Largest baleen whale mass mortality during strong El Niño event is likely related to harmful toxic algal bloom

Field surveys

The vessel-based HF24 scuba diving expedition, from Apr 15 to May 8, 2015, aimed to inventory the benthic fauna of the area between Golfo Tres Montes (Northern Golfo de Penas, 46°30′W) and Puerto Edén (49°S). By chance, VH and her team discovered recently dead baleen whales and skeletal remains in and close to the entrance of the 14 km long Estero Slight and in the Canal Castillo situated 235 km to the south (Figs. 1 and 3; Table 2). Georeferences and photographs of different views were taken, whales measured, and species and sex identified whenever possible. Between May 25 and 31, the Chilean Fisheries Service (SERNAPESCA), with the support of the Chilean Navy (Armada) and the Criminal Investigation Department of the Civil Police (PDI), organized a vessel-based trip to the location of the dead whales in Estero Slight to investigate possible anthropogenic reasons behind the mortality. During this trip, genetic samples for species identification were taken, one ear bone was extracted and stomach and intestine contents of two whales were tested for presence of PST and AST (Fiscalía de Aysén, 2015). During a subsequent aerial survey, on-board a high wing airplane Cessna 206, between Jun 23 and 27, 2015, three of us (CG, VH and FH) surveyed the coasts along the shores of Golfo de Penas. This aerial survey covered the coastal area between the Jungfrauen Islands (48°S) and Seno Newman (46°39′S) from altitudes between 100 and 850 m and at speeds between 100 and 200 km/h (Figs. 1 and 4; Table 2). Due to limited flying time (unstable weather conditions and the inability to refuel in the area), data collection was focused on counting whale carcasses, recording GPS positions and taking photographs. A GoPro camera filmed continuously until reaching Seno Newman. The researchers on the flight counted carcasses and marked their coordinates while an audio recorder captured the carcass number, position, orientation, photo number, photographer and geomorphology of the beach. Whale counts were repeated in all areas except Seno Newman due to adverse weather conditions. Since there are no landing opportunities in this remote and unpopulated area, it was not possible to take samples or close-up photos, or to search for additional whale bones.

In addition to the whale carcasses and skeletons from the two surveys, some whale carcasses and skulls were reported between Feb and Jun 2015 by boat crews navigating the west coast of Taitao Peninsula and the coast between 49°15′ and 51°S (Table 2). Between Jan 23 and Mar 1, 2016 (Expedition Huinay Fiordos 27) and between Apr 27 and May 30, 2016 (Expedition Huinay Fiordos 29), two additional vessel-based expeditions were carried out, each to Seno Escondido, Seno Newman and Estero Slight, with the aim of searching for new carcasses, taking samples for genetic and red tide analyses, and performing oceanographic transects. Data from those surveys are included here, but most of the analyses of the samples will be published in a separate paper.

Samples of marine invertebrates were collected under permit of Subsecretaria de Pesca y Acuicultura (R.EX. 1295 del 27.04.2016). Samples of cetacean carcasses were authorized by SERNAPESCA, Region de Aysen (Acta Numbers 2016-11-10 and 12).

Satellite image

A high-resolution satellite image was taken of Seno Newman on Aug 13, 2015 using the Pleiades-1 Satellite. The 16-bit ortho-rectified GeoTIFF multispectral (R-G-B-NIR) and Panchromatic files have been analyzed to count whale carcasses and determine their geographic positions (Fig. 5). The whales identified in the satellite image were compared to the photos and GPS locations obtained during the overflight, and cross-matched with reference to nearby geomorphological features.

