rhythmical repetition used during vocal communication likely enhances the probability of a call being detected and has important implications for the extent to which the seals can successfully transmit information over long distances and during times of high level background noise. See other metadata records and datasets associated with ASAC project 2122 (ASAC_2122) for further information. The fields in this dataset are: Tape/Site/File Filename Call Type Total Number of Elements Attribute Frequency Time Casey Davis

Possible communication between territorial male Weddell seals (Leptonychotes weddellii) under the ice with females on the ice was investigated. In-air and underwater recordings of underwater calls were made at three locations near Davis, Antarctica. Most underwater calls were not detectable in air, often because of wind noise. In-air call amplitudes of detectable calls ranged from 32-74 dB re. 20 microPa at 86 Hz down to 4-38 dB re. 20 microPa at 3.6 kHz. Most of these would be audible to humans. Only 26 of 582 amplitude measurements (from 230 calls) ranged from 5 dB to a maximum of 15 dB above the minimum harbour-seal (Phoca vitulina) in-air detection threshold. Seals on the ice could likely hear a few very loud underwater calls but only if the caller was nearby and there were no wind noises. The low detectability of underwater calls in air likely precludes effective communication between underwater seals and those on the ice. See other metadata records and datasets associated with ASAC project 2122 (ASAC_2122) for further information. The fields in this dataset are: Column A: G = grunt, T = trill, CT = constant freq. trill, O = tone, C = chug, AW = ascending whistle, DW = descending whistle, L = growl, R - roar Column B: frequency (Hz) Column C: underwater call level NOTE dB re 20 uPa Column D: in-air call level dB re 20 uPa Column E: in-air background noise level at this frequency dB re 20 uPa Column F: water - air difference dB Column G: location, 1-3, see paper for code Column H: seal in-air threshold dB re 20 uPa Column I: human in-air threshold dB re 20 uPa Column J: seal in-air threshold at this frequency dB re 20 uPa

Underwater vocalisations of Weddell seals were recorded at Casey (1997) and Davis (1992 and 1997) Antarctica. The goal of the study was to determine if it would be possible to identify geographic variations between the Casey and Davis seals using easily measured, narrow bandwidth calls (and not broadband or very short duration calls). Two observers measured the starting and ending frequency (Hz), duration (msec) and number of elements (discrete sounds) of four categories of calls; long duration trills, shorter descending frequency whistles, ascending frequency whistles and constant frequency mews. The statistical analyses considered all calls per base, single and multiple element calls, and individual call types. Except for trills, discriminant function analysis indicated less variation between the call attributes from Davis in 1992 and 1997 than between either of the Davis data sets and Casey 1997. The data set contains measures from 2966 calls; approximately 1000 calls per base and year. Up to 100 consecutive calls were measured from each recording location per day of recording so the data set indicates the relative occurrence of each of the call types per base and year. There were very few ascending whistles at Casey. All of the trills and mews contained a single element. This data set was published in Bioacoustics 11: 211-222. The fields in this dataset are: Observer Station Location Time Call Number Call Type Frequency Duration Elements Overlap In 2011, another download file was added to this record, providing recording locations made during the project in 2010. Furthermore: In 1997 Daniela Simon made some opportunistic recordings for the project near Casey. The recording locations were: Berkley Island 110 38'E, 66 12' 40"S Herring Island 110 40'E, 66 25'S O'Brien Bay 110 31'E, 66 18' 30"S Eyres Bay 110 32'E, 66 29" 20"S The Davis sites: IN 1990 THERE WAS ONLY ONE RECORDING SITE - 78 12.5' E, 68 31.6' S IN 1997 RECORDINGS WERE MADE AT THE FOLLOWING SITES EAST SIDE OF WEDDELL ARM - 78 07.55' E 68 32.17' S PARTIZAN ISLAND - 78 13.66' E 68 29.57' S LONG FJORD - 78 18.95' E 68 30.24' S TOPOGRAV ISLAND - 78 12.40' E 68 29.33'S OFFSHORE - 77 58.73'E 68 26.35'S TRYNE BAY - 78 26.25'E 68 24.87'S LUCAS ISLAND - 77 57.00'E 68 30.36'S WYATT EARP ISLANDS - 78 31.51'E 68 21.31'S ================================================================================ The attached document is "a listing of the Weddell seal breeding locations near Mawson where Patrick Abgrall in 2000 and Phil Rouget in 2002 made underwater recordings". The sound recording effort in 2000 was not as high as it was in 2002, hence fewer locations are listed. The Abgrall sites are referred to in the paper 'Variation of Weddell seal underwater vocalizations over mesogeographic ranges' that Abgrall, Terhune Burton co-authored, published in Aquatic mammals in 2003. This paper also refers to the Casey and Davis sites above. The Rouget sites relate to the metadata record 'Weddell Seal underwater calling rates during the winter and spring near Mawson Station, Antarctica' Entry ID: ASAC_1132-1 In general the seals can create breathing holes in areas where tide cracks form, namely close to grounded icebergs, the shoreline and islands. I doubt that they could/would create breathing holes through solid 2 m ice.