Taxonomic analysis

Whales were identified in situ during the vessel-based expedition based on morphological characteristics. The species identification of the specimens from which tissue was sampled during the SERNAPESCA expedition to Estero Slight was confirmed genetically (Fiscalía de Aysén, 2015). A 675 bp fragment of mitochondrial DNA control region was amplified using the primers using the primers M13 Dlp1.5 5′-TGTAAAACGACAGCCAGTTCACCCAAAGCTGRARTTCTA-3′ and 8G 5′GGAGTACTATGTCCTGTAACCA (Dalebout et al., 2005) and sequenced in both directions. Amplification reactions were performed in a total volume of 25 μl with 5 μl PCR buffer 10×, 2 μl MgCl₂ 50 mm, 1 μl of each primer, 2 μl dNTP 200 mm and 0.3 μl Taq DNA polymerase (Invitrogen Life Technologies, Carlsbad, CA, USA) and 50 ng DNA. The PCR temperature profile was as follows: a preliminary denaturing period of 2 min at 94 °C followed by 30 cycles of denaturation for 30 s at 94 °C, primer annealing for 40 s at 56 °C and polymerase extension for 40 s at 72 °C. A final extension period for 10 min at 72 °C was included.

Taphonomy

Analysis was carried out, following biostratinomic criteria, on different subsets of the whale remains recorded during the overflight and the vessel-based surveys. Characterization of the depositional state of the carcasses was based on a post hoc analysis of the assemblage, exclusively through photographs, classifying the carcasses into three taphonomic classes according to previous studies of biostratinomic processes in marine mammals (Pyenson et al., 2014, Liebig, Taylor & Flessa, 2003; Liebig, Flessa & Taylor, 2007; Schäfer, 1972). The aspects considered were anatomic position of the carcasses (ventral, dorsal or lateral side-up, n = 201), deposition site (rocky or sandy, n = 295), and the disarticulation and degree of decay of the carcasses. These final two aspects were sorted into classes to estimate the sequence of disarticulation/decay addressing two aspects: time since death (n = 245) and drift time/distance of the carcass (as a proxy to estimate the relative location of death, n = 151).

To assess the time since death, three categories were defined, reflecting a straightforward order from the least decomposed to the most disarticulated carcass/skeleton. “Class 1” refers to carcasses in the lowest to relatively medium state of decomposition for these assemblages. Included in this category are complete carcasses with skin, complete carcasses without skin, and complete carcasses with partially exposed bones (see Fig. 6A). “Class 2” includes carcasses in a relatively greater state of decomposition but still maintaining their longitudinal axis, although some bones may be scattered (see Fig. 6B). Finally, “Class 3” refers to isolated skeletal remains with no soft tissue, such as skulls, dentaries or postcranial remains (see Fig. 6C). Thus, the sequence of “time since death” should reflect ranges from less than three months (Class 1), several months, but probably less than six months (Class 2), to a year or more (Class 3).

The analysis of the location of death, namely whether the carcasses are para-autochthonous or allochthonous was addressed by evaluation of the time that the carcasses had remained floating in the water column and at the surface (see Schäfer, 1972). For this, we defined two classes, depending of the presence or absence of the skull, as a proxy for the time floating and the potential distance between the site of mortality and the observed site of deposition (Fig. 7) (Toots, 1965; Voorhies, 1969; Behrensmeyer, 1973; Holz & Simões, 2002; Liebig, Taylor & Flessa, 2003; Simões & Holz, 2004). Thus, “Class A” includes carcasses that preserve the skull and “Class B” includes those without a skull. For this analysis, we excluded skeletons, which were considered older than a year (minimum age, based on field observations of AVT from 2016 expedition to the site of the mortality).

A geomorphological analysis was made using photographs and Google Earth (Terrametrics, 2015). We classified the type of depositional locality (i.e., sand/pebble dominated beach or rocky outcrop) (Table 2) in order to assess the relationship between these aspects and the taphonomic categories mentioned above; for instance, whether carcasses that had been transported further and disarticulated (allochthonous) were more prevalent at high energy sites (i.e., rocky outcrops) and articulated (para-autochthonous) carcasses more prevalent in low energy environments (i.e., sandy beaches).

To compare the density of the death assemblages at Golfo de Penas with known extinct and extant death assemblages recorded in the literature, we measured linear dimensions of the geomorphological units (i.e., length and width of the beach), through the measure tool in Google Earth, using the highest resolution satellite images available, at sites where assemblages were found. In this manner, the geographic areas corresponding to the death assemblages were calculated and the density determined by dividing the number of specimens in each assemblage by its area.

Analysis of the petrotympanic complex (ear bone)

We studied the bones of the middle and inner ear of one whale, collected during the SERNAPESCA expedition. A volumetric computed tomography in the Morita tomography (box of 60 mm, 500 cuts) was carried out. The images were visualized with Osirix Dicom viewer v 5.6 32-bit in search for fractures or micro-fractures, which would appear as black gaps in the bony tissue.

Analysis for toxins (PST/AST)

Bivalve tissue was sampled in Estero Slight on Apr 22 and on May 25, 2015 (two samples in total), and in Estero Slight, Seno Newman and Seno Escondido between Jan 23 and Mar 1, and Apr 27 and May 30, 2016 (22 samples in total). The stomach content and intestine content of two whales from Estero Slight were sampled on May 25, 2015. On Feb 2016, one sample of duodenum content was obtained from a freshly dead whale observed in Estero Slight. At the same period, one sample of surface-swimming Munida spp. was collected at 46°29.730′S, 74°55.722′W. All samples were analyzed in situ for presence of PST using the protocol already described for the shellfish tissue and stomach content samples. The tissue was homogenized using a blender and mixed in a 1:1 ratio with a field extraction fluid composed of 2.5 parts of rubbing alcohol (70%) to one part white vinegar. The mixture was then homogenized manually and filtered through a paper filter (paper filter #4). The extract obtained after filtration was then used to detect the presence of toxins through rapid field test kits from scotia rapid testing for PST and AST. For this, 100 μl of the extract was placed in a test tube containing running buffer, mixed and then 100 μl of this mixture was placed in a lateral flow enzyme-linked immunosorbent assay (ELISA) test strip with antibodies specific for PST (saxitoxin and its derivative toxins) and AST (domoic acid). These tests were left to develop for 1 h before the results were read.

Twenty-two phytoplankton samples were collected in Estero Slight, Seno Newman and Seno Escondido between Jan 23 and Mar 1, and Apr 27 and May 30, 2016, using a 20 μm mesh size plankton net in a vertical tow from 15 m depth. The phytoplankton present in these samples was concentrated using the net, and a 100 μl subsample was placed in a tube with 0.1M acetic acid and mixed. About 100 μl of this mixture were then added to a test tube-containing running buffer and an aliquot of this mixture of the same volume was placed in an ELISA test strip for PST and left to develop for 1 h before results were read.

These qualitative PST test strips are extremely sensitive due to the local toxin profile, which is high in GTX2/3, resulting in detection limits below 32 μg STX Eq/100 g of tissue. The detection limit for the AST tests was reduced to 2 ppm of domoic acid by modifying the standard sample preparation protocol by eliminating the dilution of the sample before mixing it with the buffer.

A graphical analysis of the geographic and temporal distribution of PSP events, presence of harmful microalgae and environmental variables in the affected region (43°S–51°S) from 2007 to Jul 2015 and from Mar 2016 was performed with the data obtained from the red tide monitoring program conducted by the SERNAPESCA (R.S. Galdames, 2015, personal communication), in which mytilid samples are analyzed at several stations throughout Chilean Patagonia approximately once a month by the “Laboratorios SEREMI Salud,” from Aysén and Magallanes regions at Southern Chile.

Drift model

Floating objects are directly affected by surface currents, wind and waves. Wind both drives the Ekman drift of surface water (Ardhuin et al., 2009) and exerts a direct drag on the emerged surface of an object (Breivik et al., 2012). Stokes drift, the net forward transport due to non-closed particle trajectories resulting from passing waves, also contributes to the transport of floating objects. The drift of whale carcasses was simulated by parameterizing the contribution of these components, based on objects of a similar size from search and rescue models (Breivik et al., 2012; Peltier et al., 2012). Due to the large uncertainty in carcass drift characteristics, parameters were varied stochastically within a wide range of possible values.