Publication of these results is currently in progress with the Journal of Animal Ecology. Summary 1.An efficient method of describing change in Antarctic marine ecosystems is long-term monitoring of land-breeding marine predators. High-level predators are used to index the state of environment on the notion that perturbations in the ecosystem will affect their diet, reproductive performance and other demographics. For this purpose, Weddell seals breeding at the Vestfold Hills have been marked and re-sighted for the past 28 years (1973 - 2000). 2.Successful reproduction requires considerable energetic resources. The difference between rates of conception and rates of parturition suggests pregnant females abort reproductive attempts when their energy stores are low. In this way, annual rates of reproduction (i.e. parturition) are a measure of foraging efficiency. 3.Previous attempts to estimate Weddell seal reproduction have been biased by different rates of re-sighting breeding and non-breeding females. We used multistate mark and re-sight models to account for this and other variables when estimating reproductive rate. 4.The amplitude of temporal variation was much greater for reproduction than for survivorship, indicating that parous (breeding) females maximised survival by reproducing less. This strategy could be successful in fluctuating environments because seals live longer and experience more reproductive occasions. 5.The population had low reproductive rates from 1983 to 1985 and throughout the 1990s. In those years, potential recruitment into breeding groups was reduced to 50 - 60 % of the cohort before viable pups were even born. 6.Even in years of low reproductive rate, typically half (52%) of the breeding females produced pups. It seemed that individuals differed in their foraging success and thus body condition and / or their functional response to this. 7.There was no evidence for costs of reproduction. We infer that the seals responded to environmental conditions prior to parturition, as opposed to proceeding with reproduction when inadequately resourced and depleting energy resources such that they had lower probability of surviving or reproducing the following year. 8.Synthesis and applications: This study demonstrates a method of estimating reproductive rate that overcomes bias inherent in traditional methods. Estimated in this way, we propose that reproductive rate is the best indicator of the state of marine ecosystems that can be indexed for Weddell seals. The fields in this dataset are: Year Standard Error Upper confidence interval Lower confidence interval Breeding probability Upper error bar Lower error bar

Underwater recordings of vocalisations of Weddell seals were obtained at 8 locations within the Vestfold Hills (7) and Larsemann Hills (1). The recordings were made near groups of seals on the ice during the mid to late part of the breeding season. Recordings were obtained using a variety of hydrophones and both Sony Digital Audio Tape (130 during 1992 season) and standard analogue cassette (60 during 1991 season) formats. Over 11,000 vocalizations were analyzed. The calls were classified into 12 major call types (Pahl et al. 1997 Australian Journal of Zoology 45:171-187). The underwater repertoire is different than that of the seals at McMurdo Sound or the Palmer Penninsula (Thomas et al. 1988 Hydrobiologica 165:279-284). The Weddell seals at the Vestfold Hills do not exhibit the between-fjord vocal differences reported by Morrice et al. (1994 Polar Biology 14:441-446). The relative usage of each call type did not vary between the earlier and later recordings (Pahl et al. 1996 Australian Journal of Zoology 44:75-79). The recordings are currently being used to support other studies on Weddell seal vocalizations. Legend for ASAC_556.csv - csv text format. The following legend describes the 39 variables in this file. The codes for some of the variables are presented in the 1997 publication: Pahl, B.C., Terhune, J.M., and Burton, H.R. 1997. Repertoire and