Use was made of existing current and wave products, the HYCOM daily 1/12 simulation (Wallcraft, Metzger & Carroll, 2009), and waves from ECMWF ERA-Interim reanalysis (Dee et al., 2011). Winds were taken from a custom downscaling of NCEP NFL boundary conditions using the WRF model (Skamarock & Klemp, 2008) to a sub-4 km grid size. Drift scenarios were run by stepping forward in time from hypothetical sites and times of mortality. All of these sites were in shallow water, since carcasses resulting from mortality in deep water have a tendency to sink and not resurface (Smith et al., 2015). A horizontal diffusion coefficient of 10 m²s⁻¹ was included in drift tracks to represent unresolved physical processes. While the resolution of the current and wave datasets is inadequate to represent detailed coastline or seabed geometry, or the interior of the fjords, the drift model does clarify the expected distribution and spread of carcasses from localized sources.

Large-scale wind stress

The large-scale tendency toward upwelling or downwelling provides a key driver of coastal ecosystems. This was assessed using ECMWF ERA-Interim reanalysis data (Dee et al., 2011). It is the alongshore component of wind stress that drives Ekman transport normal to the coast and consequent upwelling or downwelling. Since upwelling and downwelling are cumulative processes, a time-integrated wind stress was calculated (Pierce et al., 2006) from a base time of the vernal equinox (September 21). Stress was estimated from reanalysis winds at 10 m elevation according to Large & Pond (1981). The large-scale change in coastal orientation was taken into account in extracting the alongshore wind component, although localized inlets, bays (including the Golfo de Penas) and islands were not considered.

Field surveys and toxicity tests

Of the total of dead whales observed in all expeditions and reports in 2015 (367), 35 recently dead whales and 12 skeletal remains were discovered during the HF24 expedition: 31 carcasses and 12 skeletal remains were found in and close to the entrance of the 14 km long Estero Slight and four carcasses in Canal Castillo, situated 235 km to the south, as well as many whale bones on different beaches (Fig. 3; Table 2). Three hundred and five carcasses were mapped during the overflight between the Jungfrauen Islands (∼48°S) and Seno Newman (46°39′S). In addition to this total of 284 whale carcasses and 21 skeletons from the two surveys, 51 whale carcasses and 11 whale skulls were reported between Feb and Jun 2015 by boat crews navigating the west coast of Taitao Peninsula and the coast between 49°15′ and 51°S (Table 2; Fig. 4).

On some photos what could have been carcasses of smaller animals (possibly dolphins and/or sea lions) were seen, but due to the flying altitude, speed and weather conditions, the photo quality and resolution did not allow their conclusive identification as actual carcasses. In Estero Slight, one dead pinniped was found on the shore from the vessel. During the SERNAPESCA expedition, one Otariidae skull was found and photographed in the same channel but the correspondence of the carcass and the skull could not be established.

The 28 whale carcasses that could be identified unambiguously to species level were all sei whales (Balaenoptera borealis); 15 of these identifications were confirmed genetically. Seven specimens could be identified as males and ten as females. One hundred and twenty-nine carcasses were identified as baleen whales of the Balaenopteridae family or rorquals. The 30 whales examined in detail in Estero Slight during the vessel-based expedition were between 6 and 15 m long, hence included both juvenile and fully grown specimens.

None of the examined whales showed any evidence of disease or traumatic damage. The anatomic structures of the ear bone were in good condition showing no damage; the stapes were articulated in place, and the bony tissue showed no fractures (Fig. 8). The analysis of locally collected mytilids in Apr and May 2015 and of the stomach and intestine content of two whales in May 2015 showed presence of PST and AST.

In 2016, 16 fresh carcasses were observed during the HF27 and HF29 vessel-based expeditions to Golfo Tres Montes; five further were reported by boat crews navigating the Southern part of Chilean Patagonia. None of the examined whales showed any evidence of disease or traumatic damage. Thirty-six rapid tests on PST were run using mussels (12 tests), Munida (two tests), and phytoplankton (22 tests) in Seno Escondido, Seno Newman and Estero Slight. Most of the samples collected during the 2016 expeditions proved to be negative for the presence of PST, nevertheless, both expeditions detected the presence of PSP in the phytoplankton collected at the entrance of Seno Newman. A sample collected at the head of Seno Newman was negative for PST, indicating that the toxic phytoplankton was preferentially located at the mouth of this inlet and nearby areas of the Canal Chaicayán.

Biostratinomic analysis

Of the 367 dead whales observed in 2015, 305 carcasses were mapped between Seno Newman (46°39′S) and Jungfrauen Islands (∼48°S). Those carcasses could be grouped into five assemblages (Figs. 1 and 9; Table 2), defined as a group of carcasses in close proximity. The assemblages were called Golfo de Penas, Jungfrauen Islands, Seno Escondido, Seno Newman and Estero Slight.

Some carcasses were floating (11), but most (284) were deposited ashore (Figs. 3–5). The greater proportion of carcasses were deposited in a lateral position and to a lesser extent in the ventral-up position reflecting the hydrodynamics of the body in the sea as determined by the inflation of the abdominal region and mainly of their tongues, as observed in a recently dead individual and in some decayed carcasses at Golfo de Penas (Fig. 10). In general, they were tide-oriented (parallel to the coastline) and all of the classified carcasses from the overflight were lying on their back or side (ventral-up, 44.3%; lateral-up, 55.7%) (Table 3; Fig. 11C), while only one specimen (from HF24) was found in a dorsal-up position (data not included in analysis due to different time of observation).

With respect to the classification of “time since death,” 68.8% of the carcasses were classified in Class 1 (less than three months), 24.9% in Class 2 (less than six months) and 6.3% in Class 3 (more than a year) (Figs. 11A and 11B; Table 4). With respect to “time at sea,” 147 (87%) of the carcasses were classified in Class A (short time/distance of drift), while only four (13%) were identified as Class B (long time/distance of drift) (Fig. 11C; Table 4). There was no pattern relating the geomorphological unit (sandy: 34%, pebble: 27%, rocky beach: 34%) to the taphonomic classes.

The carcasses found in April in Estero Slight were classified in stage 2 of Geraci & Lounsbury (2005) indicating a few days to weeks since death; this would be classified as Class 1 in the taphonomic classes of the present study.

The density of whale carcasses was in average 1,050/km², considering all assemblages recognized (Table 5).

Carcass drift and potential source locations

The distribution of beached carcasses was simulated from four illustrative source locations (Figs. 12A–12D). In each case, calculations tracked 13,000 hypothetical carcasses, reflecting source times spanning a two-month period from mid-February to mid-April 2015 and a range of drift model parameters. The spread of stranding locations therefore represents variability of the current, wind and wave environment during this period as well as the uncertainty in model parameters and a diffusive component to the drift tracks. While each of the illustrated source locations leads to strandings distributed over several hundred kilometers of coastline, there are important differences in these distributions. A simulated source in Golfo Tres Montes (Northern Golfo de Penas) leads to strandings throughout the Golfo de Penas (Fig. 12A), including in the Golfo Tres Montes itself. No other source location (Figs. 12B–12D) leads to strandings in Golfo Tres Montes due to the direction of prevailing currents and the sheltering effect of Peninsula Taitao. Similarly, only a source to the north of Peninsula Taitao leads to strandings in that region (Fig. 12B). Carcasses originating in the Golfo de Penas have a tendency to be transported to the south by prevailing currents (Figs. 12A, 12C and 12D).

Inter-annual variation in upwelling or downwelling

Comparison between the cumulative alongshore wind stress for the year in question and the previous 20 years (Fig. 13) reveals that the months immediately prior to the mortality event were anomalous. North of the study area, at 45°S, there was an anomalously strong tendency toward upwelling (an upward trend in Fig. 13), making this one of the most upwelled years of the period. At the latitude of Golfo de Penas and further south there was a net tendency to downwelling (a downward trend in Fig. 13), but punctuated by upwelling events, making this one of the least downwelled years of the period.