geographic variation in underwater vocalisations of Weddell seals (Leptonychotes weddellii, Pinnipedia: Phocidae) at the Vestfold Hills, Antarctica. Australian Journal of Zoology 45: 171-187. The fields in this dataset are: VariableSubject or code 1LOCATION; recording location; see AJZ article, Figure 1 2DATE; reference day, (date of day 1 has been lost) 3YEAR; 1 = 1991, 2 = 1992 4CASSETTE; cassette number, identifies individual recordings 5CALNO; call number, case numbers of each call, sequential 6CTYPE; call type, provisional call type, subjective initial classification (see below) 7NOELM; number of elements (discrete sounds) in the call 8EL_NO; element within that call relating to next 12 variables, for variable 8, only data from the first element is used 9WVFRM; waveform of element, see AJZ article for codes 10CLSHP; call shape, see AJZ article, Figure 2 for codes 11E_D; duration of the first element (seconds) 12IND1; duration of the interval between the end of the first element and the start of the second element (seconds) 13CALLD; total duration of the call (all elements; seconds) 14INCD; duration between sequential calls (seconds) 15O_LAP; overlap, is call overlapped by another call? 0 = no, 1 = yes 16S2STM; unknown measure 17SFREQ; frequency at start of first element (Hz) 18EFREQ; frequency at end of first element (Hz) 19HFREQ; highest frequency of first element (Hz) 20LFREQ; lowest frequency of first element (Hz) 21E_NO; element number, half way through the call. Data for the next 9 variables relate to this element, applies only to multiple element calls 22CLSHP; call shape of the middle element, same code as variable 10 23WVFRM: waveform of the middle element, same code as variable 9 24E_D; duration of the middle element (seconds) 25IND1; duration of the inter-element interval before the middle element 26IND2; duration of the inter-element interval after the middle element 27SFREQ; frequency at start of the middle element (Hz) 28EFREQ; frequency at end of middle element (Hz) 29HFREQ; highest frequency of middle element (Hz) 30LFREQ; lowest frequency of middle element (Hz) 31E_NO; element number of the last element of the call. Data for the next 8 variables relate to this element, applies only to multiple element calls 32CLSHP; call shape of the last element, same code as variable 10 33WVFRM: waveform of the last element, same code as variable 9 34E_D; duration of the last element (seconds) 35IND2; duration of the inter-element interval before the last element 36SFREQ; frequency at start of the last element (Hz) 37EFREQ; frequency at end of last element (Hz) 38HFREQ; highest frequency of last element (Hz) 39LFREQ; lowest frequency of last element (Hz) Codes for call types (variable 6). The provisional call types were amalgamated into 50 call types that were arbitrarily numbered from 201 to 250. These were subsequently classified into 13 broad categories (Pahl et al. 1997). The amalgamation of the provisional call types of variable 6 into the 50 call types presented in Pahl et al. (1997) is as follows: Call TypeProvisional Call Types (variable 6) 2011 7, 24, 36, 72, 31, 40, 73, 77, 107, 110, 31, 136 2023, 46, 54, 128, 33, 13, 140, 10, 25, 9, 139, 88, 46, 27, 126, 67, 91, 27, 126, 135 20359 204113 20514, 48, 69, 64, 49, 19, 92, 43, 75, 127, 99 206122, 124 2072, 41, 58, 93 20847, 138 20962, 132 210102 211115 21221, 23, 45, 35 21368, 80, 84 214114 2154 216118 21752, 78 2185, 6, 11 219104 22017, 22, 65, 97, 32, 26 22128 22283, 100, 101, 111, 105 22329, 30, 42, 51, 44, 94, 95 22487 22512 22682 2278 22818, 20, 57, 108 229109, 119 23034, 70, 130, 53, 121 23163 23298, 120 23389 23490 23556, 117 23671, 106 23785 238103 23974 24096 24176, 123, 133 24281, 86 24315 244112 24538 24679 24739, 127, 129, 55, 60 24816, 37, 50 249116 25066 For additional information or clarification, please contact Dr. J. Terhune, Dept. of Biology, University of New Brunswick, P.O. Box 5050, Saint John, NB, Canada E2L 4L5, terhune@unbsj.ca or +1 506 648 5633. See the link below for public details on this project.